AU712544B2 - DNA sequences of banana bunchy top virus - Google Patents

Description

I I WO 96/38564 PCT/AU95/00311 1

TITLE

"DNA SEQUENCES OF BANANA BUNCHY TOP VIRUS" FIELD OF INVENTION THIS INVENTION relates to DNA sequences of banana bunchy top virus (BBTV) which may be utilised for diagnosis of infection and which also may be utilised for generating virus resistant plants.

BACKGROUND ART The banana (Musa spp.) is the world's largest fruit crop by value. As with most important food crops, bananas are affected by a 1 0 number of serious diseases, particularly fungal and viral diseases.

Banana bunchy top disease (BBTD) is the most serious of the viral diseases. It occurs in all banana growing regions other than the Americas and the Caribbean. The disease virtually destroyed the banana industry in Australia and particularly in south-east Queensland and northern New South Wales in the 1920's. It is now controlled in those areas through stringent phytosanitory regulations as reported in Dale, 1987, Advances in Virus Research 33 301-325. However, through many Asian, African and South Pacific countries, the disease is not controlled and is a serious limitation to production. At this time, there are no reports of resistance within any of the Musa germplasm tested.

The disease has always been assumed to be caused by a virus and this presumed virus was classified as a possible luteovirus based on biological properties i.e. the disease is transmitted by aphids (Pentalonia nigronervosa) in a persistent manner, the phloem of infected plants is damaged, the major symptoms are marginal yellowing, stunting and leaf bunching and the disease is not sap transmissible. There is further evidence supporting the association of a luteovirus with BBTD.

Dale et al., 1986, Journal of General Virology 67 371-35 reported the presence of dsRNAs in BBTD-infected but not healthy bananas. DsRNAs of similar sizes have been reported in luteovirus-infected plants. More recently, Iskra et 1989, in Fruits 44 63-66 reported the purification of WO 96/38564 PCT/AU95/00311 2 28 nm isometric virus-like-particles (VLPs) from BBTD infected plants.

Wu and Su in ASPAC Food and Fertiliser Technology Center Taipei Technical Bulletin No. 115 (1989) reported the purification of 20-22 nm isometric VLPs which contained ssRNA of Mr 2.0 x 106 and suggested that these particles represented the virions of a "small" luteovirus. Wu and Su, 1990, Journal of Phytopathology 128 153-160 also described purification and characterisation of BBTV. Thus, at this time the evidence for the association of a luteovirus with BBTD was substantial though not conclusive.

In 1991, Harding et al., 1991, Journal of General Virology 72 225-230 and Thomas and Dietzgen, 1991, Journal of General Virology 72 217-224 reported the purification of isometric VLPs 18-20 nm in diameter.

These VLPs contained ssDNA of about 1 kb and a single protein of Mr 20,100.

BBTV has characteristics different from those of any described plant virus group; it is most similar to the geminiviruses, which have ssDNA as their genomic nucleic acid but their DNA is about 2.7 to kb, their particles are usually geminate and they are transmitted by leafhoppers or whiteflies. However, subterranean clover stunt virus (SCSV) as reported in Chu et al., 1988, Virology 167 38-49 and to a lesser extent coconut foliar decay virus (CFDV) as reported in Randles et al., 1987, Journal of General Virology 68 273-280 have characteristics very similar to BBTV. Both of these viruses have 18 to 20 nm isometric particles which contain circular ssDNA of approximately 1 kb. SCSV is persistently transmitted by aphids, whereas CFDV is transmitted by a planthopper. Furthermore, the genome of SCSV is composed of at least seven distinct ssDNA molecules each containing one large open reading frame (ORF) as reported in Chu et al., 1990, VIllth International Congress of Virology Abstracts p125. However, it is important to emphasize that none of the sequences of the ssDNA molecules have been published and therefore common characteristics of such sequences cannot be WO 96/38564 PCT/AU95/00311 3 determined. Such ssDNA molecules also contain a strong stem-loop sequence as described in Waterhouse et al., 1990, Vllith International Congress Virology, Berlin. SCSV was further characterized in Surin et al., 1993, IXth International Congress of Virology Abstracts, Glasgow, Scotland whereby each component contained a conserved stem and loop structure and a non coding region in at least five of the components. The sequence of one ssDNA molecule of CFDV has been determined as discussed in Rohde et al., 1990, Virology 176 648-651 and one of the ORFs encodes a putative replicase. There are also three animalinfecting viruses that have similar characteristics to BBTV, SCSV and CFDV, namely porcine circovirus (PCV), chicken anaemia virus (CAV) and psittacine beak and feather disease virus (PBFDV) as reported in Todd et al., 1991, Archives of Virology 117 129-135.

Reference also may be made to Harding et al., 1993, 1 5 Journal of General Virology 74 323-328 which referred to the cloning and sequencing of one component of the BBTV genome which was present as a circular ssDNA comprising of 1111 nucleotides. It contains one large open reading frame (ORF) of 858 nucleotides in the virion sense; this ORF encodes a putative replicase based on the presence of a dNTPbinding motif (GGEGKT). Two smaller ORFs (249 and 366 nucleotides), in the complementary orientation, could not be assigned any obvious function. Neither of these ORFs had significant sequence homology with any known DNA plant virus gene or gene product. Computer analysis of this component predicted a strong stem-loop structure in the virion sense putative untranslated region and a nonanucleotide sequence in the loop was found to be nearly identical to the nonanucleotide invariant loop sequence of geminiviruses and coconut foliar decay virus. This reference also mentioned that there was strong evidence that the genome of BBTV consists of more than one component because no ORF was found that would encode a protein the size of the BBTV coat protein. BBTV has some characteristics in common with geminiviruses but cannot be WO 96/38564 PCT/AU95/00311 4 classified as such. Rather, BBTV probably belongs to an undescribed plant virus group which could also include subterranean clover stunt virus and coconut foliar decay virus.

In Burns et al., 1994, Arch Virol. 137 371-380, it was ascertained that a 93 nucleotide sequence was found to be strongly conserved between two ssDNA genomic components of BBTV. Two outwardly extending degenerate primers were designed from this sequence and used in a polymerase chain reaction (PCR) with DNA extracted from purified BBTV virions. PCR amplified products consisting of at least seven distinct bands all approximately 1 kb and possibly representing full-length BBTV dsDNA were resolved. The PCR amplified products were cloned and the clones screened by restriction enzyme analysis. Four distinct restriction analysis groups were identified. This reference concluded that the genome of BBTV contains at least five components and more probably seven components and that BBTV belongs to a previously undescribed group of plant viruses which may also contain subterranean clover stunt virus.

In Karan et 1994, Journal of General Virology 75 3541- 3546, mention is made of BBTV component 1 from isolates from different countries being cloned and sequenced and the sequences were subsequently aligned and compared. This analysis indicated two groups: the South Pacific group (isolates from Australia, Burundi, Egypt, Fiji India, Tonga and Western Samoa) and the Asian group (isolates from the Philippines, Taiwan and Vietnam). The mean sequence difference within each group was 1.9 to 3.0% and between isolates from the two groups were approximately 10%, but some parts of the sequences differed more than others. However, the protein encoded by the major open reading frame differed by approximately The region from the beginning of the stem-loop sequence to the potential TATA box was identical in all isolates except for a two nucleotide change in the Western Samoan isolate and a single change in that of the NSW isolate. These results, together with WO 96/38564 PCT/AU95/00311 other evidence, suggest that BBTV has spread to bananas after the initial movement of bananas from the Asian Pacific regions to Africa and the Americas.

In Xie et al., 1995, Phytopathology 85 339-347, the Hawaiian isolate of BBTV was purified from infected banana cultivar Williams. Three single-stranded DNA (ssDNA) components were cloned and sequenced; they were named component 1, 3 and 4 respectively.

Component 1 is 1,110 nucleotides in length and shares 98% nucleotide sequence identity with the BBTV DNA component 1 of the Australian isolate as described in Harding et al. (1993) above. This component contains two open reading frames (ORF) capable of encoding a protein of 33.5 kDa, which may function as a replicase, and a protein about 15.2 kDa, with unknown functions. Component 3 is 1,057 nucleotides in length and does not contain any ORFs larger than 10 kDa. Component 4 is 1,017 nucleotides in length and potentially encodes a protein of 18.9 kDa.

All three ssDNA components have a stem-loop sequence and have a conserved non-coding region. The sequence of each of these three components is different from that of BBTV DNA components of two Taiwanese isolates. BBTV-specific clones were used in dot-blot hybridisation assays for detection of BBTV in plants using radioactive and non-radioactive probes. A polymerase chain reaction (PCR) assay was developed for detection of BBTV in banana samples and single aphids.

Dot-blot hybridisation assays were as sensitive as enzyme-linked immunosorbent assay (ELISA) while PCR was 1,000 times more sensitive than dot-blot and ELISA assays for detection of BBTV in bananas.

Thus, as a summary of the above prior art, three components of BBTV have now been prepared and sequenced, i.e.

component 1 as discussed above in the Harding et al. (1993) reference and components in the abovementioned Xie et al. (1995) reference. In addition, the Xie et al. (1995) reference has indicated that all three ssDNA components have a stem-loop sequence and have a conserved noncoding region.

WO 96/38564 PCT/AU95/00311 6 It is also noted from the Burns et al. (1994) reference above that a region was identified between components 1 and 2 of high sequence homology and hence it was postulated that as this region was located in the probable non-encoding region of component 1, 3' to the putative ORF and 5' to the major stem loop structure that such region could be conserved in all other probable BBTV genomic components.

SUMMARY OF THE INVENTION It is an object of the invention to provide DNA sequences of additional DNA components of the BBTV genome and also fragments thereof.

A further three components of BBTV have now been cloned and sequenced and it has been found that:- Component 3 consists of about 1,075 base pairs; Component 4 consists of about 1,043 base pairs; and Component 6 consists of about 1,089 base pairs It has also been determined from these sequences that each component contains:a stem/loop structure that is strongly conserved between all components; (ii) one large open reading frame in the virion sense DNA strand in each component which varies in size and sequence between components; and (iii) a second conserved region of 92 nucleotides that is conserved between all components.

The invention thus includes within its scope the sequence or part thereof of BBTV DNA component 3 of the BBTV genome and sequences that are complementary to this sequence. Component 3 consists of approximately 1075 nucleotides. This sequence is shown in FIG. 1.

The invention also includes within its scope the sequence or part thereof of BBTV DNA component 4 of the BBTV genome and WO 96/38564 PCT/AU95/00311 7 sequences that are complementary to this sequence. Component 4 consists of approximately 1043 nucleotides. This sequence is shown in FIG. 2.

The invention also includes within its scope the sequence or part thereof of BBTV DNA component 6 of the BBTV genome and sequences that are complementary to this sequence. Component 6 consists of approximately 1089 nucleotides. This sequence is shown in FIG. 3.

The virion sense DNA strand of each components 3, 4 and 6 1 0 includes an open reading frame.

Each ORF in the components 3, 4 and 6 have a potential TATA box and one or two potential polyadenylation signals associated with it and each polyadenylation signal had an associated GC-rich region containing the trinucleotide sequence TTG. In contrast, a number of ORFs were identified in component 2 reported in the Xie et al (1995) reference above but none of these had appropriately located potential TATA boxes and polyadenylation signals associated with them. None of the ORF amino acid sequences nor the full DNA sequences of any of the components 3, 4 and 6 had significant sequence homology with any known protein or rucleic acid sequences. The ORF of component 3 has a nucleotide sequence of 525 nucleotides encoding an amino acid sequence of approximately 175 amino acid residues for a 20.11KDa protein. The ORF of component 3 appears to encode the BBTV coat protein. The ORF of component 4 has a nucleotide sequence of 351 nucleotides encoding an amino acid sequence of 117 amino acid residues for a protein of 13.74KDa. The ORF of component 4 encoded a residue hydrophobic domain which may indicate that this ORF encoded a transmembrane protein. The ORF of component 6 has a nucleotide sequence of 462 encoding 154 amino acid sequence of a 17.4KDa protein.

Surprisingly, each of the components 3, 4 and 6 are substantially non-homologous both with each other as well as WO 96/38564 PCT/AU95/00311 8 components referred to in the Xie et al. reference mentioned above. This will be clearly demonstrated by reference to FIGS. 1, 2 and 3.

Each of component 3, 4 and 6 contained a conserved stem-loop structure and a nonanucleotide potential TATA box which was 5' of the large virion sense ORF. The stem-loop structures were incorporated in a common region (CR-SL) of 69 nucleotides which was 62% homologous between components.

Each of components 3, 4 and 6 also contained a major common region (CR-M) which was located 5' of the CR-SL in each component, in the non-coding region and was 76% homologous over 92 nucleotides. Each CR-M contained a near complete 16 nucleotide direct repeat and a GC-box which was similar to the rightward promoter element found in wheat dwarf geminivirus.

The invention also covers DNA sequences that can hybridize to any one or part thereof of the sequences of components 3, 4, 6 or their complementary sequences wherein sequences varying within can hybridise under standard stringency conditions. Suitable hybridisation procedures and stringency conditions are given in Burns et al., 1994, Arch Virol. 137 371-380. The invention also includes synomonous DNA sequences that encode the same protein as encoded by components 3, 4 and 6, and amino acid sequences encoded any of the aforementioned DNA sequences.

A further aspect of the present invention is the use of the abovementioned sequences or part thereof, of their complementary sequences or part thereof, of variations of these sequences within 35% of any one of the sequences or part thereof, of variations of their complementary sequences within 35% of any one of the sequences or part thereof as either DNA or RNA. Use includes:hybridization with other sequences for the purposes of detection or identification of these sequences, their variants, their complements or parts thereof using such techniques as for example, Southern WO 96/38564 PCT/AU95/00311 9 hybridisation, Northern hybridisation, dot blot hybridization, liquid or solution hybridisation or the polymerase chain reaction; diagnosis of disease from detection or identification of any one of the aforementioned sequences in banana plants or parts thereof; and insertion into plant cells or other organisms either transiently or stably for the purposes of utilising the transcribed and/or translation products of these sequences or part thereof or for the purposes of utilising these sequences to alter or facilitate the transcription and/or translation of other nucleic acid sequences (for example, for use as promoters, enhancers or termination signals).

The nucleic acid sequences of the invention may be inserted into a plasmid such as pBin19 or pUC19 between a cauliflower mosaic virus 35S promoter and a cauliflower mosaic virus 35S terminator.

These plasmids may be introduced into cells of plants. Introduction may be achieved by using Agrobacterium tumefaciens (pBin19 construct) or microprojectile bombardment (pBin19 or pUC19 constructs). Within the plant cells, the inserted nucleic acid sequences may be suitably transcribed and/or translated.

Alternatively, the sequences covered by this invention may be inserted upstream or downstream of a second nucleic acid sequence wherein the transcription and/or translation of the second nucleic acid sequence is altered or facilitated by the nucleic acid sequences of the invention.

The term "consisting essentially of' as used herein and in the appended claims refers to DNA sequences having variations within 35% as described above or limits of variability as also discussed above including synomonous DNA sequences and complementary

DNA

sequences. A similar meaning also refers to polypeptides discussed WO 96/38564 PCT/AU95/00311 hereinafter.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 illustrates the nucleotide sequence of Component 3; FIG. 2 illustrates the nucleotide sequence of Component 4; FIG. 3 illustrates the nucleotide sequence of Component 6; FIGS. 4a, 4b and 4c illustrate the nucleotide sequence of Components 3, 4 and 6 respectively and deduced amino acid sequences of their ORFs; FIG. 5 illustrates the determination of the virion-sense orientation of BBTV DNA components 2 to 6.; FIG. 6 illustrates the aligned stem-loop common regions (CR-SL) of BBTV DNA component 1 to 6; FIG. 7 illustrates the aligned major common regions (CR-M) of BBTV DNA components 1 to 6; FIG. 8 illustrates the diagrammatic representation of the proposed genome organisation of BBTV; and FIG. 9 illustrates the N-terminal sequencing of BBTV coat protein.

EXPERIMENTAL

EXAMPLE 1: BBTV COMPONENTS

METHODS

Synthesis and cloning of cDNA. Bananas with characteristic symptoms of banana bunchy top disease were collected from the Nambour region of S-E Queensland. BBTV particles were purified as described by Wu Su, 1990, Journal of Phytopathology 128 153-160 and Thomas Dietzgen, 1991, Journal of General Virology 72 217-224. Nucleic acid was extracted from virions as described by Francki Randles, 1973, Virology 54 359-368. Double stranded DNA was synthesised as described by Gubler Hoffman, 1983, Gene 25 263-269 using random hexamers (Bresatec) to prime first strand synthesis. The dsDNA was treated with mung bean nuclease (Promega) and ligated into I I WO 96/38564 PCT/AU95/00311 11 Smal digested plasmid vector pUC18 (Upcroft Healey, 1987, Gene 51 69-75). The plasmid was then used to transform Escherichia coli strain JM109 (Hanahan, 1983, Journal of Molecular Biology 166 557-580) and potential recombinant clones were identified by screening on X-gal substrate (Vieira Messing, 1982, Gene 19 259-268) Plasmids were isolated using the alkaline lysis method (Sambrook et al., 1989, Molecular Cloning: A Laboratory Manual. New York: Cold Spring Harbor Laboratory). Inserts were excised by digestion with EcoRIIHindlll, electrophoresed in agarose gels and capillary blotted 1 0 onto Hybond N+ (Amersham) using 0.4 M NaOH (Sambrook et al., 1989, Molecular Cloning: A Laboratory Manual. New York: Cold Spring Harbor Laboratory). Inserts for use as DNA probes were purified from agarose gels using a Gene-Clean kit (Bresatec). DNA probes were labelled using a Ready-To-Go labelling kit (Pharmacia) as recommended by the manufacturer. Prehybridisations and hybridisations were done as described by Burns et al., 1994, Archives of Virology 137 371-380.

Sequencing and sequence analysis. Minipreparations of respective BBTV clones were prepared by alkaline lysis followed by polyethylene glycol precipitation (Hattori Sakaki, 1986, Analytical Biochemistry 152 232-238). Sequencing was done using 35 SdATP and a Sequenase kit (US Biochemicals) as recommended by the manufacturer.

Reaction products were electrophoresed in 8% polyacrylamide gel containing 7 M urea. Gels were fixed, dried and exposed to X-ray film.

Primers used for sequencing were either universal sequencing primers or 17-30 nt primers complementary to appropriate regions of the cloned viral DNA synthesised using an Applied Biosystems (ABI) PCR Mate and processed as recommended by the manufacturer.

PCR products for sequencing were purified from agarose gels using a Gene-Clean kit (Bresatec). DNA was sequenced using a Sequenase kit (USB) essentially as described by the manufacturer.

Denaturation of template DNA (500 ng) was done by boiling following the m WO 96/38564 PCT/AU95/00311 12 addition of DMSO and 3 pmoles of sequencing primer.

Nucleotide sequences were analysed using the GCG analysis package version 8 available through the ANGIS computing facility at the University of Sydney, Australia. Nucleotide and amino acid sequences were aligned using the Clustal V software package (Higgins et al., 1991, CABIOS 8 189-191). Four DNA databases (GenBank, GenBank Weekly Updates, EMBL and EMBL weekly updates) and five protein databases (SwissProt, SwissProt Weekly Updates, PIR, GenPeptide Proteins and GenPeptides Weekly Updates) were searched 1 0 for sequence homologies with BBTV nucleotide and deduced amino acid sequences using two database search analysis programs, FASTA (Pearson Lipman, 1988, Proceedings of the National Academy of Sciences, USA 85 2444-2448) and BLAST (Altschul et al., 1990, Journal of Molecular Biology 215 403-410).

PCR: Analysis and cloning. Using the respective nucleotide sequences of clones pBTRP-11, 20, 80 and 88 and (ii) pBTRP-P1 and P2 and the nucleotide sequences of BBTV components 1 (Harding et al., 1993, Journal of General Virology 74 323-328) and 2, three sets of immediately adjacent outwardly extending primers primer A: 5' GCATCCAACGGCCCATA primer B: CTCCATCGGACGATGGA (ii) primer C: 5' TATTAGTAACAGCAACA primer D: 5' CTAACTTCCATGTCTCT (iii) primer E: CGGGa/tATa/cTGATTGt/gGT and primer F: TACa/tTTTGTCATAGc/tGT were synthesised and used in a PCR with BBTV DNA as template as described by Burns et al., 1994, Archives of Virology 137 371-380). The amplified products were cloned using the TA cloning kit (Invitrogen) into the plasmid vectors pCRII or pCR2000 as recommended by the manufacturer or into T-tailed pUC19 and Bluescript (Marchuk et al., 1990, Nucleic Acids Research 19 1154). Recombinant clones were selected using X-gal substrate on Luria Bertani (LB) agar containing the appropriate antibiotic and plasmids isolated using the WO 96/38564 PCT/AU95/00311 13 alkaline lysis method (Sambrook et al., 1989, Molecular Cloning: A Laboratory Manual. New York: Cold Spring Harbor Laboratory). Clones with apparent full-length inserts (approximately 1 kb) were selected for sequencing.

Polarity of virion ssDNA. BBTV ssDNA was extracted, electrophoresed in agarose and capillary blotted onto duplicate nylon membranes (Harding et al., 1993, Journal of General Virology 74 323- 328).

For components 3, 4 and 6, strand-specific RNA probes were used. Full-length RNA transcripts of full-length BBTV clones of each of the four components were synthesised using a riboprobe in vitro transcription kit (Promega) as recommended by the manufacturer.

RESULTS

Cloning and sequencing of genomic components 3, 4 and 6 These genomic components of BBTV were cloned and sequenced from two libraries, a random primed library and (ii) a PCR library.

Randon Primed Library A random primed library was generated from BBTV ssDNA extracted from purified virions. The resultant dsDNA was treated with mung bean nuclease, blunt-end ligated into Smal cut pUC18 and cloned into E. coli JM109. This library was screened with 32 P-labelled DNA from BBTV virions, healthy bananas and the insert from pBT338 which was a partial clone of BBTV DNA component 1 (Harding et al., 1991, Journal of General Virology 72 225-230; Harding et al., 1993, Journal of General Virology 74 323-328).

BBTV DNA Component 6: Two clones from the random primed library, pBTRP-P1 and P2, hybridised with labelled BBTV virion DNA but not with DNA from healthy bananas or pBT338. However, the inserts of these clones, both of approximately 1 kb, were digested with EcoRV whereas neither components 1 nor 2 had EcoRV sites. The two WO 96/38564 PCT/AU95/00311 14 clones were partially sequenced using universal forward and reverse primers. The sequences of both clones were identical but clearly different to those of components 1 and 2. Again, two immediately adjacent, outwardly extending primers, primers C and D, were designed from the sequence and synthesised. BBTV virion ssDNA was used as a template with these two primers in a PCR reaction and the resultant product cloned into a T-tailed Bluescript vector. One apparent full length clone, pBT- P2A1, was selected and sequenced in both directions from subclones generated by exonuclease III digestion and universal forward and reverse, and sequence-specific primers. The final component 6 sequence of 1089 bp was then compiled and a deduced amino acid sequence of the major ORF is shown in FIG. 4c.

(ii) PCR Library When the sequences of components 1 and 2 were compared, two regions of homology were identified. The first region, later defined as the stem-loop common region (CR-SL) included the potential stem/loop sequence previously identified in component 1 (Harding et al., 1993, Journal of General Virology 74 323-328); the second region, which was contained within the region later defined as the major common region was a sequence of approximately 66 nucleotides 5' to the stemloop sequence. It was hypothesised that all BBTV genomic components should contain a CR-M and therefore two immediately adjacent, outwardly extending degenerate primers, primers E and F, were designed from this region, synthesised and extended by PCR using BBTV virion ssDNA as a template (Burns et al., 1994, Archives of Virology 137 371-380). Seven products, each of approximately 1 kb, were resolved by polyacrylamide gel electrophoresis. The products were cloned into pCRII. The resultant clones were divided into three groups, groups B, C and D, on the basis that they hybridised with BBTV virion DNA but not DNA from healthy bananas and that each group had restriction patterns different to the other two groups and to components 1, 2 and 6 (Burns et al., 1994, WO 96/38564 PCT/AU95/00311 Archives of Virology 137 371-380). One group, group A, had a restriction pattern indistinguishable from that of component 2 and it was later confirmed by sequencing that group A clones represented clones of component 2.

Each group of clones was assumed to represent a new and unique BBTV DNA component. For each group of clones, three clones (component 3) or four clones (components 4 and 5) were partially sequenced using universal forward and reverse primers. In each instance, all the clones within a group had identical sequences where these sequences overlapped except for one or two single nucleotide changes. Further, the sequences from each group were different to each other group and to the sequences of components 1, 2 and 6. One clone from each group or component was selected and fully sequenced in both directions. Importantly, each of these groups of clones were generated using degenerate primers covering a sequence of 34 nucleotides derived from the conserved CR-M of components 1 and 2. The CR-M from components 1 and 2 was not fully conserved and thus it was expected that the hypothesised CR-M sequence would vary between other components. Therefore, converging primers unique to each component were designed and used to amplify a sequence including CR-M for each component from BBTV virion ssDNA. The resultant PCR product was sequenced directly using the two component specific converging primers.

BBTV DNA Component 3: Component 3 (Group C clone pBTP-64) was sequenced in both directions from the original clone and from subclones generated by exonuclease III digestion or restriction fragments using universal forward and reverse primers and three sequence-specific primers. Two additional converging primers were designed from this sequence to amplify a 380 bp product including the CR-M. The sequence of this product was identical to that of pBTP-64 except for five single nucleotide changes, four of which were in the sequence covered by the original degenerate primers and one outside this sequence at nucleotide 947. The final component 3 sequence of i WO 96/38564 PCT/AU95/00311 16 1075 bp was then compiled as shown in FIG. 4a.

BBTV DNA Component 4: Component 4 (Group D clone pBTP-62) was sequenced in both directions from the original clone and from subclones generated by exonuclease III digestion using universal forward and reverse primers and three sequence-specific primers. Two additional converging primers were designed from this sequence to amplify a 350 bp product including the CR-M. The sequence of this product was identical to that of pBTP-62 except for two single nucleotide changes in the sequence covered by the original degenerate primers.

1 0 The final component 4 sequence of 1043 bp was then compiled as shown in FIG. 4b.

Orientation of genomic components and association with banana bunchy top disease We have previously shown that the BBTV genome is encapsidated as single-stranded DNA (Harding et al., 1991, Journal of General Virology 72 225-230; Harding et al., 1993, Journal of General Virology 74 323-328). To determine the orientation of each component in virions, strand-specific DNA or RNA probes specific for each component were synthesised and hybridised with BBTV virion DNA. Component 2 specific probes were two 3' end-labelled 30mer oligonucleotides whereas probes specific for components 3, 4, 5 and 6 were SP6, T3 or T7 promoted 32 P-labelled RNA transcripts. For each component, the probes whose sequences were complementary to the component sequences presented in FIGS. 4a, 4b and 4c hybridised strongly to BBTV virion DNA whereas the probes whose sequences were the same as the FIGS. 4a, 4b and 4c sequences did not hybridise (FIG. This indicated that each component was encapsidated as ssDNA and only in one orientation, that presented in FIGS. 4a, 4b and 4c.

Further, the strand- and component-specific probes that hybridised with BBTV virion ssDNA were used as probes to demonstrate that each component was associated with banana bunchy top disease.

WO 96/38564 PCT/AU95/00311 17 Plant DNA extracts from three (for component 2) or four (for components 3, 4, 5 and 6) different BBTV isolates and DNA from four healthy bananas was Southern blotted and hybridised with each probe. Each componentspecific probe hybridised with a low molecular weight DNA of expected size in all the extracts from BBTV-infected bananas but did not hybridise with the extracts from healthy bananas (results not shown). This indicated that each component was clearly associated with the disease and the virus. In FIG. 5, the data for component 5 is shown for the sake of comparison.

Analysis of the BBTVgenomic components The sequences of the BBTV genomic components presented here and the sequence of component 1 (Harding et al., 1993, 2) were aligned and compared. Each of the six sequences were different except for two significant regions which had varying degrees of homology between all six components.

Stem-loop common region: We have previously identified a potential stem-loop structure in BBTV component 1 (Harding et aL., 1993, Journal of General Virology 74 323-328) which had a loop sequence almost identical to the invariant loop sequence of geminiviruses (Lazarowitz, 1992, Critical Reviews in Plant Sciences.ll 327-349). An equivalent stem/loop structure was also found in components 2 to 6 (FIG.

Each component had an 11 nucleotide loop sequence of which 9 consecutive nucleotides were conserved between all components. Each component also had a 10 bp stem sequence of which 14 nucleotides were fully conserved. However, when all six components were compared, the region of homology extended up to 25 nucleotides 5' of the stem-loop structure and up to 13 nucleotides 3' of the stem/loop structure. The 5' nucleotides were fully conserved between components 1, 3, 4 and There were apparently two deletions in both components 2 and 6. In component 2, eight nucleotides were fully conserved with components 1, 3, 4 and 5 whereas in component 6, 16 nucleotides were conserved with WO 96/38564 PCT/AU95/00311 18 these other components. The 13 nucleotides 3' of the stem-loop were fully conserved between all six components except for an apparent single nucleotide deletion in component 2. The sequence of up to 69 nucleotides including the stem-loop sequence was termed the stem-loop conserved region or CR-SL.

Major common region: The second common region was located at various distances 5' of the CR-SL and was called the Major Common Region or CR-M. This region varied in size from 65 nucleotides in component 1 to 92 nucleotides in component 5 (FIG. Component 1 apparently had the first 26 nucleotides of the CR-M deleted as well as a further single nucleotide deletion. Components 2, 3 and 4 had two single nucleotide deletions and component 6 had one single nucleotide deletion.

Forty-five nucleotides were conserved between all components and 23 of the first 26 nucleotides, deleted in component 1, were conserved between 1 5 components 2 to 6. Also in components 2 to 6, there was an almost complete 16 nucleotide direct repeat (ATACAAc/gACa/gCTATGA) from nucleotides 4 to 20 and 21 to 36. Further, a 15 nucleotide GC rich sequence (average of 86% GC) was located from nucleotides 78 to 92 and was 93% conserved between all components.

The sequence between the last nucleotide of the CR-M and the first nucleotide of the CR-SL varied in length from 22 nucleotides in component 1 to 233 nucleotides in component 2 (FIG. Interestingly, this sequence of 175 nucleotides in components 3 and 4 was 97% conserved between these two components.

Potential TATA boxes: A potential TATA box was identified in BBTV component 1 and was located 20 nucleotides 3' of the last nucleotide of the stem-loop and 43 nucleotides 5' of the start codon of the putative replicase gene (Harding et al., 1993, Journal of General Virology 74 323-328). Similar potential TATA boxes were also identified in components 2 to 6. In each of these components, the potential TATA box was a nine nucleotide sequence, CTATa/taltAtlaA, and was located downstream from the stem-loop sequence (FIGS. 4a, 4b and 4c).

WO 96/38564 PCT/AU95/00311 19 However, the sequence between the last 3' nucleotide of the stem-loop sequence and the potential TATA box was considerably longer in components 2 to 6 than in component 1 and varied from 157 nucleotides in component 5 to 227 nucleotides in component 4 (FIG. 8).

Analysis of components for open reading frames One large ORF coding for a putative replicase was identified in the virion sense of BBTV component 1 which had a potential TATA box 43 nucleotides 5' of the ATG start codon and a polyadenylation signal 13 nucleotide 5' of the stop codon (Harding et al., 1993, Journal of General 1 0 Virology 74 323-328). Components 2 to 6 were therefore analysed for ORFs in both the virion and complementary sense that could code for proteins of more than 25 amino acids. Numerous such ORFs were identified in both orientations in all five components. However, only four ORFs were identified that had associated with them a potential 5' TATA 1 5 box and an appropriately located polyadenylation signal. Components 3 to 6 each had one such ORF in the virion sense. These ORFs were (i) 525 nucleotides potentially coding for a 175 amino acid protein of 20.11K in component 3, (ii) 351 nucleotides potentially coding for a 117 amino acid of 13.74K in component 4, (iii) 483 nucleotides potentially coding for a 161 amino acid protein of 18.97K in component 5 and (iv) 462 nucleotides potentially coding for a 154 amino acid protein of 17.4K in component 6 (FIGS. 4a, 4b and 4c).

Nine ORFs were identified in component 2, four in the virion sense and five in the complementary sense (Table However, none of these ORFs had appropriately located nonanucleotide potential TATA boxes and polyadenylation signals and therefore were unlikely to be transcribed.

This data shows that there is clearly wide disparity between the ORFs of component 5 on one hand and components 3, 4 and 6 on the other.

Analysis of potential polyadenylation signals Six potential polyadenylation signals were identified _i___lL_/ilT _~l~li ~i~_~ll WO 96/38564 PCT/AU95/00311 associated with the 3' end of the major ORFs of components 3 to 6. A GTrich region of 10 to 17 nucleotides was located between 0 and 23 nucleotides 3' of each of these polyadenylation signals and each GT-rich region contained the trinucleotide sequence TTG (FIGS. 4a, 4b and 4c).

Only one potential polyadenylation signal identified in component 2 had a corresponding GT-rich region with the trinucleotide sequence TTG and this was located 233 nucleotides 3' of the nonanucleotide potential TATA box in the virion sense.

Sequence comparisons and analysis 1 0 Four DNA data bases were searched using the complete nucleotide sequences of components 2 to 6 and five protein databases were searched using the putative amino acid sequences derived from the major ORFs of components 3 to 6 as well as the smaller ORFs. No significant sequence homology was found either at the nucleotide or amino acid level. Further, no motifs or other signals were identified that would suggest possible functions for any of the four putative proteins other than 30 hydrophobic residues toward the N-terminus of the major ORF translation product of component 4 suggesting the presence of a transmembrane domain (Boulton et al., 1993, Virology 192 85-93). The ORFs of components 3 and 5 encoded putative proteins with molecular weights close to the 20.1KDa molecular weight of the BBTV coat protein.

However, there is no other evidence to suggest either of these ORFs encoded the coat protein.

CONCLUSIONS

We have cloned and sequenced a further three ssDNA components of the BBTV genome. The CR-SL incorporated the conserved stem-loop structure. The loop sequence of 11 nucleotides was conserved in all BBTV components with the exception of two nucleotides and was similar to that present in nine geminiviruses (Lazarowitz, 1992, Geminiviruses: genome structure and gene function. CFDV (Rohde et al., 1990. Virology 176 648-651). A model for implicating the loop WO 96/38564 PCT/AU95/00311 21 sequence in rolling circle replication has been described for geminiviruses (Saunders et al, 1993, DNA forms of the geminivirus African cassava mosaic virus consistent with the rolling circle mechanism of replication. IXth International Congress of Virology, Glasgow, August, 1993. Abstract P60-18). It is possible that the loop sequence in BBTV has a similar function. The stem-loop sequences were also highly conserved in all BBTV components and contained the pentanucleotide sequence TACCC which has been shown to be the site for initiation of viral strand DNA synthesis in wheat dwarf geminivirus 1 0 (Heyraud et al., 1993, EMBO Journal 12 4445-4452).

The major common region (CR-M) was identified in all components and was located 3' of the major ORF (except for component 2 where no major ORF was identified) and 5' of the CR-SL (FIG. 8).

Hexanucleotide repeats were identified within the CR-M in all three 1 5 components. However, no function could be directly attributed to these repeats but they may be associated with, or part of promoter sequences.

The CR-M also contained a 15 nt GC-rich sequence located at the 3' end and had the potential to form a small stem-loop structure. This GC-rich sequence also contained two direct GC-repeats which resembled the Spl binding sites found in promoters of genes in animal cells and viruses (Fenoll et al., 1990, Plant Molecular Biology 15 865-877). A similar promoter in the monocot-infecting maize streak geminivirus has been shown to be required for maximal rightward transcription and also appeared to bind maize nuclear factors in a non-cooperative manner (Fenoll et al., 1990, Plant Molecular Biology 15 865-877).

The nucleotide length and sequence between the CR-M and CR-SL was dissimilar in four of the six components. However, in components 3 and 4, this 175 nucleotides region was 97% homologous and the 334 nucleotides from the 5' end of the CR-M to the 3' end of the CR-SL were 98% homologous. A similar large common region of 300 nucleotides has been found in geminiviruses and is identical between the WO 96/38564 PCT/AU95/00311 22 A and B components of individual bipartite geminiviruses (Lazarowitz, 1992, Geminiviruses: genome structure and gene function. 2).

Components 3, 4 and 6 each had one large ORF in the virion sense, 3' of the CR-SL. Each of these ORFs had potential conserved TATA boxes and polyadenylation signals associated with them (FIG. The potential TATA boxes highly conserved with the nonanucleotide sequence CTATa/taltAaltA which was essentially similar to that described by Bucher et al., 1990, Journal of Molecular Biology 212 563-578. The distance between the potential TATA box and the 1 0 translation initiation codon varied in each component from 13 nucleotides in component 3 to 102 nucleotides in component 1. An ATGG translation initiation codon was identified in the five components encoding large ORFs. However, two possible translation initiation codons were identified in component 3, the first at nucleotide 213 (ATGT) and the second at nucleotide 227 (ATGG); the second initiation codon was in frame with the first. This would suggest that the second initiation codon is the correct codon; this could be verified by 5' RACE or N-terminal sequencing of the ORF translation product (see Example GT-rich regions were identified 0 to 24 nucleotides 3' of each of the polyadenylation signals in components 3, 4 and 6. Each of these GT-rich regions contained the nucleotide sequence TTG. Both the polyadenylation signals in components 3 and 6 had these sequences. The combination of a consensus polyadenylation signal (Aa/tTAAa/t) and a 3' proximal GT-rich region containing the trinucleotide sequence TTG were associated with the single major virion sense ORF in components 3, 4 and 6 and were not identified elsewhere in these sequences suggesting that each of these components encoded a single gene. Similar sequences have been associated with many polyadenylation signals and may be required for efficient termination (Gil Proudfoot, 1984, Nature 312 473-474; Conway Wickens, 1985, Proceedings of the National Academy of Sciences USA 82 3949-3953).

WO 96/38564 PCT/AU95/00311 23 None of major ORFs of components 3, 4 or 6 had significant sequence homology either at the DNA or protein level with any other available sequences and no functions could be assigned to the putative proteins. However, the ORFs of components 3 and 5 encoded proteins of a size similar to that of the BBTV coat protein, 20.1KDa (Harding et al., 1991, Journal of General Virology 72 225-230). Further, a sequence of hydrophobic amino acids were identified near the N-terminal end of the putative protein of the component 4 ORF which is characteristic of inor trans-membrane domains. Similar hydrophobic domains have been identified in the proposed movement proteins of cereal-infecting geminiviruses but have not been identified in any other virus movement proteins including those of dicot-infecting geminiviruses (Boulton et al., 1993, Virology 192 85-93). No other ORFs were identified in components 3, 4 and 6 in either virion or complementary sense which potentially encoded proteins greater than 10K and had appropriately located potential TATA boxes and polyadenylation signals.

EXAMPLE 2: IDENTIFICATION OF THE COAT PROTEIN GENE OF BANANA BUNCHY TOP VIRUS Semi-purified preparation of BBTV was used as the source of protein for direct sequencing of viral coat protein by Edman Degradation.

The protocol followed was as follows:- A. Identification of BBTV coat protein in semi-purified virus preparation 1. Semi-purified BBTV was prepared from 1 kg of infected midrib tissue using the extraction method of Harding et al., 1991, Journal of General Virology 72 225-230, omitting the final casesium sulphate density gradient step.

2. The virus preparation was mixed with SDS-gel sample buffer (Tris, %glycerol) and electrophoresed in duplicate 4%/12% polyacrylamide gels (BioRad Mini-Protean II).

WO 96/38564 PCT/AU95/00311 24 3. One of the gel was blotted across onto protran nitrocellulose membrane (Schiller Schull) and probed with BBTVspecific monoclonal antibody (Thomas and Dietzgen, 1991, Journal of General Virology 72 217-224). The second gel was stained with coomassie blue C-250 (Bio-Rad).

4. The BBTV-specific band was identified by mapping the immunoreactive band with the banding pattern on coomassie-stained polyacrylamide gel.

The result from the western-immunoblot identified a single 1 0 band of approximately 20kD in the virus preparation. This is most likely to be the viral coat protein as it specifically binds BBTV-specific antibody which was prepared with semi-purified virus preparation and (ii) the size of this protein was estimated to be around 20kD which is the reported size of BBTV coat protein.

1 5 B. N-terminal sequencing of BB TV coat protein 1. The semi-purified BBTV was electrophoresed on 4%/12% polyacrylamide gel.

2. The gel was then blotted onto immobilon PSQ PVDF membrane (Millipore). The transfer was carried out in CAPS buffer (10 mM3-[cyclohexylamino]-1-propane sulfonic acid, pH 11, 0.001% SDS) at 70V/16 hr.

3. The PVDF was stained lightly with coomassie blue R-250 (BioRad) and destained with 12% v/v methanol.

4. BBTV-specific band was then excised and sequenced by Edman degradation method.

A total of 18 amino acid residues were obtained from this experiment.

C. Analysis of sequence data 1. The amino acid sequence obtained was compared to the major open reading frames (ORF) of the six known BBTV component.

This comparison revealed very high homology between the WO 96/38564 PCT/AU95/00311 amino acid sequence obtained and one component of BBTV. On the amino acid level, the 18 residues obtained from direct sequencing has identity to residues 2-19 of the protein encoded by BBTV-3 major ORF. Up to 85% identity level was obtained when the first residue of the unknown sequence, which is always the least accurate residue, was discounted (see FIG. 9).

2. The protein encoded by BBTV-3 major ORF (see FIG. 4a) was used in a search of protein database using Blast (Altschul et al., 1990, Journal Molecular Biology 215 403- 410).

The result from the search revealed similarities between this protein and the major gene product of subterranean clover stunt virus DNA component-5 (SCSV5), which has been identified as the coat protein. SCSV is also a multipartite, single-stranded DNA virus which is 1 5 believed to be closely related to BBTV.

CONCLUSIONS

The experimental result suggests that BBTV-3 encodes the coat protein of BBTV. The amino acid sequence obtained by direct sequencing of the protein which has been shown to bind BBTV-specific antibody, shows very high homology to the protein encoded for by BBTV- 3. Furthermore, the BBTV-3 encoded protein also show considerable homology to the reported coat protein of subterranean clover stunt virus, a virus probably belonging to the same group of new ssDNA plant viruses as BBTV.

WO 96/38564 PCT/AU95/00311 26

TABLES

Table 1. The open reading frames contained within potentially coding for proteins of more than 2K.

BBTV component 2 ORF STRAND LOCATION SIZE(nt) SIZE(aa) MW (K) 2V1 virion 42-116 75 25 2.83 2V2 virion 55-159 105 35 3.87 2V3 virion 143-406 264 88 10.38 2V4 virion 660- 743 84 28 3.52 2C1 comp.

1 43-820 282 94 10.83 2C2 comp. 900- 820 81 27 3.12 2C3 comp. 630-415 216 72 8.42 2C4 comp. 361-215 147 49 5.65 comp. 173-54 120 40 4.67 WO 96/38564 PCT/AU95/00311 27

LEGENDS

TABLE 1 1 complementary strand FIG. 1 The nucleotide sequence of component 3 (Queensland) FIG. 2 The nucleotide sequence of component 4 (Queensland) FIG. 3 The nucleotide sequence of component 6 (Queensland) 1 0 FIGS. 4a, 4b and 4c The nucleotides sequences of BBTV DNA components 3, 4 and 6 and the deduced amino acid sequences of the major ORF of components 3, 4 and 6. The potential TATA boxes are in bold and double underlined; the potential polyadenylation signals are in bold and underlined; the stemloop structure is in italics and underlined, with the stem sequence arrowed; the CR-SL is underlined; the CR-M is in bold and italics; and the ORF is in bold.

FIG. 4a corresponds to Component 3; FIG. 4b corresponds to Component 4; FIG. 4c corresponds to Component 6.

FIG. Determination of the virion-sense orientation of BBTV DNA components 2 to 6. Each blot was separately probed with either 32 P-labelled oligonucleotides (component 2) or full length RNA transcripts (components 3 to 6) specific for the virion- or complementary-sense strands of each respective component. In panel blots were hybridised with probes complementary to the component sequences presented in FIGS. 4a, 4b and 4c and in panel blots were hybridised with probes of the same sequences presented in FIGS. 4a, 4b and 4c. Lane 1: full length clone of each respective component; Lane 2: healthy banana nucleic acid; Lane 3: DNA extracted from purified BBTV virions.

WO 96/38564 PCT/AU95/00311 28 FIG. 6 The aligned stem-loop common regions (CR-SL) of BBTV DNA component 1 to 6. The stem-loop structure in each component is underlined and the loop sequence is in italics. Asterisks indicate nucleotides that are conserved between all components. Dots have been included in some sequences to maximise sequence alignment.

FIG. 7 The aligned major common regions (CR-M) of BBTV DNA components 1 to 6. The 15 nucleotide GC-rich sequence is underlined. Asterisks 1 0 indicate nucleotides that are conserved between all components and triangles indicate nucleotides that are conserved between components 2 to 6 in the first 26 nucleotides covering the deletion in component 1. Dots have been included in some sequences to maximise sequence alignment and the imperfect repeat sequences are shown in italics.

FIG. 8 Diagrammatic representation of the proposed genome organisation of BBTV. The general organisation of all components and a linear representation of each component.

FIG. 9 N-terminal sequencing of BBTV coat protein. Translation of BBTV-3 major ORF (170 amino acids). Amino acid sequence obtained from Edman degradation is displayed in italics. Identical residues are marked with

Claims (9)

1. An isolated DNA molecule which includes a DNA sequence that is substantially similar to a sequence shown in FIG.1.

2. An isolated DNA molecule which includes a DNA sequence that is substantially similar to a sequence shown in FIG.2.

3. An isolated DNA molecule which includes a DNA sequence that is substantially similar to a sequence shown in FIG.3.

4. An isolated DNA molecule which includes a DNA sequence which encodes a polypeptide substantially similar to an ORF region of component 3 as shown in FIG. 4.

5. An isolated DNA molecule which includes a DNA sequence which encodes a polypeptide substantially similar to an ORF region of component 4 as shown in FIi.

6. An isolated DNA molecule which includes a DNA sequence which encodes a polypeptide substantially similar to an ORF region of component 6 as shown in FIG.4.

7. A purified polypeptide including an amino acid sequence substantially similar to an ORF region of component 3 as shown in FIG.4a.

8. A purified polypeptide including an amino acid sequence substantially similar to an ORF region of component 4 as shown in FIG.4b.

9. A purified polypeptide including an amino acid sequence substantially similar to an ORF region of component 6 as shown in FIG.4c.