Virus Evolution, 2015, 1(1): 1–16
doi: 10.1093/ve/vev009
Research article
The global distribution of Banana bunchy top virus
reveals little evidence for frequent recent,
human-mediated long distance dispersal events
Daisy Stainton,1 Darren P. Martin,2,† Brejnev M. Muhire,2 Samiuela Lolohea,3
Mana’ia Halafihi,4 Pascale Lepoint,5 Guy Blomme,6 Kathleen S. Crew,7
Murray Sharman,7 Simona Kraberger,1 Anisha Dayaram,1 Matthew Walters,1
David A. Collings,1 Batsirai Mabvakure,8 Philippe Lemey,9,‡
Gordon W. Harkins,8 John E. Thomas,10,* and Arvind Varsani1,2,11,*,§
1School of Biological Sciences and Biomolecular Interaction Centre, University of Canterbury, Private Bag 4800,
Christchurch, 8140, New Zealand, 2Department of Clinical Laboratory Sciences, University of Cape Town,
Cape Town, South Africa, 3Tonga College, Tongatapu, Kingdom of Tonga, 4Ministry of Agriculture and Food,
Forests and Fisheries, Kingdom of Tonga, 5Bioversity International, PO Box 18937, Bujumbura, Burundi,
6Bioversity International Uganda Office, Naguru, Kampala, Uganda, 7Queensland Department of Agriculture,
Fisheries and Forestry, Ecosciences Precinct, GPO Box 267, Brisbane, QLD 4001, Australia, 8South African
National Bioinformatics Institute, MRC Unit for Bioinformatics Capacity Development, University of the
Western Cape, Bellville, 7535, South Africa, 9KU Leuven, University of Leuven, Department of Microbiology
and Immunology, Rega Institute for Medical Research, Clinical and Epidemiological Virology,
Minderbroedersstraat 10, B-3000 Leuven, Belgium, 10The University of Queensland, Centre for Plant Science,
Queensland Alliance for Agriculture and Food Innovation, Ecosciences Precinct, PO Box 46, Brisbane, QLD,
4001, Australia and 11Department of Plant Pathology and Emerging Pathogens Institute, University of Florida,
Gainesville, FL 32611, USA
*Corresponding author: E-mail: Arvind.varsani@canterbury.ac.nz; john. thomas@daff.qld.gov.au; j.thomas2@uq.edu.au
†http://orcid.org/0000-0002-8785-0870
‡http://orcid.org/0000-0003-2826-5353
§http://orcid.org/0000-0003-4111-2415
Abstract
Banana bunchy top virus (BBTV; family Nanoviridae, genus Babuvirus) is a multi-component single-stranded DNA virus,
which infects banana plants in many regions of the world, often resulting in large-scale crop losses. We analyzed 171
banana leaf samples from fourteen countries and recovered, cloned, and sequenced 855 complete BBTV components
including ninety-four full genomes. Importantly, full genomes were determined from eight countries, where previously no
full genomes were available (Samoa, Burundi, Republic of Congo, Democratic Republic of Congo, Egypt, Indonesia, the
Philippines, and the USA [HI]). Accounting for recombination and genome component reassortment, we examined the
geographic structuring of global BBTV populations to reveal that BBTV likely originated in Southeast Asia, that the current
VC The Author 2015. Published by Oxford University Press.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/),
which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
1
2 | Virus Evolution, 2015, Vol. 1, No. 1
global hotspots of BBTV diversity are Southeast Asia/Far East and India, and that BBTV populations circulating elsewhere in
the world have all potentially originated from infrequent introductions. Most importantly, we find that rather than the current
global BBTV distribution being due to increases in human-mediated movements of bananas over the past few decades, it is
more consistent with a pattern of infrequent introductions of the virus to different parts of the world over the past 1,000 years.
Key words: phylogeography; Banana bunchy top virus; Nanoviridae; babuvirus; recombination; reassortment.
1 Introduction (SPG) and the Asian group (AG) (Karan, Harding, and Dale 1994).
Despite these phylogenetic groups having been defined based
Bananas are grown in over 130 countries and are ranked fourth, on the geographic origins of genomic component sequences
after wheat, rice, and maize, in importance as a food crop in the available in the mid-1990s, subsequently determined BBTV se-
world (Perrier et al. 2011; http://www.fao.org/docrep/i3627e/
quences have continued to phylogenetically cluster within one
i3627e/i3627e.pdf). Domesticated bananas are thought to have
or the other of these groups, with almost all sequences sampled
originated somewhere in the vicinity of New Guinea, Indonesia,
outside of Southeast Asia (SEA) falling into the SPG. Although
the Philippines, or the Southeast Asia Peninsula (Perrier et al. the SPG and AG have also been, respectively, referred to as the
2011) between 7,000 and 10,000 years ago (Denham et al. 2003). Pacific/Indian Ocean and the SEA groups (Yu et al. 2012), here
Banana cultivation subsequently spread to other parts of the
we will continue to use their original names.
world reaching Cameroon in West Africa and the Indian Ocean
It is likely that this geographic structuring has arisen
island of Madagascar possibly as early as 3,000 years ago.
because the rates of natural and/or human-mediated long-
During the period between 1,500 and 700 years ago, different ba- distance BBTV movement have been low enough for geo-
nana varieties were likely introduced and reintroduced to Africa graphically separated populations of these viruses to have
and the south-west Indian Ocean Islands many times (Lejju,
differentiated from one another. It remains unknown, however,
Robertshaw, and Taylor 2006; Randrianja and Ellis 2009).
whether the current geographical distribution of BBTV variants
Banana bunchy top disease (BBTD) is one of the most impor-
arose (1) concomitantly with the slow, pre-historic spread of ba-
tant diseases of banana, causing severe crop losses in many ba- nana cultivation across the Pacific, the Indian Ocean, Asia, and
nana-growing regions outside the Americas (Dale 1987; Rybicki Africa, (2) during the pre-globalization ebb and flow of banana
and Pietersen 1999; Rybicki 2015). Banana plants apparently dis- varieties across the Pacific and Indian Oceans between 100 and
playing BBTD symptoms were described in Fiji as early as the 1,500 years ago, or (3) during the modern globalization era as a
1880s (Magee 1927). In the 1930s, the banana aphid, Pentalonia consequence of poorly regulated agricultural trade. It is addi-
nigronervosa, was found to transmit the disease in a persistent tionally plausible that the current distribution of BBTV might
manner (Magee 1940). However, it was not until the 1990s that have arisen during this entire span of time. Importantly, the de-
an icosahedral single-stranded DNA virus with six genome grees of geographic structure evident within contemporary ge-
components was identified as the causative agent. This virus, nomic sequence data might be high enough to yield insights
Banana bunchy top virus (BBTV) (Harding, Burns, and Dale 1991; into when and from where the BBTV populations in particular
Thomas and Dietzgen 1991; Harding et al. 1993; Burns, Harding, continents, countries, or territories were founded. Such insights
and Dale 1994, 1995), is now recognized as the type member of would be extremely valuable in determining, for example,
the genus Babuvirus in the family Nanoviridae. whether modern movements of banana germplasm across the
The six genome components of BBTV are each approximately globe have had an appreciable impact on BBTV distributions.
1,000 nt long and are called DNA-R, DNA-U3, DNA-S, DNA-M, The potential for human-mediated dissemination of BBTV is
DNA-C, and DNA-N (formerly DNA-1 to DNA-6, respectively) high since cultivated bananas are sterile and are propagated
(King et al. 2012). DNA-R encodes a replication-associated protein vegetatively. Also, a banana plant infected with BBTV can take
(rep), DNA-S a capsid protein (cp), DNA-M a movement protein between 25 and 85 days to develop visible symptoms (Hooks
(mp), DNA-C a cell-cycle link protein (Clink), and DNA-N a nuclear et al. 2008) meaning that infected but symptomless banana
shuttle protein (nsp) genes (Hafner et al. 1997b; Aronson et al. propagules could be inadvertently transferred to regions where
2000; Wanitchakorn, Harding, and Dale 2000; Wanitchakorn et al. P. nigronervosa is present. The BBTV variants within infected
2000). The function of DNA-U3 is currently unknown. All compo- propagules might then be successfully transmitted and estab-
nents of individual viruses contain two sequence elements which lish new BBTV populations within bananas or wild hosts.
are highly similar across the components: a common region It is also likely that, as is the case with other related single-
stem-loop (CR-SL) element and a common region major (CR-M) stranded DNA viruses (Duffy and Holmes 2008; Duffy,
element (Burns, Harding, and Dale 1995). The CR-SL is involved in Shackelton, and Holmes 2008; van der Walt et al. 2008; Firth
replication and contains both a hairpin structure with a highly et al. 2009; Harkins et al. 2009, 2014; Grigoras et al. 2010;
conserved non-anucleotide sequence (TATTATTAC) and three re- Kraberger et al. 2013), BBTV is evolving at a sufficient rate that
peated five nucleotide long sequences, called iterons, that are evidence of such movement events should be encoded within
likely involved in the recognition and/or binding of Rep to the vi- the phylogenetic relationships of genomic component se-
rion strand origin of replication (v-ori) (Burns, Harding, and Dale quences sampled from extant BBTV populations.
1995; Herrera-Valencia et al. 2006). The CR-M is thought to be in- Phylogenetic inference of BBTV movement dynamics might,
volved in transcription (Burns Harding, and Dale 1995) and also however, be confounded by two other evolutionary processes
contains most of the binding sites for a primer molecule that is that occur in BBTV, genome component reassortment, and ho-
involved in complementary strand DNA synthesis (Hafner, mologous recombination. Because of genome components each
Harding, and Dale 1997). being packaged individually into separate virions, new infec-
Components of BBTV isolates broadly fall into two geograph- tions that are propagated from mixed BBTV infections will
ically well-defined phylogenetic groups, the South Pacific group frequently contain an assortment of different genome
D. Stainton et al. | 3
components. BBTV isolates that have genome components de- DNA extractions, amplification, and sequencing of BBTV ge-
rived from two or more different parental viruses have been in- nome components were carried out as described previously
ferred using a variety of phylogenetic (Hu et al. 2007; Yu et al. (Stainton et al. 2012). Briefly, sampled leaf material (fresh or
2012) and statistical recombination detection methods (Martin dried) was homogenized, and total DNA was extracted using an
et al. 2010; Stainton et al. 2012). Similar examples of component Epoch plant purification kit (Epoch Life Science Inc., USA).
reassortment have also been found in a number of other nano- Circular DNA was preferentially enriched using the TempliPhi
virus species (Grigoras et al. 2014; Savory and Ramakrishnan amplification kit (GE Healthcare, USA) as described previously
2014). (Owor et al. 2007; Shepherd et al. 2008). BBTV genome compo-
The known sequences of many individual BBTV genome nents were polymerase chain reaction amplified using compo-
components also carry evidence of homologous recombination nent-specific back-to-back primers described in Stainton et al.
(Hyder et al. 2011; Stainton et al. 2012; Wang et al. 2013; (2012). The resulting amplicons were resolved on an agarose gel,
Banerjee et al. 2014). Although the accuracy of phylogenetic re- gel purified, cloned and single transformed plasmid clones were
constructions for individual genome components could be sig- sequenced at Macrogen Inc. (South Korea). Sequence contigs
nificantly undermined by homologous recombination, both were assembled using DNA Baser Sequence Assembler v4
recombination and reassortment will undermine the accuracy (Heracle Biosoft SRL, Romania). Where possible, all six compo-
of full-genome phylogenetic reconstructions (Schierup and Hein nents were sequenced from each sample, although for some
2000; Posada and Crandall 2002). samples we were unable to recover all of the components
Both to gain a more detailed view of global BBTV diversity and (Fig. 2, Supplementary Table S1). Sequences from this study
to assess the geographical structuring of BBTV populations at have been deposited in GenBank (accession numbers
higher resolution than has previously been achievable, we deter- KM607005 - KM607859).
mined the sequences of 855 full BBTV genome components from
samples collected from across much of the known BBTV 2.2 Datasets
geographic range (Fig. 1, Supplementary Table S1). Accounting for
recombination, reassortment, and inferred rates of BBTV evolu- All full components sequenced as part of this study, along with
tion, we find that the diversity and phylogeographic structure of all full BBTV component sequences available in GenBank
contemporary-known BBTV populations is entirely consistent (downloaded 1st March 2014, see Supplementary Table S1 for
with there having been infrequent introductions of the virus to isolate information) were split into individual component-spe-
different parts of the world over the past 1,000 years. cific datasets (CSD) all starting at the ‘TATTAC’ region of the
nonanucleotide sequence motif. These sequences were aligned
using MUSCLE (Edgar 2004) implemented in MEGA5 (Edgar 2004;
2 Materials and methods Tamura et al. 2011). Aligned component sequences identified
2.1 Extraction and sequencing from the same sample were then concatenated into a single se-
quence. As outlined in Stainton et al. (2012), blank sequences
Samples were collected from 171 banana plants displaying (i.e., composed entirely of ‘-’ characters) were used where com-
stunting, bunched-up leaves, and Morse-code like streaking be- ponent sequences were not available both to maintain the com-
tween leaf margins and the midrib: all of which are symptoms ponent order and for alignment purposes. These concatenated
characteristic of BBTD. Samples were collected between 1989 sequences were labeled as the concatenated dataset (CD)
and 2012 from Australia (n¼ 40 isolates), four African countries (Supplementary Table S1 contains information on cognate
(n¼ 23 isolates), three Pacific island groups (n¼ 69 isolates), two BBTV components). From this dataset, a new BBTV genome
Indian Subcontinent countries (n¼ 8 isolates), and four SEA/Far dataset called the full-genome dataset (FGD) was created con-
East countries (n¼ 31 isolates), summarized in Fig. 2 and taining all available full-genome sequences (all six components
Supplementary Table S1. sequenced). Abaca bunchy top virus (ABTV) sequences were used
Complete genomes CG
South Pacific group components SPG
Asian group components AG Indian subcontinent
Components R U3 S M C N CG = 11
SPG:42 23 34 16 14 15
AG: Japan
2 - 1 2 1 1
China
Egypt
Africa India Southeast Asia
CG = 24 Taiwan CG = 24
SPG:27 27 35 26 26 25 Pakistan Vietnam SPG: 1 3 2 - 1 1 USA (Hawaii)
AG: 1 - - 1 1 - Myanmar AG: 69 38 51 32 32 34
Thailand Pacific Islands
Philippines CG = 40
Cameroon Sri Lanka SPG:61 68 66 69 65 67
Gabon Rwanda AG: - - - - - 1
Congo Burundi
Democratic Republic of Congo Malawi Indonesia
Samoa
Australia
Fiji Tonga
CG = 22
SPG:37 29 34 37 38 35
AG: - - - - - -
Figure 1. Geographical distribution of BBTV isolates. Summaries of component numbers and full genomes are provided for different regions. Accession numbers and
specific component information can be found in Fig. 2 and Supplementary Table S1.
4 | Virus Evolution, 2015, Vol. 1, No. 1
Study code SG R U3 S M C N Study code SG R U3 S M C N Study code SG R U3 S M C N Study code SG R U3 S M C N
Australia EF470243 Philippines TOS68*
482P2* GQ374514 AF148068 TOS71*
B2844* GU559702 AF416469 TOS72*
KP7* GU559703 10ph TOS78*
737* GU559704 13ph TOS82*
1429A* GU559705 14ph TOS85*
1429B* GU559706 15ph TOS87*
B2820* HM212635 MS14* TOS88*
B2826* Q529-5* 768* 536* C1
B2827* U97525 522A* D5 35to C1
B2828* Q529-1* 522B* D5 36to C1
B2832* Q529-6* 571-1* D5 37to C1
KP8* Q529-2* 571-2* D5 38to C1
B2818* 21cn D1 MS15* D5 39to C1
B2823* 23cn D4 MS16* D5 40to C1
B2830* 62cn D2 MS17* D5 41to C1
B2833* 63cn D3 MS18* D5 42to C1
B2834* Q529-4* E1 MS6* D5 43to C1
B2846* Egypt MS7* D5 44to C1
602* AF102784 Pakistan 45to C1
2557* C3 AF416465 AM418534 46to C1
1900A* C3 HQ259074 AM418535 KP4* C1
1900B* C3 34eg AM418573 Q276* C1
482-96* C3 9-150510* AM418539 Q277* C1
482-97* C3 8-150510* A1 AM418565 Q278* C1
482-98* C3 Fiji AM418567 Q570* C1
4au C3 32fj AY996563 TOS16* C1
B2817* C3 Gabon HE864318 TOS2* C1
B2819* C3 JF755981 HE864319 TOS20* C1
B2821* C3 JF755982 HE864320 TOS21* C1
B2822* C3 Indonesia 50pk TOS22* C1
B2824* C3 JN003631 52pk TOS25* C1
B2825* C3 JN003632 53pk TOS29* C1
B2829* C3 JN003633 54pk TOS39* C1
B2845* C3 16id 55pk TOS42* C1
B2847* C3 17id 56pk TOS48* C1
KP14* C3 18id 57pk TOS56* C1
KP15* C3 Q568-3* 58pk TOS60* C1
KP16* C3 520* D5 59pk TOS63A* C1
KP17* C3 Q568-1* D5 61pk TOS65* C1
KP18* C3 India 2pk TOS83* C1
KP6* C3 AF416470 48pk TOS90* C1
Burundi AY845437 49pk TOS91* C1
AF148943 AY884172 60pk TOS93* C1
547* AY884173 1pk Taiwan
526* C2 AY953429 26pk C2 626M*
548* C2 DQ656118 Rwanda AF148942
549* C2 DQ656119 19rw C2 AF416468
Democratic Republic of Congo EU046323 20rw C2 DQ826392
JF755984 FJ009240 Thailand EF095162
JF755986 HM120718 KC581796 FJ773283
JF755987 27in Tonga 626*
BU1* C2 28in AF416467 MP1*
BU10* C2 22in TOS53* 625I*
BU11* C2 523-6A* TOS77* 625*
BU12* C2 Q524-2* TOS79* 24tw
BU13* C2 64in TOS92* 5tw
BU14* C2 33in C2 TOS34* 627* D8
BU15* C2 3in C2 TOS45* 765* D5
BU16* C2 51in C2 TOS46* 25tw D5
BU17* C2 523-6B* D8 TOS7* MP2* D6
BU18* C2 66in B1 TOS74* Q1160* D5
BU19* C2 736-4* C2 TOS86* Q623* D7
BU2* C2 Q524-1* C2 TOS69* Q624* D5
BU20* C2 Q524-3* C2 TOS70* United States of America [Hawaii]
BU6* C2 Japan TOS76* 6us
BU7* C2 AB108454 TOS80* 527* C1
BU9* C2 AB108455 TOS15* KP9* C1
Congo AB108457 TOS43* Vietnam
550* C2 AB108458 TOS5* AF148945
Cameroon 7jp TOS57* AF416464
GQ249344 8jp TOS59* AF416472
JF755978 9jp TOS61* AF416473
China Sri Lanka TOS67* AF416474
AF110266 KP5* TOS89* AF416475
AF238874 65lk C2 TOS14* AF416476
AF238875 Q553* C2 TOS19* AF416477
AF238876 Myanmar TOS4* AF416478
AF238877 29mm TOS40* AF416479
AF238878 30mm TOS49* 11vn
AF238879 31mm TOS55* 12vn
AF246123 Malawi TOS58* Samoa
AF349568 JF755980 TOS62* Q280*
AY264347 47mw C2 TOS63B* Q279* C1
AY266417 TOS64* Q281* C1
Figure 2. Overview of sequenced BBTV components. Isolates are grouped by country, with each individual component depicted with a blue or green square depending
on whether the component falls phylogenetically into the AG (blue) or SPG (green) group (phylogenetic tree not shown). Isolates with asterisks were sequenced as part
of this study. Full-genome subgroups (SG), based on percentage pairwise identities, are shown for all isolates with all six components sequenced. Full isolate informa-
tion, including accession numbers and references, are in Supplementary Table S1.
D. Stainton et al. | 5
as an out group for all datasets except for the generation of the origins could have credibly been derived from different BBTV
maximum clade credibility (MCC) tree. Following recombination components.
and reassortment analysis, a recombination-free CSD (RF-CSD)
for each component and a recombination and reassortment-
free FGD (RF-FGD) were constructed from the CSDs and FGD, re- 2.5 Phylogenetic analysis of BBTV geographic
spectively (see below for recombination and reassortment distributions
details). Maximum likelihood (ML) phylogenetic trees were constructed
using the recombination-free CSDs with PHYML 3 (Guindon
2.3 Pairwise nucleotide sequence identity analyses et al. 2010) applying the best fit nucleotide substitution model
for each dataset determined using jModelTest (Posada 2008)
Percentage pairwise nucleotide identities of complete BBTV ge-
with 100 bootstrap replicates to determine branch support. For
nomes (in the context of the FGD) and of individual components
the recombination and reassortment-free FGD, an ML tree was
(in the context of the CSDs) were determined using Sequence
constructed using RAxML (Stamatakis 2014) with 100 bootstrap
Demarcation Tool (SDT) v1.2 (Muhire, Varsani, and Martin 2014)
replicates. All phylogenetic trees were rooted with ABTV se-
with the MUSCLE-based alignment option. CSDs and the FGD
quences, and branches with <60 per cent bootstrap support
were all analyzed with SDT without accounting for recombina-
were collapsed using Mesquite v2.75 (http://mesquiteproject.
tion. Distributions of pairwise nucleotide identities of the FGD
org/). RAxML was used for these particular trees rather than
and CSDs were used to tentatively classify BBTV genomes into
PHYML because it has been specifically optimized to construct
groups and subgroups based on the majority of the components
phylogenetic trees from sequences containing large amounts
in a similar way to Varsani et al. (2014).
of missing data (Izquierdo-Carrasco, Smith, and Stamatakis
All CSD were split into AG and SPG based on neighbor-
2011).
joining trees reconstructed using the Jukes–Cantor model as im-
To assess the time scales over which BBTV movements have
plemented in MEGA 5 (Tamura et al. 2011) (data not shown), and
likely occurred and identify the locations of the ancestral se-
percentage pairwise identity was calculated for each group us-
quences involved, we analyzed the recombination and reassort-
ing SDT v1.2.
ment-free CD dataset (RF-CD; n¼ 224 dated sequences) under a
Percentage pairwise identities were determined for the FGD
constant population size and strict-clock discrete diffusion phy-
and CSD sequences for five geographic regions: Africa (Burundi,
logeographical model with stochastic search variable selection
Cameroon, Congo, Democratic Republic of Congo, Egypt, Gabon,
(BSSVS) (Lemey et al. 2009) implemented in BEAST v1.8.1
Malawi, and Rwanda), the Indian Subcontinent (India,
(Drummond and Rambaut 2007). Seven geographical locations
Myanmar, Pakistan, and Sri Lanka), SEA/Far East Asia (China,
were considered: Africa (Burundi, Cameroon, Congo, and
Indonesia, Japan, Philippines, Taiwan, Thailand, and Vietnam),
Democratic Republic of Congo), the Indian Subcontinent (India,
Pacific Islands (Fiji, Hawaii, Kingdom of Tonga, and Samoa), and
Pakistan, and Sri Lanka), SEA (China, Indonesia, Philippines,
Australia.
Taiwan, and Thailand), Australia, Hawaii, Samoa, and Tonga.
The Pacific Islands were classified as independent locations as
2.4 Recombination and reassortment analyses we had a specific interest in determining whether statistically
All recombination and reassortment events were detected using supported movements between the different Pacific island
RDP4.27 (Martin et al. 2010), a recombination detection program states and the rest of the world could be reliably inferred.
which implements the following detection methods: RDP Bayes factor (BF) tests were used to determine the approxi-
(Martin and Rybicki 2000), GENECONV (Padidam, Sawyer, and mate statistical support for the inferred BBTV dispersal path-
Fauquet 1999), Bootscan (Martin et al. 2005), Maxchi (Smith ways, where a BF of less than five is not well supported, a BF of
1992) Chimera (Posada and Crandall 2001), SiScan (Gibbs, more than five implies substantial support, and BFs of between
Armstrong, and Gibbs 2000), and 3Seq (Boni, Posada, and 10 and 100 are indicative of strong support (Kass and Raftery
Feldman 2007). Recombination events were considered credible 1995; Lemey et al. 2009). Ten replicate runs of the Markov chain
when an event was identified by at least three detection meth- were run until the effective sample sizes for all of the model
ods with an associated P value< 0.05 and with at least one parameters were more than 200 and checked for convergence
method having an associated P value< 0.001 coupled with sup- using TRACER v1.6 (http://tree.bio.ed.ac.uk/software/tracer/).
porting phylogenetic evidence. Reassortment events were con- To determine whether sampling bias due to uneven sam-
sidered credible when, along with phylogenetic evidence, an pling sizes from the different locations considered has system-
event was identified by at least two detection methods with an atically influenced our parameter estimates, the BEAST
associated P value< 0.05, with at least one method having an analyses were also carried out with the location states of the se-
associated P value< 0.001. Intra-component recombination quences randomized and the location state probabilities of the
events were identified using the single CSDs. Reassortment root node compared with those determined for the datasets an-
events were identified using the CD. Specifically, recombination alyzed without randomization. SPREAD (Bielejec et al. 2011) was
events identified by RDP4.27 that had associated breakpoints used to produce a graphical animation in keyhole markup lan-
which spanned an entire component were identified as reas- guage (kml) format illustrating the historical spatio-temporal
sortment events. movement dynamics of BBTV that can be viewed in Google
Because of the large number of sequences being analyzed, Earth.
as well as issues with accurately aligning all six components,
a dataset containing all sequences from all components
was not used to detect evidence of possible inter-component 3 Results and discussion
recombination. Therefore, all intra-component recombinant re- 3.1 Sample collection and sequencing
gions with an unknown minor parent were further analyzed us-
ing BLASTn (Altschul et al. 1990) to determine whether Although BBTV populations seriously constrain banana produc-
transferred sequence fragments identified as having unknown tion throughout much of the eastern hemisphere, the
6 | Virus Evolution, 2015, Vol. 1, No. 1
worldwide genetic diversity of BBTV remains poorly under- global diversity of BBTV could be best inferred by increased
stood. We therefore amplified, cloned, and sequenced 855 com- sampling effort in these regions. DNA-M diversity is highest
plete BBTV components (DNA-R, n¼ 137; DNA-U3, n¼ 138; within the Indian subcontinent, whereas for DNA-U3, the diver-
DNA-S, n¼ 146; DNA-M, n¼ 146; DNA-C, n¼ 143; DNA-N, n¼ 145) sity is highest in SEA (Table 1).
from 171 BBTV infected banana pants from fourteen countries
spanning the known geographical range of this virus (Fig. 2,
Supplementary Table S1, accession numbers KM607005– 3.3 Reassortment analyses
KM607859). Given that BBTV genome components are individually encapsi-
A subset of the newly determined genome component se- dated, mixed infections will often result in genome component
quences constitute 94 complete BBTV genomes (i.e., instances reassortment (Hu et al. 2007; Stainton et al. 2012). To ensure the
where all six components have been sequenced from a single accuracy of our FGD phylogenetic analyses, it was vital that we
sample). These 94 genomes include those sampled in countries/ identified and removed from our datasets genome components
territories from which either no BBTV sequence data were pre- that had been acquired by reassortment. Towards this end, the
viously available (Congo and Samoa) or for which no full ge- CD was analyzed for evidence of reassortment using RDPv4.27
nomes have previously been sequenced (Burundi, Democratic (Martin et al. 2010), with manual identification of reassortment
Republic of Congo, Egypt, Indonesia, the Philippines, and events as detected recombination events that had inferred
Hawaii). This new sequence data more than doubles the num- breakpoint locations spanning entire components (Stainton
ber of publically available BBTV full-genome component se- et al. 2012). Given that this analysis involved almost four times
quences. All GenBank accession numbers for these and other more full genomes than previous BBTV reassortment analyses,
publically available BBTV sequences used in this study can be it is not surprising that of the seventy-five isolates detected as
found in Supplementary Table S1. reassortants, only 10 had been detected previously (Hu et al.
In total, 1,191 BBTV and 13 ABTV component sequences 2007; Stainton et al. 2012).
(ABTV DNA-M n¼ 3, all other components n¼ 2) were assembled These seventy-five isolates carried evidence of forty differ-
into seven datasets: CD (n¼ 317), CSD DNA-R (n¼ 242), CSD ent reassortment events (Fig. 3, Supplementary Table S2). All
DNA-U3 (n¼ 190), CSD DNA-S (n¼ 225), CSD DNA-M (n¼ 186), components were represented among these events, albeit with
CSD DNA-C (n¼ 180), and CSD DNA-N (n¼ 181). These sequences some components having been transferred more than others.
have collectively been recovered from a total of 317 plant sam- Component DNA-U3 was found to be the most commonly trans-
ples (170 in this study) from twenty-five countries (fourteen ferred component (eleven events), followed by DNA-M (eight
sampled in this study; Fig. 2 and Supplementary Table S1). An events), DNA-S and DNA-N (both with seven events), DNA-C
FGD containing isolates with all six component sequences was (five events), and DNA-R (two events).
assembled from the CD and contained 121 full BBTV genomes Similar reassortment analyses in Cardamom bushy dwarf virus
and two full ABTV genomes. (CBDV) and viruses in the genus Nanovirus that lack a DNA-U3
component have also found that DNA-M and/or DNA-N are
among the most frequently transferred nanovirus components
3.2 Classification of the genome segments and full during reassortment (Grigoras et al. 2014; Savory and
genomes Ramakrishnan 2014). However, DNA-U3, which is only present
The DNA-U3 components were most diverse, sharing >74 per in Babuviruses (BBTV, ABTV, and CBDV), was not found to be
cent pairwise identity followed by DNA-S and DNA-M (both among the most frequently transferred genome components
sharing >82% pairwise identity), and the DNA-N, DNA-C, and during CBDV reassortment (Savory and Ramakrishnan 2014),
DNA-R components that shared >83 per cent, >85 per cent, and suggesting that patterns of component transfer are not abso-
>88 per cent pairwise identity, respectively. Collectively, the lutely conserved between different species.
segments in the FGD shared >85 per cent pairwise identity. For Of the seventy-five reassortant genomes that we detected,
the FGD, the genome sequences which shared >85 per cent but thirty-four had one detectable reassortment event, thirty-three
<94 per cent pairwise identity were subdivided into groups A–E. had two, and eight had three. Overall, 38 per cent of all isolates
Within these groups, genomes with >98 per cent pairwise iden- with at least three sequenced components (75/196) show evi-
tity were further divided into subgroups A1, B1, C1–3, D1–8, and dence of at least one component having been acquired by reas-
E1 (Fig. 2; Supplementary Table S1). sortment. Crucially, twelve of the forty reassortment events
With the exception of DNA-S, the genetic diversity among were each detected in multiple genomes. This strongly suggests
the currently sampled AG genome components is generally that these events occurred in an ancestor of these genomes and
greater than that among the corresponding SPG components: therefore that reassortment yielded viable viruses that went on
DNA-R (AG> 91%, SPG> 94%), DNA-U3 (AG> 76%, SPG> 81%), to become epidemiologically relevant.
DNA-M (AG> 89%, SPG> 91%), DNA-C (AG> 89%, SPG> 94%), Our detection of reassortment events between AG and SPG
and DNA-N (AG> 89%, SPG> 91%) components. In the case of genomes sampled in Egypt, China, India, and Taiwan (Fig. 2) is
DNA-S, the AG sequences are >92 per cent identical, whereas consistent with the geographic range of the AG and SPG line-
the SPG sequences are >87 per cent identical. ages overlapping in these regions. This overlap suggests that
The percentage pairwise identities of genome components the Indian/Southeast Asian/Far Eastern region is likely the geo-
sampled from five major regions of the world (Africa, the Indian graphic hotspot of BBTV diversity and might even be the region
subcontinent, SEA/Far East, the Pacific Islands, and Australia) where the most recent common ancestor of all currently sam-
indicated that the greatest degree of BBTV sequence diversity pled BBTV isolates originated.
occurs within the SEA/Far East/Indian subcontinent regions
(Table 1). This is true for the FGD and all individual components.
The significant diversity observed in Africa is contributed 3.4 Recombination analyses
mainly by the AG-like DNA-R, -M, and -C components of the A number of studies have identified potential recombination
Egyptian isolate, 8–150,510 (Fig. 2). This suggests that the true events in BBTV (Fu et al. 2009; Islam et al. 2010; Hyder et al.
D. Stainton et al. | 7
Table 1. Percentage pairwise identities of individual BBTV genome components that have been sampled from different geographical regions.
Full genome/ Region Pairwise Number of pairwise Average pairwise Standard
component identity (%) comparisons identity (%) deviation (%)
Full genome Africa >91.3 276 98.2 2.0
Australia >97.3 231 99.0 0.5
Indian subcontinent >87.1 66 95.2 3.4
Pacific islands >97.3 780 97.9 0.7
SEA >90.1 276 96.5 2.6
DNA-R Africa >90.1 378 98.2 2.2
Australia >98.6 666 99.5 0.3
Indian subcontinent >89.5 946 97.9 2.6
Pacific islands >95.5 1,830 98.7 0.9
SEA >89.1 2,415 96.8 2.2
DNA-U3 Africa >87.2 351 96.3 3.4
Australia >96.7 406 98.5 0.8
Indian subcontinent >82.5 253 93.1 4.1
Pacific islands >91.3 2,278 91.3 1.8
SEA >75.9 820 90.8 6.4
DNA-S Africa >96.8 595 98.6 0.6
Australia >88.9 561 98.5 2.0
Indian subcontinent >87.8 595 97.2 2.4
Pacific islands >96.9 2,145 98.6 0.6
SEA >88.2 1,378 96.7 2.9
DNA-M Africa >82.1 351 97.3 4.2
Australia >92.4 666 97.8 2.2
Indian subcontinent >83.4 183 94.6 5.7
Pacific islands >91.8 2,346 96.6 2.1
SEA >89.2 496 95.9 3.3
DNA-C Africa >86.7 351 97.8 3.0
Australia >97.6 703 99.0 0.4
Indian subcontinent >86.4 105 96.3 3.4
Pacific islands >97.5 2,080 98.7 0.4
SEA >86.3 528 95.9 3.5
DNA-N Africa >97.2 300 98.9 0.5
Australia >98.9 595 99.5 0.2
Indian subcontinent >85.0 120 96.3 4.4
Pacific islands >83.8 2,278 98.0 2.4
SEA >84.7 595 95.9 4.1
2011; Stainton et al. 2012; Banerjee et al. 2014), and, as with reas- -S, and -C. As has been found in previous nanovirus recombina-
sortment, it was important to account for these events during tion studies (Hyder et al. 2011; Stainton et al. 2012; Grigoras
our subsequent phylogenetic analyses. We analyzed the CSDs et al. 2014; Savory and Ramakrishnan 2014), we detected similar
to identify recombinant sequences, the locations of recombina- numbers of recombination breakpoints within the non-coding
tion breakpoints, and the identities of likely parental viruses. and coding regions (twenty-four and twenty breakpoints,
These analyses revealed that all components displayed at respectively).
least some evidence of recombination (Figs 4 and 5; Supplemen- In total, thirteen events resulted in recombinant genes that
tary Tables S3–S8), with the greatest number of recombination could express chimeric proteins. However, all thirteen of these
events being detected in DNA-U3 (twelve events) and the fewest events involved recombination between closely related BBTV
in DNA-M (two events). variants, meaning that these recombination events would have
All components carried evidence of recombinant regions in- had only a minimal impact on encoded protein amino acid se-
volving the CR-SL region (with breakpoints falling within and/or quences (Lefeuvre et al. 2009). Another possible sign of protein
on either side of this region) but only DNA-U3 and -N have re- coding sequences having an impact on recombination patterns
combination regions involving the CR-M, all of which had in BBTV is that the DNA-U3 component, which has no con-
breakpoints falling on either side of this region. Of the 18 recom- firmed protein coding function, has a higher concentration of
bination events that were identified within multiple isolates, detectable recombination breakpoints than those of the known
nine are seen within isolates from multiple countries. As with protein coding genes of other components. Interestingly DNA-
the reassortment events that are observed in multiple different U3 is also the component that appears to be most frequently ex-
genomes, these recombination events apparently occurred changed by reassortment in BBTV. High frequencies of recombi-
within genomes that were ancestral to two or more of the se- nation in this component might reflect the fact that it is mostly
quences analyzed here and indicate that at least some BBTV re- evolving neutrally with no risk that recombinants might ex-
combinants are epidemiologically relevant. press defective chimeric proteins (Lefeuvre et al. 2009) and that
Twenty-two recombination events were detected within the there is therefore little conservation of coevolved epistatic inter-
components encoding genes of known function, DNA-R, -M, -N, actions within this component.
8 | Virus Evolution, 2015, Vol. 1, No. 1
Reassortant R U3 S M C N Reassortant R U3 S M C N
627-TW-1996 D8 1 B2826-AU-2011 36
Q623-TW-1996 D7 2 B2832-AU-2011 36
Q529-6-CN-1990 4 MP2-TW-1996 D6 37
TOS88-TO-2010 6 BU11-CD-2012 C2 38
523-6B-IN-1991 D8 9 KP5-LK-2003 39
625I-TW-1995 11 9-150510-EG-2010 16 39
736-4-IN-1997 C2 12 1429A-AU 34 7
TOS14-TO-2010 18 B2833-AU-2011 23 13
TOS71-TO-2010 19 62cn-CN D2 14 15
Q529-4-CN-1990 E1 22 1429B-AU 27 24
KP8-AU-1989
Q529-2-CN-1990 22 44to-2010 C1
45to-2010 C1 20 28
46to-2010 C1
B2823-AU-2011 23
B2827-AU-2011 23 B2818-AU-2011 23 33
523-6A-IN-1991 25 B2834-AU-2011 23 26
KP7-AU-1989 27 All C3 23 33
TOS40-TO-2010 28 B2819-AU-2011 C3 23 8 33
TOS49-TO-2010
42to-TO-2010 C1 8-150510-EG-2010 A1 40 17 3
43to-TO-2010 C1
TOS39-TO-2010 C1 28
TOS48-TO-2010 C1 24tw-TW 10 31 5
TOS90-TO-2010 C1
B2820-AU-2011 33 63cn-CN-D3 14 32 30
B2828-AU-2011
B2846-AU-2011 33 527-US-1992 C1
KP9-US-1990 C1
Q279-WS-1989 C1 16 21 29
3in-IN-2007 C2 35 Q281-WS-1989 C1
602-AU-1996 36 Non reassortant sequence
B2830-AU-2011
Reassortant sequence
737-AU-1997 36 No sequence available
Figure 3. Detected reassortment events. As not all reassortant isolates consist of full genomes, circles depict component sequences, which are available and a dash in-
dicates where no component sequence is available. Components are shown as either non-reassortant sequences (white filled circles) or as reassortant sequences
(black circles) with the corresponding reassortant event number. Further information on reassortment events can be found in Supplementary Table S2.
D. Stainton et al. | 9
Recombination event Graphical representation Recombinant sequence(s) Detection P-value
methods
DNA-R
R1 AF416476-R-VN MCT 1.70x10-05
AF416477-R-VN
AF416478-R-VN
R2 MP2-R-TW-1996-D6 RGBT 8.04x10-05
R3 5tw-R-TW MCS 3.68x10-04
625I-R-TW-1995
R4 21cn-R-CN-D1 RGB 2.95x10-03
62cn-R-CN-D2
63cn-R-CN-D3
R6 MCS 3.70x10-03
6us-R-US KP9-R-US-1990-C1
527-R-US-1992-C1 Q279-R-WS-1989-C1  
Q281-R-WS-1989-C1 
DNA-S
S1 626-S-TW-1996 MP2-S-TW-1996-D6 RGT 3.96x10-09
626M-S-TW-1995
S2 10ph-S-PH AF148942-S-TW RGMCST 3.72x10-06
11vn-S-VN AF148945-S-VN
12vn-S-VN AF238876-S-CN 
13ph-S-PH AF238877-S-CN
14ph-S-PH MP1-S-TW-1996 
15ph-S-PH MS14-S-PH-2008 
16id-S-ID Q529-1-S-CN-1990
17id-S-ID Q529-2-S-CN-1990
18id-S-ID Q529-5-S-CN-1990
5tw-S-TW All D1 
625-S-TW-1996 All D2 
625I-S-TW-1995 All D3 
626-S-TW-1996 All D4 
626M-S-TW-1995 All D5 
7jp-S-JP All D6 
768-S-PH-1995 All D8 
8jp-S-JP All E1 
9jp-S-JP
AF148068-S-PH
S5 B2846-S-AU-2011 RGB 6.90x10-04
S7 JF755981-S-GA-2008 JF755984-S-CD-2008 RGB 2.90x10-03
S8 5tw-S-TW 625I-S-TW-1995 RGB 7.27x10-03
625-S-TW-1996
DNA-M
M1 66in-M-IN-2012-B1 GBMS 9.97x10-05
M5 ABTV3-M-MY GBS 2.19x10-04
DNA-C
C1 3in-C-IN-2007-C2 RGMCT 1.03x10-07
C2 8-150510-C-EG-2010-A1 Q624-C-TW-1996-D5 RGT 1.41x10-06
625-C-TW-1996 All D6
765-C-TW-1996-D5 All D7
C3 Q529-4-C-CN-1990-E1 Q529-2-C-CN-1990 MCS 4.42x10-06
C4 526-C-BI-1992-C2 GBT 6.65x10-03
Open reading frame Common region stem-loop Common region major Recombinant region
Figure 4. Recombination events detected in DNA-R, DNA-S, DNA-M, and DNA-C. Methods which detected the event are shown by abbreviations: R, RDP; G, GENCONV;
B, BOOTSCAN; M, MAXCHI; C, CHIMERA; S, SISCAN; T, 3SEQ. The highest detected P value is shown with the detection method marked in bold. Further information on
recombination events can be found in Supplementary Tables S3 and S5–S7.
Eighteen of the detected recombination events apparently S3–S8 for details. All of the recombinationally derived genome
involved the acquisition by BBTV isolates of genetic material de- regions were analyzed using BLASTn (Altschul et al. 1990), with
rived through either inter-component recombination or recom- four of these regions—those transferred in U7 (in DNA-U3), S1
bination with non-BBTV babuvirus species (DNA-R, n¼ 2; -U3, (in DNA-S), C2 (in DNA-C), and N3 (in DNA-N)—having poten-
n¼ 8; -S, n¼ 3; -C, n¼ 2; -N, n¼ 3) see Supplementary Tables tially involved inter-component sequence transfers. BLASTn
10 | Virus Evolution, 2015, Vol. 1, No. 1
Recombination event Graphical representation Recombinant sequence(s) Detection P-value
methods
DNA-U3
U2 22in-U3-IN TOS67-U3-TO-2010 RGMCST 1.17x10-17
24tw-U3-TW TOS68-U3-TO-2010 
28in-U3-IN-2012 TOS72-U3-TO-2010 
34eg-U3-EG-1997 TOS78-U3-TO-2010 
6us-U3-US TOS79-U3-TO-2010 
602-U3-AU-1996 TOS80-U3-TO-2010 
9-150510-U3-EG-2010 TOS82-U3-TO-2010 
AY884173-U3-IN TOS85-U3-TO-2010 
B2818-U3-AU-2011 TOS87-U3-TO-2010 
B2823-U3-AU-2011 TOS88-U3-TO-2010 
B2827-U3-AU-2011 TOS89-U3-TO-2010 
B2828-U3-AU-2011 All A1
B2830-U3-AU-2011 All C1 except 6 
B2833-U3-AU-2011 [35to-U3-TO-2010-C1 
B2834-U3-AU-2011 36to-U3-TO-2010-C1 
EU046323-U3-IN 38to-U3-TO-2010-C1
Q524-2-U3-IN 44to-U3-TO-2010-C1 
Q529-6-U3-CN-1990 TOS56-U3-TO-2010-C1 
TOS14-U3-TO-2010 TOS83-U3-TO-2010-C1]
TOS19-U3-TO-2010 3in-U3-IN-2007-C2 
TOS40-U3-TO-2010 33in-U3-IN-2002-C2 
TOS43-U3-TO-2010 51in-U3-IN-C2
TOS49-U3-TO-2010 65lk-U3-LK-2010-C2
TOS55-U3-TO-2010 Q524-1-U3-IN-C2
TOS63B-U3-TO-2010 Q524-3-U3-IN-C2
TOS64-U3-TO-2010 
U4 25tw-U3-TW-D5 MS14-U3-PH-2008 RGMST 7.77x10-12
U5 AY996563-U3-PK-2007 BU10-U3-CD-2012-C2  RGT 4.07x10-08
19rw-U3-RW-2009-C2 BU17-U3-CD-2012-C2 
20rw-U3-RW-2009-C2  BU9-U3-CD-2012-C2
BU1-U3-CD-2012-C2  
U6 TOS93-U3-TO-2010-C1 BMT 1.22x10-04
U7 5tw-U3-TW GMST 5.93x10-19
U8 625-U3-TW-1996 Q568-3-U3-ID-1995 RGMCST 7.39x10-11
625I-U3-TW-1995 All D2
DQ826392-U3-TW All D4
FJ773283-U3-TW  All D5
GU559703-U3-CN-2008 All D6
MP1-U3-TW-1996 All D7
MS14-U3-PH-2008 
U10 5tw-U3-TW RGBST 7.21x10-05
U12 Q529-2-U3-CN-1990 Q529-4-U3-CN-1990-E1 RGB 1.86x10-03
U17 AY884173-U3-IN 65lk-U3-LK-2010-C2 MCST 6.68x10-05
Q529-6-U3-CN-1990 Q524-1-U3-IN-C2 
33in-U3-IN-2002-C2 Q524-3-U3-IN-C2
51in-U3-IN-C2 
U19 66in-U3-IN-2012-B1 RMCS 1.07x10-04
U21 8-150510-U3-EG-2010-A1 MCS 1.85x10-05
U22 TOS43-U3-TO-2010 RGB 2.47x10-03
DNA-N
N1 TOS53-N-TO-2010 RGB 2.95x10-10
N2 TOS40-N-TO-2010 42to-N-TO-2010-C1 RGMCST 2.26x10-08
TOS49-N-TO-2010 45to-N-TO-2010-C1
TOS58-N-TO-2010 46to-N-TO-2010-C1
36to-N-TO-2010-C1 TOS39-N-TO-2010-C1
37to-N-TO-2010-C1 TOS48-N-TO-2010-C1
41to-N-TO-2010-C1 TOS90-N-TO-2010-C1
N3 ABTV2-N-PH RMST 5.26x10-22
N4 BU6-N-CD-2012-C2 RGB 7.42x10-06
N6 MP1-N-TW-1996 MP2-N-TW-1996-D6 RGB 1.03x10-03
N7 TOS16-N-TO-2010-C1 TOS61-N-TO-2010 RGB 7.28x10-03
TOS22-N-TO-2010-C1 BU13-N-CD-2012-C2
TOS56-N-TO-2010-C1
Open reading frame Common region stem-loop Common region major Recombinant region
Figure 5. Recombination events detected in DNA-U3 and DNA-N. Methods which detected the event are shown by abbreviations: R, RDP; G, GENCONV; B, BOOTSCAN;
M, MAXCHI; C, CHIMERA; S, SISCAN; T, 3SEQ. The highest detected P value is shown with the detection method marked in bold. Further information on recombination
events can be found in Supplementary Tables S4 and S8.
D. Stainton et al. | 11
(Altschul et al. 1990) analysis of the U7 recombinant region indi- constructed ML phylogenetic trees for all of these datasets
cated that this had likely involved a BBTV satellite (accession (Supplementary Figs S1–S6, Fig. 6). In order to date possible
no. EU366175): a result that corroborates the finding of Fu et al. BBTV movement events and identify the likely origins of BBTV
(2009). Although BLASTn analyses of events S1 and C2 indicated isolates in particular geographical regions, we additionally per-
that the most likely sources of the recombinationally acquired formed a Bayesian Monte Carlo Markov chain (MCMC) analysis
sequences were BBTV DNA-M components, analysis of event N3 and constructed an MCC tree (Fig. 7) from a recombination and
(which was detected in ABTV) indicated that it had likely in- reassortment-free dataset (RF-CD) representing 224 BBTV iso-
volved a sequence transfer from an ABTV DNA-S component. late sequences.
Our analyses indicated the remaining fourteen detected re- Although all of the sequences in the DNA-R, -S, -M, -C, and -
combination events with unknown parents likely involved ho- N trees fell within clearly defined SPG and AG clades
mologous recombination between BBTV and viruses belonging (Supplementary Figs 1 and S3–S6), the DNA-U3 tree contains
either to currently unsampled babuvirus species or to divergent some sequences that cannot be convincingly classified into ei-
currently unsampled BBTV strains. This suggests that there ther the SPG or AG clades (Supplementary Fig. S2).
may exist a far greater diversity of BBTV-like babuvirus species Given the low degrees of support for most of the branches
(or perhaps divergent BBTV strains) than is presently known. within the RF-CSD trees, we opted to focus on the RF-FGD ML
Also, the fact that recombination events that are inferred to and RF-CD MCC trees, in our assessment of finer scale geo-
involve currently unsampled babuvirus species are primarily graphic structure within each of the AG and SPG clades.
evident in BBTV isolates sampled in SEA/Far East region (nine For the MCC tree, the MCMC analysis under a constant popu-
of fourteen events) further suggests that this region is likely lation size, strict-clock, discrete diffusion model produced an
a major hotspot of ongoing recombination-driven BBTV estimate of the BBTV mean nucleotide substitution rate of
diversification. 2.916 104 substitutions/site/year (95% highest posterior den-
For recombination to occur between any particular pair of vi- sity [HPD] 2.148 104–3.755 104), which is approximately
ruses, the viruses must have overlapping geographic ranges, double that reported in the analysis by Almeida et al. (2009)
host ranges, and cell tropisms. A number of plants, which are based on non-coding (500 nts) regions of the five components
also hosts of Pentalonia spp. (Pentalonia caladii and P. nigroner- (DNA-R, DNA-S, DNA-M, DNA-C, and DNA-N) in Hawaiian iso-
vosa), have been suggested as potential alternative hosts for lates sampled between October and December 2005 (1.4 104
BBTV including Canna indica (canna lily), Hedychium coronarium substitutions/site/year) but lower than that obtained under
(white ginger lily), and Colocasia esculenta (taro) (Foottit et al. experimental conditions (3.9 104 substitutions/site/year)
2010; Duay et al. 2014). BBTV can be transmitted by Pentalonia (Almeida et al. 2009). The time since the most recent common
spp. from an infected banana plant into Co. esculenta (asymp- ancestor of the BBTV sequences represented in the MCC tree
tomatic) and then back into a healthy banana plant to cause dis- was 1,086 years (95% HPD 812—1,399 years).
ease (Ram and Summanwar 1984). Canna indica and H. It is immediately evident from both the RF-FGD ML (Fig. 6)
coronarium have also shown weak to moderate reactions in and RF-CD MCC (Fig. 7) trees that there is a high degree of geo-
BBTV-specific ELISA tests (Su, Wu, and Tsao 1992). However, al- graphic clustering among the sub-clades within both the SPG
though Pinili et al. (2013) reported the successful transmission and AG. That is, there are many well-supported monophyletic
of an Okinawan BBTV isolate to C. indica, Co. esculenta, and groups containing viruses all sampled from the same country.
Alpinia zerumbet, further studies have failed to confirm that This clustering is particularly strong among the SPG isolates.
these species are suitable hosts for other BBTV strains (Hu et al. For example, although the Tongan and Hawaiian SPG isolates
1996; Geering and Thomas 1997; Manickam et al. 2002). form single well-supported monophyletic groups in both the
Regardless of the actual BBTV host-range, our results indi- MCC and ML trees, the Australian SPG isolates form a single
cate that an increased sampling effort targeting uncultivated cluster in the ML tree and two separate clusters in the MCC tree.
species in SEA/Far East and possibly India may lead to the iden- This degree of clustering is indicative of these viruses all having
tification of both alternative BBTV host species and numerous originated from one or two founder viruses, a pattern that
other epidemiologically relevant babuvirus species. Indeed, the strongly supports the occurrence of infrequent BBTV introduc-
only other known babuvirus species have been identified from tion events into each of these countries/territories.
this region: CBDV from India (Mandal et al. 2013) and ABTV The Bayesian MCMC analysis depicted in the MCC tree indi-
from the Philippines and Malaysia (Sarawak) (Sharman et al. cated that the present global distribution of BBTV isolates could
2008). be accounted for by as few as fourteen individual movement
events between eight statistically supported (i.e., with an asso-
ciate BF> 5.0) pairs of locations (Fig. 7; Supplementary Fig. S7).
3.5 Analysis of geographical structure within BBTV Although the first of these events likely involved a movement
phylogenies from SEA to the Indian subcontinent approximately 1,000 years
Banana domestication is thought to have occurred on the ago, the most recent involved a movement from SEA to Egypt
Southeast Asian peninsula or its adjacent islands between 7 approximately 30 years ago (Fig. 7, Supplementary Fig. S7,
and 10,000 years ago (Denham et al. 2003; Perrier et al. 2011). Supplementary Data—animated kml files).
Given the potential for human-mediated dissemination of BBTV Although SEA is a BBTV diversity hotspot and is identified in
via the subsequent worldwide movements of infected banana our analysis as the most probable location of the BBTV MRCA
propagules, we aimed to examine the degree to which the phy- (probability¼ 0.557; Fig. 7), the Indian subcontinent was inferred
logenetic relationships between BBTV isolates reflected their to be the most highly connected location (involved in five of the
geographic origins. After removing genome components in the eight statistically supported links) and is inferred to be a major
FGD dataset that had been derived through reassortment and hub of long-distance BBTV movements: it is both the major do-
fragments of components that had been derived through re- nor location for BBTV dispersal events to other parts of the
combination in both the FGD and CSDs (to, respectively, yield world (seven of the fourteen supported movements) and the
recombination-free datasets, RF-FGD and RF-CSDs), we major recipient location of virus introductions (four of the
12 | Virus Evolution, 2015, Vol. 1, No. 1
66in-IN-2012 B1
41to-TO-2010 C1 Branch support
44to-TO-2010 C1
45to-TO-2010 C1 >95%
46to-TO-2010 C1
TOS90-TO-2010 C1 >80 - 95%
42to-TO-2010 C1
43to-TO-2010 C1 >60 - 80%
KP4-TO-1990 C1
Q570-TO-1990 C1
TOS48-TO-2010 C1
TOS39-TO-2010 C1 Pacific Islands
TOS63A-TO-2010 C1
TOS91-TO-2010 C1
TOS65-TO-2010 C1 USA (Hawaii)
536-TO-1993 C1
TOS42-TO-2010 C1 Tonga
TOS60-TO-2010 C1
Q277-TO-1989 C1
Q276-TO-1989 C1 Fiji
Q278-TO-1989 C1
TOS20-TO-2010 C1
TOS25-TO-2010 C1 Samoa
TOS21-TO-2010 C1
38to-TO-2010 C1
TOS29-TO-2010 C1 Australia
39to-TO-2010 C1
35to-TO-2010 C1 C1
TOS83-TO-2010 C1
36toTO-2010 C1 Southeast Asia
37to-TO-2010 C1
TOS2-TO-2010 C1 Indonesia
40to-TO-2010 C1
TOS93-TO-2010 C1 
TOS56-TO-2010 C1 Vietnam
TOS22-TO-2010 C1
TOS16-TO-2010 C1 Thailand
26pk-PK-2004 C2
51in-IN C2
Q281-WS-1989 C1 Taiwan
Q279-WS-1989 C1
KP9-US-1990 C1
527-US-1992 C1 Philippines
4au-AU C3
1900A-AU-2006 C3 China
1900B-AU-2006 C3
KP17-AU-2010 C3
KP18-AU-2010 C3 Japan
B2847-AU-2011 C3
B2845-AU-2011 C3
2557-AU-2010 C3 Indian subcontinent
482-96-AU-1996 C3
482-98-AU-1998 C3 Myanmar
482-97-AU-1997 C3
KP15-AU-2009 C3
KP16-AU-2009 C3 Pakistan
KP6-AU-2011 C3
KP14-AU-2009 C3
B2824-AU-2011 C3 India
B2819-AU-2011 C3
B2817-AU-2011 C3 Sri Lanka
B2829-AU-2011 C3
B2822-AU-2011 C3
B2821-AU-2011 C3 Africa
B2825-AU-2011 C3
3in-IN-2007 C2
65lk-LK-2010 C2 Egypt
33in-IN-2002 C2
Q524-1-IN C2 Gabon
Q524-3-IN C2
47mw-MW-2008 C2
Q553-LK-1995 C2 Cameroon
550-CG-1995 C2
BU19-CD-2012 C2 Congo
BU18-CD-2012  C2
BU14-CD-2012 C2
BU16-CD-2012 C2 DR Congo
BU20-CD-2012 C2
BU10-CD-2012 C2
BU15-CD-2012 C2 Burundi
BU11-CD-2012 C2 
BU17-CD-2012 C2 Rwanda
BU13-CD-2012 C2
BU12-CD-2012 C2
BU6-CD-2012 C2 Malawi
BU7-CD-2012 C2
BU2-CD-2012 C2
526-BI-1992 C2
549-BI-1995 C2
736-4-IN-1997 C2
548-BI-1995 C2
BU1-CD-2012 C2
BU9-CD-2012 C2
20rw-RW-2009 C2
19rw-RW-2009 C2
Q529-4-CN-1990 E1
21cn-CN D1
62cn-CN D2
63cn-CN D3
23cn-CN-2008 D4
MP2-TW-1996 D6
627-TW-1996 D8
Q1160-TW-1995 D5
Q568-1-ID-1995 D5
Q623-TW-1996 D7
Q624-TW-1996 D5
765-TW-1996 D5
25tw-TW D5
8-150510-EG-2010 A1
520-ID-1995 D5
571-1-PH-1993 D5
523-6B-IN-1991 D8
0.09 nucleotide substitutions per site MS6-PH-2008 D5
MS16-PH-2008 D5
MS18-PH-2008 D5
MS15-PH-2008 D5
MS17-PH-2008 D5
MS7-PH-2008 D5
571-2-PH-1993 D5
522A-PH-1991 D5
522B-PH-1991 D5
Figure 6. An RAxML tree of the FGD after all recombination and reassortment sequences were removed. ABTV was used to root the phylogenetic tree. Branches with
<60 per cent bootstrap support have been collapsed. Full genomes are shown with isolate name, two letter country code, year of collection, and group name. GenBank
accession numbers of the components which constitute each full genome can be found in Supplementary Table S1.
SPG AG
D. Stainton et al. | 13
A KC581796-TH-2012
Q529-2-CN-1990
Q529-4-CN-1990
Q529-6-CN-1990
Q529-5-CN-1990
Q529-1-CN-1990
GQ374514-CN-2009
23cn-CN-2008
625I-TW-1995
MP1-TW-1996
MP2-TW-1996 D6
626M-TW-1995
Q568-1-ID-1995
Q568-3-ID-1995
Q280-R-WS-1989
1977 JN003633-ID-2010
JN003632-ID-2010
JN003631-ID-2010
765-TW-1996
8-150510-EF-2010
625-TW-1996
Q624-TW-1996 D5
1982 Q623-TW-1996 D7
Q1160-TW-1995 D5
571-1-PH-1993 D5
523-6B-IN-1991 D8
571-1-PH-1993 D5
1976 520-ID-1995
523-6A-IN-1991 D8
MS6-PH-2008 D5
1980 MS16-PH-2008 D5
MS15-PH-2008 D5
MS18-PH-2008 D5
MS17-PH-2008 D5
MS14-PH-2008
522A-PH-1991 D5
571-2-PH-1993 D5
522B-PH-1991 D5
768-PH-1995
MS7-PH-2008 D5
28in-IN-2012
66in-IN-2012 B2
TOS59TO-2010
TOS70-TO-2010
TOS5TO-2010
TOS63A-TO-2010 C1
926 TOS69-TO-2010
TOS93-TO-2010 C1 
44to-TO-2010 C1
45to-TO-2010 C1
46to-TO-2010 C1
41to-TO-2010 C1
TOS64TO-2010
1882 TOS55TO-2010
KP4-TO-1990 C1
Q570-TO-1990 C1
TOS77TO-2010
TOS90-TO-2010 C1
42to-TO-2010 C1
43to-TO-2010 C1
TOS63B-TO-2010
TOS87-TO-2010
35to-TO-2010 C1 C1
TOS83-TO-2010 C1
TOS43-TO-2010
TOS42-TO-2010 C1
TOS89-TO-2010
TOS74-TO-2010
TOS67-TO-2010
TOS57-TO-2010
TOS65-TO-2010 C1
TOS91-TO-2010 C1
TOS80-TO-2010
TOS71-TO-2010
TOS72-TO-2010
TOS40-TO-2010
TOS92-TO-2010
TOS48-TO-2010 C1
TOS39-TO-2010 C1
TOS34-TO-2010
TOS45-TO-2010
TOS49-TO-2010
1619 Q276-TO-1989 C1
Q277-TO-1989 C1
TOS68-TO-2010
USA (Hawaii) Q278-TO-1989 C1
Posterior branch TOS88-TO-2010
TOS78-TO-2010
TOS86-TO-2010
Tonga support 536-TO-1993 C1
TOS60-TO-2010 C1
TOS85-TO-2010
>0.95 TOS19-TO-2010
40to-TO-2010 C1
TOS76-TO-2010
Samoa >0.90 - 0.95 38to-TO-2010 C1
39to-TO-2010 C1
TOS29-TO-2010 C1
TOS7-TO-2010
TOS20-TO-2010 C1
TOS25-TO-2010 C1
TOS21-TO-2010 C1
Australia TOS22-TO-2010 C1
TOS16-TO-2010 C1
TOS56-TO-2010 C1
TOS61-TO-2010
TOS62-TO-2010
TOS82-TO-2010
Southeast Asia TOS14-TO-2010
TOS79-TO-2010
TOS15-TO-2010
37to-TO-2010 C1
TOS46-TO-2010
TOS2-TO-2010 C1
Indian subcontinent TOS58-TO-2010
TOS4-TO-2010
TOS53-TO-2010
1941 36toTO-2010 C1
737-AU-1997
Africa 1900B-AU-2006 C3
1735 1900A-AU-2006 C3
602-AU-1996
KP17-AU-2010 C3
KP18-AU-2010 C3
2557-AU-2010 C3
B2846-AU-2011
B2847-AU-2011 C3
B2844-AU-2011
B2845-AU-2011 C3
482-96-AU-1996 C3
482P2-2011
482-97-AU-1997 C3
482-98-AU-1998 C3
B2834-AU-2011
B2832-AU-2011
B 0.6 KP16-AU-2009 C3
KP15-AU-2009 C3
KP14-AU-2009 C3
KP6-AU-2011 C3
1843 B2822-AU-2011 C3
B2828-AU-2011
B2821-AU-2011 C3
B2830-AU-2011
0.5 B2825-AU-2011 C3
B2826-AU-2011
B2824-AU-2011 C3
B2829-AU-2011 C3
B2817-AU-2011 C3
B2833-AU-2011
B2823-AU-2011
B2827-AU-2011
0.4 1925 B2818-AU-2011
B2819-AU-2011 C3
B2820-AU-2011
65lk-LK-2010
33in-IN-2002 C2
KP5-LK-2003
9-150510-EG-2010
Q281-WS-1989 C1
0.3 Q279-WS-1989 C1
1915 527-US-1992 C1
KP9-US-1990 C1
3in-IN-2007
HM120718-IN-2009
1961 AM418565-PK-2004
AM418567-PK-2004
0.2 AM418535-PK-2004
AM418534-PK-2004
1pk-PK-2004
AM418537-PK-2004
AM418539-PK-2004
2pk-PK-2004
50pk-PK-2007
52pk-PK-2007
0.1 48pk-PK-2007
54pk-PK-2007
59pk-PK-2007
57pk-PK-2007
53pk-PK-2007
26pk-PK-2004 C2
61pk-PK-2007
58pk-PK-2007
0.0 49pk-PK-2007
1825 1955 56pk-PK-2007
55pk-PK-2007
27in-IN-2006
FJ009240-IN-2007
64in-IN-2009
KP7-AU-1989
KP8-AU-1989
Q553-LK-1995 C2
550-CG-1995 C2
47mw-MW-2008 C2
Bayes factor 1929 BU11-CD-2012 C2
BU12-CD-2012 C2
C BU19-CD-2012 C2 
Africa - Indian subcontinent 2374.8 1936 BU18-CD-2012  C2
BU10-CD-2012 C2
BU17-CD-2012 C2
Asia - Indian subcontinent 1186.3 BU13-CD-2012 C2
BU16-CD-2012 C2
BU20-CD-2012 C2
Australia - Indian subcontinent 180.7 BU15-CD-2012 C2
BU14-CD-2012 C2
BU6-CD-2012 C2
USA (Hawaii) - Samoa 101.2 BU2-CD-2012 C2
BU7-CD-2012 C2
19rw-RW-2009 C2
Asia - Samoa 37.5 20rw-RW-2009 C2
BU1-CD-2012 C2
BU9-CD-2012 C2
Indian subcontinent - Samoa 16.0 526-BI-1992 C2
549-BI-1995 C2
736-4-IN-1997 C2
Indian subcontinent - Tonga 12.2 547-BI-1995
1972 548-BI-1995 C2
Africa - Asia 5.3 JF755978-CM-2008
GQ249344-CM-2008
34eg EG-1997
JF755986-CD-2008
1934 JF755980-MW-2008
JF755987-CD-2008
HQ259074-EG-2008
JF755982-GA-2008
912 1012 1262 1512 1762 2012
Year
Figure 7. (A) An MCC tree constructed using the BBTV recombination-free CD (RF-CD) with a constant population size, strict-clock, discrete diffusion model. Branches
are colored according to locations and the inferred dates (red font) of the statistically supported BBTV movement events are indicated with red arrows. Black circle on
nodes indicate posterior branch support with >0.95 support and gray circles with >0.9–0.95 support. (B) Location probabilities for the root node of the tree are provided
in the color-coded bar graph, and those obtained with randomization of the tip locations are shown as gray bars for each location. (C) The statistically supported epide-
miological linkages between locations and their associated BF support values inferred using the Bayesian discrete diffusion phylogeography model are summarized.
root location probability
14 | Virus Evolution, 2015, Vol. 1, No. 1
fourteen movements). These include two dispersal events from Acknowledgements
the Indian subcontinent to Sub-Saharan Africa between 1825
and 1934, and one to Egypt (between 1929 and 1936), Australia We thank the students of Tonga College for their help in
(two events between 1843 and 1974), Tonga (one event between some of the sample collection in the Kingdom of Tonga. D.S
1735 and 1882), and Samoa (one event between 1915 and 1934), was supported by a postgraduate scholarship from the
and introduction events from SEA (the oldest between 926 and Marsden Fund of New Zealand (UOC0903). S.K. was sup-
1619, and two more recent events between 1976 and 1991) and ported by a scholarship from the School of Biological
Africa (between 1972 and 1997). Some of these movements may Sciences (University of Canterbury, New Zealand). D.P.M.,
have been indirect. For example, reports indicate that BBTV G.W.H., and A.V. are supported by the National Research
most likely reached Australia from Fiji shortly before 1913 Foundation of South Africa. P.L. acknowledges support from
(Magee 1927). However, there were strong colonial and cultural the European Union Seventh Framework Programme [FP7/
links between India, Fiji, and Australia, and it is feasible that 2007-2013] under Grant Agreement no. 278433-PREDEMICS
movement to Australia was from the Indian sub-continent to and ERC Grant agreement no. 260864. J.T., M.S., and K.C. ac-
Australia, via Fiji. However, Fijian sequences are poorly repre- knowledge the support of Horticulture Innovation Australia
sented in this study and may have failed to reveal such an inter- (formerly Horticulture Australia Limited). This work was
mediate step. supported by the Marsden Fund Council from Government
Other notable statistically supported viral dispersal events funding, administered by the Royal Society of New Zealand
include one from Samoa to Hawaii between 1961 and 1978 (first (grant UOC0903) awarded to A.V.
disease report in Hawaii was in 1989) and SEA to Africa (be-
tween 1982 and 2010). Conflict of interest: None declared.
Therefore, although the global patterns of geographic struc-
ture that are evident within the BBTV phylogeny are not at References
all consistent with BBTV having been spread during the pre-
historic dissemination of bananas across the globe, neither Almeida, R. P. et al. (2009) ‘Spread of an Introduced Vector-Borne
are they consistent with frequent, human-mediated trans- Banana Virus in Hawaii’, Molecular Ecology, 18: 136–46.
continental BBTV movements during the past few decades. Altschul, S. F. et al. (1990) ‘Basic Local Alignment Search Tool’,
They are, however, entirely consistent with more gradual, natu- Journal of Molecular Biology, 215: 403–10.
ral, or human-facilitated movements of the virus (via infected Aronson, M. N. et al. (2000) ‘Clink, a Nanovirus-Encoded Protein,
propagules and/or aphids) over the past 300 years from its cen- Binds Both pRB and SKP1’, Journal of Virology, 74: 2967–72.
ters of diversity in India, and SEA/Far East across the banana Banerjee, A. et al. (2014) ‘Identification and Characterization of
growing regions of the world. Additionally, there is clear evi- a Distinct Banana Bunchy Top Virus Isolate of Pacific-
dence within our RF-FGD ML and RF-CD MCC trees, of frequent Indian Oceans Group from North-East India’, Virus Research,
shorter distance BBTV movements. For example, in the AG 183: 41–9.
clade sequences sampled from Taiwan, China, the Indian Bielejec, F. et al. (2011) ‘SPREAD: Spatial Phylogenetic
Subcontinent and Indonesia intermingle with one another Reconstruction of Evolutionary Dynamics’, Bioinformatics, 27:
within well-supported clades: a pattern which suggests that 2910–12.
during the past 100 years there must have been multiple BBTV Boni, M. F., Posada D., and Feldman M. W. (2007) ‘An Exact
movements between these locations. Nonparametric Method for Inferring Mosaic Structure in
Sequence Triplets’, Genetics, 176: 1035–47.
4 Concluding remarks Burns, T. M., Harding R. M., and Dale J. L. (1994) ‘Evidence that
Banana Bunchy Top Virus has a Multiple Component
Here, we have studied the landscape of global BBTV diversity to Genome’, Archives of Virology, 137: 371–80.
reveal that the Indian Subcontinent, SEA, and the Far East are ——, ——, and ——. (1995) ‘The Genome Organisation of Banana
the current BBTV diversity hotspots. Accounting for recombina- Bunchy Top Virus: Analysis of Six ssDNA Components’, Journal
tion and reassortment, we phylogenetically analyzed 855 newly of General Virology, 76: 1471–82.
sequenced full BBTV genome components together with all Dale, J. L. (1987) ‘Banana Bunchy Top—An Economically
available complete genome component sequences presently Important Tropical Plant-Virus Disease’, Advances in Virus
available in GenBank. Research, 33: 301–25.
We find that the global distribution of BBTV genotypes is Denham, T. P. et al. (2003) ‘Origins of Agriculture at Kuk Swamp
highly structured at the continental scale, suggesting that hu- in the Highlands of New Guinea’, Science, 301: 189–93.
man-mediated inter-continental transfers of epidemiologically Drummond, A. J., and Rambaut A. (2007) ‘BEAST: Bayesian
important BBTV genotypes have occurred relatively infre- Evolutionary Analysis by Sampling Trees’, BMC Evolutionary
quently. Rather than the current global distribution of BBTV be- Biology 7: 214.
ing attributable to either frequent long-distance movements of Duay, J. A. M. et al. (2014) ‘Pentalonia Nigronervosa Coquerel and
diseased banana planting material over the past few decades or Pentalonia caladii van der Goot (Hemiptera: Aphididae) and
the pre-historic spread of BBTV along with the spread of banana Their Relationship to Banana Bunchy Top Virus in Micronesia’,
cultivation, our results suggest that the current distribution of Pacific Science, 68: 359–64.
the virus was primarily attained through infrequent movement Duffy, S., and Holmes E. C. (2008) ‘Phylogenetic Evidence for
events over the past 300 years primarily from its diversity hot- Rapid Rates of Molecular Evolution in the Single-Stranded
spots in India and SEA. DNA Begomovirus Tomato Yellow Leaf Curl Virus’, Journal of
Virology, 82: 957–65.
Supplementary data ——, Shackelton L. A., and Holmes E. C. (2008) ‘Rates of
Evolutionary Change in Viruses: Patterns and Determinants’,
Supplementary data is available at VEVOLU Journal online. Nature Reviews Genetics, 9: 267–76.
D. Stainton et al. | 15
Edgar, R. C. (2004) ‘MUSCLE: Multiple Sequence Alignment with Izquierdo-Carrasco, F., Smith S. A., and Stamatakis A. (2011)
High Accuracy and High Throughput’, Nucleic Acids Research, ‘Algorithms, Data Structures, and Numerics for Likelihood-
32: 1792–7. Based Phylogenetic Inference of Huge Trees’, BMC
Firth, C. et al. (2009) ‘Insights into the Evolutionary History of an Bioinformatics, 12: 470.
Emerging Livestock Pathogen: Porcine Circovirus 2’, Journal of Karan, M., Harding R. M., and Dale J. L. (1994) ‘Evidence for Two
Virology, 83: 12813–21. Groups of Banana Bunchy Top Virus Isolates’, Journal of General
Foottit, R. G. et al. (2010) ‘The Identity of Pentalonia Virology, 75: 3541–6.
Nigronervosa Coquerel and P. caladii van der Goot (Hemiptera: Kass, R. E., and Raftery A. E. (1995) ‘Bayes Factors’, Journal of
Aphididae) Based on Molecular and Morphometric Analysis’, American Statistical Association, 90: 773–95.
Zootaxa, 2358: 25–38. King, A. M. et al. (2012) Virus Taxonomy: Ninth Report of the
Fu, H. C. et al. (2009) ‘Unusual Events Involved in Banana Bunchy International Committee on Taxonomy of Viruses. Burlington:
Top Virus Strain Evolution’, Phytopathology, 99: 812–22. Elsevier Science.
Geering, A. D. W., and Thomas J. E. (1997) ‘Search for Alternative Kraberger, S. et al. (2013) ‘Evidence that Dicot-Infecting
Hosts of Banana Bunchy Top Virus in Australia’, Australasian Mastreviruses are Particularly Prone to Inter-Species
Plant Pathology, 26: 250–4. Recombination and have Likely been Circulating in Australia
Gibbs, M. J., Armstrong J. S., and Gibbs A. J. (2000) ‘Sister- for Longer than in Africa and the Middle East. Virology, 444:
Scanning: A Monte Carlo Procedure for Assessing Signals in 282–91.
Recombinant Sequences’, Bioinformatics, 16: 573–82. Lefeuvre, P. et al. (2009) ‘Widely Conserved Recombination
Grigoras, I. et al. (2010) ‘High Variability and Rapid Evolution of a Patterns among Single-Stranded DNA Viruses’, Journal of
Nanovirus’, Journal of Virology, 84: 9105–17. Virology, 83: 2697–707.
—— et al. (2014) ‘Genome Diversity and Evidence of Lejju, B. J., Robertshaw P., and Taylor D. (2006) ‘Africa’s Earliest
Recombination and Reassortment in Nanoviruses from Bananas?’, Journal of Archaeological Science, 33: 102–13.
Europe’, Journal of General Virology, 95: 1178–91. Lemey, P. et al. (2009) ‘Bayesian Phylogeography Finds Its Roots’,
Guindon, S. et al. (2010) ‘New Algorithms and Methods to PLoS Computational Biology, 5: e1000520.
Estimate Maximum-Likelihood Phylogenies: Assessing the Magee, C. J. P. (1927) ‘Investigation on the Bunchy Top Disease of
Performance of PhyML 3.0’, Systematic Biology, 59: 307–21. the Banana’, Bulletin of the Council for Scientific and Industrial
Hafner, G. J., Harding R. M., and Dale J. L. (1997) ‘A DNA Primer Research in Australia, 30.
Associated with Banana Bunchy Top Virus’, Journal of General —— (1940) ‘Transmission Studies on the Banana Bunchy-Top
Virology, 78: 479–86. Virus’, Journal of the Australian Institute of Agricultural Science, 6:
—— et al. (1997) ‘Nicking and Joining Activity of Banana Bunchy 109–10.
Top Virus Replication Protein in vitro’, Journal of General Mandal, B. et al. (2013) ‘Nine Novel DNA Components Associated
Virology, 78: 1795–9. with the Foorkey Disease of Large Cardamom: Evidence of a
Harding, R. M., Burns T. M., and Dale J. L. (1991) ‘Virus-Like Distinct Babuvirus Species in Nanoviridae’, Virus Research, 178:
Particles Associated with Banana Bunchy Top Disease Contain 297–305.
Small Single-Stranded DNA’, Journal of General Virology, 72: Manickam, K. et al. (2002) ‘Early Detection of Banana Bunchy Top
225–30. Virus in India Using Polymerase Chain Reaction’, Acta
—— et al. (1993) ‘Nucleotide Sequence of One Component of the Phytopathologica et Entomologica Hungarica, 37: 9–16.
Banana Bunchy Top Virus Genome Contains a Putative Martin, D., and Rybicki E. (2000) ‘RDP: Detection of
Replicase Gene’, Journal of General Virology, 74: 323–8. Recombination amongst Aligned Sequences’, Bioinformatics, 16:
Harkins, G. W. et al. (2009) ‘Experimental Evidence Indicating 562–3.
that Mastreviruses Probably did not Co-Diverge with Their Martin, D. P. et al. (2005) ‘A Modified Bootscan Algorithm for
Hosts’, Virology Journal, 6: 104. Automated Identification of Recombinant Sequences and
—— et al. (2014) ‘Towards Inferring the Global Movement of Beak Recombination Breakpoints’, Aids Research and Human
and Feather Disease Virus’, Virology, 450-451: 24–33. Retroviruses, 21: 98–102.
Herrera-Valencia, V. A. et al. (2006) ‘An Iterated Sequence in the —— (2010) ‘RDP3: A Flexible and Fast Computer Program for
Genome of Banana Bunchy Top Virus is Essential for Efficient Analyzing Recombination’, Bioinformatics, 26: 2462–3.
Replication’, Journal of General Virology, 87: 3409–12. Muhire, B. M., Varsani A., and Martin D. P. (2014) ‘SDT: A Virus
Hooks, C. R. R. et al. (2008) ‘Effect of Banana Bunchy Top Virus Classification Tool Based on Pairwise Sequence Alignment and
Infection on Morphology and Growth Characteristics of Identity Calculation’, PLoS One, 9: e108277.
Banana’, Annuals of Applied Biology, 153: 1–9. Owor, B. E. et al. (2007) ‘Successful Application of FTAVR Classic
Hu, J. M. et al. (2007) ‘Reassortment and Concerted Evolution in Card Technology and Use of Bacteriophage Phi 29 DNA
Banana Bunchy Top Virus Genomes’, Journal of Virology, 81: Polymerase for Large-Scale Field Sampling and Cloning of
1746–61. Complete Maize Streak Virus Genomes’, Journal of Virological
Hu, J. S. et al. (1996) ‘Use of Polymerase Chain Reaction (PCR) to Methods, 140: 100–5.
Study Transmission of Banana Bunchy Top Virus by the Padidam, M., Sawyer S., and Fauquet C. M. (1999) ‘Possible
Banana Aphid (Pentalonia nigronervosa)’, Annals of Applied Emergence of New Geminiviruses by Frequent
Biology, 128: 55–64. Recombination’, Virology, 265: 218–25.
Hyder, M. Z. et al. (2011) ‘Evidence of Recombination in the Perrier, X. et al. (2011) ‘Multidisciplinary Perspectives on
Banana Bunchy Top Virus Genome’, Infection Genetics and Banana (Musa spp.) Domestication’, Proceedings of the
Evolution, 11: 1293–300. National Academy of Sciences of the United States of America, 108:
Islam, M. N. et al. (2010) ‘Genetic Diversity and Possible 11311–8.
Evidence of Recombination among Banana Bunchy Top Virus Pinili, M. S. et al. (2013) ‘Cross-Transmission and New Alternate
(BBTV) Isolates’, International Research Journal of Microbiology, 1: Hosts of Banana Bunchy Top Virus’, Tropical Agriculture and
001–12. Development, 57: 1–7.
16 | Virus Evolution, 2015, Vol. 1, No. 1
Posada, D. (2008) ‘jModelTest: Phylogenetic Model Averaging’, Reassortment in Banana Bunchy Top Virus’, Journal of General
Molecular Biology and Evolution, 25: 1253–6. Virology, 93: 1103–19.
—— and Crandall, K. A. (2001) ‘Evaluation of Methods for Stamatakis, A. (2014) ‘RAxML Version 8: A Tool for Phylogenetic
Detecting Recombination from DNA Sequences: Computer Analysis and Post-Analysis of Large Phylogenies’,
Simulations’, Proceedings of the National Academy of Sciences of Bioinformatics, 30: 1312–3.
the United States of America, 98: 13757–62. Su, H.-J., Wu R.-Y., and Tsao L.-Y. (1992) ‘Ecology of Banana
—— and —— (2002) ‘The Effect of Recombination on the Bunchy-Top Virus Disease’, in Proceedings of the International
Accuracy of Phylogeny Estimation’, Journal of Molecular Symposium on Recent Developments in Banana Cultivation
Evolution, 54: 396–402. Technology, Los Baños, Philippines: INIBAP-ASPNET, pp. 308–12.
Ram, R. D., and Summanwar A. S. (1984) ‘Colocasia esculenta (L.) Tamura, K. et al. (2011) ‘MEGA5: Molecular Evolutionary Genetics
schott. A Reservoir of Bunchy Top Disease of Banana’, Current Analysis Using Maximum Likelihood, Evolutionary Distance,
Science, 53: 145–6. and Maximum Parsimony Methods’, Molecular Biology and
Randrianja, S., and Ellis S. (2009) Madagascar: A Short History. Evolution, 28: 2731–9.
Chicago, IL: University of Chicago Press. Thomas, J. E., and Dietzgen R. G. (1991) ‘Purification,
Rybicki, E. P. (2015) ‘A Top Ten List for Economically Important Characterization and Serological Detection of Virus-Like
Plant Viruses’, Archives of Virology, 160: 17–20. Particles Associated with Banana Bunchy Top Disease in
—— and Pietersen G. (1999) ‘Plant Virus Disease Problems Australia’, Journal of General Virology, 72: 217–24.
in the Developing World’, Advances in Virus Research, 53: van der Walt, E. et al. (2008) ‘Experimental Observations of Rapid
127–75. Maize Streak Virus Evolution Reveal a Strand-Specific
Savory, F. R., and Ramakrishnan U. (2014) ‘Asymmetric Patterns Nucleotide Substitution Bias’, Virology Journal, 5: 104.
of Reassortment and Concerted Evolution in Cardamom Bushy Varsani, A. et al. (2014) ‘Establishment of Three New Genera in
Dwarf Virus’, Infection, Genetics and Evolution, 24: 15–24. the Family Geminiviridae: Becurtovirus, Eragrovirus and
Schierup, M. H., and Hein J. (2000) ‘Consequences of Turncurtovirus’, Archives of Virology, 159: 2193–203.
Recombination on Traditional Phylogenetic Analysis’, Genetics, Wang, H. I. et al. (2013) ‘Application of Motif-Based Tools on
156: 879–91. Evolutionary Analysis of Multipartite Single-Stranded DNA
Sharman, M. et al. (2008) ‘Abaca Bunchy Top Virus, a New Viruses’, PLoS One, 8: e71565.
Member of the Genus Babuvirus (family Nanoviridae)’, Wanitchakorn, R., Harding R. M., and Dale J. L. (2000) ‘Sequence
Archives of Virology, 153: 135–47. Variability in the Coat Protein Gene of Two Groups of Banana
Shepherd, D. N. et al. (2008) ‘A Protocol for the Rapid Isolation of Bunchy Top Isolates’, Archives of Virology, 145: 593–602.
Full Geminivirus Genomes from Dried Plant Tissue’, Journal of —— et al. (2000) ‘Functional Analysis of Proteins Encoded by
Virological Methods, 149: 97–102. Banana Bunchy Top Virus DNA-4 to -6’, Journal of General
Smith, J. M. (1992) ‘Analyzing the Mosaic Structure of Genes’, Virology, 81: 299–306.
Journal of Molecular Evolution, 34: 126–9. Yu, N. T. et al. (2012) ‘Cloning and Sequence Analysis of Two
Stainton, D. et al. (2012) ‘Evidence of Inter-Component Banana Bunchy Top Virus Genomes in Hainan’, Virus Genes, 44:
Recombination, Intra-Component Recombination and 488–94.