Complete nucleotide sequence and genome organization of Grapevine leafroll-associated virus 3, type member of the genus Ampelovirus

Footnotes

Recently, the International Committee on Taxonomy of Viruses (ICTV) study group on closteroviruses and allied viruses revised the family Closteroviridae by creating a new genus, Ampelovirus (from ampelos, Greek for grapevine), with Grapevine leafroll-associated virus 3 (GLRaV-3) as its type member (Martelli et al., 2002). The initial proposal to create a new genus for mealy-bug-transmitted closteroviruses was suggested based on the analysis of the recent molecular and biological information in a review (Karasev, 2000). The revised family Closteroviridae now consists of three genera, Closterovirus, Ampelovirus and Crinivirus. The genus Closterovirus with type species BYV has a positive-sense single-stranded RNA (ssRNA) genome and contains viruses that are transmitted by aphids. The genus Ampelovirus with type member GLRaV-3 also has a positive-sense ssRNA with a larger coat protein (3539 kDa) and is transmitted by mealy bugs (coccid or pseudococcid). The genome for viruses in the genus Crinivirus is generally divided into two ssRNA molecules, which are separately encapsidated in virions. All members of the genus Crinivirus are transmitted by whiteflies.

The objectives of the current study were to complete the genome sequence and to understand the genome organization of GLRaV-3. The NY1 isolate of GLRaV-3 (Hu et al., 1990) was used. dsRNA was extracted from phloem tissue of virus-infected grapevines collected from a central New York vineyard according to the method described by Hu et al. (1990). High-molecular-mass dsRNA (about 18 kb) was purified by electrophoresis in low-melting-point agarose gel and extracted using the phenol/chloroform method described in Sambrook et al. (1989).

cDNA synthesis and Lambda ZAPII cDNA library were carried out as described by Ling et al. (1997). Duplicate nylon membranes containing recombinant phage DNA were prepared and used for subsequent screening. A clone walking strategy was used to extend the nucleotide sequence from the previously sequenced portion of the GLRaV-3 genome (Ling et al., 1998). The clone closest to the 5' end of the contig was selected as a probe to screen the cDNA library (Ling et al., 1997) for clones that contained inserts extending further towards the 5' terminus. Using this strategy we were able to walk along the genome step by step to obtain most of the sequence for GLRaV-3. The RT-PCR gap-bridging strategy was applied to fill a sequence gap using sense primer 97-36 (5'-GGTAGAGGGGAGGAATGTGTA-3') and reverse complement primer 98-7 (5'-TAGACTGTTGGTGAAAGACA-3') derived from both ends of sequenced contigs. RT-PCRs were prepared with purified dsRNA as a template and the PCR product was either sequenced directly or cloned into pBlue T-vector (Novagen) and sequenced.

To determine the exact 5' end sequence, poly(A) was added to purified GLRaV-3-specific dsRNA by yeast poly(A) polymerase (US Biochemical) and reverse transcribed using oligo(dT) primer [KSL95-7; 5'-GGTCTCGAG(T)₁₅-3'] and Moloney murine leukemia virus reverse transcriptase, similar to the method that was used to obtain the 3'-terminal sequence (Ling et al., 1998). Using this newly synthesized cDNA as the template, an RT-PCR product was amplified with oligo(dT) primer and the GLRaV-3 specific primer 97-47 (5'-AGGAAGTGGTACGTGGACGC-3'). The RT-PCR product was then cloned into pBlue T-vector for sequencing.

pBluescript SK(+) DNA inserts of selected clones were initially sequenced using T3 or T7 primers. Internal nucleotide sequences were obtained with virus-specific primers. Plasmid DNA was prepared according to the manufacturer's instruction for mini alkaline-lysis/PEG precipitation (Applied Biosystems); the sequencing reaction of the cloned DNA was prepared using ABI Taq DyeDeoxy Terminator cycle sequencing kit (Applied Biosystems) and sequenced by using an ABI 373 automatic sequencer at Cornell University, Geneva, NY.

Nucleotide sequences were assembled and analysed using the DNASTAR sequence analysis package (Madison, WI). Nucleotide sequences were entered into EditSeq and assembled using the SeqMan program. The MegAlign program was used to depict amino acid sequence similarity of GLRaV-3 with respect to other closteroviruses. The CLUSTAL W method of MegAlign was used to compare multiple sequences, to identify consensus sequences and to generate putative phylogenetic relationships. The 272 aa residue region starting from aa residue 460 of GLRaV-3 ORF1a that contains the MT conserved motifs was used for multiple alignment. The tentative phylogenetic tree was generated using CLUSTAL W method, followed by the neighbour-joining method in the MegAlign program with PAM250 residue weight table. To assess further the phylogenetic relationship, bootstrap was used to obtain a consensus tree and Tobacco mosaic virus (TMV) tomato strain (L) with protein ID CAA26085.1 (GenBank accession no. X02144) was selected as an outgroup member. Other sequences that were taken for alignment were derived from GenBank for BYV with protein ID CAA51871.1 (X73476), CTV with protein ID AAC59623.1 (U16304), LIYV with protein ID AAA61797.1 (U15440), LChV with protein ID CAA71285.1 (Y10237), GLRaV-2 with protein ID AAC40855.1 (AF039204) and PMWaV-2 with protein ID AAG13938.1 (AF283103).

The GLRaV-3 genome encompasses 13 open reading frames (ORFs) with 5' and 3' UTRs of 158 and 277 nt, respectively. Designations of ORF1a, 1b and ORF213 were done using the convention for closteroviruses (Dolja et al., 1994). Previously, we reported on the 13 154 nt sequence of the 3'-terminal two-thirds of the genome (Ling et al., 1998). In this work, the 5'-terminal 4765 nt sequence was completed to make a total of 17 919 nt for the entire genome of GLRaV-3, NY1 isolate (Fig. 1). Ten representative cDNA clones were selected for this sequencing project. The criteria used for clone selection were based on each clone's sequence overlapping with the existing sequence contig and extending toward the 5'-terminal region of the genome. Their cDNA insertion sizes (ranging from 370 to 1940 bp) and location relative to the genome are presented in Fig. 1.

The 5'-terminal sequence was determined with sequences derived from two cDNA clones (namely #24 and #31) and the RT-PCR product amplified from dsRNA after 3' polyadenylation (PCR 9-22). cDNA sequences matched one another and were confirmed by the sequence generated from the RT-PCR product. GLRaV-3 had the longest 5' UTR (158 nt) among closteroviruses sequenced to date, followed by BYV and CTV (107 nt), LIYV (97 nt) and LChV-1 (76 nt). The 5'-terminal nucleotide was assumed to be a single C after removal of an extra nucleotide (C) from the dsRNA-minus strand. The assumption of an extra nucleotide on the dsRNA-minus strand was based upon the report for CTV (Karasev et al., 1995), where an additional nucleotide (G) was removed from its dsRNA-minus strand. Similar to other ssRNA viruses, the 5' UTR of GLRaV-3 had low G+C content (30 %, 54 out of 158 nt). The resulting low degree of secondary structure would facilitate ribosome binding in the 5' UTR and initiate an efficient translation process. Pairwise comparison of the GLRaV-3 5' UTR with that of BYV or CTV showed significant nucleotide sequence similarity, with 50 % to BYV and 41 % to CTV (data not shown). However, the three viruses did not show significant consensus sequences in the 5' UTR. High sequence similarities in the 5' UTR regions may be due to AT-rich stretches.

ORF1a started from the first ATG at positions 159161, which had a favourable context for translational initiation with G at the +4 and 3 positions. The ORF1a terminated at nt 68706872 and encoded a large polypeptide of 245 277 Da. In a multi-sequence alignment of GLRaV-3 with another ampelovirus (PMWaV-2), using 92 aa residues located from amino acid position 280 to 371 of GLRaV-3, ORF1a revealed a papain-like protease characteristic signature similar to catalytic cysteine and histidine residues (Fig. 2). The GLRaV-3 p-protease was predicted to cleave ORF1a polypeptide between residues Gly³⁷¹ and Gly³⁷², similar to the BYV papain-like protease. This would generate an N-terminal peptide of 371 aa of 41 963 Da. Unlike GLRaV-2 (Zhu et al., 1998) and CTV (Karasev et al., 1995), which possess two papain-like proteases, GLRaV-3 only had a single papain-like protease.