Arch Virol (2007) 152: 14091416

DOI 10.1007/s00705-007-0946-9
Printed in The Netherlands

Brief Report
Molecular characterization of banana streak acuminata Vietnam
virus isolated from Musa acuminata siamea (banana cultivar)
F. Lheureux1 , N. Laboureau1 , E. Muller1, B. E. L. Lockhart2 , and M.-L. Iskra-Caruana1
1
2

CIRAD=UMR BGPI TA A54=K, Montpellier, France

Department of Plant Pathology, University of Minnesota, St. Paul, MN, U.S.A.

Received November 14, 2006; accepted January 22, 2007; published online April 16, 2007
# Springer-Verlag 2007

Summary

An isolate of banana streak virus (BSV) that does

not also occur as an integrant in the Musa balbisiana
genome was sought in order to investigate the biological role of BSV in the evolution of either the
Musa genome or of the virus itself. We isolated
BSV virions from a Musa acuminata siamea accession from Vietnam and sequenced the entire viral
genome. The molecular organization is similar to
that described for other BSV but slightly larger
(7801 bp vs. 16117568 bp), and ORF I has a
non-conventional start codon. This genome was
sufficiently different to propose it as a member of
a distinct species named Banana streak virus strain
acuminata Vietnam (BSAcVNV).


Banana streak virus (BSV) is one of the most
widely distributed viruses of plantains and bananas
[21, 22] causing chlorosis leaf streak disease.
BSV is a pararetrovirus belonging to the genus
Authors address: Marie-Line Iskra-Caruana, CIRAD=
UMR BGPI TA A54=K Campus International de Baillarguet
F-34398 Montpellier, Cedex 5, France. e-mail: marie-line.
caruana@cirad.fr

Badnavirus, family Caulimoviridae [14]. Like all badnaviruses, BSV has bacilliform-shaped virions 30
150 nm in size and a circular dsDNA genome which
is replicated by reverse transcription involving an
RNA intermediate [12, 19]. Its genome contains
three open reading frames (ORFs) on one strand.
ORF I and ORF II potentially encode two small
proteins of unknown function of 20.8 and 14.5 kDa,
respectively. ORF III encodes a polyprotein of
208 kDa consisting of the putative cell-to-cell
movement protein, the coat protein, the aspartic
protease, and the viral replicase, which has reverse
transcriptase (RT) and ribonuclease H (RNase H)
functions [12]. This polyprotein is thought to be
post-translationally cleaved into functional units
by the aspartic protease. There is pronounced antigenic variability as well as molecular variability
among BSV isolates, as reported for species of the
genus Badnavirus. Lockhart and Olszewski [20]
recognized at least three distinct serotypes among
four different isolates based on PCR amplification
with degenerate primers followed by DNA hybridization analyses. Geering et al. [6] observed 21.8
33.6% sequence differences in the ribonuclease
H and reverse transcriptase coding regions of three
BSV isolates in Australia. Virus isolates showing

1410

such a degree of difference in this conserved region

are now considered to be members of different BSV
species according to current ICTV criteria (2005)
for species demarcation in the genus Badnavirus.
The high degree of molecular variability observed
makes reliable detection of BSV difficult and raises
questions about the taxonomic designations within
this genus.
Fifteen years ago, an increasing record of BSV
outbreaks were observed among banana breeding
lines and micro-propagated interspecific Musa hybrids [2, 18, 26] worldwide. These infections were
mainly linked to the presence of viral DNA sequences in the Musa balbisiana genome (denoted
B). These viral integrants, called endogenous pararetroviruses (EPRV), have also been described in
the nuclear genome of several other species including tobacco, petunia and rice [10, 17, 30]. Although
integration is not an essential step in the replication
cycle of pararetroviruses, it is assumed that, under
stress conditions, some EPRVs could become infectious by reconstituting a complete replicationcompetent viral genome [4, 13, 18, 27]. Two types
of BSV integrants exist in banana. The first type
comprises non-functional sequences present in both
common Musa species, Musa acuminata (denoted
A) and Musa balbisiana (denoted B) [8]. Integrants
of the other type are infectious, they contain the
complete viral genome and are only present in the B
chromosome. Three widespread BSVs, banana streak
Obino lEwai virus (BSOlV), banana streak Imov
e
virus (BSImV) [7] and banana streak Golfinger virus (BSGfV) are known to occur as infectious integrants in the M. balbisiana chromosome ([15,
18], Iskra-Caruana, unpublished results). However,
even though such integrations are known to be infectious, their presence is not sufficient to induce
infection, and wild M. balbisiana diploids such as
Pisang Klutug Wulung (PKW) harboring pathogenic BSV EPRVs are resistant to the multiplication of
BSV ([18], Lheureux and Lockhart, unpublished
results). Their results were obtained by exploiting
the BAC library of diploid M. balbisiana genitors
and BSV expression among the triploid inter-specific progeny from virus-free PKW (BB) and IDN
110 T (AAAA) crosses ([18], Iskra-Caruana, unpublished results). It is now assumed that EPRVs have

F. Lheureux et al.

integrated into the host genome through a process

of illegitimate recombination and appear to be
relics of ancient infection events [8].
It is essential to provide baseline information
needed to tackle the problem of the biological role
of EPRVs in the evolution of either the Musa
genome or BSV by identifying and characterizing
non-integrated BSV species. Because episomal
BSV species described so far are all associated with
M. balbisiana genome integrations, we only searched
for BSV isolates in M. acuminata accessions. The
aim of this study was to characterize the genome of
one such isolate, banana streak acuminata Vietnam
virus (BSAcVNV) and to provide preliminary evidence that partial viral integrations in the host genome are restricted to the M. acuminata genome.
Twenty BSV-positive Musa accessions were
checked by immunosorbent electron microscopy
(ISEM) using partially purified leaf extracts [18]
and then by immunocapture PCR (IC-PCR) using
five specific BSV species primers [6]. ITC accession
number 1431 M. acuminata siamea from Vietnam,
found to be ISEM positive and IC-PCR negative,
was used as the source for this study. The BSV ITC
1431 isolate was first transmitted by Planococcus
citri mealybug to three banana plants of the cv
Cavendish (AAA) in order to obtain enough infected plant material to purify and characterize
the virus. Three weeks later, the inoculated plants
showed specific yellow streak symptoms, and bacilliform viral particles were observed under an electron
microscope, confirming BSV infection. Molecular
characterization of the genome consisted of sequencing the full-length genome of BSAcVNV.
A short fragment of 1.5 kbp was first partially
amplified by PCR using the set of degenerate primers Badna T and Badna 2 [20] on DNA extracted
from a purified virus preparation [1, 6], then sequenced and used to design primers to perform
one-step PCR amplification of the total viral DNA
as previously described [25]. The following outward primers were designed to overlap at their
50 end extremity corresponding to the sequence
of a PstI restriction site VnPst1 50 CTG CAG
AGG CAA CTA TGG TCT 30 (anti-sense), VnPst2
50 CTG CAG GGG GTA TGT CAA GG 30
(sense). The ExpandTM Long Template PCR sys-

Banana streak acuminata Vietnam virus

tem (Boehringer Mannheim) was then used according to the manufacturers instructions, resulting in
amplification of the total viral DNA. The template,
consisting of 2.05.0 ml of virion DNA, was amplified following one cycle at 94  C for 2 min, 10
cycles at 94  C for 15 s, 60  C for 30 s, 68  C for
5 min 30 s, 20 cycles at 94  C for 15 s, 60  C for
30 s, 68  C for 5 min 30 s with 5 s increment=cycle
and a final extension at 72  C for 7 min. The 7.8 kbp
PCR products were gel-purified using 0.7% lowmelting agarose in TAE and GELase (Epicentre)
and cloned using the Topo XL cloning Kit (Invitrogen). Positive clones were selected by Pst1 digestion.
Plasmid DNA was then purified for sequencing using the QIAprep spin maxiprep kit (QIAGEN).
Sequencing was done by Gene Walking by Genome
Express.
Sequence analysis was done using the Vector NTI
software package, and sequences were compared
with other viral sequences of the NCBI Database
using BLAST programs. Sequences were aligned
using the CLUSTAL W algorithm, and phylogenetic neighbor-joining trees were generated by the
Darwin 4 program [28]. Robustness of trees was
determined by bootstrap sampling of multiple sequence alignment (1000 sets).
The total length of the BSAcVNV genome was
7801 bp (accession AY750155), which is greater
than those of previously characterized badnaviruses
(BSOlV: 7389 bp, BSGfV: 7263 bp, Commelina
yellow mosaic virus ComYMV: 7489 bp, cacao
swollen shoot virus Agou1 isolate CSSV-Agou1:
7161 bp, sugarcane bacilliform Mor virus SCBMV:
7568 bp, Citrus mosaic virus CMBV: 7559 bp).
The main difference occurs in the intergenic region,
which is longer (1177 bp), and in ORF I, which is
shorter (435 bp) than those recorded for the members of other BSV species. Like the other BSV
isolates and members of the family Caulimoviridae
in general, the tRNAmet binding site corresponding
to the putative 50 minus-strand priming site was
identified, and the 50 nucleotide of this site was
designated position one in the viral genome. The
BSAcVNV genome consists of three open reading
frames (ORFs) located on the plus DNA strand
(Fig. 1A). Unlike the other BSV species, BSAcVNV
has no ORF overlaps. ORF I, ORF II and ORF III

1411

are separated by 92 nt and one codon, respectively.

ORF II (nts 12831678) and ORF III (nts 1682
7375) are representative of badnavirus in general.
ORF II encodes a protein of 132 amino acid residues with a molecular weight of 14.4 kDa and does
not resemble any other protein sequence for which
the function is known. The long ORF III encodes a
putative 219.9-kDa polyprotein of 1898 amino acid
residues. The amino acid sequence of the ORF III
product has motifs suggestive of movement proteins (MP), RNA binding, aspartate protease (AP),
reverse transcriptase (RT) and ribonuclease H
(RNase H) (Fig. 2A). In contrast to the other ORFs,
ORF I (nts 7541191) begins with a non-standard
AUG start codon. ORF III and ORF I are separated
by a large intergenic region (LIGR) of 1177 nt
containing 15 small ORFs (sORF) that could be
involved in a ribosome shunt process like that proposed for members of the family Caulimoviridae.
Indeed, the fold in this region revealed a large hairpin structure of 37 nt in front of the putative ORF I
start codon. As in the case of the BSOlV LIGR, a
sORF of six codons including one internal in-frame
AUG terminates 6 nt upstream of this hairpin
(Fig. 1B). As observed with the first ORF of rice
tungro bacilliform virus (RTBV), cauliflower mosaic
virus (CaMV) and banana streak Mysore virus
(BSMyV) [5, 9], and based on the shunt model [29],
we assume that the CUG codon (nt 754) acts as the
actual start codon of ORF I. This codon is in
an optimal initiation context (ACUCUGGA) with,
firstly, a purine at position 3 with respect to
the first nucleotide of the codon and a G at position
4 [5]; secondly, an A at position 5, for which a
strong stimulatory effect on initiation at a non-ATG
start codon has been demonstrated [3, 11]. Finally,
ORF I is characterized by the absence of any
stop codon for 145 codons. The predicted putative
protein of about 17.1 kDa does not resemble any
known protein. LIGR analysis revealed a small
potential TATA box (TATAA) at nt 7, 625 to promote transcription, and, 76 nt downstream, a possible polyadenylation signal (AATAAA) similar to
those found in members of other BSV species. This
position enables the production of a terminally
redundant pre-genomic RNA (so termed full-length
transcript) that is useful as a template for both the

Fig. 1. A Proposed strategy for the polycistronic translation of the pregenomic (pg) RNA of BSAcVNV. The pgRNA leader is depicted as a thick line, open reading
frames (ORFs) as boxes, and numbers show the position in the virus genome. Symbols indicate the following: large arrowhead, the putative polyadenylation signal;
triangle, the primer-binding site (PBS) for reverse transcription; inwardly pointing, small arrows, the complementary sequences that form the base of the large stem-loop
shown in Fig. 1B. B Predicted secondary structure of the pgRNA leader. Predicted MFold, large stem-loop structure of the leader are shown schematically at right.
Stability of the structure is given in kcal=mol. The 50 and 30 sequences flanking the main structure are shown in open conformation. The stable structure element (circled
at the stem base) and adjacent regions are enlarged alongside. Nucleotide numbering refers to genome positions.

F. Lheureux et al.: Banana streak acuminata Vietnam virus

1413

Fig. 2. A Alignment of aspartate protease, reverse transcriptase, and RNase H domains of the ORF III polyprotein of various
badnavirus. The badnaviruses are banana streak virus Ol, AcVN, Gf, and My, species (BSOlV, BSCavV, BSAcVNV, and
BSMyV, respectively), Commelina yellow mottle virus (ComYMV), cacao swollen shoot virus (CSSV), sugarcane bacilliform Mor virus (SCBMV), Citrus mosaic virus (CMBV). B Putative transcriptional elements present in the intergenic region
of BSV isolates.
Indicates identical amino acid

translation and reverse transcription phase of replication (Fig. 2B). Analysis of the sequence upstream
of the TATA box revealed the presence of putative
upstream promoter cis-elements that enabled effective transcription: one putative inverted CCAAT
box was present 45 nt from the TATA box. Studies

of the adenovirus major late promoter revealed the

importance of this box, which plays a role in transcription initiation [23]. Already identified in members of other BSV species and plant pararetroviruses,
the activation one sequence 1 (as-1) motif has also
been detected in BSAcVNV [12, 16, 24].

Fig. 3. Phylogenetic neighbor-joining trees generated by the Darwin 4 program based on complete nucleotide sequences
of BSV species, other badnaviruses, and two RTBV isolates (RTBV-AP and RTBV-Phil1). Numbers at the nodes of
the branches represent percentage bootstrap values (1000 replicates) when greater than 60. The badnaviruses are banana
streak virus (BSV) species BSAcVNV, BSOlV, BSAcYuV, BSMyV, ComYMV, SCBMV, sugarcane bacilliform Mor virus
(SCBMV), CSSV-Agou1, CSSV-Wobe12, CSSV-N1A, CMBV, taro bacilliform virus (TaBV) and Dioscorea alata bacilliform virus (DaBV)

F. Lheureux et al.: Banana streak acuminata Vietnam virus

1415

To explore whether BSAcVNV is integrated in the

Musa genome, specific PCR primers were designed
on the BSAcVNV sequence to amplify genomic DNA
from 195 healthy banana genotypes from the openfield CIRAD Musa collection in Guadeloupe. This
collection is representative of the diversity of the
Musa genotype: intra- and interspecies diploids
(AA, BB), triploids (AAA, BBB, AAB, ABB) and
tetraploid (AAAA, AAAB). We did not obtain the
expected PCR product from any M. balbisiana genotypes, whereas we observed specific BSAcVNV PCR
amplification from several M. acuminata and interspecies genotypes. These results indicate that partial
BSAcVNV integrants are restricted to M. acuminata.
To investigate the relationship between BSAcVNV
and other BSV isolates with reliable full-length
genome sequences, other known badnaviruses, and
RTBV, phylogenetic analysis was performed using
the neighbor-joining method (Fig. 3). BSAcVNV
grouped with banana streak acuminata yunnan virus-BSAcYuV (accession DQ092436), where 86.6%
sequence identity exists between the full-length sequences. The pairwise combination of BSAcVNV
and members of other BSV species has sequence
identities of 57.3% (BSMyV) to 60.3% (BSOlV)
over the entire genome. This intermediate value is
comparable to a minimum of 70.6% sequence identities among badnaviruses that infect the same plant
such as CSSV and a maximum of 54% of sequence
identities between badnaviruses that infect different
plants such as CSSV, BSV and SCBV.
In conclusion, we believe the genome of BSV
virions isolated from a M. acuminata siamea accession from Vietnam is sufficiently different to be
classified as belonging to a distinct BSV species,
and we consequently named it banana streak virus
strain acuminata Vietnam (BSAcVNV). However,
the result of phylogenetic analysis suggests that the
BSVs are more closely related to each other than to
members of any of the other badnaviruses species.
One possible explanation is that the intermediate
values observed (Fig. 3) are due to a particular feature of BSV since the episomal BSV viral particle
could originate both from episomal and from integrated forms. According to ICTV criteria for 2005,
four distinct BSV subgroups nevertheless appear to
be sufficiently distant from one another to warrant

recognition as four distinct BSV species. Consequently, classification criteria for the badnavirus
group need to be amended to define the exact limits
of the badnavirus group.
If it is assumed that BSV EPRVs are molecular
fossils of viruses that exist in populations of wild
plants, as reported by Geering et al. [8], to be able
to analyze the evolution of BSV, we need to know
the exact origin of the molecular diversity observed. According to Geerings hypothesis, the four
distinct BSV species identified in this study (Fig. 3)
could result from at least four distinct integrated
viral sequences, which testify to the diversity of
the viral ancestry. In this case, each BSV species
must correspond to an integrated viral sequence and
the slight diversity observed within one subgroup
(BSAcYuV and BSAcVNV) must result from natural viral diversification. A wider knowledge of
non-integrated viral sequences should help clarify
the biological role of BSV in the evolution of either
the Musa genome or of the virus itself.
Acknowledgments
We thank Anne Sophie Dielen for checking of BSAcVNV
integrants in Musa genotypes during her Master in 2005.

References
1. Ahlawat Y, Pant R, Lockhart B, Srivastav M,
Chakraborty N, Varma A (1996) Association of Badnavirus with Citrus mosaic disease in India. Plant Dis 80:
590592
2. Bakry F (2001) Banana genetic diversity and perspectives for breeding. New varieties: why and how? C R
Acad Agric Fr 87: 129141
3. Boeck R, Kolakofsky D (1994) Positions 5 and 6
can be major determinants of the efficiency of nonAUG initiation codons for protein synthesis. EMBO J
13: 36083617
4. Dallot S, Acu~
na P, Rivera C, Ramirez P, C^
ote F,
Lockhart BEL, Caruana M-L (2001) Evidence that
the proliferation stage of micropropagation procedure
is determinant in the expression of Banana streak virus
integrated into the genome of the FHIA 21 hybrid
(Musa AAAB). Arch Virol 146: 21792190
5. Futterer J, Potrykus I, Bao Y, Li L, Burns T, Hull R,
Hohn T (1996) Position-dependent ATT initiation during plant pararetrovirus rice tungro bacilliform virus
translation. J Virol 70: 29993010

1416
6. Geering A, McMichael L, Dietzgen R, Thomas J (2000)
Genetic diversity among Banana streak virus isolates
from Australia. Phytopathology 90: 921927
7. Geering A, Parry J, Zhang L, Olszewsk IN, Lockhart B,
Thomas J (2000) Is banana streak virus strain Onne the
only activateable virus integrant in the Musa genome?
In: 2nd International Symposium on the Molecular
and Cellular Biology of Banana. Byron Bay Australia,
INIBAP
8. Geering A, Olszewski N, Harper G, Lockhart B, Hull R,
Thomas J (2005) Banana contains a diverse array of
endogenous badnaviruses. J Gen Virol 86: 511520
9. Geering A, Pooggin M, Olszewski N, Lockhart B,
Thomas J (2005) Characterisation of Banana streak
Mysore virus and evidence that its DNA is integrated
in the B genome of cultivated Musa. Arch Virol 150:
787796
10. Gregor W, Mette MF, Staginnus C, Matzke MA, Matzke
AJM (2004) A distinct endogenous pararetrovirus family in Nicotiana tomentosiformis, a diploid progenitor of
polyploid tobacco. Plant Physiol 134: 11911199
11. Grunert S, Jackson R (1994) The immediate downstream codon strongly influences the efficiency of
utilization of eukaryotic translation initiation codons.
EMBO J 13: 36183630
12. Harper G, Hull R (1998) Cloning and sequence analysis
of banana streak virus DNA. Virus Genes v: 271278
13. Harper G, Osuji JO, Heslop-Harrison P, Hull R, Harper
G, Lockhart B (1999) Integration of Banana streak
badnavirus into the Musa genome: molecular evidence.
Virology 255: 207213
14. Hull R (1999) Classification of reverse transcribing elements: a discussion document. Arch Virol 144: 209213
15. Iskra-Caruana M-L, Lheureux F, Teycheney P-Y (2003)
Les pararetrovirus endogenes (EPRV), voie nouvelle de
transmission des virus de plantes. Virologie 7: 255265
16. Krawczyk S, Thurow C, Niggeweg R, Gatz C (2002)
Analysis of the spacing between the two palindromes of
activation sequence-1 with respect to binding to different TGA factors and transcriptional activation potential.
Nucleic Acids Res 30: 775781
17. Kunii M, Kanda M, Nagano H, Uyeda I, Kishima Y,
Sano Y (2004) Reconstruction of putative DNA virus
from endogenous Rice tungro bacilliform virus-like
sequences in the rice genome: implications for integration and evolution. BMC Genomics 5: 80
18. Lheureux F, Carreel F, Jenny C, Lockhart BEL, IskraCaruana M-L (2003) Identification of genetic markers

F. Lheureux et al.: Banana streak acuminata Vietnam virus

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

linked to Banana streak disease expression in interspecific Musa hybrids. Theor Appl Genet 106: 594598
Lockhart B (1990) Evidence for a double-stranded
circular DNA Genome in a second Group of plant
Viruses. Phytopathology 80: 127131
Lockhart B, Olszewski N (1993) Serological and genomic heterogeneity of banana streak badnavirus: implication for virus detection in Musa germplasm. In:
Breeding Banana and plantain for resistance to disease
and pest. INIBAP, Montpellier, France, pp 105113
Lockhart B (1995) Banana streak badnavirus infection
in Musa: epidemiology, diagnosis and control. ASPAC
Food and fertilizer technology center (Taiwan). Techn
Bull 143: 111
Lockhart B, Jones DR (2000) Banan streak. In: Jones
DR (ed) Disease of banana, abaca and ensete. CABI,
wallingford, pp 263274
Lu H, Reach M, Minaya E, Young C (1997) The initiator
element of the adenovirus major late promoter has an
important role in transcription initiation in vivo. J Virol
71: 102109
Medberry L, Lockhart B, Olszewski N (1992) The
Commelina yellow mottle virus promoter is a strong
promoter in vascular and reproductive tissues. Plant Cell
4: 185192
Muller E, Sackey S (2005) Molecular variability analysis of five new complete Cacao swollen shoot virus
genomic sequences. Arch Virol 150: 5366
Ndowora T, Lockhart B, Olszewski N (1997) Relationship between integrated and episomal genomic sequences in Musa. Virologie 87: S69
Ndowora T, Dahal G, LaFleur D, Harper G, Hull R,
Olzsewski N, Lockhart BEL (1999) Evidence that
badnavirus infection in Musa can originate from
integrated pararetroviral sequences. Virology 255:
214220
Perrier X, Flori A, Bonnot FI (2003) Data analysis
methods. In: Genetic diversity of cultivated tropical
plants. CIRAD, Montpellier, pp 4376
Pooggin M, Futterer J, Skryabin K, Hohn T (1999) A
short open reading frame terminating in front of a
stable hairpin is the conserved feature in pregenomic
RNA leaders of plant pararetroviruses. J Gen Virol 80:
22172228
Richert-Poggeler K, Noreen F, Schwarzacher T, Harper
GTH (2003) Induction of infectious petunia vein clearing (pararetro) virus from endogenous provirus in petunia. EMBO J 22: 48364845