Rumen Ecology 
Research Planning
Proceedings of a Workshop held at ILRI
Addis Ababa, Ethiopia
13?18 March 1995
Edited by
R.J. Wallace
A. Lahlou-Kassi
THE INTERNATIONAL LIVESTOCK RESEARCH INSTITUTE
BOX 30709, NAIROBI, KENYA ? BOX 5689, ADDIS ABABA, ETHIOPIA
Production Editor: Peter Werehire
This publication was typeset on a microcomputer and the final pages produced on a laser printer
at ILRI, P.O. Box 30709, Nairobi, Kenya, telephone 254 (2) 630743, fax 254 (2) 631499, e-mail
ILRI-KENYA@CGNET.COM. Printed by ILRI, P.O. Box 5689, Addis Ababa, Ethiopia.
Copyright ? 1995 by the International Livestock Research Institute
ISBN 92-9146-005-2
Correct citation: Wallace R.J. and Lahlou-Kassi A. 1995. Rumen Ecology Research Planning.
Proceedings of a Workshop Held at ILRI, Addis Ababa, Ethiopia, 13?18 March
1995. ILRI (International Livestock Research Institute), Nairobi, Kenya. 270
pp.
The International Livestock Research Institute (ILRI) began operations on 1
January 1995. The institute incorporates the resources, facilitites and major
programmes of two former CGIAR centres founded about two decades
ago?the International Laboratory for Research on Animal Diseases, in
Nairobi, Kenya, and the International Livestock Centre for Africa, in Addis
Ababa. The research and outreach programmes of ILRAD and ILCA have
been consolidated, streamlined and reoriented to support an expanded
mandate. ILRI will conduct strategic research in the biological, animal and
social sciences to improve livestock productivity in sustainable agricultural
systems throughout the developing world.
The objectives of the new institute?s research programme are to improve
animal health, nutrition and productivity (milk, meat, traction) in ways that
are sustainable over the long term, to characterize and conserve the genetic
diversity of indigenous tropical forage species and livestock breeds, to
promote sound and equitable national policies for animal agriculture and
natural resource management, and to strengthen the animal husbandry
research programmes of developing countries.
Contents
Preface 
Openingaddress
Improvingfibreutilisationandproteinsupplyinanimals
fedpoorqualityroughages:ILRInutritionresearchandplans 
P.O.Osuji,S.Fern?ndez-RiveraandA.Odenyo
Potentialapplicationofrumenecologymanipulation
toanimalnutritionindevelopingcountries
Nitrogenmetabolismintherumen:
biotechnologicalproblemsandprospects 
R.I.MackieandM.Morrison
AntimicrobialfactorsinAfricanmultipurposetrees
S.M.ElHassan,A.Lahlou-Kassi,C.J.Newbold andR.J.Wallace
AustralianMeatResearchCorporation?srumenbacteria
manipulationkeyprogramme:rationaleandprogress1990?1995 
I.D.Johnsson
Stateofartinmanipulationoftherumen
microbialecosystem
Ruminantsandrumenmicroorganismsintropicalcountries 
H.Kudo,S.Imai,S.Jalaludin,K.FukutaandK.-J.Cheng
Animportantroleforruminalanaerobicfungiinthevoluntary
intakeofpoorqualityforagesbyruminants 
G.L.R.Gordon,C.S.McSweeneyandM.W.Phillips
Rumenfungiindomesticandwildherbivores
J.Kopecn?
Factorslimitingproliferationofdesirablegroupsofbacteria
intherumenofanimalsfedpoorqualityfeedsofhighfibrecontent 
N.O.vanGylswyk
Exploitationofrumenmicrobialenzymestobenefit
ruminantandnon-ruminantanimalproduction 
L.B.Selinger,K.-J.Cheng,L.J.Yanke,H.D.Bae,T.A.McAllisterandC.W.Forsberg
Clostridiumparaputrificumvar.Ruminantium:colonisationand
degradationofshrimpcarapacesinvitroobservedbyscanning
electronmicroscopy 
M.A.CobosandM.T.Yokoyama
Comparativeabilitiesoffeedintakeandmicrobialdigestion
intheforestomachsofcamelidsandruminants 
J.P.Jouany,J.P.DulphyandC.Kayouli
Optimisingrumenenvironmentforcellulosedigestion
E.R.?rskov
Foragelegumesasproteinsupplementstopoorqualitydiets
inthesemi-aridtropics 
J.H.Topps
Stateofartinbioengineeringofrumenmicroorganisms
Applicationofmolecularbiology
intheproductionoffoodfromruminants 
R.M.Cook
Engineeringtherumenforenhancedanimalproductivity 
J.D.Brooker
Geneticmanipulationofrumenbacteria:nowareality
K.Gregg,D.Schafer,C.CooperandG.Allen
Workinggroupreports,conclusionsandrecommendations
Termsofreference 
ReportofWorkingGroup1:Potentialresearchagenda
forILRIonrumenmanipulation:potential
collaboratorsandmodeofcollaboration 
ReportofWorkingGroup2:Improvedefficiency
ofuseoffeedresourcesbytropicallivestock
Conclusionsandrecommendations
Listofparticipants 
iv Preface
Preface
There is increasing evidence that manipulation of microbiological activities in the
rumen and bioengineering of rumen organisms could lead to the development of
technologies to improve animal productivity from the available feed resources.
The merging of the International Livestock Centre for Africa (ILCA) and the Inter-
national Laboratory for Research on Animal Diseases (ILRAD) expertise in animal
nutrition, physiology and animal biotechnology open to the International Livestock
Research Institute (ILRI) a new perspective in developing a research agenda on rumen
ecology.
However, there is a need to assess the possible uses of available technology, identify
applications of direct relevance to ILRI research, define areas where ILRI would have
a comparative advantage, and identify partners for collaborative research projects.
The  planning workshop is proposed to bring together experts from the rapidly
advancing field of rumen microbiology to assist ILRI staff to identify these opportunities
and set priorities.
Workshop objectives
? Identify and prioritise areas of rumen ecology which are promising for their potential
impact on improving nutrition status of tropical ruminants.
? Develop a rumen ecology research programme for ILRI based on relevance to
developing countries and ILRI?s comparative advantage vis-a-vis other institutions.
? Identify potential collaborators in advanced research institutes and define mode(s) of
collaboration.
Workshop organising committee
A. Lahlou-Kassi (Coordinator)
P. Fajersson
K. Getahun
Opening address
H.A. Fitzhugh
Director General, International Livestock Research Institute
On behalf of the Institute, Dr Lahlou-Kassi, who organised this workshop, and all ILRI
staff, I welcome you who have come such long distances to help us develop a plan for
research in rumen ecology.
I would like to start by introducing you to this new international research institute:
the International Livestock Research Institute. The buildings may not be new, but that
is because this new institute is something of a chimera, being built on the foundation of
two international livestock research institutes.
The International Laboratory for Research on Animal Diseases was established in
1973 with a global mandate to develop effective control measures for livestock diseases
that seriously limit world food production. ILRAD?s research programme focused on
animal trypanosomiasis and tick-borne diseases, particularly theileriosis (East Coast
fever).
The International Livestock Centre for Africa (ILCA) was established in 1974 with
a mandate to focus on livestock production systems in Africa. ILCA followed an
interdisciplinary approach combining social and environmental sciences with the ani-
mal sciences. The founders of ILCA believed that technologies to improve traditional
animal production systems were available from the laboratories and research stations
of Australia, Japan, North America and Europe and that an understanding of those
livestock production systems and their constraints would reveal how modern technolo-
gies might be best fitted to meet the needs for improving livestock production in the
developing world. However, those traditional farming systems were evolving very
rapidly, especially during the 1980s, along with social, political and cultural changes
throughout Africa. Technologies available off the shelf from stations and laboratories
elsewhere were often not suitable, and were sometimes inappropriate, for livestock
producers in sub-Saharan Africa. The special characteristics of livestock production in
tropical environments provided new challenges to research.
ILRAD and ILCA were supported by a consortium of donors known as the Consult-
ative Group for International Agricultural Research, or the CGIAR. This Group was
established in the early 1970s. The number of CGIAR institutes increased through the
1970s and more were added in the early 1990s. A few years ago, this consortium, which
had increased in funding from 20 or so million dollars to about 300 million dollars, went
through a crisis of confidence that all of us involved in research felt. Many donors lacked
confidence that research was the way toward solving important development environ-
mental  problems.  The CGIAR,  therefore,  began to reassess whether international
agricultural research was worth supporting. Fortunately the answer was ?yes?. In 1995
we believe that we have regained the confidence of the donors. But during this time of
reassessment some changes were made within the system. One was the establishment
of the International Livestock  Research  Institute to combine animal health and
production research.
Both ILRAD and ILCA were focusing on African problems. Studies conducted by
the donor community that supported these international research institutes indicate that
there is a strong demand and need for livestock research in Asia and, to a somewhat
lesser extent, in Latin America, West Asia and North Africa or the Mediterranean region.
One priority for livestock research in Asia are the major economic developmental and
environmental problems that have arisen due to intensification of agriculture.
Because of the location of our campuses and our access to the problems of developing
countries, ILRI is able to serve as a transfer agent, not only by transferring technologies
but also by providing access and  contacts  between the laboratories of  developed
countries, which are substantially better endowed, and the laboratories and research
centres of developing countries. Perhaps more important is the research-based informa-
tion and knowledge that are being transferred and the increasing opportunities to apply
this knowledge and experience to the problems of developing countries.
International research centres make up only a small component in the whole spectrum
of research on problems of agriculture in developing countries. The largest players are
the national research institutes in the developing countries the world over. Many of these
institutes are underfunded but their scientists are increasingly well trained. Through
collaborative research, we can employ the large scientific capacity in the national
institutes.
Most research at ILRI is done in collaboration with other scientists and institutes in
both developing and developed countries. We plan to expand our collaborative activities
and that is one of the reasons why we brought you here, so that we can become better
acquainted. This will help us to identify useful collaborations. This institute has a very
broad set of programme activities ranging from work in the laboratories in molecular
biology to work in the field. We are able to cover this very broad base of research
activities only because of our ability to work and collaborate with a host of other
scientists and institutions.
I remind you that the appropriate type of research for an international institute lies at
the upper end of the applied spectrum rather than that of adapting technologies to meet
the needs of farmers and their livestock. Our research must be development- oriented.
We must be clear as to how our research will serve development needs. Ninety-eight
percent of our funding comes from the development vote. And so even though our
research is fundamental in character, we must be able to articulate a clear use of the
research products to further development.
One of our research areas is utilisation of feed resources. Any analysis of the
constraints to livestock production in developing countries consistently identifies feed
resources as a primary constraint. The primary constraint to livestock production may
be the production of feed resources or their utilisation, depending on the specific
production system and the particular environment.
The goal of the feed utilisation programme is to improve efficiency of use of feed
resources by tropical ruminants. We look to our sister centres to focus their work on
production which leaves us to focus on utilisation. It is usually the intersection between
production and utilisation that is the major problem. Forage production, for example,
will often be facilitated by genes that promote within the plant characteristics that
viii H.A. Fitzhugh
protect it from pests or predators. Sometimes selecting for rapid growth rate will
generate qualities that are anti-nutritional in character. There is, therefore, need to match
what is good for production of the plant with what is good for utilisation of the products
of plant growth.
There were opportunities for real synergies in the types of research conducted in
Nairobi and Addis Ababa. ILCA was carrying on some work in the area of rumen
ecology while ILRAD was carrying on work at the microbial and molecular level. So
we felt that there would be good opportunities for synergisms by putting these two
institutions together and we quite deliberately delayed having this workshop until we
could have scientists from both former ILCA and former ILRAD working together.
The objectives of this workshop are very straightforward. We will look for a very
specific recommendation coming out of this workshop as to whether or not ILRI as an
international institute has a comparative advantage in conducting research in this broad
field  of rumen ecology. If at the  end of the  week you decide ILRI does have  a
comparative advantage for this type of research, we will ask that you be specific as to
exactly where this comparative advantage lies. While doing this, we ask you not to think
of ILRI as just a new institute, but as a set of resources and opportunities that can be
developed as part of a much larger global network of institutes, many of which you
yourselves represent.
Opening address ix
Improving fibre utilisation and protein
supply in animals fed poor quality roughages:
ILRI nutrition research and plans
P. O . O s u j i
1
, S. Fern?ndez-Rivera
2
and A. Odenyo
1
1. International Livestock Research Institute, P.O. Box 5689, Addis Ababa, Ethiopia
2. International Livestock Research Institute, ICRISAT Sahelian Centre, BP 12404, Niamey, Niger
Abstract
Inadequate nutrition is the main cause of low productivity of livestock in
sub-Saharan Africa. The primary feed resources include fodder trees, natural
pastures and crop residues. This paper presents results from studies conducted
at ILRI in evaluations of these resources. The effect of supplementing teff straw
with lablab or cowpea in cattle on microbial protein supply was evaluated using
urinary purine derivatives. There was no effect of type of supplement but the
level of supplementation significantly improved microbial protein supply.
Similarly, supplementing maize stover with oilseed cake increased total micro-
bial protein flow in sheep. Teff straw supplemented with Sesbania sesban
increased dry matter degradation and liquid passage rates. Degradation of
Sesbania, Lucaena, Chamaecytisus and Vernonia foliage was also evaluated.
Sesbania was degraded rapidly and Vernonia slowly. The effects of anti-nutri-
ent extracts from Vernonia amygdalina, Chamaecytisus palmensis, Sesbania
sesban and Acacia angustissima were tested on pure cultures of cellulolytic
rumen microbes and A. angustissima prolonged the lag phase. A. angustissima
killed sheep and Tephrosia caused massive rumen stasis. Studies conducted in
semi-arid West Africa indicated that the microbial digestive activity, as meas-
ured by the disappearance in sacco of a standard forage, varies both seasonally
and across animal species. Genetic variation in feeding value of crop residues,
forage legumes and fodder trees were assessed. The variation among varieties
of sorghum and pearl millet were relatively small and inconsistent across years.
Relatively larger differences were observed in forage legumes. Preliminary
results suggest possibilities for identifying geographical areas that produce
better quality fodder trees. Future work should include the biochemical basis
of the interactions between rumen microbes and the chemicals contained in
fodder trees, and the seasonal variations of microbial populations in various
ruminants across agro-ecological zones.
Introduction
Inadequate year round nutrition is a major smallholder production constraint. Ruminants
owned by smallholders in all ILRI mandate regions will continue to subsist, for the
foreseeable future, on unimproved native pastures and crop residues. These poor quality
roughages are bulky, high in fibre, poorly degraded in the rumen, low in nitrogen and
minerals (Table 1) resulting in very low intakes. Population pressure and urbanisation,
particularly in Africa, will limit the quantity of grain available for animal feeding.
However, these deficiencies could be corrected by  the addition of fodder  trees,
herbaceous legumes or multipurpose trees (MPTs) (Tables 2 and 3).
Several factors affect the utilisation of poor quality roughages by ruminants. These
include rumen environment (where conditions should be pH > 6.2, NH
3
> 3.5 mmol/l),
microbial adhesion, particle size reduction, passage rates of both particulate and liquid
digesta, roughage degradation rate and volatile fatty acid production, adequate supply
of iso-acids for microbial protein production and the availability of by-pass protein.
ILRI nutrition research has focused on how to enhance fibre degradation and microbial
protein supply by rumen manipulation aimed at optimising the above conditions. The
use of MPTs as a means to correct nutrient imbalances and improve the conditions in
the rumen has been a priority area of research (ILCA 1995; Osuji 1994). This paper
describes some of the nutrition research work done at ILRI. First, results obtained in
Table 1. Nitrogen (N) and neutral detergent fibre (NDF) concentrations of low quality roughages used in
ILRI research.
g/kg dry matter (DM)
Roughage N NDF
Maize stover 5.4 755
Teff straw 4.9-7 798
Oat straw 3.8 727
Oat hay 7.9 815
Barley straw 5.4 829
Debre Zeit hay 12.0 722
Pea straw 7.5 768
Sorghum stover 8.4 773
Sululta hay 9.4 738
Cynodon hay 11.8 737
Wheat straw 4.6 772
Napier grass 4.8 720
Pearl millet 8-11 704
Table 2. Characteristics of MPTs: Degradation rates of N for the fresh leaves (FL) and dry leaves (DL) of
two accessions of Sesbania sesban and Acacia siberiana pods.
Sesbania Readily soluble N degradation
ILCA Acc. No. fraction (g/kg) rate (/h)
10865 DL (dry leaves) 298 0.065
10865 FL (fresh leaves) 468 0.110
15021 DL 314 0.075
A. siberiana (pods) 433 0.055
2 P.O. Osuji, S. Fern?ndez-Rivera and A. Odenyo
the use of fodder or multipurpose trees (MPTs) and herbaceous legumes in animal feeds
are presented. Research aimed at identifying and overcoming constraints to rumen
function in grazing animals is discussed. The selection of crops and trees for improved
feeding value is addressed. Finally, research plans on rumen ecology are presented.
Supplementation with MPTs
and herbaceous legumes
Ruminants depend on microbial protein production to meet their N requirements. In
ruminant production systems where poor quality roughages constitute the principal
source of feed there are two major objectives aimed at optimising their use. There is
need to enhance the fermentation of the roughage to ensure adequate energy supply,
mainly as VFA. Secondly, given the scarcity of protein in such systems, there is a need
to maximise the supply  of microbial protein. Several strategies are available to
manipulate the rumen. The strategy of choice depends on the desired output. One
strategy that has often been used, particularly when poor quality forages are fed, is
supplementation. The principal objective of supplementation is to increase the supply
of nutrients, mainly energy and protein, such as to create favourable conditions in the
rumen which result in better fermentation and microbial protein supply (Figure 1).
Supplementation with MPTs can do this in several ways?MPTs can increase the
amount of energy supplied by the basal feed by:
1. Alleviating a deficiency hampering microbial fermentation of the basal feed e.g. N
or S.
Table 3. Chemical characteristics of MTPs.
Low High
(g/kg DM)
Nitrogen Dichrostachys cinerea (13.9) Leucaena (40.5)
Fibre (NDF) Sesbania (206) D. cinerea (498)
NDF-N Sesbania (2.4) Tagasaste (9.2)
Soluble tannins Carissa edulis (21.5) A. siberiana (327)
Condensed tannins (absorbance
units/g NDF at 500nm) Sesbania (13) Tagasaste (56)
Phosphorus Vernonia (0.7) Sesbania (2.7)
Calcium Tagasaste (10.5) Leucaena (20.0)
Sulfur Tagasaste (1.5) Leucaena (2.3)
(mg/kg)
Iron Sesbania (360) Tagasaste (520)
Manganese Leucaena (66) Tagasaste (200)
Copper Leucaena (13) Vernonia (20)
Zinc Leucaena (19) Tagasaste (39)
NDF: Neutral detergent fibre.
NDF-N: Nitrogen bound to NDF.
Improving fibre utilisation and protein supply in animals fed poor quality roughages 3
Figure 1. Effect of rate of degradation of browses on (A) rate of passage of particulate matter, (B) efficiency
of microbial N supply and (C) substitution rate.
0.041 0.076
(A)
0
0.01
0.02
0.03
0.04
Passage
rate
(/h)
0.041 0.076
(B)
0
5
10
15
20
Microbial
ef
ficiency
(g/kg
DOM)
0.041 0.076
(C)
0
0.05
0.1
0.15
0.2
0.25
0.3
Substituti
on
rate
(g/g)
4 P.O. Osuji, S. Fern?ndez-Rivera and A. Odenyo
2. Improving the rumen environment (e.g. pH, rumen NH
3
or rumen-degradable
protein) to ensure increased fermentation of the basal roughage diet. For example,
increased numbers of cellulolytic bacteria will increase the invasion of and adhe-
sion onto the fibrous feed.
3. Improving the rate and extent of particle size reduction and thus increasing the
passage rates of both liquid and particulate matter, leading to increased feed intake.
4. Supplements such as MPTs are themselves sources of energy. In fact some MPTs
are fed at certain times of the year as the sole feed.
Thus MPTs, as supplements, can increase protein supply to the host animal by
increasing the supply of both degradable and undegradable protein, and by creating a
favourable rumen environment resulting in enhanced fermentation of the basal roughage
and thus increased microbial protein synthesis.
Nutrition research at ILRI has involved in vivo, in sacco and in vitro studies. In vivo
work has looked into the effects of various factors on microbial protein supply. In one trial
(Abule 1994), supplementation of teff straw with graded levels of cowpea or lablab
significantly increased microbial N supply in calves. The type of supplement had no
significant effect on microbial N supply but the level of inclusion did (Table 4). When
sheep were given a basal diet of maize stover supplemented with oilseed cakes, total
microbial protein flow into the intestine was significantly affected (P < 0.01) and was
increased further by maize supplementation (Osuji et al 1993). When poor quality
roughages were supplemented with MPTs, the rate of degradation of the basal feed, the
fractional outflow rate of liquid matter from the rumen and the efficiency of microbial
protein supply were increased (Figure 2). Rapidly degraded MPTs like Sesbania increased
passage rate and the efficiency of microbial protein supply and are better supplements.
The feeding of rapidly degraded MTPs has resulted in reduced substitution of the basal
roughage diet (Bonsi et al 1995).
In another set of trials sheep were given a basal diet of teff straw supplemented with
0, 175, 245 or 315 g of Sesbania to create different rumen environments. Dry matter
degradation and liquid passage rates were estimated. Figure 2 shows clearly that both
dry matter degradation and liquid passage rates increased with increasing level of
Sesbania up to 245 g/head per day and then declined. Improved or faster liquid outflow
rate may result in increased intake of the basal roughage as new material is eaten to
replace what has left the rumen (Bonsi et al 1995).
Table 4. Microbial N supply of calves fed a basal diet of teff straw supplemented with lablab hay
?
.
Inclusion rate Microbial N
(g/kg LW) (g/day)
0 4.6
5 12.4
10 20.4
15 24.9
?
No differences were observed between cowpea and lablab.
Improving fibre utilisation and protein supply in animals fed poor quality roughages 5
Effect of plant parts
With most MPTs there is the choice of feeding the leaves alone or the leaves plus the
fruits or the fruits alone. Sesbania sesban 10865 leaves were compared with Acacia
albida pods. The rate of degradation of fresh leaves was much higher than that for the
albida pods. Bulkiness of the MPT seems to be an important additional factor affecting
the intake of fibrous feeds (Osuji 1994).
Degradation properties of MPTs and interaction with rumen
microbes
In sacco degradation techniques (?rskov and MacDonald 1979) have been used at ILRI
to study factors which affect the interaction of MPTs with rumen microbes. The results
of these studies indicate that several factors affect the interaction of the MPT with the
rumen microbes. The type of MPT, its form and the quantity fed affect roughage
utilisation. When MPTs of the genera Sesbania, Leucaena, Chamaecytisus and Ve r no ni a
were evaluated using the in sacco method, the rates of both DM and N degradation
varied significantly among the four MPTs (Figure 3). Sesbania was degraded most
rapidly while Ve r no ni a was degraded relatively more slowly. It has been found that the
rate of degradation of MPTs varies with the basal diet (Figure 4). Results from several
trials, where different MPTs were incubated in the rumens of sheep, showed that the
form of the MPT affected the rate of DM degradation. Fresh forms of the MPTs were
degraded much faster than the dry forms (Figure 5). Both the fast-degrading and the
slow-degrading MPTs could be harnessed to nutritional advantage (Bonsi et al 1995;
Nsahlai et al 1994; Siaw et al 1993; Umunna et al 1995).
0 175 245 315
Levels of Sesbania (g/d)
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
Degradation rate
Liquid passage rate
Figure 2. Effect of Sesbania supplementation on the degradation of teff straw and on liquid passage rate.
6 P.O. Osuji, S. Fern?ndez-Rivera and A. Odenyo
Sesbania Leucaena Tagasaste Vernonia
Sun dried MPTs
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
Rate
of
N
d
egradation
Figure 3. Effect of the type of MPT (sun dried) on rate of nitrogen degradation.
123
Basal diet
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
Rate
of degradation
DM N Gas production
1 = Teff straw alone
2 = Teff straw + 180 g Sesbania
3 = Teff straw + 180 g Leucaena
Figure 4. Effect of different basal diets on rate of degradation of dry matter and nitrogen, and gas
production (per hour) of four MPTs (Sesbania, Leucaena, Tagasaste and Vernonia).
Improving fibre utilisation and protein supply in animals fed poor quality roughages 7
Figure 5. Effect of different basal diets on rate of dry matter and nitrogen degradation.
Fresh
leaves
Sun dried
leaves
Form
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
Rate
of
DM
degradation
Fresh
leaves
Sun dried
leaves
Form
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
Rate
of
N degradation
8 P.O. Osuji, S. Fern?ndez-Rivera and A. Odenyo
Effect of quality of basal roughage
For roughages of different N content supplemented with MPTs, the degradation rate of
the unsupplemented basal roughage increased with the N content and the effect of the
MPT supplement on roughage degradation rate was most marked in the basal roughage
with the lowest N content. This clearly demonstrates that smallholders who depend on
poor  quality  roughages (native pastures and crop residues) will benefit most  by
supplementing their animals with fodder trees. Therefore a farmer with limited MPT
supplement would be better off feeding it to animals on maize stover or teff straw rather
than to animals fed better quality Napier grass.
Anti-nutritional factors
MPTs could also contribute to creating an unfavourable rumen environment, e.g. by
lowering rumen pH which would result in a reduction in the number of rumen fibrolytic
organisms and thus reduce cellulolysis (Table 5).
There is a wide range of anti-nutritional factors found in MPTs (Table 5) though the
phytochemistry  and mode of  action  of  these anti-nutritional  factors are not fully
understood. There are indications, however, that some of them have defaunation
qualities, some are bactericidal, others bacteriostatic. Furthermore, poisons from MPTs
can act either on the rumen microbes or on the host animal itself. This makes the
evaluation of MPTs rather difficult. Figure 6 demonstrates the effect of anti-nutritional
factors on dry matter intake. A. angustissima is very high in proanthocyanidins com-
pared to S. sesban 10865. Dry matter intake was depressed to such an extent in sheep
fed A. angustissima that many of the animals died (Kurdi 1994).
Attempts have been made to harness anti-nutritional factors (e.g. tannins) to nutri-
tional advantage. In this regard, oilseed cakes were fed in association with tanniferous
browses (1994). Only minor depressions in the rate of fermentation of oilseed cake
were observed. The changes due to this associative feeding of oilseed cakes and browses
on growth rate were equally small. Attempts to match roughages with protein supple-
ments have not demonstrated any benefits that could be attributed to the interaction
between the tannin content of sorghum stovers and the oilseed cakes (Osafo 1993). What
is clear is that cottonseed cake promoted better gains than noug cake.
The consumption of tannins caused a shift in the paths of nitrogen excretion from the
urine to faeces. In addition, there was an increase in the insoluble fraction of faecal
nitrogen. These effects of tannins  can  lead  to decreased volatilisation  losses  and
mineralisation rates, which may have important effects in the process of nutrient cycling
and crop production (Powell et al 1994).
Effects of MPTs on microbes
The effects of MPTs on rumen microbes have been studied at ILRI using the gas
production technique (Menke et al 1979). The effects of MPTs on mixed rumen microbes
seem to be different from their effects on pure cultures (El Hassan 1994). In vitro trials
suggested that S. sesban reduced the number of protozoa, suggesting that S. sesban
Improving fibre utilisation and protein supply in animals fed poor quality roughages 9
10865 may have defaunating properties. However this was not confirmed in an in vivo
trial where S. sesban 10865 was compared to Medicago sativa (El Hassan 1994). No
particular efforts were made in this trial to isolate the animals used as the numbers of
rumen microbes were similar for both treatments. Cross contamination might have taken
place.
Mueller-Harvey et al (1988) found that extracts of Ethiopian browses increased the
lag time and reduced the growth rate of Streptococcus bovis. They found that extracts
from Acacia nilotica were particularly detrimental to rumen bacteria. Alternatively, the
rumen population may have adapted to metabolise the anti-protozoal agent.
The effects of extracts of Vernonia amygdalina, C. palmensis, S. sesban 10865 and
A. angustissima on the growth of pure cultures of Fibrobacter succinogenes, Rumino-
coccus flavefaciens and R. albus were investigated (El Hassan 1994). Ve r no ni a
amygdalina, C. palmensis and S. sesbania 10865 extracts affected the growth of only
R. albus by prolonging the lag phase. On the other hand A. angustissima at lower
Table 5. Toxic substances found in some MPTs
Toxic Part
MPT component of plant Concentration References
Sesbania Tannins Foliage 0.62% Reed (1986)
National Academy
of Sciences (1979)
Saponins Foliage ? Kinghorn
Alkaloids Foliage ? and Smolenski
Amines Foliage ? (1981)
Acacia Fluoroacetic acid Foliage/pods ? Topps (1992)
Tannins Foliage/pods 3.12% Ahn et al
Cyanide Foliage/pods (1989)
Oxalate Foliage/pods ? Goodchild
Saponins Foliage/pods ? and McMeniman (1987)
Gliricidia Tannins Foliage 2.05% Ahn et al (1989)
Dicoumarol Leaves ? ILCA (1991)
Albizia Dicoumarol Leaves 2.25% ILCA (1991)
Calliandra Dicoumarol Leaves 7.91 ILCA (1991)
Tagasaste Tannins Foliage ? Borens and
Alkaloids Foliage ? Poppi (1990)
Leucaena Mimosine Leaf 1?12% Tangendjaja
Seed 3.3?14.5 et al (1990)
Tannins Leaf 3?14% D?Mello and
Seed 7.1% Fraser (1981)
3-Hydroxy-4(1H)- Leaf 5.1?8.2% D?Mello and
pyridone (DHP) Seed ND Taplin (1978)
Trypsin inhibitors Leaf ? Acamovic and
Seed ? D?Mello (1984)
Saponins Leaf 2?11% Tangendjaja
Seed 2?11% and Lowry (1984)
Flavonols Leaf 3?6% Lowry
Seed 2?11% et al (1984)
10 P.O. Osuji, S. Fern?ndez-Rivera and A. Odenyo
inclusion rates greatly reduced the growth of Ruminococcus species and slowed the
growth of F. succinogenes. Extracts of L. leucocephala prolongedthelagphaseof
cellulolytic bacteria. Ruminococcus flavefaciens and R. albus are very important for
fibre digestion.
Figure 6. Effect of tannin content on total DM intake.
1000
A.angustissima S.sesban
0
200
400
600
800
T
o
tal
D
M
i
ntake
(
g/d)
A.angustissima S.sesban
0
5
10
15
20
25
30
35
T
annins
Soluble phenolics
(% DM)
Proanthocyanidins
(A500/g NDF)
Improving fibre utilisation and protein supply in animals fed poor quality roughages 11
Using the gas production technique, it was found that gas production rates decreased
substantially when some MPTs were  incubated with rumen fluid in vitro. Acacia
angustissima suppressed fermentation and the extent of suppression depended on the
amount of MPT included (El Hassan 1994) (Figure 7). The reverse was true for leucaena.
These effects were illustrated in vivo when animals fed A. angustissima died and those
fed Tephrosia had massive rumen stasis.
More work on the utilisation of fodder trees as livestock feed is needed. This should
aim at refining rapid methods of evaluation, expansion of the work on (1) anti-nutrition
factors to other compounds other than tannins and (2) the effects of various fodder trees
on specific rumen microbes. These would add to the tools needed to use fodder trees
effectively in rumen manipulation aimed at enhancing ruminant production.
Seasonal variation in the rumen environment
under traditional feeding systems
Fibre digestion and microbial protein synthesis in the rumen are greatly affected by the
conditions that prevail in the rumen. In general, when low quality forages are consumed,
the conditions needed for efficient digestion of fibrous feeds and microbial growth are
not met. In addition, the ingestion of feeds containing anti-nutritional  factors is
Figure 7. Gas production (ml) in vitro from different inclusion levels of Acacia angustissima with teff straw.
0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96
Time (h)
0
10
20
30
40
50
60
Gas
production (ml/200
mg
DM)
0% 15% 30% 60%
12 P.O. Osuji, S. Fern?ndez-Rivera and A. Odenyo
associated with decreased microbial activity. Studies on the environmental conditions
in the rumen in traditional feeding systems can help identify the nature of the constraints
to  rumen function. Their alleviation can result  in substantial improvement in the
efficiency of  use of the feed  resources available.  If nutrient imbalances  are also
overcome, the economic response to supplementation can be improved.
Studies on environmental conditions in the rumen of grazing animals are under way
at the zonal programmes of ILRI in Niger and Nigeria. In these studies, the nylon bag
technique is used to evaluate the quality of the ruminal conditions as judged by the
disappearance rate of a standard forage. Some results obtained with cattle, sheep and
goats in Niger (Fern?ndez-Rivera 1994) are presented here. Six intact females, four
ruminally fistulated males and four oesophageally fistulated males of each species were
used in the study. All animals grazed pearl millet residue fields (December 1992 to
March 1993) and Sahelian rangelands (March 1993 to December 1993). A standard
forage (pearl millet stover leaves) was incubated monthly in the rumens of the fistulated
animals to determine the rate and extent of disappearance of organic matter (OMD).
Rumen fluid samples were taken monthly (mean of am and pm values) from the intact
animals and analysed for NH
3
-N. Extrusa samples were collected monthly from the
oesophageally fistulated animals and analysed for crude protein (CP) and OMD.
All variables were influenced by species and month of measurement. During most of
the year, sheep and goats selected a diet with similar CP, but the NH
3
-N concentration
in the ruminal fluid was lower in goats than in sheep, which reflects differences in
solubility of the nitrogen consumed by these species (Figure 8). Cattle selected a diet
lower in CP and had a lower concentration of NH
3
-N in the rumen than small ruminants.
The concentration of NH
3
-N in the rumen fluid of cattle was less than 70 mg/l for most
of the dry season, whereas in small ruminants NH
3
-N did not appear to limit rumen
function at any of the sampling times. In spite of the lower concentration of NH
3
-N in
the rumen fluid of cattle, the potentially digestible insoluble fraction of the standard
forage was higher and less variable than in small ruminants (Figure 9). Large seasonal
variations in the magnitude of the potentially digestible insoluble fraction were ob-
served. The rate of OM digestion and the lag time were affected by month of measure-
ment but not by species (P = 0.08). The lag time was longer in the dry season, whereas
no clear trends were observed for the monthly variation in digestion rate. The reduction
in potentially degradable fraction (B) with advancing dry season probably reflects
increasing cell wall and thus decreased soluble carbohydrates. However, it is not clear
to what extent the reduction in the B fraction may have been compensated by the
increased rate of degradation of this fraction resulting from increased N supply from
browses available at this time.
These results suggest that the environmental conditions and the digestive microbial
activity in the rumen of cattle, goats and sheep grazing year-round on crop residue fields
and Sahelian rangelands vary seasonally. This variation appears to be associated with
the dietary selectivity of the three species. The potentially digestible fraction is generally
thought to depend on the nature of the feed. However, these results suggest that
anti-nutritional factors may decrease the potential extent of digestion of fibrous feeds.
Since the different animal species differ in dietary selectivity, the constraints to rumen
Improving fibre utilisation and protein supply in animals fed poor quality roughages 13
function appear to be different across species. The variation in the ruminal environment
observed under traditional feeding regimes in the tropics must be considered when
attempting to manipulate the rumen microbial population. Special consideration should
be given to the influence of toxic substances and the competition among microbes or
their ability to tolerate periods of starvation.
These results also have implications for defining supplementation priorities. For
instance, during crop residue grazing, high levels of intake and diet quality were
observed only during the first three or four weeks of the season, when the amount of
above-ground leaf exceeded 400 kg/ha (ILCA 1993). After the first month of grazing
the amounts of protein- and energy-yielding nutrient appears to be insufficient for
efficient microbial activity and the maintenance of the animals. Therefore a response to
rumen degradable (RDP) and undegradable  (RUDP) protein and energy could be
expected. In a supplementation experiment, the response to metabolisable energy (ME)
followed a diminishing return pattern and decreased as the initial live weight of the
animals increased. The sheep grew faster when supplemented with RUDP and this
response rose as ME intake  increased, which suggests that  the microbial  protein
Figure 8. Ammonia concentration in rumen fluid of cattle, goats and sheep grazing crop residues and
rangeland in Niger.
DJ FMAMJJASON
40
80
120
160
200
240
280
NH
3
-N,
mg/l
Sheep
Goats
Cattle
CROP
RESIDUES
RANGELAND
14 P.O. Osuji, S. Fern?ndez-Rivera and A. Odenyo
synthesis was not sufficient to meet sheep protein requirements. Similarly, the influence
of ME was also dependent on the consumption of RUDP. Further studies should
determine the economically optimum levels of supplementation.
Improving fibre digestion through
genetic selection of crops and trees
Fibre digestion in the rumen is constrained by retention time, digestion rate and the
magnitude of the indigestible fraction (Allen and Mertens 1988; ?rskov 1991). These
constraints could be alleviated by increasing rumen volume, improving the environ-
mental conditions in the rumen or increasing the fibrolytic capacity of the rumen
microbes. The constraints to fibre digestion in the rumen are determined by the
chemical and physical nature of the fibre, which in turn depends on the genetic make-up
of the plant and the environment where it grows. Therefore the selection of cultivars
that produce high levels of grain and better quality residues offer possibilities for
improving fibre breakdown in the rumen (Bartle and Klopfenstein 1988; ?rskov 1991).
Three different avenues for improving fibre digestion through plant selection have
been followed by ILCA/ILRI, in collaboration with crop-oriented institutions. The first
is the exploitation of variation across varieties in cereal crops (sorghum and millet) and
forage legumes (cowpea and groundnut). The second is the introduction of genetically
controlled quality related traits, as the low-lignin brown mid-rib (bmr) and trichomeless
Figure 9. Degradation (b fraction) in sacco of a standard forage incubated in the rumen of cattle, goats
and sheep grazing crop residues and rangeland in Niger.
DJFMAMJJASON
350
400
450
500
550
B
FRACTION, g OM/kg
OM
Cattle
Sheep
Goats
CROP
RESIDUES
RANGELAND
Improving fibre utilisation and protein supply in animals fed poor quality roughages 15
(tr) traits in pearl millet. The third is the identification of geographic locations in the
semi-arid zone that could produce higher quality trees. In the last approach, studies are
under way in Niger to evaluate the variation in feed potential of several species.
Influence of genotype on feeding value of cereal crops
Bird-resistant and non-bird-resistant varieties of sorghum grown in Ethiopia did not differ
in neutral detergent fibre (NDF) concentration of leaves (Reed et al 1987, 1988). However,
leaves from non-bird-resistant varieties were more digestible. Most of the variation in leaf
NDF digestibility was accounted for by differences in lignin or insoluble proanthocy-
anidins. Further studies by Osafo (1993) showed that varieties differed only in the
potentially degradable fraction (?rskov 1991) of leaves but not in the readily soluble
fraction. Since no differences were observed in other plant parts, the degradation rates and
the potential degradation of the whole stover were similar across varieties.
In 12 millet varieties grown in Niger, Reed et al (1988) found differences in NDF
and lignin of leaf and stem, although the range in variation appeared to be smaller than
the inter-varietal differences that they had observed for sorghum. However, these
differences do not appear to be consistent over the years. In a three-year collaborative
study with the International Crop Research Institute for the Semi-arid Tropics
(ICRISAT), the forage quality of the stover from nine pearl millet varieties grown under
three phosphorus fertiliser levels (0, 30, 60 kg P
2
O
5
/ha) was evaluated. In the three
years of the study, the millet varieties differed (P < 0.01) in stover yield. These
differences were not affected by phosphorus fertiliser and were consistent across years.
The concentration of NDF in leaf and stem was affected by variety in two of the three
years, whereas digestibility of leaf was influenced by variety (P < 0.01) in one year
and that of stem (P < 0.01) in two years. The annual means of OM digestibility of leaf
were 442, 447 and 490 g/kg and those of stem were only 235, 227 and 217 g/kg.
Digestibility was lowest in the highest grain yielding year, but no correlation between
grain yield and digestibility was detected within years. This study indicated that
differences among varieties are large for residue yield, but small for quality traits.
Grain yield is highly positively correlated with stover yield but appears not to be
correlated with digestibility or NDF of the stover.
In another collaborative study with ICRISAT, the effects of the brown mid-rib (bmr)
and trichomeless (tr) traits on the quality of millet stover were studied (ILCA 1994a).
At grain harvest, stover leaves and stems from 120 progenies of either bmr or normal
millet and 20 progenies of either tr or normal millets were collected and sequentially
analysed for ashless NDF, ADF and lignin, as well as for organic matter (OM) and in
sacco OM disappearance. Leaf from bmr millet had less lignin (31 versus 44 g/kg DM,
standard error of the mean, sem = 2.0) and was more digestible than normal millet.
Digestible OM (g/kg DM, i.e. D value) was 602 for bmr and 572 for normal millet
(sem = 3.8). Similarly, millet bearing the bmr trait produced stems with less lignin (62
versus 87 g/kg DM, sem = 5.5) and a higher D value (479 versus 419 g/kg, sem = 4.9)
than their normal counterparts. Although leaf yield (g DM/plant) was not different
16 P.O. Osuji, S. Fern?ndez-Rivera and A. Odenyo
between genotypes (88.8 g for bmr and 85.2 g for normal millet, sem = 3.9), bmr millet
produced less stem than normal millet (85.6 versus 102.1 g DM/plant, sem = 4.4).
Pearl millet plants bearing the tr trait tended (P = 0.09) to produce more leaf (123.0
versus 95.3 g DM/plant, sem = 12.2) and stem (191.7 versus 152.4 g DM/plant,
sem = 16.4) than their normal counterparts. No differences were observed in fibre
constituents of leaves from both millet types, but tr millet leaves had a higher OM
concentration (918 versus 902, sem = 3.8, P < 0.01) possibly owing to the chemical
nature of the trichomes, and a higher OMD (670 versus 645 g/kg, sem = 6.1, P < 0.01)
than normal millet leaves. Stems from tr millet had less (P < 0.01) NDF (667 versus
726 g/kg DM, sem = 17.6), ADF (463, versus 537 g/kg DM, sem = 19.8) and lignin (87
versus 103 g/kg, sem = 5.4) and a higher OMD (521 versus 441 g/kg, sem = 9.1) than
stems from normal millet.
These results confirmed that the effects of the bmr trait on lignin and digestibility of
residues from millet grown for grain production and harvested at advanced stages of
maturity are similar to those found in forage millet. An increased residue quality of bmr
genotypes could be obtained at the expense of decreased stem yield. They also suggest
that the tr trait might improve the feed quality of millet stover without compromising
stover yield.
In summary, in the semi-arid zone the differences in feeding value of the stover
among sorghum and pearl millet varieties appear to be small and variable over the
years. However, there are possibilities for introducing quality-related genes that can
modify the composition of the fibre and increase digestibility. Further studies must
evaluate  the  influence  of such traits  on the  adaptability, grain yield and disease
resistance of the crops. Preliminary data suggest that the bmr trait reduces grain yield
and promotes lodging in pearl millet. However, these effects appear to vary across
parent lines (A. Kumar, personal communication), which offer additional possibilities
for selection.
Influence of genotype on feeding value of forage legumes
In collaboration with the International Institute for Tropical Agriculture (IITA) and
ICRISAT, the differences in feed quality of hay from cowpea varieties grown as sole
crop (20 entries) or inter-cropped with millet (16 entries) were studied.
Varieties grown as sole crops differed in leaf yield (range = 134 to 1115, sem = 145
kg DM/ha), stem yield (range = 157 to 1131, sem = 177 kg DM/ha) and litter leaf as
proportion of leaf produced (range = 0.221 to 0.772, sem = 0.09). No differences were
observed in OMD (mean = 720 g digestible OM/kg DM, P > 0.48) or in crude protein
(mean = 197 g/kg DM, P > 0.16) of leaf. However, high variation among varieties was
observed in OMD (range = 459 to 611, sem = 20 g/kg) and CP concentration of stems
(range = 108 to 190, sem = 6.6 g/kg DM). Similar trends were observed in the study
with varieties grown intercropped with millet, but in this experiment differences among
entries were also observed for CP of leaf.
Similar on-going trials with groundnut varieties show large differences across varie-
ties in feed value of the stem but little or no differences in feed quality of the leaf. In
Improving fibre utilisation and protein supply in animals fed poor quality roughages 17
summary, these studies suggest that the selection of cowpea and groundnut varieties for
feed quality and yield of hay is promising. This may result in important income or
production benefits, since these are the most important cash crops of the semi-arid zone
and are used widely in the more intensive feeding systems.
Influence of genotype on feeding value of multipurpose trees
In a collaborative study with the International Center for Research in Agroforestry
(ICRAF) the genetic variation in fodder quality of Combretum aculeatum, among other
species, was studied in 60 half-sib families. Seeds were collected in a 100 km long and
20 km wide area along the Niger River. The collection zone was divided into
provenances. From each provenance several half-sib families were collected, grown in
the nursery and transplanted to the fields (J. Weber, personal communication).
During the dry season, the concentration (g/kg DM) of N in C. aculeatum foliage
varied from 15.7 to 36.2 (mean = 22.4, SD = 3.1), that of phosphorus from 0.7 to 2.7
(mean = 1.5, SD = 0.4), and that of organic matter from 798 to 994 (mean = 918,
SD = 29). Although the samples were taken in early physiological stages, the results
suggest that C. aculeatum is a valuable source of protein for dry season feeding.
Preliminary analyses suggested that variation exists among families of C. aculeatum
for concentration of N, P and OM. Nitrogen concentration was positively correlated
with that of P (r = 0.41, P < 0.01) and OM (r = 0.25, P < 0.01). Digestibility data are
not yet available. Although preliminary, these results support the hypothesis that
geographical areas that produce superior genotypes of trees could be selected for
fodder yield and quality. Studies with other species of importance in the semi-arid zone
are on-going.
Suggested activities for ILRI rumen ecology research
Based on the results of ILRI research summarised above, future ILRI rumen ecology
work is planned:
1. To evaluate other in vitro methods (Odenyo et al 1991; Theodorou et al 1994) as
procedures for early assessment of MPTs for negative effects.
2. To extract and partially purify anti-nutrients from MPTs. These semi-purified
extracts will be used to isolate microbes that can degrade such substances.
3. To eventually identify MPTs that are not toxic to the rumen microbes and may be
safely used as supplements when basal roughage diets are fed. Both classical and
molecular methods will be used to evaluate effects of MPTs on rumen microbes.
4. To follow the microbial population dynamics according to seasonal variations.
5. To identify and characterise microbes from indigenous ruminants that effectively
degrade fibre, and those that can tolerate or degrade anti-nutrients from fodder trees
by classical and molecular methods. These microbes may be transferred to rumi-
nants not adapted to such diets. It is also possible to manipulate these microbes
using recombinant DNA technology to enhance fibre-degrading capabilities of
already established rumen microbes.
18 P.O. Osuji, S. Fern?ndez-Rivera and A. Odenyo
A collaborative programme  among ILRI, Australia and Indonesian scientists is
planned. The work will focus mainly on isolation of tannin tolerant or degrading
microbes from indigenous ruminants. The microbes will also be characterised geneti-
cally to make it possible to develop nucleic acid probes for future persistence studies
in the new host.
Conclusion
Results of ILRI work summarised above strongly suggest that the focus of future work
should include attempts to explain the role of rumen microbes in the utilisation of high
fibre feeds supplemented with multipurpose trees. Furthermore it is important to explain
the biochemical basis of the interactions between rumen microbes and the chemicals
contained in MPTs, especially anti-nutritional factors. There is need to do more research
on the seasonal variations of microbial populations in various ruminants across
agro-ecological zones. Improving fibre digestion through genetic selection of crops and
trees is another promising area of research.
References
Abule Ebro 1994. ?Supplementation of Friesian ? Boran Crossbred Calves Fed Teff (Eragrostis
tef) Straw with Graded Levels of Cowpea (Vigna unguiculata) and Dolichos lablab (Lablab
purpureus) Hay.? MSc thesis, Alemaya University of Agriculture, Alemaya, Ethiopia. 141
pp.
Acamovic T.J. and D?Mello J.P.F. 1984. Trypsin inhibition by leucaena leaf meal, leucaena seeds
and mimosine. Leucaena Research Reports 5:74?75.
Ahn J.H., Robertson B.M., Elliot R., Guteridge R.C. and Ford C.W. 1989. Quality assessment
of tropical browse legumes: Tannin content and protein degradation. Animal Feed Science
and Technology 27:147?156.
Allen M.S. and Mertens D.R. 1988. Evaluating constraints on fibre digestion by rumen
microbes. Journal of Nutrition 118:261?270.
Bartle S.J. and Klopfenstein T.J. 1988. Non-chemical opportunities for improving crop residue
feed quality. Journal of Production Agriculture 1:356?361.
Bonsi M.L.K., Osuji P.O., Nsahlai I.V. and Tuah A.K. 1994. Graded levels of Sesbania sesban
and Leucaena leucocephala as supplements to teff straw given to Ethiopian Menz sheep.
Animal Production 59:235?244.
Bonsi M.L.K., Osuji P.O. and Tuah A.K. 1995. Effect of supplementing teff straw with different
levels of leucaena or sesbania leaves on the degradabilities of teff straw, sesbania, leucaena,
tagasaste and vernonia and on certain rumen and blood metabolites in Ethiopian Menz sheep.
Animal Feed Science and Technology (in press).
Borens F.M.P. and Poppi D.P. 1990. The nutritive value for ruminants of Tagasaste
(Chamaecytisus  palmensis), a leguminous tree. Animal Feed Science and Technology
28:275?292.
D?Mello J.F.P. and Fraser K.W. 1981. The composition of leaf meal from Leucaena leuco-
cephala. Tropical Science 23:75?78.
D?Mello J.F.P. and Taplin D.E 1978. Leucaena leucocephala in poultry diets for the tropics.
World Review of Animal Production 24:41?47.
Improving fibre utilisation and protein supply in animals fed poor quality roughages 19
El Hassan S.M. 1994. ?Yeast Culture and Multipurpose Fodder Trees as Feed Supplements for
Ruminants.? PhD thesis, University of Aberdeen, Scotland, UK. 263 pp.
Fern?ndez-Rivera S. 1994. Seasonal variation in the microbial digestive activity, rumen ammo-
nia concentration and diet ingested by grazing ruminants in the Sahel. Proceedings of the
Society of Nutrition and Physiology 3:324.
Goodchild A.V. and McMeniman N.P. 1987. Nutritive value of acacia foliage and pods for
animal feeding. In: Turnbull J.W. (ed), Australian Acacias in Developing Countries. Pro-
ceedings of  an International Workshop held at the Forestry Training Centre, Gympie,
Queensland, Australia, 4?7 August 1986. ACIAR Proceedings 16. ACIAR (Australian Centre
for International Agricultural Research), Canberra, Australia. pp. 101?106.
ILCA (International Livestock Centre for Africa). 1991. Gliricidia, the rat killer. ILCA Weekly
News 106:2?3. ILCA, Addis Ababa, Ethiopia.
ILCA (International Livestock Centre for Africa). 1993. Feeding systems for cattle, goats and
sheep in semi-arid West Africa. ILCA Annual Programme Report 1992 25?26. ILCA, Addis
Ababa, Ethiopia.
ILCA (International Livestock Centre for Africa). 1994a. Feed resource use and nutrition of
ruminants in crop-livestock systems of semi-arid West Africa. ILCA Annual Programme
Report 1993 5?7. ILCA, Addis Ababa, Ethiopia.
ILCA(International Livestock Cenre for Africa). 1994b. ILCA Annual Programme Report 1993.
ILCA, Addis Ababa, Ethiopia. pp. 95?98.
ILCA (International Livestock Centre for Africa). 1995. Feed resource use and nutrition of
ruminants in crop-livestock systems of semi-arid West Africa. ILCA Annual Programme
Report 1994 3?6. ILCA, Addis Ababa, Ethiopia.
Khalili H. and Varvikko T. 1992. Effect of replacement of concentrate mix by wilted sesbania
(Sesbania sesban) forage on diet digestibility, rumen fermentation and milk production in
Friesian ? Zebu (Boran) crossbred cows fed low quality native hay. Animal Feed Science
and Technology 36:275?286.
Kinghorn A.D. and Smolenski S.J. 1981. Alkaloids of the papilionoideae. In: Polhill R.M. and
Raven P.H. (eds), Advances in Legume Systematics. Part 2. Royal Botanical Gardens, Kew,
UK. pp. 585?598.
Kurdi O.I. 1994. Final Post-Doctoral Report. ILCA (International Livestock Centre for Africa),
Addis Ababa, Ethiopia.
Lowry J.V., Cooke N. and Wilson R. 1984. Flavonol glycoside distribution in cultivars and hybrids
of leucaena leucocephala. Journal of the Science of Food and Agriculture 35:401?407.
Menke K.H., Rabb L., Salewski A., Steingass H., Fritz F. and Schneider W. 1979. The estimation
of the digestibility and metabolisable energy content of ruminant feeding stuffs from the gas
production when they are incubated with rumen liquor in vitro. Journal of Agricultural
Science (Cambridge) 93:217?222.
Mueller-Harvey I., McAllan A.B., Theodorou A.K. and Beever D.E. 1988. Phenolics in fibrous
crop residues and plants and their effects on the digestion and utilisation of carbohydrates
and proteins in ruminants. In: Reed J.D., Capper B.S. and Neate P.J.H. (eds), Plant Breeding
and the Nutritive Value of Crop Residues. Proceedings of a Workshop Held at ILCA, Addis
Ababa, Ethiopia, 7?10 December 1987. ILCA (International Livestock Centre for Africa),
Addis Ababa, Ethiopia. pp. 97?132.
National Academy of Sciences. 1979. Tropical Legumes: Resources for the Future. National
Academy Press, Washington, DC, USA. 331 pp.
20 P.O. Osuji, S. Fern?ndez-Rivera and A. Odenyo
Nsahlai I.V., Siaw D.E.K.A. and Osuji P.O. 1994. The relationships between gas production and
chemical composition of 23 browses of the genus Sesbania. Journal of the Science of Food
and Agriculture 65:13?20.
Odenyo A.A., Mackie R.I., Fahey G.C. Jr. and White B.A. 1991. Degradation of wheat straw
and alkaline hydrogen peroxide-treated wheat straw by Ruminococcus albus 8andRumino-
coccus flavefaciens FD-1. Journal of Animal Science 69:819?826.
?rskov E.R. 1991. Manipulation of fibre digestion in the rumen. Proceedings of the Nutrition
Society 50:187?196.
?rskov E.R. and McDonald I. 1979. The estimation of protein degradability in the rumen from
incubation measurements weighted according to rate of passage. Journal of Agricultural
Science (Cambridge) 92:499?503.
Osafo E.L.K. 1993. ?Sorghum Stover as a Forage: Cultivar Effects on Yield and Effects of
Chopping, Amount Offered, Supplementation and Variety on Intake, Selection and Live-
weight Gain in Ethiopian Sheep and Cattle.? PhD thesis, University of Reading, Reading,
UK. 347 pp.
Osuji P.O. 1994. Nutritional and anti-nutritional values of multipurpose trees used in agro-for-
estry systems. In: Copland J.W., Djajanegra A. and Sabrani M. (eds), Agroforestry and Animal
Production for Human Welfare: Proceedings of an International Symposium Held in Asso-
ciation with the 7th AAAP Animal Science Congress, Bali, Indonesia, 11?16 July 1994.
ACIAR Proceedings 55. ACIAR (Australian Centre for International Agricultural Research),
Canberra, Australia. pp. 82?88.
Osuji P.O., Sibanda S. and Nsahlai I.V. 1993. Supplementation of maize stover for Ethiopian
Menz sheep: effects of cottonseed, noug (Guizotia abyssinica) or sunflower cake with or
without maize on the intake, growth, apparent digestibility, nitrogen balance and excretion
of purine derivatives. Animal Production 57:429?436.
Powell J.M., Fern?ndez-Rivera S. and H?fs S. 1994. Effects of sheep diet on the nutrient cycling
in mixed farming systems of semi-arid wet Africa. Agriculture, Ecosystem and the Environ-
ment 48:263?271.
Reed J.D. 1986. Relationships among soluble phenolics, insoluble proanthro-cyanidins and
fibre in East Africa browse species. Journal of Range Management 39:5?7.
Reed J.D., Tedla A. and Kebede Y. 1987. Phenolics, fibre and fibre digestibility in the crop
residue from bird resistant and non-bird-resistant sorghum varieties. Journal of the Science
of Food and Agriculture 39:113?121.
Reed J.D., Yilma Kebede and Fussell L.K. 1988. Factors affecting the nutritive value of sorghum
and millet crop residues. In: Reed J.D., Capper B.S., Neate P.J.H. (eds), Plant Breeding and
the Nutritive Value of Crop Residues. Proceedings of a Workshop Held at ILCA, Addis Ababa,
Ethiopia, 7?10 December 1987. ILCA (International Livestock Centre for Africa), Addis
Ababa, Ethiopia. pp. 233?251.
Siaw D.E.K.A., Osuji P.O. and Nsahlai I.V. 1993. Evaluation of multipurpose tree germplasm:
The use of gas production and rumen degradation characteristics. Journal of Agricultural
Science (Cambridge) 120:319?330.
Tangendjaja B. and Lowry J.B. 1984. Usefulness of enzymatic degradation of mimosine in
leucaena leaf for monogastric animals. Leucaena Research Reports 5:55?56.
Tangendjaja B., Rahardjo Y.C. and Lowry J.B. 1990. Leucaena leaf meal in the diet of growing
rabbits: Evaluation and effect of a low-mimosine treatment. Animal Feed Science and
Technology 29:63?72.
Improving fibre utilisation and protein supply in animals fed poor quality roughages 21
Theodorou M.K., Williams B.A., Dhanoa M.S., McAllan A.B. and France J. 1994. A simple gas
production method using a pressure transducer to determine the fermentation kinetics of
ruminant feeds. Animal Feed Science and Technology 48:185?197.
Topps J.H. 1992. Potential composition and use of legume shrubs and trees as fodders for
livestock in the tropics. Journal of Agricultural Science (Cambridge) 118:1?8.
Umunna N.N., Osuji P.O., Nsahlai I.V., Khalili H. and Mohamed-Saleem M.A. 1995. The effect of
supplementing oat hay with lablab, sesbania, tagasaste or wheat middlings on the voluntary intake
nitrogen utilisation and gain of Ethiopian Menz sheep. Small Ruminant Research (in press).
22 P.O. Osuji, S. Fern?ndez-Rivera and A. Odenyo
Potential application of rumen
ecology manipulation to animal
nutrition in developing countries
Nitrogen metabolism in the rumen:
biotechnological problems and prospects
R.I. Mackie
1
and M. Morrison
2
1. Department of Animal Sciences and Division of Nutritional Sciences
University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA
2. Department of Animal Sciences and Center for Biotechnology, University of Nebraska-Lincoln,
Lincoln, NE 68583-0908, USA
Abstract
Ruminant nutritionists have recognised the central importance of ammonia
in nitrogen metabolism in the rumen for many years. This was emphasised
in studies in which growth, reproduction and lactation were obtained from
dairy cows on protein-free, purified diets and were thus solely dependent on
the products of ruminal fermentation. The ruminant mode of digestion in
which  microbial  fermentation  precedes  the hydrolytic  digestive  process
allows ruminants to utilise dietary non-protein nitrogen efficiently, to up-
grade low quality dietary protein to high quality microbial quality protein,
and to survive on low intakes of dietary nitrogen by efficiently utilising
nitrogen recycled to the rumen via the saliva. Since ruminal ammonia not
utilised for microbial growth is absorbed through the ruminal walls and
converted to urea in the liver, it follows that nitrogen economy of ruminant
animals is dependent on proper balances of degradable and bypass protein.
If nitrogen supply to the rumen bacteria is inadequate, rumen function is
affected in a number of ways. Digestion of starch and cellulose and other
polysaccharides can be depressed. Bacterial synthetic effort can be diverted
from protein to storage polysaccharide and adenosine triphosphate (ATP) can
be diverted to uncoupled fermentation. Death and lysis of bacteria will probably
be increased and the reduced growth rate means that a larger proportion of ATP
is likely to be used for maintenance and less for protein synthesis. In the rumen
ecosystem the rate and efficiency of microbial protein synthesis is determined
by assimilation of the key enzymes of ammonia or the energy required for this
process. The three primary enzymes of ammonia assimilation are glutamate
dehydrogenase, glutamine synthetase and glutamate synthase and the synthesis
and activity of these enzymes are central to the regulation of the intracellular
nitrogen pool and the control of ammonia assimilation. These enzymes of
ammonia assimilation play a critical role in the rate and efficiency of bacterial
protein synthesis in the rumen. The recent biotechnology initiative on the
genetic manipulation of the ruminal ecosystem has focused entirely on fibre
degradation. The enzymes cloned from ruminal bacteria into Escherichia coli
encode enzymes involved in the hydrolysis of plant cell wall material.
Introduction
Feedstuffs consumed by ruminants are all initially exposed to the fermentative activity
in the rumen prior to gastric and intestinal digestion. Dietary polysaccharides and
protein are generally degraded by the ruminal microorganisms into characteristic end
products, which in turn provide nutrients for metabolism by the host animal. The extent
and type of transformation of feedstuffs thus determines the productive performance of
the host. Fermentation of feedstuffs in the rumen yields short-chain volatile fatty acids
(VFA) (primarily acetic, propionic and butyric acids), carbon dioxide, methane, ammo-
nia and occasionally lactic acid. Some of the change in free energy (?G
0
?)isusedto
drive microbial growth, but heat also is evolved. Ruminants use the organic acids and
microbial protein as sources of energy and amino acids, respectively, but methane, heat
and ammonia can cause a loss of energy and nitrogen (N). The quality and quantity of
rumen fermentation products is dependent on the types and activities of the microor-
ganisms in the rumen. This, in turn, will have an enormous potential impact on nutrient
output and performance of ruminant animals. It is only with a thorough understanding
of the mechanisms involved that this system can by successfully manipulated and fully
exploited (Mackie and White 1990).
Ammonia plays a central role as an intermediate in the degradation and assimilation
of dietary nitrogen by rumen bacteria. Ammonia is the major end-product of digestion
of dietary protein and non-protein nitrogen (urea and amino acids) as well as the major
source of nitrogen for synthesis by ruminal bacteria. A comparison of peptide and amino
acid utilisation showed that peptides were more effectively incorporated into bacterial
protein while a greater proportion of amino acids were fermented to VFA. Ruminal
digestion results in the production of VFA and bacterial cells which are used as the major
energy and protein sources, respectively, for metabolism by the host animal. As a result,
nitrogen metabolism in the rumen is intimately related to the metabolism and utilisation
of nitrogen by ruminant tissues. Growth and production in ruminants is dependent on
bacterial protein synthesised in the rumen and ammonia is of central importance in this
process. The integration of metabolism demands that consideration be given to protein
and energy interactions in the rumen in order to achieve a balanced supply of nutrients
at the duodenum. For the ruminant animal, the quantity of microbial plus undegraded
dietary protein arriving at the duodenum has a great influence on productivity. The
potential of enhancing the productivity and lean tissue growth of ruminants by growth
promotants and hormone manipulations is determined by a balanced supply of protein
and energy to the animal (MacRae and Lobley 1986).
Nitrogen metabolism in the rumen
Proteases, proteolysis and protein degradation
In ruminants, the amino acid requirements are provided by microbes synthesised in the
rumen and from dietary protein that is not degraded in the rumen but is intestinally
digestible (bypass or escape protein). A large but variable proportion (60% to 90%) of
26 R.I. Mackie and M. Morrison
the dietary protein is degraded by rumen microorganisms and it is the rate at which the
different proteins can be hydrolysed that controls the extent of their degradation before
they pass out of the rumen (Leng and Nolan 1984; Mackie and Kistner 1985; Tamminga
1979). This has an important influence on the proportions of undegraded dietary protein
and microbial protein that are presented to the small intestine for digestion by the host
animal. Furthermore, this forms the basis of all modern systems for evaluating and
predicting protein utilisation by ruminants. Although much attention has been focused
on physical and chemical methods of controlling the rate of protein degradation in the
rumen, little research has been done on the factors influencing the proteolytic activity
of ruminal bacteria despite the nutritional significance of this activity.
For ruminants receiving diets containing protein, the change in quantity and pattern
of amino acids, which results from the conversion to microbial protein in the rumen,
can be an advantage or a disadvantage, depending on the composition of the food
protein. If the latter is of good quality, biological value is reduced because the microbial
protein is of lower digestibility and is accompanied by nucleic acids. Under these
conditions, it would be advantageous to limit degradation of dietary protein, provided
this does not lead to a reduction of the microbial population and their activities, such as
fibre digestion and VFA production (Smith 1979). However, proteolysis during ruminal
fermentation may benefit the host animal if the microbial protein synthesised from the
products is of higher biological value than the feed proteins (Tamminga 1979).
Recent work confirms that the proteolytic activity in the rumen is almost entirely
associated with bacterial cells and that cell-free rumen fluid and protozoa have little
activity toward soluble proteins (Nugent and Mangan 1981). However, the protozoa
play an important role in the engulfment of bacteria and particulate matter and hence
degradation of insoluble proteins (Coleman 1979). Proteolytic activity in the rumen is
not confined to a single bacterium but is a variable property possessed by many different
bacteria  that may be  active in the  degradation of other feed constituents, mainly
carbohydrates. Furthermore, the predominant proteolytic bacteria will differ depending
on diet (Mackie 1982). Although extracellular enzymes are usually produced by Gram-
positive bacteria, the most important protease-producing bacteria in the rumen are
Gram-negative, including species of Prevotella, Selenomonas, and Butyrivibrio (Cotta
and Hespell 1986). Other species, probably of less significance, are Megasphaera
elsdenii, Streptococcus bovis, Clostridium spp., Eubacterium spp., Lachnospira multi-
parus, Succinivibrio dextrinosolvens, and the Spirochaetes (Allison 1970). Russell et
al (1981) demonstrated that the Gram-positive S. bovis played a predominant role in
ruminal proteolysis, especially on high concentrate diets.
Rumen microbial proteolytic activity has a broad optimum in the neutral range
(Blackburn and Hobson 1960; Brock et al 1982; Kopecny and Wallace 1982). In general,
rumen bacteria have a mixture of proteases based on the chemistry of their active site
and the specificity of their hydrolytic activity (Brock et al 1982; Kopecny and Wallace
1982; Prins et al 1983; Wallace 1983; Wallace and Brammall 1985). This activity is
predominantly of the cysteine protease type with large contributions from the serine-
and metallo-protease type. Aspartic acid proteases were of minor importance in rumen
bacteria. Protozoal proteases were mainly of the cysteine type and aspartic activity was
Nitrogen metabolism in the rumen: biotechnological problems and prospects 27
significant (Forsberg et al 1984). In contrast, the proteases of Neocallimastix frontalis,
a prominent member of the anaerobic rumen fungi, had significant metallo-protease
activity with trypsin-like specificity (Wallace and Joblin 1985).
Proteases have been studied in few pure cultures of rumen bacteria. Prevotella rumini-
cola, which can utilise peptide N instead of ammonia, had proteolytic activity in batch
culture that was maximal and mostly (>90%) cell-associated during the midexponential
phase of growth (Hazlewood and Edwards 1981; Hazlewood et al 1981). The proteolytic
activity comprised a mixture of serine, cysteine and aspartic acid proteinases with the
possibility that some of the activity is dependent on the presence of metal ions (Mg
2+
).
Ruminobacter amylophilus produces both cell-bound and cell-free proteolytic activity,
which consistently amounts to 80 and 20%, respectively, of the total activity during
exponential growth, but the proportion of cell-free activity increases in stationary phase
cultures (Blackburn 1968). This cell-free activity may be truly extracellular or it might
indicate that a proportion of bacteria are lysing during growth.
In contrast, proteolytic activity of Butyrivibrio fibrisolvens was essentially all ex-
tracellular, regardless of stage of growth. The number and approximate molecular
weight (MW) of extracellular proteases produced by Butyrivibrio fibrisolvens H17c
were determined by gelatin-PAGE. Nine bands of protease activity with apparent MW
of approximately 101,000, 95,000, 87,000, 80,000, 76,000, 68,000, 63,000, 54,000 and
42,000 were found in supernatants from exponential phase sultures (Strydom et al 1986).
The proteases were stable and survived unchanged in stationary phase cultures. The
activity of all 10 exoproteases was optimal between pH 6.0 and 7.5 and a temperature
of 55?C. The activities of all 10 protease fragments were inhibited by the serine protease
inhibitor phenylmethylsulfonyl fluoride (PMSF). Their activities were not affected by
inhibitors of trypsin-like enzymes or metallo-, sulfhydryl- and carboxylproteases, thus
confirming that they are all serine proteases. Production of the 10 exoproteases was not
subject either to nitrogen or carbohydrate catabolite repression, and all 10 exoproteases
were produced under conditions of nitrogen and carbohydrate excess. Production of
serine exoproteases by Butyrivibrio fibrisolvens H17c did not require a specific inducer
and exoproteases were constitutively produced in various media. Production of exopro-
teases  was positively  correlated  with growth  and  would therefore be growth-rate
dependent.
However, additional strains of these organisms and other proteolytic species require
urgent examination before a clear understanding of the proteolytic activity in the rumen
will be achieved.
Peptide transport and utilisation
Proteolysis results in oligopeptide production. These oligopeptides then undergo deg-
radation to smaller peptides and amino acids. A comparison of peptide and amino acid
used showed that peptides were more effectively incorporated into bacterial protein
while a greater proportion of amino acids were fermented to VFA (Cotta and Hespell
1986; Wright 1967). Thus, free amino acids are not incorporated into microbial protein
per se but undergo rapid deamination providing ammonia for bacterial growth. The
28 R.I. Mackie and M. Morrison
deamination and degradation of specific amino acids are of special relevance to bacterial
growth in the rumen. The most important of these is the conversion of leucine, isoleucine
and valine to isovalerate, 2-methylbutyrate and isobutyrate, respectively. These
branched-chain fatty acids are either required or highly stimulatory to the growth of
many ruminal bacteria, particularly the fibrolytic species (Bryant and Robinson 1962,
1963).
It has been well documented that many rumen bacteria can utilise peptides as a source
of amino acids; for several bacterial systems, peptide transport into the cell followed by
intracellular cleavage has been demonstrated. Because transport mechanisms for free
acids appear to be absent in many rumen bacteria, the transport of peptides into the cell
represents an important means of obtaining N for the synthesis of cellular components,
especially those lacking the ability to utilise ammonia. Pittman and Bryant (1964)
showed that Prevotella ruminicola was able to utilise peptide and ammonia N but not
amino N for growth. In further studies, Pittman et al (1967) showed that the same strain
of Prevotella ruminicola did not transport [
14
C]proline or glutamic acid and very little
valine, whereas peptides of MW up to 2000 were transported. Russell (1983) also
showed that the addition of Trypticase to cultures of Prevotella ruminicola B
1
4signifi-
cantly improved growth yields, although peptides alone could not support the growth
of this organism. However, the actual transport of peptides by rumen bacteria has only
been demonstrated using a radiolabelling technique that is unable to distinguish between
possible extracellular hydrolysis of labelled proteins, or possible efflux of labelled
amino acids from the intracellular pool.
Studies on the transport of peptides of varying chain length and amino acid compo-
sition were carried out with Streptococcus bovis using a dansylation procedure (West-
lake and Mackie 1990; Westlake et al 1987). Under these conditions the transport of the
majority of components of Trypticase, Leu-Trp-Met-Arg-Phe, and Phe-Arg has been
demonstrated, although the defined structure peptides were completely hydrolysed by
extracellular peptidases to smaller units before transport. The pentapeptide was hydro-
lysed to eight different components in the transport buffer and all were present within
the cell after 2 min of incubation. The Phe-Arg was hydrolysed extracellularly to
phenylalanine and a second compound, but not alanine. However, both alanine and
phenylalanine were present within the cell after 2 min of incubation. These studies
emphasise the limitations in the use of
14
C-labelled substrates in transport experiments
and the need to account for possible extracellular hydrolysis of labelled compounds,
which in the case of Streptococcus bovis appears to be considerable. Streptococcus
bovis, a Gram-positive organism, has been shown to play a major role in ruminal
proteolysis on high concentrate diets (Russell et al 1981).
Ammonia assimilation in rumen bacteria
Depending upon the diet, 60?90% of the daily nitrogen intake by the ruminant is
converted to ammonia and 60?80% of bacterial N is derived from ammonia, with the
balance derived from N-containing compounds (peptides mainly) which do not equili-
brate with the NH
3
-N pool (Pilgrim et al 1970; Mathison and Milligan 1971; Nolan and
Nitrogen metabolism in the rumen: biotechnological problems and prospects 29
Leng 1972). Bryant and Robinson (1961, 1962, 1963) found that 92% of ruminal
bacterial isolates could utilise ammonia as the main source of N, while it was essential
for the growth of 25% of all isolates tested. Little amino acid-nitrogen is available for
microbial growth and amino acids are rapidly deaminated to ammonia, VFA and CO
2
before being utilised for microbial protein synthesis. The deamination and degradation
of specific amino acids are of special relevance to bacterial growth in the rumen. The
most important of these is the conversion of leucine, isoleucine and valine to isovalerate,
2-methylbutyrate and isobutyrate, respectively. These branched-chain fatty acids are
either required or highly stimulatory to the growth of many ruminal bacteria, particularly
the fibrolytic species (Bryant and Robinson 1961, 1962, 1963). Although our knowledge
about other mammalian gastrointestinal ecosystems is not as extensive or definitive as
that of the rumen, results indicate that ammonia is the major N source supporting
bacterial growth in the cecum and colon of pigs, horses and humans (Allison et al 1979;
Herbeck and Bryant 1974; Maczulak et al 1985; Robinson et al 1981; Takahashi et al
1980; Varel and Bryant 1974; Wozny et al 1977). Thus, the rumen provides an ideal
model for the study of these processes since it has been extensively characterised.
Ammonia produced in the rumen that is not incorporated into microbial cells is largely
absorbed through the wall of the reticulo-rumen and converted into urea by the liver.
This is excreted in the urine leading to environmental pollution in the vicinity of
intensive animal production operations. Between 60 and 75% of the N in excreted
manure is converted into NH
3
, of which 25 to 40% is lost during storage, and an
additional 20 to 60% is lost during spreading (Tamminga and Verstegen 1992).
The concentration of ammonia in the rumen varies widely ranging from 2?40 mM.
Satter and Slyter (1974) found that when the concentration of NH
3
-N is lower than 3.5
mM, microbial growth decreased significantly. Levels of ammonia in dairy cows were
found to range from 7?13.5 mM (Wohlt et al 1976). Thus, under adequate feeding
regimens, prevailing ammonia concentrations should always be adequate for optimal
growth of rumen bacteria. Earlier research with sheep fed low-protein (2.9% CP)
Eragrostis tef hay showed NH
3
-N concentrations of 1.8 mM immediately after feeding
which went down to 0.05 mM 8 h after feeding (Schwartz and Gilchrist, Internal
Progress Report, Digestion and Metabolism in the Ruminant Unit, Veterinary Research
Institute, Onderstepoort, South Africa). Grazing ruminants could be expected to have
similar low ammonia levels during winter or dry periods not only in the USA but
elsewhere in the world. It is worth noting that most of the world?s ruminant population
live and produce under grazing conditions that exist in the subtropics and tropics
(Gilchrist and Mackie 1984). When environmental ammonia concentrations are greater
than 1 mM, the glutamate dehydrogenase pathway of ammonia assimilation predomi-
nates, whereas when concentrations of ammonia are less than 0.5 mM the high affinity
glutamine synthetase pathway is more important. Schaefer et al (1980) reported ammo-
nia saturation constants for 10 species of predominant ruminal bacteria which ranged
from 5 to 50 uM which suggests that these organisms could attain 95% of their maximal
growth rate in the presence of 1.0 mM ammonia and an unlimited energy supply.
Although many species of ruminal bacteria are efficient scavengers of ammonia, both
pathways may be used simultaneously in different microorganisms. Thus, irrespective
30 R.I. Mackie and M. Morrison
of the prevailing environmental ammonia concentration, enzymes of ammonia assimi-
lation are critical in procuring this essential nutrient for bacterial cell growth.
Theoretically, optimal ruminal NH
3
concentrations required for maximal microbial
growth are in the range of 2?3 mM; a concentration below 1 mM is limiting. However,
this assumes that there is no other limiting nutrient and is an oversimplification since
many other factors need to be considered. Thus bacteria adherent on surfaces may be
exposed to environmental NH
3
concentrations that are less than optimal. Additionally,
NH
3
concentrations that facilitate maximal levels of fibre degradation may be lower
than those that maximise microbial protein synthesis and feed intake. It is also possible
that, in grazing ruminants consuming low protein forages where ruminal microbes are
dependent mainly on recycled NPN sources, fermentation uncoupled from bacterial
growth may occur. This area of research needs further focus.
Enzymes of ammonia assimilation
Since ammonia is the preferred source of N for most rumen bacteria, enzymes of
ammonia assimilation are essential to the growth of most rumen microorganisms.
Glutamate dehydrogenase (GDH) and the dual enzyme system glutamine synthetase
(GS) and glutamate synthase (GOGAT) are the two most important routes by which
ammonia may be assimilated.
Glutamate dehydrogenase (GDH; EC 1.4.1.2 and 1.4.1.4) catalyses the following
reaction:
GDH
2-Ketoglutarate + NH
3
+ NAD(P)H + H
+
?? Glutamate + NAD(P)
+
High levels of both NADH- and NADPH-linked GDH activity have been observed
in ruminal contents (Chalupa et al 1970) and in continuous culture of mixed ruminal
bacteria (Erfle et al 1977). NADH-linked GDH has been reported in Ruminococcus
albus and Megasphaera elsdenii, whereas most other ruminal bacteria possess NADPH-
linked activity (Joyner and Baldwin 1966). Selenomonas ruminantium was shown to
possess both types of activity in a relatively constant activity ratio (0.25?0.35) under
both glucose- and ammonia-limited growth conditions (Smith et al 1980). In general,
NADP-linked GDH is assumed to have a biosynthetic role and functions efficiently at
high ammonia concentrations as evidenced by its high K
m
for ammonia. The NAD-
linked enzyme has a catabolic function associated with limiting levels of ammonia and
growth on amino acids (glutamate) (Brown 1980; Dalton 1979). Smith et al (1980)
showed that NADP-linked GDH activity was optimal at ca 400 mM monovalent salt
concentration (Na
+
,K
+
or NH
4
+
)inS. ruminantium. Apparent K
m
values for this
enzyme of 6.7 and 23 mM were estimated from a double-reciprocal plot of the enzyme
activity as a function of NH
4
Cl concentration. It is also of interest that GDH activity in
Bacteroides thetaiotaomicron was influenced by salt  concentration.  NADP-linked
activity increased by 73% and NAD-linked activity decreased by 50% in the presence
of 0.1 M NaCl (Glass and Hylemon 1980).
The second major pathway for ammonia assimilation in ruminal bacteria is mediated
by the concerted action of two enzymes. Glutamine synthetase (GS; EC 6.3.1.2) fixes
ammonia in the amide of glutamine by an ATP-dependent reaction. Then glutamate
Nitrogen metabolism in the rumen: biotechnological problems and prospects 31
synthase (GOGAT; EC 1.4.1.13), the second enzyme of this system, catalyses the
reductive transfer of the glutamine amide to 2-ketoglutarate (Meers et al 1970).
GS
Glutamate + ATP + NH
3
? Glutamine + ADP + P
GOGAT
2-Ketoglutarate + Glutamine + NADPH + H
+
? 2 Glutamate + NADP
+
Glutamine synthetase activity can be assayed by several different methods (Hespell
1984) which makes the observation of GS activity in whole rumen contents (Chalupa
et al 1970) and during continuous culture of mixed ruminal bacteria (Erfle et al 1977)
difficult to interpret since each organism has different conditions for optimal enzyme
activity. In S. ruminantium, GS activity was measured using the biosynthetic assay since
?-glutamyltransferase activity was absent (Smith et al 1980). Furthermore, GS activity
was not under covalent control by the adenylylation-deadenylylation reaction and
feedback inhibition commonly found in other bacteria. In contrast, both Ruminobacter
(Bacteroides) amylophilus (Jenkinson et al 1979) and Succinivibrio dextrinosolvens
(Patterson and Hespell 1985) contain ?-glutamyltransferase activity which was under
control of adenylylation-deadenylylation Escherichia coli type of regulation for GS.
Glutamate synthase activity has been demonstrated in only one strain of the gut
anaerobe S. ruminantium strainD(Smithetal1981)andwasunusualinthatneither
NADPH or NADH were suitable as electron donors for the reaction. The enzyme was
reductant-dependent and used dithionite-reduced methyl viologen as the electron carrier
system in the assay. Thus some unidentified low-potential electron carrier is required
and may help to explain problems involved in measuring GOGAT activities in other
species of rumen bacteria.
The Ruminococci are mesophilic, Gram-positive bacteria which play an important
role in ruminal cellulose digestion. Ruminococcus albus and R. flavefaciens are among
the predominant cellulolytic bacteria isolated from the rumen of animals fed a variety
of diets and ammonia is essential for their growth since they are unable to utilise nitrogen
from amino acids or peptides (Bryant and Robinson 1961, 1962, 1963). Reports of GDH
activity in Ruminococci are scarce. Joyner and Baldwin (1966) reported the presence
of NADH-linked GDH activity in R. albus. This was confirmed in R. albus 7 grown
under  glucose limitation in chemostat (Kistner  and  Kotze 1973). Cell  associated
NADP-linked GDH activity in R. flavefaciens 67 was approximately three-fold higher
in cells grown under ammonia limitation and was independent of growth rate (D =
0.06?0.15 h
?1
) (Pettipher and Latham 1979).
We have focused our studies on ammonia assimilation by R. flavefaciens FD-1 in
both batch and continuous culture (Duncan and Mackie 1989; Duncan et al 1991, 1992).
Initially the principal routes of ammonia assimilation and the activities of the corre-
sponding enzymes were examined. Results from batch culture confirm that GS and
GDH are active routes of NH
3
assimilation and that their relative activities depend on
the concentration of NH
3
in the growth medium. GS activity increased under N-limiting
conditions but activity was inhibited 50% by NH
3
shocking. The adenylylation form of
32 R.I. Mackie and M. Morrison
GS modification common in many bacteria (Magasanik 1982; Merrick 1988) does not
occur in R. flavefaciens (Duncan et al 1992).
All GS and most of the GDH activity present in whole cells was sedimented by
ultracentrifugation (328,000 ? g; 37 min; 4?C). Streicher and Tyler (1980) reported a
rapid purification of GS from a variety of bacteria based on differential centrifugation
which depends on the integrity of the GS-DNA complex in cell extracts. However, in
contrast to their findings, prior treatment of the cell extracts of R. flavefaciens with
DNAse or RNAse had no effect on sedimentation. However, harvesting and fractiona-
tion of cells in the presence of a detergent (CTAB) reduced the amount of sedimentable
activity of both GS and GDH and the balance was recovered in the supernatant. This
provides preliminary evidence that GS and some of the GDH activity may be mem-
brane-associated. The nature of the interaction with the membrane fraction is being
studied using different detergents and salt washes combined with PAGE and activity
gel assays to locate activity of GS and GDH.
Reaction kinetics of NADP-linked GDH were assayed in cell fractions (membrane
pellet and supernatant) to address the possible existence of multiple forms of GDH with
different physiological roles and kinetic characteristics. Michaelis-Menten kinetics
were demonstrated and Lineweaver-Burk plots were linear for the substrate concentra-
tion range (1?50 mM). The affinity for ammonia was higher in the membrane pellet
fraction (K
m
= 14.2mM) than supernatant (K
m
= 47.9 mM). Further research is required
to determine if these are different enzymes or different forms (isozymes) of the same
enzyme.
Glutamate dehydrogenase (L-glutamate: NADP
+
oxidoreductase, deaminating, E.C.
1.4.1.4; GDH) from R. flavefaciens FD-1 has been purified and characterised in our
laboratory (Duncan et al 1992). The native enzyme and subunits are 280 and 48 kDa,
respectively, suggesting that the native enzyme is a hexamer. This is characteristic of
bacterial and fungal NADP-linked GDHs (Britton et al 1992). The enzyme requires 0.5
M KCl for optimal activity, and has a pH optimum of 6.9?7.0. The K
m
?s for ammonia,
?-ketoglutarate and glutamate are 19, 0.41 and 62 mM, respectively. The sigmoidal
NADPH saturation curve revealed positive cooperativity for the binding of this coen-
zyme. The first residue in the N-terminal sequence was alanine, suggesting that the
protein may be modified post-translationally. Comparison of the N-terminal sequence
with those of E. coli, S. typhimurium and Clostridium symbiosum revealed only 39%
amino acid homologies. Sequence comparisons for GDHs from a diverse range of
sources show that the hexameric enzymes are structurally similar. However, the simi-
larity is poor for the N-terminal 50 residues and then rises significantly with long
stretches of highly similar sequences over the next 350 residues (Britton et al 1992).
Regulation of ammonia assimilation in bacteria
A thorough understanding of the process of ammonia assimilation by bacteria requires
identification  of the possible pathways of assimilation for different N  sources  in
different organisms and characterisation of the enzymes involved in these pathways.
This knowledge must then be integrated with information on the physiological state of
Nitrogen metabolism in the rumen: biotechnological problems and prospects 33
cells growing under different conditions. Finally, a genetic analysis is required both of
the structural genes which encode the enzymes of ammonia assimilation and of the
regulatory genes which are essential in the control and coordination of the level of these
enzymes in the cell. Such a complete description has been achieved in very few systems
but the most comprehensive description available thus far is in the enteric bacteria,
namely E. coli, S. typhimurium, Klebsiella pneumoniae and K. aerogenes (Kustu et al
1986; Magasanik 1982, 1988; Merrick 1988). However, in rumen bacteria we have no
information on the mechanisms controlling expression of the primary ammonia assimi-
lation enzymes (GS, GOGAT and GDH). Rapid developments in these systems can be
expected due to the availability of gene probes and promoter sequence information from
the enterics.
Of the three  primary enzymes of ammonia  assimilation  only GS expression is
markedly regulated in response to changes in the availability of ammonia. GS from
enteric bacteria is comprised of 12 identical subunits each with 50 kDa molecular
weight. In E. coli, GS activity is regulated by covalent modification involving reversible
adenylylation of a specific tyrosine residue on each subunit. Maximum biosynthetic
activity is obtained when the enzyme is completely unadenylated. The adenylation of
GS is carried out by an adenylyl transferase (ATase) encoded by the glnE gene. ATase
can act in reverse to deadenylate GS and the activity of ATase is determined by its
interaction with a small regulatory protein P
II
. If ammonia is abundant, ATase acts to
adenylylate GS and reduces its activity, while under ammonia limitation GS is deadeny-
lated  by  ATase. The regulatory protein  P
II
is the product  of the glnB gene. The
unmodified P
II
stimulates adenylylation of GS by ATase while uridylylated P
II
enhances
the deadenylylation reaction. The covalent attachment of a UMP residue to P
II
is caused
by uridylyl transferase (UTase), the product of the glnD gene. Regulation of GS activity
is therefore achieved by a complex cascade of events which are ultimately controlled
by the relative levels of glutamine and 2-ketoglutarate inside the cell. A high glu-
tamine:2-ketoglutarate ratio causes UTase to deuridylylate P
II
which in turn stimulates
adenylylation of GS by ATase. The reverse sequence occurs when the ratio is low. GS
levels are also regulated transcriptionally such that in ammonia-limited cultures, the
level of GS is higher than that measured in cells grown with excess ammonia. The
system which mediates this control is known as the nitrogen regulation (ntr) system.
The gene which encodes GS (glnA) has been cloned from the enterics as well as several
other bacteria (Merrick 1988; Southern et al 1986; Usdin et al 1986).
In contrast GOGAT and GDH are apparently subject to far less control and even in
the enterics, the precise nature of their regulation is presently unknown (Merrick 1988).
However, recent research on the nac gene suggests a second tier of nitrogen regulation
which intercedes between ntr and nitrogen-controlled systems including GOGAT and
GDH. It is possible that nac is specifically involved in the regulation of systems which
can provide the cell with carbon as well as nitrogen (Bender 1991). Cloning and
sequencing of nac and the analysis of promoters under nac control have recently been
published (Maculuso et al 1990). This type of system may be important in rumen
bacteria where catabolite regulatory mechanisms appear to be involved in regulation
(Russell and Baldwin 1978; Russell et al 1990).
34 R.I. Mackie and M. Morrison
Bacterial ammonium transport
With few exceptions, the metabolism of a compound starts with its transport across the
cell membrane, mediated in most cases by specific transport proteins or carriers. This
uptake is frequently strictly regulated. However, regulation of transport is almost
impossible if a compound crosses the membrane by non-specific diffusion. With respect
to bacterial membranes, there is ample evidence that transport systems for most ions
and large polar molecules (M
r
approx >100) exist. If the cell membrane was not
basically impermeable, it would be difficult for cells to retain ionic nutrients as these
could also diffuse out into the surrounding dilute aqueous environment. However, the
exception may be those ions that are in equilibrium with a non-ionic, relatively lipid
solubleformsuchastheNH
4
+
-NH
3
pair. The important question is which form of
ammonia passes across the membrane, since the pKa for the dissociation
NH
4
+
?NH
3
+H
+
is 9.25 at 24?C. It is obvious that at physiological pH, the vast bulk of substrate is present
as NH
4
+
. However, it is possible that under conditions of high substrate repression and
inhibition of scavenging active transport systems, passive diffusion across the cell
membrane could still occur. A case could also be made for a facilitated diffusion
mechanism, although net flux would only occur if the solute passes down a concentra-
tion gradient.
The occurrence of NH
4
+
transport systems is widely distributed amongst many
species of bacteria with varying physiology and ecology. In general, the affinity of the
NH
4
+
carriers is high, with K
m
values of 5?50 ?M (Kleiner 1985). Most of the
ammonium transport systems have been detected with
14
CH
3
NH
3
+
. With the exception
of the cyanobacteria, the NH
4
+
gradient is similar to the methylamine gradient and
ranges from 50 to 200 fold. The cyanobacteria are able to generate NH
4
+
gradients as
high as 3000 fold (Boussiba et al 1984). Another important feature apart from concen-
trative uptake is energy  dependence.  Energy dependent NH
4
+
transport  has been
inferred from studies which demonstrate dependence on an energy source such as ATP,
inhibition by inhibitors of energy metabolism, and inhibition by compounds which
decrease PMF. Taken together, the available evidence favours a component of the PMF,
the membrane potential (??), as the driving force for NH
4
+
transport in most bacterial
carriers (Kleiner 1985).
The synthesis of the bacterial NH
4
+
carriers is repressed when grown in media with
high levels of NH
4
+
. This is similar to the high affinity GS enzyme system for ammonia
assimilation described earlier. These enzymes are under the nitrogen control (ntr)
regulatory system (Magasanik 1982; Merrick 1988) which comprises three regulatory
genes ntrA (or glnF), ntrB (glnL) and ntrC (glnG). The first indication that the ntr system
may also regulate ammonium transport (Amt) was derived from a regulatory mutant of
Klebsiella pneumoniae (Kleiner 1982) which was unable to transport methylamine. This
mutant was also defective in glutamine synthetase (Gln
-
), nitrogenase (Nif
-
)and
histidine utilisation (Hut
-
). A more detailed study carried out with E. coli strains
containing mutations in different ntr genes showed that both the ntr A (glnF) and the
ntrC (glnG) gene products were required to activate synthesis of the NH
4
+
carrier, while
Nitrogen metabolism in the rumen: biotechnological problems and prospects 35
the ntrB (glnL) gene product played a role in its repression (Servin-Gonzalez and
Bastarrachea 1984). Similar effects were observed on the synthesis of GS so that the
Gln
-
phenotype also was always Amt
-
.
Under conditions of derepression (synthesis) of the NH
4
+
carrier, bacteria generally
assimilate NH
4
+
via the GS/GOGAT pathway (Magasanik 1982; Merrick 1988). The
putative regulator of NH
4
+
transport is thought to be glutamine since intracellular levels
respond rapidly to external NH
4
+
concentrations (Kleiner 1976, 1979; Kleinschmidt and
Kleiner 1981). It serves as a feedback inhibitor of several enzymes strongly affecting the
modification of GS (Kleiner et al 1981; Tyler 1978) and acts as a key regulator of the ntr
system (Magasanik 1982; Merrick 1988) . Recently, the regulatory effect of glutamine on
Amt was studied in glnL mutants of E. coli which constitutively express Amt (Castroph
and Kleiner 1984; Jayakumar et al 1987). These results showed that Amt was regulated
by the internal glutamine pool via feedback inhibition. Both CH
3
NH
3
+
and NH
4
+
transport
of many bacterial strains is strongly inhibited by methionine sulfoximine, a potent inhibitor
of GS, indicating the existence of a regulatory glutamine binding site at the carrier (Kleiner
1985). Recent cloning and sequencing of the E. coli amtA gene, which codes for one of
the soluble components of the ammonium transport system, should lead to rapid progress
in the field (Fabiny et al 1991).
However, repeated attempts in our laboratory to demonstrate active transport of NH
4
+
using
14
CH
3
NH
3
+
as the probe have been unsuccessful in Ruminococcus flavefaciens
FD-1. These experiments involved cells grown under N-limitation from both batch and
chemostat grown cultures (P.A. Duncan and R.I. Mackie, unpublished results). Valida-
tion of the transport assay showed that R. flavefaciens exhibited energy dependent,
concentrative uptake of
14
C-labelled 2-deoxyglucose. Further experiments with
14
CH
3
NH
3
+
showed that E. coli actively transported this ammonium probe. We have
had most success monitoring NH
4
+
depletion from the external medium over time in
30-min assays with R. flavefaciens but this technique lacks sensitivity and is unsuitable
for determining bioenergetic aspects of ammonium transport. However, initial rates of
ammonia depletion were clearly different between cells grown under ammonium-lim-
ited as compared to carbon-limited growth (11.2 versus 0.8 nmol/min/mg cell protein,
respectively). This is consistent with the higher specific activity or GDH and GS found
for this organism when grown under ammonium limitation.
Biotechnological problems and prospects
Protein and peptide catabolism
Protein and peptide metabolism have been studied extensively and proteolytic enzymes
from bacteria, fungi and protozoa have been demonstrated. However, our knowledge
and understanding of the biochemistry and regulation of proteinase, peptidase and
deaminase enzymes which are involved in the sequential catabolism of protein in the
rumen is superficial. This process is complicated by the wide distribution of these
activities between the different microbial groups and also that activities are predomi-
nantly cell associated. Much attention has been focussed on the association of rumen
36 R.I. Mackie and M. Morrison
microbes with plant surfaces and cereal grains mainly from the standpoint of fibre
degradation or carbohydrate utilisation but not from the protein utilisation and degra-
dation standpoint. Similarly, our understanding of the molecular basis for recognition
and adhesion to specific sites remains largely unknown although these studies have been
initiated in relation to cellulose binding by fibre degrading bacteria. This microbial
association with surfaces is a logical ecological adaptation and could possibly be
exploited to manipulate the kinetics of protein degradation in the rumen.
Quantitative estimates of intraruminal recycling indicate that as much as 30 to 50% of
the daily N intake may be recycled. These values may be underestimates since turnover
of peptide-N independent of the NH
3
-N pool may not be accounted for by this technique.
Further studies of this nature are warranted under a variety of feeding conditions which
are known to provide different ruminal microbial populations. An obvious target is
protozoal activity which is thought to be a major contributing factor to intraruminal N
recycling. Defaunation invariably results in improved protein flow from the rumen
although most protocols for this procedure are suitable experimentally but not practically.
Novel approaches to control protozoal activity using the ?smugglin? concept, secretion of
antiprotozoal antibodies or delivery of such compounds by bacteria specifically engi-
neered for this purpose using recombinant DNAtechnology are potential methods. Further
ecological studies which describe bacterial and fungal population changes accompanying
defaunation are an integral component of this approach.
It has long been recognised that a diverse population of bacteriophage is maintained
in the rumen and some research has been carried out on this neglected group of microbes.
However, our knowledge of their biology and molecular bases of infection, lysogeny
and lysis are limited. Studies on quantitating phage DNA to estimate fluctuations in
phage populations will provide useful information relating to intraruminal nitrogen
recycling and also to bacterial population changes associated with phage blooms.
Bacteria also produce a range of antibacterial compounds which are active in the
gastrointestinal tract. The first report of a bacteriocin-like compound produced by
Ruminococcus albus 8 has recently been published. Documentation of these intimate
relationships is dependent on utilisation of modern molecular ecology techniques.
Although methods exist for controlling protein degradation, most of these strategies
are applicable to dietary supplements and intensive production systems but probably
have little current application to extensive, forage-based production systems. Ionopho-
res such as monensin increase peptide and amino-N flow from the rumen but its
inhibitory spectrum is probably too broad for application in grazing and forage fed
ruminants.  Pretreatment  of  protein sources, and more recently  amino acids,  have
provided improvements in protein supply to the animal but are expensive and require
increased management inputs. The interaction between dietary tannins, particularly in
browse species, and protein requires further study. Most improvements with high tannin
levels have been documented in temperate, highly degradable forages where protein
degradation is retarded by this association.
Future research efforts should focus on highly selective and specific methods for
inhibiting microorganisms responsible for proteinase, peptidase and deaminase activity
in the rumen.
Nitrogen metabolism in the rumen: biotechnological problems and prospects 37
Conclusions
Two quotes from pioneers in the field of rumen microbial ecology and metabolism are
still pertinent today. In 1960, Hungate proposed that an analysis of an ecological habit
required elaboration of the numbers and kinds of microorganisms present, an analysis
of the activities of these microbes (in vitro), and finally examination of the extent to
which these activities are expressed (in vivo). Allison (1970) concluded his review on
nitrogen metabolism of ruminal microorganisms as follows: ?Exciting possibilities for
improving the nitrogen economy of the ruminant may be ahead if measures for
protecting protein and amino acids from degradation in the rumen are perfected. It may
be, however, that more will be done towards alleviation of the protein shortages in the
world by  imaginative exploitation of the ability of ruminants to employ ruminal
microbes for synthesis of good quality protein from non-protein nitrogen.?
Modern analytical, biochemical and molecular techniques applied at both the organ-
ismal and whole animal level will afford scientists an excellent opportunity to provide
the first complete description of this system and not only an extension of our knowledge
and understanding. Further improvements in ruminant nitrogen metabolism will require
innovative approaches to manipulate the degradative and bisynthetic activities of the
rumen microorganisms.
References
Allison M.J. 1970. Nitrogen metabolism in rumen micro-organisms. In: Phillipson At. T. et al
(eds), Physiology of Digestion and Metabolism in the Ruminant. Oriel Press, Newcastle upon
Tyne, UK. pp. 456?473.
Allison M.J., Robinson I.M., Bucklin J.A. and Booth G.D. 1979. Comparison of bacterial
populations of the pig cecum and colon based upon enumeration with specific energy sources.
Applied and Environmental Microbiology 37:1142?1151.
Bender R.A. 1991. The role of the NAC protein in the nitrogen regulation of Klebsiella
aerogenes. Molecular Microbiology 5:2575?2580.
Blackburn T.H. 1968. Protease production by Bacteroides amylophilus H18. Journal of General
Microbiology 53:27?36.
Blackburn T.H. and Hobson P.N. 1960. Proteolysis in the sheep rumen by whole and fractionated
rumen contents. Journal of General Microbiology 22:272?281.
Boussiba S., Dilling W. and Gibson J. 1984. Methylammonium transport in Anacystis nidulans
R-2. Journal of Bacteriology 160:204?210.
Britton K.L, Baker P.J., Rice D.W. and Stillman T.J. 1992. Structural relationship between the
hexameric and tetrameric family of glutamate dehydrogenases. European Journal of Bio-
chemistry 209:851?859.
Brock F.M., Forsberg C.W. and Buchanan-Smith J.G. 1982 Proteolytic activity of rumen
microorganisms and effects of proteinase inhibitors. Applied and Environmental Microbiol-
ogy 44:561?569.
Brown C.M.M. 1980. Ammonia assimilation and utilization in bacteria and fungi. In: Payne
J.W. (ed), Microorganisms and Nitrogen Sources. John Wiley, Chichester, UK. pp. 511?535.
Bryant M.P. and Robinson I.M. 1961. Studies on the nitrogen requirements of some ruminal
cellulolytic bacteria. Applied Microbiology 9:96?103.
38 R.I. Mackie and M. Morrison
Bryant M.P. and Robinson I.M. 1962. Some nutritional characteristics of predominant culturable
ruminal bacteria. Journal of Bacteriology 84:822?828.
Bryant M.P. and Robinson I.M. 1963. Apparent incorporation of ammonia and amino acid carbon
during growth of selected species of rumen bacteria. Journal of Dairy Science 46:150?154.
Castroph H. and Kleiner D. 1984. Some properties of a Klebsiella pneumoniae ammonium
transport negative mutant (Amt). Archives of Microbiology 139:245?247.
Chalupa W., Clark J., Opliger P. and Larker R. 1970. Ammonia metabolism in rumen bacteria
and mucosa from sheep fed soy protein or urea. Journal of Nutrition 100:161?169.
Coleman G.S. 1979. Rumen ciliate protozoa. In: Biochemistry and Physiology of Protozoa.
Second edition. Vol. 2. Academic Press, New York, USA. pp. 381?408.
Cotta M.A. and Hespell R.B. 1986. Proteolytic activity of the ruminal bacterium Butyrivibrio
fibrisolvens. Applied and Environmental Microbiology 52:51?58.
Dalton H. 1979. Utilization of inorganic nitrogen by microbial cells. International Review of
Biochemistry 21:227?266.
Duncan P.A. and Mackie R.I. 1989. Activities of ammonia assimilatory enzymes of Rumino-
coccus flavefaciens FD-1. Abstracts of Conference on Rumen Function 20:55.
Duncan P.A., White B.A. and Mackie R.I. 1991. Partial characterization of ammonia assimilatory
enzymes of Ruminococcus flavefaciens FD-1. Abstracts of ASM Meeting, Dallas, Texas, USA.
Duncan P.A., White B.A. and Mackie R.I. 1992. Purification and properties of NADP-dependent
glutamate dehydrogenase from Ruminococcus flavefaciens FD-1. Applied and Environ-
mental Microbiology 58:4032?4037.
Erfle J.D., Sauer F.D. and Mahadevan S. 1977. Effect of ammonia concentration on activity of
enzymes of ammonia assimilation and on synthesis of amino acids by mixed rumen bacteria
in continuous culture. Journal of Dairy Science 60:1064?1072.
Fabiny J.M., Jayakumar A., Chinault A.C. and Barnes E.M. 1991. Ammonium transport in
Escherichia coli: localization and nucleotide sequence of the Amt A gene. Journal of General
Microbiology 137:983?989.
Forsberg C.W.,  Lovelock  L.K.A.,  Krumholz L.  and Buchanan-Smith J.G. 1984. Protease
activities of rumen protozoa. Applied and Environmental Microbiology 47:101?110.
Gilchrist F.M.C. and Mackie R.I. 1984. Herbivore Nutrition in the Subtropics and Tropics. The
Science Press, Craighall, South Africa. 779 pp.
Glass T.L. and Hylemon P.B. 1980. Characterization of a pyridine nucleotide-nonspecific
glutamate dehydrogenase from Bacteroides thetaiotaomicron. Journal of Bacterioloy
141:1320?1330.
Hazlewood G.P. and Edwards R. 1981. Proteolytic activities of a rumen bacterium, Bacteroides
ruminicola R8/4. Journal of General Microbiology 125:11?15.
Hazlewood G.P., Jones G.A. and Mangan J.L. 1981. Hydrolysis of leaf fraction 1 protein by the
proteolytic rumen bacteria Bacteroides ruminicola R8/4. Journal of General Microbiology
123:223?232.
Herbeck J.L. and Bryant M.P. 1974. Nutritional features of the intestinal anaerobe Ruminococ-
cus bromii. Applied Microbiology 28:1018?1022.
Hespell R.B. 1984. Influence of ammonia assimilation pathways and survival strategy on rumen
microbial growth. In: Gilchrist F.M.C. and Mackie R.I. (eds), Herbivore Nutrition in the
Subtropics and Tropics. The Science Press, Craighall, South Africa. pp. 346?358.
Hungate R.E. 1960. Microbial ecology of the rumen. Bact. Rev. 24:353?364.
Jayakumar A., Hong J.-S. and Barnes E.M. 1987. Feedback inhibition of ammonium (methy-
lammonium) on transport in Escherichia coli by glutamine and glutamine analogs. Journal
of Bacteriology 169:553?557.
Nitrogen metabolism in the rumen: biotechnological problems and prospects 39
Jenkinson H.F., Buttery P.J. and Lewis D. 1979. Assimilation of ammonia by Bacteroides
amylophilus in chemostat cultures. Journal of General Microbiology 113:305?313.
Joyner A.E. and Baldwin R.L. 1966. Enzymatic studies of pure cultures of rumen microorgan-
isms. Journal of Bacteriology 92:1321?1330.
Kistner A. and Kotze J.P. 1973. Enzymes of intermediary metabolism of Butyrivibrio fibrisol-
vens and Ruminococcus albus grown under glucose limitation. Canadian Journal of Micro-
biology 19:1119?1127.
Kleiner D. 1976. Ammonium uptake and metabolism by nitrogen-fixing bacteria. II. Klebsiella
pneumoniae. Archives of Microbiology 111:85?91.
Kleiner D. 1979. Regulation of ammonium uptake and metabolism by nitrogen fixing bacteria.
III. Clostridium pasteurianum. Archives of Microbiology 120:263?270.
Kleiner D. 1982. Ammonium (methylammonium) transport by Klebsiella pneumoniae. Bio-
chimica et Biophysica Acta 688:702?708.
Kleiner D. 1985. Bacterial ammonium transport. FEMS Microbiology Reviews 32:87?100.
Kleiner D., Phillips S. and Fitzke E. 1981. Pathways and regulatory aspects of N
2
and NH
4
+
assimilation in N
2
-fixing bacteria. In: Bothe H. and Trebst A. (eds), Biology of Inorganic
Nitrogen and Sulfur. Springer-Verlag, Berlin, Germany. pp. 131?140.
Kleinschmidt J.A. and Kleiner D. 1981. Relationships between nitrogenase, glutamine syn-
thetase, glutamine and energy changes in Azotobacter vinelandii. Archives of Microbiology
128:412?415.
Kopecny J. and Wallace R.J. 1982. Cellular location and some properties of proteolytic enzymes
of rumen bacteria. Applied and Environmental Microbiology 43:1026?1033.
Kustu S., Sei K. and Keener J. 1986. Nitrogen regulation in enteric bacteria. In: Booth I.R. and
Higgins C.R. (eds), Regulation of Gene Expression. Cambridge University Press, Cambridge,
UK. pp. 139?154.
Leng R.A. and Nolan J.V. 1984. Nitrogen metabolism in the rumen. Journal of Dairy Science
67:1072?1089.
Mackie R.I. 1982. Proteolytic activity and protein degradation in the rumen. In: Schwartz H.M.
(ed), Nitrogen Metabolism in the Ruminant. National Chemical Research Laboratory, Coun-
cil for Scientific and Industrial Research, Pretoria, South Africa. pp. 33?57.
Mackie R.I. and Kistner A. 1985. Some frontiers of research in basic ruminant nutrition. South
African Journal of Animal Science 15:72?85.
Mackie R.I. and White B.A. 1990. Recent advances in rumen microbial ecology and metabo-
lism: potential impact on nutrient output. Journal of Dairy Science 73:2971?2995.
MacRae J.C. and Lobley G.E. 1986. Interactions between energy and protein. In: Milligan L.P.,
GrovumW.L.andDobsonA.(eds),Control of Digestion and Metabolism in Ruminants.
Prentice Hall, New Jersey, USA. pp. 367?385.
Maculuso A., Best E.A. and Bender R.A. 1990. Role of the nac gene product in the nitrogen
regulation of some NTR-regulated operons of Klebsiella aerogenes. Journal of Bacteriology
172:7249?7255.
Maczulak A.E., Dawson K.A. and Baker J.P. 1985. Nitrogen utilization in bacterial isolates from
the equine cecum. Applied and Environmental Microbiology 50:1439?1443.
Magasanik B. 1982. Genetic control of nitrogen assimilation in bacteria. Annual Review of
Genetics 16:135?168.
Magasanik B. 1988. Reversible phosphorylation of an enhancer binding protein regulates the
transcription of bacterial nitrogen utilization genes. Trends in Biochemical Sciences 13:475?479.
Mathison G.W. and Milligan L.P. 1971. Nitrogen metabolism in sheep. British Journal of
Nutrition 25:351?366.
40 R.I. Mackie and M. Morrison
Meers J.L., Tempest D.W. and Brown G.M. 1970. Glutamine (amide): oxoglutarate amino
transferase (NADP), an enzyme involved in the synthesis of glutamate by some bacteria.
Journal of General Microbiology 64:187?194.
Merrick M.J. 1988. Regulation of nitrogen assimilation by bacteria. In: Cole J.A. and Ferguson
S.J. (eds), The Nitrogen and Sulphur Cycles. Cambridge University Press, Cambridge, UK.
pp. 331?361.
Nolan J.V. and Leng R.A. 1972. Dynamic aspects of ammonia and urea metabolism in sheep.
British Journal of Nutrition 27:177?194.
Nugent J.H.A. and Mangan J.L. 1981. Characteristics of the rumen proteolysis of protein 1 (18S)
leaf protein from lucerne (Medicago sativa L.). British Journal of Nutrition 46:39?58.
Patterson J.A. and Hespell R.B. 1985. Glutamine synthetase activity in the ruminal bacterium
Succinivibrio dextrinosolvens. Applied and Environmental Microbiology 50:1014?1020.
Pettipher G.L. and Latham M.J. 1979. Characteristics of enzymes produced by Ruminococcus
flavefaciens which degrade plant cell walls. Journal of General Microbiology 110:21?27.
Pilgrim A.F., Gray F.V., Weller R.A. and Belling G.B. 1970. Synthesis of microbial protein from
ammonia in the sheep?s rumen and the proportion of dietary nitrogen converted into microbial
nitrogen. British Journal of Nutrition 24:589?598.
Pittman K.A. and Bryant M.P.  1964.  Peptides and other nitrogen sources for growth of
Bacteroides ruminicola. Journal of Bacteriology 88:401?410.
Pittman K.A., Lakshmanan S. and Bryant M.P. 1967. Oligopeptide uptake by Bacteroides
ruminicola. Journal of Bacteriology 93:1499?1508.
Prins R.A., van Rheenen D.L. and van?t Klooste A.T. 1983. Characterization of microbial
proteolytic enzymes in the rumen. Antonie van Leeuwenhoek Journal of Microbiology
49:585?595.
Robinson I.M., Allison M.J. and Bucklin J.A. 1981. Characterization of the cecal bacteria of
the pig. Applied and Environmental Microbiology 41:950?955.
Russell J.B. 1983. Fermentation of peptides by Bacteroides ruminicola B
1
4. Applied and
Environmental Microbiology 45:1566?1574.
Russell J.B. and Baldwin R.L. 1978. Substrate preferences in rumen bacteria: evidence of
catabolite regulatory mechanisms. Applied and Environmental Microbiology 36:319?329.
Russell J.B., Bottje and Cotta M.A. 1981. Degradation of protein by mixed cultures of rumen
bacteria. Identification of Streptococcus bovis as an actively proteolytic rumen bacterium.
JournalofAnimalScience53:242?252.
Russell J.B., Strobel H.J. and Martin S.A. 1990. Strategies of nutrient transport by ruminal
bacteria. Journal of Dairy Science 73:2996?3012.
Satter L.D. and Slyter L.L. 1974. Effect of ammonia concentration on rumen microbial protein
production in vitro. British Journal of Nutrition 32:199?208.
Schaefer D.M., Davis C.L. and Bryant M.P. 1980. Ammonia saturation constants for predomi-
nant species of rumen bacteria. Journal of Dairy Science 63:1248?1263.
Servin-Gonzalez L. and Bastarrachea F. 1984. Nitrogen regulation of synthesis of the high
affinity methylammonium transport system of Escherichia coli. Journal of General Micro-
biology 130:3071?3077.
Smith C.J., Hespell R.B. and Bryant M.P. 1980. Ammonia assimilation and glutamate formation
in the anaerobe Selenomonas ruminantium. Journal of Bacteriology 141:593?602.
Smith C.J., Hespell R.B. and Bryant M.P. 1981. Regulation of urease and ammonia assimilatory
enzymes in Selenomonas ruminantium. Applied and Environmental Microbiology 42:89?96.
Smith R.H. 1979. Synthesis of microbial nitrogen compounds in the rumen and their subsequent
digestion. JournalofAnimalScience49:1604?1614.
Nitrogen metabolism in the rumen: biotechnological problems and prospects 41
Southern J.A., Parker J.A. and Woods D.R. 1986. Expression and purification of glutamine
synthetase cloned from Bacteroides   fragilis. Journal   of   General Microbiology
132:2827?2835.
Streicher S.L. and Tyler B. 1980. Purification of glutamine synthetase from a variety of bacteria.
Journal of Bacteriology 142:69?78.
Strydom E., Mackie R.I. and Woods D.R. 1986. Detention and characterization of extracellular
proteases in Butyrivibrio  fibrisolvens H17c. Applied Microbiology and Biotechnology
24:214?217.
Takahashi M., Benno Y. and Mitsuoka T. 1980. Utilization of ammonia nitrogen by intestinal
bacteria from pigs. Applied and Environmental Microbiology 39:30?35.
Tamminga S. 1979. Protein degradation in the forestomachs of ruminants. Journal of Animal
Science 49:1615?1630.
Tamminga S. and Verstegen M.W.A. 1992. Implications of nutrition of animals on environ-
mental pollution. In: Garnsworthy P.C., Haresign W. and Cole D.J.A. (eds), Recent Advances
in Animal Nutrition. Butterworth-Heinmann, Oxford, UK. pp. 113?130.
Tyler B. 1978. Regulation of the  assimilation of  nitrogen compounds. Annual Review of
Biochemistry 47:1127?1162.
Usdin K.P., Zappe H., Jones D.T. and Woods D.R. 1986. Cloning, expression, and purification
of glutamine synthetase from Clostridium acetobutylicum.  Applied  and Environmental
Microbiology 52:413?419.
Varel V.H. and Bryant M.P. 1974. Nutritional features of Bacteroides fragilis subsp. fragilis.
Applied Microbiology 18:251?257.
Wallace R.J. 1983. In vitro studies of potential inhibitors of protein degradation by rumen
bacteria. In: Arnal M., Pion R. and Bomin D. (eds), Protein Metabolism and Nutrition. Vo l .
2. INRA (Institut National de Recherche Agronomique), Paris, France. pp. 219?222.
Wallace R.J. and Brammall M.L. 1985. The role of different species of bacteria in the hydrolysis
of protein in the rumen. Journal of General Microbiology 131:821?832.
Wallace R.J. and Joblin K.N. 1985. Proteolytic activity of a rumen anaerobic fungus. FEMS
Microbiology Letters 29:19?25.
Westlake K. and Mackie R.I. 1990. Peptide and amino acid transport in Streptococcus bovis.
Applied Microbiology and Biotechnology 34:97?102.
Westlake K., Mackie R.I. and de Beer G. 1987. Peptide transport in Streptococcus bovis. In:
Rose M.J. (ed), Herbivore Nutrition Research. Australian Society of Animal Production,
Brisbane, Australia. pp. 119?120.
Wohlt J.E., Clark J.H. and Blaisdell F.S. 1976. Effect of sampling location, time and method of
concentration of ammonia nitrogen in rumen fluid. Journal of Dairy Science 59:459?464.
Wozny M.A., Bryant M.P., Holdeman L.V. and Moore W.E.C. 1977. Urease assay and urease-
producing species of anaerobes in the bovine rumen and human feces. Applied and Environ-
mental Microbiology 33:1097?1104.
Wright D.E. 1967. Metabolism of peptides by rumen microorganisms. Applied Microbiology
15:547?550.
42 R.I. Mackie and M. Morrison
Antimicrobial factors in African
multipurpose trees
S.M. El Hassan
1,2
, A. Lahlou-Kassi
1
,C.J.Newbold
2
and R.J. Wallace
2
1. International Livestock Research Institute, P.O. Box 5689, Addis Ababa, Ethiopia
2. Rowett Research Institute, Bucksburn, Aberdeen AB2 9SB, UK
Abstract
Samples and extracts of foliage from African multipurpose trees (MPT) were
screened for their effects on rumen protozoa and bacteria. The MPT species
tested were Acacia aneura, Acacia angustissima, Chamaecytisus palmensis,
Brachychiton populneum, Flindersia maculosa, Sesbania sesban, Leucaena
leucocephala and Vernonia amyedalina. S. sesban was highly toxic to rumen
protozoa from sheep reared in Aberdeen. The toxic factor was associated with
the saponin-containing fraction of the plant. S. sesban may therefore be
useful in suppressing protozoa and thereby improving protein flow from the
rumen; however, protozoa from the rumen of sheep raised in Ethiopia were
resistant to S. sesban, indicating that resistance, either endogenous or ac-
quired, may be a problem in exploiting this antiprotozoal effect. The Acacia
species were toxic to rumen bacteria, particularly the cellulolytic species of
Ruminococcus. The cause of this toxicity was not identified.
Introduction
In order to meet the increasing demand for animal products in Africa, it is necessary to
enhance the utilisation of available fodder and agricultural byproducts and also to look
for alternative feed resources. Multipurpose trees (MPT) and shrubs may help in this
respect and in addition may provide other benefits, in terms of increasing soil fertility
by nitrogen fixation (Woodward and Reed 1989), providing green manure and mulch
(McKell 1980), protecting soil from erosion by stabilising surface soil, control of desert
encroachment (McKell 1980; Woodward and Reed 1989) and providing fuel and shelter
(McKell 1980; Woodward 1988). MPT are a rich source of protein, carotenoid precur-
sors of vitamin A and many minerals, but generally have a low phosphorus content (Le
Hou?rou 1980; Kumar and Vaithiyanathon 1990; Olasson and Welin-Berger 1991). In
many areas of Africa they provide an available supplement to grasses in the dry season
(Le  Hou?rou  1980), lessening the dependence on expensive protein  supplements.
However, the presence of antinutritional compounds in MPT limits their utilisation as
feedstuffs.
Plants produce a large number of chemicals, arbitrarily categorised as primary or
secondary metabolites, that are essentially defence mechanisms for the plant against
animal and insect predators and microbial infection (Deshhande et al 1986; Feeny
1970). Several deleterious compounds have been identified in forages and fodder
trees, including alkaloids, non-protein amino acids, mycotoxins, terpenoids, steroids,
lectins and protease inhibitors (Cheeke and Shull 1985; Lowry 1990; D?Mello 1992).
Many of the African browse species contain high levels of polyphenolics and insoluble
condensed tannins (proanthocyanidins) that bind to neutral detergent fibre (El Hassan
1994;  Reed  1986; Reed et  al 1990; Woodward and Reed 1989).  Some phenolic
compounds are toxic to rumen bacteria (Akin 1982; Borneman et al 1986; Chesson et
al 1982; Martin and Akin 1988), rumen fungi (Akin and Rigsby 1987) and rumen
protozoa (Akin 1982).
Oxalate and the alkaloid perloline inhibited cellulose degradation by the rumen
microbial population (Hemken et al 1984; James et al 1967). Thus MPT may have
components that are toxic to the micro-organisms of the rumen as well as to the host
animal.
The aim of the work described here was to determine if seven potentially useful MPT
contained factors that were antimicrobial and therefore detrimental to rumen fermenta-
tion. One particular MPT, A. aneura, proved toxic to rumen bacteria and therefore
inhibitory; however, S. sesban contained an antiprotozoal factor that may be useful in
enhancing rumen productivity by suppressing wasteful breakdown of bacterial protein
by rumen protozoa.
Antiprotozoal effects of the foliage
of multipurpose trees
Protozoa are responsible for over 90% of bacterial protein turnover in the rumen
(Wallace and McPherson 1987). Thus they cause a major decrease in the microbial
protein flowing from the rumen (Weller and Pilgrim 1974). The degradation of labelled
bacterial protein by protozoa can be used as a measure of protozoal activity (Wallace
and McPherson 1987). This method was used here as an assay of the antiprotozoal
activity of MPT, using samples of rumen fluid taken from sheep in Aberdeen, UK, which
received a mixed grass hay/concentrate diet, and from sheep in Debre Zeit, Ethiopia,
receiving teff straw.
Effects of MPT on protozoa from sheep in Aberdeen
The breakdown of the rumen bacterium Selenomonas ruminantium, in the presence
and absence of added MPT, was measured in rumen fluid taken from sheep in Aberdeen
(Table 1). The rate of breakdown of S. ruminantium was measured on different samples
of rumen fluid, each with a slightly different protozoal population, thus accounting for
the variation in the rate of breakdown in the absence of MPT, which ranged from 7.3
to 13.3% h
?1
. All of the MPT caused a progressive decrease in the observed rate of
release of
14
C-leucine from S. ruminantium, most likely due to a physical interference
with the predatory activity of the protozoa as well as a degree of toxicity. However,
the effect with S. sesban was particularly pronounced, causing a 59% decrease at 6 mg
ml
?1
and complete abolition of predatory activity at 24 mg ml
?1
(Table 1).
44 S.M. El Hassan, A. Lahlou-Kassi, C.J. Newbold and R.J. Wallace
Identification of the active component of S. sesban
All of the MPT species contained phenolics (El Hassan 1994), so the possibility that
tannins or phenolics were responsible for the antiprotozoal effect of S. sesban was tested.
Tannins can be removed from solution by polyethylene glycol or polyvinyl pyrrolid-
ine (PVP; Garrido et al 1991). When PVP was added to the incubation containing S.
sesban and
14
C-labelled S. ruminantium, the precipitation of tannins had no effect on
the antiprotozoal property of S. sesban (Figure 1). Thus it was concluded that tannins
were not responsible for the observed antiprotozoal property of S. sesban.
Butanol extraction of plant material causes the removal of saponins (Headon et al
1991; Wall et al 1952). When an aqueous extract of S. sesban was extracted with
n-butanol and the butanol extract and remaining aqueous phase were tested for their
effects on the degradation of S. ruminantium, the antiprotozoal activity was removed to
the butanol phase (Figure 2; Newbold et al 1994). The remaining aqueous extract had
no effect (Figure 2).
TheeffectsofS. sesban were compared in rumen liquor obtained from sheep in
Aberdeen, UK, and rumen liquor collected from sheep at ILRI?s Debre Zeit Research
Station, Ethiopia. It is not known if the latter group had previously been exposed to S.
sesban. As before, the degradation of
14
C-labelled S. ruminantium by rumen protozoa
from Aberdeen sheep was completely inhibited in the presence of S. sesban, but S.
sesban had no effect on the activity of the lower numbers of protozoa in the Ethiopian
sheep (Figure 3). Either the protozoa in Ethiopian sheep were intrinsically resistant to
the toxic material, or the bacterial population had adapted to metabolise the material
and thus remove its toxicity.
It can be concluded that the component of S. sesban toxic to rumen protozoa is butanol
extractable and therefore likely to be part of the saponin fraction. Wallace et al (1994)
showed that other sources of saponins were toxic, but less so than extract from S. sesban,
suggesting that it is a specific type of saponin or saponin-like substance in S. sesban
that is responsible for antiprotozoal effect. The observation that saponins are toxic to
Table 1. Influence of MPT on breakdown of
14
C-labelled S. ruminantium in rumen liquor taken from sheep
in Aberdeen.
Rate of degradation (% h
?1
) at inclusion level of
?
MPT 0mgml
?1
6mgml
?1
12 mg ml
?1
24 mg ml
?1
SED
Acacia aneura 8.6
a
7.5
ab
6.8
ab
4.6
b
1.80
Brachychiton populneum 11.7
a
10.1
ab
8.8
bc
7.5
c
0.96
Chamaecytisus palmensis 7.3
a
6.4
ab
5.6
bc
4.3
c
0.72
Flindersia maculosa 13.3
a
12.7
a
11.6
a
9.4
b
0.84
Leucaena leucocephala 7.8
a
6.9
a
6.3
a
3.7
b
0.90
Sesbania sesban 11.4
a
4.7
b
0.2
c
0
c
1.14
Vernonia amyedalina 8.6
a
7.6
a
6.2
a
3.9
b
1.03
?
The balance was made up by wheat straw to give a total substrate concentration of 40 mg ml
?1
.
a, b, c
Values with different superscripts are different (P < 0.05).
Antimicrobial factors in African multipurpose trees 45
rumen protozoa is consistent with earlier studies in dairy cows (Valdez et al 1986).
Recent results from Colombia indicate that the antiprotozoal effects of other plants,
including Sapindus saponaria, are similarly due to saponins (Navas-Camacho et al
1994). Other antiprotozoal factors in plants have been known for a long time. Eadie et
al (1956) showed that certain terpenes and other substances present in plant material
had marked toxic properties toward rumen protozoa. Warner (1962) found that minor
plant constituents such as terpenes or alkaloids may have specific effects on individual
species of rumen micro-organisms. It is unlikely that terpenes or alkaloids would be
extracted into butanol by the method used here. Akin (1982) observed that p-coumaric
acid reduced the motility of entodiniomorphid protozoa, but again it is unlikely that
phenolic acids would be extracted from S. sesban into butanol.
Thus the indications are that it is the saponin fraction of S. sesban that is toxic to
rumen protozoa, although more work is required to identify the precise component.
Selective elimination of protozoa, if it were to persist, would enhance the flow of
microbial protein from the rumen and improve the nutrition of the animal. Defaunation
has been reported to increase microbial outflow (Leng et al 1981) and increase the
efficiency of feed utilisation and hence growth rate of cattle on diets low in bypass
Figure 1. Influence of removal of tannins from S. sesban on breakdown of Selenomonas ruminantium by
rumen protozoa. The rumen bacterium, S. ruminantium, was labelled with
14
C-leucine and incubated with
rumen liquor removed from sheep in Aberdeen as described previously (Wallace and McPherson 1987). The
radioactivity released into acid-soluble material gives a measure of protozoal activity. S. sesban was present
at a concentration of 20 mg ml
?1
. PVP, which precipitates tannins, was added to the incubation mixture at
a concentration of 20 mg ml
?1
.
01234
Incubation time (h)
0
5
10
15
20
25
30
35
Release
of radioactiv
ity
(%)
S. ruminantium + S. sesban
S. ruminantium + S. sesban + PVP
S. ruminantium
46 S.M. El Hassan, A. Lahlou-Kassi, C.J. Newbold and R.J. Wallace
protein (Bird and Leng 1978). Similarly, the rate of wool growth was increased in
defaunated lambs (Bird et al 1979). Provided the loss of protozoa does not compromise
fibre breakdown, which can occur under some circumstances (Ushida et al 1991), the
antiprotozoal effects of some MPT maybe useful in enhancing feed utilisation over and
above the nutrient content of the feed. There may even be scope for exporting appro-
priate plant extracts for use in other countries. Possible problems in the use of these
materials must be investigated. Indications are that toxicity to the host animal will not
be a problem (El Hassan 1994; Navas-Camacho et al 1994), but more work is required.
The finding that Ethiopian sheep were resistant may mean that the method has limited
applicability in regions where the MPT are indigenous species. It also has implications
for their use elsewhere, depending on the nature and time-scale of the development of
resistance. These factors should be investigated.
Antibacterial effects of the foliage
of multipurpose trees
In contrast to rumen protozoa, there is no assay that can be applied specifically to
rumen bacteria. To give a broad view of the effects of MPT on mixed rumen bacteria,
Figure 2. Influence of butanol extraction of an aqueous extract of S. sesban on its inhibition of the
breakdown of Selenomonas ruminantium by rumen protozoa. Protozoal activity was measured as described
in Figure 1.
01234
Incubation time (h)
0
2
4
6
8
10
12
14
16
18
Release
of
radioactiv
ity
(%)
Control
S. sesban
Butanol extract
Aqueous extract
Antimicrobial factors in African multipurpose trees 47
rumen liquor from which protozoa had been removed by centrifugation was assayed
for adenosine 5?-triphosphate (ATP) as an indication of live biomass (Wallace and West
1982). Individual species of rumen bacteria were then examined to determine if the
toxicity of MPT was selective with respect to different species of bacteria.
Influence of MPT on ATP concentrations
in mixed rumen bacteria
The effect of MPT on the mixed bacterial population were investigated in protozoa-free
rumen liquor (PFRL) prepared from sheep in Aberdeen. ATP was analysed in acid
extracts as described by Wallace and West (1982) after 4 and 24 h incubation in vitro of
a mixture of PFRL with MPT and wheat straw. The ATP pools of the different samples
of PFRL without addition of MPT were variable and tended to decline between 4 and
24 h, presumably reflecting some cell death over the longer incubation (Table 2). Toxic
effects of MPT were seen at 24 h, but not 4 h. Indeed, C. palmensis increased ATP at 4
h, most likely because it contained fermentable substrate, giving rise to a greater
bacterial biomass. The only evidence of toxicity was after 24 h, when the ATP pool was
significantly decreased by A. aneura, S. sesban and V. amyedalina (Table 2).
Figure 3. Comparison of the effects of S. sesban on rumen protozoa from rumen fluid taken from sheep in
Aberdeen and Ethiopia.
01234
0
5
10
15
20
25
30
35
Release
of radioactivity
(%)
Control, Aberdeen, UK
S. sesban, Aberdeen, UK
Control, Debre Zeit, Ethiopia
S. sesban, Debre Zeit, Ethiopia
48 S.M. El Hassan, A. Lahlou-Kassi, C.J. Newbold and R.J. Wallace
These effects of MPT on bacterial ATP are indicative only, because they represent a
balance between ATP increases caused by fermentable materials in the MPT and
decreases resulting from toxicity. Nevertheless, some indication of toxicity was ob-
tained, which was greatest with A. aneura.
Influence of MPT on growth of individual
species of rumen bacteria
The influence of MPT on the growth of nine major species of rumen bacteria was
examined by adding aqueous extracts of the plants to pure cultures of bacteria and
determining the effects on growth by monitoring optical density. Only two examples
are given here, illustrating typical differences between effects observed with cellulolytic
and non-cellulolytic species. Full details are given elsewhere (El Hassan 1994).
A. angustissima and L. leucocephala were most toxic to non-cellulolytic bacteria,
with the other species having little or no effect; the effect was relatively minor, however,
as illustrated here with Streptococcus bovis (Figure 4). Relatively minor effects were
seen with cellulolytic bacteria as well, except for A. angustissima, which abolished
growth of Ruminococcus flavefaciens (Figure 5) and R. albus. The third major cellu-
lolytic species commonly found in the rumen, Fibrobacter succinogenes, wasmuchless
sensitive to A. angustissima, and responded more like the other non-cellulolytic species
(data not shown).
A toxic effect of A. angustissima against two of the three cellulolytic rumen bacteria
would be expected to have major consequences for fibre breakdown in ruminants
consuming this MPT. The results presented here were consistent with results obtained
with the Menke gas production technique, in which A. angustissima produced the most
Table 2. Influence of MPT on ATP content of mixed rumen bacteria incubated in vitro.
Incubation ATP (mM) at inclusion level of
?
MPT time (h) 0 mg ml
?1
6mgml
?1
12 mg ml
?1
24 mg ml
?1
SED
Acacia aneura 4 21.1 22.1 21.6 20.1 2.07
24 10.0
a
8.6
ab
6.6
b
3.7
c
1.78
Brachychiton populneum 4 34.7 32.5 31.8 30.8 10.02
24 28.2 25.8 25.7 23.9 7.74
Chamaecytisus palmensis 4 29.1
a
30.4
a
35.1
ab
41.1
b
4.55
24 21.1 18.7 17.5 16.5 2.64
Flindersia maculosa 4 35.6 34.3 33.3 36.2 8.59
24 25.2 21.2 20.4 17.2 3.78
Leucaena leucocephala 4 29.0 28.8 30.7 29.8 7.73
24 23.9 20.6 20.3 16.3 4.63
Sesbania sesban 4 24.9 28.8 29.4 27.1 3.97
24 27.1
a
21.9
b
20.4
b
15.4
c
2.14
Vernonia amyedalina 4 32.3 32.7 32.5 32.2 6.35
24 26.9
a
24.3
ab
21.9
ab
18.6
b
2.61
?
The balance was made up by wheat straw to give a total substrate concentration of 40 mg ml
?1
.
abc
Values with different superscripts are different (P < 0.05).
Antimicrobial factors in African multipurpose trees 49
pronounced effect on gas production among the MPT examined (El Hassan 1994).
Mueller-Harvey et al (1988) noted that all extracts from Ethiopian browse increased the
lag time and reduced biomass yields of S. bovis and they observed differences in the
extent of these antimicrobial effects that were consistent with the concentration of
phenolics from browse. The toxicity of phenolics has been demonstrated in a wide range
of bacteria, including cellulolytic and xylanolytic species (Akin 1982; Chesson et al
1982; Stack and Hungate 1984). However, no species specificity in the antibacterial
effects were noted previously. The cause of the inhibition of Ruminococcus species
caused by A. angustissima is not clear. They may have been due to soluble polypheno-
lics, which may contain simple phenolics and proanthocyanidins (Reed 1986). A.
angustissima contained the highest content of soluble polyphenolics among MPT
examined (El Hassan 1994). However, much work remains to be done to identify the
anti-Ruminococcus factor present in A. angustissima.
Figure 4. Influence of aqueous extracts of MPT on the growth of the non-cellulolytic rumen bacterium,
Streptococcus bovis. Foliage from the MPT (5 g) was extracted with 70% aqueous acetone and the residue
after drying resuspended in 50 ml of water, filtered, sterilised, and 0.3 ml was added to 10 ml of culture
medium. Growth was determined turbidimetrically.
50 S.M. El Hassan, A. Lahlou-Kassi, C.J. Newbold and R.J. Wallace
Conclusions and implications
Some MPT, such as S. sesban, may be useful as protein supplements and as defaunating
agents in ruminants in sub-Saharan Africa. These MPT may have some commercial
value as defaunating agents in other regions, including Europe and North America.
Other MPT had relatively minor effects on rumen micro-organisms, except for A.
angustissima, which had a specific toxic effect on Ruminococcus species. Unless a
means can be found of removing the toxic factor from A. angustissima, its use is not
recommended due to likely detrimental effects on fibre breakdown.
Figure 5. Influence of aqueous extracts of MPT on the growth of the cellulolytic rumen bacterium,
Ruminococcus flavefaciens. Foliage from the MPT (5 g) was extracted with 70% aqueous acetone and the
residue after drying resuspended in 50 ml of water, filtered, sterilised, and 0.3 ml (or 0.2 or 0.1 ml of A.
angustissima) was added to 10 ml of culture medium. Growth was determined turbidimetrically.
Antimicrobial factors in African multipurpose trees 51
References
Akin D.E. 1982. Forage cell wall degradation and p-coumaric, ferulic and sinapic acids.
Agronomy Journal 74:424.
Akin D.E. and Rigsby L.L. 1987. Mixed fungal populations and lignocellulosic tissue degrada-
tion in the bovine rumen. Applied and Environmental Microbiology 53:1987.
Bird S.H. and Leng R.A. 1978. The effect of defaunation on the rumen on the growth of cattle
on low protein high energy diets. British Journal of Nutrition 40:163.
Bird S.H., Hill M.K. and Leng R.A. 1979. The effects of defaunation on the rumen on the growth
of lambs on low-protein-high energy diets. British Journal of Nutrition 42:81.
Borneman W.S., Akin D.E. and Van Eseltine W.P. 1986. Effect of phenolic monomers on ruminal
bacteria. Applied and Environmental Microbiology 52:1331.
Cheeke P.R. and Shull L.R. 1985. Natural Toxicants in Feeds and Poisonous Plants. Avi
Publishing Co. Inc., Westport, USA.
Chesson A., Stewart C.S. and Wallace R.J. 1982. Influence of plant phenolic acids on growth
and cellulolytic activity of rumen bacteria. Applied and Environmental Microbiology 44:597.
D?Mello J.P.F. 1992. Chemical constraints to the use of tropical legumes in animal nutrition.
Animal Feed Science and Technology 38:237.
Deshpande S.S., Cheryan M. and Salunke D.K. 1986. Tannin analysis of food products. CRC
Critical Reviews in Food Science and Nutrition 24:401.
Eadie J.M., Mann S.O. and Oxford A.E. 1956. Asurvey of physically active organic infusoricidal
compounds and their soluble derivatives with special reference to their action on the rumen
microbial system. Journal of General Microbiology 14:122.
Feeny P. 1970. Seasonal changes in oak leaf tannins and nutrients as a cause of spring feeding
by winter moth caterpillars. Ecology 51:565.
Garrido A., Gomez-Cabrera A., Guerrero J.E. and Van der Meer J.M. 1991. Effect of treatment
with polyvinylpyrridone and polyethylene glycol on faba bean tannins. Animal Feed Science
and Technology 35:199.
El Hassan S.M. 1994. ?Yeast Culture and Multipurpose Fodder Trees as Feed Supplements for
Ruminants.? PhD thesis, University of Aberdeen, Aberdeen, UK.
Headon D.R., Buggle K., Nelson A. and Killeen G. 1991. Glycofractions of the yucca plant and
their role in ammonia control. In: Lyons T.P. (ed), Biotechnology in the Feed Industry:
Proceedings of Alltech?s Fifth Annual Symposium. Alltech Technical Publications, Nicholas-
ville, USA. p. 95.
Hemken R.W., Jackson J.A. and Boling J.A. 1984. Toxic factors in tall fescue. Journal of Animal
Science 58:1011.
Le Hou?rou H.N. 1980. Browse in northern Africa. In: Le Hou?rou H.N. (ed), Browse in Africa:
The Current State of Knowledge. ILCA (International Livestock Centre for Africa), Addis
Ababa, Ethiopia. pp. 55?82.
James L.F., Street J.C. and Butcher J.E. 1967. In vitro degradation of oxalate and of cellulose
by rumen ingesta from sheep fed Halogeton glomeratus. Journal of Animal Science 26:1438.
Kumar R. and Vaithiyanathon S. 1990. Occurrence, nutritional significance and effect on animal
productivity of tannins in tree leaves. Animal Feed Science and Technology 30:21.
Leng R.A., Gill M., Kempton T.J., Rowe J.B., Nolan J.V., Stachiw S.J. and Preston T.R. 1981.
Kinetics of large ciliate protozoa in the rumen of cattle given sugar cane diets. British Journal
of Nutrition 46:371.
Lowry J.B. 1990. Toxic factors and problems: methods of alleviating them in animals. In:
Devendra C. (ed), Shrubs and Tree Fodders for Farm Animals. Proceedings of a Workshop
52 S.M. El Hassan, A. Lahlou-Kassi, C.J. Newbold and R.J. Wallace
in Denpasar, Indonesia, 24?29 July 1989. IDRC (International Development Research
Centre), Ottawa, Canada. p. 76.
Martin S.A. and Akin D.E. 1988. Effect of phenolic monomers on the growth and B-glucosidase
activity of Bacteroides ruminicola and on the carboxymethyl cellulase, B-glucosidase and xy
lanase activitiesof Bacteroides succinogenes. Applied and Environmental Microbiology 54:3019.
McKell C.M. 1980. Multiple use of fodder trees and shrubs?a worldwide perspective. In: Le
Hou?rou H.N. (ed), Browse in Africa: The Current State of Knowledge. ILCA (International
Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 141?149.
Mueller-Harvey I., McAllan A.B., Theodorou M.K. and Beever D.E. 1988. Phenolics in fibrous
crop residues and plants and their effects on the digestion and utilisation of carbohydrates
and proteins in ruminants. In: Reed J.D., Capper B.S. and Neate P.J.H. (eds), Plant Breeding
and the Nutritive Value of Crop Residues. Proceedings of a Workshop Held at ILCA, Addis
Ababa, Ethiopia, 7?10 December 1987. ILCA (International Livestock Centre for Africa),
Addis Ababa, Ethiopia. pp. 97?132.
Navas-Camacho A., Laredo M.A., Cuesta A., Ortega O. and Romero M. 1994. Evaluation of
tropical trees with high or medium saponin content as dietary alternative to eliminate ciliate
protozoa from the rumen. Proceedings of the Society of Nutrition Physiology 3:204.
Newbold C.J., El Hassan S.M., Wallace R.J., Chen X-B., Goodchild A.V. and Arthaud, L. 1994.
Influence of African multipurpose trees on activity of rumen protozoa and bacteria in vitro.
Animal Production 58:461
Olasson A.C. and Welin-Berger S. 1991. The potential of local shrubs as feed for livestock and
the mineral contents of some soils and soil-licks in Central Tanzania. Two minor field studies.
Nutrition Abstracts and Reviews (Series B) 61:164.
Reed J.D. 1986. Relationship among soluble phenolics, insoluble proanthocyanidins and fibre
in East African browse species. Journal of Range Management 39:5.
Reed J.D., Soller H. and Woodward A. 1990. Fodder tree and straw diets for sheep: intake,
growth, digestibility and effect of phenolic on nitrogen utilization. Animal Feed Science and
Technology 30:39.
Stack R.J. and Hungate R.E. 1984. Effect of 3-phenylpropanoic acid on capsule and cellulases
of Ruminococcus albus. Applied and Environmental Microbiology 48:218.
Ushida K., Jouany J.P. and Demeyer D. 1991. Effects of presence or absence of rumen protozoa
on the efficiency of utilization of concentrate and fibrous feeds. In: Tsuda T., Sasaki Y. and
Kawashima R. (eds), Physiological Aspects of Digestion and Metabolism in Ruminants.
Academic Press, San Diego, USA. pp. 3?32.
Valdez F.R., Bush L.J., Goetsch A.L. and Owens F.N. 1986. Effect of steroidal sapogenins on
ruminal fermentation and on production of lactating dairy cows. Journal of Dairy Science
69:1568.
Wall M.E., Krider M.M., Rothman E.S. and Eddy C.R. 1952. Steroidal sapogenins. I. Extraction,
isolation and identification. Journal of Biological Chemistry 198:553.
Wallace R.J. and McPherson C.A. 1987. Factors affecting the rate of breakdown of bacterial
protein in rumen fluid. British Journal of Nutrition 58:313.
Wallace R.J. and West A.A. 1982. Adenosine 5? triphosphate and adenylate energy charge in
sheep digesta. Journal of Agricultural Science (Cambridge) 98:523.
Wallace R.J., Arthaud L. and Newbold C.J. 1994. Influence of Yucca shigidera extract on
ruminal ammonia concentrations and ruminal microorganisms. Applied and Environmental
Microbiology 60:1762.
Warner A.C.J. 1962. Some factors influencing the rumen microbial population. Journal of
General Microbiology 25:129.
Antimicrobial factors in African multipurpose trees 53
Weller R.A. and Pilgrim A.F. 1974. Passage of protozoa and volatile fatty acids from the rumen
of the sheep and from a continuous in vitro fermentation system. British Journal of Nutrition
32:341.
Woodward A. 1988. ?Chemical Composition of Browse in Relation to Relative Consumption
of Species and Nitrogen Metabolism of Livestock in Southern Ethiopia.?PhD thesis, Cornell
University, USA.
Woodward A. and Reed J.D. 1989. The influence of polyphenolics on the nutritive value of
browse: A summary of research conducted at ILCA. ILCA Bulletin 35. ILCA (International
Livestock Centre for Africa), Addis Ababa, Ethiopia. p. 2.
54 S.M. El Hassan, A. Lahlou-Kassi, C.J. Newbold and R.J. Wallace
Australian Meat Research Corporation?s
rumen bacteria manipulation key programme:
rationale and progress 1990?1995
I.D. Johnsson
Meat Research Corporation, P.O. Box 498, Sydney South NSW, Australia 2000
Abstract
In 1990, the Meat Research Corporation contracted four Australian research
teams to develop genetically modified rumen bacteria with enhanced cellu-
lolytic activity. The goal was a sustained increase in digestibility by at least
5 percentage units in ruminants grazing on poor quality forage typical of the
northern Australian grasslands, and more generally of 90% of Australia?s
pastoral country. With an A$4.9 million (US$3.7 m) budget, the programme
planned to explore four alternative research options over three years (mu-
tagenesis of rumen bacteria, bacteria-bacteria gene transfer, fungi-bacteria
gene transfer, cleavage of ligno-cellulose bonds), then choose the most
promising methodologies for refinement and in vivo testing in years 4 and 5.
After a review in year 3, additional funding of A$1.3 (US$1.0 m) million was
allocated to extend the research phase for two years to June 1995. Difficulty
in successfully transforming rumen bacteria has been the major stumbling
block for two years. Agenetically modified Butyrivibrio fibrisolvens express-
ing a highly active cellulase derived from the rumen fungus Neocallimastix
patriciarum is now being tried in cattle under quarantine, but no significant
impact on digestibility has been demonstrated to date. Genes for a number
of other cellulases have been isolated and characterised for transformation
into suitable host bacteria. It is likely that further funding will be needed to
optimise expression and survival in the rumen, and to achieve transfer of
more than one desirable gene into multiple strains of host bacteria, in order
to achieve the programme goal. The industry outcome is low cost bacterial
strains that spread naturally between animals in the same herd and survive
through fluctuating seasons, conferring an advantage in fibre digestibility
across a large range of grass-based diets. Their application may be limited to
management systems where nitrogen and other limiting factors can be
supplied to ruminants in the dry season.
Introduction
The Meat Research Corporation was established by the red meat industries and the
Government of Australia to invest industry research levies, which are matched by the
government, in research and development (R&D) of benefit to the industries involved.
The MRC?s current portfolio covers 200?300 projects ranging from aspects of meat
production (new plant varieties, animal genetics, stock health, grazing management)
and processing (new abattoir equipment, carcass measurement, meat hygiene, packag-
ing, storage) to human nutrition, work-related issues, retailing practices and strategic
market research.
In 1988/1989, the corporation decided to take a more pro-active approach to R&D,
and established its ?key programme? methodology for identifying, commissioning and
managing large, integrated programmes of R&D to address agreed industry priorities.
Now more than half of our annual A$50 million (US$38 m) budget is used in key
programmes, as opposed to supporting single research proposals submitted by individ-
ual research groups.
In 1989, MRC had been supporting Dr K. Gregg, University of New England, for
five years to improve the cellulolytic capacity of rumen bacteria, and Dr J. Brooker,
University of Adelaide, to improve nitrogen efficiency in rumen bacteria. The Univer-
sity of New England work included limited screening of rumen microflora in other
species (camel, buffalo etc.) for highly cellulolytic, naturally occurring bacteria that
could be readily transferred to cattle but this was unsuccessful. Given the complexity
of  the science involved and the perception that technological developments were
beginning to accelerate, MRC decided to draw together the major research groups
involved in Australia to establish a ?critical mass? across the necessary disciplines to
address the issue of improved fibre digestion in grazing ruminants.
Over 90% of Australia?s pastoral industry relies on moderate-poor quality native
grasslands which are unlikely to be improved other than by broadcasting low-phospho-
rus tolerant legumes on better soil types and by seasonal supplementation of energy,
protein and minerals.
With highly variable rainfall, and pronounced wet/dry seasons in the tropical north,
vast areas of grassland are too low in nutritive value to support animal liveweight for a
significant part of each year, unless supplemented with nitrogen and phosphorus.
Animals can literally lose weight and die in a ?sea? of dry feed. Even in more favoured
temperate zones, with improved pasture species, livestock traditionally only maintain
or lose weight over summer/autumn, and extensive use is made of crop stubbles and
conserved hay. In assessing the economic benefits of an R&D programme to increase
fibre digestion in grazing ruminants by up to about 5 percentage units, a conservative
benefit cost analysis indicated maximum annual net benefits of up to A$180 million
(US$135 m) per year. This analysis assumed a technology that was low cost to graziers,
based on local experience with the naturally occurring rumen bacteria that was imported
into Australia to detoxify mimosine ingested by cattle grazing the tropical legume,
leucaena.
The major factors in the decision to establish the key programme were:
? Low-cost of final ?product? to end user.
? Unsophisticated, robust product (cf highly sophisticated research).
? Likely to be compatible with current management systems.
? Potential for rapid adoption by target end-users.
56 I.D. Johnsson
? Potentially applicable to a large proportion of the production sector.
? ?Enabling? technology with potentially useful spin-offs e.g., nitrogen efficiency,
detoxification of plant toxins etc.
Major concerns were, and still are:
? Can genetically modified bacteria  thrive in  the rumen (are  they advantaged or
disadvantaged by carrying novel genes)?
? Can we identify and manipulate, spatially and temporally, the key enzymes that are
limiting within the complex of enzymes responsible for cellulolysis?
? If the modified host bacteria can spread between animals, will they colonise feral pest
species (camels, goats, horses, rabbits etc.) and increase environmental impact?
? Will increased fibre digestion in ruminants lead to increased over-grazing and
?desertification??
? Will the Australian public allow the release of genetically-modified rumen bacteria?
Alternative technologies were examined but rejected for the following reasons:
? Genetic engineering of pasture plants: relatively more complex, much slower and far
less total adoption in the target sector.
? Genetic engineering of ruminants themselves: more complex, greater ethical ?prob-
lems?, very much slower adoption.
? Nutritional supplementation: mature technology, already widely known/used, deliv-
ery difficult in more extensive areas.
? Manipulation of rumen fungi: role and potential to manipulate unknown, molecular
genetics unknown, current basic work in progress funded by ?sister? corporation
(being done by Dr G. Gordon).
Key programme structure
In 1990, MRC contracted four research teams to examine various methodologies to
develop rumen bacteria with enhanced cellulolytic activity. The methodologies were
proposed by the team leaders, and the plan was to fund four ?competing? methodologies
for three years to 1993, and to then choose the most promising for optimisation and field
testing in years 4 and 5.
The overall programme objective was ?to obtain a sustainable increase in fibre
digestibility of at least 5 percentage points by establishing modified bacteria in the
rumens of sheep and cattle grazed under controlled conditions by June 1995?. Two
subprogramme objectives and the major thrust of each research team are shown in
Figure 1. After an external review in December 1992, a more integrated programme
structure was adopted, with greater emphasis on sharing of knowledge, pooling of ideas
and exchanging components of transformation systems. Although the  1993 target
deadline was obviously not going to be achieved, the MRC agreed to support the
research phase for an additional two years.
In addition to the core projects, two ancillary projects which will impact on the
progress have also been supported by MRC and are shown in Figure 1. At UNE, a team
led by Dr K. Gregg has been examining a solution to fluoroacetate toxicity in grazing
ruminants by modification of rumen bacteria. At CSIRO Division of Tropical Animal
Australian Meat Research Corporation's rumen bacteria manipulation key programme 57
UNE. 032
Fluroacetate
detoxification
using modified
bacteria
CS. ?
Screening for exotic
cellulolytic bacteria
UNE. 034
Bacteria - bacteria
gene transfer
CS. 147
Genes for
enzymatic
cleavage of
ligno cellulose
UA. 013
Mutagenesis of
rumen bacteria
CS. 143
Fungi - bacteria
gene transfer
SUBPROGRAM 1
> 5% unit increase in fibre
digestion from at least 1 modified
bacterial strain, when assessed in
contained in vivo trials
SUBPROGRAM 2:
>5% unit increase in fibre
digestion sustained
under controlled
grazing
Figure 1. Key programme structure.
58 I.D. Johnsson
Production, a new project will screen ?exotic? ruminants (particularly antelope and deer
species, such as oryx, addax, gemsbok) for novel strains of rumen bacteria. These
animals are  being run under rangeland conditions at the Tipperary Sanctuary for
Endangered Wildlife in the Northern Territory.
Progress to date
Project UA.013 (Dr J. Brooker) attempted to develop highly cellulolytic bacteria by
transposon-mediated mutagenesis of naturally cellulolytic bacteria. This proved unsuc-
cessful, although transposon-marking of specific strains (Tn 916 in B. fibrisolvens E14)
was initially useful to track populations in the rumen. A naturally occurring Streptococ-
cus spp. with some ability to degrade plant tannins was isolated from feral goats and
work was initiated on the cloning and enzymology of Clostridium spp. with the hope
of providing useful cellulase genes. This project was not extended after the mid-term
review.
Project UNE.034 (Dr K. Gregg) continued previous work to introduce cellulase genes
from highly cellulolytic rumen bacteria into other predominant, but poorly-cellulolytic
rumen bacteria. Major progress has been:
? The cloning and sequencing of genes from Ruminococcus albus AR67 (Gregg et al
1993).
? Transformation achieved for Prevotella ruminicola AR20 to reasonable efficiency (K.
Gregg, personal communication).
? Transformation achieved for Butyrivibrio fibrisolvens OB156 (Beard et al 1994).
? Development of qualitative and quantitative PCR systems to track B. fibrisolvens and
P. ruminicola strains in cattle and sheep (K. Gregg, personal communication).
? Demonstration of long-term persistence of introduced laboratory-grown strains of B.
fibrisolvens (E14, AR10) and P. r u m i n i c o l a (AR20, AR29) in the rumens of sheep or
cattle, and of their ability to transfer naturally between hosts in some cases.
Project CS.143 (Dr J. Aylward) aimed to transfer genes for highly active fungal
cellulases secreted by N. patriciarum to cellulolytic rumen bacteria to enhance their
capacity to degrade cellulose. Major progress has been:
? Cloning of five genes of varying activities from N. patriciarum (celA, celB, celC and
celD, plus xynA), after preparation of a comprehensive gene library. (Xue et al 1992;
Orpin and Xue 1993).
? Development of a transformation system for five strains of B. fibrisolvens (K.S.
Gobius, personal communication).
? Quantification of relative activities of various enzymes in vitro, alone and in combi-
nation, using standard forage substrates (J. Aylward and P. Kennedy, personal com-
munication).
Project CS.147 (Dr P. Kennedy) aimed to find enzymes which are capable, under
anaerobic conditions, of at least partial hydrolysis of ligno-cellulose bonds. Significant
progress has been:
? The development of a range of model substrates for assaying enzymes which cleave
various linkages.
Australian Meat Research Corporation's rumen bacteria manipulation key programme 59
? Demonstration of higher lignin content in tropical grasses than indicated by the
traditional acid detergent lignin technique,  and  of  higher  lignin digestion than
previously recognised (Lowry et al 1994).
? With CS.143, the cloning of acetyl xylan esterases from the N. patriciarum gene
library (B. Dalrymple, personal communication).
? A study of cinnamoyl ester hydrolases isolated from the B. fibrisolvens E14 gene
library (D. Cybinski and B. Dalrymple, personal communication).
? Development of techniques (with UNE.034) to study persistence of B. fibrisolvens
strains in cattle and characterisation of their fibre degrading capacity in vitro (C.S.
McSweeney, personal communication).
The most significant breakthroughs with regard to the attainment of the programme
goal have occurred in the last 12 months:
? In a closely related MRC-funded project, K. Gregg at UNE developed the first
potentially useful recombinant rumen bacterium, B. fibrisolvens OB156 expressing
a dehalogenase enzyme which degrades toxic fluoroacetate (Gregg et al 1994). This
recombinant strain has been shown to persist for at least five months in penned sheep,
and its efficacy in vivo is currently being tested.
? First transformation of several B. fibrisolvens strains with cloned fungal genes from
N. patriciarum and secretion of cellulase/xylanase enzymes (K.S. Gobius and G.P.
Xue, personal communication).
? First evidence that transformation of a rumen bacteria with a xylanase gene can
enhance the ability of the native strain to digest fibre when incubated in pure cultures
in vitro (C.S. McSweeney, personal communication).
The future
On the basis of these results, the MRC is likely to contribute further funding to take the
research through to ?proof of concept?, namely the demonstration of an increase of at
least 5 percentage units in digestion of poor quality grasses in ruminants inoculated with
at least one genetically-modified bacterial strain, when assessed in containment facili-
ties.
This should be achieved by December 1995, most probably using a suite of modified
bacterial strains, each expressing one of a range of novel enzymes of bacterial or fungal
origin.
Subsequent molecular biology research could then be undertaken to optimise expres-
sion, stability (is chromosomal integration essential?) and persistence, to explore combi-
nations of recombinant genes in the one host strain and to extend the library of useful genes
and candidate hosts, in addition to the detailed rumen microbiology and nutrition studies
needed to quantify the commercial value of the modified strains in diverse commercial
grazing situations. One can thus envisage ?waves? of new recombinants being produced
for animal evaluation as the scientists involved become more experienced in the manipu-
lation of rumen bacterial species and find new sources of novel enzymes.
However, from an industry viewpoint, this research is of no value if Australian society
decides not to condone release of these recombinants for commercial use. Thus a priority
60 I.D. Johnsson
for the MRC programme, once the concept is proven, will be to assess the likely
environmental impact of highly cellulolytic recombinant bacteria, in terms of transfer
to non-target species (feral rabbits, horses, camels and goats, in particular, as well as
native macropods) and of effects on grazing animal intake, and therefore grazing
pressure, on fragile landscapes.
The ultimate decision will be made on the basis of risk assessment, based on an
evaluation of the benefits and risks. Thus comprehensive efficacy data will be essential,
probably across a range of management systems. If the concerns regarding feral animal
competitiveness can be addressed, it seems likely that approval will be granted, given
that the ultimate outcome is similar to the impact of increasing stocking rate, changing
from Bos taurus to B. indicus cattle (with consequently higher survival rates) in tropical
areas or changing from a smaller breed to a larger breed of cattle. In these cases, changes
in grazing management practices have been required to ensure sustainable land use, and
these new practices are being increasingly adopted by Australian pastoralists.
References
Beard C.E., Hefford M.A., Forster R.J., Sontakke S., Teather R.M. and Gregg K. 1994. A stable
and efficient transformation system for Butyrivibrio fibrisolvens OB156. Current Microbi-
ology (in press).
Gregg K., Rowan A. and Ware C. 1993. Digestion of filter-paper by cellulases cloned from
Ruminococcus albus AR67. In: Shimada K., Hoshino S., Ohmiya K., Sakka K., Kobayashi
Y. and Karita S. (eds), Genetics, Biochemistry and Ecology of Lignocellulase Degradation.
Uni Publishers Co, Tokyo, Japan. pp. 166?172.
Gregg  K.,  Cooper  C.L, Schafer D.J.,  Sharpe  H.,  Beard C.E.,  Allen  G. and Xu J.  1994.
Detoxification of the plant toxin fluoroacetate by a genetically modified rumen bacterium.
Bio/Technology 12:1361?1365.
Lowry J.B., Conlan L.L., Schlink A.C. and McSweeney C.S. 1994. Acid detergent dispersible
lignin in tropical grasses. Journal of the Science of Food and Agriculture 65:41?49.
Orpin C.G. and Xue G. 1993. Genetics of fibre degradation in the rumen, particularly in relation
to anaerobic fungi, and its modification by recombinant DNA technology. In: Proceedings
of the XVII International Grassland Congress held in Palmerston North, Hamilton and
Lincoln, New Zealand, 13?16 February 1993. New Zealand Grassland Association, New
Zealand. pp. 1209?1214.
Xue  G.P., Orpin C.G., Gobius K.S., Aylward J.H. and Simpson G.D. 1992. Cloning and
expression of multiple cellulase cDNAs from the anaerobic rumen fungus Neocallimastix
patriciarum in E. coli. Journal of General Microbiology 138:1413?1420.
Australian Meat Research Corporation's rumen bacteria manipulation key programme 61
State of art in manipulation
of the rumen microbial
ecosystem
Ruminants and rumen microorganisms
in tropical countries
H. Kudo1, S. Imai2, S. Jalaludin3, K. Fukuta4 and K.-J. Cheng5
1. Japan International Research Center for Agricultural Sciences, Ibaraki, 305 Japan
Present address: National Institute of Animal Industry, Inashikigun, Ibaraki, 305 Japan
2. Nippon Veterinary and Animal Science University, Tokyo, 180 Japan
3. Universiti Pertanian Malaysia, Selangor, Malaysia
4. National Institute of Animal Health, Ibaraki, 305 Japan
5. Agriculture and Agri-Food Canada, Alberta, Canada T1J 4B1
Abstract
Crop residues and by-products can be utilised for animal production. Some
by-products contain toxic components and may require supplementation or
modification to reduce the toxicity.
Mimosine/DHP poisoning by Leucaena may be reduced by introducing
rumen bacteria that can degrade mimosine and DHP. The comparative study
on ciliate fauna strongly indicated the presence of geographical distribution.
For example, Eudiplodinium giganteum, Eudiplodinium kenyense,
Enoploplastron stokyi, Diplodinium africanum, Diplodinium  nanum and
Ostracodinium iwawoi have been found only in Africa. In the percentage
composition of genera, the numbers of Entodinium, which normally pre-
dominate in rumen in other areas, were very low in African ruminants. The
presence of Metadinium ypsilon and Ostracodinium trivesiculatum is char-
acteristic of Malaysian cattle and water buffaloes. The level of Metadinium
and Eudiplodinium, considered to be fermenters of cellulosic materials, was
higher in Malaysian cattle and water buffaloes. Cellulolytic bacteria isolated
from water buffaloes possess a higher ability to degrade cellulose than those
from cattle.
From the viewpoint of rumen microbiology, rumen flora and fauna of
lesser mousedeer (Tragulus javanicus), the smallest ruminant (1.3 to 1.7 kg
at adult size), are unique and very interesting. Fairly large motile bacteria,
similar to Oval and Oscillispira, are present in large numbers, contributing
to a higher microbial mass. The natural occurrence of mono-fauna protozoa
or the total absence of protozoa have also been observed. So far, no fungus
has been detected or isolated from the rumen contents. For the successful
development of rumen microbiology in the field of biotechnology, it is
important to study and include more basic characteristics of rumen micro-
organisms, including some ecosystems which have so far not received
attention.
Introduction
Today?s ruminants receive an increasing variety of feeds. In extreme cases, farmers feed
ruminants like non-ruminant animals, ignoring the fact that they are ruminants, because
of the availability of feed and the adaptability of the animals. As the human population
increases, we produce a greater variety and quantity of waste products which can be
consumed by animals but consequently the land available for both housing and agricul-
ture is reduced. The nutrient energy in most of these by-products is not readily available
and is not sufficient, even for the ruminant digestive system. This is especially the case
for cellulosic materials which have slower rates of digestion because the cellulose is in
a highly organised and complex form, and because this kind of feed generally lacks the
nitrogen necessary to balance the large amounts of available carbon. We are now trying
to manipulate the rate of digestion of various feeds to favour both humans and ruminants.
We are trying to accelerate those rates that are slow, while delaying those with a high
initial rate of nutrient release which can cause digestive disturbances, such as acidosis,
bloat and liver abscesses (Cheng et al 1991). The former objective is more interesting
and more important to tropical countries. Theoretically it is conceivable that up to 90%
of ruminant feed could be made up of by-products which have been modified to improve
digestibility.
In practical terms, the benefits from the successful attainment of the objectives
would be (1) to make underutilised by-products available as ruminant feed, (2) to
improve the digestibility and efficiency of utilisation of these by-products for animal
feed, and (3) to reduce the requirement for hay and animal feed products?thus freeing
land for more valuable crops. In Asia, ruminants are kept by small farmers on a small
scale and the production systems are closely integrated with farming systems. One of
the limiting factors for the expansion of production is the available feed resources, as
under such farming systems the farmers do not have adequate feed resources or the
capacity to increase available feed. Usually agricultural by-products are utilised for
feeding domestic ruminants. Given the present decline in the area available for forage
production, ruminants will have to depend more and more on fibrous residues and
by-products for their energy sources. These by-products still have problems to be
solved and many of them cannot be used without some detoxifying treatments.
However, considering the quantities of the by-products and the fact that tropical
ruminants are better converters of low quality feed into milk and meat (Abdullah 1987),
the by-products are likely to play an important role as potential feed sources in ruminant
nutrition.
Ruminants in Southeast Asia
Southeast Asia is rapidly expanding its manufacturing industry but agriculture continues
to be the mainstay of the rural economy. Agriculture in Southeast Asian is crop-based,
producing abundant feedstuffs (conventional  and non-conventional)  which  can be
utilised for animal production. Demand for animal products has increased due to rapid
urbanisation and improved living standards and local production is inadequate.
The slow growth of ruminant production in Southeast Asia, in spite of readily available
resources (especially feeds), may be attributed to the inefficiencies in their utilisa-
tion?the consequence of inappropriate technological development and application
(Jalaludin et al 1992).
The lesser mousedeer (Tragulus javanicus)
Successful breeding in the laboratory
The lesser mousedeer is the smallest ruminant, weighing 1.3 to 1.8 kg, and is considered
a ?living fossil? which escaped from extinction and remains in its original morphology
after 25 million years. The lesser mousedeer is found in Southeast Asia and all the Sunda
Islands. The family Tragulidae comprises the three Asian species of mousedeer or
chevrotain, genus Tragulus and the water chevrotain of Africa, Hyaemoschus aquaticus
(Anderson and Jones 1967).
The study of the nutritional physiology of herbivorous domestic animals such as the
ruminant is very important in the field of veterinary and zootechnical sciences (Goto
and Hashizume 1978; Kudo and Oki 1981, 1982, 1984). However, small herbivorous
laboratory animals are still not established for comparative purposes. From the view-
point of conservation and protection of wildlife, it is also very important to establish a
breeding colony. With permission from the Department of Protected Wild Life and
National Park, Malaysia, August 1990 the rumen microbiology laboratory of Universiti
Pertanian Malaysia has started to establish a breeding colony of mousedeer and has
initiated some studies on the reproduction, nutrition and microbial population and
activity in the rumen of mousedeer. Ten (five males and five females) mousedeer were
introduced from the jungles of Selangor and Pahang, Malaysia. We feed them with rabbit
pellets, sweet potatoes, carrots, kangkong (Ipomoae aquatica) and long beans. Hand
feeding, mainly on sweet potato, is used for better reproductive results. Despite the fact
that they are all from the wild, they take feed from hands on the same day they are
introduced into the laboratory. Foster nursed mousedeer can be toilet trained like dogs
and cats. We kept a pair of mousedeer in half of a stainless steel cage of 46 H???116W???45
cm D with a divider.
Mortality of captured lesser mousedeer is usually quite high (60?70%; National Zoo,
Malaysia, and Institute of Medical Research, Malaysia, personal communication). We
recorded about 10% mortality in our laboratory. The behaviour of the tragulids is of
particular interest because of the remarkable convergence in morphology and ecology
between the tragulids and the caviomorph rodents of South America. There are many
features that differ from the ruminants (Fukuta et al 1991; Ralls et al 1975). It ruminates
while standing, does not use its tongue for eating but instead uses it for drinking (like a
dog or a cat), sits dog-like and can scratch its head with the hind leg. Rabbit-like
stamping is often observed. Thus, they behave more like dogs and cats rather than like
ruminants. Males do not have antlers but instead they have well-developed canines
which extend beyond the upper lip. Their feeding style and habit are very selective, very
often eating only particular parts of plants. Therefore, this animal may not be
suitablefor feeding experiments. On the other hand, this feeding style may have
protected the animals from plant poisoning. For example, they like lundai leaf (Sapium
baccatum, a tropical shrub tree) very much but the appetite for this leaves is not
consistent. The lundai leaf is thought to be one of the important components of the diet
of the Malaysian mousedeer in its natural rain forest habitat but it contains a very high
level of a toxic substance, mimosine. Their small mouths permit them to select parts of
plants in a way that the larger ungulates are not able to do. The teeth type is not
particularly suited for grinding (Vidyadaran et al 1981). The rumen of the adult lesser
mousedeer is fairly large (? of the total abdominal cavity) and the ratio of reticulo-ru-
men volume is larger than that of cattle and sheep (Vidyadaran et al 1982). The omasum
is vestigial.
Mousedeer reach sexual maturity at about four to five months of age and adult size at
five months. The earliest age of the first parturition in our laboratory was 258 days. The
canines of the young male do not extend beyond the upper lip until they are about nine to
ten months old. Sexual cycle of female lesser mousedeer is estimated to be 16 days from
mounting action, the duration of mating season is two days where copulation occurs
several times and they resume estrus unless they are pregnant. In mousedeer, pregnancy
is not apparent but females tend to sit all the time for a few days before parturition. The
female mousedeer are known to have a postpartum estrus (Cadigan 1972; Davis 1965)
and copulation is observed often after 30 min to a few hours after the parturition. There is
no breeding season in the laboratory. All five pairs from jungles and the subsequent
offspring reproduce well and gave relatively regular continuous parturition which may
compensate for their relatively long gestation period. Although there are numerous reports
that the gestation period of mousedeer is 140 to 177 days (Anderson and Jones 1967;
Lekagul and McNeely 1977; Tubb 1966), we observed a shorter gestation period. The
apparent gestation period is 136 to 164 days, but actual gestation is estimated to be 134
(SE?=?2)days,judgingfromtheintervalsofcontinuousparturitions.Adelayinimplantation
due to lactation probably occurs. It was believed that reproduction in the laboratory was
quite difficult but we did not have a major problem, unlike other wild animals introduced
into the laboratory. The size of new offspring is fairly large (120 to 190 g) in relation to
the mother animal (about 10%). Infant mortality among the mousedeer at the zoo was
unusually high (?; Ralls et al 1975). Infant mortality at our laboratory was about 14%
(6/43, until weaning). Our reproductive data at Universiti Pertanian Malaysia for 3? years
is summarised in Table 1.
Rumen content of mousedeer is very much like a thick slurry with a pH range of 5.5
to 6.5. From the viewpoint of rumen microbiology, the rumen flora and fauna of
mousedeer are unique and very interesting in that they differ totally from other domestic
ruminants. A number of bacterial isolates (e.g. Fibrobacter succinogenes, Ruminococ-
cus albus and R. flavefaciens) are highly cellulolytic. Fairly large motile bacteria,
similar to Oval and Oscillospira, are present in large numbers, hence contributing to a
higher microbial biomass. The occurrence of mono-fauna protozoa or a total absence
of protozoa have also been observed. The reasons for these phenomena are not known.
Six species of protozoa belonging to two genera were detected in our laboratory (Table
2). Ciliate composition is poor as compared with other domestic ruminants (Table 3)
but the total counts are comparable (4.6???105/g of rumen contents) to other domestic
ruminants (Imai et al 1995). So far, no fungus has been detected and isolated from the
rumen contents.
Feeding systems for ruminants in Southeast Asia
Pasture utilisation
The standard of feeding should be improved along with the increase of the genetic
potential of the ruminants in Southeast Asia. Southeast Asia has a sizeable land area
covered with natural pastures, especially in Indonesia where it is more than twice the
area under cropping systems. Fresh fodder is also produced at a rate of 5?10 ton/ha from
land under crops. A major part of the roughage intake for ruminants in Southeast Asia
is provided by about 20???106 ha of permanent pastures. A conservative yield estimate
of 5 ton dry matter per year would be sufficient to supply the requirements of all large
ruminants in the region. There are at least 60 species of naturally grown forages, of
which 70% are palatable. Due to poor soil quality and rapid maturity, tropical forages
are low in nutritive value. However, with phosphorus dressing, a stocking rate of four
zebu cattle per ha is possible. Individual liveweight gain ranges from 281?448 g/day
Table 1. Reproductive data of captured lesser mousedeer at Universiti Pertanian Malaysia.
Breeding season All year round
Estrous cycle 14 to 16 cyclic days
Mating period 2 days
Earliest age of copulation (male) 166 days
Earliest age of conception (estimated) 125 days
Earliest age of the first parturition 258 days
Shortest gestation period 132 days
Estimated actual gestation period 134 (SE = 2) days
Earliest copulation after parturition (postpartum estrus) 30 min (Yes)
Delay of implantation of fertilised egg Yes
Maximum continuous parturition within 136 days of interval 3 times (136, 133 and 136 days)
Litter size 1
Infant mortality before weaning 6/43
Table 2. Ciliate protozoa in the rumen of lesser mousedeer at Universiti Pertanian Malaysia.
Total ciliate number 45.5 ? 10
4
/ml
Species identified Isotricha jalaludinii
Entodinium simplex
E. dubardi
E. anteronucleatum
E. nanellum
E. convexum
but in terms of annual liveweight gain per ha, the highest yield, at 855 kg, was obtained
on mixed grass-Leucaena pasture with a stocking rate of 7.3 animals per ha. The grazing
intensity can be raised to ten animals per ha on fertilised grass pasture, yielding 1100
kg/ha per year, if accompanied by moderate supplementation of concentrate (Jalaludin
et al 1992).
Non-conventional feeds in Southeast Asia
Non-conventional feeds, such as crop residues and by-products, are fast emerging as an
alternative source of nutrients for intensive ruminant production. In Southeast Asia, rice
cultivation and oil palm production produce the largest quality of residues and by-prod-
ucts. The total quantity of fibrous crop residues available in Malaysia is more than
300???103 tons (comprising mainly rice straw and palm press fibre (PPF). Rice straw
utilisation has been widely reported but information on the potential use of oil palm
by-products as animal feed is lacking.
In Malaysia, research to promote a viable, low cost ruminant production system based
on oil palm by-products as the main source of nutrients is being undertaken. By-products
such as PPF, oil palm trunk (OPT) and oil palm fronds (OPF) are very fibrous but
potentially rich in energy, while palm kernel cake (PKC) contains high protein. The
fibrous residues, e.g. PPF and OPT, are characterised by poor digestibility due to a high
lignin content. They are also low in crude protein. Treatment of PPF with 8% sodium
hydroxide increased the dry matter digestibility from 43.2 to 58.0%. Physical treatment
with steam at 12 kg/cm2 for 5?10 min increased the digestibility from 26 to 28% (Oshio
et al 1990). The response of OPT to chemical and physical treatment is similar to that
obtained for PPF.
Palm kernel cake is widely used as animal feed in Malaysia and can be a good feed.
However, its high concentration of copper and imbalance of Ca/P cannot be over-
Table 3. Comparative host distribution of rumen protozoa.
Lesser mousedeer Water buffalo Cattle Goat
Entodinium ++++
Isotricha
Diplodinium
Eodinium ++
Eudiplodinium ++++
Metadinium
Polyplastron ++
Elytroplastron +
Ostracodinium
Enoploplastron ++
Epidinium ++++
Epiplastron +
Ophryoscolex ++
Opisthrotrichum +
Source: Ogimoto and Imai 1981.
looked. Table 4 shows the mineral composition of PKC. Although few negative papers
regarding PKC as animal feed have been published, the author (H.K.) has observed
abnormalities of rumen flora and fauna of ruminants fed on PKC. Feeding less than
30% PKC showed no effect. Palm oil is the major agricultural industry in Malaysia
and the residues are usually burned. As oil palm bunch ash (OPBA) is strongly alkaline,
Murayama and Zahari (1993) studied OPBA as a correctant for soil acidity. Low
germination rate and abnormal growth were observed in corn grown on OPBA-applied
soil.
Characteristics of the forage material
Forages differ greatly in their composition and structure. Temperate and tropical forages
are covered on the outer surface by a thin waxy cuticle, which contains a tough basal
polyester layer. Immediately below this is the plant cell, in which the cell walls account
for the bulk of the aerial portion of the plant. The cell wall is composed primarily of
cellulose and hemicellulose, with a small portion of lignin (van Soest 1982).
The  cuticular surface layer presents a formidable barrier to invasion by rumen
microorganisms. The cutin layer appears to be totally resistant to microbial digestion
within the rumen, except for some rumen fungi which may penetrate it (Ho et al 1988a).
Cutin forms the structural component of the plant cuticle and is a polyester C-16 and
C-18 hydroxy- and hydroxyepoxy fatty acid. The cutin fraction ranges from 0.2% of the
cell wall of wheat straw to 2.4% for mature alfalfa (van Soest 1982). The ester
Table 4. Mineral composition of palm kernel cake (solvent extracted).
Palm kernel cake (PKC)
1 2 3 4 5-1 5-2 6-1 6-2
Ca (%) 0.28 0.25 0.28 0.27 0.25 0.24 0.24 0.25
P (%) 0.60 0.62 0.57 0.61 0.64 0.64 0.62 0.62
Mg (%) 0.23 0.30 0.31 0.30 0.29 0.28 0.27 0.28
K (%) 0.76 0.78 0.79 0.79 NT NT NT NT
S (%) 0.19 0.24 0.23 0.23 NT NT NT NT
Cu (ppm) 28.50 28.90 27.50 26.40 19.80 21.30 19.30 19.70
Zn (ppm) 40.50 43.60 48.30 48.40 41.60 42.30 40.40 40.70
Fe (ppm) 6130 5666 844 835 817 3486 665 1125
Mn (ppm) 201 196 204 197 38.40 147.90 79.60 73.6
Co (ppm) ? ? ? ? ? ? ? ?
Se (ppm) 0.30 0.32 0.23 0.24 NT NT NT NT
Pb (ppm) NT NT NT NT ? ? ? ?
Ni (ppm) NT NT NT NT ? ? ? ?
Cd (ppm) NT NT NT NT ? ? ? ?
Mo (ppm) NT NT NT NT ? ? ? ?
Na (ppm) NT NT NT NT 96.00 82.30 89.60
A dash (?) represents values below detectable limits; NT not tested.
PKC 5-1 and 5-2 were duplicates, as were 6-1 and 6-2; PKC 1 to 4 were analysed by Dr M. Ivan, Agriculture Canada;
PKC 5 and 6 were analysed by Dr T. Kawashima, Japan International Research Centre for Agricultural Sciences.
linkage of cutin is hydrolysed by some pathogenic fungi and some aerobic bacteria.
Most of the usable plant nutrients are internal. External plant tissues are only poorly
colonised by rumen microorganisms while the inner tissues are heavily colonised.
However physical disruption (e.g. chewing) is necessary to allow optimal microbial
access to the inner tissues which are then avidly colonised (the inside-out digestion
concept; Cheng et al 1990b).
Characteristics of rumen microorganisms
Rumen bacteria
Ruminants posses a complex stomach system, in which the stomach is divided into three
or four compartments, the first and the largest of which is the rumen. It is here that
continuous anaerobic fermentation takes place by a complex community of microor-
ganisms. The rumen is an extremely complex community of many microorganisms,
protozoa, bacteria, fungi and probably other unknown microorganisms (Figure 1). When
ruminants are born their rumen is germ-free, the unique flora and fauna start to establish
after birth. Once established, the rumen microbial community is very stable and will
change only when the nutrients are changed (Cheng and Costerton 1980). This may be
the reason that feed efficiency in ruminants has stayed the same in recent decades with
the exception of some success using feed additives such as the ionophores, monensin
and lasalocid while that of pigs and chickens has greatly improved. Although a definite
reason has not been found, it is known that some ruminants, notably the buffalo and the
yak, can utilise poor quality feed more efficiently. Our experimental results indicated
that the cellulolytic bacteria, Fibrobacter succinogenes (formerly Bacteroides succino-
genes; Montgomery, 1988) and Ruminococcus flavefaciens strains isolated from water
the buffalo possess a more active cellulase activity than strains isolated from cattle
(Tables 5 and 6). This is probably because these animals have been kept on low quality
roughage feeds for a long time. There was no difference in numbers of cellulolytic
bacteria between water buffaloes and KK cattle in Malaysia. We often detected and
isolated antibiotic-resistant strains of rumen bacteria from Indonesian and Malaysian
domestic ruminants.
Rumen protozoa
Tropical countries are known to have a wider diversity of flora and fauna, including
many toxic plants and highly active microorganisms. As it is known that the ciliate
composition varies according to the feeding, geographical distribution and/or physical
condition of the host, ciliate composition has been used as an index of the conditions in
the rumen. However, the classification of these ciliates is not easy when we study
tropical ruminants, because the taxonomy of these ciliates is still inconsistent and
furthermore most reports on rumen ciliates have described only those of the ruminants
in temperate zones. Tropical ruminants have many characteristics and varieties of
ciliates.
Protozoa in African ruminants
From the comparative study of ciliate fauna, the presence of a geographical distribution
was strongly indicated. For example, Eudiplodinium giganteum, Eudiplodinium
kenyense, Enoploplastron stokyi, Diplodinium africanum, Diplodinium nanum and
Ostracodinium iwawoi have been found only in Africa (Imai et al 1992). In the
percentage composition of genera, the numbers of Entodinium, which normally pre-
dominates in the rumen in other areas, were very low in African ruminants. In contrast,
the values of the genera belonging to the subfamily Diplodiniinae, such as
Eudiplodinium, Ostracodinium and Diplodinium, were high in African ruminants. The
ratio of genera is affected by the diet of the host (Hungate 1966). The Diplodiniinae
were observed to ingest many fragments of plants. These ciliates have been considered
Table 5. Fermentation products produced by strains of Fibrobacter succinogenes isolated from steers and
water buffaloes.
Fermentation products (mM)
?
Strain Acetate Butyrate Succinate Lactate
Steer Water buffalo Steer Water buffalo Steer Water buffalo Steer Water buffalo Steer Water buffalo
a 112 5.0 10.1 ? 1.3 9.2 11.5 ? ?
Z 116 5.2 7.3 ? 0.6 8.7 12.7 ? 1.2
b 123 4.9 6.1 ? 1.2 8.8 12.6 ? 1.1
M 127 5.5 7.1 ? 0.6 7.9 11.9 ? 1.2
C 126 5.5 5.6 ? ? 7.7 12.7 ? 1.7
A 114 5.0 8.0 ? ? 7.9 10.2 ? 1.7
D 104 5.2 8.1 ? ? 7.5 11.5 ? ?
o 121 3.9 7.4 ? 1.2 8.7 10.1 ? 0.8
O 128 4.8 4.8 ? ? 7.5 11.1 ? 2.8
d 120 4.6 4.4 ? ? 7.6 12.3 ? 1.8
V 107 4.2 7.0 ? ? 7.9 11.5 ? ?
?
Cultures were incubated for five days at 39?C in 10 ml of Scott and Dehority?s medium with 30 mg of Whatman No. 1
cellulose filter paper.
Table 6. Fermentation products produced by strains of Ruminococcus flavefaciens isolated from steers and
water buffaloes.
Fermentation products(mM)
?
Strain Acetate Lactate Succinate
Steer Water buffalo Steer Water buffalo Steer Water buffalo Steer Water buffalo
1 102 2.9 6.5 2.2 ? 4.5 8.5
2 106 3.4 6.9 2.2 ? 4.4 6.8
3 115 2.8 6.2 2.0 2.2 4.4 7.8
4 136 3.3 8.7 2.3 0.6 4.1 9.4
?
Cultures were incubated for five days at 39?C in 10 ml of Scott and Dehority?s medium with 30 mg of Whatman No. 1
cellulose filter paper.
None of the strains produced ethanol.
to possess cellulolytic activity  (Hungate 1978). Thus,  the ratio of ciliate genera
obtained in African study may reflect the fact that the zebu cattle examined had been
fed roughage. The very low ratio of Entodinium might indicate the low amount of
starch in the ration since it is known that the entodinia principally utilise starch grains
as their energy source (Abou-Akkada and Howard 1960). When comparing the ciliate
compositions in East African zebu cattle with those in India (Banerjee 1955; Kofoid
and MacLennan 1930, 1932, 1933) and Sri Lanka (Imai 1986; Kofoid and MacLennan
1930, 1932, 1933), a high similarity is found, suggesting a close phylogenetic relation-
ship between East African zebu cattle and Indian zebu cattle (Imai 1988). A similar
relation has been found in the distribution of haemoglobin-beta alleles (Namikawa
1980).
Protozoa in Southeast Asia
The ciliate fauna of zebu cattle and water buffaloes in Thailand resembled that of zebu
cattle and water buffaloes in Malaysia and the Philippines rather than that of zebu cattle
in the Indian area. Although most of the protozoan species detected from water buffaloes
and KK cattle in Malaysia have been detected from domestic ruminants and temperate
areas, the presence of Metadinium ypsilon and Ostracodinium trivesiculatum is charac-
teristic of Malaysian cattle and water buffaloes (Imai et al 1995). The level of Metad-
inium and Eudiplodinium, considered to be fermenters of cellulosic materials, was
higher in Malaysian cattle and water buffaloes. This suggests that Malaysian cattle and
water buffaloes possess a ciliate protozoal composition favourable for digestion of
Figure 1. Social structure of rumen microbial consortia
cellulosic feedstuffs. Of the species which had low incidence, Entodinium longinuclea-
tum spinolobum, E. parvum monospinosum, E. tsunodai, E. bubalum, E. fujitai and E.
javanicum have been reported only from ruminants in Southeast Asia (Imai 1985).
Rumen fungi
All the species of rumen fungi so far isolated are capable of fermenting structural
carbohydrates of plant cell walls. Rumen fungi in the rumen also contribute significantly
to the prime function of the rumen, which is the digestion of plant cell walls to provide
fermentation products for the nutrition of the host animal (Orpin and Ho 1991). One of
the characteristics of fungi is their penetration into the plant material. Microscopic study
indicates  that  the rhizoid penetrates the tissue, colonising the schlerenchyma  and
vascular tissues, eventually degrading the schlerenchyma (Ho et al 1988a). The rhizoidal
system attaches to the more recalcitrant vascular tissue, resulting in the fungus remain-
ing attached to this tissue despite the degradation of surrounding cells. In this way, the
fungus remains attached to the tissue and is not washed out of the rumen with the liquid
phase of the rumen contents. Attachment to plant cell walls within the digesta fragments
is, in some species, by way of appressoria (Ho et al 1988b).
Comparison of fungal strains between those from temperate areas and tropical areas
has been difficult, as the practical transportation of fungal strains has not yet been
established. Although we do not yet have a reliable method to determine fungal numbers,
in general, the higher the fibre content of the diet, the is higher the population density
of rumen fungi (Orpin and Ho 1991). It is interesting that as the digestibility decreased
and lignin content increased, fungal populations increased (Orpin 1977).
Plant toxins and rumen microorganisms
Leucaena is widely grown in tropical and subtropical countries. The plant is potentially
an excellent source of crude protein. However, its use as a feed has been limited because
it contains mimosine, a toxic amino acid that causes many undesirable problems (low
weight gains, poor health condition and hair loss, etc.) in both ruminants and non-rumi-
nants.  Mimosine toxicity is more acute  in non-ruminants owing to the absence of
endogenous microorganisms capable of enzymatic detoxification.Mimosine is hydrolysed
in the rumen by microbial enzymes to 3,4 DHP, a potent goitrogen, and sometimes to
2-hydroxy-3-(1H)-pyridone (2,3 DHP), a structural isomer of 3,4 DHP which is probably
as goitrogenic as 3,4 DHP.
Previous Canadian (Kudo et al 1984, 1986) and Malaysian (Kudo et al 1989b, 1989c)
studies on the in vitro metabolism of mimosine showed that degradation occurred in
the microbial fraction and that concentrate diets increased the microbial population
that degraded mimosine, but probably not sufficient to completely escape from mimos-
ine poisoning. When we examined the rate of mimosine degradation in the rumen fluid
of cattle fed five different diets, significant (P?<?0.01) differences were seen between
the highest rate on a blue grass-molasses diet and the lowest rate on corn silage.
The highest rate obtained with bluegrass (Poa prantensis)-molasses suggests
that high rates of metabolism can also be associated with inocula from diets other than
concentrates. Intermediate rates were obtained with alfalfa hay, fresh alfalfa herbage
and orchard grass hay (Dactylis glomerata) but these rates were not significantly
different. In vitro rates for mimosine and DHP degradation using ruminal contents were
conducted in Malaysia during 1988 (Kudo et al 1989b). All rumen samples from cattle,
sheep and water buffaloes produced no disruption of the heterocyclic ring, suggesting
a complete absence of DHP detoxification. In 1989, the active rumen fluid from
Indonesian goats fed on and adapted to Leucaena was infused into one of these cattle.
Not only the infused animal but also the untreated neighbouring animals gained the
ability to detoxify DHP (Kudo et al 1989c, 1990c). During this study, we established
that the ability to detoxify DHP is very rapidly passed from treated to untreated
neighbouring animals. To date, only the rumen microorganisms in Hawaii (Henke
1958; Jones 1981) and Indonesia (Kraneveld and Djaenoedin 1947) have been shown
to detoxify mimosine in vivo and protect the animals from any ill effects caused by the
plants. Thus, ruminants in some parts of the world possess microorganisms capable of
mimosine detoxification. It was demonstrated that Australian cattle could be protected
from mimosine poisoning by transferring rumen fluid obtained from Hawaiian goats
that can degrade DHP (Jones and Megarrity 1986). A rumen bacterium capable of DHP
degradation was recently isolated from the rumen contents of a goat from Hawaii
(Allison et al 1992). Recent grazing trials with Leucaena pastures have also shown that
transfers of DHP-degrading bacteria can protect cattle in both Australia and North
America (Hammond et al 1989; Pratchett et al 1991; Quirk et al 1988). We should note
that in tropical countries the quality of feed supplied, including that supplied to
laboratory animals, is unsatisfactory, varying from batch to batch. Some feeds contain
toxic substances, which cannot be degraded even with the special capabilities of rumen
microorganisms.
Factors affecting feed degradation
Microbial attachment
Recent studies have demonstrated that, for the digestion of cellulose (Figures 2 and 3)
and starch, attachment to these insoluble substrates is a pre-condition for both pure
cultures and the natural mixed population if digestion is to proceed. Cellulolytic species
differ in the nature of their attachment to insoluble substrates and in the nature of their
enzymatic attack (Kudo et al 1987b; McAllister et al 1990; Minato and Suto 1978).
When these substrates are placed in the rumen, they attach to substrates very rapidly
(< 15 min). Among the major cellulolytic bacteria in the rumen, F. succinogenes
attaches very closely to the substrate while R. flavefaciens attaches at a small distance
and R. albus remains at a much greater distance. This spatial distribution appears to be
dictated by the width of the glycocalyx structure that these species use to attach to
cellulose but, in all three of these bacterial species, their cellulases reach and digest the
colonised cellulose  substrate. Similarly, the digestion of starch is  affected by the
structure of the glycocalyx of the amylolytic organisms (Cheng et al 1990a). Amylase
Figure 2. Scanning electron micrographs (SEM) of Whatman No. 1 filter paper incubated with cellulolytic
bacteria at  39?C for  20 h  showing  extensive  bacterial adhesion to cellulose fibres. (A) Fibrobacter
succinogenes BL 2. Note that grooves (G) have formed as a result of cellulose digestion and that some
deformation of cells is evident. (B) Fibrobacter succinogenes E. Note that the bacterial cells attach avidly
to the fibre and are oriented in the same direction as the cellulose subfibres. (C) Ruminococcus flavefaciens
1. (D) Ruminococcus flavefaciens 2. Note that the extensive secretion of exopolysaccharide glycocalyx
material by these organisms has produced a fibrous network (C and D), arrows seen here after dehydration
during preparation of SEM. (Bars = 5 ?m; source: Kudo et al 1987b).
Figure 3. Scanning electron micrographs (SEM) of bacterial detachment from Whatman No. 1 filter paper.
(A?C) Filter paper was first incubated with cellulolytic bacteria at 39?C for 20 h and then treated with 0.1%
methylcellulose at 39?C for 1 h. These SEM show the very effective detachment of cellulolytic bacteria from
their cellulosic substrata that is caused by methylcellulose. (A) Fibrobacter succinogenes BL2. Note very
regularly formed parallel grooves, as evidence of fibre degradation, and the complete absence of adherent
bacteria. (B) Ruminococcus albus SY3. There is some irregular evidence of digestion, in the form of grooves
or pits, and only a few adherent cells remained. (C) Ruminococcus flavefaciens 1. Note that some cells are
still attached to the fibre by means of fibrous extracellular material that formed a network over the cellulose
surface and that no grooves or pits are formed as a result of digestion. (D) Filter paper was incubated with
cells of Fibrobacter succinogenes at 39?C for 20 h, in the presence of 0.1% (w/v) methylcellulose, and no
bacterial adhesion or evidence of digestion can be seen on the cellulose fibre surfaces. (Bars = 5 ?m; source:
Kudo et al 1987b.)
producers adhere to starch but not to cellulose, while cellulose decomposers adhere to
cellulose but not to starch (Minato and Suto 1978). Cheng and Costerton (1980)
postulated that rumen microorganisms can be classified into three groups: (i) microor-
ganisms attaching to the rumen wall; (ii) microorganisms living freely in the rumen;
and (iii) microorganisms attaching to feed particles. In the rumen, as much as 75% of
these microorganisms are attached to feed particles. Microorganisms in the rumen have
a variety of surfaces to which they may attach, and a distinct population of microorgan-
isms adheres to each different surface. From an ecological viewpoint, bacteria with the
ability to attach to feed particles have a great advantage over non-attaching microorgan-
isms, which can flow more quickly from the rumen. In addition to adhesion, ingestion
of feed particles was observed in protozoa (Figure 4). The importance of the adhesion
of microorganisms to substrates suggests that there is a possibility for better feed
efficiency if we can increase the attachment.
Digestive consortia
Pure cultures of cellulolytic bacteria and fungi digest cellulose in vitro but digestion
does not proceed at a similar rate to that seen in the rumen unless consortia are formed
with non-cellulolytic Treponema bryantii, Butyrivibrio fibrisolvens and methanogenic
bacteria. Electron microscopy of partly digested plant material shows that consortia of
F. succinogenes, R. flavefaciens and R. albus are in direct contact with the cellulose
fibres, while cells of Treponema species, Butyrivibrio species and methanogenic bac-
teria are more loosely associated. Morphological examination, by electron microscopy,
showed that the cells of T. bryantii associate with the plant cell wall materials in straw,
but that cellulose digestion occurs only when these organisms are present with cellu-
lolytic species such as F. succinogenes. These results show that cellulolytic bacteria
interact with non-cellulolytic Treponema to promote the digestion of cellulosic materials
(Kudo et al 1987a). Recent studies have shown that both protozoa and fungi may be
active members of consortia (Ho et al 1988b; Imai et al 1989; Kudo et al 1990a). We
have recently discovered that mycoplasma tend to be associated with most rumen fungal
cultures  (Kudo  et al  1990b). We  have confirmed that  the  rumen fungal cultures,
reportedly pure, were also contaminated with these cell wall-deficient bacteria. Thus,
the microbial ecology of the rumen is obviously very complex, including many inter-
actions of microbial species and even though primary cellulolytic organisms may be
present in a system, many other factors may be required to facilitate actual cellulose
digestion.
Practical applications in animal production
The results of up to date ruminant production indicate that immediate improvements
in feed efficiency in ruminants may not be expected. However, there is some potential.
Some of the forages and by-products, for example Leucaena and palm oil by-products,
may require manipulations to reduce the toxicity. PKC is used widely as animal feed
but its high concentration of copper and imbalance of Ca/P cannot be ignored.
Figure 4. Light photomicrographs showing rumen ciliates that are uniquely found in cattle fed only barley
straw. All ciliates contain fragments of barley straw (arrows) inside the cell. (A) Ostracodinium clipeolum.
(B) Ostracodinium obtusum. (C) Polyplastron multivesiculatum, in binary fission. (D?G) Polyplastron
multivesiculatum. (H) Group of Diploplastron and Diplodinium. Structures surrounding the protozoa are
considered to be molds. (I) Polyplastron multivesiculatum. Transformation of the shape of cell outward due
to the plant material ingested is noted (source: Kudo et al 1990a).
Alfalfa (Medicago sativa L.) is a valuable leguminous forage crop with a high yield
and an excellent nutritive value. However, the risk of bloat caused by the high initial
rates of microbial digestion and nutrient release from the more digestible leaves often
limits its use. We (Kudo et al 1985) have shown that through plant breeding programmes
this adverse effect can be limited in order to reduce the high initial rate of digestion and
provide an engineered forage that is equivalent to the original plant but does not have
an adverse effect on the ruminant. This concept could be applied to low-quality forages
in tropical countries by similar plant breeding programmes to increase the initial rate of
digestion. Tropical ruminants are known to be better converters of poor quality feed.
Therefore, breeding based on this fact and the better chewing characteristics of certain
individual animals may improve animal production in tropical countries.
Problems in rumen biotechnology
The present approach in biotechnology has many problems if we expect the genetic
engineered microorganisms to survive in the rumen. For the successful development of
rumen microbiology, it is very important to study and to include more basic charac-
teristics of rumen microorganisms including the ecosystem. Otherwise, further devel-
opment in rumen and ruminant studies would not be expected (Ling 1994).
Electron microscopy revealed that when grass leaves (Akin 1976a, 1976b; Akin and
Amos 1975; Akin et al 1974) or straw (Cheng et al 1984) were exposed to natural
populations of rumen bacteria, the thick highly structured cellulosic cell walls of these
plant materials were seen to be colonised by pleomorphic cells of F. succinogenes.
Attempts to isolate F. succinogenes from the rumen and to maintain active cellulolytic
activity under laboratory conditions have not always been successful (van Gylswyk
and Schwartz 1984; Stewart et al 1981). These difficulties limited the number of F.
succinogenes used in in vitro studies to only a few strains (Bryant and Doetsch 1954;
Bryant et al 1959; Dehority 1963; van Gylswyck and Schwartz 1984; Mackie et al
1978; Stewart et al 1981). Many of the cellulolytic bacteria were isolated a long time
ago. For example, F. succinogenes S85, which is one of the most popular cellulolytic
bacterium in the laboratory, was isolated over at least 40 years ago (Bryant and Doetsch
1954). Growth of fresh isolates of F. succinogenes in liquid glucose medium is very
good but some strains would not grow after eight to ten subcultures in this medium. F.
succinogenes S85 can survive in glucose medium although great loss of cellulolytic
activity is evident. Most strains of fresh isolates of Ruminococcus sp. cannot utilise
glucose (Hungate 1966) but most ?laboratory strains? of Ruminococcus sp. can utilise
glucose. Continued growth of N. patriciarum in vitro, with glucose as carbon source,
resulted in a loss of ability to utilise cellulose for growth, but little loss of activity was
evident when grown on plant tissues or cellulose (Orpin and Ho 1991). We have
confirmed this with other rumen fungi, Neocallimastix, Piromyces, Sphaeromonas and
Orpinomyces. In addition, cellulolytic microorganisms do not belong to the population
freely living in the rumen but they belong to the population attached to feed particles.
Therefore, we should use only insoluble cellulosic materials when subculturing cellu-
lolytic microorganisms.
The bacteria that comprise the wall-associated biofilm population produce very large
amounts of exopolysaccharide glycocalyx material when grown in culture immediately
after isolation from the rumen (McCowan et al 1980). However, this elaboration of
slimy material was seen to decrease sharply upon subculture and almost entirely after
five to ten subcultures. We have shown the possible mutation of some strains of
Bifidobacterium sp. which lacks extracellular glycocalyx material (Kudo et al 1989a;
Figure 5). This ?naked? Bifidobacterium sp. is probably a mutant strain produced and
adapted under laboratory conditions. To generate and maintain a glycocalyx, a bacterial
cell must expend energy, and in the protected environment of pure culture the glyco-
calyx is a metabolically expensive luxury conferring no selective advantage. Cells that
fabricate these elaborate coatings are usually eliminated from pure cultures by un-
coated mutants that can devote more of their energy budget to proliferation. The
bacterial glycocalyx was ignored because the familiar pure laboratory strains do not
need it and therefore do not fabricate it (Costerton et al 1978). Therefore, the ?naked?
laboratory strains of bacteria have an advantage in laboratory conditions as an energy
saver but not in natural conditions. Another example of mutation in the laboratory is
Streptococcus bovis. On initial isolation the rumen strains of this species are sensitive
to oxygen. Later, this sensitivity diminishes, and the cells grow aerobically (Hungate
1966).
Most researchers subculture fungi at five-day intervals. Fresh isolates of fungi
require three-day intervals of subculture if we do not want to lose any of the cultures.
Practically this is difficult and sometimes the subcultures are done at four- to six-day
intervals and we lose the strains. So finally we have the cultures which can survive
Figure 5. Transmission electron micrographs (TEM) of ruthenium red-stained pure culture of Bifidobac-
terium. (A) Very thick irregular Gram-positive cell wall (ca. 0.2 ?m). (B) The thin regular Gram-positive
cell wall (ca. 0.05 ?m) and the formation of spaces between the cell and the cytoplasm (arrows). This naked
bacterial strain is probably a mutant strain produced and adapted under laboratory conditions (source:
Kudo et al 1989a).
Figure 6. Simple isolation method for rumen cellulolytic bacteria. Rumen sample
?
: Take one part of rumen
fluid and one part of solid contents (preferably taken from the upper one-third of rumen contents), blend for
30 s at low speed, 30 s rest, 30 s at high speed, squeeze through two layers of cheese cloth or gauze and use
as inoculum. Incubation time may vary. If F. succinogenes is desired use Whatman No. 5 or 6 cellulose paper
(crystalline cellulose) and use amorphous (e.g. ball-milled) cellulose for Ruminococcus albus, which cannot
digest crystalline cellulose. Use only solid cellulosic materials in the medium for subculture and never use
soluble carbohydrates (especially glucose) in the medium as these dramatically reduce cellulase a ctivities.
five-day intervals of subculture. Thus, laboratory strains, especially cellulolytic
bacteria, that are currently used have many problems. Therefore, we have developed
a simple isolation method for  cellulolytic bacteria using medium enriched  with
Whatman cellulose filter paper (the type of the papers should be chosen according to
the target cellulolytic bacteria) as the sole selective substrate (Figure 6). Once
isolated, soluble carbohydrates (especially glucose) should not be used as carbon
sources in the medium as these reduce the enzyme activities.
The author (H.K.) criticised genetic engineering but this does not mean the denial of
biotechnology. We should study and include more basic characteristics of rumen
microorganisms, including ecosystems which have been ignored totally. Genetically
engineered microorganisms from ?laboratory strains? might not survive in the rumen if
we introduce them back into the rumen. Even laboratory strains might not survive in
the rumen any more. Generally, in case of aerobic microorganisms, substrates are
utilised effectively towards production of microbial cells and metabolic products are
mostly released as carbon dioxide and water. In contrast, anaerobic bacteria produce
less microbial cells and large amounts of organic materials (ethanol, methane, volatile
fatty acids, non-volatile fatty acids, etc.) are accumulated. Thus, rumen microorganisms
have a potential for practical applications in other industries such as food and fermen-
tation industries, although up to now they and other related microorganisms have not
been used in other areas.
Acknowledgments
We would like to thank Drs B. Tangendjaja and B. Bakrie, Research Institute for Animal
Production, Bogor, Indonesia for the preparation of rumen fluid adapted to Leucaena.
We are very grateful to Drs N. Abdullah and Y.W. Ho, Universiti Pertanian Malaysia,
Selangor, Malaysia for the cooperative work on rumen microorganisms. Mineral com-
position of PKC was analysed by Dr M. Ivan, Agriculture Canada, Canada, and Dr T.
Kawashima, Japan International Research Center for Agricultural Sciences, Japan. A
part of the work was conducted under the joint project of ?the improvement of the
productivity of tropical ruminants? between Japan International Research Center for
Agricultural Sciences (formerly called Tropical Agriculture Research Center), Japan,
Universiti Pertanian Malaysia, Malaysia and Agriculture Canada, Canada. We are
grateful to the Department of Protected Wild Life and National Park, Malaysia for
permission to capture and export mousedeer to Japan and to the Ministry of Agriculture,
Forestry and Fisheries, Japan for the permission to import mousedeer.
References
Abdullah N. 1987. Rumen bacteria and fungi of cattle and buffaloes. Pertanika 10:49?52.
Abou-Akkada A.R. and Howard B.H. 1960. The biochemistry of rumen protozoa. 3. The
carbohydrate metabolism of Entodinium. Biochemical Journal 76:445?451.
Akin D.E. 1976a. Ultrastructure of rumen bacterial attachment to forage cell walls. Applied and
Environmental Microbiology 31:562?568.
Akin D.E. 1976b. Ultrastructure of rigid and lignified forage tissue degradation by a filamentous
rumen microorganism. Journal of Bacteriology 125:1156?1162.
Akin D.E. and Amos H.E. 1975. Rumen bacterial degradation of forage cell walls investigated
by electron microscopy. Applied and Environmental Microbiology 27:1149?1156.
Akin D.E., Burdick D. and Micheals G.E. 1974. Rumen bacterial interrelationships with plant
tissue during degradation revealed by transmission electron microscopy. Applied Microbiol-
ogy 27:1149?1156.
Allison M.J., Mayberry W.R., McSweeney C.S. and Stahl D.A. 1992. Synergistes jonesii, gen.
nov.: a rumen bacterium that degrades toxic pyridinediols. System. Applied Microbiology
15:522.
Anderson S. and Jones J.K. Jr. 1967. Recent Mammals of the World. A Synopsis of Families.
Ronald Press, New York, USA. 263 pp.
Banerjee A.K. 1955. Studies on parasitic ciliates from Indian ruminants. Proceedings of the
Zoological Society, Calcutta 8:87?101.
Bryant M.P. and Doetsch R.N. 1954. A study of actively cellulolytic rod-shaped bacteria of the
bovine rumen. Journal of Dairy Science 37:1176?1183.
Bryant M.P., Robinson I.M. and Chu H. 1959. Observations on the nutrition of Bacteroides
succinogenes?a ruminal cellulolytic bacterium. Journal of Dairy Science 42: 1831?1847.
Cadigan F.C. 1972. A brief report on copulatory and perinatal behavior of the lesser Malayan
mousedeer (Tragulus javanicus). Malayan Nature Journal 25:112?116.
Cheng K.-J. and Costerton J.W. 1980. Adherent rumen bacteria: their role in the digestion of
plant material, urea, and epithelian cells. In: Ruckebush Y. and Thivend P. (eds), Digestive
Physiology and Metabolism in Ruminants. MTPPress, Lancaster, England, UK. pp. 227?250.
Cheng K.-J., McAllister T.A., Kudo H. and Costerton J.W. 1990a. The importance of adhesion
in the microbial digestion of plant materials. In: Hoshino S., Onodera R., Minato H. and
Itabashi H. (eds), The Rumen Ecosystem. Japan Scientific Societies Press, Tokyo, Japan. pp.
129?135.
Cheng K.-J., Forsberg C.W., Minato H. and Costerton J.W. 1990b. Microbial ecology and
physiology of feed degradation within the rumen. In: Tsuda T., Sasaki Y. and Kawashima R.
(eds), Physiological Aspects of Digestion and Metabolism in Ruminants. Proceedings of the
Seventh International Symposium on Ruminant Physiology, Sendai, Japan, 28 August?1
September 1989. Academic Press, San Diego, California, USA. pp. 595?624.
Cheng K.-J., Kudo H., Jakober K.D., Duncan S.H., Mesbah A., Stewart C.S., Bernalier A., Fonty
G. and Gouet P. 1991. Prevention of fungal colonization and digestion of cellulose by the
addition of methylcellulose. Canadian Journal of Microbiology 37:484?487.
Cheng K.-J., Stewart C.S., Dinsdale D. and Costerton J.W. 1984. Electron microscopy of
bacteria involved in the digestion of plant cell walls. Animal Feed Science and Technology
10:93?120.
Costerton J.W., Geesey G.G. and Cheng K.-J. 1978. How bacteria stick. Scientific American
238:86?95.
Davis J.A. 1965. A preliminary report of the reproductive behavior of the small Malayan
Chevrotain Tragulus javanicus at the New York Zoo. International Zoo Yearbook 7: Appendix
IV.
Dehority B.A. 1963. Isolation and characterization of several cellulolytic bacteria from in vitro
rumen fermentations. Journal of Dairy Science 46:217?222.
Fukuta K., Kudo H. and Fujina N.H. 1991. Behavior of captive lesser mousedeer, Tragulus
javanicus Experimenta. Herbivora 16:47?50 (in Japanese).
Goto N. and Hashizume R. 1978. Pattern of ovulation in Microtus montebelli. Journal of
Mammalogy 7:181?188 (in Japanese with English summaries).
van Gylswyck N.O. and Schwartz H.M. 1984. Microbial ecology of the rumen of animals fed
high-fibre diets. In: Gilchrist F.M.C. and Mackie R.I. (eds), Herbivore Nutrition in the
Subtropics and Tropics. The Science Press, Crighall, South Africa. pp. 359?377.
Hammond A.C., Allison M.J., Williams M.J., Prine G.M. and Bates D.B. 1989. Prevention of
Leucaena toxicosis of cattle in Florida by ruminal inoculation with 3-hydroxy-4-(1H)-pyri-
done degrading bacteria. American Journal of Veterinary Research 50:2176?2180.
Henke L.A. 1958. Value of Leucaena glauca as a feed for cattle. Proceedings of the 8th Congress
of Pacific Scientists. Vol. 8. p. 46.
Ho Y.W., Abdullah N. and Jalaludin S. 1988a. Penetrating structures of anaerobic rumen fungi
in cattle and swamp buffalo. Journal of General Microbiology 134:177?181.
Ho Y.W., Abdullah N. and Jalaludin S. 1988b. Colonization of guinea grass by anaerobic
rumen fungi in swamp buffalo and cattle. Animal Feed Science and Technology
22:161?171.
Hungate R.E. 1966. The Rumen and its Microbes. Academic Press, New York, USA. 533 pp.
Hungate R.E. 1978. The rumen protozoa. In: Kreier J.P. (ed), Parasitic Protozoa Vol. 2. Academic
Press, New York, USA. pp. 655?695.
Imai S. 1985. Rumen ciliate protozoal faunae of Bali cattle (Bos javanicus domesticus) in
Indonesia, with the description of a new species, Entodinium javanicum sp. nov. Zoological
Sciences 2 (4):591?600.
Imai S. 1986. Rumen ciliate protozoal fauna of zebu cattle (Bos taurus indicus) in Sri Lanka,
with the description of a new species, Diplodinium sinhalicum sp. nov. Zoological Sciences
3 (4):699?706.
Imai S. 1988. Ciliate protozoa in the rumen of Kenyan zebu cattle, Bos Taurus indicus, with the
description of four new species. Journal of Protozoology 35 (1):130?136.
Imai S., Abdullah N., Ho Y.W., Jalaludin S., Hussain H.Y., Onodera R. and Kudo H. 1995.
Comparative study on the rumen ciliate populations in small experimental herds of water
buffalo and Kedah Kelantan cattle in Malaysia. Animal Feed Science and Technology
52:345?351.
Imai S., Kudo H., Fukuta K., Abdullah N., Ho Y.W. and Onodera R. 1995. Isotricha jalaludinii
n. sp. found from the rumen of lesser mousedeer, Tragulus javanicus, in Malaysia. Journal
of Eukaryotic Microbiology 42: 75?77.
Imai S., Han S.S., Cheng K.-J. and Kudo H. 1989. Composition of the rumen ciliate population
in experimental herds of cattle and sheep in Lethbridge, Alberta, Western Canada. Canadian
Journal of Microbiology 35:686?690.
Imai S., Tsutsumi Y., Yumura S. and Mulenga A. 1992. Ciliate protozoa in the rumen of Kafue
lechwe, Kobus leche kafuensis, in Zambia, with the description of four new species. Journal
of Protozoology 39 (5):564?572.
Jalaludin S., Ho Y.W., Abdullah N. and Kudo H. 1992. Strategies for animal improvement in
Southeast Asia. Tropical Agriculture Research Series 25:67?76.
Jones R.J. 1981. Does ruminal metabolism of mimosine explain the absence of Leucaena
toxicity in Hawaii? Australian Veterinary Journal 57:55?56.
Jones R.J. and Megarrity R.G. 1986. Successful transfer of DHP-degrading bacteria from
Hawaiian goats to Australian ruminants to overcome the toxicity of Leucaena. Australian
Veterinary Journal 63:259?262.
Kofoid C.A. and MacLennan R.F. 1930. Ciliates from Bos indicus Linn. 1. The genus Ento-
dinium Stein. University of California Publications in Zoology 33:471?544.
Kofoid C.A. and MacLennan R.F. 1932. Ciliates from Bos indicus Linn. 2. A revision of
Diplodinium Schuberg. University of California Publications in Zoology 37:53?152.
Kofoid C.A. and MacLennan R.F. 1933. Ciliates from Bos indicus Linn. 3. Epidinium Crawley,
Epiplastron gen. nov., and Ophryoscolex Stein. University of California Publications in
Zoology 39:1?33.
Kraneveld F.C. and Djaenoedin R. 1947. Haaruitval bij dieren na voedering met lamtoro
(Leucaena glauca). Hemera Zoa 54:170?184.
Kudo H. and Oki Y. 1981. Fermentation and VFA production in the esophageal sac Microtus
montebelli fed different rations. Japanese Journal of Veterinary Science 43:299?305.
Kudo H. and Oki Y. 1982. Breeding and rearing of Japanese field vole (Microtus montebelli
Milne Edwards) and  Hungarian vole (Microtus montebelli Pallas) as new herbivorous
laboratory animal species. Experimental Animals 31:61?64.
Kudo H. and Oki Y. 1984. Microtus species as new herbivorous laboratory animals: reproduc-
tion; bacterial flora and fermentation in the digestive tracts; and nutritional physiology.
Veterinary Research Communications 8:77?91.
Kudo H., Cheng K.-J. and Costerton J.W. 1987a. Interactions between Treponema bryantii sp.
B25 and cellulolytic bacteria in the in vitro degradation of straw cellulose. Canadian Journal
of Microbiology 33:244?248.
Kudo H., Cheng K.-J. and Costerton J.W. 1987b. Electron microscopic study of methylcellu-
lose-mediated detachment of cellulolytic rumen bacteria from cellulose fibres. Canadian
Journal of Microbiology 33:267?272.
Kudo H., Cheng K.-J., Hanna M.R., Howarth R.W., Goplen B.P. and Costerton J.W. 1985.
Ruminal digestion of alfalfa strains selected for slow and fast initial rates of digestion.
Canadian Journal of Animal Science 65:157?161.
Kudo H., Cheng K.-J., Imai S., Han S.S. and Costerton J.W. 1990a. Effects of feed on the
composition of the rumen ciliate protozoal population in cattle and its relationship to
cellulolytic ciliate protozoa. Animal Feed Science and Technology 29:159?169.
Kudo H., Jakober K.D., Phillippe R.C., Cheng K.-J., Barr D.J.S. and Costerton J.W. 1990b.
Isolation and characterization of cellulolytic anaerobic fungi and associated mycoplasmas
from the rumen of a steer fed a roughage diet. Canadian Journal of Microbiology 36:513?517.
Kudo H., Tangendjaja B. and Mutalib A.R. 1990c. Mimosine degradation in the rumen. In:
Hoshino S., Onodera R., Minato H. and Itabashi H. (eds), The Rumen Ecosystem. Japan
Scientific Societies Press, Tokyo, Japan. pp. 73?81.
Kudo H., Cheng K.-J., Majak W., Hall J.W. and Costerton J.W. 1984. Degradation of mimosine
in rumen fluid from cattle and sheep in Canada. Canadian Journal of Animal Science
64:937?942.
Kudo H., Cheng K.-J., Majak W., Hall J.W. and Costerton J.W. 1986. In vitro degradation of
mimosine by microorganisms from the esophageal sac of voles (Microtus arvalis). Canadian
Journal of Animal Science 66:547?551.
Kudo H., Kimura N., Suzuki M., Cheng K.-J., Costerton J.W. and Mitsuoka T. 1989a. Electron
microscopic biochemical and physiological studies of Bifidobacterium pseudolongum SS-24
and Bifidobacterium thermophilum SS-19. Zentralblatt f r Bakteriologie 271:263?271.
Kudo H., Jalaludin S. and Mutalib A.R. 1989b. Absence of detoxification of DHP by rumen
microorganisms from Malaysian cattle, water buffaloes and sheep. Japan Agricultural Re-
search Quarterly 23:78?80.
Kudo H., Majak W., MutalibA.R., Ho Y.W. and Cheng K.-J. 1989c. Microorganisms and degradation
of deleterious principles. In: Devendra C. and Imaizumi E. (eds), Ruminant Physiology and
Nutrition in Asia. Japan Society of Zootechnical Science, Tokyo, Japan. pp. 73?87.
Lekagul B. and McNeely J.A. 1977. Mammals of Thailand. Kuruspha, Bangkok, Thailand. 97
pp.
Ling J.R. 1994. The golden anniversary of rumen and ruminant studies. Letters in Applied
Microbiology 19:291?293.
Mackie R.I., Gilchrist F.M., Roberts A.M., Hannah P.E. and Schwartz H.M. 1978. Microbio-
logical and chemical changes in the rumen during the stepwise adaptation of sheep to high
concentrate diets. Journal of Agricultural Science 90:241?254.
McAllister T.A., Rode L.M., Major D.J., Cheng K.-J. and Buchanan-Smith. 1990. Effect of
ruminal  microbial colonization on cereal grain digestion. Canadian Journal of Animal
Science 70:571?579.
McCowan R.P., Cheng K.-J. and Costerton J.W. 1980. Adherent bacterial populations of bovine
rumen wall: distribution patterns of the adherent bacteria. Applied and Environmental
Microbiology 39:233?241.
Minato H. and Suto T. 1978. Technique for fractionation of bacteria in rumen microbial
ecosystems. II. Attachment of bacteria isolated from bovine rumen to cellulose powder in
vitro and elution of bacteria attached therefrom. Journal of General and Applied Microbiol-
ogy 24:1?16.
Montgomery L., Flesher B. and Stahl D. 1988. Transfer of Bacteroides succinogenes (Hungate)
to Fibrobacter succinogenes gen. nov. International Journal of Systematic Bacteriology
38:430?435.
Murayama S. and Zahari Abu Barker. 1993. Decomposition characteristics of tropical peats.
Final Report of Joint Research Project Between TARC and MARDI. Tropical Agriculture
Research Centre, Tsukuba, Japan. 92 pp.
Namikawa T. 1980. General aspects of domestication and phylogeny in cattle. Japanese Journal
of Zootechnical Science 51:235?246 (in Japanese).
Ogimoto K. and Imai S. 1981. Atlas of Rumen Microbiology. Japan Scientific Societies Press,
Tokyo, Japan. 231 pp.
Orpin C.G. 1977. Invasion of plant tissue in the rumen by the flagellate Neocallimastix frontalis.
Journal of General Microbiology 98:423?430.
Orpin C.G. and Ho Y.W. 1991. Ecology and function of the anaerobic rumen fungi. In: Ho Y.W.,
Wong H.K., Abdullah N. and Tajuddin Z.A. (eds), Recent Advances on the Nutrition of
Herbivores. Vinlin Press, Kuala Lumpur, Malaysia. pp. 163?170.
Oshio S., Takigawa A., Abe A., Nakanishi N., Hassan O.A., Jaafar D.M. and Dahlan I. 1990.
Processing and Utilization of Oil Palm by-Products for Ruminants. MARDI-TARC collabo-
rative study. Dicetak oleh Percetakan Seremban Sdn. Bhd, Selangor, Malaysia. 110 pp.
Pratchett D., Jones R.J. and Syrch F.X. 1991. Use of DHP-degrading rumen bacteria to overcome
toxicity in cattle grazing irrigated Leucaena pasture. Tropical Grasslands 25:268.
Quirk M.F.M., Bushell J.J., Jones R.J., Megarity R.G. and Butler K.L. 1988. Live weight gains
on Leucaena and native grass pastures dosing cattle with rumen bacteria capable of degrading
DHP, a ruminal metabolism from Leucaena. Journal of Agricultural Science (Cambridge)
111:165?170.
Ralls K., Barasch C. and Minkowski K. 1975. Behavior of captive mousedeer, Tragulus napu.
Zeitschrift f r Tierpsychologie 37:356?378.
van Soest P.J. 1982. Nutritional Ecology of the Ruminants. O and B Books, Oregon, USA. 374
pp.
Stewart C.S., Paniagua C., Dinsdale D., Cheng K.-J. and Garrow S.H. 1981. Selective isolation
and characteristics of Bacteroides succinogenes from the rumen of a cow. Applied and
Environmental Microbiology 41:504?510.
Tubb J.A. 1966. Notes on breeding the mousedeer in captivity. Journal of the Siam Society of
Natural History 21:87?91.
Vidyadaran M.K., Hilmi M.  and Sirimane R.A.  1981.  Dentition of the  Malaysian  lesser
mousedeer (Tragulus javanicus). Pertanika 4:47?52.
Vidyadaran M.K., Hilmi M. and Sirimane R.A. 1982. The gross morphology of the stomach of
the Malaysian lesser mousedeer (Tragulus javanicus). Pertanika 5:34?38.
An important role for ruminal anaerobic
fungi in the voluntary intake of poor
quality forages by ruminants
G.L.R. Gordon1, C.S. McSweeney2 and M.W. Phillips1
1.?DivisionofAnimalProduction
Commonwealth Scientific and Industrial Research Organisation (CSIRO)
Prospect, Sydney, New South Wales 2148, Australia
2.?Division ofTropicalAnimalProduction
Commonwealth Scientific and Industrial Research Organisation (CSIRO)
Long Pocket, Brisbane, Queensland 4068, Australia
Abstract
Anaerobic fungi are prevalent in the rumen of grazing ruminants where they
are closely associated with plant particles. An important role for these unusual
microorganisms in ruminant nutrition has been proposed over the course of the
past decade or so. The significance of this role was based largely on in vitro
studies of anaerobic fungi. However, there was very little direct evidence from
studies with animals to support this position until recently. It is now known that
anaerobic fungi make a significant contribution to the intake of poor quality
pasture. There may also be a contribution to forage digestion in the rumen,
although much of the fermentation of structural polysaccharides in the rumen
is attributed to anaerobic bacteria, either alone or in association with ciliate
protozoa. Fungal populations in the rumen are known to be limited by a low
sulphur content of both rough pastures and poor quality crop residues. Where
appropriate, fertilisation of the growing pasture with sulphur improves the
sulphur content of feed which subsequently is eaten in greater quantities by
ruminants with ruminal populations of anaerobic fungi returned to ?normal?.
Alternatively, supplementing the feed with a suitable sulphur compound is also
effective in improving the voluntary intake of digestible feed both with crop
residues and harvested rough pastures. In addition, considerable potential exists
for the selection and intra-ruminal inoculation of superior strains of anaerobic
fungi obtained from either domesticated or wild herbivorous mammals. These
possible mechanisms for improving forage intake by ruminants grazing rough
pastures through the manipulation of anaerobic fungi in the rumen have
implications for ruminant production in Africa.
Introduction
Anaerobic chytridiomycete fungi have been known to be a component of the rumen
microbiota for 20 years (Orpin 1975). Together with the anaerobic bacteria and ciliate
protozoa, they are responsible for the production of hydrolytic enzymes which degrade
dietary polysaccharides and other carbohydrates and the fermentation of the resulting
monosaccharides. An important role for anaerobic fungi in the rumen has long been
proposed (Akin et al 1983; Bauchop 1979; Orpin 1977). Until recently, there was little
or no direct evidence for the magnitude of the contribution made by anaerobic fungi to
digestion in the rumen. However, it is now known that anaerobic fungi are extremely
important for the voluntary feed intake of poor quality, mature herbage by sheep
(Gordon and Phillips 1993).
In our paper we present some of the evidence which supports this conclusion and
discuss some of the possible reasons for it. Also, we consider some ways in which the
benefits obtainable from anaerobic fungi could be harnessed in domesticated ruminants
in semi-arid regions of Africa. There are a number of recent reviews (Fonty and Joblin
1991; Li and Heath 1993; Teunissen and Op Den Camp 1993; Trinci et al 1994; Wubah
et al 1993) which should be consulted for detailed information on the biology of
anaerobic fungi.
Contribution of anaerobic fungi to voluntary
feed intake and the nutrition of ruminants
A definite positive relationship exists between the presence of anaerobic fungi in the
rumen and the voluntary intake of herbage diets of low digestibility (Akin et al 1983;
Gordon 1985; Gordon and Phillips 1993; Morrison et al 1990; Weston et al 1988). This
is quite possibly a result of fungal attack of lignified plant tissues (Akin 1987; Akin and
Borneman 1990) with the resultant weakening of these tough plant components (Akin
et al 1983, 1989).
Table 1 shows that the removal of anaerobic fungi from the rumen of sheep reduced
the voluntary intake of poor quality feed by about 30%, with little effect on the
populations of bacteria and ciliate protozoa. Ruminal fungal activity is often accompa-
nied by increased feed digestibility in vivo (Gordon and Phillips 1993; Weston et al
1988). An oral inoculum of anaerobic fungus stimulated hay intake in early weaned
calves (Theodorou et al 1990). Aplan for selecting appropriate strains of anaerobic fungi
for inoculation into the rumen of mature ruminants at pasture has been proposed
(Gordon 1990).
Anaerobic fungi have the potential of contributing to the protein supply of the host
animal. They are proteolytic and may contribute to ruminal protein degradation (Wallace
and Munro 1986), although the extent of this contribution remains to be conclusively
determined (Bonnemoy et al 1993). Fungal cells are composed of proteins with a well
balanced combination of amino acids (Gulati et al 1989, 1990; Kemp et al 1985) which
are highly digestible and available to the ruminant host (Gulati et al 1989, 1990). Until
an accurate method of measuring the biomass in the rumen is developed (Faichney et
al 1991), the possible extent of the fungal contribution to protein supply is largely
conjecture. However, should an increase in the biomass of ruminal fungi prove to be
feasible, it is unlikely that the supply of high-quality microbial protein to the host
ruminant would be diminished.
92 G.L.R. Gordon, C.S. McSweeney and M.W. Phillips
Effect of defaunation of the rumen
on populations of anaerobic fungi
The treatment of ruminants with defaunating agents (usually ionic detergents) to
remove ciliate protozoa from the rumen (Bird 1989) has a secondary effect; it results
in increased populations of ruminal fungi (Soetanto et al 1985). Therefore an inverse
relationship between the sizes of the fungal and the ciliate populations in the rumen is
apparent, at least in herbage-fed animals. This relationship has been confirmed in
several studies (Hsu et al 1991; Romulo et al 1986, 1989). Defaunation can also be
accomplished through non-chemical methods and the same general inverse relation-
ship between anaerobic fungi and ciliates is frequently (Newbold and Hillman 1990;
Ushida et al 1990) but not always (Ushida et al 1990; Williams and Withers 1991)
observed.  The  underlying mechanism by which defaunation increases the fungal
population is possibly a reduction in the turnover of fungal protein in the rumen
(Newbold and Hillman 1990). Therefore, it is likely that the nutritional benefits due
to defaunation of animals being fed poor-quality, low-nitrogen herbage is at least in
Table 1. The effect of the removal of anaerobic fungi from the rumen of sheep fed on a straw-based diet and
the subsequent reinoculation of a Neocallimastix sp. into the rumen on voluntary feed intake, feed digestion,
rumen parameters and microbial populations.
Parameter A. Pretreatment B. No fungi C. With fungi added
Intake (g/d)
OM 894 (30.8) 628 (36.4) 877 (52.6)
ADF 390 (9.3) 264 (16.4) 373 (22.6)
Digestibility (%)
OM 53.2 (1.30) 50.3 (1.38) 57.3 (1.02)
ADF 51.2 (1.03) 46.5 (1.28) 55.0 (1.64)
Rumen pH 6.8 (0.05) 7.0 (0.05) 6.6 (0.07)
Rumen ammonia (mg N/l) 46 (4.9) 50 (10.8) 32 (9.0)
Rumen VFA (mM) 69 (5.1) 55 (10.8) 78 (2.4)
molar proportions:
acetic 0.69 0.56 0.72
propionic 0.19 0.34 0.19
butyric 0.09 0.07 0.07
branched chain 0.03 0.03 0.02
Anaerobic fungi
Rumen (zoospores/ml; ?10
3
) 7.6 (4.0) UD 19 (5.7)
Faeces (zoospores/g; ?10
3
) 5.2 (2.9) UD 11 (5.6)
Bacteria (cells/ml)
total viable (?10
9
) 0.8 (0.18) 1.4 (0.37) 1.6 (0.33)
cellulolytic (?10
8
) 0.4 (0.19) 2.6 (1.48) 1.1 (0.29)
Ciliate protozoa
(cells/ml; ?10
5
) 3.8 (0.92) 3.0 (0.53) 4.8 (0.49)
Data are mean (SE) for four sheep.
ADF: acid-dergent fibre; OM: organic matter; VFA: volatile fatty acid; UD: undetectable.
Data from Gordon and Phillips 1993.
Important role for ruminal anaerobic fungi in voluntary intake of poor quality forages 93
part a result of increased fungal numbers with their associated degradative activity
against plant fibre.
Effect of dietary sulphur
on anaerobic fungi in the rumen
Early in the study of anaerobic fungi, it was recognised that the sulphur content of hay
diets or pasture was a significant factor governing the fungal population in the rumen
(Akin et al 1983; Gordon 1985). When sulphur was present in the diet at levels of around
1.0 g S per kg organic matter or less, anaerobic fungi were apparently absent from the
rumen of sheep fed on hay made from the tropical pasture grass Digitaria pentzii (Akin
et al 1983). The size of the anaerobic fungal populations in the rumen increased
dramatically after either an application of a sulphur fertiliser to the pasture used to make
the hay (Akin et al 1983) or a sulphur supplement to the low-S hay (Gordon et al 1984).
Fertilisation of the pasture resulted in an average increase of 38% in ad libitum feed
intake (Akin et al 1983; Gordon 1985). A diet of another tropical grass hay (speargrass,
Heteropogon contortus), which had a low sulphur content, also resulted in an undetect-
able fungal population in the rumen (Morrison et al 1990). On the other hand, diets of
cereal straw (usually wheat straw) which were low in sulphur supported a low, but
detectable, population of anaerobic fungi (Gordon et al 1983; Gulati et al 1985; Weston
et al 1988). Some of the results of these studies are summarised in Table 2. The apparent
relationship between the declining sulphur content of a pasture and a declining ruminal
fungal population did not apply to a ryegrass pasture in Scotland where anaerobic fungi
increased with increasing pasture maturity which was accompanied by a declining
sulphur content (Millard et al 1987). Therefore, the form of sulphur in the feed (e.g.
inorganic-S, non protein-S, protein-S) may be as important as its total sulphur content
in determining the ruminal fungal population.
Herbage diets with a low content of sulphur can have a negative effect on the size of
the fungal population in the rumen. In some cases the fungi are apparently absent from
the rumen (Akin et al 1983; Gordon 1985; Morrison et al 1990), whereas in other cases
the fungal populations are greatly reduced (Gulati et al 1985; Weston et al 1988). In all
cases, supplementation of these straws with several different types of sulphur allowed
fungi to proliferate in the rumen and resulted in increased voluntary feed intake. At the
same time, there was little or no change in the ruminal populations of bacteria and ciliate
protozoa due to dietary supplementation (Akin et al 1983; Gulati et al 1985; Morrison
et al 1990). Diets of low-sulphur Digitaria have been successfully supplemented with
methionine and elemental sulphur (each about 1 g S/d per head; Gordon 1985; Gordon
et al 1984), and cereal straw supplemented with either methionine (Gordon et al 1983;
Gulati et al 1985) or sulphate (Weston et al 1988) supported a greatly increased number
of anaerobic fungi in the rumen. Heteropogon (speargrass) supplemented with sulphate
supported a higher fungal population in the rumen compared with the same hay when
unsupplemented (Morrison et al 1990). Anaerobic fungi grown in vitro require reduced
forms of sulphur (Orpin and Greenwood 1986; Phillips and Gordon 1991) indicating
the need for reduction of supplementary sulphate in the rumen before it can be available
94 G.L.R. Gordon, C.S. McSweeney and M.W. Phillips
Table 2. Influence of sulphur in the diet on feed intake of poor quality forages by sheep.
Increased
ruminal population of Change due to S addition
S in feed
?
S added to of fungi after S Voluntary Digestible
Low S feed (g S/kg OM) diet as: addition OM intake OM intake Reference
Digitaria pentzii 0.8?1.1 fertiliser presumed +40% +31% Rees et al (1982)
(grass hay) on pastur known +36% ND Akin and Hogan (1983)
methionine known +11% ND Gordon (1985)
sulphur presumed +6% ?4% Rees et al (1982)
Wheat straw 0.7 methionine known +6% +23% Gulati et al (1985)
(alkali-treated) sulphate known +7% +16% Weston et al (1988)
Heteropogon contortus 0.5 sulphate known +75% +96% Morrison et al (1990)
(speargrass hay)
OM: organic matter; ND: no data.
?
S content of diets increases to 1.3-1.7 g S/kg OM after either fertilisation of the pasture or use of supplement to diet.
I
m
porta
n
t
r
o
le
for
r
u
m
inal
anaer
obic
f
ungi
in
voluntary
i
nta
k
e
o
f
p
o
o
r
q
u
a
lity
forages
9
5
for anaerobic fungi. However, a sulphur supplement which is either specific for
anaerobic fungi in the rumen or relatively so is still to be discovered.
Influence of dietary phenolics on anaerobic fungi
Early studies showed that anaerobic fungi preferentially colonised the sclerenchyma
patches of tropical grass leaves (Akin and Hogan 1983; Akin et al 1983) which suggested
an affinity for lignified tissues. Subsequently, selected strains of anaerobic fungi were
found to solubilise about 35% of the label from a 14C[lignin]lignocellulose preparation
(Gordon 1987; Gordon and Phillips 1989b) whereas the consistent loss of acid-detergent
lignin from wheat straw was not demonstrated (Gordon and Phillips 1989a). However,
around 34% of the lignin component of sorghum stem was removed by the anaerobic
fungus Neocallimastix patriciarum (McSweeney et al 1994) confirming that ?core?
lignin can be attacked by at least some anaerobic fungi. Pure cultures of anaerobic fungi
are unable to mineralise lignin-derived phenolics to CO2 (Bernard-Vailh? et al 1995;
Gordon 1987).
Phenolic compounds, derived from lignocellulose within the rumen, are a potential
barrier to the degradation of structural polysaccharides in the rumen (Wilson 1994).
Ferulic and p-coumaric acids inhibited fibre degradation by mixed anaerobic fungi in
vitro (Akin and Rigsby 1985; Tanaka et al 1991). These phenolic compounds also
inhibited  fibre degradation  by pure cultures  of both monocentric and polycentric
anaerobic fungi, with polycentric strains generally being less sensitive than the mono-
centric strains (G. Gordon, H.K. Wong and M. Phillips, unpublished observations). The
inhibition by plant phenolics is potentially significant because anaerobic fungi produce
powerful hydrolytic enzymes for releasing ferulic and p-coumaric acids from plant cell
walls (Borneman et al 1990). Even though the infusion of free phenolic acids into the
rumen of sheep had no effect on digestion of herbage by sheep (Lowry et al 1993), the
localised concentrations of phenolics liberated by fungal enzymes in the rumen may be
high enough to hinder fungal growth and activity.
The deleterious effects of another form of dietary phenolics, the plant tannins, on rumen
function have been broadly characterised but the ability of the rumen microbial population
to interact with and adapt to the presence of tannins has received little attention. A better
understanding of interactions between tannins and the rumen microbial population will
enable us to capitalise on the beneficial attributes of tannin-containing plants.
The minimum inhibitory concentration of tannin appears to be much higher for fungi
than bacteria (Scalbert 1991). This would imply that anaerobic ruminal fungi may be
more tolerant of tannin than the bacteria which compete with them in the rumen. The
tolerance of rumen fungi to tannins was demonstrated by the ability of Neocallimastix
patriciarum to degrade cellulose effectively in the presence of 100 ug per ml condensed
tannin (CT; from birdsfoot trefoil, Lotus corniculatus; McAllister et al 1994).
Also, it is thought that the protein complexed with tannin is in a form unavailable to the
microorganisms of the rumen. Cleavage of a proportion of these tannin-protein complexes
may be achieved by modifying the microbial population. Protein availability in the rumen
would then be optimised while allowing sufficient complexed plant protein to bypass the
96 G.L.R. Gordon, C.S. McSweeney and M.W. Phillips
rumen for intestinal digestion. Therefore some rumen anaerobic fungi have been screened
for  their tolerance to tannin and  their ability to utilise  protein from  tannin-protein
complexes. Condensed tannin from the shrub legume Calliandra calothyrsus was solvent
extracted, purified by gel chromatography (Terrill et al 1990) and used as a model tannin
in the screening procedure. Calliandra CT was included in media which lacked protein
so that tolerance to free CT could be tested. Stable complexes of Calliandra CT and bovine
serum albumin were prepared in acidic buffer and dried before addition to media to
examine the ability of fungi to utilise the complexed protein (Table 3). Neocallimastix
patriciarum appeared to grow better in the presence of free CT compared with two other
related species and Piromyces communis and Caecomyces communis. All three Neocalli-
mastix spp. grew when nitrogen was present only as a CT-protein complex while growth
of the other fungal genera was inhibited under these conditions.
Conclusions
There have been a large number of studies on the biology of anaerobic fungi, some of
which have dealt with the contribution, both as measured and as perceived, that is made
by these unusual microorganisms to the availability of nutrients for the host animal.
Some of the essential features of these studies have been given in the present paper.
From these studies, it is apparent that anaerobic fungi are extremely important to
ruminants that are consuming diets of poor-quality, mature herbage through the mechan-
ism of increasing the voluntary intake of feed. This influence of ruminal fungi is most
apparent when additional sulphur is added to tropical pasture grasses that have a low
content of sulphur. The high lignocellulose content of tropical grasses (Wilson 1994)
together with the relatively high tolerance of anaerobic fungi to plant phenolics may be
determining factors in the effectiveness of these microorganisms. Therefore, a consider-
able potential exists for the manipulation of fungal numbers and activity in the rumen
to benefit the utilisation of poor-quality herbage by domesticated ruminants (cattle,
goats, sheep, buffalo and camels) in the semi-arid regions of Africa and other continents
for the improved production of milk, meat, hair, wool and hide.
Table 3. Growth of several strains of ruminal fungi in the presence of condensed tannins (CT) purified from
Calliandra calothyrsus (C.S. McSweeney, unpublished data). Growth was strong (+++), moderate (++),
weak (+), or absent (?) compared with controls that lacked free CT or CT-protein complexes.
Free CT
?
CT-protein complex
?
Fungus 150 ?g/ml 300 ?g/ml +NH
3
?NH
3
Neocallimastix frontalis ++ ? +++ ++
N. hurleyensis ++ ? ++ ++
N. patriciarum ++ ++ ++ +++
Piromyces communis ++ + ? ?
Caecomyces communis ??+
?
Basal medium contained minerals, glucose and ammonia.
?
Basal medium contained minerals, glucose and CT-protein complexes (8% w/v).
Important role for ruminal anaerobic fungi in voluntary intake of poor quality forages 97
References
Akin D.E. 1987. Association of rumen fungi with various forage grasses. Animal Feed Science
and Technology 16:273?285.
Akin D.E. and Borneman W.S. 1990. Role of rumen fungi in fibre degradation. Journal of Dairy
Science 73:3023?3032.
Akin D.E. and Hogan J.P. 1983. Sulfur fertilization and rumen microbial degradation of cell
walls in Digitaria pentzii Stent. Crop Science 23:851?858.
Akin D.E. and Rigsby L.L. 1985. Influence of phenolic acids on rumen fungi. Agronomy Journal
77:180?182.
Akin D.E., Gordon G.L.R. and Hogan J.P. 1983. Rumen bacterial and fungal degradation of Digitaria
pentzii grown with or without sulfur. Applied and Environmental Microbiology 46:738?748.
Akin D.E., Lyon C.E., Windham W.R. and Rigsby L.L. 1989. Physical degradation of lignified
stem tissues by ruminal fungi. Applied and Environmental Microbiology 55:611?616.
Bauchop T. 1979. Rumen anaerobic fungi of sheep and cattle. Applied and Environmental
Microbiology 38:148?158.
Bernard-Vailh? M.A., Besle J.M. and Dor? J. 1995. Transformation of 14C-lignin-labeled cell
walls of wheat by Syntrophococcus sucromutans, Eubacterium oxidoreductens, and Neocal-
limastix frontalis. Applied and Environmental Microbiology 61:379?381.
Bird S.H. 1989. Production from ciliate-free ruminants. In: Nolan J.V., Leng R.A. and Demeyer
D.I. (eds), The Roles of Protozoa and Fungi in Ruminant Digestion. Penambul Books,
Armidale, Australia, pp. 233?246.
Bonnemoy F., Fonty G., Michel V. and Gouet P. 1993. Effect of anaerobic fungi on the ruminal
proteolysis in gnotobiotic lambs. Reproduction Nutrtition Development 33:551?555.
Borneman W.S., Hartley R.D., Morrison W.H., Akin D.E. and Ljungdahl L.G. 1990. Feruloyl
and p-coumaroyl esterase from anaerobic fungi in relation to plant cell wall degradation.
Applied Microbiology and Biotechnology 33:345?351.
Faichney G.J., Brownlee A.G., Gordon G.L.R., Phillips M.W. and Welch R.J. 1991. Contribution
of protozoa and anaerobic fungi to digesta N in sheep given a pelleted hay/grain diet.
Proceedings of the Nutrition Society of Australia 16:209.
Fonty G. and Joblin K.N. 1991. Rumen anaerobic fungi: their role and interaction with other
rumen microorganisms in relation to fibre digestion. In: Tsuda T., Sasaki Y. and Kawashima
R. (eds), Physiological Aspects of Digestion and Metabolism in Ruminants. Proceedings of
the Seventh International Symposium on Ruminant Physiology, Sendai, Japan, 28 August?1
September 1989. Academic Press, San Diego, California, USA, pp. 655?680.
Gordon G.L.R. 1985. The potential for manipulation of rumen fungi. In: Leng R.A., Barker
J.S.F., Adams D.B. and  Hutchinson K.J. (eds), Biotechnology  and Recombinant DNA
Technology in the Animal Production Industries. Reviews in Rural Science, Volume 6.
University of New England, Armidale, Australia, pp. 124?128.
Gordon G.L.R. 1987. Degradation of 14C-labelled lignocelluloses by rumen anaerobic fungi.
In: Rose M. (ed), Herbivore Nutrition Research. Australian Society for Animal Production,
Brisbane, Australia, pp. 115?116.
Gordon G.L.R. 1990. Selection of anaerobic fungi for better fibre degradation in the rumen. In: Akin
D.E., Ljungdahl L.G., Wilson J.R. and Harris P.J. (eds), Microbial and Plant Opportunities to
Improve Lignocellulose Utilization by Ruminants. Elsevier, New York, USA, pp. 301?309.
Gordon G.L.R. and Phillips M.W. 1989a. Degradation and utilization of cellulose and straw by
three different anaerobic fungi from the ovine rumen. Applied and Environmental Microbi-
ology 55:1703?1710.
98 G.L.R. Gordon, C.S. McSweeney and M.W. Phillips
Gordon G.L.R. and Phillips M.W. 1989b. Comparative fermentation properties of anaerobic
fungi from the rumen. In: Nolan J.V., Leng R.A. and Demeyer D.I. (eds), The Roles of
Protozoa and Fungi in Ruminant Digestion. Penambul Books, Armidale, Australia, pp.
127?137.
Gordon G.L.R. and Phillips M.W. 1993. Removal of anaerobic fungi from the rumen of sheep
by chemical treatment and the effect on feed consumption and in vivo fibre digestion. Letters
in Applied Microbiology 17:220?223.
Gordon G.L.R., Akin D.E. and Hogan J.P. 1984. Rumen anaerobic fungi in sheep fed methionine
supplemented Digitaria pentzii hay. In: Baker S.K., Gawthorne J.M., Mackintosh J.B. and
Purser D.B. (eds), Ruminant Physiology?Concepts and Consequences. University of West-
ern Australia, Perth, Australia, p. 174.
Gordon G.L.R., Gulati S.K. and Ashes J.R. 1983. Influence of low-sulphur straw on anaerobic
fungal numbers in a sheep rumen. Proceedings of the Nutrition Society of Australia 8:188.
Gulati S.K., Ashes J.R. and Gordon G.L.R. 1990. Comparative digestibility of sulphur amino
acids in rumen bacteria and fungal proteins by sheep. Proceedings of the Nutrition Society
of Australia 15:128?131.
Gulati S.K., Ashes J.R., Gordon G.L.R. and Phillips M.W. 1985. Possible contribution of rumen
fungi to fibre digestion in sheep. Proceedings of the Nutrition Society of Australia 10:96.
Gulati S.K.,  Ashes J.R.,  Gordon G.L.R.,  Connell  P.J. and  Rogers  P.L. 1989. Nutritional
availability of amino acids from the rumen anaerobic fungus Neocallimastix sp. LM1 in
sheep. Journal of Agricultural Science 113:383?387.
Hsu J.T., Fahey G.C., Merchen N.R. and Mackie R.I. 1991. Effects of defaunation and various
nitrogen supplementation regimens on microbial numbers and activity in the rumen of sheep.
Journal of Animal Science 69:1279?1289.
Kemp P., Jordan D.J. and Orpin C.G. 1985. The free- and protein-amino acids of the rumen
phycomycete fungi Neocallimastix frontalis and Piromonas communis. Journal of Agricul-
tural Science 105:523?526.
Li J.L. and Heath I.B. 1993. Chytridiomycetous gut fungi, oft overlooked contributors to
herbivore digestion. Canadian Journal of Microbiology 39:1003?1013.
Lowry J.B., Sumpter E.A., McSweeney C.S., Schlink A.C. and Bowden B. 1993. Phenolic acids
in the fibre of some tropical grasses, effect on feed quality, and their metabolism by sheep.
Australian Journal of Agricultural Research 44:1123?1133.
McAllister T.A., Bae H.D., Yanke L.J., Cheng K.J. and Muir A. 1994. Effect of condensed
tannins from birdsfoot trefoil on endoglucanase activity and the digestion of cellulose filter
paper by ruminal fungi. Canadian Journal of Microbiology 40:298?305.
McSweeney C.S., Dulieu A., Katayama Y. and Lowry J.B. 1994. Solubilization of lignin by the
ruminal anaerobic fungus Neocallimastix patriciarum. Applied and Environmental Microbi-
ology 60:2985?2989.
Millard P., Gordon A.H., Richardson A.J. and Chesson A. 1987. Reduced ruminal degradation
of ryegrass caused by sulphur limitation. Journal of the Science of Food and Agriculture
40:305?314.
Morrison M., Murray R.M. and Boniface A.N. 1990. Nutrient metabolism and rumen micro-
organisms in sheep fed a poor-quality tropical grass hay supplemented with sulfate. Journal
of Agricultural Science 115:269?275.
Newbold C.J. and Hillman K. 1990. The effects of ciliate protozoa on the turnover of bacterial
and fungal protein in the rumen of sheep. Letters in Applied Microbiology 11:100?102.
Orpin C.G. 1975. Studies on the rumen flagellate Neocallimastix frontalis. Journal of General
Microbiology 91:249?262.
Important role for ruminal anaerobic fungi in voluntary intake of poor quality forages 99
Orpin C.G. 1977. Invasion of plant tissue in the rumen by the flagellate Neocallimastix frontalis.
Journal of General Microbiology 98:423?430.
Orpin C.G. and Greenwood Y. 1986. Nutritional and germination requirements of the rumen
chytridiomycete Neocallimastix patriciarum. Transactions of the British Mycological Soci-
ety 86:103?109.
Phillips M.W. and Gordon G.L.R. 1991. Growth responses to reduced sulphur compounds
of a ruminal fungus, Neocallimastix sp. LM1. In: Zahari M.W., Tajuddin Z.A., Abdullah
N. and Wong H.K. (eds), Proceedings of the Third International Symposium on the
Nutrition of Herbivores. Malaysian Society of Animal Production, Serdang, Selangor,
Malaysia, p. 26.
Rees M.C., Davis P. and Hogan J.P. 1982. The effect of fertilizer and supplementary sulphur on
the digestion of Digitaria pentzii in sheep. Proceedings of the Australian Society of Animal
Production 14:681.
Romulo B., Bird S.H. and Leng R.A. 1986. The effects of defaunation on digestibility and rumen
fungi counts in sheep fed high-fibre diets. Proceedings of the Australian Society of Animal
Production 16:327?330.
Romulo B., Bird S.W. and Leng R.A. 1989. Effects of defaunation and protein supplementation
on intake, digestibility, N-retention and fungal numbers in sheep fed straw-based diets. In:
Nolan J.V., Leng R.A. and Demeyer D.I. (eds), The Roles of Protozoa and Fungi in Ruminant
Digestion. Penambul Books, Armidale, Australia, pp. 285?288.
Scalbert A. 1991. Antimicrobial properties of tannins. Phytochemistry 30:298?305.
Soetanto H., Gordon G.L.R., Hume I.D. and Leng R.A. 1985. The role of protozoa and fungi in
fibre digestion in the rumen of sheep. In: Efficient Animal Production for Asian Welfare.
Proceedings of the Third Congress of the Asian-Australian Association of Animal Production
Societies, Seoul, Korea, Volume 2. Australian Association of Animal Production, Seoul,
Korea. pp. 805?807.
Tanaka H., Matsui H., Ushida K. and Kojima Y. 1991. Effect of p-coumaric acid on growth and
cellulolytic activity of rumen fungi with and without methanogens. In: Zahari M.W., Tajuddin
Z.A. and Wong H.K. (eds), Proceedings of the Third International Symposium on the Nutrition
of Herbivores. Malaysian Society of Animal Production, Serdang, Selangor, Malaysia, p. 35.
Terrill T.H., Windham W.R., Evans J.J. and Hoveland C.S. 1990. Condensed tannin concentrat-
ion in Sericea lespedeza as influenced by preservation method. Crop Science 30:219?224.
Teunissen M.J. and Op Den Camp H.J.M. 1993. Anaerobic fungi and their cellulolytic and
xylanolytic enzymes. Antonie van Leeuwenhoek International Journal of General and Mo-
lecular Microbiology 63:63?76.
Theodorou M.K., Beever D.E., Haines M.J. and Brooks A. 1990. The effect of a fungal probiotic
on intake and performance of early weaned calves. Animal Production 50:577.
Trinci A.P.J., Davies D.R., Gull K., Lawrence M.I., Nielsen B.B., Rickers A. and Theodorou
M.K. 1994. Anaerobic fungi in herbivorous animals. Mycological Research 98:129?152.
Ushida K., Kayouli C., DeSmet S. and Jouany J.P. 1990. Effect of defaunation on protein and
fibre digestion in sheep fed on ammonia-treated straw-based diets with or without maize.
British Journal of Nutrition 64:765?775.
Wallace R.J. and Munro C.A. 1986. Influence of the rumen anaerobic fungus, Neocallimastix
frontalis, on the proteolytic activity of a defined mixture of rumen bacteria growing on a
solid substrate. Letters in Applied Microbiology 3:23?26.
Weston R.H., Lindsay J.R., Purser D.B., Gordon G.L.R. and Davis P. 1988. Feed intake and
digestion responses in sheep to the addition of inorganic sulfur to a herbage diet of low sulfur
content. Australian Journal of Agricultural Research 39:1107?1119.
100 G.L.R. Gordon, C.S. McSweeney and M.W. Phillips
Williams A.G. and Withers S.E. 1991. Effect of ciliate protozoa on the activity of polysaccha-
ride-degrading enzymes and fibre breakdown in the rumen ecosystem. Journal of Applied
Bacteriology 70:144?155.
Wilson J.R. 1994. Cell wall characteristics in relation to forage digestion by ruminants. Journal
of Agricultural Science 122:173?182.
Wubah D.A., Akin D.E. and Borneman W.S. 1993. Biology, fibre-degradation, and enzymology
of anaerobic zoosporic fungi. CRC Critical Reviews in Microbiology 19:99?115.
Important role for ruminal anaerobic fungi in voluntary intake of poor quality forages 101
Rumen fungi in domestic and wild herbivores
J. KopeBb9n?
Institute of Animal Physiology and Genetics, Czech Academy of Sciences
Prague 10, UhBcd?nB8dves, 104 00, Czech Republic
Abstract
Anaerobic fungi play an important role in fibre degradation in herbivorous
animals. Therefore, 35 fungal strains were isolated from the digestion tract
of domestic ruminants (3 species) and 86 strains from wild herbivores kept
in Czech zoos (29 species). In total 44 herbivorous animal species were
tested. Some of the  isolated fungal strains were tested for cellulolytic,
hemicellulolytic and pectinolytic activities. Pectinolytic activity was studied
in detail. In addition to the pectin lyase described, pectate lyase and poly-
galacturonase activities were found in some isolates. No pectin esterase was
detected in any isolate tested. Generally, relatively few of the isolates (ca.
20%) were able to degrade pectin to any extent.
Development of the fungal population may  be limited in the rumen
ecosystem. A negative effect of ruminococci has been reported (Stewart et al
1992). We found another inhibiting?an action of chitinolytic bacteria in the
rumen. These bacteria can reduce fungal growth and fungal cellulose degra-
dation in pure cultures. Their highest effect was observed in presence of
Caecomyces. Positive interactions of fungi with rumen bacteria have been
studied. In the presence of intermediate-utilising (Megasphaera elsdenii and
Eubacterium limosum) bacteria, the polycentric rumen fungus Orpinomyces
joyonii A
4
was able to increase cellulose degradation and biomass produc-
tion. There was a shift in the fermentation pattern too. Systematics of rumen
fungi is based on spore and mycelium morphology. Unfortunately, the shape
of mycelium and sporangium is dependent on growth conditions. An attempt
to distinguish different genera by their DNA fingerprints was made. DNA
fingerprints of three fungi were tested. In our view, analysis of genetic
information is the only reliable approach to rumen fungal systematics.
Introduction
Anaerobic fungi represent a special group of microorganisms inhabiting the rumen
ecosystem and possess a life cycle alternating between a motile reproductive flagel-
lated form (zoospore) and a non-motile vegetative form (thallus). The history of their
discovery is interesting. The motile form was first observed in 1912 in the fresh water
crustacean Cyclops stenuus and Callimastix cyclopis was described as a parasitic
polyflagellate protozoan by Weissenberg (1912). The next year a very similar organism
was observed by  Braune (1913)  in  the rumen  of  sheep and  was  described  as  a
Callimastix frontalis. In the mid-1960s, it was found that Callimastix cyclops had a
cell ultrastructure similar to fungi and the species was transferred to chytrid fungi in
the Blastocladiales (Vavra and Joyon 1966). The rumen ?Callimastix? species remained
in Protozoa in a new genus, Neocallimastix, with Neocallimastix frontalis as the type
species. In 1975, Orpin described the life cycle of N. frontalis consisting of motile and
non-motile stages. The genus was transferred to fungal class Chytridiomycetes (Heath
et al 1983). On the basis of the zoospore ultrastructural characteristics, a new family,
Neocallimasticaceae (Barr 1988), was established, now with six subdivisions of the
family.
monocentric Neocallimastix
Piromyces
Caecomyces
and polycentric Orpinomyces
Ruminomyces
Anaeromyces
The most important feature of these fungi is their plant fibre degrading ability. They
are attracted to fibre which they degrade and expose plant material to bacterial attack.
All isolates produce cellulases and hemicellulases. In very few isolates is either of these
activities suppressed (our observation). The fibrolytic activity produced by pure strains
of rumen fungi is in generally comparable or higher than the fibrolytic activity of pure
cultures of rumen bacteria (Orpin and Joblin 1988; Wood et al 1986). Fungi colonise
even lignified cell walls which are not attacked by rumen bacteria (Akin and Rigsby
1987). This is probably facilitated by their high activity of p-coumaroyl- and feruloyl-
esterases (Borneman et al 1992).
This study focuses on the occurrence of anaerobic fungi in the digestive tract of
different herbivorous animals, their hydrolytic properties and on interactions with other
rumen microbes with the aim of increasing their importance in rumen fermentation.
Materials and methods
Chemicals
Citrus pectin (GENU Pektin, Denmark) with 65% esterification was used for all experiments.
Pectin was washed twice with 80% (v/v) ethanol to remove soluble sugars and freeze dried.
It contained 3.50% (w/w) galactose, 0.41% (w/w) rhamnose and 0.27% (w/w) mannose.
Content of arabinose, xylose and glucose was lower than 0.05% (w/w). Polygalacturonic
(pectic) acid was prepared from the same pectin by the method of Kertesz (1957).
Colloidal chitin used as a substrate for all experiments was prepared from crab shell
chitin (Sigma) according to Shimahara and Takiguchi (1988). Chitin suspension was
autoclaved and stored at 4?C.
Microorganisms (interactions with chitinolytic bacteria)
A strain of rumen fungus Orpinomyces joyonii A
4
was isolated from the rumen fluid of
a camel according to Joblin (1981).
104 J. KopeBddn?
Microorganisms (microbial interactions)
Megasphaera elsdenii L
2
was isolated in our laboratory from rumen fluid of a sheep on
PY medium with lactate. The strain L
2
utilised fructose, glucose, maltose, mannitol and
lactate. Eubacterium limosum ATCC 8486 is a type strain from The American Type
Culture Collection (Rockford, MD, USA). Of the sugars tested, only fructose and
glucose were utilised by this strain.
Microorganisms (pectin degradation)
Four fungal strains showing pectinolytic activity were selected from 24 isolates obtained
in our laboratory according to Joblin (1981). Orpinomyces joyonii A
4
was isolated from
rumen fluid of a camel, Neocallimastix sp. JL3 from faeces of a red deer, and
Neocallimastix sp. OC and H15 from sheep rumen.
Cultivation methods
Rumen bacteria and fungi were grown anaerobically at 38?C in 100 ml flasks with
butyl-rubber stoppers in modified medium M10 (Caldwell and Bryant 1966). Medium
was enriched with 0.1% (v/v) vitamin mixture (pyridoxine 200 mg, riboflavin 200 mg,
thiamine 200 mg, nicotine amide 200 mg, pantothenic acid 200 mg, p-aminobenzoic
acid 1 mg, biotin 0.5 mg, cobalamin 0.5 mg in 100 ml of water), 0.1% (v/v)
microminerals (NiCl
2
, 25.3 mg; H
3
BO
3
, 33.3 mg; Na
2
MoO
4
.2H
2
O, 50 mg;
FeSO
4
.7H
2
O, 1.1 mg; MnSO
4
.4H
2
O50,mg;ZnSO
4
.7H
2
O, 50 mg; CuSO
4
.5H
2
O,25.3
mg; CoCl
2
.6H
2
O, 25.3; EDTA, 10 mg; KAl(SO
4
)
2
.12H
2
O, 10 mg; NaOH, 5.06 g in
100 ml H
2
O, pH 7), 20% (v/v) clarified rumen fluid and 4 g of substrate (chitin, glucose,
microcrystalline  cellulose,  amorphous cellulose or  cellobiose  instead  of a mix  of
glucose, cellobiose and starch). Medium was boiled and after cooling 0.5 g of cysteine
and4gofNa
2
CO
3
was added (Holdeman et al 1977). Reduced medium was transferred
into flasks in an oxygen-free atmosphere (70% N
2
,25%CO
2
,5%H
2
) and autoclaved.
Isolation of anaerobic fungi
Anaerobic fungi were isolated according to Kudo et al (1990) from serial dilutions
(10
?1
?10
?3
) of rumen fluid or faeces in medium M10 with agar, cellobiose and
antibiotics (penicillin, streptomycin and chloramphenicol) in an anaerobic atmosphere.
After three days colonies were reisolated and then transferred into liquid medium.
Culture purity was checked microscopically. Animal faeces were transported under a
CO
2
atmosphere at a temperature of 30 to 40?C.
Isolation of chitinolytic rumen bacteria
Chitinolytic rumen bacteria were isolated on plates with the same medium M10
containing 0.5% colloidal chitin and 1.1% agar. Samples of rumen fluid were diluted in
anaerobic flasks and inoculated into agar. Cultivation was done at 40?C under anaerobic
atmosphere for 48 h. Cleared zones indicated chitinolytic strains.
Rumen fungi in domestic and wild herbivores 105
Total counts
Anaerobic bacterial counts were estimated on PYG medium with 15% rumen fluid under
anaerobic conditions.
Analytical methods
Volatile fatty acids (VFA) were determined by gas chromatography after diethylether
extraction on a column of DB-Wax Megabore (J and W, USA). Succinate was estimated
on the same column but after esterification with methanol and extraction with chloro-
form. Ethanol was measured directly by GLC on a column with 10% SP1200 with 1%
phosphoric acid on Chromosorb WAW. Formate was estimated colorimetrically (Sleat and
Mah 1984). Lactate was oxidised to acetaldehyde and measured in microdiffusion cham-
bers (Marounek and Bartos 1987). Protein concentration was measured according to
Lowry et al (1951) and Herbert et al (1971).
Enzyme activity (chitin)
N-acetyl-?-glucosaminidase was assayed with p-nitrophenyl-?-N-acetyl-glucosaminide
(pNAG, Sigma) according to Bidochka at al (1992). The reaction mixture contained 100
?l of enzyme solution, 100 ?l of 4 mM pNAG and 200 ?l of 100 mM phosphate buffer
pH 6.5. Reaction was stopped with 800 ?l of 2% Na
2
CO
3
. After centrifugation, released
p-nitrophenol was measured at 410 nm in the supernatant.
Chitinase was determined by a colorimetric assay with p-dimethylaminobenzaldehyde
(DMAB) (Boller and Mauch 1988) or with p-hydroxybenzoic acid hydrazide (Lever 1977).
Enzyme activity (pectin)
Activity of pectin esterase (EC 3.1.1.11) was measured by continuous titration with NaOH
(Shejter and Marcus 1988). Pectate (EC 4.2.2.2) and pectin lyase (EC 4.2.2.10) activity were
measured by recording absorbance at 232 nm during incubation with polygalacturonic acid
and pectin (Tagawa and Kaji 1988). Polygalacturonase (EC 3.2.1.15) activity was estimated
according to Tagawa and Kaji (1988) by release of reducing sugars. Sample (0.5 ml) was
incubated with 0.5 ml of 0.5% polygalacturonic acid and 1 ml of 100 mM phosphate buffer
pH 6.0 with 6 mM EDTA for 10?60 min at 40?C. Reaction was stopped in a boiling water
bath for 5 min. After cooling and centrifugation, reducing sugars were estimated in the
supernatant (Lever 1977).
Results and discussion
Isolation of anaerobic fungi from the digestion
tract of herbivores
Anaerobic fungi were isolated from rumen fluid and faeces of different animals over a
period of two years (Tables 1 and 2). The distribution of fungal species changed during
106 J. KopeBddn?
the year and even within one animal species individuals did not bear the same types of
fungal strains. Some unknown types of anaerobic fungi were observed and we are sure
there exist several undescribed species. Such a wide spectrum of fungi was not observed
in other animals kept in a zoo (Teunissen et al 1991).
Metabolic tests were done with some isolated fungi (Table 3). Only alditols, mannose,
sucrose and inulin were of diagnostic importance. Fibrolytic enzymes were measured
in most isolates. The range of activities varies up to three orders (Table 4). Since
extracellular ?-xylosidase and ?-glucosidase activity were not measured  in many
samples we did not include the results in Table 4. However, these enzymes were always
detected in the medium. The most important is the activity of cellulases and hemicellu-
lases. The efficiency of the cellulolytic and hemicellulolytic complex is so high that the
fungi are able to degrade soft wood (Joblin and Naylor 1989).
Pectin degradation by anaerobic fungi
Pectinolytic strains isolated in our laboratory belong to two genera of anaerobic fungi.
Strains OC,H15,JL3shared characteristics similar to the monocentric members of the
genus Neocallimastix (Borneman et al 1989; Mountfore and Asher 1989).
The strain A
4
is a polycentric fungus and had properties typical of Orpinomyces
joyonii (Li et al 1991). Similar to other anaerobic fungi, all pectinolytic isolates utilised
glucose, cellobiose, cellulose, fructose, xylose and lactose. Other sugars tested were
metabolised by some strains only. No growth was observed in the presence of galactose,
galacturonic acid or arabinose. Fermentation products of these pectinolytic isolates
grown on glucose and pectin were similar (Table 5). Amounts of individual fermenta-
Table 1. Occurrence of anaerobic fungi in rumen fluid of domestic and wild animals.
Animal Number of Fungi Fungal
samples isolated strains
Domestic animals
Sheep 12 12? Caecomyces
Neocallimastix
Cattle 15 15? Caecomyces
Neocallimastix
Goat 8 8? Neocallimastix
Wild animals
Red deer (Cervus elaphus) 33 Neocallimastix
Mouflon (Ovis musimon) 1?
Hungarian Steppe Cattle 3 3? Neocallimastix
Swamp Buffalo (Bos bubalis) 33 Orpinomyces
Chamois (Rupicarpa rupicarpa) 1?
Cattle 1 ?
Sheep 2 2? Neocallimastix
Deer (Pseudaxis sika) 1?
Goat (Carpa aegargus hircus)
Roe-deer (Capreolus capreolus) 11? Neocallimastix
Rumen fungi in domestic and wild herbivores 107
tion products except formate were greater with glucose as growth substrate. No meas-
urable amount of methanol was detected in any media tested.
Pectinase activities found in intracellular and extracellular fractions of Neocalli-
mastix sp.H15,JL3,OCandO. joyonii A
4
are given in Table 6. In all strains the highest
Table 2. Occurrence of anaerobic fungi in faeces of domestic and wild animals.
Animal Number of Fungi Fungal
inocula isolated strains
Wild animals Prague Zoo
Adax (Addax nasomaculatus) 33?Orpinomyces
Antelope bongo (Boocercus eyryceros) 55Orpinomyces
Bison (Bibos bison) 88?Caecomyces
Orpinomyces
Porcupine (Hystrix cristata) 3?
Hippopotamus (Hippopotamus amphitius)
Hippopotamus (Choeropsis liberiensis)
Yak (Bibos gruniens) 32?Caecomyces
???
Deer (Rucervus duvanceli) 44Caecomyces
Deer (Rucervus eldi) 55?Caecomyces
Piromyces
Deer (Elaphurus davidianus) 31Caecomyces
Red kangaroo (Macropus rufus) 2?
Kiang (Equus hemionus kiang) 42?Neocallimastix
Goat (Capra falconeri) 33Neocallimastix
Goat (Capra caucasica) 31????
Kulan (Equus hemionus-asianus) 42Caecomyces
Przewalski horse (Equus cabal.przewalskii) 66?Orpinomyces
Neocallimastix
Llama pako (Lama vicugna pacos) 44Orpinomyces
Llama guanaco (Lama guanicoe glama) 99?Caecomyces
???
Rhinoceros (Diceros bicornis) 43Piromyces
Neocallimastix
Somalian ass (Equus somaliensis) 11????
Sheep (Ammotragus lervia) 32???
Chilean pudu (Pudu pudu) 11????
African elephant (Loxodonta africana) 33Caecomyces
Indian elephant (Elephas maximus) 32?Orpinomyces
Bactrian camel (Camelus bactrianus) 44Caecomyces
Indian cattle (Hemitragus sp.) 3 1? Orpinomyces
Urson (Urson dorsatus) 2?
Grevy zebra (Equus grevyi) 66Orpinomyces
European bison (Bibos bonasus) 44?Neocallimastix
Orpinomyces
Rothschild giraffe (Giraffa rotchildi) Orpinomyces
Neocallimastix
Ruminomyces
108 J. KopeBddn?
activities observed were intracellular pectin lyase (EC 4.2.2.10) and polygalacturonase
(EC 3.2.1.15). Enhanced polygalacturonase activities were recorded when EDTA
replaced CaCl
2
in the assay, indicating the absence of catalytic requirements for
divalent ions (Table 7). The pH optima of these enzymes were tested in intracellular
and extracellular fractions (Table 8). All extracellular polygalacturonases had a pH
optimum at pH 6.0, except in strain H15 which had another peak at pH 7.5. On the
contrary, in intracellular fraction this enzyme possess other optima beside the main
one at pH 6. These data are in agreement with observations reported by Collmer at al
(1988).
Our pectinolytic isolates also produced pectate lyase (EC 4.2.2.2). In comparison with
the two enzymes mentioned above, the level of this activity was lower in the intracellular
fraction (Table 2). No detectable pectinesterase (EC 3.1.1.11) was observed (data not
shown). Several species of bacteria and protozoa are responsible for the degradation
and fermentation of pectin in the rumen (Orpin 1984; Szymanski 1981). These microbes
produce mainly exo-polygalacturonase, endo- and exo-pectate lyase and pectin esterase
Table 3. Carbohydrates utilised by isolated rumen fungi.
Isolate
N. frontalis C. communis O. joyonii
Substrate OC H15 JL2 JL3 CB3 G1 B1 ZU4 JB1 SC2 M8 B5 A4 A1
Glucose + +++++++ ++++++
Amygdalin N + + + + + ? ? ? + + + ? +
Galactose ? ? ? ? ? ? ? ? ? ? ? ? ? ?
Rhamnose N ? + + ? + ? ? ? ? ? ? ? +
Dulcitol N ? + + ? ? ? ? ? ? ? ? ? +
Sorbitol N ? + + ? ? + ? ? ? + ? ? +
Cellobiose + + ++++++++++++
Lactose N +++++++ ?+N?++
Mannose N ? ? ? ++++?? ????
Sucrose + ? + + ? ? ? + ? ? + ? + +
Xylose + + ? + + + ++++++++
Arabinose ? ? ? ? ? ? ? ? ? ? ? ? ? ?
Xylan ++ ??+??? ?+++++
Hemicellulose N + + +++++?+++++
Maltose + +++++++ ?++?++
Starch + ? + + + ? ? + ? + + ? + +
Pectin + + ? + ? + ? + ? ? ? ? + +
Fructose + + + + + + + +++++++
Inulin + + + ? ? ? ? + ? ? ? ? + +
Source of fungal strains: Sheep H15; OC; SC2; Bison ZU4; Red deer JB1; JL2; JL3; Mouflon M8; Cattle B1; B5; CB3;
Camel A1; A4 Goat G1.
Rumen fungi in domestic and wild herbivores 109
(Paster and Canale-Parola 1985). Contrary to the former findings of Phillips (1989),
extracellular pectinase activity was observed in mono- and polycentric rumen fungi
(Gordon and Phillips 1991). The predominant pectinolytic activity in the monocentric
fungus Neocallimastix sp. was represented by an endoacting pectin lyase with pH
optimum at 8.0 (Gordon and Phillips 1992). We found that this enzyme was located
mainly in the intracellular fungal fraction. The mixture of pectin degrading enzymes
(pectin and pectate lyase and polygalacturonase) should be able to degrade natural
pectins in  a  cooperative manner in the  absence  of  pectin esterase (Tsuyumu and
Table 4. Range of cellulolytic and hemicellulolytic activities of anaerobic fungi from wild herbivores.
Activity?B6dg hexose or pentose/h per ml
Enzyme Substrate Intracellular Extracellular
CBH MC 42?942 31?280
EG CMC 63?1360 16?293
?-GLU PNP-?-1,4-glu 561?5107 ND
XYL Xylan 10?1465 157?1360
?-XYL PNP-?-1,4-xyl 738?9450 ND
ND: not determined.
Table 5. Fermentation products of rumen fungi grown on pectin and glucose
?
.
Fermentation Substrate Fungal strain
Products
?
(4 g/l) A
4
OC H15 JL
3
Formate glucose 32.9 29.7 24.5 26.3
pectin 26.0 35.0 27.8 27.5
Acetate glucose 6.4 14.0 12.3 17.6
pectin 9.0 8.8 9.6 8.7
Propionate glucose 0 0 0.2 0.6
pectin 0.2 0.4 0.5 0.7
Butyrate glucose 0 0 0.5 0.9
pectin 0.8 0.7 0.4 0.7
iso-Butyrate glucose 0 0 0 0.1
pectin 0.1 0.2 0.1 0.4
iso-Valerate glucose 1.0 1.9 1.1 1.5
pectin 0 0 0 0
Ethanol glucose 6.8 4.7 3.4 3.3
pectin 2.4 1.6 1.4 1.1
Lactate glucose 17.1 20.8 14.2 18.1
pectin 1.7 3.8 3.8 3.1
Succinate glucose 0.2 0.5 0.4 0.5
pectin 0.2 0.2 0.3 0.2
Reducing sugars glucose 239 429 1449 757
pectin 877 897 877 1034
?
Fungi were grown for three days in modified M10 medium with glucose (4 g/l) and pectin (4 g/l) as the only energy
source. After cultivation, fermentation products were measured in centrifuged (10000 xg, 10 min, 4?C) medium.
Reducing sugars remaining in the medium after three days of incubation (mg hexose l
?1
).
?
Concentration of short chain fatty acids (SCFA) in mmol l
?1
.
110 J. KopeBddn?
Miyamoto 1986). Thus the results of this study confirmed the possibility of complex
utilisation of pectin and pectin-like substances by anaerobic fungi. Nevertheless, the
determination of the activity of specific pectic enzymes in crude pectolytic preparations
is considerably complicated by the partial suppression of the activity value due to
simultaneous effect of these enzymes on the same substrate. There is still a need for a
more detailed description of isolated pectic enzymes produced by anaerobic rumen
fungi.
Table 6. Activity of enzymes involved in pectin degradation produced by rumen fungi grown on pectin.
Enzyme Fungal strain
activity A
4
OC H15 JL
3
Polygalacturonase
?
Extracellular 46 ? 9 29 ? 6 10 ? 3 27 ? 10
Intracellular 251 ? 62 43 ? 3 435 ? 50 167 ? 11
Pectate lyase
?
Extracellular 110 ? 39 83 ? 7 29 ? 17 1 ? 2
Intracellular 69 ? 2 31 ? 1 91 ? 4 49 ? 5
Pectin lyase
?
Extracellular 142 ? 4 185 ? 89 166 ? 41 93 ? 24
Intracellular 258 ? 17 274 ? 5 294 ? 17 908 ? 22
?
Enzyme activity is expressed in ?g galacturonic acid.h
?1
.mg protein
?1
.
Table 7. The effect of calcium ions on fungal extracellular polygalacturonases.
Fungal Polygalacturonase activity
?
?g galacturonic acid.ml
?1
.h
?1
strains 10 mM CaCl
2
?
10 mM EDTA
?
A
4
56.3 ? 10.1 61.1 ? 2.1
OC 38.6 ? 0.3 61.1 ? 3.7
H15 37.0 ? 0.1 136.7 ? 10.8
JL3 ND 125.4 ? 5.8
?
Activity of extracellular polygalacturonase was estimated in the presence of 100 mM Tris HCl buffer pH 7 with
polygalacturonic acid by measuring reducing sugars.
?
Final concentration of reagents.
ND = not determined.
Table 8. Optimum pH of rumen fungal polygalacturonases
?
.
Fungal strains
Fractions A
4
OC H15 JL
3
Extracellular 6.0 6.0 6.0, 7.5 6.0
Intracellular 6.0, 8.1 5.05, 6.0 6.0, 8.1 6.5, 8.55
?
Activity was estimated with 100 mM phosphate-citrate buffer in range of pH 3?7.9 and with 100 mM borate buffer from
pH 8.1?9.0.
Rumen fungi in domestic and wild herbivores 111
Interaction of rumen fungi with anaerobic bacteria
Isolation of chitinolytic bacteria
Chitinolytic bacteria were screenedin rumenfluid of sheepand faeces ofherbivorousanimals
from the Prague Zoo. From 14 animal species, chitinolytic bacteria were present in sheep,
American bison, deer milu, przewalski horse, llama paco, Somalian ass and European bison.
Different chitinolytic strains were isolated from the rumen fluid of five cows (Table 9).
Their counts were in the range 5 ? 10
4
to 2 ? 10
8
per ml. Many chitinolytic strains were
examined. Most of them were spore-forming rods. One of the isolated strains, ChK5, was
fully characterised (Table 10) and it bore all features of Clostridium tertium (Cato et al 1984).
Table 9. Counts of chitinolytic bacteria in the bovine rumen.
?
Chitinolytic strains
?
Total counts
Animal 1 2.32 ? 10
8
8.01 ? 10
9
Animal 2 9.10 ? 10
7
9.22 ? 10
9
Animal 3 5.23 ? 10
4
1.98 ? 10
10
Animal 4 0 7.21 ? 10
9
Animal 5 0 5.43 ? 10
9
?
Cows were fed with a silage-concentrate diet.
?
Chitinolytic strains were detected on agar medium 10 with colloidal chitin (4 g/l).
Table 10. Characteristics of chitinolytic strain Cl. tertum ChK5.
General features
G ? straight to curved rods
motile,
spores oval and terminal, often released
gelatin not hydrolysed, catalase and lipase negative
Metabolic profile of the strain ChK5
?
Adonitol +++ Lactose +
Amygdaline ++ Lactate ?
Arabinose ? Maltose ++
Cellobiose ++ Mannitol +++
Cellulose wheat ? Mannose ++
Dulcitol ? Mellibiose +++
Esculine ++ Raffinose +
Fructose +++ Rhamnose +
Galacturonic acid + Ribose ?
Galactose +++ Salicin +++
Glucose +++ Sorbitol ?
Glycerol + Starch ?
Inositol + Trehalose +++
Inulin ++ Xylose ++
?
Cultures were grown anaerobically on PY medium with different carbohydrates at 24?C for 36 h.
112 J. KopeBddn?
Properties of Cl. tertium ChK5
The growth rate of ChK5 differed significantly on glucose and chitin. Exponential
growth on glucose was complete in 4?6 h in comparison with more than 30 h in case of
colloidal chitin. Chitin significantly stimulated sporulation. Fermentation products also
changed. The main product in both cases was acetate (Table 11). The other main
fermentation product on glucose was propionate. When chitin was used as a substrate,
there was almost no propionate production and increased levels of acetate, butyrate and
lactate were observed.
Production of chitinolytic enzymes was measured in the culture medium of ChK5
grown on chitin (Table 12). Endochitinase activity was usually much higher than
N-acetylglucosamidase, which is cell-associated.
The effect of Cl. tertium on rumen fungi
Cell walls of anaerobic fungi represent a substrate for chitinase that is normally present
in the rumen. A lot of chitinase activity is produced by fungi themselves. It is necessary
to maintain hyphal growth and branching in fungi. Therefore, the effect of the
chitinolytic bacterium on rumen fungi was studied.
The polycentric fungus Orpinomyces joyonii A
4
was cultured on microcrystalline
cellulose. SCFA production was measured for 10 days (Figure 1). Exponential growth
of the fungus was observed up to the fifth day of incubation. Strain Cl. tertium ChK5
was inoculated at different stages of the fungal culture and cultivated for five more days
(Figure 1). In co-cultures SCFA production was decreased. The inhibitory effect was
related to the time of Cl. tertium inoculation. The sooner the chitinolytic strain was
added the bigger the decrease of SCFA that was found.
In the same experiments chitinolytic activity in the medium and cell fraction was
measured. Addition of Cl. tertium to the fungal culture decreased total chitinolytic
activity (Figure 2). Cellulolytic activity was decreased in a similar manner (data not
Table 11. Fermentation end-products of strain Cl. tertium ChK5 grown on glucose and chitin.
?
mmol/l glucose chitin
Acetate 16.8 23.7
Propionate 19.5 0.8
Butyrate 1.4 7.4
Lactate 0.1 8.8
Succinate 0.6 0.1
?
Cultures were grown anaerobically on medium 10 with 10% rumen fluid, 4 g glucose/l or 4 g colloidal chitin.
Table 12. Activity of chitinolytic enzymes in culture medium of Cl. tertium ChK5.
Activity (ug N-acetylglucosamine/h/ml)
Endochitinases 131.7 + 29.7
N-Acetylglucosamidases 42.0 + 2.9
Rumen fungi in domestic and wild herbivores 113
shown). It would be interesting to compare the effect of rumen ruminococci and
chitinolytic bacteria on anaerobic fungi.
Fungal interactions with intermediate utilising bacteria
The polycentric fungus A
4
was assigned to the genus Orpinomyces joyonii on the basis
of its morphological and biochemical characteristics (Breton et al 1989).
In monoculture of strain A
4
, cellulose was degraded extensively and metabolised to
yield mainly formate, acetate, ethanol and lactate (Table 13). High levels of glucose and
cellodextrins released by the fungus were observed from the second day of cultivation.
The association of the fungus A
4
with E. limosum or M. elsdenii clearly increased
cellulose degradation. It was also significant that in these associations the free sugar
concentration (glucose and cellodextrins) was much lower than in the fungal monocul-
ture (Table 13). Besides the fact that bacteria profited from the association with the
fungus, the low sugar concentrations positively influenced fungal enzyme activities
(data not shown).
The end products of cellulose fermentation for the two co-cultures (Table 13) were
qualitatively similar but there were variations in the amounts of the individual products.
Association of the fungus with E. limosum as a H
2
-consuming bacterium resulted in a shift
Figure 1. The effect of Cl. tertium ChK5 on SCFA production in co-culture with O. Joyonii A
4
grown on
microcrystalline cellulose. Fungus was grown alone (solid line) and Clostridium was added at different
intervals (broken line).
0246810
Incubation (days)
0
12
24
36
48
60
mmol SCF
A
/I
A4
A4 + ChK5
ChK5 inoculation
114 J. KopeBddn?
of fermentation pattern toward more acetate and butyrate. The shift was similar to changes
observed with monocentric fungi (Bernalier et al 1993). Since the utilisation of hydrogen by
E. limosum is repressed in the presence of glucose (Genthner and Bryant 1987), cellulolysis
in our case would be enhanced mainly by consumption of free glucose by the bacterium.
Utilisation of formate by E. limosum might have some effect as well (Loubi?re et al 1992).
Although the strain did not grow on formate as a sole source of energy, it was able to utilise
it in presence of other ATP-generating mechanisms (Loubi?re et al 1987). Formate is used
for acetate and butyrate production (Genthner et al 1981). Utilisation of lactate by E. limosum
for the production of butyrate is in agreement with results of Loubi?re et al (1992). The
decrease in ethanol concentration in co-culture could be caused by bacterial degradation
(Genthner et al 1981) or by lower fungal production.
The association of the fungus strain A
4
with M. elsdenii led to a greater extent of
cellulose degradation in comparison with O. joyonii plus E. limosum co-culture. M.
elsdenii L
2
is a ruminal species with a nutritional strategy based on the utilisation of
lactate and glucose. Increases in butyrate and caproate production associated with the
disappearance of acetate allowed us to assume that the main substrate utilised by strain
L
2
was glucose. This assumption is supported by the findings of Forsberg (1978) that
acetate was essential for growth of M. elsdenii in the presence of glucose but not in the
presence of lactate. The conversion of acetate to butyrate and caproate is an energy
Figure 2. The effect of Cl. tertium ChK5 on total endochitinase activity in co-culture with O. joyonii A
4
grown on microcrystalline cellulose. Fungus was grown alone (solid line) and Clostridium was added at
different intervals (broken line).
0246810
Incubation (days)
0
300
600
900
1200
1500
Activ
ity
of endochitinase
A4
A4 + ChK5
ChK5 inoculation
Rumen fungi in domestic and wild herbivores 115
consuming reaction in which the acetate serves as an electron sink (Hino et al 1991).
Acetate alone never supported growth of our strain without glucose.
M. elsdenii utilises lactate via the acrylate pathway (Ladd and Walker 1965). This
metabolic pathway is associated with ATP production because propionate is the main
end-product of lactate fermentation. In contrast, in our co-culture with the fungus,
valerate but not propionate production was observed.
Mixed cultures had a stimulating effect on specific biomass production. Co-culture
with M. elsdenii was more efficient (Y
cellulose
= 64.9 g dry matter (mol glucose)
?1
than
with E. limosum (Y
cellulose
= 55.3 g dry matter (mol glucose)
?1
in specific production
of total dry matter. The yield of the fungus alone was Y
cellulose
= 52.3 g dry matter (mol
Table 13. Cellulose degradation and fermentation product patterns of O. joyonii alone and in co-cultures
with E. limosum and M. elsdenii grown on microcrystalline cellulose.
O. joyonii + O. joyonii +
O. joyonii E. limosum M. elsdenii
Cellulose degraded in glucose units
% 77.07 ? 1.05 85.03 ? 0.33 87.19 ? 0.61
mmol l
?1
44.60 ? 0.60
a
46.31 ? 0.18
b
47.47 ? 0.33
b
Unfermented oligosaccharides (mmol glucose units.l
?1
)
Glucose 6.53 ? 0.56
a
0.93 ? 0.86
b
1.62 ? 0.07
b
Cellodextrins 9.77 ? 1.14
a
3.16 ? 0.56
b
2.60 ? 0.67
b
Cellulose fermented (mmol glucose units.l
?1
)
28.30 ? 0.79
a
42.20 ? 0.65
b
43.28 ? 0.45
b
Metabolites
?
(mol/100 mol glucose units fermented)
Formate 98.59 ? 8.13
a
4.00 ? 3.57
b
78.09 ? 4.02
c
Acetate 64.66 ? 4.59
a
91.47 ? 6.22
b
?3.93 ? 3.97
c
Propionate ?0.07 ? 0.74
a
0.81 ? 0.32
b
?2.77 ? 0.09
c
iso-Butyrate 0.11 ? 0.18
a
2.61 ? 0.28
b
1.96 ? 0.10
b
Butyrate 0.78 ? 0.71
a
14.50 ? 2.93
b
39.51 ? 6.72
c
iso-Valerate 0.28 ? 0.25
a
3.46 ? 0.16
b
3.40 ? 0.19
b
Valerate 0.18 ? 0.28
a
0.45 ? 0.05
a
4.07 ? 0.26
b
Caproate 0.32 ? 0.04
a
ND 7.97 ? 1.21
b
Lactate 3.50 ? 0.39
a
0.18 ? 0.05
b
0.16 ? 0.05
b
Succinate 2.23 ? 0.32 1.21 ? 0.14 1.43 ? 0.12
Ethanol 43.82 ? 4.13
a
22.82 ? 2.24
b
21.97 ? 3.30
b
CO
2
142.37 ?12.08 102.09 ? 2.93 55.36 ? 1.33
Biomass(g.l
?1
) 1.48 ? 0.35 2.36 ? 0.42 2.81 ? 0.46
(100%) (159.4%) (189.9%)
Y
cellulose
52.29 55.92 64.93
(g dry matter/mol glucose unit)
Carbon recovery (%) 107.9% 109.5% 99.3%
?
Bacteria were inoculated into medium 10 with 1% of microcrystalline cellulose 24 h after the rumen fungus. The
co-cultures were incubated for a further 7 days at 40?C. Concentration of VFA (mmol.l
?1
) in original medium was the
following: acetate 8.38, propionate 1.51, isobutyrate 0.56, butyrate 0.96, iso-Valerate 0.40, valerate 0.14, succinate 0.05,
formate 0.05, lactate 0.21, ethanol 0.49. These values were subtracted from co-culture and monoculture final products.
ND = Not determined.
a, b, c
Mean + SD (P < 0.05).
116 J. KopeBddn?
glucose)
?1
. Carbon recovery higher than 100% was probably due to utilisation of
compounds from the rumen fluid.
Microbial interactions in the rumen are highly complex. Any type of improvement in
cellulose degradation will lead to a higher efficiency of anaerobic fermentation in the
rumen. Detailed studies on microbial interactions with more microbial species and more
substrates will be necessary for a better understanding of this type of fermentation.
References
Akin D.E. and Rigsby L.L. 1987. Mixed fungal populations and lignocellulosic tissue degrada-
tion in the bovine rumen. Applied and Environmental Microbiology 53:1987?1995.
Barr D.J.S. 1988. How modern systematics relate to the rumen fungi. BioSystems 21:351?356.
Bernalier A., Fonty G., Bonnemoy F. and Gouet P. 1993. Effect of Eubacterium limosum, a
ruminal hydrogenotrophic bacterium on the degradation and fermentation of cellulose by 3
species of rumen anaerobic fungi. Reproduction Nutrition Development 33:577?584.
Bidochka M.J., Tong K.I. and Khachatourians G.G. 1992. Partial purification and charac-
terization of two extracellular N-acetyl-D-glucosaminidases produced by the entomopatho-
genic fungus Beauveria bassiana. Canadian Journal of Microbiology 39:40?45.
Boller T. and Mauch F. 1988. Colorimetric assay of chitinase. In: Wood W.A. and Kellogg S.T.
(eds), Methods in Enzymology. Vol. 161. Academic Press, San Diego, California, USA. pp.
430?435.
Borneman W.S., Akin D.E. and Ljungdahl L.G. 1989. Fermentation products and plant cell wall
degrading enzymes produced by monocentric and polycentric anaerobic rumen fungi. Ap-
plied and Environmental Microbiology 55:1066?1073.
Borneman W.S., Ljungdahl L.G., Hartley R.D. and Akin D.E. 1992. Purification and partial
characterization of two feruloyl esterases from the anaerobic fungus Neocallimastix strain
MC-2. Applied and Environmental Microbiology 58:3762?3766.
Braune R. 1913. Untersuchungen uber die im wiederkauermagen vorkommenden Protozoen.
Archiv Fur Protistenkunde 32:111?115.
Breton A., Bernalier A., Bonnemoy F., Fonty G., Gaillard B. and Gouet P. 1989. Morphological
and metabolic characterization of a new species of strictly anaerobic rumen fungus: Neocal-
limastix joyonii. FEMS Microbiology Letters 58:309?314.
Caldwell D.R. and Bryant M.P. 1966. Medium without rumen fluid for non-selective enumera-
tion and isolation of rumen bacteria. Applied Microbiology 14:794?801.
Cato E.P., George W.L. and Finegold S.M. 1984. Genus Clostridium. In: Holt J.G. (ed), Bergey?s
Manual of Determinative Bacteriology. Williams  and  Wilkinson, Baltimore,  USA.  pp.
1194?1195.
Collmer A., Ried J.L. and Mount M.S. 1988. Assay methods of pectic enzymes. In: Wood W.A.
and Kellogg S.T. (eds), Methods in Enzymology. Vol. 161. Academic Press, San Diego,
California, USA. pp. 329?334.
Forsberg C.W. 1978. Nutritional characteristics of M. elsdenii. Canadian Journal of Microbi-
ology 24:981?985.
Genthner B.R. and Bryant M.P. 1987. Additional characteristics of one-carbon-compound utiliza-
tion by E. limosum and A. woodii. Applied and Environmental Microbiology 53:471?476.
Genthner B.R., Davis L.C. and Bryant M.P. 1981. Features of rumen and sewage sludge strain
of Eubacterium limosum, a methanol and H
2
-CO
2
-utilizing species. Applied and Environ-
mental Microbiology 42:12?19.
Rumen fungi in domestic and wild herbivores 117
Gordon G.L.R. and Phillips M.W. 1991. Pectin degradation by some ruminal anaerobic fungi.
In: Zahari M.W., Tajuddin Z.A., Abdullah N. and Wong H.K. (eds), Proceedings of the III
International Symposium on the Nutrition of Herbivores. Malaysian Society of Animal
Production, Serdang, Selangor, Malaysia. pp. 27.
Gordon G.L.R. and Phillips M.W. 1992. Extracellular pectin lyase produced by Neocallimastix
sp. LM1, a rumen anaerobic fungus. Letters in Applied Microbiology 15:113?115.
Heath I.B., Bauchop T. and Skipp R.A. 1983. Assignment of the rumen anaerobe Neocallimastix
frontalis to  the Spizellomycetales (Chitridiomycetes) on  the  basis  of its  polyflagellate
zoospore. Canadian Journal of Botany 61:295?307.
Herbert D., Phipps P.J. and Strange R.E. 1971. Chemical analysis of microbial cells. In: Norris
J.R. and Ribbons D.W. (eds), Methods in Microbiology 5B. Academic Press, London, UK.
pp. 210?336.
Hino T., Miyazaki K. and Kuroda S. 1991. Role of extracellular acetate in the fermentation of
glucose by a ruminal bacterium, Megasphaera elsdenii. Journal of General and Applied
Microbiology 37:121?129.
Holdeman L.V., Cato E.P. and Moore W.E.C. 1977. Anaerobe Laboratory Manual. Fourth edition.
Virginia Polytechnic Institute and State University, Blackburg, Virginia, USA. 150 pp.
Joblin K.N. 1981. Isolation, enumeration and maintenance of rumen anaerobic fungi in roll
tubes. Applied and Environmental Microbiology 42:1119?1122.
Joblin K.N. and Naylor G.E. 1989. Fermentation of woods by rumen anaerobic fungi. FEMS
Microbiology Letters 65:119?122.
Kertesz Z.I. 1957. Preparation and determination of pectic substances. In: Colowick S.P. and
Kaplan N.O. (eds), Methods in Enzymology. Vol. 3. Academic Press, New York, USA. pp.
27?30.
Kudo H., Jacober K.D., Phillipe R.C., Cheng K.J., Barr D.J.S. and Costerton J.W. 1990. Isolation
and characterization of cellulolytic anaerobic fungi and associated mycoplasmas from the
rumen of steer fed a roughage diet. Canadian Journal of Microbiology 36:513?517.
Ladd N.J. and Walker D.J. 1965. Fermentation of lactic acid by the rumen microorganism
Peptostreptococcus elsdenii. Annals of New York Academy of Sciences 119:1038?1045.
Lever M. 1977. Carbohydrate determination with 4-hydroxybenzoic acid hydrazide
(PAHBAH): effect of bismuth on the reaction. Analytical Biochemistry 81:21?27.
Li J., Heath I.B. and Cheng K.J. 1991. The development and zoospore ultrastructure of a
polycentric chyrtidiomycete gut fungus Orpinomyces joyonii comb. nov. Canadian Journal
of Botany 69:580?589.
Loubi?re P., Pacaoud S., Goma G. and Lindley N.D. 1987. The effect of formate on the
acidogenic fermentation of methanol  by E. limosum. Journal of  General  and Applied
Microbiology 33:463?470.
Loubi?re P., Gros E., Paquet V. and Lindley N.D. 1992. Kinetics and physiological implications
of the growth behaviour of Eubacterium limosum on glucose/methanol mixtures. Journal of
General Microbiology 138:979?985.
Lowry O.H., Rosebrough N.J., Farr A.L. and Randall R.J. 1951. Protein measurement with Folin
phenol reagent. Journal of Biological Chemistry 193:265?275.
Marounek M. and Bartos S. 1987. Metabolism and some characteristics of ruminal strain
Megasphaera elsdenii. Applied and Environmental Microbiology 63:233?238.
Mountfore D.O. and Asher R.A. 1989. Production of xylanase by the ruminal anaerobic fungus
Neocallimastix frontalis. Applied and Environmental Microbiology 55:1016?1022.
Orpin C.G. 1975. Studies on the rumen flagellate Neocallimastix frontalis. Journal of General
Microbiology 91:249?262.
118 J. KopeBddn?
Orpin C.G. 1984. The role of ciliate protozoa and fungi in the rumen digestion of plant cell
walls. Animal Feed Science and Technology 10:121?143.
Orpin C.G. and Joblin K.N. 1988. The rumen anaerobic fungi. In: Hobson P.N. (ed), The Rumen
Microbial Ecosystem. Elsevier Applied Science, London, UK. pp. 129?150.
Paster B.J. and Canale-Parola E. 1985. Treponema saccharophilum sp. nov., a large pectinolytic
spirochete from the bovine rumen. Applied and Environmental Microbiology 50:212?219.
Phillips M.W. 1989. In: Nolan J.V. (ed), The Roles of Protozoa and Fungi in Ruminant Digestion.
Penambul Books, Armidale, Australia. pp. 247?249.
Shejter A. and Marcus L. 1988. Isoenzymes of pectinesterase and polygalacturonase from
Botrytis cinerea Pers. In: Wood W.A. and Kellogg S.T. (eds), Methods in Enzymology. Vo l .
161. Academic Press, San Diego, California, USA. pp. 366?372.
Shimahara K. and Takiguchi Y. 1988. Preparation of crustacean chitin. In: Wood W.A. and
Kellogg S.T. (eds), Methods in Enzymology. Vol. 161. Academic Press, San Diego, California,
USA. pp. 417?423.
Sleat R. and Mah R.A. 1984. Quantitative method for colorimetric determination of formate in
fermentation media. Applied and Environmental Microbiology 47:884?885.
Stewart C.S., Duncan S.H., Richardson A.J., Backwell C. and Begbie R. 1992. The inhibition
of fungal cellulolysis by cell-free preparation from ruminococci. FEMS Microbiology Letters
9:83?88.
Szymanski P.T. 1981. Anote on the fermentation of pectin by pure strains and combined cultures
of rumen bacteria. Acta Microbiologica Polonica 30:159?163.
Tagawa K. and Kaji A. 1988. Polygalacturonase from Corticium rofsii. In: Wood W.A. and
Kellogg S.T. (eds), Methods in Enzymology. Vol. 161. Academic Press, San Diego, California,
USA. pp. 361?365.
Teunissen M.J., Op den Camp H.J.M., Orpin C.G., Huis IN`T Veld J.H.J. and Vogels G.D. 1991.
Comparison of growth characteristics of anaerobic fungi isolated from ruminant and non-ru-
minant herbivores during cultivation in a defined medium. Journal of General Microbiology
137:1401?1408.
Tsuyumu S. and Miyamoto Y. 1986. Effective production of bacterial pectinases. Symbiosis
2:103?110.
Vavra J. and Joyon L. 1966. Etude sur la morphologie le cycle evolutif et la position systema-
tique de Callimastix cyclopis Weissenberg 1912. Protistologica 2:5?16.
Weissenberg R. 1912. Callimastix cyclopis gen. n., sp.n. ein geisseltragendes Protozoon aus
dem serum von Cyclops, S.-B. Ges. Naturforsch. Fruende, Berlin 5:299?311.
Wood T.M., Wilson C.A., McCRae S.I. and Joblin K.N. 1986. A highly active extracellular
cellulase from  the rumen fungus Neocallimastix frontalis. FEMS  Microbiology Letters
34:37?40.
Rumen fungi in domestic and wild herbivores 119
Factors limiting proliferation of desirable
groups of bacteria in the rumen of animals
fed poor quality feeds of high fibre content
N.O. van Gylswyk
Department of Animal Nutrition and Management
Swedish University of Agricultural Sciences
Kungs?ngen Research Centre, S-753 23 Uppsala, Sweden
Abstract
Animal production in Africa is very reliant on natural pastures as well as on
other plant material often of poor quality. The role of fibre-digesting flora in
the rumen of animals feeding on such material is particularly important in
releasing fermentable carbohydrate both for themselves and other microbes.
The nutritional requirements of the important fibre-digesting bacteria is
examined and related to deficiencies in poor quality feeds. Low protein
content implies deficiencies of ruminal ammonia and branched chain volatile
fatty acids (BC-VFA) essential for the cellulolytic bacteria. The fate and
possible role of the BC-VFA is considered. An additional factor contributed
by protein and favouring fibre digestion is reduced sulphur. Methionine and
peptides are of importance in stimulating certain hemicellulolytic bacteria.
Foremost among mineral deficiencies on such poor feeds are those related
to phosphorus and sodium. The major cellulose-digesting bacteria, as well
as others, have absolute requirements for a range of vitamins. Low concen-
trations of these in the rumen could limit fibre digestion but little is known
about this. The results of two experiments with sheep on the effects of
supplementing low-protein hays with urea and BC-VFA are summarised. In
one of these the numbers and types of cellulolytic bacteria were determined
and their relation to fibre digestion in the rumen and animal performance is
discussed. Mechanisms for the ?protein-sparing? action of readily ferment-
able carbohydrate are also considered.
Introduction
In many parts of Africa natural grazing is often the most important or even the sole
source of nutrition for ruminants. Where available, cereal straws or stubble are often
also used as well as any type of plant material (shrubs and trees) including crop residues.
For large parts of the year the grazing can be dry and of very poor quality, characterised
by high fibre content of low digestibility and very low protein content. Such material
cannot maintain the condition of ruminants. In periods of drought even this material is
in short supply. The purpose of this paper is to look at some ways by which such low
quality material has been upgraded and attempt to relate the results to the rumen ecology,
particularly the rumen bacteria.
Important fibre-digesting bacteria of the rumen
It is natural to assume that fibre-digesting bacteria in the rumen are of primary
importance when ruminants feed on poor quality material of which cellulose and
hemicellulose can comprise about 80% in roughly equal proportions. The cellulolytic
bacteria provide fermentable energy sources both for themselves and other bacteria.
Table 1 shows the types of greatest significance.
All the species listed ferment cellobiose. These and many non-cellulolytic rumen
bacteria can ferment soluble and even poorly soluble hydrolysis products of cellulose
(Russel 1985), thus demonstrating the reliance of the non-fibre digesting bacteria on
fibrolytic ones. Fibrobacter succinogenes and the ruminococci are considered to be the
most active in cellulose digestion. The hemicellulolytic activity of F. succinogenes and
Ruminococcus flavefaciens appears to be directed mainly towards exposing cellulose
through removal of hemicellulose because the mono- and oligosaccharides of this fibre
component cannot be utilised by these bacteria and are available for others. In particular,
Table 1. Some characteristics of important rumen bacterial species active in digesting cellulose (C) and
hemicellulose (H).
Type Main
of fibre fermentation
Species digested products Comments
Fibrobacter C, H Acetate, Can not utilise pentose released
succinogenes succinate. from H. Ferments glucose but few
other carbohydrates.
Ruminococcus C, H Acetate, Most strains do not utilise pentose.
flavefaciens succinate, Some strains may ferment glucose
hydrogen. and a few other sugars.
Ruminococcus C, H Formate Strains may or may not ferment
albus acetate, glucose and a few other sugars.
ethanol, Many strains (perhaps 30%) are not
hydrogen. cellulolytic. These ferment many sugars.
Butyrivibrio C, H or Formate, Most strains non-cellulolytic.
fibrisolvens H alone (acetate), Approx. 5% are weakly cellulolytic.
butyrate, Many carbohydrates fermented.
lactate. Proteolytic.
Eubacterium C Formate, Many sugars are fermented.
cellulosolvens butyrate,
lactate,
n-valerate.
Prevotella H Acetate, Many sugars are fermented.
ruminicola propionate, Proteolytic.
succinate.
122 N.O. van Gylswyk
F. succinogenes appears to occur in high numbers (10% or more of ?total culturable?
bacteria) in the rumen of animals feeding on material containing highly crystalline forms
of cellulose such as wheat straw (Byrant and Burkey 1953) or maize straw (van Gylswyk
and van der Toorn 1986). Butyrivibrio fibrisolvens can comprise more than half of the
more active hemicellulolytic bacteria in the case of the latter diet (van der Linden et al
1984).
F. succinogenes and the ruminococci have an absolute requirement for ammonia.
Amino acids and other N sources are of little importance in their nutrition (Byrant 1973).
These species also require certain branched chain (BC) and straight chain volatile fatty
acids (VFA) for growth. They are shown in Table 2.
Protein is rapidly broken down in the rumen to amino acids and peptides and it was
shown (El-Shazly 1952; Menahan and Schultz 1963) that the BC-VFA originated from
the BC amino acids valine, leucine and isoleucine after deamination or, more usually,
deamination followed by decarboxylation. n-Valeric acid may be derived by deamina-
tion of ?-aminovaleric acid which is probably formed by means of a Stickland-type
reaction involving alanine and proline. n-Valeric acid is also produced by Eubacterium
cellulosolvens (Table 1) as a product of carbohydrate fermentation.
The cellulolytic bacteria incorporate the straight and BC-VFA mainly into n-C
13
to
n-C
17
straight and BC fatty acids and aldehydes as part of the lipid component of
bacterial cells (Allison et al 1962; Wegner and Foster 1963). Only in the case of R.
flavefaciens are significant amounts of BC-VFA used in amino acid synthesis (Allison
and Byrant 1963). However, the total ruminal flora tends to incorporate BC-VFA
predominantly into the corresponding BC amino acids (Allison and Byrant 1963). It has
been suggested (D. Palmquist, personal communication) that the long BC acids and
aldehydes lend fluidity to the lipids of the cellulolytic rumen bacteria as is the case for
unsaturated long chain straight acids in aerobic organisms. In the rumen and other
anaerobic environments there is a strong tendency towards saturation of double C bonds
due to reducing conditions (Harfoot and Hazlewood 1988). The special need for
fluidity in the lipids of the cellulolytic bacteria could indicate that it is concerned with
cellulolysis, perhaps in the transport of the large molecules of cellulolytic enzymes
Table 2. Branched and straight chain volatile fatty acids required by the actively cellulolytic rumen bacteria
(Bryant and Doetsch 1954; Wegner and Foster 1963).
5 to 8 carb.
2-Methyl- 2-Ketoiso- straight chain
Species Isobutyrate butyrate Isovalerate valerate (e.g. n-valerate)
Fibrobacter +or+ ? ? +
succinogenes
Ruminococcus + ? or + ? ?
flavefaciens
Ruminococcus + or + or + or + ?
albus
Factors limiting proliferation of desirable bacteria in the rumen 123
across membranes. Cellulases tend to be excreted in membranous vesicles (Chesson
and Forsberg 1988). In this context lipid fluidity may also be of significance.
The cellulolytic rumen bacteria have requirements for vitamins but all of the vitamins
listed in Table 3 for a particular species are not necessarily required by all strains of that
species. The growth and cellulolytic activity of R. albus is stimulated by 3-phenyl-
propionic acid (Stack and Hungate 1984). The subspecies ruminicola of Prevotella
ruminicola (hemicellulolytic) requires heme for growth (Caldwell et al 1965) because
it contains cytochromes. Strains of this species require methionine when ammonia is
the major N source (Pittman and Bryant 1964). P. ruminicola does not appear to require
vitamins.
Rumen bacteria require a range of macro- and micro-elements. In tropical regions,
the elements most likely to be deficient for grazing cattle are P followed by Cu and Co.
Deficiencies in Na and I are widespread (McDowell et al 1984). Deficiencies in P and
Na are likely to have the greatest effect on rumen bacteria. Wheat and barley straw have
very low P contents, not sufficient for the bacteria (Hungate 1966). Rumen bacteria are
somewhat halophilic, e.g. the Na concentration must be at least 100 mM to support the
growth of R. albus (Mackie and Therion 1984). Sulfur can be expected to be limiting
in diets low in protein. Sulfate is a good source of S for rumen bacteria, although not
for the protozoa (Hungate 1966). However, reduction of sulfate for utilisation  in
biosynthesis requires energy.
Feeding experiments
Hemsley and Moir (1963) conducted experiments with sheep. Milled oat hay (0.7% N)
was supplemented as shown in Table 4 which also gives some of the results. The addition
of BC-VFA increased intake and digestion of the hay over that obtained by supplement-
ing only with urea. In another experiment with sheep (van Gylswyk 1970) a poor quality
teff hay (0.5% N) resulted in increased intake on supplementation with 3% urea and a
further increase occurred when BC-VFA were added (Table 5). With the latter diet the
sheep often ate the complete ration and intake was therefore restricted. This part of the
experiment was subsequently repeated with unrestricted intake. It showed that the
urea/VFA supplement raised the quality of the hay to achieve more than maintenance.
Numbers and types of cellulolytic bacteria were determined and these are shown in
Table 6.
Table 3. Vitamin requirements of important fibre-digesting rumen bacteria.
Species Biotin Folate PABA
?
Pyridoxine Thiamin Riboflavin
F. succinogenes +?+? ??
R. flavefaciens ++++ ++
R. albus +?++? ?
B. fibrisolvens ++?+
?
p-Aminobenzoic acid.
Source: Wolin and Miller (1988).
124 N.O. van Gylswyk
The number of cellulolytic and total culturable bacteria did not change when the hay
was supplemented with urea. However, the increased rate of passage of feed particles
plus bacteria would mean that the bacteria had grown more rapidly and cellulolytic
activity increased to account for greater digestibility (Table 5). Further addition of a
BC-VFA mixture increased the numbers of both cellulolytic and non-cellulolytic bac-
teria. The percentage of cellulolytic bacteria remained the same indicating the depen-
dance of the total bacterial flora on the activity of the fibre-digesting ones. Together the
ruminococci were the predominant cellulolytic species on all three diets followed by B.
fibrisolvens and E. cellulosolvens. The proportion of ruminococci increased with
supplementation of urea and even more so when the VFA were included. Unfortunately
the method used for the primary culture of cellulolytic bacteria was not suitable for the
growth of F. succinogenes. It is most likely that this species was an important component
of the cellulolytic population because poor quality roughages such as wheat straw
(Byrant and Burkey 1953) and maize straw (van Gylswyk and van der Toorn 1986) fed
to cows and sheep, respectively, support rumen populations containing several-fold
more F. succinogenes than ruminococci.
In vitro experiments
Hemsley and Moir (1963) found that molasses and sucrose could partly replace
BC-VFA. Other workers also found that even small amounts of easily fermented
carbohydrate such as starch, added to poor quality roughage, could enhance its utilisa-
tion. It was suggested (Br?ggemann and Giesecke 1967) that lysis of bacteria not
requiring BC-VFA could release BC amino acids which are broken down to BC acids.
Evidence of this was obtained in vitro when Ruminobacter amylophilus grew in a simple
medium containing starch (Miura et al 1980). After growth had ceased, BC-amino acids
were released. When it was co-cultured with the amino acid-dependent Megasphaera
Table 4. Effects of supplementing a low protein oat hay with urea and other additions when fed to sheep .
Daily MRT Digestion of
intake of hay DMD
?
of solids
?
cotton thread
?
Treatment (g) (%) (h) (h)
Basal (hay + mineral) 613 54.5 48 84
Basal + 3% urea 839 62.2 46 38
Basal + 3% urea +
10% molasses 896 61.7 41 38
Basal + 3% urea +
6.6% sucrose 906 62.8 43 43
Basal + 3% urea +
0.56% VFA mix 989 62.0 38 38
?
Dry-matter digestibility.
?
Mean retention time of coloured particles in the rumen.
?
Time taken to achieve a 50% loss in weight of cotton thread bundles suspended in the rumen.
Source: Hemsley and Moir (1963).
Factors limiting proliferation of desirable bacteria in the rumen 125
elsdenii in a similar  medium,  also  containing  glucose,  the BC-amino  acids were
converted to BC-VFA required for the growth of R. albus. A small amount of starch in
medium containing glucose and cellobiose gave successive growth of the three species
(when mixed) in the order R. amylophilus, M. elsdenii and R. albus. A second type of
mechanism was also demonstrated (Stevenson 1978). A range of pure cultures of
different rumen bacteria were grown separately in medium containing glucose, cello-
biose, starch and (NH
4
)
2
SO
4
. All strains excreted amino acids during active growth.
The BC amino acid, valine, was generally excreted in relatively large amounts. These
two mechanisms could explain the BC-VFA-sparing action of starch or other readily
fermentable materials added to poor quality roughages. It would, of course, also be a
protein-sparing action. However, protein is a source of peptides and the growth of the
important hemicellulolytic bacterium, P. ruminicola, as well as others, is stimulated by
these. Furthermore, protein is a source of reduced sulfur.
Vitamins
Vitamins may be limiting for growth of bacteria in the rumen (van Gylswyk et al 1992).
Pure cultures of six species of common rumen bacteria were grown separately in
filter-sterilised rumen fluid from cows and sheep fed good diets. Glucose was the only
added energy source. Parallel cultures were grown with yeast extract as the only other
additive. It was found that P. ruminicola, the most abundant species in the rumen of
Table 5. Effects of supplementing low protein teff hay with urea and a mixture of branched and straight
chain VFA when fed to sheep.
Daily intake Cellulose Hemicellulose Loss in body
of hay digestibility digestibility weight over
Treatment (g) (%) (%) 12 weeks (%)
Basal (hay + minerals) 717
?
57 45 19.0
Basal + 3% urea 902
?
72 64 8.2
Basal + 3% urea +
VFA mixture 1077
?
72 67 3.8
>1500
?
ND
?
ND gain
?
Unrestricted intake.
?
Restricted intake.
?
Not done.
Source: van Gylswyk (1970).
Table 6. Mean counts and proportions (%) of cellulolytic bacteria in the rumen of sheep fed low protein teff
hay supplemented with urea and a mixture of branched and straight chain VFA.
Cellulolytic counts R. albus R. flavefaciens B. fibrisolvens E. cellulosolvens
Treatment (? 10
?7
/g ingesta) (%) (%) (%) (%)
Basal (hay + minerals) 4.0 34 26 39 1
Basal + 3% urea 3.5 42 28 22 8
Basal + 3% urea +
VFA mixture 9.8 46 46 5 4
Source: (van Gylswyk 1970).
126 N.O. van Gylswyk
animals from which rumen fluid was taken, grew almost as rapidly in rumen fluid free
of yeast extract as in that containing yeast extract. B. fibrisolvens, the second most
numerous bacterium, grew more rapidly with yeast extract than without but at a greater
rate than species occurring in much lower numbers in the rumen. Growth stimulating
factors were only slowly released from ruminal microbes and feed particles. P. r u m i n i -
cola appears to have low or no requirements for vitamins. A B. fibrisolvens strain was
subsequently found to be stimulated by pyridoxine and folic acid and peptides when
grown in rumen fluid containing glucose (Wejdemar 1994). It is considered possible
that the vitamin supply in the rumen is one of the factors that regulate the proportions
of different species in the rumen. The possibility of bloating when vitamins, together
with other nutrients, are in plentiful supply was examined recently (van Gylswyk 1994).
Minerals
In the animal experiments cited above, mineral supplements were provided, but low
quality roughages are usually deficient in minerals for rumen bacteria. Recently, the
effects of P, Ca and Mg concentrations on the growth of F. succinogenes and R.
flavefaciens were examined (Komisarczuk-Bony et al 1994). Concentrations below 15
mg/l for P and 5 mg/l for  Mg reduced growth and cellulose degradation  by R.
flavefaciens while the corresponding concentrations for F. succinogenes were 5 mg/l
and close to zero, respectively. The response to Ca concentrations was similar for both
species with inhibitions below 1 mg/l. While the Ca content of a straw was adequate
for optimal degradation by F. succinogenes, the P content was so low that dry matter
digestibility was decreased appreciably (>25%). These results suggest that the prolif-
eration of F. succinogenes is favoured more than that of R. flavefaciens in the rumen
of animals feeding on poor quality hay or straw; at least with respect to P and Mg. P.
ruminicola subsp. ruminicola, which possesses cytochromes, will require Fe. This
element is unlikely to be lacking in roughages as all plant material contains cyto-
chromes. Although the iron content of rumen fluid is low it is also quite constant. As
mentioned earlier the Na content of such diets could also limit growth and cellulose
degradation by rumen bacteria. The intake of soil will contribute to the alleviation of
mineral deficiencies.
References
Allison M.J. and Bryant M.P. 1963. Biosynthesis of branched-chain amino acids from branched-
chain fatty acids by rumen bacteria. Archives of Biochemistry and Biophysics 101:269.
Allison M.J., Bryant M.P., Katz I. and Keeny M. 1962. Metabolic function of branched-chain
volatile fatty acids, growth factors for ruminococci. II. Biosynthesis of higher branched-chain
fatty acids and aldehydes. Journal of Bacteriology 83:1084.
Br?ggemann J. and Giesecke D. 1967. The effect of urea on rumen microbiology and metabo-
lism. In: Briggs M.H. (ed), Urea as a Protein Supplement. Pergamon Press, Oxford, UK. pp.
125?154.
Bryant M.P. 1973. Nutritional requirements of the predominant rumen cellulolytic bacteria.
Federation Proceedings 32:1809.
Factors limiting proliferation of desirable bacteria in the rumen 127
Bryant M.P. and Burkey L.A. 1953. Numbers and some predominant groups of bacteria in the
rumen of cows fed different rations. Journal of Dairy Science 36:218.
Bryant M.P. and Doetsch R.N. 1954. Factors necessary for the growth of Bacteroides succino-
genes in the volatile acid fraction of rumen fluid. Science 120:944.
Caldwell D.R., White D.C., Bryant M.P. and Doetsch R.N. 1965. Specificity of the heme
requirement for growth of Bacteroides ruminicola. Journal of Bacteriology 90:1645.
Chesson A. and Forsberg C.W. 1988. Polysaccharide degradation by rumen microorganisms.
In: Hobson P.N. (ed), The Rumen Microbial Ecosystem. Elsevier Applied Science, London,
UK. pp. 251?284.
El-Shazly K. 1952. Degradation of protein in the rumen of sheep. 1. Some volatile fatty acids,
including branched-chain isomers, found in vivo. Biochemical Journal 51:640.
van Gylswyk N.O. 1970. The effect of supplementing a low-protein hay on the cellulolytic
bacteria in the rumen of sheep and on the digestibility of cellulose and hemicellulose. Journal
of Agricultural Science (Cambridge) 74:169.
van Gylswyk N.O. 1994. Aspects of the supply of nutrients to bacteria in the rumen. In:
El-Shazly K. (ed), Manipulation of Rumen Microorganisms. Proceedings of the International
Conference on Manipulation of Rumen Microorganisms, September 1992. Department of
Animal Production, Alexandria University, Egypt. pp. 90?97.
van Gylswyk N.O. and van der Toorn J.J.T.K. 1986. Enumeration of Bacteroides succinogenes
in the rumen of sheep fed maize-straw diets. FEMS Microbiology Ecology 38:205.
van Gylswyk N.O., Wejdemar K. and Kulander K. 1992. Comparative growth rates of various
rumen bacteria in clarified rumen fluid from cows and sheep fed different diets. Applied and
Environmental Microbiology 58:99.
Harfoot C.G. and Hazlewood G.P. 1988. Lipid metabolism in the rumen. In: Hobson P.N. (ed),
The Rumen Microbial Ecosystem. Elsevier Applied Science, London, UK. pp. 285?322.
Hemsley J.A. and Moir R.J. 1963. The influence of higher volatile fatty acids on the intake of
urea-supplemented low quality cereal hay by sheep. Australian Journal of Agricultural
Research 14:509.
Hungate R.E. 1966. The Rumen and its Microbes. Academic Press, New York, USA. 533 pp.
Komisarczuk-Bony S., Fonty G. and Durand M. 1994. Effects of mineral deficiencies on growth
and solid degradative activity of two ruminal cellulolytic bacteria: Fibrobacter succinogenes
and Ruminococcus flavefaciens. Proceedings of the Society of Nutrition Physiology 3:158.
van der Linden Y., van Gylswyk N.O. and Schwartz H.M. 1984. Influence of supplementation
of corn stover with corn grain on the fibrolytic bacteria in the rumen of sheep and their
relation to the intake and digestion of fibre. Journal of Animal Science 59:772.
Mackie R.I. and Therion J.J. 1984. Influence of mineral interactions on growth efficiency of
rumen bacteria. In: Gilchrist F.M.C. and Mackie R.I. (eds), Herbivore Nutrition in the
Subtropics and Tropics. The Science Press, Craighall, South Africa. pp. 455?477.
McDowell L.R., Conrad J.H. and Ellis G.L. 1984. Mineral deficiencies and imbalances, and
their diagnosis. In: Gilchrist F.M.C. and Mackie R.I. (eds), Herbivore Nutrition in the
Subtropics and Tropics. The Science Press, Craighall, South Africa. pp. 67?88.
Menahan L.A. and Schultz L.H. 1963. Effect of leucine and valine administration on volatile
fatty acids. Journal of Dairy Science 46:640.
Miura H., Horiguchi M. and Matsumoto T. 1980. Nutritional interdependence among rumen
bacteria, Bacteroides amylophilus, Megasphaera elsdenii and Ruminococcus albus. Applied
and Environmental Microbiology 40:294.
Pittman K.A. and Bryant M.P. 1964. Peptides and other nitrogen sources for growth of
Bacteroides ruminicola. Journal of Bacteriology 88:401.
128 N.O. van Gylswyk
Russell J.B. 1985. Fermentation of cellodextrins by cellulolytic and noncellulolytic rumen
bacteria. Applied and Environmental Microbiology 49:572.
Stack R.J. and Hungate R.E. 1984. Effect of 3-phenyl propanoic acid on capsule and cellulases
of Ruminococcus albus 8. Applied and Environmental Microbiology 48:218.
Stevenson I.L. 1978. The production of extracellular amino acids by rumen bacteria. Canadian
Journal of Microbiology 24:1236.
Wegner G.H. and Foster E.M. 1963. Incorporation of isobutyrate and valerate into cellular
plasmalogen by Bacteroides succinogenes. Journal of Bacteriology 85:53.
Wejdemar K. 1994. Factors that stimulate the growth of Butyrivibrio fibrisolvens in rumen fluid.
Proceedings of the Society of Nutrition Physiology 3:161.
Wolin M.J. and Miller T.L. 1988. Microbe-microbe interactions. In: Hobson P.N. (ed), The
Rumen Microbial Ecosystem. Elsevier Applied Science, London, UK. pp. 343?360.
Factors limiting proliferation of desirable bacteria in the rumen 129
Exploitation of rumen microbial enzymes
to benefit ruminant and non-ruminant
animal production
L.B. Selinger
1
, K.-J. Cheng
1
,L.J.Yanke
1
,H.D.Bae
1
,
T.A. McAllister
1
and C.W. Forsberg
2
1. Lethbridge Research Centre, Lethbridge, Alberta, Canada T1J 4B1
2. Department of Microbiology, University of Guelph, Guelph, Ontario, Canada N1G 2W1
Abstract
As competition in the livestock industry increases, producers are actively
seeking technological advances through which to increase their production
efficiency. Supplementation of diets of non-ruminant animals with fibrolytic
enzymes, such as cellulases, xylanases and ?-glucanases, increases their feed
conversion efficiency and growth rate. These enzymes enhance the release
of sugars from plant cell wall polymers, thereby increasing their availability
to the animal and eliminating some of their naturally occurring anti-nutri-
tional effects. Recent research indicates that enzyme supplementation can
also produce similar improvements in ruminant production, even though the
rumen microbial population is known to produce an extensive array of potent
endogenous fibrolytic enzymes.  Widespread adoption  of  this  promising
technology has been hampered in the past by the cost and inconvenience of
enzyme production and delivery. However, over the last decade, advances in
recombinant DNA technology have significantly improved microbial pro-
duction systems. In addition, the rumen has been recognised as a rich,
alternative source of genes for industrially useful enzymes and novel strate-
gies are being developed for effective delivery of these gene products. Thus,
the biotechnological framework is in place to achieve substantial improve-
ments in animal production through enzyme supplementation.
Introduction
There is a growing trend in the non-ruminant livestock industry to supplement diets with
fibrolytic enzymes. Amending rations with cellulases and/or xylanases increases the
feed efficiency of livestock. Enzymatic hydrolysis of cellulose and xylan to simple
sugars (e.g. glucose and xylose) provides the non-ruminant animal with carbon sources
that are normally not made available by intestinal enzymes. Furthermore, enzymes
eliminate certain forms of these polymers (e.g. arabinoxylan, found in wheat and rye;
?-glucan, in barley and oats) that may have deleterious effects on nutrient absorption
and promote intestinal disturbances by pathogenic enteric microorganisms.
Unlike non-ruminant animals, ruminants have an extensive array of microbial fibro-
lytic enzymes produced in the rumen, and these enzymes play an important role in the
ruminant digestive process. Fibrolytic activity in the rumen arises primarily from the
activities of three bacterial species: Fibrobacter succinogenes, Ruminococcus albus and
R. flavefaciens (Forsberg and Cheng 1992). Enzymes produced by other microorgan-
isms, including fungi and protozoa, also contribute to fibre degradation. Ruminal fungi
are noted for their production of potent fibrolytic enzymes and their ability to degrade
the most recalcitrant of plant cell wall polymers (Forsberg and Cheng 1992; Forsberg
et al 1993; Trinci et al 1994; Wubah et al 1993). Nonetheless, recent evidence suggests
that enzyme supplementation can also dramatically increase the average daily gain of
ruminants (T.A. McAllister and K.-J. Cheng, unpublished data).
Fibrolytic activity in the rumen is estimated to be 10 times higher than in any other
known fermentation system. The microflora of the rumen represents a rich and under-
utilised source of superior fibrolytic enzymes. In this review, the current status of the
inclusion of enzymes and microorganisms in animal feeds is outlined briefly, and some
of the recent advances in enzyme technology are highlighted. From this perspective, we
examine the microflora of the rumen as a source of enzymes and discuss the application
of these enzymes in the livestock feed industry.
Feed enzymes
Silage
Some commonly ensiled plants (e.g. grasses and, in particular, alfalfa) are low in soluble
carbohydrate and therefore often ensile poorly. The inclusion of cellulase and endoxy-
lanase enzyme mixtures from a variety of fungi or bacteria during ensiling increases the
amount of free soluble carbohydrate available for fermentation (van Vuuren et al 1989).
Consequently, conversion of carbohydrate to lactic acid is more extensive, resulting in
a lower pH in enzyme-treated silage as compared to untreated silage (Chen et al 1994;
Selmer-Olsen et al 1993). In some instances, enzyme treatment has increased the intake
(DM basis) of silage by dairy cows and improved milk production, but these responses
are less pronounced for cereal silages (Chen et al 1994) or silages with a high dry matter
content (Fredeen and McQueen 1993). Enzyme treatment of grass silages has increased
(Jacobs et al 1992), decreased (Jacobs and McAllan 1991) and had no effect (Jacobs
and McAllan 1992) on the apparent whole-tract digestibility of organic matter. In a
single study, different enzyme preparations were found to increase and decrease the
digestibility of organic matter in sheep (Chamberlain and Robertson 1989). These
inconsistencies emphasise the need for more precise definitions of enzyme preparations
and of the forage properties for which a given enzyme treatment is most likely to elicit
a positive response in animal performance.
Poultry and swine diets
The addition of glucanases and pentosanases (endoxylanases) to poultry diets contain-
ing barley, oats, wheat and rye has been shown to increase feed efficiency and to enhance
132 L.B. Selinger, K.-J. Cheng, L.J. Yanke, H.D. Bae, T.A. McAllister and C.W. Forsberg
growth rates by 5 to 17% (Campbell and Bedford 1992; Classen et al 1991; Marquardt
et al 1987; Rotter et al 1987). Protein sources such as oilseed meals (e.g. canola meal,
soybean meal), commonly incorporated into these rations, contain substantial amounts
of non-starch polysaccharides. The addition of cell-wall digesting enzymes to canola-
based diets for poultry increases the digestibility of non-starch polysaccharides from 3
to 37% (Slominski and Campbell 1990). The benefits of enzyme supplementation on
growth performance of swine are more variable with improvements in average daily
gain ranging from 0 to 15% (Inborr 1990). This variability can be partially attributed to
the age of the pigs, since in older pigs (> 15 weeks) considerable degradation of
mixed-link ?-glucans by the ileal microflora and greater endogenous enzyme secretions
may negate the beneficial effect of enzyme supplements on digestion (Johnson et al
1993).
The mechanisms by which glucanases and xylanases enhance animal performance
have been elucidated. Barley and oats contain mixed-linkage (1?3,1?4)-B62D-glucans
as a major constituent of the endosperm cell wall whereas the endosperm cell walls of
rye and wheat are rich in arabinoxylans (Pettersson and Aman 1989). During digestion
these polymers are released from cereal cell walls, become hydrated and dramatically
increase the viscosity of the digesta. An increase in viscosity impairs the activity of
digestive enzymes, reduces digestive flow and nutrient absorption and contributes to
the formation of ?sticky? faeces which is especially problematic in poultry operations,
as it adheres avidly to feathers and to eggs (Classen et al 1988, 1991; Pettersson and
Aman 1989).
Small quantities of purified ?-glucanase (Edney et al 1989) and/or endoxylanase
(GrootWassink et al 1989) in poultry diets cleaves these polymers and lowers the
viscosity of the digesta (Almirall et al 1993; Hesselman and Aman 1986; Pettersson and
Aman 1989; Rotter et al 1990). The associated improvements in nutrient utilisation and
decreases in digestive upset reduces the need for antibiotics in poultry and swine diets
(Inborr 1990; Johnson et al 1993).
The addition of microbial phytase to poultry feed increases phosphorus availability
by more than 60% and reduces phosphorus in the droppings by 50%, resulting in
significant increases both in growth rate and feed conversion (Simons et al 1990).
Similarly, the addition of phytase to diets for growing pigs increases the apparent
digestibility of phosphorus by 24?53% and lowers the amount of phosphorus in the
faeces by 35% (Ketaren et al 1993; Simons et al 1990). The inclusion of phytase in the
diet induces hydrolysis of plant phytate (inositol hexaphosphate), releasing dietary
phosphate in the gastrointestinal tract. Phytate constitutes approximately 1 to 2%
(wt/wt) of cereals and oil seeds, and phosphorus as phytate accounts for 60 to 90% of
the total phosphorus present in cereal diets (Cheryan 1980; Swick 1991). Phytate is a
strong chelating agent and binds with proteins and minerals, reducing the bioavailabil-
ity of these nutrients in feeds (Erdman 1979; Cheryan 1980). Consequently, supple-
mentation of poultry and swine diets with phytase not only improves phosphorus
retention, but often enhances the digestion of dry matter and crude protein, and the
utilisation of calcium and zinc (Lei et al 1993; Mroz et al 1994; Zyla and Koreleski
1993).
Exploitation of rumen microbial enzymes to benefit animal production 133
Ruminant diets
In the 1960s, several studies were conducted to examine the efficacy of adding enzyme
preparations to diets for ruminants (Burroughs et al 1960; Clark et al 1961; Perry et al
1966; Rovics and Ely 1962). Clark et al (1961) found that enzyme supplementation
improved average daily gains by as much as 20%. However, responses to enzyme
preparations were inconsistent (Perry et al 1966; Rust et al 1965; Theurer et al 1963),
possibly because of differences in diet composition, method of enzyme application and
the stability and activity of the enzyme preparations.
Improvements in fermentation technology and the biotechnological development of
more defined enzyme preparations have prompted a renewed interest in the use of
enzymes in ruminant diets. More recent studies have attempted to define the conditions
in a given feed that are most likely to result in a favourable animal response to enzyme
supplementation (Feng et al 1992a, 1992b; Judkins and Stobart 1988). Factors such as
substrate specificity of the enzymes, moisture level of the feed, the time required for
enzyme-substrate interaction and temperature at time of treatment are all likely to
influence the extent to which enzymes enhance the utilisation of feeds by ruminants.
The binding of enzymes to their appropriate substrates is an absolute prerequisite for
the digestion of plant cell walls in the rumen (McAllister et al 1994). It follows, then,
that treatment methods enabling adequate interaction between enzyme and substrate
prior to feeding are most likely to improve animal performance.
In light of the exceptional fibre-digesting capacity of the rumen, it is difficult to
explain why pretreatment of forages with fibrolytic enzymes prior to consumption
would further improve the utilisation of these feeds by ruminants. The enhanced
degradation in enzyme-treated forages may be related to the increase in passage rate
and reduction in the retention time of forage particles in the rumen of cattle as observed
by Feng et al (1992b). Scanning electron microscopy has shown that the surface of
plant cell walls is occasionally colonised by a single bacterial morphotype (Cheng et
al 1981, 1984). Complete digestion of plant cell walls, however, requires an array of
enzymes and therefore a single morphotype may not produce the diversity of enzymes
required to effect complete cell wall digestion. Furthermore, these primary colonisers
may limit the access of complementary bacteria to the surface of the plant cell walls.
Under such circumstances, prior treatment of forages may contribute to the digestive
process the precise enzymes which would otherwise be limiting the rate and/or extent
of plant cell wall digestion. Ultimately, enzyme cocktails should be designed to
overcome the specific constraints limiting digestion of a particular type of forage.
Component enzymes in such cocktails might vary even for a given forage, targeting
particular maturity levels or dry matter concentrations. Developments in biotechnol-
ogy make it feasible to engineer such enzyme cocktails for xylanase and ?-glucanase,
but the technology for specific production of many of the other enzymes (e.g. ferulic
acid esterase, acetylxylan esterase, arabinofuranosidase) required for cell wall diges-
tion is lacking. In the short term, then, improvements in ruminant performance arising
from enzyme treatment of forages will most likely be the result of treatment of the
forage with broad spectrum crude enzyme extracts from cellulolytic microorganisms
(e.g. Aspergillus spp., Trichoderma reesei).
134 L.B. Selinger, K.-J. Cheng, L.J. Yanke, H.D. Bae, T.A. McAllister and C.W. Forsberg
Recent advances in the development
of enzymes to enhance digestion of feedstuffs
A major factor limiting the widespread use of enzymes in the livestock industry is the
expense of traditional large-scale fermentations and downstream processing. This has
prompted manufacturers and researchers to seek more powerful enzymes and to exam-
ine alternative systems for enzyme production and delivery. Recently, application of
recombinant DNA technology has enabled manufacturers to increase the volume and
efficiency of enzyme production, and to create new products (Hodgson 1994; Ward and
Conneely 1993). The original source organism need no longer limit the production of
commercial enzymes. Genes encoding superior enzymes can be transferred from
organisms such as anaerobic bacteria and fungi, typically impractical for commercial
production, into well characterised industrial microbial production hosts (e.g. Aspergil-
lus and Bacillus spp.). These genes may also be transferred to novel plant (Pen et al
1993; van Rooijen and Moloney 1994) and animal (Hall et al 1993) expression systems.
The rumen as a source of enzymes
The rumen is increasingly being recognised as a particularly promising source of
superior fibrolytic enzymes. Cellulases and xylanases produced by ruminal fungi are
among the most active fibrolytic enzymes described to date (Gilbert et al 1992; Trinci
et al 1994). The quest to elucidate the mechanisms of fibre digestion and to find more
efficacious enzymes for industrial applications, and technological developments which
now allow the genetic manipulation of rumen microorganisms have inspired the cloning
of a growing number of genes from ruminal bacteria and fungi. At least 75 different
genes, the majority of which encode enzymes with a role in fibre digestion, have been
cloned thus far from ruminal microorganisms. Most of these have been isolated from a
small number of bacterial species, including R. albus, R. flavefaciens, F. succinogenes,
Butyrivibrio fibrisolvens and Prevotella ruminicola (reviewed by Flint 1994; Forsberg
et al 1993; Hespell 1989; Wallace 1994). Several of the bacteria possess a multiplicity
of genes coding for different cellulases and xylanases. For example, F. succinogenes
appears to possess nine non-homologous glucanases (Malburg and Forsberg 1993;
McGavin et al 1989) and four unrelated xylanases (Malburg et al 1993; Paradis et al
1993; Sipat et al 1987) in addition to a gene coding for cellodextrinase activity (Gong
et al 1989). R. flavefaciens has also been found to produce multiple xylanases (Flint et
al 1994). The presence of numerous glycanases in a single bacterium suggests that they
all may have a role during growth on glycan substrates. Verification of this hypothesis
will have to wait until mutant strains of cellulolytic and xylanolytic bacteria with null
mutations in each gene have been isolated and tested for growth on cellulose and xylan.
Researchers have only recently turned to the study of the genetics of anaerobic fungi
isolated from the rumen. Enticed by their powerful fibrolytic activity and ability to
utilise the most recalcitrant of plant cell wall polymers, researchers have cloned at least
27 genes from four fungal species. The cloned genes include five cellulases (Black et
al 1994; Selinger et al 1994; Xue et al 1992a, 1992b; Zhou et al 1994) and three
Exploitation of rumen microbial enzymes to benefit animal production 135
xylanases (Black et al 1994; Gilbert et al  1992; Tamblyn Lee et  al  1993) from
Neocallimastix patriciarum, an endoglucanase from Orpinomyces joyonii (Selinger et
al 1994), eight cellulases, four xylanases and five mannanases from Piromyces sp. (Ali
et al 1995) and a phosphoenolpyruvate carboxykinase from N. frontalis (Reymond et al
1992). A notable report (Gilbert et al 1992) of a ruminal fungal xylanase (XynA from
N. patriciarum) possessing the highest specific activity currently known (6000 ?mol of
xylose/min per milligram of XynA) will likely stimulate a more rigorous examination
of these microbes.
Gene expression and delivery systems
The cost of livestock performance-enhancing enzymes can be reduced through the use
of more effective expression and delivery systems. Recombinant bacterial and plant
systems are being examined for their efficacy in improving the ensiling of low quality
forages, and in producing and delivering enzymes on a continuous basis to livestock.
Ultimately, researchers envisage eliminating the need to use the feed as a ?middle man?
in enzyme delivery altogether by enabling transgenic animals to secrete endogenous
fibrolytic enzymes from the pancreas into the small intestine.
Silage inoculants
The cost and inconvenience of adding enzymes prior to the ensiling of forages low in
carbohydrates could be avoided if the bacterial species in the inoculant were also
capable of producing the enzymes required to release soluble sugars from plant cell
walls. However, attempts at isolating such microorganisms have proven unsuccessful,
and as a result attempts have been made to develop genetically engineered strains of
lactobacilli which produce fibrolytic enzymes. Genetically modified Lactobacillus
strains have now been developed which express cellulase and xylanase genes isolated
from other organisms (Baik and Pack 1990; Bates et al 1989; Scheirlinck et al 1989,
1990) and which have, in at least one case, been shown to exhibit competitive growth
in silage (Sharp et al 1992). Furthermore, those researchers observed that pM25, a pSA3
derivative containing a Clostridium thermocellum cellulase gene, was maintained at
high levels by the rifampin-resistant host cells (by 100% of host cells when pM25 was
present as a chromosomally integrated element and by 85%  when  present  as an
autonomously replicating plasmid element). We have observed similar stabilities of
recombinant plasmids when modified Lactobacillus plantarum constructs were applied
to the forage during the ensiling of alfalfa. Plasmid constructs pSAG (pSA3::Bacillus
subtilis endoglucanase gene) and pSAX (pSA3::B. circulans xylanase) were found to
persist in 83% and 62% of rifampin-resistant L. plantarum cells, respectively.
Persistence of newly developed recombinant lactobacilli beyond their target environ-
ments is an important, but under-studied phenomenon. In a single study, Sharp et al
(1994) demonstrated the rapid loss of both unmodified and recombinant L. plantarum
silage inoculants from the rumen. Elimination of these strains was attributed to predation
by protozoa and other undefined factors (e.g. turnover rate, cell death). These findings
are also relevant to the use of these genetically engineered microorganisms (GEMs) as
136 L.B. Selinger, K.-J. Cheng, L.J. Yanke, H.D. Bae, T.A. McAllister and C.W. Forsberg
probiotic feed additives, in which case it is essential that they establish and persist in
the gastrointestinal tract. Factors influencing environmental persistence must be iden-
tified and taken into consideration during selection of microbial strains for genetic
manipulation.
Two major drawbacks of plasmid constructs are their instability in the absence of
antibiotic selective pressure and environmental concerns arising from the potential for
horizontal transfer of antibiotic resistance genes carried on plasmids. Consequently,
methods have been sought to integrate recombinant genetic material into the L. plan-
tarum chromosome. Progress in this direction was achieved when the pSA3 plasmid
was integrated into the L. plantarum chromosome (Rixon et al 1990; Sharp et al 1992),
but inclusion of the erythromycin resistance gene was still of environmental concern.
Another method, more suitable for GEMs destined for environmental release, permits
the integration of select DNA fragments into the chromosome through a two-step
procedure involving gene inactivation and replacement (Biswas et al 1993; Hols et al
1994). In the first step, the entire plasmid construct is forced to integrate, via homolo-
gous recombination, into a predetermined locus. The process is completed by select
recombinational events between tandemly repeated elements which result in the loss of
the vector and maintenance of the desired gene.
A number of researchers (Baik and Pack 1990; Bates et al 1989; Scheirlinck et al
1989, 1990; Sharp et al 1992) have reported the introduction of cellulase and xylanase
genes into L. plantarum, but the development of efficacious recombinant silage inocu-
lants has been hampered by low levels of heterologous gene expression (Hols et al 1994;
Oosting et al 1995). The construction of hybrid genes, consisting of homologous
expression-secretion signals fused with the structural component(s) of heterologous
genes encoding fibrolytic enzymes, may provide an improvement in the levels of gene
expression. Hols et al (1994) demonstrated the effectiveness of this technique by
isolating promoter signal sequence regions from L. plantarum and using them to drive
expression and secretion of foreign amylase and levanase genes in L. plantarum. The
heterologous expression of levanase enabled the modified L. plantarum isolates to
utilise inulin as a carbon source. Future research employing recently developed ad-
vances in gene expression and integration technology should lead to the development
of genetically modified Lactobacillus strains as silage inoculants.
Probiotics
Probiotics have been defined by Fuller (1989) as live microbial feed supplements
which benefit the host animal by improving the intestinal microbial balance. Bacteria
commonly fed to domestic animals as probiotics include L. acidophilus, L. bulgaricus,
L. plantarum, L. casei, Enterococcus faecium and other enterococcal species (Jones
1991).
Lactobacillus spp. are naturally occurring inhabitants of the gastrointestinal tract,
especially of poultry, but their role in the ecology of the tract is not well understood
(Fuller 1989; Juven et al 1991). L. acidophilus is the most common lactic acid-producing
bacterium colonising the epithelial surfaces of the gastrointestinal tract. Its beneficial
effects are thought to be exerted in a number of ways, including the secretion of
Exploitation of rumen microbial enzymes to benefit animal production 137
bacteriocins or bacteriocin-like substances which inhibit both (presumably pathogenic)
Gram-positive and Gram-negative bacteria, the production of hydrogen peroxide and
the production of lactic acid, which has a strong germicidal action at low pH (Gilliland
1990). Alternate mechanisms by which various species of Lactobacillus inhibit other
microorganisms include the deconjugation of bile acids, thereby enhancing their antimi-
crobial nature, and the production of antimicrobial compounds such as acidolin and
reuterin (Juven et al 1991). In addition to inhibiting the growth of harmful bacteria,
Lactobacillus spp., as primary colonisers, also provide digestible nutrients or digestive
enzymes beneficial to the animal (Fuller 1989). Adding fibrolytic agents to the enzymes
currently produced by the lactobacilli would no doubt confer upon them a competitive
advantage for nutrients in the gut, and secondarily would benefit the host animal by
making available more of the dietary energy contained in plant polymers. Progress
toward this goal has been made. Baik and Pack (1990) reported high extracellular
glucanase activity from L. acidophilus in which they had induced expression of a B.
subtilis endoglucanase gene. To date, however, there are no reports of the effect(s) of
genetically modified lactobacilli on animal health.
Feeding probiotics has been more prevalent in non-ruminant livestock industries than
in ruminant feeding, although opportunities may exist for improving the health and
growth of ruminants in this way (Asby et al 1989; Aseltine 1991). Other reports,
however, have indicated no response (Harrison et al 1988; Mir and Mir 1994) in
ruminants fed these agents. Future studies must identify the mode of action of these feed
additives if uniform and reproducible animal responses are to be obtained.
Microflora of the gut
The manipulation of digestion in ruminants through genetic modification of anaerobic
bacteria is currently being investigated in many laboratories, and numerous reviews
have been written on this subject (Armstrong and Gilbert 1985; Flores 1989; Forsberg
and Cheng 1992; Forsberg et al 1986, 1993; Russell and Wilson 1988; Wallace 1994).
Strategies proposed for the manipulation of ruminal bacteria include expression of
heterologous genes coding for fibrolytic enzymes, artificial peptides, bacteriocins and
for detoxification or antiprotozoal factors.
The lack of functional gene transfer systems has in the past impeded progress toward
achieving expression of foreign genes in anaerobic bacteria. However, conjugal and
electroporation gene transfer systems have now been developed and refined for a
number of ruminal bacterial isolates, including P. r u m i n i c o l a (B?chet et al 1993; Flint
et al 1988; Russell and Wilson 1988; Shoemaker et al 1991; Thomson and Flint 1989;
Thomson et al 1992), Streptococcus bovis (Hespell and Whitehead 1991a; Whitehead
1992), B. fibrisolvens (Beard et al 1995; Clark et al 1994; Hespell and Whitehead 1991b;
Teather 1985; Ware et al 1992; Whitehead 1992), R. albus (Cocconcelli et al 1992),
Escherichia coli (Scott and Flint 1995) and Selenomonas ruminantium (Kopecny and
Fliegerova 1994). Gene expression in the above studies was limited to the expression
of vector-specific functions and antibiotic resistance markers. Application of gene
transfer technologies to construction of recombinant anaerobic bacteria, designed for
modification of the rumen environment, is illustrated in the following examples.
138 L.B. Selinger, K.-J. Cheng, L.J. Yanke, H.D. Bae, T.A. McAllister and C.W. Forsberg
Whitehead and Hespell (1989) cloned an endoxylanase gene from the hemicellu-
lolytic ruminal bacterium B. ruminicola 23 and, using the vector pVAL-l, subsequently
introduced it into B. fragilis and B. uniformis, where enzyme activity was increased
dramatically (Whitehead and Hespell 1990). Endoxylanase activities of the newly
constructed xylanolytic strains were up to 50 times higher than that measured in B.
ruminicola grown on xylan. The xylanase gene has since been integrated into the
chromosomal chondroitin lyase II gene of Bacteroides thetaiotaomicron to obtain
stable endoxylanase production in the absence of selective agents (Whitehead et al
1991). The activity of the enzyme in B. thetaiotamicron was still 17 times higher than
that produced by B. ruminicola 23, and was stable when the modified B. thetaiotami-
cron was grown in a carbon-limited chemostat culture for over 60 generations.
Introduction of the new strain into the rumen to increase xylan degradation was
proposed, but it remains to be determined whether or not the genetically modified
organism will persist in sufficiently high numbers to substantially increase hemicellu-
lose digestion, especially since B. thetaiotaomicron is not a naturally occurring species
in the rumen. Animal trials with this novel GEM will provide preliminary information
on its persistence in the rumen and its contribution to ruminal fibre digestion.
Many shrubs and trees native to Australia accumulate monofluoroacetate in their
shoots and are therefore poisonous to domestic livestock (Gregg and Sharpe 1991;
Gregg et al 1994; Jones and Megarrity 1986). Jones and Megarrity (1986) attempted
to detoxify fluoroacetate by introducing bacteria capable of degrading the toxin into
the rumen. An unsuccessful search for microbes indigenous to the rumen and capable
of detoxifying fluoroacetate  prompted the introduction of this phenotype into B.
fibrisolvens (Gregg et al 1994). A transcriptional fusion between the fluoroacetate
dehalogenase gene from Moraxella sp.strainBandtheerm promoter from pAM?1
was ligated to pBHerm (Beard et al 1995), a B. fibrisolvens/E. coli shuttle vector. B.
fibrisolvens OB156 isolates carrying the dehalogenase expression plasmid (pBHf)
were able to detoxify fluoroacetate in vitro and, under non-selective conditions, pBHf
was  maintained  for up  to  500  generations. It was  estimated that at the  level of
dehalogenase expression observed in B. fibrisolvens OB156 (pBHf), toxin protection
for the ruminant would be provided if B. fibrisolvens OB156 (pBHf) cells were present
in the rumen at 10
6
to 10
7
cells/mL (Gregg et al 1994). The stability and dehalogenase
activity of B. fibrisolvens OB156 (pBHf) in vivo has yet to be tested. Attempts are
under way to introduce the gene into a number of bacteria from the rumen and improve
gene expression, to increase theoretical maximum dehalogenase activity levels in
preparation for whole animal ruminal inoculation trials (Gregg et al 1994).
In the rumen microbiology laboratory at the Lethbridge Research Centre, B. fibrisolvens
has also been selected as a host for heterologous gene expression. As a naturally occurring,
ubiquitous resident in the rumen, this bacterium is an ideal candidate for the introduction
of new genetic information to the ruminal environment. Our preliminary experiments were
successful in the introduction and expression of a S. bovis amylase gene in B. fibrisolvens
H17c (Clark 1994; Clark et al 1992, 1995). The amylase gene was cloned onto an E. coli/B.
fibrisolvens shuttle vector, and the resulting construct (pUBLRSA) was introduced into
B. fibrisolvens H17c by electroporation. Expression of the S. bovis amylase was detected
Exploitation of rumen microbial enzymes to benefit animal production 139
by zymogram analysis. The amylolytic activity of S. bovis (pUBLRSA), grown on glucose
as the sole carbon source, was 2.5 times higher than that of wild type B. fibrisolvens H17c.
This is the first report of the expression of a foreign polymer-degrading enzyme in a
bacterium indigenous to the rumen. Research on construction of a more effective expres-
sion system and introduction of additional genes encoding fibrolytic enzymes into native
ruminal species is now under way in our laboratory.
Transgenic plants
Recent advances in plant biotechnology may revolutionise the commercial enzyme
industry by offering alternative, cost-effective methods of enzyme production and
delivery. Large quantities of plant biomass can be produced inexpensively through the
use of existing agricultural infrastructure. Expression of enzymes in plant species
commonly used for animal feed will minimise downstream processing as the whole or
parts of the producing plants are fed directly to livestock. The expression of a xylanase
in tobacco plants has been reported (Herbers et al 1995), as has the expression of a
phytase in tobacco and soybean (Pen 1994; Pen et al 1993; Russell 1994).
Herbers et al (1995) achieved the constitutive expression of a truncated Clostridium
thermocellum xynZ gene (Grepinet et al 1988) encoding a thermostable, high specificity
xylanase in tobacco via the cauliflower mosaic virus (CaMV) 35S promoter. The
proteinase inhibitor II signal sequence was used to target the xylanase into the apoplastic
space. C. thermocellum truncated XynZ comprised up to 4.1% of the total protein in
leaf extracts and 50% of the apoplastic proteins. The thermal stability of XynZ was
exploited to achieve a 31-fold purification through a 20 min, 60?C incubation.
In similar experiments, transgenic tobacco plants were made to express phytase
encoded by the Aspergillus niger phyA gene (van Hartingsveldt et al 1993; Pen 1994;
Pen et al 1993). Constitutive expression was driven by the CaMV 35S promoter and
targeted for the apoplast using the tobacco PR-S signal sequence. In the seeds of
transgenic tobacco plants, the expressed A. niger phytase comprised 1% of total soluble
protein. Feeding trials demonstrated that transgenic tobacco seeds were as effective at
promoting growth of broilers as was a commercial A. niger phytase product or inorganic
phosphorus (Pen 1994).
At the Lethbridge Research Centre, we are combining the utility of plant expression
systems with superior fibrolytic enzymes produced by ruminal fungal isolates. In
collaboration with researchers at the University of Calgary, we have constructed oleosin
(oil body membrane protein)-N. patriciarum xylanase gene fusions in preparation for
Agrobacterium-mediated transfer of the xylanase gene to canola. A number of peptides,
including ?-glucuronidase from E. coli, hirudin  and  interleukin  1-? (amino  acids
163?171), have already been expressed in this way (van Rooijen 1993; van Rooijen and
Moloney 1994). Yields of recombinant protein (in excess of 1% of total seed protein)
at significantly reduced cost are possible with this system.
In  addition to providing an efficient alternative to traditional microbial  systems,
transgenic plants offer the added advantage of a safe and stable formulation system in the
140 L.B. Selinger, K.-J. Cheng, L.J. Yanke, H.D. Bae, T.A. McAllister and C.W. Forsberg
form of seeds (Pen 1994). Recombinant enzymes were stable in seeds stored for up to one
year at 4?C and room temperature (Pen et al 1993; van Rooijen and Moloney 1994).
Transgenic animals
Technological developments enabling introduction of genetic material into domestic
animals (Briskin et al 1991; Ward et al 1989) have validated the concept of direct
expression of these glucanases and xylanases in the animal itself as an option to adding
microbial enzymes to the feed. To be of benefit to the animal, the enzymes should be
expressed in the appropriate tissue, secreted into the lumen of the gastrointestinal tract,
resistant to proteases, and active in environmental conditions (e.g. pH, temperature,
osmolarity) prevailing in the lumen (Forsberg et al 1993; Hall et al 1993).
Expression of fibrolytic enzymes in a non-ruminant animal was first demonstrated
by Hall et al (1993). A truncated endoglucanase E gene from Clostridium thermocellum,
under the control of the elastase I gene promoter, was expressed in the exocrine pancreas
of transgenic mice. Carboxymethylcellulase activity was detectable in small intestinal
contents, demonstrating that the recombinant enzyme was secreted from the pancreas.
Work is under way to develop transgenic mice in our laboratories as well. Twenty genes
have been screened for their suitability for introduction into non-ruminant animals. To
date, gene fusions of three selected genes with the elastase I gene or amylase amy 2.2
gene promoters have been constructed. The obvious potential of this approach is a
powerful incentive for continued efforts in this field of research.
Conclusions
The potential for improvement of the efficiency of non-ruminant and ruminant livestock
feed utilisation by enzyme supplementation is widely recognised. However, the cost of
this technology has inhibited its widespread application in the industry. Reducing the
cost of feed enzymes through improved enzymes and more efficient production systems
is the focus of much research in this area. Microbial isolates from the rumen produce a
wide array of enzymes and comprise an immense and under-utilised gene pool. As
exploration  of  this  enzymological and genetic  resource  progresses from domestic
ruminants to those adapted for survival in diverse habitats worldwide, the isolation of
even more potent enzymes is certain. The technology required for practical and safe
modification and application of these powerful agents is in place.
References
Ali B.R.S., Zhou L., Graves F.M., Freedman R.B., Black G.W., Gilbert H.J. and Hazlewood G.P.
1995. Cellulases and hemicellulases of the anaerobic fungus Piromyces constitute a mul-
tiprotein cellulose-binding complex and are encoded by multigene families. FEMS Microbi-
ology Letters 125:15?22.
Almirall M., Brufau J. and Esteve-Garcia E. 1993. Effect of intestinal viscosity on digestive enzyme
activities of intestinal content and ileal digestibilities of poultry fed barley diets at different ages
Exploitation of rumen microbial enzymes to benefit animal production 141
supplemented with ?-glucanases. In: Wenk C. and Boessinger M. (eds), Enzymes in Animal
Nutrition. Proceedings of the 1st Symposium, Kartause Ittingen, Switzerland. Institute for
Domestic Animal Sciences, Nutrition Group, Zurich, Switzerland. pp. 69?72.
Armstrong D.G. and Gilbert H.J. 1985. Biotechnology and the rumen: A mini review. Journal
of the Science of Food and Agriculture 36:1039?1046.
Asby C.B., McEwan A.D. and Wilson S.M. 1989. Effect of treatment of lambs with a probiotic
containing lactic acid bacteria. Veterinary Record 124:588?589.
Aseltine M. 1991. Certain fungal additives lead to increased ruminant production. Feedstuffs
28:18.
Baik B.-H.  and Pack M.Y.  1990.  Expression of Bacillus  subtilis endoglucanase  gene in
Lactobacillus acidophilus. Biotechnology Letters 12:919?924.
Bates E.E.M., Gilbert  H.J., Hazlewood G.P., Huckle J., Laurie J.I. and Mann S.P. 1989.
Expression of a Clostridium thermocellum endoglucanase gene in Lactobacillus plantarum.
Applied and Environmental Microbiology 55:2095?2097.
Beard C.E., Hefford M.A., Forster R.J., Sontakke S., Teather R.M. and Gregg K. 1995. A stable
and efficient transformation system for Butyrivibrio fibrisolvens OB156. Current Microbi-
ology 30:105?109.
B?chet M., Pheulpin P., Flint H.J., Martin J. and Dubourgier H.-C. 1993. Transfer of hybrid
plasmids based on the replicon pRR17 from Escherichia coli to Bacteroides and Prevotella
strains. Journal of Applied Bacteriology 74:542?548.
Biswas I., Gruss A., Ehrlich S.D. and Maguin E. 1993. High efficiency gene inactivation and
replacement systems for Gram positive bacteria. Journal of Bacteriology 175:3628?3635.
Black G.W., Hazlewood G.P., Xue G.-P., Orpin C.G. and Gilbert H.J. 1994. Xylanase B from
Neocallimastix patriciarum contains a non-catalytic 455-residue linker sequence comprised
of 57 repeats of an octapeptide. Biochemical Journal 299:381?387.
Briskin M.J., Hsu R.-Y., Boggs T., Schultz J.A., Rishell W. and Bosselman R.A. 1991. Heritable
retroviral transgenes are highly expressed in chickens. Proceedings of the National Academy
of Sciences (USA) 88:1736?1740.
Burroughs W., Woods W., Ewing S.A., Greig J. and Theurer B. 1960. Enzyme additions to
fattening cattle rations. Journal of Animal Science 19:458?464.
Campbell G.L. and Bedford M.R. 1992. Enzyme applications for non-ruminant feeds: A review.
Canadian Journal of Animal Science 72:449?466.
Chamberlain D.G. and Robertson S. 1989. The effects of various enzyme mixtures as silage
additives on feed intake and milk production of dairy cows. In: Mayne C.S. (ed), Silage for
Milk Production. Occasional Symposium 23. British Grassland Society, Reading, UK. pp.
187?189.
Chen J., Stokes M.R. and Wallace C.R. 1994. Effects of enzyme-inoculant systems on preser-
vation and nutritive value of haycrop and corn silages. Journal of Dairy Science 77:501?512.
Cheng K.-J., Fay J.P., Coleman R.N., Milligan L.P. and Costerton J.W. 1981. Formation of
bacterial microcolonies on feed particles in the rumen. Applied and Environmental Microbi-
ology 41:93?120.
Cheng K.-J., Stewart C.S., Dinsdale D. and Costerton J.W. 1984. Electron microscopy of
bacteria involved in the digestion of plant cell walls. Animal Feed Science and Technology
10:93?120.
Cheryan M. 1980. Phytic acid interactions in food systems. CRC Critical Reviews in Food
Science and Nutrition 13:297?335.
Clark J.D., Dyer I.A. and Templeton J.A. 1961. Some nutritional and physiological effects of
enzymes for fattening cattle. Journal of Animal Science 20:928.
142 L.B. Selinger, K.-J. Cheng, L.J. Yanke, H.D. Bae, T.A. McAllister and C.W. Forsberg
Clark R.G. 1994. The development of vectors and transfer systems for the rumen bacterium
Butyrivibrio fibrisolvens. PhD dissertation, University of Alberta, Edmonton, Alberta, Can-
ada. 129 pp.
Clark R.G., Cheng K.-J., Selinger L.B. and Hynes M.F. 1994. A conjugative transfer system for
the rumen bacterium Butyrivibrio fibrisolvens, based on Tn916-mediated transfer of the
Staphylococcus aureus plasmid pUB110. Plasmid 32:295?305.
Clark R.G., Cheng K.-J., Selinger L.B. and Hynes M.F. 1995. Expression of a Streptococcus
bovis amylase in the rumen bacterium Butyrivibrio fibrisolvens. Proceedings of the 95th
Annual Meeting of the American Society for Microbiology, Washington DC, USA (in press).
Clark R.G., Hu Y.J., Hynes M.F., Salmon R.K. and Cheng K.J. 1992. Cloning and expression
of an amylase gene from Streptococcus bovis in Escherichia coli. Archives of Microbiology
157:201?204.
Classen H.L., Campbell G.L. and GrootWassink J.W.D. 1988. Improved feeding value of
Saskatchewan-grown barley for broiler chickens with dietary enzyme supplementation.
Canadian Journal of Animal Science 68:1253?1259.
Classen H.L., Graham H., Inborr J. and Bedford M.R. 1991. Growing interest in feed enzymes
to lead to new products. Feedstuffs 28:22?34.
Cocconcelli P.S., Ferrari E., Rossi F. and Bottazzi V. 1992. Plasmid transformation of Ruminococcus
albus by means of high-voltage electroporation. FEMS Microbiology Letters 94:203?207.
Edney M.J., Campbell G.L. and Classen H.L. 1989. The effect of ?-glucanase supplementation
on nutrient digestibility and growth in broilers given diets containing barley, oat groats or
wheat. Animal Feed Science and Technology 25:193?200.
Erdman J.W. 1979. Oilseed phytates: Nutritional implications. Journal of the American Oil
Chemists?Society 56:736?741.
Feng P., Hunt C.W., Julian W.E., Dickinson K. and Moen T. 1992a. Effect of enzyme additives
on in situ and in vitro degradation of mature cool-season grass forage. Journal of Animal
Science (Supplement 1) 70:309.
Feng P., Hunt C.W., Julian W.E., Haenny S.C. and Pritchard G.T. 1992b. Effect of enzyme
additives to cool-season grass forage on voluntary intake and digestive function in mature
beef steers. Journal of Animal Science (Supplement 1) 70:310.
Flint H.J. 1994. Molecular genetics of obligate anaerobes from the rumen. FEMS Microbiology
Letters 121:259?268.
Flint H.J., Thomson A.M. and Bisset J. 1988. Plasmid associated transfer of tetracycline
resistance in Bacteroides ruminicola. Applied and Environmental Microbiology 54:855?860.
Flint H.J., Zhang J.-X. and Martin J. 1994. Multiplicity and expression of xylanases in the rumen
cellulolytic bacterium Ruminococcus flavefaciens. Current Microbiology 29:139?143.
Flores D.A. 1989. Application of recombinant DNA to rumen microbes for the improvement of
low quality feed utilization. Journal of Biotechnology 10:95?112.
Forsberg C.W. and Cheng  K.J. 1992. Molecular strategies to optimize forage  and cereal
digestion by ruminants. In: Bills D. and Kung S.B. (eds), Biotechnology and Nutrition.
Butterworth-Heinemann, Stoneham, MA, USA. pp. 109?147.
Forsberg C.W. Cheng K.-J., Krell P.J. and Phillips J.P. 1993. Establishment of rumen microbial
gene pools and their manipulation to  benefit fibre  digestion by domestic animals. In:
Proceedings of the VII World Conference on Animal Production, Edmonton, Alberta, Canada.
Vol. 1. University of Alberta, Edmonton, Canada. pp. 281?316.
Forsberg C.W., Crosby B. and Thomas D.Y. 1986. Potential for manipulation of the rumen
fermentation through the use of recombinant DNA techniques. Journal of Animal Science
63:310?325.
Exploitation of rumen microbial enzymes to benefit animal production 143
Fredeen A.H. and McQueen R.E. 1993. Effect of enzyme additives on quality of alfalfa/grass
silage and dairy cow performance. Canadian Journal of Animal Science 73:581?591.
Fuller R. 1989. Probiotics in man and animals. Journal of Applied Bacteriology 66:365?378.
Gilbert H.J., Hazlewood G.P., Laurie J.I., Orpin C.J. and Xue G.-P. 1992. Homologous catalytic
domains in a rumen fungus xylanase: evidence for gene duplication and prokaryotic origin.
Molecular Microbiology 6:2065?2072.
Gilliland S.E. 1990. Health and nutritional benefits from lactic acid bacteria. FEMS Microbiol-
ogy Reviews 87:175?188.
Gong J., Lo R.Y.C. and Forsberg C.W. 1989. Molecular cloning and expression in Escherichia
coli of a cellodextrinase gene from Bacteroides succinogenes S85. Applied and Environ-
mental Microbiology 55:132?136.
Gregg K. and Sharpe H. 1991. Enhancement of rumen microbial detoxification by gene transfer.
In: Tsuda T., Sasaki Y. and Kawashima R. (eds), Physiological Aspects of Digestion and
Metabolism in Ruminants. Proceedings of the Seventh International Symposium on Ruminant
Physiology, Sendai, Japan, 28 August?1 September 1989. Academic Press, San Diego,
California, USA. pp. 719?735.
Gregg K., Cooper C.L., Schafer D.J., Sharpe  H., Beard C.E., Allen G. and Xu J. 1994.
Detoxification of the plant toxin fluoroacetate by a genetically modified rumen bacterium.
Bio/Technology 12:1361?1365.
Grepinet O., Chebrou M.-C. and Beguin P. 1988. Nucleotide sequence and deletion analysis of
the xylanase gene (xynZ) of Clostridium thermocellum. Journal of Bacteriology
170:4582?4588.
GrootWassink J.W.D., Campbell G.L. and Classen H.L. 1989. Fractionation of crude pen-
tosanase (arabinoxylanase) for improvement of the nutritional value of rye diets for broiler
chickens. Journal of the Science of Food and Agriculture 46:289?300.
Hall J., Ali S., Surani A., Hazlewood G.P., Clark A.J., Simons J.P., Hirst B.H. and Gilbert H.J.
1993. Manipulation of the repertoire of digestive enzymes secreted into the gastrointestinal
tract of transgenic mice. Bio/Technology 11:376?379.
Harrison G.A., Hemken R.W., Dawson K.A., Harmon R.J. and Barker K.B. 1988. Influence of
addition of yeast culture supplement to diets of lactating cows on ruminal fermentation and
microbial populations. Journal of Dairy Science 71:2967?2975.
van Hartingsveldt W., van Zeijl C.M.J., Harteveld M.G., Gouka R.J., Suykerbuyk M.E.G.,
Luiten R.G.M., van Paridon P.A., Selten G.C.M., Veenstra A.E., van Gorcom R.F.M. and van
Den Hondel C.A.M.J. 1993. Cloning, characterization and overexpression of the phytase
gene (phyA) of Aspergillus niger. Gene 127:87?94.
Herbers K., Wilke I. and Sonnewald U. 1995. A thermostable xylanase from Clostridium
thermocellum expressed at high levels in the apoplast of transgenic tobacco has no detrimen-
tal effects and is easily purified. Bio/Technology 13:63?66.
Hespell R.B. 1989. Problems and progress with the genetics of ruminal bacteria. Developments
in Industrial Microbiology 30:13?33.
Hespell R.B. and Whitehead T.R. 1991a. Introduction of Tn916 and pAM?1intoStreptococcus
bovis JB1 by conjugation. Applied and Environmental Microbiology 57:2710?2713.
Hespell R.B. and Whitehead T.R. 1991b. Conjugal transfer of Tn916, Tn916?E, and pAM?1
from Enterococcus faecalis to Butyrivibrio fibrisolvens strains. Applied and Environmental
Microbiology 57:2703?2709.
Hesselman K. and Aman P. 1986. Effect of ?-glucanase on the utilization of starch and nitrogen
by broiler chickens fed on barley of low- or high-viscosity. Animal Feed Science and
Technology 15:83?93.
144 L.B. Selinger, K.-J. Cheng, L.J. Yanke, H.D. Bae, T.A. McAllister and C.W. Forsberg
Hodgson J. 1994. The changing bulk biocatalyst market. Bio/Technology 12:789?790.
Hols P.,  Ferain T.,  Garmyn D.,  Bernard N.  and Delcour  J.  1994.  Use of homologous
expression-secretion signals and vector-free stable chromosome integration in engineer-
ing of Lactobacillus plantarum for ?-amylase and levanase expression. Applied and
Environmental Microbiology 60:1401?1413.
Inborr J. 1990. Catalysts for pig performance. Feed Management 41:22?30.
Jacobs J.L. and McAllan A.B. 1991. Enzymes as silage additives. 1. Silage quality, diges-
tion, digestibility and performance in growing cattle. Grass and Forage Science
46:63?73.
Jacobs J.L. and McAllan A.B. 1992. Protein supplementation of formic acid- and enzyme-
treated silages. 1. Digestibilities, organic matter and fibre digestion. Grass and Forage Science
47:103?113.
Jacobs J.L., Haines M.J. and McAllan A.B. 1992. The effects of different protein supplements
on the utilization of untreated, formic acid-treated or enzyme-treated silage by growing
steers. Grass and Forage Science 47:121?127.
Johnson R., Williams P. and Campbell R. 1993. Use of enzymes in pig production. In: Wenk C.
and Boessinger M. (eds), Enzymes in Animal Nutrition. Proceedings of the 1st Symposium,
Kartause Ittingen, Switzerland. Institute for Domestic Animal Sciences, Nutrition Group,
Zurich, Switerland. pp. 49?60.
Jones F.T. 1991. Use of direct-fed microbials not new; way they work still not known. Feedstuffs
28:31?34.
Jones R.J. and Megarrity R.G. 1986. Successful transfer of DHP degrading bacteria from
Hawaiian goats to Australian ruminants to overcome the toxicity of Leucaena. Australian
Veterinary Journal 63:259?262.
Judkins M.B. and Stobart R.H. 1988. Influence of two levels of enzyme preparation on ruminal
fermentation, particulate and fluid passage and cell wall digestion in wether lambs consum-
ing either a 10% or 25% grain diet. JournalofAnimalScience66:1010?1015.
Juven B.J., Meinersman R.J. and Stern N.J. 1991. Antagonistic effects of lactobacilli and
pediococci to control intestinal colonization by human enteropathogens in live poultry.
Journal of Applied Bacteriology 70:95?103.
Ketaren P.P., Batterhan E.S., Belinda Dettmann E. and Farrell D.J. 1993. Phosphorus studies in
pigs. 3. Effect of phytase supplementation on the digestibility and availability of phosphorus
in soya-bean meal for grower pigs. British Journal of Nutrition 70:289?311.
Kopecny J. and Fliegerova K. 1994. Transformation of Selenomonas ruminantium by electro-
poration. Proceedings of the Society of Nutrition Physiology 3:150.
Lei X., Ku P.K., Miller E.R., Ullrey D.E. and Yokoyama M.T. 1993. Supplemental microbial
phytase improves bioavailability of dietary zinc to weanling pigs. Journal of Nutrition
123:1117?1123.
Malburg L.M. Jr. and Forsberg C.W. 1993. Fibrobacter succinogenes S85 possesses at least nine
distinct glucanase genes. Canadian Journal of Microbiology 39:882?891.
Malburg L.M. Jr., Smith D.C., Schellhorn H.E. and Forsberg C.W. 1993. Fibrobacter succino-
genes S85 has multiple xylanase genes. Journal of Applied Bacteriology 75:564?573.
Marquardt R.R., Fengler A.I. and Pawlik J. 1987. Improvement of the nutritional value of rye
and wheat grains through the use of crude enzymes of microbial origin. In: Lyons T.P. (ed),
Biotechnology in the Feed Industry: Proceedings of Alltech?s Fifth Annual Symposium.
Alltech Technical Publications, Nicholasville, USA. pp. 241?256.
McAllister T.A., Bae H.D., Jones G.A. and Cheng K.-J. 1994. Microbial attachment and feed
digestion in the rumen. Journal of Animal Science 72:3004?3018.
Exploitation of rumen microbial enzymes to benefit animal production 145
McGavin M.J., Forsberg C.W., Crosby B., Bell A.W., Dignard D. and Thomas D.Y. 1989.
Structure of the cel-3 gene from Fibrobacter succinogenes S85 and characteristics of the
encoded gene product, endoglucanase 3. Journal of Bacteriology 171:5587?5595.
Mir Z. and Mir P.S. 1994. Effect of the addition of live yeast (Saccharomyces cerevisiae) on
growth and carcass quality of steers  fed  high-forage or high-grain  diets and  on  feed
digestibility and in situ degradability. Journal of Animal Science 72:537?545.
Mroz Z., Jongbloed A.W. and Kemme P.A. 1994. Apparent digestibility and retention of
nutrients bound to phytate complexes as influenced by  microbial phytase and feeding
regimen in pigs. Journal of Animal Science 72:126?132.
Oosting S.J., Selinger L.B., McAllister T.A. and Cheng K.-J. 1995.  Development of
genetically engineered silage inoculants: need for higher gene expression. In: Proceed-
ings of the Joint Meeting of the Western Section of the American Society of Animal
Science and the Western Branch of the Canadian Society of Animal Science, Lethbridge,
AB (in press).
Paradis F.W., Zhu H., Krell P.J., Phillips J.P. and Forsberg C.W. 1993. The xynC from Fibrobac-
ter succinogenes S85 codes for a xylanase with two similar catalytic domains. Journal of
Bacteriology 175:7666?7672.
Pen J. 1994. Application of industrial enzymes produced in transgenic plants. In: Proceedings
of the Symposium on Production of Recombinant Proteins in Plants. University of Leicester,
Leicester, England, UK. pp. 44?49.
Pen J., Verwoerd T.C., van Paridon P.A., Beudeker R.F., van den Elzen P.J.M., Geerse K., van
der Klis J.D., Versteegh H.A.J., van Ooyen A.J.J. and Hoekema A. 1993. Phytase-containing
transgenic seeds as a novel feed additive for improved phosphorus utilization. Bio/Technol-
ogy 11:811?814.
Perry T.W., Purkhiser E.D. and Beeson W.M. 1966. Effects of supplemental enzymes on nitrogen
balance, digestibility of energy and nutrients and on growth and feed efficiency of cattle.
Journal of Animal Science 25:760?764.
Pettersson D. and Aman P. 1989. Enzyme supplementation of a poultry diet containing rye and
wheat. British Journal of Nutrition 62:139?149.
Reymond P., Geourjon C., Roux B., Durnad R. and Fevre M. 1992. Sequence of the phos-
phoenolpyruvate carboxykinase-encoding cDNA from the rumen anaerobic fungus Neocal-
limastix frontalis: comparison of the amino acid sequence with animals and yeast. Gene
110:57?62.
Rixon J.E., Hazlewood G.P. and Gilbert H.J. 1990. Integration of an unstable plasmid into the
chromosome of Lactobacillus plantarum. FEMS Microbiology Letters 71:105?110.
van Rooijen G.J.H. 1993. Molecular biology of oil body proteins in the Brassicaceae: Structure,
function and biotechnological applications. PhD dissertation. University of Calgary, Cal-
gary, Alberta, Canada. 241 pp.
van Rooijen G.J.H. and Moloney M.M. 1994. Plant seed oil-bodies as carriers for foreign
proteins. Bio/Technology 13:72?77.
Rotter B.A., Friesen O.D., Guenter W. and Marquardt R.R. 1990. Influence of enzyme supple-
mentation on bioavailable energy of barley. Poultry Science 69:1174?1181.
Rotter B.A., Marquardt R.R. and Guenter W. 1987. Optimising responses from enzyme in
poultry and pig diets: New methods for ensuring response. In: Lyons T.P. (ed), Biotechnology
in the Feed Industry: Proceedings of Alltech?s Fifth Annual Symposium. Alltech Technical
Publications, Nicholasville, USA. pp. 149?159.
Rovics J.J. and Ely C.M. 1962. Response of beef cattle to enzyme supplements. Journal of
Animal Science 21:1012.
146 L.B. Selinger, K.-J. Cheng, L.J. Yanke, H.D. Bae, T.A. McAllister and C.W. Forsberg
Russell D. 1994. Production of industrial and biopharmaceutical proteins in transgenic plants.
In: Proceedings of the Symposium  on  Production  of  Recombinant Proteins in  Plants.
University of Leicester, Leicester, England, UK. p. 43.
Russell J.B. and Wilson D.B. 1988. Potential opportunities and problems for genetically altered
rumen microorganisms. Journal of Nutrition 118:271?279.
Rust J.W., Jacobson N.L., McGilliard A.D. and Hotchkiss D.K. 1965. Supplementation of dairy
calf diets with enzymes. II. Effect on nutrient utilization and on composition of rumen fluid.
JournalofAnimalScience24:156?160.
Scheirlinck T., DeMeutter J., Arnaut G., Joos H., Claeyssens M. and Michiels F. 1990. Cloning
and expression of cellulase and xylanase genes in Lactobacillus plantarum. Applied Micro-
biology and Biotechnology 33:534?541.
Scheirlinck T., Mahillon J., Joos H., Dhaese P. and Michiels F. 1989. Integration and expression
of ?-amylase and endoglucanase genes in the Lactobacillus plantarum chromosome. Applied
and Environmental Microbiology 55:2130?2137.
Scott K.P. and Flint H.J. 1995. Transfer of plasmids between strains of Escherichia coli under
rumen conditions. Journal of Applied Bacteriology 78:189?193.
Selinger L.B., McAllister T.A., Forsberg C.W., Bae H.D., Yanke L.J. and Cheng K.-J. 1994.
Fibrolytic and amylolytic activities of ruminal fungi. In: Abstracts of the 5th International
Mycological Congress. Vancouver, British Columbia, Canada. International Mycological
Association, Baton Rouge, Louisiana, USA. p. 196.
Selmer-Olsen I., Henderson A.R., Robertson S. and McGinn R. 1993. Cell wall degrading
enzymes for silage. 1. The fermentation of enzyme-treated ryegrass in laboratory silos. Grass
and Forage Science 48:45?54.
Sharp R., Hazlewood G.P., Gilbert H.G. and O?Donnell A.G. 1994. Unmodified and recombi-
nant  strains of Lactobacillus plantarum are rapidly lost from  the  rumen by protozoal
predation. Journal of Applied Bacteriology 76:110?117.
Sharp R., O?Donnell A.G., Gilbert H.G. and Hazlewood G.P. 1992. Growth and survival of
genetically manipulated Lactobacillus plantarum in silage. Applied and Environmental
Microbiology 58:2517?2522.
Shoemaker N.B., Anderson K.L., Smithson S.L., Wang G.-R. and Salyers A.A. 1991. Conjugal
transfer of a shuttle vector from the human colonic anaerobe Bacteroides uniformis to the
ruminal anaerobe Prevotella (Bacteroides) ruminicola B
1
4. Applied and Environmental
Microbiology 57:2114?2120.
Simons P.C.M., Versteegh H.A.J., Jongbloed A.W., Kemme P.A., Slump P., Bos K.D., Wolters
M.G.E., Beudeker R.F. and Verschoor G.J. 1990. Improvement of phosphorus availability by
microbial phytase in broilers and pigs. British Journal of Nutrition 64:525?540.
Sipat A., Taylor K.A., Lo R.Y.C., Forsberg C.W. and Krell P.J. 1987. Molecular cloning of a
xylanase gene from Bacteroides succinogenes and its expression in Escherichia coli. Applied
and Environmental Microbiology 53:477?481.
Slominski B.A. and Campbell L.D. 1990. Non-starch polysaccharides of canola meal: quanti-
fication, digestibility in poultry and potential benefit of dietary enzyme supplementation.
Journal of the Science of Food and Agriculture 53:175?184.
Swick R. 1991. Phytase?a model for future feed additives? In: Proceedings of the 27th Annual
Nutrition Conference for Feed Manufacturers, Guelph, Ontario, Canada. University of
Guelph, Guelph, Ontario, Canada, pp. 107?122.
Tamblyn Lee J.M., Hu Y., Zhu H., Cheng K.-J., Krell P.J. and Forsberg C.W. 1993. Cloning of
axylanasegenefromtheruminalfungusNeocallimastix patriciarum and its expression in
Escherichia coli. Canadian Journal of Microbiology 39:134?139.
Exploitation of rumen microbial enzymes to benefit animal production 147
Teather R.M. 1985. Application of gene manipulation to rumen microflora. Canadian Journal
of Animal Science 65:563?574.
Theurer B., Woods W. and Burroughs W. 1963. Influence of enzyme supplements in lamb
fattening rations. Journal of Animal Science 22:150?154.
Thomson A.M. and Flint H.J. 1989. Electroporation induced transformation of Bacteroides rumini-
cola and Bacteroides uniformis by plasmid DNA. FEMS Microbiology Letters 61:101?104.
Thomson A.M., Flint H.J., B?chet M., Martin J. and Dubourgier H.-C. 1992. A new Escherichia
coli: Bacteroides shuttle vector, pRRI207, based on the Bacteroides ruminicola plasmid
replicon pRRI2. Current Microbiology 24:49?54.
Trinci A.P.J., Davies D.R., Gull K., Lawrence M.I., Nielsen B.B., Rickers A. and Theodorou
M.K. 1994. Anaerobic fungi in herbivorous animals. Mycological Research 98:129?152.
van Vuuren A.M., Bergsma K., Frol-Kramer F. and van Beers J.A.C. 1989. Effects of addition
of cell wall degrading enzymes on the chemical composition and in sacco degradation of
grass silage. Grass and Forage Science 44:223?230.
Wallace R.J. 1994. Ruminal microbiology, biotechnology and ruminant nutrition: progress and
problems. JournalofAnimalScience72:2992?3003.
Ward K.A., Murray J.D. and Nancarrow C.D. 1989. The insertion of foreign DNA into animal cells.
In: Biotechnology for Livestock Production. Animal Production and Health Division, Food and
Agriculture Organization of the United Nations, Plenum Press, New York, USA. pp. 17?28.
Ward P.P. and Conneely O.M. 1993. Using biotechnology to improve enzyme yields: From DNA
to the marketplace. In: Wenk C. and Boessinger M. (eds), Enzymes in Animal Nutrition.
Proceedings of the 1st Symposium, Kartause Ittingen, Switzerland. Institute for Domestic
Animal Sciences, Nutrition Group, Zurich, Switerland. pp. 16?21.
Ware C.E., Bauchop T., Hudman J.F. and Gregg K. 1992. Cryptic plasmid pBf1 from Butyrivi-
brio fibrisolvens AR10: its use as a replicon for recombinant plasmids. Current Microbiology
24:193?197.
Whitehead T.R. 1992. Genetic transformation of the ruminal bacteria Butyrivibrio fibrisolvens
and Streptococcus bovis by electroporation. Letters in Applied Microbiology 15:186?189.
Whitehead T.R. and Hespell R.B. 1989. Cloning and expression in Escherichia coli of a xylanase
gene from Bacteroides ruminicola 23. Applied and Environmental Microbiology
55:893?896.
Whitehead T.R. and Hespell R.B. 1990. Heterologous expression of the Bacteroides ruminicola
xylanase gene in Bacteroides fragilis and Bacteroides uniformis. FEMS Microbiology Letters
66:61?66.
Whitehead T.R., Cotta M.A. and Hespell R.B. 1991. Introduction of the Bacteroides ruminicola
xylanase gene into the Bacteroides thetaiotaomicron chromosome for production of xylanase
activity. Applied and Environmental Microbiology 57:277?282.
Wubah D.A., Akin D.E. and Borneman W.S. 1993. Biology, fibre-degradation, and enzymology
of anaerobic zoosporic fungi. CRC Critical Reviews of Microbiology 19:99?115.
Xue G.-P., Gobius K.S. and Orpin C.G. 1992a. A novel polysaccharide hydrolase cDNA (celD)
from Neocallimastix patriciarum encoding three multi-functional catalytic domains with
high endoglucanase, cellobiohydrolase and xylanase activities. Journal of General Microbi-
ology 138:2397?2403.
Xue G.-P., Orpin C.G., Gobius K.S., Aylward J.H. and Simpson G.D. 1992b. Cloning and
expression of multiple cellulase cDNAs from the anaerobic rumen fungus Neocallimastix
patriciarum in Escherichia coli. Journal of General Microbiology 138:1413?1420.
148 L.B. Selinger, K.-J. Cheng, L.J. Yanke, H.D. Bae, T.A. McAllister and C.W. Forsberg
Zhou L., Xue G.-P., Orpin C.G., Black G.W., Gilbert H.J. and Hazlewood G.P. 1994. Intronless
celB from the anaerobic fungus Neocallimastix patriciarum encodes a modular family A
endoglucanase. Biochemical Journal 297:359?364.
Zyla K. and Koreleski J. 1993. In vitro and in vivo dephosphorylation of rapeseed meal by means
of phytate-degrading enzymes derived from Aspergillus niger. Journal of the Science of Food
and Agriculture 61:1?6.
Exploitation of rumen microbial enzymes to benefit animal production 149
Clostridium paraputrificum var. Ruminantium:
colonisation and degradation of shrimp
carapaces in vitro observed by scanning
electron microscopy
M.A. Cobos
1
and M.T. Yokoyama
2
1. Programa de Ganaderia, Colegio de Postgraduados, Montecillo, Edo. de M?xico, 56230
2. Department of Animal Science, Michigan State University, East Lansing MI 48824.
Abstract
Previously a spore-forming bacterium identified as Clostridium paraputri-
ficum var. ruminantium was isolated from the rumen of dairy cows. Compared
with other clostridial species in the rumen, this bacterium is unable to degrade
cellulose, but rapidly degrades chitin, shrimp carapace and crab shell waste
meal. The present study was undertaken to characterise the morphology of C.
paraputrificum var. ruminantium and to improve our understanding of the
process of shrimp carapace degradation by this bacterium using scanning
electron microscopy. Bacteria growing on shrimp carapace showed vegetative
cells 3 to 5 ?m in length with rounded ends. Cells with terminal spores were
larger, averaging 7.3 ?m in length. The spores were prominent between 1 to 2
?m in length and 0.7 to 1 ?m in width. Filamentous forms were observed only
when the substrate had been depleted. Incubation of C. paraputrificum var.
ruminantium with pieces of shrimp carapace was characterised by three major
events. First, colonisation and growth was only observed on the inner surface
of the shrimp carapace. This activity was observed within the first 3 h after
incubation and extended through 24 h of incubation. Between 24 and 72 h of
incubation, bacterial growth was characterised by the degradation of the
different multiple layers of chitin that comprise the shrimp carapace to finally
reach the outer surface. After 96 h of incubation, when the shrimp carapace had
been almost totally degraded, the bacteria entered a death phase and latency,
marked by the predominance of free spores and lysed bacteria. Since chitin
degradation is initiated only after bacterial attachment occurs, a study of factors
that influence this adhesion may result in more efficient use of chitinous waste
as a feedstuff for ruminants.
Introduction
Shellfish waste is the generic name used to describe the by-products generated by the
fishing industry after meat extraction from marine crustaceans such as crab, shrimp, lobster
and prawn, and freshwater crustaceans such as crawfish. Shellfish waste consists of shells,
viscera and unextracted meat. Most of the shellfish waste, which approximates 85% of
the fresh catch, is largely disposed of by its return to the environment either on landfill
sites or by direct return to the sea and presents an environmental pollution problem.
A potential use of large quantities of shellfish waste is as a feed for livestock. For
example, shrimp carapaces dry matter contains in excess of 37% crude protein (Husby
et al 1981). The main limitation of shrimp carapaces as a feedstuff is its chitin content
(21% of DM),  which  cannot be  efficiently digested by most productive animals.
However, ruminant animals, because of the symbiotic microorganisms living in their
rumen, are considered better adapted than non-ruminants to assimilate the nutrients
containedinshellfishwaste.
The potential of chitinous waste as a feed for ruminants relies on the capacity of some
ruminal microorganisms to degrade chitin (Brundage et al 1981, 1984; Laflamme 1988;
Velez et al 1991), but the ruminal microorganisms involved in chitin degradation have
never been isolated and identified.
Recently, a spore-forming bacterium was isolated from the rumen of dairy cows
(Cobos and Yokoyama 1993). The clostridial isolate was classified as C. paraputrificum,
with 99% confidence, using the RapID II ANA system. However, at least 50 different
strains of this Clostridium have been isolated from very diverse environments such as
from soil, marine sediments, human infant faeces, and porcine, avian and bovine faeces
(Cato et al 1986). Therefore, we proposed that the ruminal strain isolated be designated
C. paraputrificum var. ruminantium. The metabolic profile of the clostridial isolate
indicates that this bacteria is highly specialised for chitin utilisation. It is unable to
degrade cellulose, a common characteristic of other clostridial species isolated from the
rumen of cattle and sheep such as C. polysaccharolyticum (van Gylswyk et al 1980), C.
chartatabidum (Kelly et al 1987) and C. longisporum (Varel 1989).
Although the intimate association of the bacterium with chitin suggested that it
attaches to chitin before degradation, similar to the attachment of other bacteria to
cellulose and starch, little was known about the actual process.
Materials and methods
Organism and culture conditions
A chitinolytic bacterium previously isolated from the rumen of dairy cows, C. paraputri-
ficum var. ruminantium, was used in this study. The bacterium was grown at 39?C in chitin
(CH) medium (CH-medium components described in Table 1). A 48 h culture of the
bacterium growing in CH-medium was used as inoculum. The bacterial concentration in
the inoculum, as estimated by the roll tube technique, was about 10
10
cells per ml.
Shrimp carapace preparation for colonisation study
Shrimp carapace (SC) were used as substrate to observe bacterial colonisation rather than
crab shell meal or pure chitin (Sigma) because in initial attempts the last two substrates
152 M.A. Cobos and M.T. Yokoyama
did not provide a good enough contrast with the scanning electron microscope. The SC
were washed in tap water to separate residues of meat, cut into pieces of about 0.5 cm
2
,
and boiled in tap water for 10 min. Five small pieces of the boiled SC were placed into
eight Dacron bags (5 ? 12.5 cm) with 20?35 ?m pore size (Ankom, Fairport, NY). Also,
in order to observe for possible cellulose colonisation and/or degradation, one small piece
of Whatman paper No. 4 (0.5 ? 2 cm) was added to each bag. The Dacron bags were placed
into a 1 litre Erlenmeyer flask, containing 500 ml of anaerobic SC medium. The SC
medium components are given in Table 2, and was prepared anaerobically according to
the method of Hungate (1969) as modified by Bryant (1972).
The SC medium in the Dacron bags was inoculated with 5 ml of a 48 h culture of C.
paraputrificum var. ruminantium, then incubated at 39.7?C. The Dacron bags were
recovered aseptically and consecutively after 0, 3, 6, 12, 24, 48, 72 and 96 h of incubation.
The Dacron bags were previously found to be very useful for a fast and aseptic recovery
method for multiple samplings and that is the main reason for their use.
Scanning electron microscopy
At appropriate time intervals, Dacron bags were removed aseptically and the residual
shrimp carapace were recovered by gently rinsing with distilled water and immediately
fixed for 1 h in 2% glutaraldehyde in 0.1 M phosphate buffer (pH 7.0). After fixation
the samples were rinsed with 0.1 M phosphate buffer. Samples were then dehydrated
with a graduated ethanol series (25, 50, 75, 90 and 100% ethanol), and critical
point-dried with CO
2
. Specimens were mounted on aluminum stubs, sputter-coated and
examined with a JEOL JSM-35C scanning electron microscope at the Center for
Table 1. Composition of the chitin medium used for growing C. paraputificum var. ruminantium.
Chitin medium
Item (per 100 ml)
Deionised water 51.6 ml
Clarified ruminal liquid
?
30.0 ml
Sodium carbonate 8% solution 5.0 ml
Mineral solution 1
?
5.0 ml
Mineral solution 2
?
5.0 ml
Cystein-sulfide solution
?
2.0 ml
Resazurin 0.1% solution 0.1 ml
Trypticase-peptone (LLB)
?
0.2 g
Yeast extract (LLB)
?
0.1 g
Chitin (Sigma) 1.0 g
?
Fresh ruminal fluid previously filtered through a cheese cloth is centrifuged at 23,420 g for 15 min at 4?C, autoclaved
20 min at 15 psi-121?C.
?
Contains (per 1000 ml) K2HPO4, 6.0 g.
?
Contains (per 1000 ml) KH2PO4, 6.0 g; (NH4)2SO4, 6 g; NaCl, 12 g; MgSO4, 2.45 g; and CaCl? 2H2O, 1.6 g (Bryant
and Robinson 1962).
?
2.5 g L-cysteine (dissolved in 15 ml of 2N NaOH); Na2S.9H2O, 2.5 g and resazurin, 0.1 ml (1% sol. in H2O) in a final
volume of 100 ml heated until the resazurin indicator is colourless, and autoclaved.
Colonisation and degradation of shrimp carapaces 153
Electron Optics of Michigan State University, following the procedures described by
Klomparens et al (1986). The same procedures were followed with the pieces of
Whatman filter paper added to the same Dacron bag for each period.
Results
Figure 1 shows structural characteristics of the shrimp carapace (SC) as to its outer
surface, inner surface and lateral view, at time zero of inoculation. The outer surface
(Figure 1a) is an even surface with the eventual appearance of small appendages about
700 ?m in length which look like the rhizoid structure of a fungus. In contrast, the inner
surface (Figure 1b) is rough with honeycomb-like compartments present in the curved
parts of the exoskeleton. Between the outer and inner surfaces, there appear to be several
layers of chitin (Figure 1c), which together with the outer and inner surface layers make
up the shrimp carapace.
Figure 2 is a series of scanning electron micrographs representing the major events
which take place during the process of colonisation and degradation of SC by C.
paraputrificum var. ruminantium at different incubation periods. The results show that
the bacterium attaches onto the SC inner surface within the first 3 h of incubation (Figure
2a), while the outer surface does not show any evidence of bacterial colonisation (Figure
2b). After 6 h of incubation, colonisation at the inner surface has spread considerably
(Figure 2c). In contrast, the outer surface still remains uncolonised at the same time,
even after degradation of the external appendages is apparent (Figure 2d). Strong
evidence of carapace digestion was observed after 9 h of incubation, as indicated by the
presence of typical digestion grooves formed on the inner surface of the shrimp carapace
(Figure 2e). At the same time, colonisation and degradation was also present on the
different multiple layers of chitin at the exposed sides (Figure 2f). The same photo
Table 2. Composition of the SC anaerobic medium used for observing bacterial attachment to and degra-
dation of shrimp carapace.
Quantity
Item per 100 ml
Deionised water 52.6 ml
Clarified ruminal liquid 30.0 ml
Sodium carbonate, 8% solution 5.0 ml
Minerals solution 1
?
5.0 ml
Mineral solution 2
?
5.0 ml
Cysteine-sulfide solution
?
2.0 ml
Resazurin, 0.1% solution 0.1 ml
Trypticase-peptone (LBB
?
) 0.2 g
Yeast extract (LBB
?
) 0.1 g
?
Contains (per 1000 ml) K2HPO4, 6.0 g.
?
Contains (per 1000 ml) KH2PO4, 6 g; (NH4)2SO4, 6 g; NaCl, 12 g; MgSO4, 2.45 g and CaCl2? 2H2O, 1.6 g (Bryant and
Robinson 1962).
?
2.5 g L-cysteine (dissolved in 15 ml of 2N NaOH); Na2S.9H2O, 2.5 g and resazurin, 0.1 ml of a 1% solution in H2Oin
a final volume of 100 ml heated until the resazurin indicator is colourless, and autoclaved.
154 M.A. Cobos and M.T. Yokoyama
Figure 1.
Structural characteristics of the shrimp carapace at its (a) outer surface (bar, 10 ?m), (b) inner
surface (bar, 10 ?m) and (c) a lateral view showing the different chitin layers that comprise its structure
(bar, 5 ?m).
Colonisation and degradation of shrimp carapaces 155
Figure 2. Scanning electron micrographs of  shrimp  carapace incubated with C. paraputrificum var.
ruminantium. (a) Morphological details of vegetative and spore-forming forms attached to the inner surface
of the shrimp carapace after 3 h of incubation (bar, 2 ?m). (b) View of the outer surface at the same incubation
period (3 h) with no signs of bacterial growth on its surface (bar, 5 ?m). (c) Degree of colonisation after 6
h of incubation (bar, 10 ?m). (d) Outer surface of the shrimp carapace still uncolonised after 6 h of incubation
(bar, 10 ?m). (e) Digestive grooves formed on the inner surface of the shrimp carapace after 9 h of incubation
(bar, 5 ?m). (f) Strong colonisation at the different layers that comprise the shrimp carapace after 9 h of
incubation. Degradation of chitin is indicated by the presence of digestion pits that penetrate into the interior
of the chitin layers (bar, 5 ?m).
156 M.A. Cobos and M.T. Yokoyama
illustrates that the bacteria have formed digestion pits that penetrate into the interior of
the chitin layers.
After 24 h of incubation, the ruminal clostridium showed profuse growth and
colonisation on the inner surface of the shrimp carapace (Figure 3a). At the outer surface
near cracked areas, bacterial growth was also evident just behind the outer layer, while
the external surface remained unaltered (Figure 3b). Between 48 and 72 h of incubation,
the presence of digestion pits that penetrate into the interior of the chitin layers was
observed (Figure 3c). By 72 to 96 h of incubation, most of the shrimp carapace had been
digested with the exception of the thin outer layer, which looked undigested, but very
fragile, due to the lack of support (Figure 3d). By 96 h of incubation, most of the
substrate was depleted, and the bacteria entered a distinct phase marked by the appear-
ance of lysed bacteria and released spores indicating the senescence of the culture
(Figure 3e and f).
When most of the shrimp carapace appeared to be degraded after 4 days of incubation,
the culture was checked to see if the bacteria would switch their enzymatic system to
cellulose degradation as an alternative source of energy. However, the Whatman paper
added was neither colonised nor degraded even by the end of the incubation period (7 d).
Cellular morphology
Vegetative cells were straight or slightly curved rods about 0.3 ?mwideand3to5?m
long with rounded ends, and occurred singly or in pairs. Rods carrying spores were
larger than the vegetative cells averaging 7.3 ?m in length. Spores were oval, terminal
and caused the cells to swell. The spores were about 1 to 2 ?m long and 0.7 to 1 ?m
wide. Sporulation seemed to be triggered by either the lack of substrate for growth or
by the adverse build up of toxic metabolites in the medium. Also filamentous forms
were observed, mainly when the shrimp carapace had been degraded.
Discussion
The results of this study confirm the ability of C. paraputrificum var. ruminantium to
bind to chitin. This characteristic is not particularly surprising if we consider the strong
adhesive characteristic of chitin (Gooday 1990), and the necessity for this attachment.
The adhesion of C. paraputrificum var. ruminantium to chitin may be essential in
aqueous environments such as the rumen ecosystem, where the substrate must be
degraded to dimers or monomers before their efficient absorption can occur. Attachment
would optimise contact between the chitinolytic enzymes and the chitin molecules. In
addition, the N-acetylglucosamine and/or chitobiose produced from chitin degradation
would be more efficiently absorbed. Without attachment, there is also the possibility
that  the chitinolytic  enzymes  may be  degraded and assimilated  by  other ruminal
bacteria, or the N-acetylglucosamine released during chitin hydrolysis may be taken up
by competing bacteria.
The mechanisms of adhesion are not known, but Montgomery and Kirchman (1993)
reported that the specific attachment to chitin particles appears to be mediated by
Colonisation and degradation of shrimp carapaces 157
Figure 3. Scanning electron micrographs of shrimp carapace incubated with C. paraputrificum var.
ruminantium representing the major events that take place between 24 and 72 h of incubation. (a) Profuse
bacterial growth and colonisation of the inner surface was observed by 24 h of incubation (bar, 1?m). (b)
At the same period of time there is evidence of prodigal bacterial growth in cracked areas behind the outer
surface; the external surface remains undigested (bar, 10 ?m). (c) Between 48 to 72 h of incubation the
degree of digestion is evident by the presence of digestion pits in most of the inner surface of the shrimp
carapace (bar, 1 ?m). (d) By 72 to 96 h of incubation, most of the chitin layers have been digested with the
exception of the outermost layer (bar, 5 ?m). (e and f) represent the senescence of the culture as indicated
by the appearance of lysed bacteria and the high proportion of free spores (bar, 1 ?m).
158 M.A. Cobos and M.T. Yokoyama
chitin-binding proteins associated with the bacterial cell membrane. Another suggested
mechanism is that lateral and polar flagella are responsible for cell attachment to binding
sites on the chitin surface (Belas and Colwell 1982).
The external surface of the shrimp carapace seems to have a protective film of
unknown structural complexity which prevents bacterial attachment and is recalcitrant
to degradation. This protective film may play a role in the defense of shrimps against
bacterial or fungal infections. It is unknown if this protective film is also present in other
shellfish wastes, but it would seem advisable to grind shellfish waste when it is used as
a feed for ruminants, to increase the attachment surface and the efficiency of degrada-
tion.
Although not well documented, it is believed that the rate of chitin degradation is
enhanced by the adherence of chitinolytic microflora (Gooday 1990). Ou and Alexander
(1974) observed that separation of chitin and chitinolytic bacteria by layers of mi-
crobeads greatly reduced the rate of mineralisation of chitin particles by pure cultures
of a soil Pseudomonas species. This phenomenon has also been observed in C. thermo-
cellum, an anaerobic bacterium which degrades ?-crystalline cellulose after the cell
adheres very strongly to the insoluble substrate (Bayer and Lamed 1986).
Since bacterial attachment to chitin seems to be essential for chitin degradation to
occur, attention has been placed on possible plant components that inhibit or promote
bacterial binding and therefore the digestion of chitinous compounds. For example,
wheat germ agglutinin (WGA), a chitin-binding lectin from wheat embryos, exhibits
binding specificity toward chitin (Bloch and Burger 1974). WGA lectin and other
chitin-binding lectins isolated from stinging nettle rhizomes are potent inhibitors of
fungal growth. Chitin chains in the fungal cell wall appear to be cross-linked and the
hydrolysis of preformed chitin chains in the process of apical growth of fungal hyphae
is inhibited (Broekaert et al 1989).
It is unknown if chitin-binding lectins are regular components of forages used as
feedstuffs. Also, it is not known if lectins are easily degraded by ruminal proteolytic
bacteria. However, it is known that the binding of lectins to chitin blocks proteolytic
and ?-N-acetyl-glucosaminidase activity against lectins and chitin, respectively (Knorr
1991). Therefore, if lectins are present in the rumen of cattle fed diets rich in chitin
by-products like shellfish waste, they may have a negative effect on the digestion of
chitinous compounds and animal performance.
Not much is known about the factors which affect the animal response to the intake
of chitinous compounds, especially those factors which affect the efficiency of the
rumen chitinolytic microorganisms. However, the observation that bacterial attachment
is essential for chitin hydrolysis opens up a new research area concerning the possible
factors that affect bacterial attachment to chitin and its consequences in the rumen
ecosystem.
References
Bayer E.A. and Lamed R. 1986. Ultrastructure of the cell surface cellulosome of Clostridium
thermocellum and its interaction with cellulose. Journal of Bacteriology 167:828?836.
Colonisation and degradation of shrimp carapaces 159
Belas M.R. and Colwell R.R. 1982. Adsorption kinetics of laterally and polarly flagellated
Vibrio. Journal of Bacteriology 151:1568?1580.
Bloch R. and Burger M.M. 1974. Purification of wheat germ agglutinin using affinity
chromatography of chitin. Biochemical and Biophysical Research Communications
58:13?17.
Broekaert W.F., Parijs J.V., Leyns F., Joos H. and Peumans W.J. 1989. A chitin-binding lectin
from stinging nettle rhizomes with antifungal properties. Science 245:1100?1102.
Brundage A.L., Husby F.M., Beardsley G.L. and Burton V.L. 1981. King crab meal in concen-
trates for lactating cows. Journal of Dairy Science 64:433?440.
Brundage A.L., Husby F.M., Herlugson M.L., Simpson W.L. and Burton V.L. 1984. Accept-
ability of Tanner crab meal in concentrates for lactation. Journal of Dairy Science
67:1965?1970.
Bryant M.P. 1972. Commentary of the Hungate technique for culture of anaerobic bacteria.
American Journal of Clinical Nutrition 25:1324?1328.
Bryant M.P. and Robinson I.M. 1962. Some nutritional characteristics of predominant culturable
ruminal bacteria. Journal of Bacteriology 64:605?613.
Cato E.P., George W.L. and Fine Gold S.M. 1986. Endospore-forming gram-positive rods and
cocci. Genus Clostridium. In: Sneath P.H.A. (ed), Bergey?s Manual of Systematic Bacteriol-
ogy. Williams and Willkins, Baltimore, Maryland, USA. pp. 1141?1200.
Cobos M.A. and Yokoyama M.T. 1993. Isolation of a ruminal Clostridium spp. which degrades
purified chitin and shellfish waste. Conference On Rumen Function. 22nd Biennial Confer-
ence on Rumen Function 22:28.
Gooday G.W. 1990. Physiology of microbial degradation of chitin and chitosan. Biodegradation
1:177?190.
van Gylswyk N.O., Morris E.J. and Els H.J. 1980. Sporulation and cell wall structure of
Clostridium polysaccharolyticum comb.nov.(FormerlyFusobacterium polysaccharo-
lyticum). Journal of General Microbiology 121:491?493.
Hungate R.E. 1969. A roll tube method for cultivation of strict anaerobes. In: Norris J.R. and
Ribbons D.W. (ed), Methods in Microbiology. Academic Press Inc., New York, USA. pp.
117?132.
Husby F.M., Brundage A.L. and White R.L. 1981. Utilization of shellfish meals in domestic
feeds. In: Otwell W.S. (ed), Seefood Waste Management in the 1980s: Conference Proceed-
ings. Report 40. Florida Sea Grant College, Orlando, Florida, USA. pp. 230?260.
Kelly W.J., Asmundson R.V. and Hoperoft D.H. 1987. Isolation and characterization of a strictly
anaerobic, cellulolytic spore former: Clostridium chartatabidum sp. nov. Archives of Micro-
biology 147:169?173.
Klomparens  K.L., Flegler S.L. and Hooper G.R. 1986. Procedures for Transmission and
Scanning Electron Microscopy for Biological and Medical Science. Center for Electron
Optics, Michigan State University. pp. 85?106.
Knorr D. 1991. Recovery and utilization of chitin and chitosan in food processing waste
management. Food Technology 45:116?122.
Laflamme L.F. 1988. Utilization of crab meal fed to young beef cattle. Canadian Journal of
Animal Science 68:1237?1244.
Montgomery M.T. and Kirchman D.L. 1993. Role of chitin-binding proteins in the specific
attachment of the marine bacterium Vibrio harveyi to chitin. Applied and Environmental
Microbiology 59:373?379.
Ou L. and Alexander M. 1974. Effect of glass microbeads on the microbial degradation of chitin.
Soil Science 118:164?167.
160 M.A. Cobos and M.T. Yokoyama
Varel V.H. 1989. Reisolation and characterization of Clostridium longisporum, a ruminal
sporeforming cellulolytic anaerobe. Archives of Microbiology 152:209?214.
Velez S.A., Allen J.C., Keery C.M. and Adkinson R.W. 1991. Evaluation of crab and crawfish
waste meals as protein sources for growing dairy heifers. Journal of Dairy Science
74:234?242.
Colonisation and degradation of shrimp carapaces 161
Comparative abilities of feed intake
and microbial digestion in the forestomachs
of camelids and ruminants
J.P. Jouany
1
,J.P.Dulphy
1
and C. Kayouli
2
1. I.N.R.A., Station de Recherches sur la Nutrition des Herbivores, Centre de Recherches de Clermont
Ferrand-Theix, 63122 Saint Gen?s Champanelle, France
2. INAT, 43 Avenue Charles Nicolle, 1002 Tunis Belv?d?re, Tunisie
Abstract
There are large differences in anatomy and digestive physiology between the
forestomachs of camelids and ruminants. The most voluminous compartment
corresponding to the rumen is named compartment 1 in camelids. Its ventral
part is composed of a cranial sac and a caudal sac separated by a transverse
muscular pillar. Each sac is covered by a band of several rows of glandular
sacs made of musculomembranous cells. Compartment 2 is small and opens
widely onto compartment 1. Compartment 3 is pear-shaped and composed
of two distinct parts. Only the pyloric area secretes acid and enzymes. The
ingestion time, the length of the main meals and the number of meals are not
different between llamas and sheep. Llamas have fewer but longer ruminat-
ing periods and are more efficient masticators than sheep (2 g vs 1.3 g
DM/min). Rumination occurs mainly at night in camelids. The liquid phase
turnover is higher in dromedaries and llamas than in ruminants.
Camelids  have lower  water requirements and  excrete  less  water and
nitrogen in urine than ruminants. Both dromedaries and llamas have a high
capacity for recycling nitrogen. Bicarbonates and phosphates are extensively
secreted by the mucosa of the forestomachs and in the glandular sacs. This
explains the higher buffering capacity of the digesta in camelids. Also, these
secretions are probably responsible for the higher and more stable osmolality
of the digesta of llamas compared to ruminants. Ammonia concentration is
lower and more variable in compartment 1 than in the rumen. Total VFA
concentration is more difficult to characterise. Butyrate concentration is
frequently higher in camelids. The microbial ecosystems are basically the
same in camelids and ruminants, but their glycolytic activities are different.
The in sacco degradation of plant cell walls is higher in camelids than in
ruminants. The amount of digested cell wall carbohydrate is much larger in
camelids than in ruminants. The lower temperature of the digesta in the
forestomachs of camelids could indicate that microbes produce less extra
heat and that more ATP is available for microbial growth. These charac-
teristics of ingestive and digestive abilities of camelids show that their rumen
microbiota has a higher digestive activity.
Introduction
Ruminants and camelids both belong to the order Artiodactylae in the mammalian class.
The family Camelidae belongs to the suborder Tylopoda which appeared 50 million
years ago. The family Bovidae (cow, sheep, goat) in the suborder Ruminantia developed
more recently, only 3 million years ago (Romer 1966). The family Camelidae is
composed of 6 species: 2 of the genus Camelus, 1 of the genus Vicugna and3ofthe
genus Lama (llama, guanaco, alpaca).
Camelids, like ruminants, developed large forestomachs harbouring a dense and
complex microbial ecosystem which makes them efficient in the use of forages as a
main source of nutrients. However, there are large differences between camelids and
ruminants in the anatomical organisation of their forestomachs and the histology of their
mucosa which give them some specificities in the digestive abilities of their microbial
ecosystem. Our paper will compare feeding abilities and microbial digestive abilities
of camelids and ruminants. Dromedaries and South American camelids will be consid-
ered in these comparisons.
Anatomical and physiological differences in the
forestomachs of camelids and ruminants
The ruminant stomach is composed of four compartments (rumen, reticulum, omasum,
abomasum) located between two sphincters: the cardia at the end of the oesophagus and
the pylorus at the beginning of the duodenum. Following ingestion, solid feeds are
submitted to an intense microbial digestion in the first two compartments. Large
exchanges of digesta occur between rumen and reticulum. A sphincter located between
reticulum and omasum controls the retention time of feed particles in the main fermen-
tation chamber, the reticulo-rumen. The mucosa of the first three compartments is
keratinised and covered by papillae. The omasum contains numerous lamellae which
act as a filter for large particles and absorb the end products of microbial fermentations.
The camelid stomach is composed of three compartments. Compartment 1 could be
compared to the rumen while compartment 2 could be compared to the reticulum.
Compartment 3 is completely different from the omasum in its function and organisa-
tion. It is pear-shaped and internally covered by a mucosa which secretes hydrochloric
acid and enzymes only in the pyloric region.
In contrast to the rumen, which is composed of six separate sacs, compartment 1 is
made of a cranial and a caudal sac separated by a muscular transverse pillar. In contrast
to true ruminants, the epithelium of the camelid forestomach is devoid of papillae and
is made of a smooth stratified epithelium. Two rows of glove-shaped glandular sacs are
located on the wall of compartment 1; one row is found on the cranial sac and the other
on the caudal sac. They are composed of glove-shaped cavities opening onto the inside
of compartment 1. Each cavity has a muscular sphincter opening at its fundus. They
have been described as a place for large bicarbonate secretions and a subsequent
absorption of volatile fatty acids (VFA) as shown by Cummings et al (1972) and Luciano
et al (1979). Papillae are absent on the wall of the first two compartments, but glandular
164 J.P. Jouany, J.P. Dulphy and C. Kayouli
cells cover the entire ventral part of compartment 1, compartment 2 and the upper part
of compartment 3. These secretive cells produce a bicarbonate-rich mucus and could
be involved in VFA absorption as indicated previously for glandular sacs.
Differences in ingestive capacities
between camelids and ruminants
Richard (1989) reviewed the available data on the voluntary intake of dromedaries in
natural conditions. The mean consumption varied from 14?15 g DM/kg liveweight (LW)
for straw and poor roughages to 23?24 g DM/kg LW for good quality forages. Data
concerning straw are comparable to those noted for cattle in tropical areas, and slightly
higher than for cattle in Europe (12 g/kg DM with heifers).
Experiments on direct comparisons between animals show that the camelid intake was
lower than sheep intake (El-Shaer 1981; Gauthier-Pilters 1979; Gihad et al 1989; Kandil
1984; Shawket 1976). Kayouli et al (1995) found that mean intake for three diets was 12
g/kg LW for dromedaries and 25 g/kg for sheep. Gihad et al (1989) also noted that
dromedaries ingested 11.4 g DM of three different diets per kg LW while the mean
consumption by sheep was 17.9 g/kg. In a comparison carried out on four dromedaries,
four rams and two goats fed on oat-vesce hay, H. Rouissi, D.I. Demeyer, C. Kayouli and
van C.J. Nevel (unpublished data) showed that goats ate twice as much as dromedaries
(25 vs 13 g DM/kg LW respectively), while rams had an intermediary intake (Table 1).
Few data involving llamas are available. From 11 direct comparisons, we observed
that intake of dry forages, supplemented or not with concentrate, was 15.9 g DM/kg LW
(49.3 g/kg LW
0.75
) in llamas and 18.4 g DM/kg LW (51.7 g/kg LW
0.75
)insheep
(Cordesse et al 1992; Dulphy et al 1994a, 1994b; Lemosquet et al 1995; Warmington
et al 1989). When data are computed in a figure, it is noteworthy that llamas ingest less
good quality forage, on a metabolic liveweight basis, than sheep (Figure 1).
Comparisons of feeding activities
Ingestion and rumination times during the day are not different between camelids and
ruminants. According to Gauthier-Pilters (1979) and Khorchani et al (1992), it took
400?500 min/d for ingestion at pasture. At trough, Abdouli and Kraiem (1990) noted
that ruminating time was 500?600 min/d. Kaske et al (1989) observed that rumination
occurs principally during the night in camels. In a comparative trial, Lemosquet et al
(1995) noted that ingestion time was the same for llamas and sheep (368 vs 389 min/d,
respectively) while the former had a shorter ruminating time (468 vs 538 min/d). The
lengths of main meals were similar (160 min and 151 min for llamas and sheep
respectively) as well as the number of meals (7.0 and 7.9/d, respectively). However,
llamas had fewer (7.5 vs 12.1) but longer ruminating periods which took place mainly
during the night. Llamas masticated 2 g/min while sheep were less efficient (1.3 g/min).
Such behaviour could explain why there are more large particles (>1 mm) in the faeces
of llamas than in sheep (Warmington et al 1989). Also, Lechner-Doll et al (1991)
observed that the critical size for particles to move out of the rumen is 3 mm in llamas
Comparative abilities of feed intake and microbial digestion in camelids and ruminants 165
Table 1. Comparative ingestive capacities of adult camelids and ruminants.
Liveweight DM intake (g/kg LW) DM intake (g/kg LW
0.75
)
Authors Animals (kg) Straw Hay Straw Hay
Dulphy et al (1994a) Sheep 60 11.2 15.5 18.0 20.5 23.0 31 43 50 57 64
Llamas 100 13.0 14.5 15.5 16.5 17.5 41 46 49 52 55
Rouissi (1994) Dromedaries 220 13.0 11 ? ? ? 53 63 ? ? ?
Sheep 46 19.0 20 ? ? ? 49 51 ? ? ?
Goats 34 25.0 17 ? ? ? 60 44 ? ? ?
166
J
.
P
.
Jouany
,
J
.
P
.
D
ulph
y
a
n
d
C.
Kayouli
while it is 1 to 2 mm in sheep and cattle. This could be due to the absence of the filter
effect of omasal lamellae in llamas.
Turnover rates of digesta in forestomachs
Mean retention time of feed particles is longer in camelids than in ruminants (Lechner-
Doll et al 1991; Kayouli et al 1993a) and this could explain the lower ingestion capacity
of camelids. Such differences are not explained only by the differences in animal body
weights. Using llamas and sheep with similar LW, Dulphy et al (1994b) and Lemosquet
et al (1995) found a mean retention time of 44 h and 27 h respectively for solid particles
in the rumen.
These differences are mainly due to forestomach anatomy, to the motility of the
digestive tract and to the feeding behaviour of animals. For instance, the low ruminating
activity during the light part of the day will limit the comminution of solid particles and
maintain them in compartment 1 for a longer period of time in camelids. As described by
Figure 1. Comparisons of DMI (g/kg LW) in llamas and sheep (n = 11 comparisons; from Dulphy et al
1994a).
0 5 10 15 20 25 30 35 40
Sheep
0
5
10
15
20
25
30
35
40
Liamas
Comparative abilities of feed intake and microbial digestion in camelids and ruminants 167
Engelhardt and H?ller (1982), Heller et al (1984) and Engelhardt et al (1986a), forestom-
ach motility is different in dromedaries and llamas to that described in ruminants. There
are more type B contractions in camelids, which could explain the efficiency of the mixing
effect inside compartment 1 and the lower propulsive effect outside the compartment
(Gregory et al 1985). According to Engelhardt et al (1986b), such a motility is favourable
to a better separation of the liquid and solid parts of digesta. As a consequence, the turnover
of liquids and solutes is greater in camelids compared to ruminants (Kayouli et al 1993a).
From the results of Maloiy (1972), Farid et al (1979) and Kayouli et al (1993a), the mean
retention time of the liquid phase in the rumen is 6 h in dromedaries and 8 h in sheep. In
a direct comparison, Lemosquet et al (1995) noted that the liquid remained for 11 h in
compartment 1 of llamas vs 14 h in the rumen. The higher turnover rate of liquids and the
longer retention time of solids both explain the higher DM content of digesta in compart-
ment 1 of camelids as observed by Dardillat et al (1994a).
Farid et al (1979), Lechner-Doll et al (1990) and Kayouli et al (1993a) observed that
the weight of fresh rumen content is higher in dromedaries than in ruminants; it
represented on average 14.7 and 10.5% of the LW of dromedaries and sheep, respec-
tively, fed the same diets. However, according to Dulphy et al (1994b) and Lemosquet
et al (1995), the weight of ruminal contents of digesta was not different between llamas
and sheep and represented 18.7% of LW when fed inside, at trough.
Water intake, salivation and secretions in the forestomachs
Dromedaries, and llamas to a lesser extent, drink and urinate less than ruminants.
According to Gihad et al (1989), water intake per kg DM intake in dromedaries equals
65% that of sheep. On the same basis, Warmington et al (1989) also noted that llamas
ingest 57% of water intake by sheep.
Kay and Maloiy (1989) showed that dromedaries secrete larger amounts of saliva
than cattle and ovines. According to Engelhardt and H?ller (1982), camelid saliva is
richer in bicarbonates and phosphates, although Engelhardt et al (1984) considered, in
another paper, that the production and composition of saliva is not very different
between ruminants and camelids.
Camelids have a high capacity for recycling nitrogen through their forestomach
mucosa (Ali et al 1977; Emmanuel et al 1976; Engelhardt and H?ller 1982; Engelhardt
et al 1986b; Farid et al 1979), thus limiting urinary losses. The low rate of glomerular
filtration by kidneys saves both water and urea (Yagil 1985), and explains why urea
concentration can be lower in urine than in blood (Schmidt-Nielsen et al 1957; Wilson
1984). This originality makes the camelids more sensitive to the toxicity of diets with
a high content in soluble nitrogen. Similar observations were made on llamas (Hinderer
and Engelhardt 1975). Camelids therefore have a greater ability to recycle nitrogen in
their forestomachs than ruminants.
Bicarbonates and phosphates are secreted inside compartment 1 via the mucosa and
more intensively in the glandular sacs. This mechanism contributes efficiently to the
stabilisation of physico-chemical conditions in compartment 1 of camelids and maintains
the activity of microbes  in extreme feeding situations  (Jouany and  Kayouli 1989;
168 J.P. Jouany, J.P. Dulphy and C. Kayouli
Lemosquet et al 1995). Engelhardt and H?ller (1982) consider that the secretion of
bicarbonates greatly improves the absorptive capacity of VFA by mucosa of the camelid
forestomach. They estimated that the rate of absorption is two to three times higher in
camelids than in ruminants.
Physico-chemical conditions in the forestomach
Physico-chemical conditions are more stable in the forestomachs of camelids as indicated
by Kayouli et al (1991, 1993b), Dardillat et al (1994a), Smacchia et al (1995) and
Lemosquet et al (1995). Compared to ruminants, pH remained higher in dromedaries and
llamas after feeding as previously indicated by Maloiy (1972) and Vallenas and Stevens
(1971). The substitution of 25% of hay by barley had no effect on pH in llamas while it
decreased the pH in sheep (Lemosquet et al 1995), indicating a higher buffering capacity
of digesta in camelids. Even when fed a poor hay, pH remained higher and more stable in
dromedaries than in sheep or goats (H. Rouissi, D.I. Demeyer, C. Kayouli and van C.J.
Nevel, unpublished data; Figure 2). NH
3
-N concentration was lower and less variable in
dromedaries as confirmed by Farid et al (1979) and H. Rouissi, D.I. Demeyer, C. Kayouli
and van C.J. Nevel (unpublished data) (Figure 3). However, Lemosquet et al (1995) noted
that the concentration of NH
3
-N is higher in llamas than sheep at the end of the night,
before the morning meal. VFA concentrations are more difficult to characterise since they
are sometimes higher (Kayouli et al 1993a; H. Rouissi, D.I. Demeyer, C. Kayouli and van
C.J. Nevel, unpublished data) or lower (Farid et al 1979) in dromedaries when compared
to sheep. On a molar proportion basis, the amount of butyrate is higher in camelids than
in ruminants. This indicates that a shift in metabolic pathways occurs in camelids. As a
consequence, VFAs are more efficiently absorbed since rate of absorption is related to the
length of acid chains (R?mond et al 1995). According to Smacchia et al (1995), the
buffering capacity of digesta in the forestomachs of camelids disappears when digesta are
transferred into in vitro fermenters. It means that this capacity is a direct effect of animals.
Secretion of bicarbonate by the wall of compartment 1 and the subsequent higher rate of
VFA absorption explain the major part of this effect. The higher turnover rate of liquid
associated with digesta in camelids also contributes to the elimination of the end products
of fermentations. Such stable conditions are favourable to the growth of cellulolytic
bacteria and fungi which are both sensitive to high acidic conditions.
Because of the high recycling capacity of NH
3
-N through saliva and rumen wall,
camelids are able to supply rumen microbes with their needs in soluble nitrogen even
in low nitrogen feeding conditions. Cellulolytic bacteria which have high needs in
soluble nitrogen can thus maintain their activity when animals are fed poor roughage
diets with a low nitrogen content. This also means that camelids excrete small amounts
of nitrogen in urine which makes them very sensitive to overdosages of soluble nitrogen
in diets and limits their contribution to soil pollution.
Considering these results, it can be assumed that camelids are better adapted than
ruminants to using diets supplemented with energy concentrates. Such complementary
energy will be fermented in compartment 1 to give higher production of VFA available
for animals without any negative effect on cellulolytic bacteria and fungi. The subsequent
Comparative abilities of feed intake and microbial digestion in camelids and ruminants 169
increase in ATP production associated with higher NH
3
-N availability will serve protein
synthesis and therefore, supply camelids with more amino acids.
Lemosquet et al (1995) showed that osmolality of the fluid in the digesta of forestom-
achs was significantly higher in llamas (277 mOsm l
?1
) than in sheep (259 mOsm l
?1
)
fed three different diets. This result confirms the higher content of electrolytes due to
secretions in the digesta of camelids. Also, osmolality was more stable in camelids all
throughout the day.
Microbial ecosystems and their activities in the forestomach
Only a few differences could be shown by scientists in the microbial population found in
the forestomachs of camelids and ruminants. According to Williams (1963) and Ghosal et
al (1981), the dominant species of bacteria are the same and their numbers are not different
between camelids and ruminants. Several observations made by Jouany and Kayouli
(1989) and Kayouli et al (1991, 1993a) showed that the total population of protozoa tended
to be lower in llamas and dromedaries, while the distribution of the different genera is the
same. We found that entodinia represented 90% of the number of ciliates and that the large
Entodiniomorphida protozoa are only of type B in camelids (see the classification of Eadie
1962), whereas some animals harbour type A and others type B in ruminants. Also, the
fungal population seems to be higher in camelids (Fonty, personal communication).
Figure 2. Evolution of pH in the forestomachs of camelids and ruminants.
012345678
Time (hrs)
5.8
6
6.2
6.4
6.6
6.8
7
7.2
pH
Goats
Sheep
Dromedaries
170 J.P. Jouany, J.P. Dulphy and C. Kayouli
However, these slight differences in the microbial ecosystems could not explain the higher
cellulolytic activities noted in the microbial population of camelids.
Kayouli and Jouany (1990), Kayouli et al (1991, 1993a, 1993b) and Dardillat et al
(1994a) observed that the in sacco degradations of the cell wall part of several roughages
Figure 3. NH
3
-N concentrations in the forestomachs of camelids and ruminants?a) animals were fed
oat-vesce hay; b) animals were fed a mixed diet.
0 258
Times (hrs)
0
10
20
30
40
50
60
70
80
NH3-N(mg1-1)
Dromedaries
Goats
Sheep
a.
02 5 8
Times (hrs)
0
20
40
60
80
100
120
140
160
180
200
NH3 - N(mg1-1)
b.
Comparative abilities of feed intake and microbial digestion in camelids and ruminants 171
was higher in camelids after 24 h of retention time in the rumen than in ruminants. The
more indigestible the substrate, the higher the difference in favour of camelids (Kayouli
et al 1991). To confirm these results, Kayouli et al (1993b) and Dardillat et al (1994b)
showed that the indigestible in sacco residue of straw from the sheep was digested again
when introduced into compartment 1 of llamas. The rate of cell wall degradation was
not different between animals. Using the same technique, H. Rouissi, D.I. Demeyer, C.
Kayouli and van C.J. Nevel (unpublished data) showed that the percentage of ADF
degradation of oat-vesce hay at 72 h was not different in goats and dromedaries (37 and
39%, respectively); it was significantly higher in these animals than in sheep (27%). No
differences were noted in proteolytic  activity of microbes in the forestomachs of
camelids and ruminants.
Comparisons on digestibilities
and microbial protein synthesis
The data obtained by Farid et al (1979) and Gihad et al (1989) suggests that the
digestibility of the organic matter (dOM) of five different diets was on average 56.2%
in dromedaries vs 52.4% in sheep. Using data from Hintz et al (1973), Warmington et
al (1989), Cordesse et al (1992) and Lemosquet et al (1995), we showed that the dOM
of 11 diets was 58.5 in llamas vs 54.3% in sheep. Differences between llamas and sheep
(4 points) remain constant regardless of the nature of the diet (Figure 4). It is noteworthy
that differences are much higher when the digestibility of plant cell walls is considered,
8.7 points on average in favour of dromedaries from five comparisons and 7.1 points in
favour of llamas from eight comparisons. These results clearly indicate that camelids
are far more efficient than ruminants in the use of cell wall carbohydrates. This is
explained both by a higher cellulolytic activity of microbes due to the physico-chemical
conditions and to the higher retention time of feed particles in compartment 1 as
discussed before. Addition of 25% barley to a hay diet had no negative effect on the
cellulolytic activity in the forestomach of camelids (Lemosquet et al 1995). The absence
of disturbance of the microbial ecosystem could explain why llamas reduce hay intake
less than ruminants when additional concentrate is offered. This ability of camelids has
to be confirmed with higher starch additions.
Differences between camelids and ruminants in the digestibility of nitrogen are
negligible (Cordesse et al 1992; Farid et al 1985; Gihad et al 1989; Kayouli et al 1993a;
Maloiy 1972). Only Hintz et al (1973) found that the digestibility of nitrogen was 5
points higher in llamas than in sheep. There is no available information in the literature
on the efficiency of microbial protein synthesis, but there are some indications that
camelids show a greater efficiency. They recycle more nitrogen in the fermentative
chamber, they produce more VFA and make more ATP available for microbial growth,
pH is maintained in a range favourable for maximum growth, and the increase in the
turnover rate of liquids in the fermenter stimulates the growth rate (Harrison et al 1975).
The temperature in the fermenter is 2?C lower than in the rumen (Lemosquet et al 1995)
indicating smaller losses in extra heat by microbes and therefore a better use of energy
for growth.
172 J.P. Jouany, J.P. Dulphy and C. Kayouli
These comparisons show that camelids are able to use poor roughages better than
domesticated ruminants in temperate climates. They use energy supplements efficiently
without any negative effect on the digestion of the basic diet. However, some care has
to be taken when diets rich in soluble nitrogen are given.
References
Abdouli H. and Kraiem K. 1990. Intake, digestion and feeding behaviour of one-humped camel
stall-fed straw-based diets. Livestock Research for Rural Development 2:23?129.
Ali K.E.,  Hinderer S., Engelhardt W., Wipper E., Becker G. and Engelhardt W.V. 1977.
Harmstoff recycling bein eiweis-stoffwechsel des lamas. Umschan 77:338?339.
Cordesse R., Inseta M. and Gaubert J.L. 1992. Intake and digestibility of four forages by llamas
and sheep. Annales de Zootechnie 41:91?92.
Cummings J.F., Munnell F.F. and Vallenas A. 1972. The mucigenous glandular mucosa in the complex
stomach of two new-world camelids, the llama and guanaco. Journal of Morphology 137:71?110.
Dardillat C., Baumont R. and Jouany J.P. 1994a. Contenus gastriques du lama et du mouton.
Annales de Zootechnie 43:279.
Figure 4. Comparisons of dMO (%) in llamas and sheep (n = 11 comparisons).
40 45 50 55 60 65 70 75 80
dMO sheep
40
50
60
70
80
dMO Llamas
Comparative abilities of feed intake and microbial digestion in camelids and ruminants 173
Dardillat C., Dulphy J.P., Jouany J.P., Kayouli C. and Lemosquet S. 1994b. Comparison of in
sacco degradation and physico-chemical characteristics of gastric fermenter in camelids and
ruminants. Proceedings of the Society of Nutrition Physiology 3:321.
Dulphy J.P., Dardillat C., Jailler M. and Jouany J.P. 1994b. Comparison of the intake and
digestibility of different diets in llamas and sheep: a preliminary study. Annales de Zootechnie
43:379?387.
Dulphy J.P., Jouany J.P., Martin-Rosset W. and Theriez M. 1994a. Aptitudes compar?es de
diff?rentes esp?ces d?herbivores domestiques ? ing?rer et dig?rer des fourrages distribu?s ?
l?auge. Annales de Zootechnie 43:11?32.
Eadie J.M. 1962. Inter-relationships between certain rumen ciliate protozoa. Journal of General
Microbiology 29:579?588.
El-Shaer H.M. 1981. ?A comparative nutrition study of sheep and goats grazing southern Sina?
desert range with supplements.? PhD thesis, Faculty of Agriculture, Alexandria University,
Egypt.
Emmanuel B., Howard B.R. and Emady M. 1976. Urea degradation in the camel. Canadian
JournalofAnimalScience56:595?601.
Engelhardt W.V. and H?ller H. 1982. Salivary and gastric physiology of camelids. Verh. Deuts.
Zool. Gesel. pp. 195?204.
Engelhardt W.V., Lechner-Doll M., Heller R., Scharwtz H.J., Rutagwenda T. and Schultka W.
1986a. Physiology of the forestomachs in camelids with particular reference to adaptation
to extreme dietary conditions. A comparative approach. Zoologische Beitrage 30:1?15.
Engelhardt W.V., R?bsamen K. and Heller R. 1984. The Digestive physiology of camelids. In:
Cockrill W.R. (ed), The Camelid: An All-purpose Animal. Vol. I. Proceedings of the Khartoum
Workshop on Camels, December 1979. Scandinavian Institute of African Studies, Uppsala,
Sweden. pp. 323?346.
Engelhardt W.V., Weyreter H., Heller R., Lechner-Doll, Schwartz H.J., Rutagwenda R. and
Schultka W. 1986b. Adaptation of indigenous sheep, goats and camels in harsh grazing
conditions. In: Nuclear and Related Techniques for Improving Productivity of Indigenous
Animals in Harsh Environments. Proceedings of an Advisory Group Meeting on Improving
the Productivity of Indigenous Animals in Harsh Environments with the Aid of Nuclear
Techniques, Held in Ankara, 3?8 June 1985. IAEA (International Atomic Energy Agency),
Vienna, Austria. pp. 105?113.
Farid M.F.A., Shawket S.M. and Abdel Rahman M.H.A. 1979. The nutrition of camels and sheep
under stress. In: Cokrill W.R. (ed), The Camelid: An All-purpose Animal. Vol 1. Proceedings
of the Khartoum Workshop on Camels, December 1979. Scandinavian Institute of African
Studies, Uppsala, Sweden. pp. 293?322.
Farid M.F.A., Sooud A.O. and Hassan N.I. 1985. Effect of type of diet and level of protein intake
on feed utilization in camels and sheep. In: Proceedings of the 3rd AAAP Animal Science
Congress, 6?10 May 1985. Vol 2. pp. 781?783.
Gauthier-Pilters H. 1979. Aspects of dromedary ecology and ethology. In: Cokrill W.R. (ed), The
Camelid: An All-purpose Animal. Vol. 1. Proceedings of the Khartoum Workshop on Camels,
December 1979. Scandinavian Institute of African Studies, Uppsala, Sweden. pp. 293?322.
Ghosal A.K., Tanwar R.K. and Dwaraknath P.K. 1981. Note on rumen microorganisms and
fermentation pattern in camel. Indian Journal of Animal Sciences 51:1011?1012.
Gihad E.A., El Gallad T.T., Sooud A.E., Abdou E.L., Nasr H.M. and Farid M.F.A. 1989.
Feed and water intake digestibility and nitrogen utilization by camels compared to sheep
and goats fed low protein desert by-products. In: Options M?diterran?ennes, S?rie
S?minaires 2:75?81.
174 J.P. Jouany, J.P. Dulphy and C. Kayouli
Gregory P.C., Heller R. and Engelhardt W.V. 1985. Control of stomach motility in the llama
(Lama guanacoe F. glama). Quarterly Journal of Experimental Physiology 70:51?61.
Harrison D.G., Beever D.E., Thomson D.J. and Osbourn D.F. 1975. Manipulation of rumen
fermentation in sheep by increasing the rate of flow of water from the rumen. Journal of
Agricultural Science (Cambridge) 85:93?101.
Heller R., Gregory P.C. and Engelhardt W.V. 1984. Pattern of motility and flow of digesta in the
forestomachs of llama (Lama guanacoe F. glama). Journal of Comparative Physiology
154B:529?533.
Hinderer S. and Engelhardt W.V. 1975. Urea metabolism in the llama. Comparative Biochem-
istry and Physiology 52A:619?622.
Hintz H.F., Schryver H.F. and Halbert M. 1973. A note on the comparison of digestion by new
world camels, sheep and ponies. Animal Production 16:303?305.
Jouany J.P. and Kayouli C. 1989. La digestion microbienne chez les cam?lid?s. In: Options
M?diterran?ennes, S?rie S?minaires 2:89?96.
Kandil H.M. 1984. ?Studies on camel nutritionfs22 146. PhD thesis, Faculty of Agriculture,
Alexandria University, Egypt.
Kaske M., Osman T., Lechner-Doll M., Larsson M. and Engelhardt W.V. 1989. Circadian
changes of forestomach motility and rumination in camels. A.J.A.S. 2:301?302.
Kay R.N.B. and Maloiy G.M.O. 1989. Digestive secretions in Camels. Options M?diterran?en-
nes, S?rie S?minaires 2:83?87.
Kayouli C. and Jouany J.P. 1990. Comparison of the hydrolytic activity of microorganisms in
the forestomachs of dromedaries and sheep. Reproduction Nutrition Development (Supple-
ment 2):193S.
Kayouli C., Dardillat C. and Jouany J.P. 1993b. Comparative study of the muralytic activity of
rumen microbes measured in situ in llamas and sheep. Annales de Zootechnie 42:84.
Kayouli C., Jouany J.P. and Ben Amor J. 1991. Comparison of microbial activity in the
forestomachs of the dromedary and the sheep measured in vitro and in sacco on Mediterra-
nean roughages. Animal Feed Science and Technology 33:237?245.
Kayouli C., Jouany J.P., Dardillat C. and Tisserand J.L. 1995. Particularit?s physiologiques du
dromadaire: cons?quences sur son alimentation. S?minaire Dromadaire, Douz, Tunisie,
Octobre 1992. In: Options M?diterran?ennes (in press).
Kayouli C., Jouany J.P., Demeyer D.I., Ali Ali, Taoueb H. and Dardillat C. 1993a. Comparative
studies  on the degradation and mean retention time of solid and liquid phases in the
forestomachs of dromedaries and sheep fed on low-quality roughages from Tunisia. Animal
Feed Science and Technology 40:343?355.
Khorchani T., Abdouli H., Nefzaoui A., Neffati M. and Hamadi M. 1992. Nutrition of the
one-humped camel. II. Intake and feeding behaviour on arid ranges in southern Tunisia.
Animal Feed Science and Technology 39:303?311.
Lechner-Doll M, Rutagwenda T., Schwartz H.J., Schultka W. and Engelhardt W.V. 1990.
Seasonal changes of ingesta mean retention time and forestomach volume in indigenous
grazing camels, cattle, sheep, goats on a thornbush savannah pasture. Journal of Agricultural
Science (Cambridge) 115:409?420.
Lechner-Doll M., Kaske M. and Engelhardt W.V. 1991. Factors affecting the mean retention
time of particles in the forestomachs of ruminants and camelids. In: Tsuda T., Sasaki Y.,
Kawashima R. (eds), Physiological Aspects of Digestion and Metabolism in Ruminants.
Proceedings of the Seventh International Symposium on Ruminant Physiology, Sendai,
Japan, 28 August?1 September 1989. Academic Press, San Diego, California, USA. pp.
445?482.
Comparative abilities of feed intake and microbial digestion in camelids and ruminants 175
Lemosquet S., Dardillat C., Jailler M. and Dulphy J.P. 1995. Voluntary intake and gastric
digestion of two hays by llamas and sheep; influence of concentrate supply. Journal of
Agricultural Science (submitted).
Luciano L., Voss-Wermbter W.V., Behnke M., Engelhardt W.V, and Reale E. 1979. Die struktur
der mangescheimhaut beim lama (Lama Guanaco und Lama Lamae). I. Vorn?gen Gegenbaurs
Morphol. Jahrb. 125:519?549.
Maloiy G.M.O. 1972. Comparative studies on digestion and fermentation rate in the forestom-
ach of the one-humped camel and the zebu steer. Research in Veterinary Science 13:476?481.
R?mond D., Ortigues I. and Jouany J.P. 1995. Energy substrates for ruminal epithelium. In:
Metabolic Fuel Selection, Symposium Lavoisier, 13?16 Juillet 1994. CNAM, Paris, France
(in press).
Richard D. 1989. Ingestibilit? et digestibilit? des aliments par le dromadaire. Options M?diter-
ran?ennes, S?rie s?minaires 2:55?59.
Romer A.S. 1966. Vertebrate Paleontology. Third edition. University of Chicago, Chicago Press,
Chicago, USA.
Rouissi H. 1994. ?Etude Comparative de l?activit? Microbienne dans le Rumen des Dro-
madaires, des Ovins et des Caprins.? PhD thesis, Universiteit Gent, Belgique. 114 pp.
Schimdt-Nielsen B., Schmidt-Nielsen K., Houpt T.R. and Jarnum S.A. 1957. Urea secretion in
the camel. American Journal of Physiology 188:477?484.
Shawket S.M. 1976. ?Studies on Rumen Microorganisms.? MSc thesis, Faculty of Agriculture,
Alexandria University, Egypt.
Smacchia A.M., Dardillat C., Papon Y. and Jouany J.P. 1995. Comparative studies on the
microbial fermentations in the forestomachs of llamas and sheep. Annales de Zootechnie (in
press).
Vallenas A. and Stevens C.E. 1971. Volatile fatty acid concentrations and pH of llama and
guanaco digesta. Cornell Veterinarian 61:239?252.
Warmington B.G., Wilson G.F. and Barry T.N. 1989. Voluntary intake and digestion of rye grass
straw by llama ? guanaco crossbreds and sheep. Journal of Agricultural Science (Cambridge)
113:87?91.
Williams V.J. 1963. Rumen function in the camel. Nature 197:1221.
Wilson R.T. 1984. The Camel. Longman Group Ltd, Essex, UK. 223 pp.
Yagil R. 1985. The Desert Camel: Comparative Physiological Adaptation. Karger, Basel,
Switzerland. 159 pp.
176 J.P. Jouany, J.P. Dulphy and C. Kayouli
Optimising rumen environment
for cellulose digestion
E.R. ?rskov
Rowett Research Institute, Bucksburn, Aberdeen, Scotland, AB2 9SB
Abstract
The most common constraints in the rumen ecosystem which prevent fer-
mentation of cellulosic feeds from accruing as at maximal low rate are:
deficiencies of N and to a lesser extent S; number of bacteria in solution in
the rumen fluid to invade new substrate; antimicrobial factors; and low rumen
pH. The latter occurs almost exclusively with high concentrate feeding which
is seldom economical in less industrialised countries.
Deficiency of N is common and can be alleviated by adding a source of
NH
3
-yielding material, usually urea. This can be added in the form of
molasses blocks for rangeland and in other suitable packages for stabled
animals. The number of bacteria in solution can be enhanced by feeding small
amounts of easily digestible fibre where poor quality roughages are used.
Examples are brans, sugar beet pulp, grass and leaves from MPT. Antimicro-
bial factors contained in some MPT can in small quantities have a positive
effect on the rumen ecosystem, they can also be complexed by protein and
other phenolic binding agents. An optimal rumen environment generally
ensures maximal microbial yield and energy utilisation.
Introduction
Regardless of animal product, be it meat, milk, draught power or maintenance, as many
livestock  serve an important  function  for security and risk aversion  (?rskov and
Viglizzo 1994), it is important that the feed resources are utilised as efficiently as
possible. It can also be said that it is important that the potential nutritive value of
roughages are extracted. Since most of the ruminant livestock in developing countries
feed mainly on cellulosic crop residues, pasture, trees and bushes, it is important that
the fermentation process be unimpeded.
Effect of ammonia concentrations
In many crop residues, the NH
3
generated in the rumen from degraded protein is often
too low to ensure an efficient digestion process. In such cases addition of NH
3
-yielding
compounds can greatly enhance both intake and digestibility. Since straw ferments
slowly, it is important that the N supplement is also provided at a slow rate. If, for
instance, urea is used for pasture grazing it is common to include it in blocks for the
animals to lick. For stall-fed animals it is generally more convenient and cheaper to
spray the solution on the straw so that it is consumed at the same rate as straw. Apart
from poor utilisation of urea when it is given rapidly, it can also result in urea poisoning
in extreme cases. Urea is also provided from recycling and there is evidence that some
breeds of cattle recycle more urea to the rumen than others due to difference in renal
reabsorption.
The problem of inhibition of degradation is illustrated in Figure 1, which shows how
the same feed in an optimal environment (A) ferments faster than in a less than optimal
rumen environment (B). Where feeds are described by the exponential model p = a +
b(1 ?e
ct
) (?rskov and McDonald 1979), it can be clearly seen that the potential (a + b)
is the same. In both environments the asymptote will be reached but in the impeded
environment (B) the rate constant is less than in A. As a result, for instance, if a normal
rumen retention is 48 h there could be substantial differences in digestibility leading to
substantial reductions in feed intake.
Effect of number of microbes in solution
This phenomenon has only recently been given some prominence and is as yet poorly
understood. One observation that intrigued us and led to a greater understanding on use
Figure 1. Effect of rumen environment on degradation rate of the same feed. A represents rumen environ-
ment in which cellulolysis is maximal; B represents a rumen environment less than maximal. Note that the
intercepts and the asymptotes are similar.
Hours
4 8 16 24 48 72 96
0
10
20
30
40
50
60
A
B
Degradability
178 E.R. ?rskov
of supplements was a trial in which rumen-fistulated sheep were given ad libitum
untreated straw + urea or ammoniated straw (Silva and ?rskov 1988).
The urea was added to the untreated straw at feeding time in order to ensure that the
degradation rate was not impeded by rumen ammonia concentrations. Subsequently
untreated  straw and  ammoniated straw  was incubated  in  their  rumens. The  most
significant aspect was that in spite of optimal pH conditions and ammonia concen-
trations the degradation rate of untreated straw proceeded more rapidly in the rumens
of animals receiving ammoniated straw than in the rumens of animals eating untreated
straw. This is illustrated in Table 1 with the 48 h degradability. Subsequent to this
observation, Silva et al (1989) attempted to find supplements which when given to the
untreated straw gave a rumen environment similar to the ammoniated straw and found
that it was generally supplements containing a source of easily fermentable fibre which
elicited a response in degradation rate. Sugar beet pulp, for instance, was successful as
shown in Table 2. Inclusion of 15% of sugar beet in the diet had the effect of increasing
straw intake by about 25%. There was no increase in intake of ammoniated straw
indicating that the process of ammoniation itself made available a certain amount of
easily fermentable fibre so that no further improvement came about.
In the tropics and subtropics there is a great interest in understanding how best to utilise
multipurpose trees (MPT). MPT contain an excess of N for rumen microbes so that they
can be used as a source of N. They also contain easily fermentable fibre. An interesting
experiment was carried out by Pathirana et al (1992) when they fed untreated or urea-sup-
plemented rice straw to sheep and supplemented with different increments of Gliricidia
(Table 3). Gliricidia is a leguminous tree which grows well in many areas. It contains
Table 1. The effect of feeding untreated but urea-supplemented or ammoniated straw to fistulated animals
on degradation of untreated straw in two rumen environments.
Rumen NH
3
Degradation
Diet Rumen pH (mg/litre) at 48 h. (%)
Untreated straw + urea 6.9 268 45.3
Ammoniated straw 6.8 244 53.1
Table 2. Effect on straw intake when urea-supplemented, untreated or an ammonia-treated straw diet was
supplemented with sugar-beet pulp.
Sugar beet pulp Straw intake Total intake Live-weight gain
(g/kg straw) (g/d) (g/d) (g/d)
Untreated straw 0 414 414 ?65
Untreated straw 150 505 589 ?6
Ammoniated straw 0 729 729 1
Ammonia-treated straw 150 717 852 51
Silva et al (1989).
Optimising rumen environment for cellulose digestion 179
about 20% crude protein. Here it can be seen that the animals ate considerably more when
Gliricidia forage was added to unsupplemented than to urea-supplemented straw. It would
appear that the response to the untreated straw was mediated both by relieving an N
deficiency and by providing a source of easily fermentable fibre while the response to the
supplemented rice was mediated only by the source of fermentable fibre. This is of course
speculative and more work is needed to clarify these issues. If no supplements are available
it is possible to make some, for instance by ammoniation. In many parts of Africa crop
residues are grazed in situ and not gathered. If a small amount is gathered, ammoniated
and fed at night, it might benefit utilisation of grazed crop residues.
Manyuchi et al (1992) carried out a small trial to look at the effect of adding a small
amount of ammoniated straw to untreated but urea-supplemented straw. Here again the
results (Table 4) indicate that intake of untreated straw was considerably enhanced when
a supplement of treated straw was provided. It is apparent that there is a great deal to
learn about supplements and how they can best support utilisation of basal feeds as well
as being utilised as energy and protein sources in their own right.
Effect of low rumen pH
Problems in the rumen caused by low pH due to feeding of large quantities of grains or
starch-based diets, often aggravated by over-processing and intermittent feeding, is a
Table 3. Effect of supplementing a rice straw diet with Gliricidia leaves on straw intake with or without
supplement of urea.
Urea level Gliricidia levels Intake of straw Total intake
(%) (%) (g/d) (g/d)
0 0 574 574
0 5 673 708
0 15 848 998
0 30 721 1030
2 0 690 690
2 5 727 765
2 15 880 1035
2 30 715 1021
Pathirana et al (1992).
Table 4. Effect of supplementing untreated straw with ammoniated straw on total intake by sheep.
Intake Digestible
of untreated Intake of Total dry-matter
Treatment straw (g/d) NH
3
straw (g/d) intake (g/d) intake (g/d)
Untreated straw 641 0 641 324
Untreated straw + 200 g NH
3
straw 814 187 1001 452
Untreated straw + 400 g NH
3
straw 667 352 1019 563
NH
3
straw ? 1152 1152 673
180 E.R. ?rskov
problem generally associated with industrialised countries where intensively fed live-
stock are given large grain supplements. Low rumen pH not only in part but sometimes
totally abolishes cellulose fermentation; it also exposes the animals to problems of
acidosis, acetonaemia, laminitis and other feed-associated problems. Since this problem
does not occur with mainly roughage diets, which is by far the prevalent feed in
developing countries, it will not be discussed here in more detail.
Problems of anti-nutritive factors
Anti-nutritive factors are in effect plant defence mechanisms against attack by animals,
insects or microbes. Some of the anti-nutritive factors are targeted to host animals and
some are targeted to the microbial population in the rumen. Most of the microbial
anti-nutritive factors are phenol related or tannins.
A higher understanding of issues relating to microbial anti-nutritive factors is needed.
Many chemical determinants of tannins have proved inadequate to quantify the severity
of phenolic anti-nutritive factors. In some instances a high tannin content seems to be
tolerated while in others a small amount can cause a large depression in intake (see
Makkar et al 1989; Mueller-Harvey and Reed 1992). New biological methods making
use of the gas-production techniques with or without a tannin-complexing compound
such as polyvinylpyrrolidone or polyethylene glycol appear promising (Khazaal and
?rskov 1994) and need more development. There is also a need to identify accurately
the active compounds so that they can be targeted or eliminated by genetic manipulation
or other means.
Conclusion
An increase in understanding of rumen ecology has the potential to make a major impact
on feed utilisation in general and in particular in the parts of the world where cellulosic
roughages are likely to be the most important feed for ruminants and in areas where
higher quality supplements, e.g. brans and oilseed cakes, are scarce. It is also of
particular interest in areas where trees and bushes form important feed ingredients.
A better understanding of rumen ecology can lead to:
1. A better strategic use of scarce supplements so that they can be utilised not only as
supplements in their own right but also provide support for basal feeds.
2. Greater production of rumen microbial protein so that animals can grow better on
basal feeds alone without a need for protein supplements. A change in the ratio of
volatile fatty acids to microbial biomass has the further advantage that it will
increase capture of fermented energy and reduce methane production.
References
Khazaal K. and ?rskov E.R. 1994. The in vitro gas production technique: an investigation on its
potential use with insoluble polyvinylpolypyrrolidone for the assessment of phenolics-related
antinutritive factors in browse species. Animal Feed Science and Technology 47: 305?320.
Optimising rumen environment for cellulose digestion 181
Makkar H.P.S., Singh B. and Negi S.S. 1989.  Relationship of rumen degradability with
microbial colonisation, cell wall constituents and tannin levels in some tree leaves. Animal
Production 49:299?303.
Manyuchi B., ?rskov E.R. and Kay R.N.B. 1992. Effect of feeding small amounts of ammonia
treated straw on degradation rate and intake of untreated straw. Animal Feed Science and
Technology 38:293?304.
Mueller-Harvey I. and Reed J. 1992. Identification of phenolic compounds and their relation-
ships to in vitro digestibility of sorghum leaves from bird-resistant and non-bird-resistant
varieties. Journal of the Science of Food and Agriculture 60:179?196.
?rskov E.R. and McDonald I. 1979. The estimation of protein degradability in the rumen from
incubation measurements weighted according to rate of passage. Journal of Agricultural
Science (Cambridge) 92:499?503.
?rskov E.R. and Viglizzo E. 1994. The role of animals in spreading farmers risk. A new
paradigm for animal science. Outlook on Agriculture 23:81?89.
Pathirana K.K., Mangalika U.L.P. and Gunarante S.S.N. 1992. Straw Based Supplementation
in a Low Output System of Zebu Heifers. IAEA. TEC Doc-691. IAEA (International Atomic
Energy Agency), Vienna, Austria.
Silva A.T. and ?rskov E.R. 1988. Fibre degradation in the rumens of animals receiving hay,
untreated or ammonia treated straw. Animal Feed Science and Technology 19:277?287.
Silva Ayona T., Greenhalgh J.F.D. and ?rskov E.R. 1989. Influence of ammonia treatment and
supplementation on the intake, digestibility and weight gain of sheep and cattle on barley
straw diets. Animal Production 48:99?108.
182 E.R. ?rskov
Forage legumes as protein supplements
to poor quality diets in the semi-arid tropics
J.H. Topps
Department of Animal Science, University of Zimbabwe
P.O. Box MP167, Mount Pleasant, Harare, Zimbabwe
Introduction
In the semi-arid tropics, the natural pasture has a low nutritive value during the dry
season when the standing grass is mature and dormant. After the first two or three months
of the dry season, crop residues become available following the harvesting of grain, but
most of the residues are from cereal crops which have a nutritional value similar to that
of the mature natural grass. Residues from leguminous crops are also available but in
small quantities. They have a quality better than that of cereal residues and may be used
as protein supplements to the poorer quality residues and to the mature pasture. In
addition there is now increasing interest in, and use of legumes, either herbaceous or
shrub species, from which the foliage can be harvested and used as protein supplements.
When ruminants consume poor quality diets, food intake is low so the supply of nutrients
is insufficient to meet requirements for maintenance.
The main objective of providing a forage legume as a protein supplement is to
increase both the total food intake and that of the basal poor quality diet. Early work on
the use of forage legumes was mainly concerned with the need to improve the nitrogen
content of diets based on poor quality roughages in order to overcome a deficiency of
nitrogenous substrates for the rumen microorganisms. More recent evidence indicates
that other changes occur in this supplementation situation which result in an enhance-
ment of intake and digestibility of the diet. This review will assess the likelihood of
such changes and the effects they may have on animal production.
Food intake
In the experiments reviewed (Minson and Milford 1967; Siebert and Kennedy 1972;
Mosi and Butterworth 1985; Smith et al 1990; Kitalyi and Owen 1993), the forage
legumes were fed in discrete amounts and separately from the poor quality basal diets,
which were given ad libitum. In all experiments, the legume supplement increased total
food intake but the response varied widely between experiments. According to pub-
lished observations, the legume supplements were consumed without any refusals, so
the differences in dietary intake were mainly due to differences in the intake of the
basal diet.
In addition, there are large differences between experiments in the change in intake
of the poor quality forage in response to supplementation. In general, if the level of
supplementation was less than 30 to 40% there was an increase in the intake of the basal
diet, but in some experiments the intake was unchanged or slightly reduced. A very large
increase in intake of mature pangola grass (Digitaria decumbens) was found in the study
by Minson and Milford (1967). Above 40% forage legume, the intake of the basal diet
fell below that seen with the unsupplemented control diet, except for that observed by
Siebert and Kennedy (1972) when there was no change when lucerne contributed 45%
of the total dry matter intake. In contrast, Mosi and Butterworth (1985) found that the
intake of the basal diet on the supplemented treatments was always less than that of the
unsupplemented control diet. The reasons for such differences appear to be related to
differences in quality of the forage legume and of the basal diet and other factors.
Substitution of the basal diet
Before these factors are considered, it is worth examining the extent of substitution of
the basal diet by the forage legume and its practical implications. As indicated earlier,
substitution usually occurs when the legume supplement contributes at least 30 to 40%
of the total dry matter intake, and it is due to the bulk effect of the forage legume.
However, assessing substitution on a 1:1 basis is misleading since the two components
of the diet will have different digestibilities and different bulking effects in the reticulo-
rumen. Since the forage legume is likely to be more digestible, it will occupy less space
than the poor quality roughage. The poor quality roughage equivalent of the forage
legume can be calculated by multiplying the intake of the forage legume by the ratio of
the digestibility of the poor quality roughage given alone to the calculated digestibility
of the forage legume in the mixed diets. This calculated digestibility is obtained as the
intercept of the linear regression of diet digestibility on the proportion of the poor quality
roughage in the diet. Anticipated intake of the poor quality roughage can be obtained
by subtracting the poor quality roughage equivalent of the forage legume from that of
the poor quality roughage when not supplemented. Any change in intake is then
calculated as the difference between anticipated and observed intakes of the poor quality
roughage. According to the work of Manyuchi (1994), when this condition for differ-
ences in digestibility is made the observed is nearly always greater than the anticipated
intake. Although this derivation is informative and reassuring, in practice any substitu-
tion on a 1:1 basis should be avoided by keeping the inclusion rate of forage legume to
a third or less of the total dry matter. This recommendation has the merit of recognising
the nutritional quality of the forage legume, especially if it has been grown on a small
farm where the growing of the legume may be limited by the land and labour available.
Quality of the basal diet
If the basal diet has a low nitrogen content, the addition of a forage legume will increase
the nitrogen content of the total diet, which is likely to increase food intake and the rate
of degradation of the basal diet in the rumen. Such positive associative effects are well
known and at one time were thought to be either the main or only effect of providing
protein-rich supplement. Anumber of recent reports have shown that responses to forage
184 J.H. Topps
supplements are not entirely due to an increase in dietary nitrogen (Smith et al 1990;
Getachew et al 1994; Manyuchi 1994).
However, when different basal diets are compared any difference in quality, of which
the nitrogen content is a part, will be reflected in the size of the effect of the legume
supplement on intake. Mosi and Butterworth (1985) carried out a study in which
Trifolium tembense hay was used as a supplement to four different cereal straws, maize,
oat, teff and wheat. The straws differed appreciably in crude protein content (23 to 62
g/kg) and in acid detergent fibre (710 to 780 g/kg). It was found that the extent of
substitution of the basal diet tended to be greater with the apparently better quality
straws (oat and maize). Part of this difference was probably due to overcoming a
deficiency of nitrogen in the straws.
Quality of the forage supplement
Very few comparisons have been made between forage legumes that differed in their
nutritional characteristics. From the early work on the need for nitrogen supplemen-
tation (Elliott and Topps 1963) it may be deduced that forages with a high content of
rumen degradable nitrogen will elicit greater responses in food intake than those with
a low content. The results of a study by Smith et al (1990) support this hypothesis.
Three forage legumes, pigeon pea (Cajanus cajan), cowpea (Vigna unguiculata) and
lablab (Lablab purpureus) hay which differed in nitrogen content and in sacco DM
degradability were compared as supplements to maize stover. Cowpea, which had the
highest nitrogen content, promoted the greatest intake of maize stover. However, the
extent of substitution was inversely related to the 48 h degradability of the legumes
measured in sacco. In a similar study, Getachew et al (1994) compared three other
legumes (Desmodium intortum, Macrotyloma axillare and Stylosanthes guianensis)
as supplements to maize stover. S. guianensis had the highest nitrogen and lowest
neutral detergent fibre content and promoted the greatest intake of maize stover.
Recently in the author?s laboratory, Masama et al (1995) found differences between
four shrub legumes in the response produced as supplements. Although the nitrogen
content appears to be a dominating factor, it should be recognised that legumes with
a high nitrogen content are likely to contain less fibre and the total organic matter
may be more easily fermented. It may be that the energy available to the microorgan-
isms is a factor equal in significance to the supply of degradable nitrogen. In fact with
certain tropical legumes the presence of anti-nutritional factors may obviate any
causal relationship between nitrogen content and increase in food intake (Topps
1992).
Digestibility and rumen degradability
Forage legumes invariably increase the digestibility of the total diet which is some-
what expected. The digestibility would probably increase as the proportion of the
forage legume is increased, which reflects the contribution of the legume with its
higher digestibility. In most studies, the proportional change in digestibility is equal
Forage legumes as protein supplements to poor quality diets 185
to or less than the fractional increase of the high quality forage in the diet. This
indicates the absence of any positive associative effect of the forage legume on the
digestibility of the basal diet. Only a few studies have recorded a positive associative
effect of the forage legume e.g. Minson and Milford (1967) and Mosi and Butterworth
(1985).
If the presence of a forage legume increases the activity of rumen microorganisms, a
concomitant increase in degradation of fibre should be observed. Several reports have
described such an effect, but it may not occur with certain recalcitrant poor quality
roughages. McMeniman et al (1988) studied five legumes as supplements to rice straw
and observed that degradation of the straw was increased by the presence of the legume.
Similarly, Ndlovu and Buchanan-Smith (1985) found that lucerne increased the rate of
degradation of barley straw, brome grass and maize cobs. In contrast, Manyuchi (1994)
found that groundnut hay did not alter the in sacco degradation of poor quality grass
hay. It is likely that any change in the degradation of the basal diet as a result of an
increase in microbial activity may depend on the number of available sites for microbial
attachment (Akin 1989). With some roughages the cuticule layer and extent of lignifi-
cation are barriers to microbial colonization, so that an increase in rumen microbial
population may not be manifested in an increase in rate of degradation.
Rumen environment
Very few studies have been carried out in which changes in the rumen environment
have been measured when forage legumes are fed with poor quality basal diets.
Inevitably poor quality forages provide insufficient degradable nitrogen and ferment-
able energy to sustain optimum digestion of fibre. Forage legumes are relatively good
sources of degradable nitrogen and fermentable energy so their inclusion in the diet is
expected to increase the rumen population of cellulolytic microbes. Such an effect was
seen in a related study by Kolankaya et al (1985) when a higher growth of selected
cellulolytic microorganisms was observed with ammonia-treated compared with un-
treated straw.
Concentrations of rumen ammonia increase following supplementation with a forage
legume (McMeniman et al 1988; Getachew et al 1994; Manyuchi 1994), the increase
being a function of the degradability of the nitrogen in the forage legume. In a study by
Said and Tolera (1994), the legume with the lower nitrogen content (M. axillare) gave
higher rumen ammonia concentrations than D. intortum, which had more crude protein
but with a lower degradability. For certain forage legumes, especially certain species of
shrubs, the availability of the nitrogen compounds would be limited by tannins (Mangan
1988). Forage legumes increase the total concentration of volatile fatty acids (VFA)
without affecting the relative proportions and the rumen pH. These results indicate that
forage legumes are likely to maintain  a stable fermentation pattern.  Ndlovu and
Buchanan-Smith (1985) observed that the feeding of a lucerne supplement increased
the proportion of branched chain VFA and suggested that this increase may stimulate
the growth of cellulolytic microorganisms.
186 J.H. Topps
Kinetics of digesta
A potential increase in digestibility when forage legumes are added to poor quality basal
diets may be offset by a reduction of retention time of digesta. Such an effect was
recorded by Ndlovu and Buchanan-Smith (1985) when lucerne was fed with maize cobs,
and by Vanzant and Cochran (1994) when lucerne was fed at different levels with low
quality prairie forage. Similar results were obtained by Manyuchi (1994) (Table 1). The
addition of groundnut hay increased the fractional outflow rate of rumen solids without
altering the pool size of the rumen digesta. Manyuchi concluded that the increase in
food intake following supplementation with a forage legume was largely facilitated by
an increase in rate of passage of digesta. The mechanism by which this occurs is not
fully known. Changes in osmotic pressure of the rumen may be implicated but the results
of Manyuchi (1994) do not confirm the presence of a relationship.
Animal performance
Increases in food intake and in digestibility of the total diet arising from supplementation
with forage legume should be manifested in a significant increase in animal perform-
ance. There are very few reports of a marked improvement of production when tropical
legumes are fed. Work carried out in Kenya with leucaena leucocephala and Napier
grass (Bana grass) by Muinga et al (1992, 1993) has established that sizeable responses
in milk yield can occur (Tables 2 and 3). In both experiments there was a small increase
in the intake of Napier grass when leucaena was given as a supplement and a significant
increase (P < 0.05) in total DM intake. These enhanced intakes resulted in significant
increases (P < 0.05) in milk yields and some curtailment of weight losses. In the work
published in 1992 it is interesting to note that the response with the poorer quality was
greater than that with the higher quality Napier grass.
Table 1. Intake, digestibility and kinetics of rumen digesta in sheep fed on veld hay plus 1% urea alone
(Control) or veld hay supplemented with 100, 200 or 300 g of groundnut hay (100 Gnut, 200 Gnut, 300
Gnut) or groundnut hay alone.
100 200 300
Control Gnut Gnut Gnut Gnut SED
DM intake (g/d)
Groundnut hay 0 90 180 270 1656
Veld hay 912 896 905 785 0
Total 912 986 1085 1055 1656 58.0
DM digestibility 0.48 0.52 0.53 0.55 0.64 0.02
Rumen liquid
Volume (l) 12.48 13.14 12.60 12.54 11.70 1.68
Fractional outflow (%/h) 6.53 6.49 6.45 6.64 6.70 0.67
Rumen solids
Pool size (g) 2036 1763 2324 2272 2378 410
Fractional outflow (%/h) 2.59 2.90 3.14 2.98 3.79 0.24
Forage legumes as protein supplements to poor quality diets 187
Conclusions
Forage legumes can be used as a supplement to poor quality diets but the responses
produced are unpredictable. Much of the nature of the response relates to changes that
occur in the rumen environment. It is axiomatic that substantial improvement in the
rumen microbial activity especially that of the cellulolytic organisms will lead to greater
food intake and increases in digestibility of the diet. These changes in turn will improve
animal performance. A knowledge of the factors controlling the changes in rumen
microbial activity is essential.
In Zimbabwe, the work of Manyuchi (1994) has shown that groundnut hay is an
effective supplement when given with either Napier grass hay or poor quality grass hay.
Groundnut hay may therefore be a useful model for future studies. In addition, ground-
nuts are a dual purpose crop in that both the nuts and the forage are harvested, which is
attractive to farmers. To successfully encourage small farmers to grow forage legumes
such considerations are important. Acceptability of crops such as cowpeas, pigeon peas
and groundnuts means that the widespread establishment of these crops is more likely
than that of legumes which provide forage only. For this reason intensive studies are
needed on certain legumes to ascertain the best way of harvesting, storing and using the
forage as a supplement to poor quality roughages.
Table 2. Intake and milk production of Jersey cows given Napier fodder ad libitum without (Control) and
with (Treated) 5 kg of fresh leucaena per day.
Control Treated
Mean SE Mean SE
DM intake (kg/d)
Napier 5.20 0.27 5.70 0.24
Total 5.20 0.27 6.90 0.24
Milk yield (kg/d) 4.20 0.20 5.20 0.18
Table 3. Intake of dry matter and milk production (kg/d) of crossbred dairy cows given ad libitum Napier
grass cut at either 1.0 or 1.5 m height and supplemented with 0, 4 or 8 kg fresh leucaena per day from day
15 to 112 of lactation
Napier harvest height (m)
Leucaena 1.5 1.0
supplementation (kg) 0 4 8 0 4 8 SED
Intake
Napier 6.5 6.6 7.1 9.1 9.6 9.3 0.51
Total 6.5 7.8 9.3 9.1 10.7 11.5 0.51
Milk yield 6.1 6.9 7.8 8.5 8.5 8.8 0.55
188 J.H. Topps
References
Akin D.E. 1989. Histological and physical factors affecting digestibility of forages. Agronomy
Journal 81:17?25.
Elliott R.C. and Topps J.H. 1963. Voluntary intake of low protein diets by sheep. Animal
Production 5:269?276.
Getachew G., Said A.N. and Sundstol F. 1994. The effect of forage legume supplementation and
body weight gain by sheep fed a basal diet of maize stover. Animal Feed Science and
Technology 46:97?108.
Kitalyi A. and Owen E. 1993. Sorghum stover and lablab bean haulm [???] as foods for lactating
cattle in the agro-pastoral system of central Tanzania. In: Gill M., Owen E., Pollott G.E. and
Lawrence T.L.J. (eds), Animal Production in Developing  Countries. BSAP Occasional
Publication 16. BSAP (British Society of Animal Production), Edinburgh, UK. pp. 170?171.
Kolankaya N., Steward C.S., Duncan S.H., Cheng K.J. and Costerton J.W. 1985. The effect of
ammonia treatment on solubilization and growth of cellulolytic rumen bacteria. Journal of
Applied Bacteriology 58:371?379.
Mangan J.L. 1988. Nutritional effect of tannins in animal feeds. Nutritional Research Review
1:209?232.
Manyuchi B. 1994. High quality forages as supplements to low quality forages for ruminants.
Effect on  intake, digestibility and  rumen  digesta dynamics. PhD thesis, University of
Aberdeen, Aberdeen, UK.
Masama E., Topps J.H., Ngongoni N.T. and Maasdorp B.V. 1995. The differential effects of
supplementation with fodder from different tree legumes on feed intake and digestibility of
nutrients in sheep fed on maize stover (in press).
McMeniman N.P., Elliot R. and Ash A.J. 1988. Supplementation of rice straw with crop
by-products.  1. Legume straw  supplementation. Animal Feed  Science and  Technology
19:43?53.
Minson D.J. and Milford R. 1967. The voluntary intake and digestibility of diets containing
different proportions of legume and mature pangola grass (Digitaria decumbens). Australian
Journal of Experimental Agriculture and Animal Husbandry 7:545?551.
Mosi A.K. and Butterworth M.H. 1985. The voluntary intake and digestibility of cereal crop
residues and legume hay for sheep. Animal Feed Science and Technology 12:241?251.
Muinga R.W., Thorpe W. and Topps J.H. 1992. Voluntary food intake, live-weight change and
lactation performance of crossbred dairy cows given ad libitum Pennisetum purpureum
(Napier grass var. Bana) supplemented with leucaena forage in the lowland semi-humid
tropics. Animal Production 55:331?337.
Muinga R.W., Thorpe W. and Topps J.H. 1993. Lactational performance of Jersey cows given
Napier fodder (Pennisetum purpureum) with and without protein concentrates in the semi-
humid tropics. Tropical Animal Health and Production 25:118?128.
Ndlovu L.R. and Buchanan-Smith J.G. 1985. Utilization of poor quality roughage by sheep:
Effects of alfalfa supplementation on ruminal parameters, fibre digestion and rate of passage
from the rumen. Canadian Journal of Animal Science 65:693?703.
Said A.N. and Tolera A. 1993. The supplementary value of forage legume hays in sheep feeding:
feed intake, nitrogen retention and body  weight change. Livestock Production Science
33:229?237.
Siebert B.D. and Kennedy P.M. 1972. The utilization of speargrass (Heteropogon contortus). 1.
Factors limiting intake and utilization by sheep. Australian Journal of Agricultural Research
23:35?44.
Forage legumes as protein supplements to poor quality diets 189
Smith T., Manyuchi B. and Mikayiri S. 1990. Legume supplementation of maize stover. In:
Dzowela B.H., Said A.N., Asrat Wendem-Agenehu and Kategile J.A. (eds), Utilization of
Research Results on Forage and Agricultural By-product Materials as Animal Feed Re-
sources in Africa. Proceedings of the Joint Workshop Held in Lilongwe, Malawi, 5?9
December 1988. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia.
pp. 303?320.
Topps J.H. 1992. Potential, composition and use of legume shrubs and trees as fodder for
livestock in the tropics. Journal of Agricultural Science (Cambridge) 118:1?8.
Vanzant E.S. and Cochran R.C. 1994. Performance and forage utilization by beef cattle receiving
increasing amounts of alfalfa hay as a supplement to low-quality, tall grass-prairie forage.
JournalofAnimalScience72:1059?1067.
190 J.H. Topps
State of art in bioengineering
of rumen microorganisms
Application of molecular biology
in the production of food from ruminants
R.M. Cook
Michigan State University, East Lansing, Michigan 48824, USA
Abstract
Several methods are available to alter end-products of the rumen fermenta-
tion. These include glycopeptides which increase propionate and amino acids
and decrease methane and ammonia nitrogen production. Monensin in-
creases propionate and decreases methane and acetate production. Isoacids
(e.g. isobutyrate) increase acetate and decrease ammonia production. Vir-
giniamycin increases propionate production and, in some studies, decreases
butyrate, acetate and lactate. The changes in rumen fermentation products
brought about by these chemicals can have positive effects on animal pro-
duction depending on whether growth, fattening or milk production is the
desired product. The effect of these chemicals at the molecular level on the
rumen ecosystem is poorly understood. Understanding their mechanism of
action can lead to better methods for controlling rumen fermentation. Of
equal importance in improving food production by ruminants is a knowledge
of utilisation of rumen fermentation products at the molecular level. Each
metabolic event requires transcription of genes. It is critical to understand
gene regulation in adipose, muscle and the mammary gland.
Identifying the important genes regulating metabolism in food producing
animals is in its infancy. It may be that the activity of some genes should be
increased while for others it should be lowered. Some current genes of
interest are fatty acid synthase  (fat deposition) and calpastatin  (muscle
metabolism). The cloning of Holstein mammary gland acetyl CoAsynthetase
gene is an example of the molecular approach to studying utilisation of rumen
fermentation products for milk production. The product of this gene catalyses
the initial step in the utilisation of rumen acetate for energy production and
for fat synthesis. Acetyl CoA synthetase is believed to play a key regulatory
role in ruminants. Understanding molecular mechanisms regulating the ru-
men fermentation and utilisation of rumen fermentation products will pro-
vide a means to enhance food animal production.
Introduction
An important problem in rumen ecology is to define conditions required for optimum
fermentation of feedstuffs, especially low protein, highly fibrous plant material. There
are several chemicals that alter end-products of the rumen fermentation. These include
glycopeptide antibiotics, isoacids and monensin. Our laboratory has conducted both in
vitro and in vivo studies of the effects of glycopeptides, isoacids and monensin on rumen
fermentation (Brondani et al 1991; Quispe et al 1991). Acetate is a major end-product
of the rumen fermentation. Currently we are studying the acetyl CoA synthetase gene
whose product catalyses the initial step in acetate utilisation. These studies will be
described.
Glycopeptide antibiotics
Teichomycin A2 is an antibiotic produced by Actinoplanea teichomyceticus. Avoparcin
is an antibiotic produced by a strain of Streptomyces candidus (Kunstmann et al 1968).
Their antimicrobial properties (growth and bacteria) stem from their ability to interfere
with cell wall synthesis (Parenti et al 1978; Redin and Dornbush 1969).
Cattle fed avoparcin have lower feed intake and improved daily gains and feed
efficiency (DeLay et al 1978; Johnson et al 1979). Changes in the rumen include
increases in propionate and decreases in acetate (Chalupa et al 1981; Froetschel et al
1983) and butyrate (Froetschel et al 1983) molar percent and ?-amino-N.
In vitro studies with teichomycin A2 indicate that this glycopeptide alters rumen
fermentation similar to avoparcin (Brondani 1983; D.J. Phillips and J.M. Tadman,
unpublished data). However, studies directly comparing the two antibiotics under in
vivo conditions have not been conducted. Additionally, most studies measured only
propionate production rates. We determined the effects of teichomycin A2 and avoparcin
on in vivo production of acetate, propionate and butyrate, and rumen nitrogen variables
(Brondani 1986).
In each of two trials, six rumen-cannulated crossbred ewes were kept in individual
pens. Average body weight for animals in Trials 1 and 2 were 40.6 and 39.2 kg,
respectively. In Trial 1 (high roughage diet), animals were fed 1000 g/d of alfalfa hay
and had free access to trace mineral salt. In Trial 2 (low roughage diet), animals received
900 g/d of a ration composed of 43% alfalfa meal, 10% ground corn cobs, 20% ground
shelled corn, 20% ground oats, 5% H
2
O, 1% trace mineralised salt, 0.5% dicalcium
phosphate and 0.5% of a vitamin premix (vitamin A, 4400 IU/g; vitamin D, 500 IU/g;
and vitamin E, 44 IU/g). Daily feed was provided in two portions (0800 and 1600) with
free access to water.
In each trial, animals were divided into three groups of two sheep each and randomly
allocated to one of the treatments: control, 30 ppm teichomycin A2, or 30 ppm avoparcin
(dry matter basis). Antibiotics were injected directly into the rumen through a fistula
immediately following feeding. On days 28 and 29, samples for the determination of
rumen VFA and rumen nitrogen constituents were collected. Rumen fluid samples were
collected prior to morning feeding and then at hourly intervals for the next 8 h.
Measurements of rates of acetate production and liquid flow were performed on days
31 and 37. Three hours after the morning feeding, 100 ml of water containing 10 g of
polyethylene glycol (PEG, M.W. 4000) and 100 ?Ci of Na-[1-
14
C]acetate were added
to the rumen. Serial rumen fluid samples were collected every 20 min for the next 4 h.
The rates of production of propionate and butyrate were measured on days 32 and 39
194 R.M. Cook
and days 33 and 40, respectively. The procedure was essentially the same as that used
for acetate (Cook 1966; Knox et al 1967).
Volatile fatty acid concentrations in rumen fluid were determined by gas liquid
chromatography according to the procedure of Ottenstein and Bartley (1971a, 1971b).
Fractionation of rumen fluid for the determination of nitrogen components was
according to the procedure of Bergen et al (1968) and Isichei (1980). ?-Amino-N and
protein were determined using a Technicon Auto-Analyser System (Technicon Instru-
ment Corp., Terrytown, New York). Ammonia-N analysis was according to the method
of Okuda et al (1965) as modified by Kulasek (1972).
The effects of teichomycin A2 and avoparcin on concentrations of rumen VFA in sheep
fed high and low roughage diets are summarised in Table 1. Volatile fatty acids as a percent
of total moles are shown in Table 2. In the high roughage ration, glycopeptides did not
change the concentrations of acetate (P < 0.10) but increased propionate. Increases over
control were 26.0 and 30.8% for teichomycin A2 and avoparcin (P < 0.10), respectively.
Glycopeptides tended to lower butyrate concentrations. Total VFA concentrations were
not changed by the treatments (P < 0.10). Teichomycin A2 decreased the concentrations
of isobutyrate and isovalerate in rumen fluid (P < 0.05), whereas avoparcin decreased
isovalerate (P < 0.05). The glycopeptides did not affect the molar percent of acetate and
butyrate (Table 2). Propionate molar percent was increased by the glycopeptides relative
to control (P < 0.05). As a result, the acetate to propionate ratio was lower for the treated
groups in the high roughage diet (P < 0.05).
For the low roughage ration, the trends for VFA concentrations were similar to those
in the high roughage ration. Propionate concentrations increased (P < 0.10) and butyrate
concentrations decreased (P < 0.10) due to glycopeptide supplementation. Concentra-
tions of acetate and total VFA were not affected by treatments (P < 0.10). The lower
concentrations of isoacids caused by the glycopeptides in the high roughage ration were
not detected in the low roughage ration. Teichomycin A2 and avoparcin increased
propionate (P < 0.01) and decreased butyrate (P < 0.05) molar percent with respect to
Table 1. Effects of teichomycin A2 (TE-A2) and avoparcin (AVO) on concentrations of rumen VFA in sheep
fed high and low roughage diets (mmoles/dl)
??
.
High roughage Low roughage
Variable Control TE-A2 AVO SE
??
Control TE-A2 AVO SE
??
Acetate 9.94 9.75 6.51 0.57 6.31 5.89 6.12 0.63
Propionate 1.77
??
2.23
??
2.31
??
0.12 1.74
??
2.38
??
2.75
??
0.13
Isobutyrate 0.19
?
0.11
?
0.14
??
0.02 0.21 0.19 0.17 0.03
Butyrate 0.81 0.73 0.65 0.14 0.97
??
0.63
??
0.71
??
0.07
2-M-butyrate 0.16 0.11 0.13 0.03 0.18 0.14 0.19 0.04
Isovalerate 0.14
?
0.09
?
0.07
?
0.01 0.16 0.15 0.13 0.03
Valerate 0.17 0.14 0.16 0.04 0.19 0.21 0.18 0.04
Total VGA 10.26 10.16 9.97 0.96 9.76 9.59 9.75 1.13
Total higher and isoacids 0.66
??
0.45
??
0.50
??
0.04 0.74 0.69 0.67 0.06
?
36 observations per mean.
?
Means in a row within a ration without a common superscript differ:
?,?
P < 0.05;
??,??
P < 0.10.
??
Standard error of each mean in a row within a ration.
Application of molecular biology in the production of food from ruminants 195
control. Acetate molar percent was slightly lower in the treated groups but differences
were not significant. This resulted in the lower acetate to propionate ratios in the treated
groups (P < 0.01).
The effect of teichomycin A2 and avoparcin on the rates of production of acetate,
propionate and butyrate in sheep fed high and low fibre diets are shown in Tables 3?5.
Animals fed the high roughage diet had a higher rumen fluid turnover rate than animals
fed low roughage. These results are in line with those reported for sheep (Froetschel et
al 1983). Froetschel et al (1983) reported an increase in rumen volume in animals fed
avoparcin. However, the present study did not show increases for either the avoparcin
or the teichomycin A2 groups.
In the high roughage ration, there was a trend for lower acetate production in the
treated groups (Table 3). However, differences were not significant (P > 0.10). Although
acetate pools were not affected by treatments (P > 0.10), the lower rate of acetate
Table 2. Effects of teichomycin A2 (TE-A2) and avoparcin (AVO) on molar percent distribution of rumen
VFA in sheep fed high and low roughage diets
??
.
High roughage Low roughage
Variable Control TE-A2 AVO SE
??
Control TE-A2 AVO SE
??
Acetate 68.0 66.3 65.2 0.89 64.6 61.4 62.8 0.82
Propionate 17.3
C
22.0
?
23.2
?
0.61 17.8
?
24.8
?
23.1
?
0.59
Isobutyrate 1.9 1.1 1.5 0.11 2.1 2.0 1.7 0.16
Butyrate 7.9 7.2 6.5 0.49 9.9
??
6.6
??
7.3
??
0.51
2-M-butyrate 1.6 1.1 1.3 0.13 1.8 1.5 1.9 0.09
Isovalerate 1.4 1.4 1.7 0.11 1.9 2.2 1.8 0.13
Valerate 1.9 1.4 1.7 0.11 1.9 2.2 1.8 0.13
A:P ratio 3.9
??
3.0
??
2.8
??
0.13 3.6
?
2.5
?
2.7
?
0.17
?
36 observations per mean.
?
Means in a row within a ration without a common superscript differ:
?,?
P < 0.05;
??,??
P < 0.10.
??
Standard error of each mean in a row within a ration.
Table 3. Effects of teichomycin A2 (TE-A2) and avoparcin (AVO) on acetate production and rumen fluid
variables in sheep fed high and low roughage diets
?
.
High roughage Low roughage
Variable Control TE-A2 AVO SE
?
Control TE-A2 AVO SE
?
Acetate
Pool size (moles) 0.36 0.34 0.39 0.02 0.31 0.27 0.29 0.03
Turnover time (min) 110.3
?
112.4
?
134.1
?
6.9 103.5 97.6 102.8 9.1
Production (moles/d) 4.74 4.41 4.12 0.27 4.39 3.89 4.18 0.36
Rumen fluid
Volume (litres) 5.1 4.4 4.6 0.47 4.4 3.8 4.0 0.31
Turnover time (h) 8.7 9.2 8.5 1.39 14.3 16.1 13.1 0.98
Turnover rate (%/h) 11.1 10.5 11.9 1.03 6.7 6.3 7.7 0.62
?
Four observations per mean.
?,?
Means in a row within a ration without a common superscript differ (P < 0.10).
?
Standard error of each mean in a row within a ration.
196 R.M. Cook
production due to avoparcin resulted in significantly less acetate pool turnover times
relative to control (P < 0.10). In the low roughage ration, trends were the same as those
found in the high roughage ration. Acetate production and acetate pool sizes tended to
be lower in the treated groups, but differences were not significant (P > 0.10).
Propionate production rates were increased by glycopeptide supplementation in both
rations (Table 6). In the high roughage diet, increases in propionate production relative
to controls were 48 (P < 0.05) and 38.7% (P < 0.05) for teichomycin A2 and avoparcin,
respectively. In the low-roughage ration, increases in propionate production due to
teichomycin A2 and avoparcin were 34.8 (P < 0.10) and 36.8% (P < 0.10), respectively.
The increase in propionate production resulted in higher propionate pool sizes for both
antibiotics in both rations. However, significant differences from control were detected
only for avoparcin in the low roughage diet (P < 0.10). Butyrate production (Table 5)
was not affected by treatments in either ration (P > 0.10).
Table 4. Effects of teichomycin A2 (TE-A2) and avoparcin (AVO) on propionate production, rumen volume,
and rumen fluid turnover rate in sheep fed high and low roughage diets
?
.
High roughage Low roughage
Variable Control TE-A2 AVO SE
?
Control TE-A2 AVO SE
?
Propionate
Pool size (moles) 0.12 0.16 0.14 0.02 0.13
?
0.17
??
0.21
?
0.03
Turnover time (min) 136.80 123.50 115.20 9.73 121.70
?
119.80
?
114.4
?
7.12
Production (moles/d) 1.24
?
1.84
?
1.72
?
0.12 1.55
?
2.09
?
2.12
?
0.14
Rumen fluid
Volume (litres) 4.40 5.10 4.70 0.36 3.50 3.90 4.40 0.42
Turnover time (h) 9.30 8.40 10.10 1.24 15.20 13.40 14.30 0.86
Turnover rate (%/h) 10.80 12.10 9.10 0.97 6.40 7.30 6.80 0.74
?
Four observations per mean.
?,?
Means in a row within a ration without a common superscript differ (P < 0.10).
?
Standard error of each mean in a row within a ration.
Table 5. Effects of teichomycin A2 (TE-A2) and avoparcin (AVO) on butyrate production, rumen volume,
and rumen fluid turnover rate in sheep fed high and low roughage diets
??
.
High roughage Low roughage
Variable Control TE-A2 AVO SE
?
Control TE-A2 AVO SE
?
Butyrate
Pool size (moles) 0.051 0.044 0.047 0.004 0.067 0.071 0.059 0.003
Turnover time (min) 150.60 138.90 143.80 13.10 213.20 214.60 199.30 17.50
Production (moles/d) 0.475 0.453 0.461 0.06 0.443 0.469 0.429 0.09
Rumen fluid
Volume (litres) 5.40 4.80 5.10 0.49 4.30 5.10 3.80 0.39
Turnover time (h) 8.10 8.50 9.20 1.19 12.20 14.40 12.90 0.96
Turnover rate (h) 12.10 11.40 10.20 0.97 8.50 6.80 7.50 0.71
?
Four observations per mean.
?
Effects of treatment are not significant in either ration (P < 0.10).
?
Standard error of each mean in a row within a ration.
Application of molecular biology in the production of food from ruminants 197
Changes in VFA production rates found in the present study followed similar trends
for both glycopeptides. In general, supplementation with teichomycin A2 and avoparcin
resulted in significant increases in propionate concentrations and production rates. For
acetate and butyrate, there were trends toward lower concentration and production rates
but differences were not significant (P > 0.10). This is in contrast with results from
studies in sheep (Froetschel et al 1983) and cattle (Chalupa et al 1981). Froetschel et al
(1983) reported higher propionate but lower acetate, butyrate and total VFA concentra-
tions in sheep fed high and low fibre diets supplemented with avoparcin. Chalupa et al
(1981) reported that avoparcin supplementation to cattle increased propionate and
butyrate and decreased acetate. Total VFA concentrations were not affected by the
antibiotic. However, our results are in line with those obtained in feedlot trials by
Johnson et al (1979). In these experiments, avoparcin consistently increased propionate
without affecting acetate or butyrate concentrations.
One point of importance is the time required for the patterns of rumen VFA to return
to pre-treatment values when feeding of antibiotics is discontinued. In contrast with
ionophores, the effects of glycopeptides seem to be persistent (Brondani 1983; D.J.
Phillips and J.M. Tadman, unpublished data). Ten days after glycopeptide supplemen-
tation was discontinued, the molar proportion of rumen VFA had not returned to normal.
These results question the validity of using Latin squares or reversal designs to study
the effects of glycopeptides on rumen metabolism unless there are long adaptation
periods to the new treatment to avoid carryover effects.
The effect of teichomycin A2 and avoparcin on rumen nitrogen constituents in sheep
fed high and low fibre diets is shown in Table 6. Soluble, bacterial and protozoal protein
fractions were not affected by treatments (P > 0.10).
In the high roughage ration, increases in ?-amino-N relative to control were 81.6
(P < 0.05) and 118% (P < 0.05), respectively, for teichomycin A2 and avoparcin. In the
low roughage diet, teichomycin A2 and avoparcin increased ?-amino-N concentrations
by 50.1 (P < 0.05) and 48.6% (P < 0.05), respectively. Ammonia-N in the high roughage
diet was decreased by 10 (P < 0.05) and 23.8% (P < 0.05) by teichomycin A2 and
avoparcin. In the low roughage diet, decreases in ammonia-N due to teichomycin A2
and avoparcin were 33.2 and 22.7%, respectively (Table 6).
Table 6. Effects of teichomycin A2 (TE-A2) and avoparcin (AVO) on rumen nitrogen constitutents in sheep
fed high and low roughage diets (mg/dl)
?
.
High roughage Low roughage
Variable Control TE-A2 AVO SE
??
Control TE-A2 AVO SE
??
Soluble protein 46.19 52.1 39.43 4.76 37.09 44.03 49.12 4.11
Bacterial protein 57.29 65.8 61.20 5.12 63.14 54.28 63.19 4.63
Protozoal protein 61.34 72.13 65.23 6.17 265.19 234.29 251.12 19.25
Alpha-amino-N 2.94
?
5.34
?
6.41
?
0.48 3.19
?
4.81
??
4.74
??
0.42
Ammonia-N 16.59
?
13.21
??
11.19
?
1.19 17.35
?
10.24
?
11.86
?
1.28
?
36 observations per mean.
?,?
Means in a row within a ration without a common superscript differ, P < 0.05;
?,??
P < 0.05.
??
Standard error of each mean in a row within a ration.
198 R.M. Cook
Results for ammonia-N and ?-amino-N indicate that teichomycin A2 and avoparcin
exert a marked effect on rumen nitrogen metabolism. These results are in agreement
with in vitro results for teichomycin A2 (Brondani 1983) and in vivo results for avoparcin
(Chalupa et al 1981; Froetschel et al 1983). The decrease in ammonia-N concentration
concomitant with increases in ?-amino-N suggests that teichomycin A2 and avoparcin
reduce deamination. Additional evidence is provided by the decline in the concentration
of isoacids in treated animals (Table 1). Although some rumen bacterial species can
synthesise branched chain acids de novo (Bryant and Doetsch 1955; Miura et al 1980),
the majority of isoacids in the rumen comes from the degradation of dietary protein.
Therefore, the decline in isoacids suggests that amino acid breakdown was depressed
by the glycopeptides.
Both compounds increase propionate production but do not affect acetate and butyrate
production. Both glycopeptides increase ?-amino-N and decrease ammonia-N concen-
trations in rumen fluid, suggesting less protein and/or amino acid degradation in the
rumen.
Isoacids
Ruminal bacteria require isoacids (isobutyrate, 2-methylbutyrate, isovalerate and valer-
ate), NH
3
and hydrogen sulfide for growth (Brondani et al 1991). Little information is
available concerning the interaction of these three nutrients, particularly when high fibre
diets are fed. To study the interaction of isoacids, NH
3
and sulfur, a 2
3
factorial crossover
(three factors each of three levels) conducted in two 4 ? 4 Quasi-Latin squares was used
(Brondani et al 1991).
Double blocking criteria were animals and time (non-random repeated measurements
were obtained from each subject assigned to a sequence of treatment combinations). In
each trial, eight rumen-fistulated adult Tabasco sheep were divided into two groups.
Sheep in Groups 1 and 2 weighed 27 and 36 kg, respectively, and were assigned
randomly to rows of 1 square, corresponding to a predetermined sequence of treatment
combinations. The three factors were isoacids, N, and S each at two levels. Composition
of the basal diets for Trials 1 and 2 is given in Table 7. Corn stover was chopped with
a silage chopper to lengths of 1.26 cm. Both sugar cane bagasse and stover were
air-dried.
Based on results of previous experiments, an equal weight mixture of isoacids
(isobutyrate, 2-methylbutyrate, isovalerate and valerate) was administered at 0.1 and
0.2 g/kg body weight (BW) per day. To achieve two levels of NH
3
-N in the rumen (about
5 and 15 mg/dl), the basal diet was fed either alone or supplemented with urea (high N
treatment) at 1.5% of DM. To achieve two levels of sulfide in the rumen (4 and 8 ?g/ml),
the basal diet was fed alone or supplemented with elemental S at 0.2% of DM. Sulfur
and N supplementation was designed to provide four different N:S ratios. Final N:S
ratios were approximately 3:1, 5:1, 8:1 and 12:1. Isoacids, urea and S were premixed
weekly with part of the sorghum and then incorporated into the totally mixed diet. The
daily ration was fed at 0700 and feed intake was recorded daily for each animal. Water
Application of molecular biology in the production of food from ruminants 199
was provided ad libitum. Animals were adapted to a given diet for 14 days prior to
measurements.
In vivo production rates of acetate were measured by a single injection radioisotope
technique (Cook 1966; Rogers and Davis 1982). On day 15, 3 h after morning feeding,
each animal received an intraruminal injection of 100 ?Ci of Na-[1-
14
C]acetate with
100 ml of a 10% solution of polyethylene glycol (molecular weight 3350). Following
infusion, ruminal fluid samples were collected every 20 min for the next 3 h. Samples
were strained through four layers of surgical gauze and stored frozen at ?20?C. Ruminal
hydrogen sulfide was determined using a hydrogen sulfide-sensing electrode (Lazar
model GS-136). Isolation of
14
C-acetate from rumen fluid was as described above under
glycopeptides. Determinations of acetate production, ruminal NH
3
-N and acetate were
described previously (Cook 1966).
In a third trial, sheep in Groups 1 (25 kg) and 2 (35 kg) were each fed 1215 and 1700
g/d of pineapple tops and 20 and 27 g/d of minerals, respectively (Quispe et al 1991).
Pineapple tops were 6.6 and 0.13% CP and S, respectively.
Pineapple tops were chopped in a silage cutter, air-dried to 20% moisture, and finally
ground for formulating rations. Isoacids (equal weights of isobutyrate, 2-methyl-
butyrate, isovalerate and valerate) were mixed into rations at two levels to provide 0.07
and 0.14 g kg BW
?1
d
?1
. In order to achieve ruminal NH
3
-N levels of approximately 5
and 15 mg/dl, pineapple tops were fed without or with urea at 0.43 g kg BW
?1
d
?1
.
Similarly, either 0 or 0.086 g S kg BW
?1
d
?1
was provided by adding elemental sulfur.
Experimental periods were one week. Acetate production was measured on the last day
of each period.
For Trials 1 and 2, results are presented in view of the existence of two significant
two-way interactions, i.e. the interactions between N with S and N with isoacids. The
Table 7. Composition and analysis of basal diets
?
.
Item Trial 1 Trial 2
Ingredient (%)
Corn stover ? 25.0
Corn cobs ? 25.0
Sugarcane bagasse 44.0 ?
Sorghum grain 55.0 49.0
Bone meal 0.5 0.5
Trace-mineralized salt
?
0.5 0.5
Analysis
?
CP (%) 7.2 7.7
Crude fibre (%) 22.0 18.2
Digestable energy (mcal/kg) 2.6 2.9
Total N (%) 1.15 1.23
Total S (%) 0.14 0.18
N:S 8.0 6.8
?
DM basis.
?
Composition: Zn = 0.35%; Mn = 0.20%; Fe = 0.20%; Cu = 0.03%; Co = 0.005%; I = .007%; and NaCl = 96%.
?
Calculated NRC values.
200 R.M. Cook
interaction of isoacids and S was not significant. Therefore, there was no specific
examination of combination means involving those two factors. When two factors
interact, it is appropriate to examine the means of two-way combinations (averaged over
other factors, if any) to determine the nature of the interaction. For the two-way
combinations between N and S (Table 8), DMI did not differ among treatment groups
in either trial. Urea increased (P < 0.01) ruminal NH
3
-N in both trials. Similarly, S
supplementation resulted in higher (P < 0.01) levels of sulfide-S in ruminal fluid of
sheep on both trials. In both trials, acetate production was not changed by either urea
or S when fed separately. Supplementation of both resulted in a 44% increase in acetate
production in Trial 1 and a 63% increase in Trial 2. Although the lack of an increase in
acetate production in the group with low N and high S may be explained by the low
level of ruminal NH
3
-N (Table 8), the results for the group with high N and low S were
not expected. The calculated N:S ratio and the percentage of S in the basal diets (Table
8) were within the ranges commonly recommended for adult sheep (Bray and Till 1975;
?rskov 1982). However, S intake by animals in the group with high N and low S may
have been limiting. Hume and Bird (1970) reported that S intake of 1.95 g/d supported
maximal ruminal microbial protein synthesis in sheep. In the present study, intake of
total S in the group with high N and low S averaged 0.65 and 0.74 g/d, respectively, for
Trials 1 and 2. For the group with high N and high S, average daily intake of S for sheep
in Trials 1 and 2 was 1.66 and 1.80 g/d, respectively. The precise level at which ruminal
sulfide concentration limits microbial growth and fermentation has not been defined
clearly, but 1 ?g of sulfide-S/ml of ruminal fluid has been suggested to be the lower
limit for optimal fermentation (Harrison and McAllen 1981). Concentrations of sulfide-
S in the groups with high N and low S were greater than 2 ?g/ml but may not have been
high enough for maximal fermentation. These results suggest that S supplementation
Table 8. Effects of N and S on feed intake and ruminal fermentation variables in sheep fed high fibre diets.
Acetate
DMI Feed NH
3
N Sulfide-S production
(g/d) N:S (mg/dl)
?
(?g/ml)
?
(mol/d)
Trial 1 (sugarcane bagasse diet)
Low N, low S 387 8.1 6.11 2.43 2.18
?
High N, low S 454 12.7 12.19 2.16 1.94
?
Low N, high S 442 3.1 5.14 5.84 2.25
?
High N, high S 462 5.0 13.11 5.51 3.16
?
S.E. 16 ? 0.84 0.33 0.24
Trial 2 (corn stover diet)
Low N, low S 430 7.0 5.28 2.97 2.24
?
High N, low S 412 10.7 14.19 3.34 2.82
?
Low N, high S 461 3.2 6.93 5.26 2.31
?
High N, high S 442 4.7 13.11 4.81 3.67
?
S.E. 15 ? 0.71 0.28 0.21
?
Significant effect of N (P < 0.01).
?
Significant effect of S (P < 0.01).
?,?
Means in the same column within a trial with different superscripts differ (P < 0.10).
Application of molecular biology in the production of food from ruminants 201
based solely on the N:S ratio or on the percentage of S in DM may not be adequate when
high fibre diets containing low N are fed. The total amount of S that would allow
maximal microbial protein synthesis should be considered first. Once that is provided,
supplementing N to attain an N:S ratio of about 10 (?rskov 1982) should result in
maximal fermentation efficiency.
Increasing the level of isoacids in the diet resulted in higher (P < 0.01) concentrations
of these acids in the rumen for both trials (Table 9). In Trial 1, acetate production was
higher when N was supplemented along with low isoacids (P < 0.05). Increasing the
amount of isoacids in the diet, without adding N, did not change acetate production.
When both were present at the high level, acetate production was further increased
(P < 0.05) by 30%. These results show that rates of fermentation in the rumen can be
only as high as the availability of the most limiting nutrient. In Trial 1, ruminal NH
3
-N
provided by the basal diet was more limiting than isoacids. Once N supply was increased
by the addition of urea to the diet, isoacids became limiting.
Trends in acetate production found in Trial 2 were similar to Trial 1 (Table 9). However,
increases (P < 0.05) in acetate production were found only when both N and isoacids were
fed at a high level. This suggests that ruminal NH
3
-N provided by the basal diet in Trial
2 was adequate to support microbial fermentations as efficiently as the diet that was
supplemented with N. However ruminal NH
3
-N was not sufficiently high to permit
utilisation of the higher amounts of isoacids supplied by the high isoacid treatment.
In Trial 3, two different ruminal levels of H
2
S and NH
3
-N were achieved by supple-
menting pineapple tops with elemental sulfur and urea (Table 10). Ruminal H
2
S averaged
7to9?g/ml when S was added compared to 0.7 to 4 ?g/ml with no S. Similarly, ruminal
NH
3
-N was 13 to 16 mg/dl when urea was added compared to 5 to 7 mg/dl without urea.
Table 9. Effects of isoacids (Iso) and N on feed intake and ruminal fermentation variables in sheep fed high
fibre diets.
Total Acetate
DMI NH
3
N isoacids production
(g/d) (mg/dl)
?
(mmol/dl)
?
(mol/d)
Trial 1 (sugarcane bagasse diet)
Low Iso, low N 384 6.82 0.28 1.91
?
Low Iso, high N 421 15.10 0.29 2.86
?
High Iso, low N 438 7.14 0.49 1.97
?
High Iso, high N 417 14.36 0.53 3.74
??
S.E. 16 0.64 0.03 0.17
Trial 2 (corn stover diet)
Low Iso, low N 431 8.31 0.26 2.64
?
Low Iso, high N 456 17.11 0.28 2.95
?
High Iso, low N 471 7.34 0.47 2.88
?
High Iso, high N 448 15.21 0.48 3.87
??
S.E. 15 0.71 0.02 0.21
?
Significant effect of N (P < 0.01).
?
Significant effect of isoacids (P < 0.01).
?,?,??
Means in the same column within a trial with different superscripts differ (P < 0.05).
202 R.M. Cook
Ruminal concentrations of isoacids were not increased with high levels of isoacids.
This may be due to rapid utilisation in the rumen. The ruminal N:S ratio (Table 10)
reflected the dietary N:S ratio.
For acetate production, interactions were only significant between isoacids and S.
Therefore, the Bonferroni t-test was used to test differences among treatment means
within each level of isoacids or S. These results show that acetate production was
increased (P < 0.05) with S supplementation. Similarly, acetate production was higher
when adding S and high isoacids (P < 0.01).
Acetate production was higher with a dietary N:S ratio of 5:1 relative to 3:1, 8:1 or
12:1 (Figure 1). This ratio is narrower than the ratio usually considered adequate for
sheep (Bray and Till 1975). The dietary N:S ratio of 5:1 resulted in ruminal NH
3
-N to
H
2
S ratio of 17:1, which is close to the N:S ratios for ruminal bacteria of approximately
13:1 proposed by Bray and Till (1975).
Acetate production was more correlated to N:S ratios (r
2
= 14.6%) than to ruminal
NH
3
-N concentration (r
2
= 3.2%). There was a trend towards more acetate production
as ruminal NH
3
-N increased, but differences were not significant.
In general, acetate production was 42% greater when higher levels of isoacids and S
were added to the feed ration. Adding urea to diets containing high levels of isoacids
and S increased acetate production by another 33%. It is concluded that 0.14 g of
isoacids, 0.43 g urea and 0.086 g S kg BW
?1
gave the maximum fermentation of
pineapple tops. These treatments gave a dietary N:S ratio of 5:1.
Because production of VFA from carbohydrates in the rumen is coupled with microbial
growth, maximal microbial yield can be attained only if precursors for protein synthesis
are made available to the microbiota simultaneously and in adequate quantities. This study
suggests that high fibre diets low in N are utilised better when all three factors (N, isoacids
and S) are adequate.
Table 10. Effects of interaction of isoacids, urea, and sulfur, each at 2 levels, on ruminal acetate production,
H
2
S, and NH
3
-N in sheep fed pineapple tops.
Acetate
H
2
SNH
3
-N Rumen Ration (mol h
-1
Treatments
?
(?g/ml)
?
(mg/dl)
?
(NH
3
-N:H
2
S) (N:S) per sheep)
A
L
B
L
C
L
0.7 6.7 90 8 0.21
A
H
B
H
C
L
4.0 14.8 36 12 0.22
??
A
H
B
L
C
H
7.8 7.4 9 3 0.33
?
A
L
B
H
C
H
7.3 12.8 18 5 0.25
?
A
L
B
L
C
H
7.0 7.0 10 3 0.17
?
A
L
B
H
C
L
2.9 15.9 54 12 0.27
A
H
B
L
C
L
3.5 4.8 14 8 0.28
??
A
H
B
H
C
H
9.1 13.9 16 5 0.44
?
?
A, B and C are levels of isoacids, urea, and sulfur, respectively. L and H are low and high levels.
?
S.E. for H2S and NH2-N is 0.6 and 1.2, respectively.
?,?,??
Two pooled treatment contrasts were different: ? vs. ? (P < 0.05), positive effect of isoacids when sulfur was present
and ? vs. ?? (P < 0.1), positive effect of sulfur when isoacids were high. All other treatment comparisons were not different
(P > 0.01).
Application of molecular biology in the production of food from ruminants 203
N:S Ratios
0 1:1 2:1 3:1 4:1 5:1 6:1 7:1 8:1 9:1 10:1 11:1 12:1
0
0.1
0.2
0.3
0.4
Ace
tat
ePr
oduc
tio
n(
mo
les/
hr
xs
heep)
Optimum N:S ratio
x
x
x
x
Figure 1. Effects of N:S ratios in the ration on mean rumen acetate production in sheep (SE = 0.05).
204
R.M.
Coo
k
Monensin
Monensin has been used extensively in diets for feedlot cattle for several years. Isoacids
have been used as a nutritional supplement for lactating dairy cattle. Both monensin and
isoacids affect rumen fermentation. However, little is known about the effects of the
combination of the two compounds in the rumen. In vitro rumen fermentations have shown
that monensin decreases the acetate:propionate ratio (Chalupa 1977; Richardson et al
1976). These results have been attributed to a toxic effect of the ionophore on ruminococci
(Henderson et al 1981). Unlike monensin, isoacids (isobutyric, isovaleric, 2-methylbutyric
and valeric acids) increase rumen acetate production, probably due to an enhanced growth
of ruminococci (Allison and Bryant 1958; Bryant 1973; Felix et al 1980).
An in vitro system was used to study the effects of monensin and isoacids on the
rumenfermentation(Koneetal1989).Ruminalfluidwasobtained3hpost-feeding
from a mature non-pregnant and non-lactating rumen-fistulated Holstein cow that
weighed 550 kg. Samples from different parts of the rumen were strained through two
layers of surgical gauze into 1 litre glass bottles kept at 40?C.
The semicontinuous culture technique of Short (1978) was used with minor modifi-
cations  (Figure 2). The  culture  was  maintained  in the same  flask throughout  the
experiment. The substrate was timothy hay.
Isobutyric, 2-methylbutyric, isovaleric and valeric acids were mixed at equimolar
concentrations and neutralised with sodium hydroxide. The final concentrations of the
Figure 2. Daily rumen VFA concentration in vitro using timothy hay as substrate.
12345678910112
Days of Incubation
0
1
2
3
4
5
6
7
8
9
10
11
12
VF
A
(mmoles/dl)
total vfa
acetate
propionate
butyrate
Application of molecular biology in the production of food from ruminants 205
isoacid mixture in the incubation flasks were 10, 15 and 20 mg/dl of media. Monensin
(Sigma Chemical Co., S. Louis, MO) was first dissolved in 10 mg of methanol and then
diluted with water. The final medium concentrations were 100, 150 and 200 ?g/dl.
Experiment 1 investigated the effect of isoacids at 10, 15 and 20 mg/dl of final
incubation medium. Experiment 2 investigated the effect of monensin at 100, 150 and
200 ?g/dl of final incubation medium. For Experiment 3, monensin was fixed at 150
?g/dl, and isoacids were added to the incubation medium at 10 or 15 mg/dl of the final
incubation.
Isoacids at 10 and 15 mg/dl increased acetate, total VFAand propionate concentration,
but the increase was not as great at 20 mg/dl. As expected, branched-chain fatty acid
concentrations increased in proportion to the amounts added in the medium. In contrast
to VFA production, CH
4
,CO
2
and H
2
were not affected by isoacids (Table 11).
The effects of different concentrations of monensin are in Table 12. In contrast to
isoacids, acetate concentration decreased at all concentrations of monensin tested.
Propionate concentration was significantly increased, whereas butyrate concentration
decreased at all concentrations tested. Monensin decreased methane production. Neither
CO
2
nor H
2
was affected. The results from the monensin experiment agree with similar
studies conducted by Chalupa (1977). Monensin consistently decreased acetate and
butyrate but increased propionate production.
The addition of monensin caused a decrease in acetate and propionate after 24 h of
incubation, but after 48 h, propionate was higher than control values, whereas acetate
remained lower than the controls. The addition of isoacids at 10 and 15 mg/dl to flasks
containing monensin increased acetate concentration compared with monensin alone
(P < 0.05; Figure 3). However, the addition of isoacids to flasks containing monensin
did not alter propionate concentration (Figure 4).
Table 11. Effect of isoacids on rumen VFA concentration and gas composition in vitro using inoculum from
a cow fed timothy hay.
Isoacid concentration, mg/dl
Variable Control 10 15 20 SEM
mmol/dl
VFA 7.87
?
8.72
?
8.93
?
8.41
?
0.08
Acetate 5.48
?
6.07
?
6.17
?
5.61
?
0.07
Propionate 1.55
?
1.69
?
1.63
?
1.52
?
0.02
Butyrate 0.60 0.58 0.66 0.65 0.03
Isoacids 0.25
?
0.40
?
0.49
?
0.59
?
0.01
Isobutyrate 0.03
?
0.06
?
0.07
?
0.08
?
0.001
2-methylbutyrate 0.06
?
0.10
?
0.13
?
0.15
?
0.003
Isovalerate 0.04
?
0.10
?
0.12
?
0.13
?
0.003
Valerate 0.12
?
0.14
?
0.16
?
0.22
?
0.006
Percent
CH
4
35.15 36.15 34.38 33.08 1.99
CO
2
49.02 54.94 49.03 47.37 3.15
H
2
3.83 4.00 3.67 3.83 0.2
?,?,?,?
Means in a row with different superscripts differ (P < 0.05).
206 R.M. Cook
It is interesting to speculate on the mode of action of the combination of isoacids and
monensin. According to Wolin and Miller (1983), monensin affects the rumen fermen-
tation by selecting for organisms that participate in the production of relatively more
propionate and against those that contribute to the production of acetate, butyrate and
precursors of methane. The growth of ruminococci and butyrivibrios is inhibited by very
low concentrations of monensin. These genera are important producers of acetate,
butyrate and the substrate for methanogens, H
2
and CO
2
. Selenomonads are very
insensitive, whereas bacteroids, although sensitive, rapidly become resistant to the
ionophore. Both organisms are important in the production of propionate. The addition
of isoacids to the culture containing monensin may cause an outgrowth of bacteroids
and perhaps ruminococci and methanogenic bacteria, which require isoacids, resulting
in more acetate, succinate and formate. Succinate would be decarboxylated to propion-
ate by selenomonads and formate would be used as an energy source by methanogens.
The end result would be an increase in acetate and probably more substrate degraded
as indicated by the higher VFA in Experiment 3 (Figures 3 and 4).
The combination of isoacids and monensin may be of practical application in cattle
rations. The addition of monensin alone to the diet of growing steers increases the
efficiency of growth by increasing propionate production. Because of the importance
of propionate in glucose metabolism and insulin secretion in ruminants, the increase in
ruminal propionate production will promote growth. Isoacids increase acetate, total VFA
and microbial synthesis. Because the addition of isoacids to cultures containing monen-
sin does not alter the effect of monensin on propionate production, the additional energy
and microbial protein may further improve growth.
The changes in rumen fermentation products brought about by glycopeptides,
monensin and isoacids can have positive effects on animal production. In addition,
there are other chemicals such as virginiamycin and lasolacid that affect rumen VFA
production. These chemicals provide a tool to study regulation of microbial growth at
the molecular level. Knowledge in this area is critically needed to provide a basis for
Table 12. Effect of monensin on rumen VFA concentration and gas composition in vitro using inoculum
from a cow fed timothy hay.
Monensin, ?g/dl
Variable Control 100 150 200 SEM
mmol/dl
VFA 8.54
?
7.22
?
6.84
?
6.88
?
0.14
Acetate (A) 6.02
?
4.28
?
3.74
?
3.60
?
0.11
Propionate (P) 1.73
?
2.27
?
2.28
?
2.42
?
0.03
Butyrate 0.52
?
0.46
?
0.42
?
0.35
?
0.02
Isoacids 0.18 0.16 0.16 0.19 0.01
A:P 3.50
?
1.90
?
1.75
??
1.61
?
0.07
Percent
CH
4
36.33
?
27.07
?
26.77
?
26.00
?
0.91
CO
2
49.83 48.33 42.50 47.50 3.2
H
2
3.50
?
3.5
?
3.5
?
4.00
?
?,?,?
Means in a row with different superscripts differ (P < 0.05).
Application of molecular biology in the production of food from ruminants 207
developing methods to control the rumen ecosystem under a variety of feed manage-
ment systems.
Utilisation of rumen VFA
In addition to developing better methods to control the rumen ecosystem, it is equally
important to understand utilisation of rumen VFA for production of meat, milk and fibre.
Of course, molecular mechanisms regulating utilisation of substrates in ruminants are
not known. Our laboratory is engaged in a study of the regulation of acetate for milk
synthesis at the molecular level. Acetate, as acetyl coenzyme A (acetyl-CoA) is a major
substrate for milk synthesis and occupies a central position in metabolism of Holstein
mammary gland as well as in both prokaryotic and eukaryotic cells. Acetyl-CoA is
generated via the acetate activation reaction catalysed by acetyl coenzyme A synthetase
(ACS).
Our work on characterisation of ACS in ruminants has determined the tissue distri-
bution, intracellular  localisation, physical and catalytic properties and  changes in
mammary gland throughout lactation. In addition, a bovine mammary gland ACS cDNA
has been cloned (Raafat 1994). ACS is most active in heart, kidney, adipose tissue and
mammary gland at peak lactation (Cook and Qureshi 1972; Cook et al 1969, 1975).
There is marginal activity in brain and skeletal muscle. Unlike non-ruminants, acetate
Figure 3. Effect of the combination monensin/isoacids on daily rumen acetate concentration by rumen
microbes in vitro using timothy hay as substrate.
4 5 6 7 8 9 10 11 12
Days of incubation
3.6
4.2
4.8
5.4
6
6.6
Acetate
(mmoles/dl)
monensin
M + 1-15
M + 1-10
isoacids
control
208 R.M. Cook
is not utilised by ruminant liver (Cook 1970; Cook and Miller 1965) and ACS activity
is not found in liver (Ricks and Cook 1981a). However, in vitro, acetate can be oxidised
by liver because propionyl CoA synthetase can activate acetate at a low rate (Ricks and
Cook 1981a). The apparent molecular weight of the mammary gland enzyme is 63,000
(Qureshi and Cook 1975). ACS is a glycoprotein and the carbohydrates have been
identified (Ricks and Cook 1981b; Stamoudis and Cook 1975).
In early lactation, blood growth hormone is relatively high and insulin is low. As
lactation advances and milk production declines, blood levels of growth hormone
decrease and insulin increases. Also, high producing cows have lower blood insulin and
higher growth hormone than do low producing cows (Ghirardi and Cook 1987). Insulin
inhibits and growth hormone stimulates ACS activity in lactating goat mammary gland.
As lactation advances, ACS activity decreases sharply but can be partially reinstated by
injecting a combination of growth hormone, prolactin and dexamethasone (Marinez et
al 1976). The existence of 5? regulatory site hormonal response elements (e.g. gluco-
corticoids) in the ACS gene might explain these coordinated regulatory patterns.
In cattle, ACS activity is marginal in a dry mammary gland, increases from parturition
to peak milk production and then declines in activity as lactation advances (Marinez et
al 1976). ACS activity is directly correlated with milk production. Thus, as lactation
advances, less acetate can be utilised.
Figure 4. Effect of the combination monensin/isoacids on daily rumen propionate concentration by rumen
microbes in vitro using timothy hay as substrate.
4 5 6 7 8 9 10 11 12
Days of incubation
1.66
1.82
1.98
2.14
2.30
2.46
Propionate
(momoles/dl
)
M +1-15
monensin
M + 1-10
isoacids
control
1.50
Application of molecular biology in the production of food from ruminants 209
Given the central role of ACS in ruminant metabolism and milk production, cDNA
cloning of the ACS gene was initiated using poly(A+)RNA isolated from bovine
mammary tissue, taken at peak lactation. Mammary gland mRNA was reverse tran-
scribed with AMV reverse transcriptase (Promega Inc., Madison, WI), according to
the manufacturers?instructions. Following second strand synthesis and EcoRI adaptor
addition, cDNAs were cloned into the EcoRI site of lambda gt11. Initial screening
with rabbit anti-ACS sera, prepared by injection of purified bovine ACS, yielded a
truncated cDNA clone, AR8. AR8 remained positive through several rounds of plaque
purification. Subsequent screening of 750,000 plaques using AR8 yielded 1 clone,
designated ATC5, which remained positive through several rounds of plaque purifi-
cation. ATC5 contains a 4.2 Kbp insert, more than sufficient to encode the complete
ACS protein. The insert from ATC5 was excised by EcoRI digestion and subcloned
into the EcoRI site of pUC19 (generating pATC5) for further analysis and sequencing.
Prior to initiation of full-scale DNA sequence analysis, several steps were taken to
ensure pATC5 represented a true copy of ACS gene coding sequences.
1. Rabbit anti-ACS was affinity purified using the fusion protein produced by induc-
tion of pATC5 with IPTG. Purified anti-ACS removed over 90% of ACS activity
from a preparation of partially purified bovine mammary ACS. In contrast, preim-
mune sera did not adversely affect ACS activity.
2. Anti-ACS, affinity purified using LacZ-ACS fusion protein from clone pAR8, was
used in Western blots of mitochondrial extracts from heart, liver, kidney, mammary
gland and spleen. Proteins with the expected molecular weight and tissue distribu-
tion as ACS were recognised.
3. Preliminary DNA sequencing of pAR8 and pATC5 revealed an open reading frame
with 25% homology to other synthetases (ATP and HSCoA binding proteins). No
homology to common structural proteins was observed.
4. The pattern of RNA expression detected in bovine tissues is consistent with known
protein expression patterns of ACS. Specifically, heart and mammary gland contain
significant amounts of ACS RNA while liver and kidney contain very little.
A restriction map of the 4.2 Kbp insert of pATC5 has been prepared and the insert
partially sequenced. To date, over 1200 bp have been sequenced, revealing a single long
open reading frame. Northern blots of poly(A+) RNA from bovine heart, liver, mam-
mary gland and kidney were probed with [alpha-
32
P]-ATC5. Northern analyses clearly
show multiple forms of ACS mRNA in most tissues. For example, three distinct
transcripts were detected in heart and mammary gland while two transcripts were
detected in kidney and one in liver. ACS transcripts range in size from 0.8 to 5.2 kb.
Multiple forms of ACS mRNA were not unexpected, based on results from other
synthetases, and may be the result of multiple promoters and/or alternative splicing. It
is curious, however, that a faint ACS mRNA (4.2 kb) would be present in liver where
no ACS activity can be detected. This suggests that either the liver transcript is not
translated, leads to translation of an inactive molecule, or is regulated at the post-trans-
lational level. It is also possible that our ACS probe hybridised with another closely
related mRNA (e.g. propionyl CoA synthetase).
210 R.M. Cook
The expression of ACS activity in ruminants is unique. ACS activity is constitutive in
heart, controlled by nutrition and physiological state in the mammary gland but not
expressed in liver (Marinez et al 1976; Mellenberger et al 1973; Ricks and Cook 1981a).
Ruminant liver is not lipogenic (Emery 1980). Therefore, acetate is not utilised by liver.
The liver utilises propionate as the major source of glucose (Amaral et al 1990; Reynolds
et al 1988). Since there is a paucity of glucose in ruminants, this pattern of tissue utilisation
of acetate spares glucose for other vital metabolic functions (Ricks and Cook 1981a).
ACS is the first committed step in acetate oxidation which is a major source of energy
for milk synthesis. Also, in the ruminant mammary gland, ACS is the first committed
step in fatty acid synthesis. The activity of ACS is marginal in a dry gland, increases to
peak lactation and then declines in activity as lactation advances (Marinez et al 1976).
In rat mammary gland, ACC is generally considered to be the first committed step in
fatty acid synthesis (Ponce-Castaneda et al 1991). Acetyl-CoA in rat mammary tissue
is from glucose via the ATP citrate lyase reaction and this source of acetyl-CoA is
generally considered to be relatively constant.
Since we find multiple forms of ACS mRNA, acetyl-CoAsynthetase could be regulated
by tissue-specific promoters. This would explain why the enzyme is active in heart (three
forms of mRNA) but not liver (one form of mRNA). Also, there may be an alternative
promoter in the mammary gland that is activated by nutrition and/or physiological state.
However, tissue specific alternative splicing mechanisms for acyl-CoA synthetases that
produce protein isoforms specific for individual fatty acids cannot be ruled out at this point.
Understanding regulation of ACS is critical because the reaction catalysed by this enzyme
is the initial step in fatty acid synthesis and acetate oxidation in ruminants.
It is important to target other  genes regulating  metabolism.  In ruminants,  this
endeavour is in its infancy (Bergen et al 1995). However, the genomic regulation of
protein and fat deposition promises to be the next advance in food producing animals.
The genes encoding fatty acid synthase and calpastatin are of current interest and a
knowledge of their regulation can provide better methods to regulate nutrient composi-
tion and tenderness of meat products.
In summary, there are several chemical methods available to regulate the rumen
ecosystem such that end-products of feed fermentation can be altered to improve growth
and fattening or milk production. A knowledge of the mechanism of action of these
chemicals at the molecular level will provide new and better methods to regulate the rumen
ecosystem. However, this is just one aspect of improving feed utilisation by ruminants. A
knowledge of utilisation of rumen fermentation products by extraruminal tissue is para-
mount if improved efficiency in food producing animals is to be achieved. This will require
detailed knowledge of the regulation of genes important in substrate utilisation. Also, the
availability of probes for these important genes will provide additional opportunities for
genetic modification of ruminants using marker-assisted selection programmes.
References
Allison M.J. and Bryant M.P. 1958. Volatile fatty acid growth factor for cellulolytic cocci of
bovine rumen. Science 129:474.
Application of molecular biology in the production of food from ruminants 211
Amaral D.M., Veenhuizen J.J., Drackley J.K., Cookey J.H., McGilliard A.D. and Young J.W.
1990. Metabolism of propionate, glucose and carbon dioxide as affected by exogenous
glucose in dairy cows at energy equilibrium. Journal of Dairy Science 73:1244.
Bergen W.G., Dickerson-Weber P.S., Mater M.K. and Kramer S.A. 1995. Approaches to gene
analysis, gene transfer into cells and gene targeting strategies. Journal of Animal Science (in
press).
Bergen W.G., Purser D.B. and Cline J.K. 1968. Determination of limiting amino acids of rumen
isolated microbial proteins fed to rats. Journal of Dairy Science 51:1698.
Bray J.C. and Till A.R. 1975. Metabolism of sulphur in the gastrointestinal tract. In: McDonald
I.W. and Warner A.I.C. (eds), Digestion and Metabolism in the Ruminant. University of New
England Publications Unit, Armidale, New South Wales, Australia.
Brondani A.V. 1983. ?Effects of the Antimicrobial Agent Teichomycin A2 on Rumen Fermen-
tation.? MSc thesis, Michigan State University, East Lansing, Michigan, USA.
Brondani A.V. 1986. ?A Study of the Action of Teichomycin A2, Avoparcin and Isoacids in
Sheep.? PhD thesis, Michigan State University, East Lansing, Michigan, USA.
Brondani A., Towns R., Barradas H., Chou K. and Cook R.M. 1991. Effects of isoacids, urea and
sulfur on rumen fermentation in sheep fed high fibre diets. Journal of Dairy Science 74:2724.
Bryant M.P. 1973. Nutritional requirements of the predominant rumen cellulolytic bacteria.
Federation Proceedings 32:1809.
Bryant M.P. and Doetsch R.N. 1955. Factors necessary for the growth of Bacteroides succino-
genes in the volatile acid fraction of the rumen fluid. Journal of Dairy Science 38:340?350.
Chalupa W. 1977. Manipulating rumen fermentation. Journal of Animal Science 45:585.
Chalupa W., Oppegard C., Williams H.C., Bloch B. and Perkins G. 1981. Effect of avoparcin
on rumen environment and fermentation. 73rd Annual Meeting of American Society of Animal
Science (Abstract). p. 387.
Cook R.M. 1966. Use of C
14
to study utilization of substrates in ruminants. Journal of Dairy
Science 49:1018?1023.
Cook R.M. 1970. Utilization of volatile fatty acids in ruminants. II. Comparison of the metabolism
of acetate, propionate and butyrate when injected into the jugular vein, portal vein or the rumen.
Biochimica et Biophysica Acta 201:91.
Cook R.M. and Miller L.D. 1965. Utilization of volatile fatty acids in ruminants. I. Removal of
VFA from portal blood by the liver. Journal of Dairy Science 48:1339.
Cook R.M. and Qureshi S. 1972. Utilization of volatile fatty acids in ruminants. 4. Relative
activity of acetyl CoA synthetase and acetyl CoA hydrolase in bovine mitochondria and
intracellular localization of acetyl CoA synthetase. Journal of Agricultural and Food Chem-
istry 20:91.
Cook R.M., Liu S.C. and Quraishi S. 1969. Utilization of volatile fatty acids in ruminants. III.
Comparison of mitochondrial acyl CoA synthetase activity and substrate specificity in
different tissues. Biochemistry 8:2966.
Cook R.M., Simon S. and Ricks C.A. 1975. Utilization of volatile fatty acids in ruminants. VI.
Purification of acetyl CoA from the mitochondria of lactating goat mammary gland. Journal
of Agricultural and Food Chemistry 23:561.
DeLay R.L., Zimmer P.R. and Simkins K.L. 1978. Effect of avoparcin on the performance of
feedlot cattle. 70th Annual Meeting of the American Society of Animal Science (Abstract). p.
414.
Emery R.S. 1980. Mobilization, turnover and disposition of adipose tissue lipids. In: Rucke-
buschY.andThivendP.(eds),Digestive Physiology and Metabolism in Ruminants. Avi
Publishing Co., Westport, Connecticut, USA. p. 541.
212 R.M. Cook
Felix A., Cook R.M. and Huber J.T. 1980. Isoacids and urea as a protein supplement for lactating
cows fed corn silage. Journal of Dairy Science 63:1103.
Froetschel M.A., Croom W.J. Jr., Gaskins H.R., Leonard E.S. and Whitacre M.D. 1983. Effects
of avoparcin on ruminal propionate production and amino acid degradation in sheep fed high
and low fibre diets. Journal of Nutrition 113:1355?1362.
Ghirardi G.G. and Cook R.M. 1987. Insulin and glucose plasma levels in cows producing 32
and20kgFCMinearlylactation.Journal of Dairy Science (Supplement 1):49.
Harrison D.G. and McAllan A.B. 1981. Factors affecting microbial growth yields in the
reticulo-rumen. In: Ruckebush Y. and Thivend P. (eds), Digestive Physiology and Metabolism
in Ruminants. AVI Publishing Co. Inc., Westport, Connecticut, USA.
Henderson C., Stewart C.S. and Nekrep F.V. 1981. The effect of monensin on pure and mixed
cultures of rumen bacteria. Journal of Applied Bacteriology 51:159.
Hume I.D. and Bird P.R. 1970. Synthesis of microbial protein in the rumen. IV. The influence
of the level and form of dietary sulphur. Australian Journal of Agricultural Research 21:315.
Isichei C.O. 1980. ?The Role of Monensin on Protein Metabolism in Steers.? PhD thesis,
Michigan State University, East Lansing, Michigan, USA.
Johnson R.J., Herlugson M.L., Ogikutu L.B., Cordova G., Dyer I.A., Zimmer P. and DeLay R.
1979. Effect of avoparcin and monensin on feedlot performance of beef cattle. Journal of
Animal Science 48:1338.
Knox K.L., Black A.L. and Kleiber M. 1967. Some kinetic characteristics of rumen short-chain
fatty acids as measured by the isotope dilution method. Journal of Dairy Science 50:1716.
Kone P., Machado P.F. and Cook R.M. 1989. Effect of the combination of monensin and isoacids
on rumen fermentation. Journal of Dairy Science 72:2767.
Kulasek G. 1972. A micromethod for determination of urea in plasma, whole blood and blood
cells using urease and phenol reagent. Pol. Arch. Wet. 15:801.
Kunstamann M.P., Mitscher L.A., Porter J.W., Shay A.J. and Darken M.A. 1968. LL-AV290, a
new antibiotic. I. Fermentation, isolation and characterization. Antimicrobial Agents and
Chemotherapy 1968:242.
Marinez D.I., Ricks C.A. and Cook R.M. 1976. Utilization of volatile fatty acids in ruminants.
VIII. Acetate activation by mammary tissue. Journal of Agricultural and Food Chemistry
24:927.
Mellenberger B.W., Bauman D.E. and Nelson D.R. 1973. Metabolic adaptations during lacto-
genesis. Fatty acid and lactose synthesis in cow mammary tissue. Biochemical Journal
136:741.
Miura H., Horiguchi M. and Matsumoto M. 1980. Nutritional interdependence among rumen
bacteria, Bacteroides amylophilus, Megasphaera elsdenii and Ruminococcus albus. Applied
and Environmental Microbiology 40:294.
Okuda H., Fujii S. and Kawashima Y. 1965. A direct colorimetric determination of blood
ammonia. Tokushima Journal of Experimental Medicine 12:11.
?rskov E.R. 1982. Protein Nutrition in Ruminants. Academic Press, New York, New York, USA.
160 pp.
Ottenstein D.M. and Bartley D.A. 1971a. Separation of free acids C2-C5 in dilute aqueous
solution column technology. Journal of Chromatographic Science 9:673.
Ottenstein D.M. and Bartley D.A. 1971b. Improved gas chromatography separation of free acids
C2-C5 in dilute solutions. Analytical Chemistry 43:952.
Parenti F., Beretta G., Berti M. and Arioli V. 1978. Teichomycins, a new antibiotic from
Actinoplanes teichomyceticus Nov. Sp. I. Description of the producer strain, fermentation
studies and biological properties. Journal of Antibiotics 31:276.
Application of molecular biology in the production of food from ruminants 213
Ponce-Castaneda M.V., Lopez-Casillas F. and Kim K.-H. 1991. Acetyl-coenzyme Acarboxylase
messenger ribonucleic acid metabolism in liver, adipose tissues and mammary glands during
pregnancy and lactation. Journal of Dairy Science 74:4013.
Quispe M.E., Barradas H. and Cook R.M. 1991. Effects of isoacids, urea and sulfur on ruminal
fermentation in sheep fed pineapple tops. Small Ruminant Research 6:49.
Qureshi S. and Cook R.M. 1975. Utilization of volatile fatty acids in ruminants. 5. Purification
of acetyl CoA from  the mitochondria of lactating bovine mammary gland. Journal of
Agricultural and Food Chemistry 23:555.
Raafat A.M. 1994. ?Cloning and Characterization of a Bovine Mammary Gland cDNAEncoding
Acetyl-CoA Synthetase.? PhD thesis, Michigan State University, East Lansing, Michigan,
USA.
Redin G.S. and Dornbush A.C. 1969. LL-AV290, a new antibiotic. II. Antibacterial efficacy in
mice in vitro. Antimicrobial Agents and Chemotherapy 1968:246.
Reynolds C.K., Huntington G.B., Tyrrell H.F. and Reynolds P.J. 1988. Net metabolism of
volatile fatty acids, D-beta-hydroxybutyrate, nonesterified fatty acids, and blood gases by
portal-drained viscera and liver of lactating Holstein cows. Journal of Dairy Science 71:2395.
Richardson L.F., Raun A.P., Potter E.L., Cooley C.O. and Rathmacher R.P. 1976. Effects of
monensin on rumen fermentation in vitro and in vivo. Journal of Animal Science 3:657.
Ricks C.A. and Cook R.M. 1981a. Regulation of volatile fatty acid uptake by mitochondrial
acyl CoA synthetases of bovine liver. Journal of Dairy Science 64:2324.
Ricks C.A. and Cook R.M. 1981b. Regulation of volatile fatty acid uptake by mitochondrial
acyl CoA synthetases of bovine heart. Journal of Dairy Science 64:2236.
Rogers J.A. and Davis C.L. 1982. Effects of intraruminal infusions of mineral salts on volatile
fatty acid production in steers fed high grain and high roughage diets. Journal of Dairy
Science 65:953.
Short D.E. 1978. ?Rumen Fermentation and Nitrogen Metabolism as Affected by Monensin.?
PhD thesis, University of Illinois, Urbana, USA.
Stamoudis V. and Cook R.M. 1975. Utilization of volatile fatty acids in ruminants. 7. Acetyl
CoA synthetase?a glycoprotein. Journal of Agricultural and Food Chemistry 23:563.
Wolin M.J. and Miller T.L. 1983. Interactions of microbial populations in cellulose fermentation.
Federation Proceedings 42:109.
214 R.M. Cook
Engineering the rumen for enhanced
animal productivity
J.D. Brooker
Department of Animal Science, University of Adelaide, Glen Osmond, 5064, Australia
Abstract
The rumen can be manipulated in several ways. Plasmid vectors that can
transform rumen bacterial species have been developed and some that shuttle
between different species have been described. One of these, pMU1328, has
been used to transform Streptococcus bovis and to form the basis of construc-
tion of a suicide vector for integration into the S. bovis chromosome. This
vector, pJDB9, replicates in Escherichia coli but not in S. bovis due to
removal of the ori-S. Insertion of a 0.47 kb fragment of Tn916 into pJDB9
generated pJDB0.5 that can integrate into its chromosomal Tn916 target
sequence in S. bovis and be maintained in a stable fashion. The amylase gene
promoter has been inserted in these plasmids to promote expression of
various genes, including ?-1,4-endoglucanase, xylanase, pea albumin and
tannin acyl hydrolase in S. bovis and related species.
Population control in the rumen has also been demonstrated. Transfer of
lactic acid utilising bacteria from grain-adapted animals to livestock just
prior to the introduction of a grain-based feeding regime reduced the potential
lactate concentrate from 100 mM to less than 10 mM and avoided the onset
of acidosis. Bacterial transfers from feral goats to domestic sheep and cattle
also demonstrated the viability of bacterial transfers in improving the N
balance of animals feeding on tannin-containing Acacia. Improved weight
gain and wool growth in sheep is reported. These examples demonstrate the
feasibility of rumen manipulation to improve productive performance of
domestic ruminants browsing fibrous forages.
Introduction
Limitations to ruminant production include poorly digestible fibrous plant material,
nitrogen losses and toxic compounds present in the feed. In tropical and arid regions,
these limitations can be so great as to become critical to the survival of ruminant
animals in that environment. The fibre content of many tropical grasses is high and
digestibility of the feed is low compared with improved Mediterranean-type pastures
(McMeniman  et  al 1981). Feed supplements are often unavailable or expensive.
Anti-nutritive components are also very prevalent in tropical plants and these restrict
the forage range for domestic livestock. Many of these limitations may be overcome,
or at least alleviated by manipulation of rumen microbial populations.
The rumen can be manipulated in two fundamental ways: firstly by the application
of genetic techniques to modify the functional capacity of specific bacterial species;
secondly, through modification of rumen ecology by introducing novel bacterial species
or by selectively enhancing populations of existing species.
Considerable advances have been made in developing techniques for the genetic
manipulation of rumen bacteria and the predominant species have been transformed, albeit
at low frequencies in some cases. Various plasmid vectors have been designed, and for
some species such as Streptococcus, recombination of genes into the chromosome of the
host has been demonstrated. Recombinant strains of rumen bacteria expressing various
foreign genes have been produced and some recombinant strains have been returned to
the rumen and have been shown to persist in the absence of direct selection.
Modification of microbial ecology by population control has also been demonstrated
to be an effective way to improve rumen function. Lactic acidosis in feedlot animals can
be controlled by ruminal inoculation of lactate-utilising bacteria. Tannin toxicity can be
alleviated and productivity of animals browsing tannin-containing tropical shrubs has been
enhanced by the cross inoculation of rumen fluid from wild to domestic ruminants.
There is now ample evidence that animal production, especially in marginal areas, can
be improved by manipulation of the rumen. Further work on the isolation of novel bacterial
species (Brooker et al 1994) and genes that may be transferred between bacterial species
will be valuable additions to studies on animal nutrition that will ultimately lead to
increased productivity in domestic ruminants in tropical and arid regions.
Methods and materials
Bacterial strains
Bacterial strains used were E. coli K12, strain DB11 (B. White, University of Illinois,
USA) and S. bovis strain WI-1 (a local ruminal isolate). Plasmid pMU1328 was obtained
from B. Davidson, Melbourne University, Australia. All subcloning and plasmid manipu-
lation steps were carried out using pUC19. E. coli was grown in LB medium. S. bovis
WI-1 was grown in nutrient medium (YTGS) containing 5 g yeast extract, 15 g trypticase,
10 g glucose and 15 g starch per litre of water. S. bovis WI-1::Tn916 was obtained by
conjugation between Enterococcus faecalis::Tn916 and S. bovis WI-1 (Brooker and Lum
1993). Transformation of E. coli and S. bovis was by electroporation using the Biorad
Gene Pulser under conditions previously described (Whitehead 1992) except that cells
were grown to A
660
of 0.8, washed in sterile distilled water and suspended in 10% glycerol.
Most genetic techniques are as described by Sambrook et al (1989).
Preparation of anaerobic medium
Anaerobic medium was prepared by boiling for 5 min then bubbling with O
2
-free CO
2
(passed over copper turnings at 300?C) for 15 min. Solutions were dispensed into
Hungate tubes or 100 ml crimp-top bottles under CO
2
before autoclaving. Media were
stored in a Coy anaerobic hood under an atmosphere of 95% CO
2
and  5%  H
2
.
216 J.D. Brooker
Virginiamycin (Stafac 500) was kindly supplied by Dr. J.B. Rowe (University of New
England, Australia).
Fermenter conditions
A Bioflow IIC microprocessor-controlled laboratory scale fermenter (New Brunswick
Scientific Co., Inc.) with a 600 ml working volume was used to develop an in vitro
model for acidosis. The pH was maintained at 6.0 by the automatic pumping of either
1 M HCl or 1 M KOH and the fermenter was sparged at a rate of 20 ml/min with CO
2
and N
2
gases in a ratio of 3:1. The temperature was maintained at 39?C. Cultures were
mixed continuously at 100 rpm for pure culture studies and at 50 rpm for crude rumen
fluid incubations.
Metabolism studies in rumen fluid-inoculated sheep
Twelve three- to four-year-old merino wethers and two female domestic angora goats
were treated with a broad-spectrum anthelmintic (Valbazen: Smith Kline and French,
Australia) and placed in individual metabolism cages. During the first seven days, the
animals were allowed to acclimatise to their surroundings and were fed 1 kg lucerne
pellets daily. Thereafter Acacia (mulga) was harvested daily from trees 4?6 m in height,
the leaves were separated as described by McMeniman et al (1981) and fed ad libitum
at 0800 h each morning.
Two days after commencing the mulga diet the first of four ten-day metabolism
studies (period 1) was conducted. Daily urine and faeces output were measured and
samples were collected and bulked for N analysis. Daily dry matter intake (DMI) was
measured and liveweight (LW) change was recorded periodically. The second metabo-
lism study (period 2) commenced 25 days after starting the mulga diet.
Results
Genetic manipulation in the rumen
Various vectors have been designed for rumen bacteria. Some of these are based on
endogenous  plasmids,  e.g.  pJDB216 for Selenomonas  ruminantium (Attwood and
Brooker 1992). Others are based on compatible plasmids from other bacterial species,
e.g. pBS42 from Bacillus subtilis that functions in Butyrivibrio fibrisolvens. Transfor-
mation systems have been developed for several ruminal species. This work has now
been extended to the development of a recombination system to insert foreign genes
directly into the host chromosome.
Development of an expression system for Streptococcus spp
To develop an expression vector capable of mediating expression of foreign genes in
Streptococcus spp, firstly we cloned the non-secreted ?-amylase gene from S. bovis into
Engineering the rumen for enhanced animal productivity 217
E. coli (DB11). The gene was isolated from a pUC19 plasmid library by screening for
starch degradation on nutrient agar plates using the iodine reaction to localise positive
colonies. The cloned DNA was mapped and 500 bp of the DNA was sequenced and
compared with previously published data. The gene corresponded to that recently
described in S. bovis by T.R. Whitehead (personal communication). Using PCR primers,
constructed so that a synthetic HindIII site was present on either end, the promoter
region and ribosomal binding site for the gene was synthesised. The promoter fragment
was cloned in pMU1328 and sequenced to verify its structure and orientation. This
fragment has been placed in the mcs of the suicide vector, pJDB9 to drive expression
of new genes inserted nearby. Genes to be expressed include a ?-1,4-endoglucanase,
xylanase, tannin acyl hydrolase and pea albumin.
Construction of vectors for recombination in S. bovis
Plasmid pMU1328 was digested with PstI to delete the streptococcal replicon (ori-S),
self ligated and electroporated into E. coli DB11. This plasmid, pJDB9, was isolated
from single recombinant colonies and mapped. The ori-S deletion was confirmed
although a minor rearrangement involving inversion of the ori-C-containing fragment
had also occurred (Figure 1). This rearrangement appeared to have no adverse effect on
subsequent manipulations.
To construct a vector capable of integration into the S. bovis WI-1::Tn916 chromo-
some, we inserted a PCR-generated 0.47 kb fragment of transposon Tn916 into the
multiple cloning site of pJDB9. This was achieved by cloning the tet M-containing 4.8
kb Hinc II fragment of Tn916 into the Sma I site of pUC19 to form pJDB4.8. A 0.47 kb
fragment of the tet M gene, between nucleotides 70 and 540, was PCR amplified using
primers synthesised from the published sequence (Burdett 1990) but with an EcoRI
restriction site added to each end. The fragment was digested with EcoRI and inserted
into the EcoRI site of pJDB9 to form pJDB0.5 (Figure 1).
To establish homologous recombination of pJDB0.5 in S. bovis WI-1::Tn916, cells
were electroporated with pJDB0.5 and the cells were then incubated for various times
on agar containing 5 mg/ml erythromycin. Plasmids pJDB9 and pMU1328 were
electroporated into S. bovis WI-1 cells as controls. S. bovis WI-1 cells were only
transformed by pMU1328. S. bovis WI-1::Tn916 cells were transformed by pJDB0.5
and pMU1328; pJDB9 failed to transform either host. Erythromycin resistant S. bovis
WI-1::Tn916 transformants generated with pJDB0.5 were tetracycline sensitive, con-
sistent with integration of pJDB0.5 into the tet M target site. pMU1328 transformants
of S. bovis WI-1::Tn916 were resistant to both erythromycin and tetracycline. The
frequency of recombination was estimated to be 3 ? 10
?8
/cell.
Analysis of S. bovis::Tn916 recombinants
To demonstrate chromosomal integration of pJDB0.5, total DNA from recombinant S.
bovis WI-1::Tn916 was digested with Bgl II or Pvu I, enzymes for which there are no
sites within pMU1328 or the tet M gene, and analysed by gel electrophoresis and
218 J.D. Brooker
Southern blot. A single high molecular weight (> 23 kb) band was observed that
hybridised with pMU1328-specific probe, and, after washing and reprobing, with the
tet M-specific  probe. Control S. bovis WI-1::Tn916 DNA hybridised with the tet
M-specific probe but not pMU1328.
To examine the organisation of the integrated DNA, chromosomal DNA from two S.
bovis WI-1::Tn916 recombinants, JDB-2 and JDB-3, was digested with HindIII, EcoRI or
HindIII + EcoRI. Fragments were analysed by gel electrophoresis and Southern blot using
radiolabelled pMU1328 as a probe. The results (Figure 2) show that in each digest, several
positive bands were detected, and in the single digests, one band in each track displayed a
stronger hybridisation signal than the other bands and was the same size as the original
linearised plasmid. The same pattern was observed for both recombinants tested.
Stability of recombinants
Due to regeneration of the Tn916 target sequence during homologous recombination of
pJDB0.5, spontaneous excision of the vector sequence may occur in the absence of
erythromycin selection. To assess the stability of recombinants, a culture of JDB-2 was
passaged for approximately 100 generations in the absence of erythromycin. Analysis
Figure 1. Derivation of suicide vector pJDB9 and recombination vector pJDB0.5 from pMU1328.
pJDB4.8
7.3 kb
pJDB9
6.0 kb
pJDB0.5
6.5 kb
pMU 1328
7.5 kb
mcs
Ori C
Ori S
EcoR1
Pst 1
Hind 111
Em
r
Pst 1Pst 1
Cm
r
Pst 1
EcoR1
Pst 1
Cm
r
Hind 111
Pst 1
Em
r
PCR
Am
r
0.47 kb fragment
Ori C
Em
r
Hind 111
Cm
r
4.8 kb Tet M-containing
fragment from Tn916
Pst 1
EcoR1
Pst 1
mcs
Ori C
Engineering the rumen for enhanced animal productivity 219
of total and erythromycin-resistant viable cell counts showed that even after extensive
passaging in the absence of selection, between 75 and 84% of total viable cells still
maintained the pMU1328 em
r
sequence.
Population control in the rumen
Bacterial transfer to control acidosis
Acidosis in ruminants is caused by rapid fermentation of cereal starch and the production
of excess lactic acid in the rumen. This can be avoided by slow adaptation to the cereal
diet. During adaptation to this diet Sel. ruminantium subsp lactilytica was the predomi-
nant lactic acid utilising bacterium present. We therefore tested the effectiveness of a
fresh isolate of this species to prevent the short-term development of acidosis when the
animals were acutely fed with wheat rations.
In control sheep, ruminal pH fell to 4.7 (P < 0.001) within 8 h of grain feeding and
remained around this value for the duration of the experiment. In contrast, sheep
inoculated with Sel. ruminantium subsp lactilytica strain JDB201 maintained a stable
ruminal pH which did not drop below pH 6.3. Total ruminal VFA in control animals fell
to less than 20 mM in 8 h and to less than 10 mM by 16 h. In inoculated animals, total
ruminal VFA increased significantly (P < 0.001) from 50 to 92 mM within 8 h, but this
value declined to 65 mM by 24 h. These changes in inoculated sheep were the result of
a twofold increase in propionate and a fourfold increase in butyrate compared with a
decrease in propionate and butyrate in control sheep (Table 1). In control sheep, ruminal
lactate increased to 109 mM within 8 h of grain engorgement whereas lactate was
undetectable in inoculated animals. Blood lactate concentrations were similar for both
groups of animals and varied between 1.7 and 2.6 mM.
Over the 24 h experimental period, sheep inoculated with Sel. ruminantium subsp
lactilytica strain JDB201 appeared to suffer no ill effect from acutely administered grain
whereas control sheep exhibited symptoms of physiological distress, diarrhoea, listless-
ness and lack of appetite. However, in treated sheep, if grain feeding was continued past
24 h without further bacterial treatment, symptoms of acidosis (distress, lack of appetite)
became evident. At the completion of the 24 h experiment, inoculated sheep immedi-
ately resumed normal feeding on the original lucerne diet, whereas control sheep took
another 3 days before they resumed feeding.
Ruminal inoculation with a bacterial mixture
Since Sel. ruminantium subsp lactilytica was only effective in controlling acidosis for
24 h, we tested a combination of Sel. ruminantium subsp lactilytica and Megasphaera
elsdenii. To help monitor the introduced bacteria in one of the three treated sheep, the
bacterial inoculum was contained within a dialysis bag with a pore size of approximately
10,000 Daltons. Control sheep were not tested beyond one day to avoid acute acidosis
in these animals.
Within one day of intraruminal administration of wheat slurry, ruminal pH in control
sheep decreased from 6.0 to 4.6. In contrast, all inoculated sheep maintained a ruminal
220 J.D. Brooker
pH (P < 0.05) greater than 5.5 throughout the four-day experiment. Grain administration
also significantly (P < 0.05) elevated the ruminal L-lactic acid concentration of control
sheep from 0 to 99.0 mM within 24 h. In inoculated sheep, L-lactic acid remained at
less than 1 mM for at least four days. Total VFA values were greater than 100 mM in
all animals before administration of grain. Within one day of grain feeding, total VFA
values in control sheep fell sharply to 21 mM. In animals inoculated with Sel. ruminan-
tium subsp lactilytica strain JDB201 and M. elsdenii strain JDB301, total VFAdecreased
transiently to 65?73 mM and then increased to greater than 80 mM.
In vitro model for lactic acidosis
In a continuous culture experiment in vitro, S. bovis was grown to produce lactic acid.
The effectiveness of Sel. ruminantium subsp lactilytica strain JDB201 and M. elsdenii
strain JDB 301 in controlling lactic acid accumulation was then tested, and contrasted
with the effect of administering 0.75 ?g/ml of virginiamycin.
In a control culture of S. bovis, when starch was added to the medium, lactic acid
concentrationincreasedrapidly from13mMto35mMwithin4handreached48mM
by8h.WhenanestablishedcultureofS. bovis was inoculated with Sel. ruminantium
subsp lactilytica strain JDB201, lactic acid increased to 60 mM within 24 hr. In contrast,
when an established culture of S. bovis was inoculated with M. elsdenii strain JDB301,
the lactic acid concentration was reduced by 65% to less than 20 mM. A combination
of Sel. ruminantium subsp lactilytica strain JDB201 and M. elsdenii strain JDB301 was
not as effective as M. elsdenii alone. In S. bovis cultures treated with 0.75 ?g/ml of
Table 1. Effect of Sel. ruminantium subsp lactilytica inoculation on individual VFA in wheat-fed sheep.
VFA Sampling time Animal treatments
(mM) (h) Control (SED)
?
Inoculated (SED)
?
Acetate 0 31.0 (3.65) 35.9 (3.21)
ns
8 10.7 (1.13) 53.8 (8.06)
?
16 5.1 (0.96) 42.6 (8.39)
?
24 10.7 (2.50) 31.7 (20.0)
?
Propionate 0 8.8 (0.92) 9.2 (1.03)
ns
8 3.4 (0.38) 21.4 (2.39)
??
16 0.7 (0.42) 21.1 (2.53)
??
24 0 17.0 (0.90)
Butyrate 0 3.2 (0.50) 3.4 (0.24)
ns
8 0.5 (0.03) 13.6 (1.95)
??
16 0.28 (0.05) 13.7 (1.13)
??
24 0.47 (0.04) 14.5 (1.36)
??
?
Mean (standard error mean of four replications).
?
Mean (standard error mean of three replications).
?
Significant (P < 0.05) difference.
?
Significant (P < 0.01) difference.
??
Highly significant (P < 0.001) difference.
ns
not significant.
Engineering the rumen for enhanced animal productivity 221
virginiamycin alone, lactic acid accumulation was prevented for up to 12 h, but by 24
h lactate levels had increased to the level of the untreated controls. In cultures treated
with Sel. ruminantium subsp lactilytica strain JDB201 plus virginiamycin or M. elsdenii
strain JDB301 plus virginiamycin, the concentration of lactic acid in the mixed cultures
was reduced by 50% and > 90% of untreated controls respectively and remained at these
levels for at least 24 h.
The combination of virginiamycin and M. elsdenii strain JDB301 was tested for its
ability to control lactic acid accumulation for up to three days in strained crude rumen
fluid cultures incubated in the presence of soluble starch. Starch was added 12 h before
the antibiotic/bacteria combination. At 24 h (12 h after introduction of virginiamycin
and bacteria), lactic acid in treated cultures was approximately 20% of that in the control.
However, by 36 h lactate had increased threefold and by 72 h was 30 mM compared
with 50 mM in the controls. When the combination of virginiamycin and M. elsdenii
strain JDB301 was introduced at 12 and again at 36 h, lactate concentrations remained
below 15 mM for the duration of the experiment. Surprisingly, a combination of Sel.
ruminantium subsp lactilytica strain JDB201, M. elsdenii strain JDB301 and vir-
giniamycin administered together at 12 and 36 h was the most effective treatment and
the concentration of lactate in the fermenter never exceeded 5 mM.
Bacterial transfer to control tannin toxicity in sheep
Feral goats thrive on a diet of Acacia (mulga) containing up to 14% by weight of
condensed tannin. Domestic ruminants do not do so well. Metabolism studies were
carried out on sheep inoculated with feral goat rumen fluid (FGRF) and fed a diet of
mulga. Crude protein and CT levels of mulga leaf for the four measurement periods in
this experiment were 142, 139, 146 and 141 g/kg and 139, 124, 120, 120  g/kg,
respectively.
During the dietary adaptation stage of the experiment, N digestibility was signifi-
cantly less efficient (P < 0.05) during period 2 than during period 1. No differences were
apparent in DMI, N balance or DMD. Similarly, there were no differences between the
two periods with any of the parameters measured in the domestic goats.
Following administration of FGRF, inoculated sheep had measurably higher DMI, N
balance and N digestibility values in both post-inoculation periods compared with the
pre-inoculation periods (Table 2). N digestibility was significantly higher during the ten
days immediately following inoculation (period 3 ver. period 2). This increase was
sustained in period 4 and was accompanied by significant increases in DMI and N
balance compared to periods 2 and 3. Weight loss was reduced following inoculation.
Dry  matter digestibility  did  not change significantly during dietary adaptation or
following FGRF inoculation. Over the test period, the untreated sheep lost approxi-
mately 40% more weight than the treated sheep. This provided inoculated sheep with a
weight advantage of approximately 2.5 kg over the uninoculated sheep at the end of the
experimental period.
Domestic goats inoculated with FGRF  also  exhibited  increased N digestibility
(P < 0.05) and DMI. N balance increased six-fold. Mean DMI, N balance and N
222 J.D. Brooker
digestibility for feral goats consuming mulga were 496 (SED = 60.8) g/day, 0.75
(SED = 0.580) g/day and 544 (SED = 62.9) g/kg respectively. For both sheep and goats,
in vitro rumen fluid gas production increased following inoculation. This increase was
significant (P < 0.05) for sheep.
Wool growth in mulga-fed sheep
Both inoculated groups of sheep grew more wool per unit area than the uninoculated
group for three of the four periods. However, these differences were not significant.
Providing a mineral supplement to the inoculated sheep produced a slight non-signifi-
cant beneficial effect, reducing the rate of wool growth decline. Linear wool growth did
not differ significantly between treatments.
Identification of bacteria in FGRF
Fractionation of FGRF into individual bacterial colonies on tannic acid containing plates
revealed relatively few different species. One of these, subsequently named S. caprinus,
was capable of growth in up to 3% w/v tannic acid and promotes the hydrolysis of
protein-tannin complexes in vitro. The related species, S. bovis was sensitive to less than
0.5% w/v tannic acid. S. caprinus is now known to be widespread amongst animals
browsing tannin-rich feed, but not to be normally present in domestic livestock. Another
bacterium, possibly a selenomonad, expressed tannin acyl hydrolase activity and could
utilise condensed tannin or tannic acid as a sole carbon source. We are currently cloning
the tannin acyl hydrolase gene for possible transfer and expression in other rumen bacteria.
Discussion
Genetic engineering of rumen bacteria
Vectors for transformation and recombination in several different species of rumen
bacteria have been designed and tested. A plasmid vector for gene expression in S. bovis
has been developed, using the promoter sequence from ?-amylase to drive new genes
Table 2. Effect of crude goat rumen fluid on sheep consuming mulga.
Pre-inoculation Post inoculation
Parameter Period 1 Period 2 Period 3 Period 4 s.e.m.
DMI (g/day) 510
a
524
a
535
a
611
b
17.1
NBAL (g/day) ?0.61
a
?0.55
a
0.51
ab
1.04
b
0.4
ND (g/kg) 450
a
417
b
450
a
463
a
8.8
DMD (g/kg) 440 452 443 445 11.4
in vitro GP (?l/48 h) nm 890
a
nm 1483
b
108.7
LW change (g/day) ?97
a
?47
b
13.9
Within rows, means with different superscripts differ significantly (P < 0.05).
nm?not measured.
Engineering the rumen for enhanced animal productivity 223
ligated nearby. We have also described the construction of a suicide vector that integrates
in a site-specific manner into S. bovis WI-1::Tn916. Recombination into the S. bovis
chromosome was only observed when pJDB0.5 was introduced into S. bovis WI-1::Tn916.
As no erythromycin resistant transformants were obtained with S. bovis WI-1, it can be
concluded that the plasmid is unable to replicate in this host and requires the presence of
the chromosomal copy of Tn916 to provide a site for homologous recombination. Screen-
ing is unequivocal because integration inactivates the host tet M gene, as well as permitting
expression of the vector-borne antibiotic resistance marker. The patterns from the Southern
blot indicate that twp copies of the plasmid were integrated into the chromosome. Since
there were two copies of the transposon integrated (Brooker and Lum 1993), this could
mean that each transposon sequence was interrupted by integration of the plasmid. This
result is also in keeping with the loss of tetracycline resistance.
Figure 2. Analysis of pJDB0.5 integration into S. bovis WI-1::Tn916. Chromosomal DNA from strains
JDB-2 and JDB-3 was digested with HindIII (lanes 1 and 4), EcoRI (lanes 2 and 5), or HindIII + EcoRI
(lanes 3 and 6) and analysed by agarose gel electrophoresis and Southern blot. Hybridisation was with a
32
P-dCTP labelled pMU1328 DNA.
224 J.D. Brooker
These results clearly demonstrate that expression of foreign genes in rumen bacteria
can now be achieved for some species, and is likely to be achieved for others in due
course. Limitations such as restriction barriers, nuclease activity and extracellular
polysaccharide can be overcome. It is now possible to examine the effect of introducing
and over-expressing genes such as cellulases, tannases and proteases on microbial
ecology, competitiveness and function in the rumen.
Population control in the rumen
The introduction of microbial populations into the rumen can effect animal productive
performance. Ruminal inoculation with lactic acid utilising bacterial species has been
shown to prevent accumulation of lactic acid and thereby reduce the possibility of
acidosis occurring in acutely grain-fed ruminants. The effect was sustainable for at least
four days and suggests that this may provide an alternate treatment for acidosis than use
of ionophores or antibiotics. The results show that at least two different bacterial species
were required. This is a consequence of different metabolic directions. Under appropri-
ate conditions, S. ruminantium grows rapidly but ferments glucose to produce lactic
acid. M. elsdenii utilises lactic acid although it is slower growing. The combination of
bacteria is required to maintain a reasonable growth rate and deal with lactic acid as it
is synthesised.
Condensed tannins inhibit the viability and growth of many different species of
microorganism (McLeod 1974). Evidence now suggests that microorganisms present
in crude rumen samples from tannin-adapted animals are transferable to non-adapted
ruminants to tolerate the tannins, provide microbial biomass and improve digestibility
of tannin-containing forage. The results from these experiments also indicate that one
organism is not sufficient and that a consortium of several different microbial species
is involved in vivo.
Overall, these results support the view that effective manipulation of the rumen may
be achieved by genetic engineering, but population control is a practical and effective
alternative. This approach has certain advantages over the genetic route, including the
ability to utilise the abilities of a consortium of bacteria rather than having to rely on an
individual species. In the two cases above, effective control of an anti-nutritive situation
was only achieved by a mixed bacterial population.
In contrast, genetic manipulation of individual species does provide opportunities for
the substantial enhancement of bacterial activity and the ready transfer of the technology
to other ruminants. There are advantages and disadvantages of both approaches. Notwith-
standing this, manipulation of the rumen ecosystem is achievable by various routes, and
will hopefully lead to increased ruminant productivity, particularly in marginal areas.
Acknowledgements
Parts of the work reported here were carried out by S. Miller, D. Lum, I. Skene, L.
O?Donovan, K. Wiryawan and J. Durham, supported by grants from the Australian Wool
Corporation, ARC and AIDAB.
Engineering the rumen for enhanced animal productivity 225
References
Attwood G.T. and Brooker J.D. 1992. Complete nucleotide sequence of a Selenomonas rumi-
nantium plasmid and definition of a region necessary for its replication in Escherichia coli.
Plasmid 28:123?129.
Brooker J.D. and Lum D.K. 1993. Use of transposon Tn916 as a genetic marker in the rumen.
Letters in Applied Microbiology 17:224?227.
Brooker J.D., O?Donovan L.A., Skene I., Clarke K., Blackall L. and Muslera P. 1994. Strepto-
coccus caprinus sp nov, a tannin-resistant ruminal bacterium from feral goats. Letters in
Applied Microbiology 18:313?318.
Burdett V. 1990. Nucleotide sequence of the tet (M) gene of Tn916. Nucleic Acids Research
18:6137.
McLeod M.N. 1974. Plant tannins?their role in forage quality. Nutrition Abstracts and Reviews
44:803?15.
McMeniman N.P., Niven D.R., Crowther D., Gartner R.J.W. and Murphy G.M. 1981. Studies
on  the supplementary  feeding of sheep consuming mulga (Acacia aneura). Australian
Journal of Experimental Agriculture and Animal Husbandry 21:404?409.
Sambrook J., Fritsch E.F. and Maniatis T. 1989. Molecular Cloning: A Laboratory Manual. Cold
Spring Harbor Laboratory, New York, USA. 1546 pp.
Whitehead T.R. 1992. Genetic transformation of the ruminal bacteria Butyrivibrio fibrisolvens
and Streptococcus bovis by electroporation. Letters in Applied Microbiology 15:186?189.
226 J.D. Brooker
Genetic manipulation of rumen
bacteria: now a reality
K. Gregg
1
,D.Schafer
2
,C.Cooper
2
and G. Allen
2
1. Institute of Biology, Armidale NSW 2350 Australia
2. Institute of Biotechnology, University of New England, Armidale 2351 Australia
Abstract
Techniques are now available for addition of novel genes to rumen bacteria,
using plasmid vectors and high-voltage electroporation. Experiments have
shown that Butyrivibrio fibrisolvens strain AR10 isolated from sheep can
effectively colonise cattle on unimproved pastures. Strains of B. fibrisolvens
isolated from reindeer (E14) and Canadian white-tail deer (OB156) have also
been shown to colonise sheep and cattle. Bacterial populations fluctuate in
number in different samples from the same animal. Tracking of four separate
bacterial strains in the same two sheep has confirmed that individual strains
vary in number independently of the others within any  single sample.
Bacterial numbers were observed during changes of diet, including a week
in which the sheep were fed ground wheat straw. All four strains remained
present during the 149 days of the experiment.
Strain OB156 of B. fibrisolvens was genetically modified by insertion of
a non-transmissable plasmid containing the gene for fluoroacetate dehalo-
genase. The modified bacterium retained the new gene after 500 generations
of non-selective in vitro growth. A growing population of modified bacteria
was able to detoxify 10 nmol fluoroacetate per minute per mg of bacterial
protein. This is calculated to be a level capable of offering some protection
against the toxin for the host ruminant. Over the 149 days of the experiment,
the genetically modified OB156 was detectable in most samples at levels
between 10
5
?10
7
cells/ml, occasionally dropping to below detectable levels
(~10
3
cells/ml). There was no evidence of loss of the plasmid during major
population changes, or of its transfer to other organisms. These data indicate
that many of the reservations expressed about rumen bacterial genetic
manipulation do not necessarily apply in practice. Expectations for the
application of this technology are now considerably more confident. Equally
important, the molecular techniques involved in these studies have applica-
tion to a variety of ruminal ecology studies.
Introduction
Biotechnology is finding its way into many areas of agriculture and animal production
through feed and crop plants, through the genetic pool of animals, and through manipu-
lation of microbes that influence the growth and development of both animals and
plants. The direct genetic manipulation of microbes that inhabit the rumen has been
proposed as a possible mechanism for enhancing animal production, although progress
in this area has been slow.  Nevertheless,  the application of molecular  biological
techniques is beginning to contribute to our understanding of the rumen through the
precise methods now available for making in vivo observations. Molecular techniques
now allow individual micro-organisms to be tracked in the rumen, identification of
organisms by genetic criteria is rapidly becoming an essential part of taxonomy, and
?tagging? bacteria with genetic markers has greatly facilitated studies of their popula-
tions in vivo (Gregg et al 1993).
As studies of ruminal ecology become more precise, the contribution of biotechnol-
ogy methods to animal production will become continually more significant. However,
in the specific area of rumen bacterial genetic manipulation, the most significant
developments  are very recent and have so  far made  little impact  upon how this
technology is regarded. There has been considerable reservation about the contributions
that can be made by manipulating the genetics of rumen bacteria (Armstrong and Gilbert
1985; Egan et al 1992) and the basis for this reservation includes at least four areas of
concern.
1. Genetic manipulation of rumen bacteria has not been demonstrated as a reproduc-
ible and reliable technique.
2. Bacteria grown in the laboratory are unlikely to retain their competitive features
and are unlikely to return successfully to the rumen.
3. Genetic changes to bacteria are likely to impair their competitive fitness, particu-
larly if the changes do not directly benefit the bacterium itself.
4. Addition of an artificially manipulated bacterium to the rumen may upset the
balance of the rumen, with detrimental effects on the host animal.
In the absence of data to the contrary, these are all reasonable reservations. However,
data are now being accumulated to indicate that, at least in some cases, these difficulties
may have been overestimated.
Gene transfer systems for rumen bacteria
There are now several systems by which new genetic material can be introduced into
rumen bacteria as plasmids (e.g. Beard et al 1995; Cocconcelli et al 1992; Thomson et
al 1992). During the past ten years many different approaches have been made to the
transformation of rumen bacteria, and most have shown very limited success. The
technique that has proven generally useful for inserting new DNA is that of electropo-
ration. This process uses an electrical pulse to generate openings in the cell wall, through
which DNA can enter (Dower et al 1988). Combining this with plasmids originating
from the bacterial species to be modified has led to at least two of the more successful
processes reported (Beard et al 1995; Thomson et al 1992).
Methods that successfully achieve plasmid insertion do not necessarily provide
evidence of successful genetic manipulation because there remain complications in
making and transferring plasmid/gene constructs. It has been our experience that the
228 K. Gregg, D. Schafer, C. Cooper and G. Allen
feasibility of adding any particular new gene to a bacterium can only be assessed
empirically, even with a stable and efficient vector system. As a primary example, the
usefulness and stability of the transformation system described by Beard et al (1995)
for inserting genes of practical benefit into B. fibrisolvens is described below. Some
aspects of that example demonstrate unexpected contradictions of the results predicted
from basic biological and ecological principles.
Recolonisation of the rumen by laboratory-grown bacteria
The ability of B. fibrisolvens strain AR10 to colonise the rumen after several years
storage in a laboratory has been described previously (Gregg et al 1993). Strain AR10
was isolated originally from a sheep and had been shown, through a survey of 71 animals
from 14  grazing properties, to be absent from the  Central  Queensland region of
Australia. Figure 1 shows data from an experiment in which 5 cows out of a herd of 40,
on unimproved tropical pastures at  one of the properties  previously  tested,  were
inoculated with strain AR10. There was rapid spread through the herd (19 out of the 24
animals tested showing easily detectable levels of AR10 after 28 days post-inoculation).
Figure 1. The proportion of pasture-fed cattle showing the presence of Butyrivibrio fibrisolvens AR10 at
various times after inoculation. At days 7, 14 and 168, only the 5 inoculated animals were tested. At days
28 and 70, the same 5 animals were tested together with 20 others chosen at random from the herd. At day
268, 33 animals were tested and at day 404, a total of 45 animals were tested, some of which were small
calves at the beginning of the experiment, with 13 showing positive detection for AR10. The vertical lines
show rainfall in millimeters.
B. fibrisolvens AR10
In Cattle at Pasture
Genetic manipulation of rumen bacteria: now a reality 229
The conclusion from this work is that bacteria isolated and cultured in the laboratory
can retain their ability to cope with rumen conditions.
A more important test for recolonisation of the rumen is to examine the ability of a
bacterium to which genetic changes have been made in the laboratory to return
successfully to the rumen. Strain E14 of B. fibrisolvens was obtained from a reindeer
and provided to our laboratory by Dr C. Orpin. This strain had been ?tagged? by the
addition of a tetracycline resistance gene using transposon Tn916 by Dr J. Brooker
(Waite Institute, South Australia). E14 was inoculated into the rumens of four sheep in
the UNE animal house and rumen samples were tested over a period of seven weeks by
plating dilutions on selective plates. Colonies that grew on the plates were tested by
DNA hybridisation, at high stringency, with genomic DNA from E14. The levels
observed in individual animals fluctuated significantly, sometimes falling below detect-
able levels (10
3
cells/ml rumen fluid). On average, however, E14 was detectable in all
four animals at an average density between 10
6
?10
7
cells/ml (Figure 2).
The fluctuation of bacterial populations from day to day was observed in both of these
experiments, but it was not possible to be certain whether it arose from genuine
population changes or from sampling variability.
Figure 2. Persistence of Butyrivibrio fibrisolvens E14 in the sheep rumen after transposon modification to
impart tetracyline resistance. The addition of Tn916 required the bacterium to express a foreign gene from
which it obtained no benefit in vivo. Modified E14 recovered from the rumen was plated on agar containing
rumen-fluid medium and tetracycline, and the identity of colonies obtained was confirmed by hybridisation
to chromosomal DNA of E14 at high stringency. In each animal, E14 became undetectable at various times.
The line shows the mean value obtained from the four sheep.
01020304050
Days Post-Inoculation
0
1
2
3
4
5
6
7
8
Cells
per
ml
Rumen
Fluid
(log)
B. fibrisolvens E 14
Colonization of the Sheep Rumen
230 K. Gregg, D. Schafer, C. Cooper and G. Allen
Bacterial growth efficiency and the effect of an extra genetic load
Experiments described by Russell and Wilson (1988) indicated that bacteria carrying a
plasmid grew approximately 30% slower than the same strain without the plasmid. It
can be reasonably proposed that sustaining an additional genetic load could impede the
population growth of an organism, especially under conditions in which competition
for nutrients is high. However, experiments comparing the population growth of
bacteria with and without a multicopy plasmid have demonstrated that harbouring a
novel plasmid that has no survival benefit to the bacterium need not necessarily reduce
its population growth (Gregg et al 1993; Figure 3). It is important to note, however, that
this experiment was performed in vitro and did not address the question of direct
competition for nutrients against other, non-plasmid-bearing strains.
Addition of new organisms and the ecological
balance of the rumen
The concern that a modified bacterium could upset the ruminal balance, leading to other
(undefined) problems, may be allayed somewhat by recent findings in rumen bacterial
Figure 3.
Population growth in vitro of Prevotella ruminicola AR20 in the presence (B6c) and absence (o)of
plasmid pBA. Plasmid-bearing bacteria were grown in the presence of clindamycin (10 mg/ml) to ensure
expression of the foreign gene. Supporting plasmid replication and clindamycin resistance gene expression
did not cause any detectable decrease in growth efficiency.
51015 20 25
Growth time (hours)
0.0
0.5
1.0
1.5
2.0
Optical
density; 550
n
m
Genetic manipulation of rumen bacteria: now a reality 231
taxonomy. It has been shown that phenotypically defined ?species? may actually com-
prise multiple, distantly related, or even unrelated, genetic groups. Such genetic diver-
sity has been demonstrated for B. fibrisolvens (Hudman and Gregg 1989; Mannarelli
1988), Prevotella ruminicola (Hudman and Gregg 1989) and Ruminococcus albus
(Ware et al 1989). Thus it has been known for some time that, although a ?species? such
as B. fibrisolvens may represent 10% of the ruminal population, any individual ?strain?
is more likely to represent 0.1?1.0%. This is consistent with the population levels
observed for strains that have been specifically tracked in the rumen (Gregg et al 1993).
The prospect of replacing, for example, 10?20% of the rumen population with a
genetically modified organism is bound to raise suggestions of major ruminal disrup-
tion. However, it is now clear that modification of one or a few rumen bacteria is not
likely to effect this proportion of the total microbial population. It now appears that to
alter 10% of the ruminal population will require the manipulation of many individual
genetic strains. While providing a possible ameliorating factor for one aspect of this
technology, this generates a perhaps equally difficult question, i.e. whether the relatively
small contribution made by a single bacterial strain can provide significant biochemical
effect for a modified bacterium to influence the biology of the host animal. Discussion
on this point (Gregg and Ware 1990) has offered the suggestion that the effectiveness
of genetic manipulation will depend to some extent upon the nature of the changes that
are attempted.
Therefore, some of the reservations about rumen bacterial manipulation have been
addressed and the evidence indicates that the system need not be constrained by these
factors, although it remains likely that they will have relevance to a proportion of cases.
The work described here addresses some of the final reservations, constituting a model
system for rumen bacterial genetic manipulation.
Genetic manipulation of rumen bacteria?an example
Our recent work has been directed towards practical goals for solving problems in
ruminant production. In doing so, it has also addressed the questions of how stable
genetic modifications to rumen bacteria may be, and how the ability of the bacterium
to survive ruminal competition may be affected. These questions became very important
in a project that has attempted to protect ruminants against poisoning by fluoroacetate.
Background to the problem
The compound monofluoroacetate is highly toxic to aerobic organisms, being con-
verted metabolically to fluorocitrate which blocks the action of the enzyme aconitase
(Elliot and Kalnitsky 1950). This toxin occurs naturally in at least 40 species of trees
and shrubs in Australia and for many years has been known to be a problem in Africa
(Marais 1943) and Central America (DeOliveira 1963). Livestock may be killed in
large numbers by the toxic trees, depending to some extent upon seasonal conditions
(McCosker 1989).
232 K. Gregg, D. Schafer, C. Cooper and G. Allen
The problem of livestock poisoning occurs despite the existence of widespread
capabilities for microbial degradation of the toxin. Many soil organisms are capable of
defluorinating the compound to produce glycolate, and of using this as a carbohydrate
nutrient. Perhaps because the toxin acts upon aerobic metabolism and because it has
poor nutritive value under anaerobic conditions, the microbes of the rumen do not appear
to have acquired the same defluorinating capability as their counterparts in the soil
(Gregg and Sharpe 1991).
The soil organism Moraxella spp. strain B has high activity for dehalogenating
fluorinated and chlorinated compounds (Kawasaki et al 1981). A gene encoding the
enzyme fluoroacetate dehalogenase (haloacetate halidohydrolase, EC 3.8.1.3) has been
cloned and sequenced (Kawasaki et al 1984, 1992). We repeated the cloning and
sequencing after obtaining the bacterium from Prof H. Kawasaki (Osaka University) to
examine the DNA regions responsible for regulation of the gene. It was shown that the
dehalogenase gene was the second open reading frame (ORF2) in an operon, with the
gene promoter approximately 1.4 kilobase pairs (kb) upstream (Figure 4). Removal of
ORF1 and attachment of a different gene promoter demonstrated that ORF2 was
sufficient for fluoroacetate detoxification.
Experimental approach and results
The dehalogenase gene (H1) was isolated and spliced to a gene promoter from the
erythromycin resistance gene (erm
r
) of broad host range plasmid pAM?1 (LeBlanc and
Lee 1984) using polymerase chain reaction (Yulov and Zabara 1990). The chimaeric
gene was then attached to plasmid pBHerm for transfer to B. fibrisolvens strain OB156
(Figure 5). The erm
r
gene within plasmid pBHerm allowed transformed bacteria to be
selected on culture plates containing erythromycin (Gregg et al 1994).
The genetically modified OB156 was shown to produce dehalogenase enzyme with
activity of 10 nmol/min per milligramme of bacterial protein, which was calculated to
be sufficient to make a significant decrease in toxicity to the host ruminant, provided
the bacterium survived return to the rumen.
Figure 4. Diagram of the dehalogenase gene operon from Moraxella spp. The function of ORF1 has not
been determined, but its removal does not affect dehalogenase expression.
TSI =transcription initiation site; Restriction site: A: AvaI; B: BsshII E: EcoRI; H: HpaI; P: PstI; S: SalI; Sa: SacI;
Sp: SphI.
0 1 233.6
E
S H
A
Sa S
Sp
B
PA Sa E
TIS
ORF1 ORF2
Kilobase pairs
TSI = transcription initiation site; Restriction site: A: AvaI; B: BsshII E: EcoRI; H: HpaI; P: PstI; S: SalI; Sa: SacI;
Sp: SphI.
Genetic manipulation of rumen bacteria: now a reality 233
Figure 5. Construction of the chimaeric shuttle plasmid containing the dehalogenase gene. The dehalo-
genase gene was joined to the promoter of the erm
r
gene promoter of plasmid pAM?1 to ensure expression
in B. fibrisolvens.
OVERLAP
PRIMER
P2
PROMOTER erm
r
GENE
P1
PRODUCTS MIXED
MELTED AND
ANNEALED
PCR
DEHAL. OPERON ORF2
P3
P4
P1
P4
PCR
PROMOTER/DEHAL CLONED IN pUC18
EXCISED WITH HindIII HALF END-FILLED
LIGATED
PROM/DEHAL.
pBHF
10.3 kb
pRJF1
pUC118
erm
r
CLEAVED
WITH Xbal
HALF END-FILLED
PbhERM
9.3 kb
PCR
234 K. Gregg, D. Schafer, C. Cooper and G. Allen
The stability of this plasmid-bearing organism was tested by growth for 500 genera-
tions in culture medium without the selective antibiotic. After each 100 generations,
cells were plated onto duplicate selective and non-selective plates. No difference was
observed between bacterial numbers on the two media and it was concluded that no loss
of plasmid occurred within the cultures.
To examine the competitive ability of this organism in the rumen, it was inoculated
into the rumens of two sheep, together with an unmodified strain of B. fibrisolvens
(AR10) and two strains of P. r u m i n i c o l a (AR20 and AR29). Over the following 149
days, the animal?s diet was changed several times to encourage competition between
rumen microbes and samples were removed for estimation of bacterial numbers.
Bacteria were detected by polymerase chain reaction amplification of part of their
chromosome (Figure 6) and densitometer scanning gel photographs to quantitate PCR
product formation (Figure 7). Plasmid within the modified strain was also tracked by
PCR of plasmid sequences.
Figure 6. Agarose gel showing PCR product from tracking Butyrivibrio fibrisolvens OB156. The vertical
row of numbers shows the days at which samples were taken and the diet was changed as indicated on the
right of the figure, to encourage competition between bacteria for nutrients.
DIET
Oat Chaff
+
luceme
Oat chaff
Wheat straw
Oat chaff
+
luceme
Luceme
pDS63
(10
6
/ml)
AR10 B. fibrisolvens
G8 G9
SHEEP
Genetic manipulation of rumen bacteria: now a reality 235
The results of this test showed:
1. That all four strains recolonised the rumen and remained present for the full test
period. This confirmed that the ability to recolonise the rumen is common in
laboratory strains and that the genetically modified bacterium was not significantly
disadvantaged by its modification.
2. The levels detected for each organism fluctuated independently of the others,
indicating that variation in bacterial numbers seen in previous experiments was
probably genuine fluctuation, and not the result of sampling variability.
3. The plasmid detection reactions showed that levels of plasmid and host bacterium
co-varied, indicating that the plasmid was not transferred to other organisms at any
detectable level. This was predicted, since pBHerm is a non-conjugative plasmid.
The results also indicate that there was no significant loss of plasmid from the
OB156 population.
4. Plasmid was extracted from bacteria after they were recovered from the rumen and
was shown to be structurally unchanged.
Discussion
This work proved that the genetic manipulation of rumen bacteria is feasible. Most
importantly, the altered bacteria were shown to be competitive in the rumen, colonising
to the same extent as three other strains that were tracked over the same period. Diet
was shown to influence population levels, as expected, and the competition for
nutrients under poor dietary conditions did not appear to disadvantage the modified
bacterium.
Figure 7. Tracking genetically modified Butyrivibrio fibrisolvens OB156 in the sheep rumen, in parallel
with an unmodified strain AR10 and two unmodified strains of Prevotella ruminicola (AR20 and AR29).
Polymerase chain reaction products were separated by agarose-gel electrophoresis and quantitated by
densitometer scanning the negative of a photograph from the ethidium bromide-stained gel.
236 K. Gregg, D. Schafer, C. Cooper and G. Allen
Examination of the sheep during the trial period by a veterinarian confirmed that there
was no change in the animals? health, apart from the predicted effects of feeding a diet
of ground wheat straw for about a week.
It is now clear that rumen bacteria can be genetically modified; the modification can
be stable even when carried on a plasmid; carrying plasmid does not necessarily
disadvantage the bacterium energetically; laboratory-grown bacteria can recolonise the
rumen; and genetically modified bacteria can do so as efficiently as their unmodified
counterparts.
These observations will not necessarily apply to other genetically modified organisms
that may be constructed. Each example must be tested individually before these features
can be assumed. However, this work has shown that predictions based upon basic
biological principles cannot substitute for testing the case empirically.
It is interesting that this modification appears stable when the dehalogenase gene
confers no known advantage to the bacterium. Rationally, this represents a ?most likely
to be lost? case for genetic manipulation. It appears, therefore, that the need for a
modification to be useful to the modified organism is secondary in importance to the
inherent stability of the plasmid or whichever other gene transfer system is used. There
is currently no proven explanation for the ability of a bacterium to stably maintain a
genetic feature from which it gains no direct benefit. However, we can propose that this
ability stems from the fact that each bacterium maintains about 1000 genes, some of
which are constitutively expressed. During the natural changes of conditions that occur
within the rumen, some of these genes will inevitably be of no immediate benefit to the
organism at one stage or another. If there were serious disadvantages to the bacterium
in this situation, then many strains could be lost from the rumen during relatively minor
changes in ruminal conditions. Our dietary change tests were for limited periods only,
but indicate that the loss of bacterial strains that could be caused by such change may
be more likely to occur when dietary conditions vary over much longer periods.
Biotechnology in developing countries
While the efficiency and viability of ruminant production is important to the economies
of some industrialised countries, the importance for developing nations with less
overseas purchasing power are more far-reaching. Where ruminants fill essential needs
for  food, fibre  and draught power, improvements in production and  maintenance
efficiency contribute directly to the living standards of a high proportion of the
population. Much of the nutritional improvement possible through supplementation or
processing of animal feed may be difficult to implement in these situations. It is vital,
therefore, that improvements in ruminant production should make maximum use of the
feed sources currently or potentially available, rather than depend on adding new
elements that need to be purchased.
Where a high proportion of animal feed is the residue from grain or other staple crops,
digestibility is frequently the limiting factor in ruminant production. In those cases
where cash crops (e.g. cocoa) leave residues with feed potential, there may be serious
limitations created by the secondary compounds present in that material. There may be
Genetic manipulation of rumen bacteria: now a reality 237
considerable value, therefore, in systems that help in fibre breakdown or that detoxify
anti-nutritive factors.
The capability of genetic manipulation to make rumen microbes digest plant fibres
more efficiently remains to be proven. For this aim, there may be greater and more
immediate gain in examining how different feed plants can be combined to supplement
each other and optimise microbial growth and nutrient yield within the rumen. Microbes
isolated from the fermentative organs of non-domestic animals (e.g. giraffes, camels,
leaf-eating monkeys, etc.) may be directly adaptable if they are able to colonise domestic
ruminants. In our laboratory we have shown that bacteria from sheep (strain AR10) will
successfully colonise cattle. A strain of B. fibrisolvens from reindeer (strain E14) has
successfully colonised both sheep and cattle and OB156 from a Canadian white-tail
deer successfully colonised sheep. These three cases are the only ones we have tested
to date, but all have been successful, suggesting that they are unlikely to be atypical.
Attempts to transfer microbes between animal species will be facilitated by the precise
molecular methods now available to track individual organisms within complex mix-
tures. Without precise and quantitative tracking systems, the effectiveness of any
particular organism cannot be clearly demonstrated.
The potential  for removing  or  inactivating toxins and  anti-nutrients  by genetic
engineering of rumen bacteria has been strongly supported by the fluoroacetate case.
The ease with which this approach may be applied will depend upon the complexity of
reactions required for detoxification. Fluoroacetate represents the simplest system, with
a single toxin inactivated by a single enzyme encoded by a single gene. The detoxifi-
cation by non-oxidative means was also important for this process to be functional in
the rumen. Effective removal of the goitrogen from Leucaena leucocephala (Jones and
Megarrity 1986) by a bacterium that grows to only 10
5
cells/ml (C. Orpin, personal
communication) indicates that a small number of modified strains should be sufficient
to perform detoxification reactions. Modification of only a single strain might be
expected to have variable success, because of the population fluctuations seen in our
monitoring work. However, this may turn out, like some of the other predictions about
this work, to be overly pessimistic. Trials to test this aspect of the fluoroacetate work
are currently in progress.
The development of molecular mechanisms for ruminant studies, including methods
for microbial strain identification, estimation of population numbers and measurement of
changes, has now reached a stage where less affluent nations should be able to benefit
from them. Future research will indicate which of the possible applications for genetic
manipulation of rumen bacteria are likely to be successful. However, judgement of which
technologies will truly benefit the developing economies remains difficult, and the subject
of much conflict of opinion. If agencies assisting the developing nations are to adopt this
technology, then it should preferably be for the solution of problems that are well defined.
Acknowledgements
The work described here was funded by Applied Biotechnology Ltd (Qld) and the
Australian Meat Research Corporation.
238 K. Gregg, D. Schafer, C. Cooper and G. Allen
References
Armstrong D.G. and Gilbert H.J. 1985. Biotechnology and the rumen: a mini review. Journal
of the Science of Food and Agriculture 36:1039?1046.
Beard C.E., Hefford M.A., Forster R.J., Sontakke S., Teather R.M. and Gregg K. 1995. A stable
and efficient transformation system for Butyrivibrio fibrisolvens OB156. Current Micro-
biology 30:105?109.
Cocconcelli P.S., Terrari E., Rossi F. and Bottazzi V. 1992. Plasmid transformation of Ruminococcus
albus by means of high-voltage electroporation. FEMS Microbiology Letters 941:203?208.
DeOliveira M.M. 1963. Chromatographic isolation of monofluoroacetic acid from Palicourea
marcgravii. Experientia 19:586?587.
Dower W.J., Miller J.F. and Ragsdale C.W. 1988. High efficiency transformation of E. coli by
high voltage electroporation. Nucleic Acids Research 16:6127?6145.
Egan A.R., Hogan J.P. and Leng R.A. 1992. Effect of manipulation of rumen fermentation on
ruminant production. In: Manipulation of Rumen Micro-Organisms. Animal Production
Department, University of Alexandria, Egypt. pp. 70?83.
Elliot W.B. and Kalnitsky G. 1950. A mechanism for fluoroacetate inhibition. Journal of
Biological Chemistry 186:487?493.
Gregg K. and Sharpe H. 1991. Enhancement of microbial detoxification by gene transfer. In:
Tsuda T., Sasaki  Y. and Kawashima R. (eds), Physiological Aspects  of Digestion and
Metabolism in Ruminants. Proceedings of the Seventh International Symposium on Ruminant
Physiology, Sendai, Japan, 28 August?1 September 1989. Academic Press, San Diego,
California, USA. pp. 718?735.
Gregg K. and Ware C.E. 1990. Importance of genetic diversity in manipulation of rumen
bacteria. In: Akin D., Ljungdahl L., Wilson J. and Harris P. (eds), Microbial and Plant
Opportunities for Improvement of Lignocellulose Utilization by Ruminants. Academic Press,
New York, USA. pp. 357?366.
Gregg K., Allen G., Bauchop T., Klieve A.V. and Lincoln M. 1993. Practical genetic manipula-
tion of rumen bacteria. In: Farrell D.J. (ed), Recent Advances in Animal Nutrition in Australia.
University of New England, Armidale, Australia. pp. 13?21.
Gregg  K., Cooper C.L., Schafer D.J., Sharpe  H., Beard  C.E., Allen  G. and Xu J. 1994.
Detoxification of the plant poison fluoroacetate by genetically manipulated rumen bacteria.
Bio/Technology 12:1361?1365.
Hudman J.F. and Gregg K. 1989. Genetic diversity among strains of bacteria from the rumen.
Current Microbiology 19:313?318.
Jones R.J. and Megarrity R.G. 1986. Successful transfer of DHP degrading bacteria from
Hawaiian goats to Australian ruminants to overcome the toxicity of Leucaena. Australian
Veterinary Journal 63:259?262.
Kawasaki H., Tone N. and Tonomura K. 1981. Plasmid determined dehalogenation of halo-
acetates in Moraxella species. Agricultural and Biological Chemistry 45:29?34.
Kawasaki H., Tsuda K., Matsushita J. and Tonomura K. 1992. Lack of homology between two
haloacetate dehalogenase genes encoded on a plasmid from Moraxella species strain B.
Journal of General Microbiology 138:1317?1323.
Kawasaki H., Yahara H. and Tonomura K. 1984. Cloning and expression in Escherichia coli of
the haloacetate dehalogenase genes from Moraxella plasmid pUO1. Agricultural and Bio-
logical Chemistry 48:2627?2632.
LeBlanc D.J. and Lee L.N. 1984. Physical and genetic analyses of stroptococcal plasmid pAM?1
and cloning of its replication region. Journal of Bacteriology 157:445?453.
Genetic manipulation of rumen bacteria: now a reality 239
Mannarelli B.M. 1988. Deoxyribonucleic acid relatedness among strains of the species Butyrivi-
brio fibrisolvens. International Journal of Systematic Bacteriology 38:340?347.
Marais J.C.S. 1943 The isolation of the toxic principle potassium cymonate from gifblaar
Dichapetalum cymosum (Hook) Engl. Onderstepoort Journal of Veterinary Science and
Animal Husbandry 18:203?206.
McCosker T.H. 1989. Ruminal detoxification of fluoroacetate. Agricultural Science (New
Series) 2:46?47.
Russell J.B. and Wilson J.D. 1988. Potential opportunities and problems for genetically altered
rumen micro-organisms. Journal of Nutrition 118:271?279.
Thomson A.M., Flint H.J., Bechert M., Martin J. and Dubourgiuier H.-C. 1992. A new Es-
cherichia coli/Bacteroides shuttle vector, pRRI207, based on the Bacteroides ruminicola
plasmid replicon pRRI2. Current Microbiology 24:49?54.
Ware C.E., Bauchop T. and Gregg K. 1989. Isolation and characterization of cellulase genes
from two strains of Ruminococcus albus. Journal of General Microbiology 135:921?930.
Yulov A.A. and Zabara Z.A. 1990. Constructing DNA by polymerase recombination. Nucleic
Acids Research 18:3983?3986.
240 K. Gregg, D. Schafer, C. Cooper and G. Allen
Working group
reports, conclusions
and recommendations
Terms of reference
1. Which are the potential research areas in rumen ecology?
2. Among these areas, which would be the preferred research areas for ILRI?
a) What are the comparative advantages for ILRI in each area?
b) Which areas require collaboration between ILRI and other institutions?
3. For the areas identified to be potential ILRI research areas, which are the resource
needs?
a) With regard to human resources.
b) With regard to facilities.
4. a) Which are the suggested modes of collaboration within the previously identified
areas?
b) Who are the potential collaborators in each area?
5. Which are the time frames for expected output from respective research areas in
ILRI?
It is important to remember the potential impact of the respective research areas
within ILRI during working group discussions.
242. Working group
Report of Working Group 1
Potential research agenda for ILRI on rumen
manipulation: potential collaborators
and mode of collaboration
Chair: R.J. Wallace
Rapporteur: S. Fern?ndez-Rivera
Day 1 (Tuesday, 14 March 1995)
The group identified the following ten research areas pertaining to rumen manipulation.
1. Fibre breakdown
2. Adhesion to fibre particles
3. Anti-nutritional factors
4. Protein breakdown
5. Microbial protein production
6. Variation in microflora/microfauna
7. Methane production
8. Defaunation
9. Feed characteristics
10. Interactions between rumen microbes
General issues
It was established that given the global mandate of the institute, the research areas to
suggest for ILRI?s agenda in rumen manipulation should have a broad applicability. One
role of ILRI in these areas should be the development of tools robust enough to be used
successfully by the national research systems of developing countries (NARS).
Fibre breakdown
It was agreed that this is an important problem that could be addressed either by
improving the environmental conditions that facilitate fibre digestion, by promoting the
establishment of fungi or by genetic engineering. Several felt that in the short term the
manipulation of the microbial population is more likely to succeed than the high-tech
approach.
Several participants felt that the data presented in the workshop on the role of fungi
in fibre digestion encourage the inclusion of this group of rumen micro-organisms in
the research agenda.
One topic mentioned as researchable was the process of digestion as the fibre becomes
more lignified. There was a suggestion that the mechanisms involved in the improve-
ment of fibre digestion through supplementation needs to be understood. However, there
was also the recognition that the influence of supplementation on fibre breakdown could
be addressed by NARS.
There was a consensus that a main objective in this research area is the charac-
terisation of indigenous microbes to identify superior fibrolytic strains. The approach
suggested was the isolation and purification of strains. Superior strains could then be
used in high-tech approaches. Some technical difficulties were identified (e.g. purity of
CO
2
available in some countries, transport of microbes, etc.). However, it was consid-
ered that many of these difficulties could be overcome by involving NARS in the process
of preparing the pure cultures and by using polymerase-chain reaction (PCR). Several
NARS were identified as having capabilities for isolating microbes and preparing pure
cultures.
Attachment to fibre particles
Participants agreed that attachment of bacteria to fibre particles is a requirement for
fibre digestion and could be considered as an integrated part of the first research area.
The problem of reduced fibre digestion when high levels of concentrate are used was
discussed. There was a feeling that under these conditions there is no need for attach-
ment because of preferential use of readily fermentable carbohydrates as substrate. The
low pH can also affect the attachment of bacteria to fibre. It was mentioned that
possibilities for manipulating microbial attachment need to be investigated. Finally it
was concluded that adhesion was not a separate facet of fibre digestion and should not
be addressed as a priority area.
Anti-nutritional factors (ANFs)
The group considered that the institute has some expertise in this area. There is a broad
base of knowledge already available that needs to be transferred and put into practice.
It was suggested that phenolics constitute the most important group of ANFs, but that
there are many other types of compounds that have anti-nutritional effects. The detoxi-
fication of ANF and the breakdown of protein-tannin complexes were suggested as the
main researchable problems related to ANF.
One specific area of research considered during the group discussion was the
screening for the presence of ANF factors in ILRI?s germplasm bank, especially in
multipurpose trees (MPTs). Several participants felt that the screening should be made
at early stages in the process of evaluating the feed potential of promising trees. Gas
release, VFA and microbial mass production in vitro, as well as bioassays with small
animals (e.g. the field vole, mouse deer) were mentioned as possible screening criteria.
However, several participants questioned the need for a rumen microbiologist in these
activities. It was unclear how a microbiologist would fit in this project. It was argued
that the microbiologist should be a problem solver. This would require the identification
of a specific problem caused by ANF. In addition, the problem should have broad
relevance (i.e. not local or site-specific problems). Feasibility for being solved through
microbiological techniques was other requirement for its inclusion in ILRI?s agenda.
244 Report of Working Group 1
Several specific problems related to the utilisation of Acacia siberiana pods, Ta-
gasaste, Calliandria and Tephrosa were mentioned as potential researchable issues.
However, several participants felt that problems needing solution should be further and
more clearly defined and should have a broader relevance. Chosen species should have
a high feed potential and a wide distribution. The nature of the problem, i.e. the plant
compound that causes the toxic effects, should first be identified and isolated. The
causes of the anti-nutritional or toxic effects may well be at the systemic level and the
solution might not be found through microbiological approaches. The possibility of a
need of input from a toxicologist or a phytochemist to identify the nature of the problem
was noted.
There was consensus that possibilities for transferring indigenous micro-organisms
from animals that have survived and adapted under the presence of specific ANFs in
their diets to domesticated animals should be addressed. The input from the microbi-
ologist would be the isolation and culture of the microbes that detoxify the targeted
ANF.
Day 2 (Wednesday, 15 March 1995)
General issues
1. The feasibility for isolating and preparing pure cultures in NARS was revisited. There
was concern that the technical skills required are relatively high. Therefore it might
be difficult to find effective collaborating NARS in this area. Several participants felt
that in general, with some basic training and the provision of some key materials not
readily available in some countries, this should not be a major problem.
2. A framework for identifying research priorities was proposed. The framework was
based on the nature of the constraints to livestock production, i.e. nutritional and
non-nutritional factors. The latter (economic, social, political, disease, genetic
factors) would influence the appropriateness of each potential nutritional solution
proposed in a given farming system. The nutritional constraints were arranged in a
matrix with the limiting factors on one axis and possible solutions on the other. It
was suggested that  rumen microbiologists  had a potential  contribution in the
identification of which areas of the matrix could be focused on.
3. It was considered that the proposed activities on rumen ecology should be viewed
as part of ILRI?s feed utilisation programme. The group should focus on how
expertise in rumen microbiology will contribute and enhance the relevance of the
other research activities in this programme. It was suggested that a great deal of
specificity or focus was needed to make reasonable progress in five years. Aspecific
area that was suggested was the possibility for identifying natural defaunating
agents. Others replied that such a topic would be the subject of a PhD thesis, rather
than the core of a research programme. It was suggested that the feasibility of
transferring defaunating technology either commercially or through trees grown
on-farm should be considered. It might be possible that even if defaunating agents
exist, the indigenous protozoans have developed a resistance to these agents.
Report of Working Group 1 245
4. As the discussion on ANFs continued, the need for identifying the relative impor-
tance of ANFs (i.e. the farming systems, regions, seasons where they act as limiting
factors), as well as the chemical nature and mechanisms of action was indicated.
On the question of how much progress ILRI had made on these aspects, there was
a response that there was some information on their effects, but the chemical
characterisation would require the input of a phytochemist.
5. Among the ten research areas listed during the first group discussion, the priority
areas for ILRI?s research agenda were identified. It was considered that several areas
(e.g. methanogenesis, protein supply) actually derive from fibre breakdown; others
(e.g. defaunation) could be addressed through collaboration with other institutes,
but would not be broad enough to be the core of the programme; others (e.g. feed
characteristics) are part of other activities in feed evaluation or fibre breakdown.
There was consensus that the two most important areas are fibre breakdown and
ANFs. The identification of microorganisms that have superior fibrolytic activity
or capabilities to degrade ANFs was identified as the major contributions of rumen
ecology to ILRI?s research work.
Fibre digestion: needs, objectives and outputs
The feasibility of measuring the ability to degrade fibre in pure cultures was discussed.
It was mentioned that the isolated microbes could be characterised on the basis of
substrate utilisation, particularly in terms of relative rates of fibre digestion. Several
participants argued that this is already known while others suggested that the substrate
used for screening should be made up of locally available and important fibrous feeds.
This would ensure relevance of comparisons among strains. It was also suggested that
fungi should be included in the characterisation of indigenous microbes. Although the
collection of micro-organisms was important it was not seen as the end or objective of
the project.
The possibility for storing ?ecosystems?, i.e. mixed rumen populations, was also
discussed. Although technically it appears possible (transport of rumen fluid, use of
glycerol + soluble carbohydrates, dilution series, etc.), some participants questioned the
value of this approach.
Several members of the group mentioned the possibility that in a reasonable period
of time (three years) the programme could: a) assess the fibrolytic activity of few
species, b) assess how well these species are fitted to digest specific substrates and
locally available feeds, and c) describe the indigenous microbial populations in some
of the mandate areas. It was argued that the programme should be problem-oriented and
that more work to identify the constraints to fibre digestion, similar to the experiments
conducted in Niger, should be made. In reply, it was noted that such work identifies the
problem, not the solution. There is the possibility that the constraints are nutrients and
that the problem might not require microbiological techniques to solve. There is a need
to establish if the constraints are nutrient deficiencies. If this is the case, interventions
to overcome these constraints should be tested along with measures of how the fibrolytic
bacteria respond to such interventions.
246 Report of Working Group 1
The influence of synchrony in the availability of nutrients to microbes on fibre
digestion was suggested as a research issue. Several participants thought that this is not
an activity for a microbiologist and noted that it was not clear how this work would
serve the feed utilisation programme.
General objective
There was consensus that the general objective of the research activities on fibre
breakdown is ?the use of rumen microbiological techniques for alleviating constraints
to fibre digestion breakdown in ILRI mandate areas?.
Aims
It is expected that in a period of five years the research activities on fibre breakdown
will result in:
? Characterisation of fibrolytic activity of microbial populations.
? Selection of isolates with enhanced fibrolytic activity towards relevant fibrous
feeds.
? Assessment of cross inoculation and pure culture inoculation for improved fibre
breakdown.
Outputs
In a period of five years, the research on fibre breakdown should produce the following
outputs:
? Inoculants for enhanced fibre utilisation
? Improved feed utilisation
? Improved skills and techniques
? Collection of potential bio-resources
ANFs and MPTs
The need for microbiological work on MPTs was discussed at length. ILRI would not
have a comparative advantage from a microbiological point of view. The comparative
advantage of the institute is the access to the materials and the agronomic knowledge
base. Since more progress appears to be taking place in Latin America, it was suggested
that ILRI should focus on Africa and Asia and learn about the on-going work in Latin
America.
It was suggested that the most important group of ANFs is constituted by phenolics
and that there is need for knowledge on how the microbiological system in the rumen
responds to these compounds. Some group members thought that too much work is
already being done on phenolics at the expense of investigating the role of other
important ANFs.
Mechanisms of action of ANFs (i.e. at systemic or microbiological level) and the
needs for toxicological work were revisited. Several participants thought that this might
Report of Working Group 1 247
not be a microbiological problem. The need for a sharp definition of the problem to be
solved and the MPT species to be worked on was mentioned again. It was suggested
that trees to be used in ANF work should rank highly in relevance (broadly distributed),
feed potential (consumption by livestock, concentration of nutrients), yield potential
and severity (economic importance) of the problem caused. The specific problems
observed in Debre Zeit with Tagasaste, Tephrosa and Calliandria were rediscussed. It
was argued that some of the problems with MPTs are related to feed characterisation
and not rumen ecology. Some participants thought that it might be easier to make an
impact at farm level, through interventions including MPTs, than through enhancing
fibre digestion.
The group was reminded that work on mimosine required a lot of manpower and it
was unlikely that, with limited resources, ILRI could make a substantial impact in this
area in the short term.
After long discussions the group agreed that rumen manipulation work should focus
on fibre breakdown, but that the rumen microbiologist should collaborate with and
support the work on MPTs as the opportunities arise. Two potential areas that were
suggested as worth looking into were the characterisation of microbes in animals that
have evolved when forages high in ANFs are consumed and the possibilities for cross
inoculation to domesticated animals.
Resources needed for rumen manipulation work
The group advised that the rumen microbiologist would need back-up in three important
areas.
1. Analytical chemistry
2. Nutrition
3. Molecular biology
The group was informed about ILRI?s present analytical capabilities in Addis Ababa.
It was considered that the present facilities would be sufficient for only the most
common analytical needs. Some additional equipment would have to be purchased (see
below), but more advanced analytical needs demanding cutting-edge technology and
expertise would be contracted to advanced institutions and/or consultants.
The group was also informed that ILRI?s facilities for molecular biology in Nairobi
are state-of-the-art but are heavily used by other programmes. There is no expertise in
rumen microbes. If the rumen ecology work were based in Addis Ababa, the needs for
back-up in molecular biology could be met either in institutions already working in
rumen microbes or in Nairobi.
The group considered that at least a senior scientist and a post-doc with two full time
well qualified technicians are needed to cover the proposed rumen microbiology work.
Given that the work on rumen ecology would be part of the feed utilisation pro-
gramme, it was considered that it would be best placed in Ethiopia, where most of the
work on nutrition research is taking place. The laboratory should be placed where
cannulated animals are readily accessible.
248 Report of Working Group 1
The following equipment was considered necessary for the programme on rumen
manipulation: anaerobic cabinet and gas mixers, high-speed refrigerated centrifuge,
incubators, shaking water bottles, spectrophotometer (VIS/UV) and thermal block, GC
(FID and TZ), four-place balances, microscope and camera, liquid scintillation counter,
furnace and ovens, autoclave, colony counter, and automatic pipettes. Subsequent
discussion indicated that a micro-plate reader would be extremely useful in any
programme for the screening of ANFs.
Collaboration
The group agreed that collaboration with NARS and advanced institutions would be
critical for the success of the programme. No specific institutions for collaboration were
proposed. The group felt that this decision should be taken at a later stage and would
depend, at least in part, on the appointee(s).
Report of Working Group 1 249
Report of Working Group 2
Improved efficiency of use of feed
resources by tropical livestock
Chair: R.I. Mackie
Rapporteur: N. Murphy
The group identified the following seven constraints to feed resource utilisation by
tropical livestock.
1. Shortage of feed resources
2. Seasonal fluctuations
3. Feed quality
4. Toxic compounds
5. Applicability of methodologies
6. Environmental and health aspects
7. Low input/access
Opportunities
In considering the constraints, opportunities which could arise through a focused
research effort in feed utilisation and rumen microbiology were discussed and the
following seven identified.
1. Increase conversion efficiency
2. Enhance degradation
3. Detoxification
4. Provision and evaluation
5. Biodiversity
6. Ameliorate stress
7. Technology transfer
Potential research areas in rumen
microbiology and ecology
The agreed seven opportunities were discussed in turn with the focus on researchable
areas, available information, constraints and potential outputs. The group did not place
major emphasis on budgetary or personnel constraints that might exist in the institute
during this discussion, as information on this was lacking.
1. Conversion efficiency
? particle reduction
? improved growth efficiency (C, N, minerals)
? engineer protein quality
? intrarumenal N recycling/defaunation
? reduce methanogenesis/increase acetogenesis
2. Enhance degradation
? enzymatic enhancement (fibre, protein)
? alleviate nutritional limitations
? microbial interactions
3. Detoxification
? identification and characterization (plant and bugs)
? biochemical pathways
? relative contribution (bugs versus host)
? strategies for intervention
4. Evaluation
? chemical, microbial and animal
? seasonal and environmental variation
? develop appropriate methodology
5. Microbial diversity
?numbersandtypes
? choice of animal model
? adaptation (bugs versus animal)
6. Ameliorate stress
? nutrition/immune interaction
? heat generation by fermentation
? biocontrol of parasites
7. Technology transfer is a requirement
? consultation
? collaboration
? Preferred research area
? Focus on forage legumes which include multi purpose trees (MPTs).
Hypothesis
increased fibre degradation
?
better nutrient supply
?
enhanced biotransformation of ANFs
?
improved livestock efficiency
Why ILRI?
Advantages
? germplasm resources (ILRI, CIAT, ICRAF)
? research experience (characterisation)
? strong nutrition programme and facilities
252 Report of Working Group 2
? location (plant and animal)
? access/biodiversity
? global relevance
? global application (research)
Challenges
? scale (selection)
? target
Resource needs
Human
? two teams
? one rumen microbiologist/ecologist
? one rumen plant biochemist
? six to eight associates (allocation variable
? integrated with other expertise
Facilities
? space (three laboratories)
?
microbiology
?
molecular
?
biochemistry
It was agreed that there would be extensive interaction between both research groups
and that, therefore, it would be preferable for them to be located at the same place.
Further discussion on this was deferred until the two working groups met to discuss the
final recommendations.
Modes of collaboration
? science to drive collaboration
? collaborative grants
? international network
? contract research
? PhD students and post-docs
? review meetings
? potential collaborators unspecified
Time frame
? scope of work with five-year milestone
? mid-term review
Report of Working Group 2 253
Conclusions and recommendations
The International Livestock Research Institute (ILRI) inherits considerable research
experience on the development of livestock feeds in Africa and their utilisation by tropical
livestock. The institute is in the process of formulating a defined programme on feed
utilisation, the overall goal of which is the improved efficiency of use of feed resources
by tropical ruminants. The broad programme objectives are to (i) evaluate digestibility
and nutritional quality of traditional feedstuffs through animal responses; (ii) determine
how feed utilisation can be enhanced; and (iii) define adaptation mechanisms of tropical
ruminants to low quality feeds and other stresses (work, heat, water, etc.).
Feedstuffs consumed by ruminants is exposed to the fermentative activity in the
rumen prior to gastric and intestinal digestion. The fermentation yields characteristic
products which provide nutrients for metabolism by the host animal. The quality and
quantity of these products are dependent on the types and activities of the microbes in
the rumen. This, in turn, has major impacts on nutrient output and productive perform-
ance of ruminants.
It was therefore considered that manipulation of rumenal function, particularly by
influencing the microbiological activity of the rumen, may have a significant impact on
feed utilisation. Research in this area therefore might contribute directly to the second
objective of the feed utilisation programme.
The Rumen Ecology Research Planning Workshop was therefore convened with the
following objectives:
? to identify and prioritise areas of rumen ecology research which are promising for
their potential impact on improving the nutritional status of tropical ruminants;
? to develop, if appropriate, a rumen ecology research programme for ILRI based on
relevance to developing countries and ILRI?s comparative advantage vis-a-vis other
institutions; and
? to identify potential collaborators in advanced research institutes and define mode(s)
of collaboration.
Potential research areas in rumen
microbiology and ecology
The workshop participants collectively identified the following research areas:
1. Improvement of conversion efficiency by:
? better microbial adhesion to fibre particles
? reduction in particle size
? improved microbial growth efficiency (C, N, minerals)
? capture of limiting nutrients (NH
3
,BCVFA)
? engineering of protein quality
? intrarumenal N recycling/defaunation
? reduced methanogenesis/increased acetogenesis
2. Enhancement of degradation of feeds by:
? enzymatic enhancement of fibre and protein degradation
? alleviation of nutritional limitations
? optimisation of microbial interactions
3. Detoxification of toxic and anti-nutritional compounds in feeds by:
? identification and characterisation of toxins and their effects (plant and microbial)
? identification of their biochemical pathways
? understanding the relative effects of toxins/anti-nutritional factors (ANFs) on
rumen microflora and the host animal
? determination of strategies for intervention
4. Evaluation of toxins/ANFs by:
? development of appropriate methodologies for their determination
? chemical, microbial and animal assays
? examination of seasonal and environmental variation
5. Determination of microbial diversity by:
? determination of numbers, types and inter-strain variation
? choice of animal model
? adaptation (microbial flora vs animal host)
6. Amelioration of stress by:
? improving nutrition/immune interactions
? reduction of heat generation by fermentation
? biocontrol of parasites
7. Technology transfer is a requirement through
? consultation
? collaboration
? technologies appropriate to end-users
Within these research areas the participants evaluated the requirements of ILRI and
ILRI?s comparative advantage in undertaking research in any of the areas. The workshop
determined that any development of capacity should be through enhancement of the
existing feed utilisation programme; on the basis that:
? the most promising strategy is the exploitation of supplementary plant materials (this
research is in progress and represents a comparative advantage of ILRI);
? but the presence of ANFs in these potential supplementary plant feeds represents a
major constraint.
? nutritional evaluation programmes, largely for fodder trees, are currently based on
chemical analyses, degradation studies and feeding trials.
The workshop therefore developed two recommendations, the final wording of which
was determined in plenary session.
Recommendation 1
Within the existing programme, develop additional screening procedures at ILRI
incorporating microbiological, chemical and animal assessments of these supplemen-
tary plant materials relevant to their effects on ruminant productivity.
The expected outcome will be:
1. Improved methods of assessment of the potential utility of fodder trees and shrubs
(FTS) for enhancing the quality of low quality diets.
256 Conclusions and recommendations
2. More effective use of resources by concentrating on those FTS that do not have
significant ANFs.
3. Development of research strategies to overcome effects of ANFs for these acces-
sions that are otherwise of high merit.
4. Identification and selection of superior plant accessions as supplementary feeds.
Recommendation 2
Establish capacity in rumen microbiology/ecology at ILRI to capitalise on institu-
tional and international programmes aimed at increasing the utilisation of low quality
feeds through improved rumen function with emphasis on FTS to supply limiting
nutrients.
The expected outcome will be:
1. Assessment and exploitation (as appropriate) of indigenous or engineered microbes,
or changed population balance, on rumenal efficiency and the adaptation of tropical
livestock.
2. More complete characterisation of selected FTS and combination with available
feeds to optimise rumen function (especially with respect to detoxification, fibre
digestion and defaunation) and subsequent animal productivity.
3. Provision of more comprehensive and quicker attainment of the goals of the feed
utilisation programme.
4. Methodologies and materials for transfer to national programmes for improved
animal productivity on locally available feedstuffs.
Resource needs
The participants collectively agreed that the following human and material resources
would be required to carry out the recommendations when integrated with ILRI?s
existing resources.
1. Human?Two distinct but related efforts would require:
? one rumen microbiologist/ecologist
? one plant biochemist
? six to eight associates (at the postdoctoral and postgraduate levels)
? integration with other technical expertise and linkage to collaborating institutions
2. Facilities
? three discrete laboratory areas for:
?
plant biochemistry (with appropriate chromatographic and analytical facilities)
?
microbiology of anaerobic organisms
?
molecular biology
1
? isolation facilities (e.g. for transfer/adaptation experiments involving rumenal
microbes)
Conclusions and recommendations 257
1 It should be noted that the collective view was that ILRI, at this stage, would require molecular biology
facilities for DNA probe preparation, and microbial typing and quantification by PCR methods.
Genetic engineering of rumen microflora is not presently anticipated.
? animal facilities including cannulation capacity
? experimental surgery
3. Arrangement?Juxtaposition of the two research efforts (in phytochemistry and
rumen ecology) is clearly desirable. Workshop participants were of the view that
some disaggregation of component parts may be feasible but did not seek to make
a recommendation as to how this would be managed by the institute.
4. Modes of collaboration?An ILRI programme of scientific excellence developing
applicable research findings will attract collaborators. This could be supported
through the pursuit of collaborative grants, exchange of postdoctoral and PhD
students and, potentially, contract research. This interaction, the current workshop
and scientific review meetings would help establish the ILRI programme within the
worldwide network of laboratories active in the field of rumen ecology. Because
collaborations will be established after programme development and in response to
changing research needs, the workshop did not see the need to establish a list of
potential collaborators, although the workshop participants and their associated
institutes provide a research pool from which ILRI might draw.
5. Actions and time scale?Two senior appointments be made as soon as practicable
(start 1996). Development of chemical, microbiological  and animal/tissue for
screening ANFs in plant materials on a large scale will be a priority (variable but
within one to two years).
Early work will require establishment of additional methods of identifying and
quantifying important rumen microbes, including fungi (some assays are available,
others to be developed by associate scientific support or collaboration within two
years). Concurrently, a search can be made for suitable rumen microbial populations
from other mammals able to tolerate FTS known to contain toxins or ANFs. These
capacities will allow ILRI to interact with advanced research institutes (ARIs)
conducting complementary and supporting research on the rumen ecosystem.
A strategy of research for improving the utilisation of selected FTS through the
above could be expected between year 2 and 3 of the research (when a further
small-scale, peer review of the sub-programmes is suggested) with impact from
selected improvements estimated within five years of establishment of the pro-
gramme in ILRI.
258 Conclusions and recommendations
Appendix: list of participants
Workshop participants.
List of participants
Anindo, D.
International Livestock Research Institute
P.O. Box 5689
Addis Ababa
Ethiopia
Fax: (251-1) 61 18 92
Brooker, J.D.
Department of Animal Science
University of Adelaide
Glen Osmond, 5064 Australia
Fax: (618) 303 7114
Cobos, M.A.
Programa de Ganaderia
Colegio de Postgraduados
Montecillo, M?xico 56230
Fax: 52 595 45 265
Cook, R.M.
Michigan State University
East Lansing
Michigan 48824, USA
Fax: (517) 353 1699
Dolan, T.T.
International Livestock Research Institute
P.O. Box 30709
Nairobi
Kenya
Fax: (254-2) 63 14 99
Fern?ndez-Rivera, S.
International Livestock Research Institute
c/o ICRISAT Niger
P.O. Box 12404
Niamey
Niger
Fax: (227) 73 43 29
Fitzhugh, H.A.
International Livestock Research Institute
P.O. Box 5689
Addis Ababa
Ethiopia
Fax: (251-1) 61 18 92
Gardiner, P.
International Livestock Research Institute
P.O. Box 30709
Nairobi
Kenya
Fax: (254-2) 63 14 99
Gordon, G.R.
Division of Animal Production
Commonwealth Scientific and Industrial
Research Organisation (CSIRO)
Prospect, Sydney, New South Wales 2148
Australia
Fax: (61-2) 840 2940
Gregg, K.
Institute of Biotechnology
Armidale, NSW
2351 Australia
Fax: 67 72 8235
van Gylswyk, N.O.
Department of Animal Nutrition
and Management
Swedish University of Agricultural
Sciences
Kungs?ngen Research Centre
S-753 23 Uppsala, Sweden
Fax: 018 67 29 95
Hanson, J.
International Livestock Research Institute
P.O . B ox 56 89
Addis Ababa
Ethiopia
Fax: (251-1) 61 18 92
Johnsson, I.D.
Meat Research Corporation
P.O . B ox 49 8
Sydney South NSW 2000
Australia
Fax: 02 380 0699
Kaitho, R.
International Livestock Research Institute
P.O. Box 5689
Addis Ababa
Ethiopia
Fax: (251-1) 61 18 92
KopeB85n?, J.
Institute of Animal Physiology
and Genetics
Czech Academy of Sciences
Prague 10, UhBcd?nB8dves, 104 00
Czech Republic
Fax: (42-2) 759 182
Kudo, H.
Department of Animal Physiology
National Institute of Animal Industry
Ministry of Agriculture, Forestry
and Fisheries
Inashikigun, Ibaraki
305 Japan
Fax: 0298 38 8606
Lahlou-Kassi, A.
International Livestock Research Institute
P.O. Box 5689
Addis Ababa
Ethiopia
Fax: (251-1) 61 18 92
Mackie, R.I.
Department of Animal Sciences
and Division of Nutritional Sciences
University of Illinois at Urbana-
Champaign
Urbana, Illinois 61801
USA
Fax: 217 333 8804
Mohamed-Saleem, M.A.
International Livestock Research Institute
P.O. Box 5689
Addis Ababa
Ethiopia
Fax: (251-1) 61 18 92
Mpairuse, D.
International Livestock Research Institute
P.O . B ox 56 89
Addis Ababa
Ethiopia
Fax: (251-1) 61 18 92
Murphy, N.
International Livestock Research Institute
P.O. Box 30709
Nairobi
Kenya
Fax: (254-2) 63 14 99
Nsahlai, I.V.
International Livestock Research Institute
P.O . B ox 56 89
Addis Ababa
Ethiopia
Fax: (251-1) 61 18 92
Odenyo, A.
International Livestock Research Institute
P.O . B ox 56 89
Addis Ababa
Ethiopia
Fax: (251-1) 61 18 92, (251-1) 33 87 55
?rskov, E.R.
Rowett Research Institute
Bucksburn, Aberdeen
Scotland, AB2 9SB
Fax: 44 224 716 687
Osuji, P.
International Livestock Research Institute
P.O . B ox 56 89
Addis Ababa
Ethiopia
Fax: (251-1) 61 18 92, (251-1) 33 87 55
Topps, J.H.
Department of Animal Science
University of Zimbabwe
P.O. Box MP167
Mount Pleasant, Harare
Zimbabwe
Fax: 263 4333 407
262 List of participants
Umunna, V.
International Livestock Research Institute
P.O. Box 5689
Addis Ababa
Ethiopia
Fax: (251-1) 61 18 92
Wallace, J.
Rowett Research Institute
Bucksburn, Aberdeen
Scotland AB2 9SB
Fax: 44 224 716 687
Zerbini, E.
International Livestock Research Institute
P.O . B ox 56 89
Addis Ababa
Ethiopia
Fax: (251-1) 61 18 92
List of participants 263