Novel structure of the genome of Rice yellow stunt virus: identification of the gene 6-encoded virion protein

Footnotes

The nucleotide sequence of the RYSV genomic region spanning the untranscribed intergenic spacer 5' to the G gene to that 3' to the trailer was obtained by sequencing 10 overlapping cDNA clones isolated from a cDNA library of the RYSV genomic RNA (Fang et al., 1994) using a genome-walking strategy initiated with a probe from the G gene. This region consisted of 6560 nucleotides (nt) and each nucleotide was determined by sequencing both strands of at least two different clones. Sequence analysis using the DNASIS software (Hitachi Software Engineering) revealed two ORFs on the vc strand each of which is bordered by the stretch of nucleotides 5'-AAAUAAAACCCCAACA-3', similar to the gene junction sequences found between the other RYSV genes.

The large ORF located at the 3' region of the vc strand could encode a protein of 1967 aa with a deduced molecular mass of 223·6 kDa, and probably encodes the L protein. The transcription initiation sequence of the L gene has been determined to be 5'-AACA-3' by 5'RACE analysis performed on poly(A⁺) RNA from RYSV-infected rice plants (Luo & Fang, 1998). This sequence motif is identical to the 5' end sequence of other RYSV genes (N, P, M and G) and similar to that of gene 3 (5'-AACU-3') or gene 6 (see below). As in the case of genes G, M and 6 of RYSV, non-viral nucleotides were found at the 5' terminus of the mRNA for the L gene preceding the initiation sequence (Luo & Fang, 1998), suggesting the possibility that the initiation of transcription of RYSV genes proceeds via a cap snatching mechanism as originally demonstrated for influenza virus (Krug, 1981). To determine the exact 3' end of the L gene, 3'RACE was performed as described by Fang et al. (1994). The termination sequence of the L gene was defined as 5'-AAAUAAAAA-3', which is consistent with the conserved termination sequence of other RYSV genes. We thus conclude that the L gene contains a 39 nt 5'-untranslated sequence and a 52 nt 3'-untranslated sequence, and in total is composed of 5988 nt extending from positions 7860 to 13847 relative to the 3' end of the RYSV genomic RNA.

The RYSV L protein is the smallest of all characterized non-segmented negative-strand RNA viruses (NNSV) except for the 1608 aa L protein of Borna disease virus (Briese et al., 1994). Nevertheless, the RYSV L protein harbours multiple functional domains typical of the RNA polymerases of NNSV, e.g. the catalytic domain, the RNA template-binding site and a metal-binding motif (data not shown). A phylogenetic tree was generated by comparison of the amino acid sequence of the RYSV L protein with sequences of the L protein of 32 NNSV (Fig. 1). It is clear that the RYSV L protein is most closely related to the L protein of SYNV, also a nucleorhabdovirus (Choi et al., 1992), with a sequence similarity of 36·9 %. However, the RYSV L protein is distinct from the L proteins of other rhabdoviruses and most members of other families of the Mononegavirales in that it is an acidic protein with a calculated isoelectric point of 6·22 (Fig. 1). Inspection of the RYSV L protein sequence revealed an overall Asp+Glu composition of 12·1 % and a Lys+Arg content of 10·9 %. More significantly, the N-terminal 110 amino acids contained 30 % Asp+Glu and 7·3 % Lys+Arg. This acidic domain is not present in the L proteins of the order Mononegavirales, and its function is unknown.

The 3' end of gene 6 was determined by 3'RACE. Gene 6 terminates with the sequence 5'-AAAUAAAA-3' followed by a tetranucleotide CCCC as an untranscribed intergenic spacer before the L gene (Fig. 2). Thus gene 6 is flanked by two junction sequences which are homologous to the RYSV intergenic consensus sequence. Since 5' and 3'RACE have provided evidence for the presence of the mRNA for gene 6 in RYSV-infected rice plants, this confirms that RYSV encodes an extra gene located between the G and L genes that is unique to plant rhabdoviruses.

RYSV P6 is an acidic protein with an isoelectric point of 3·49, and has five potential phosphorylation sites (consensus pattern S/T-X-X-D/E) and one possible aspartic protease motif (D-T-G). Homology analysis and pair-wise comparison of the P6 amino acid sequence was conducted against GenBank/EMBL and SWISSPROT entries, but no clear similarity was found with any protein from these databases. Small non-virion (NV) genes preceding the L gene are present in all characterized novirhabdoviruses (Basurco & Benmansour, 1995; Schutze et al., 1995; Johnson et al., 2000). The RYSV P6 has very limited sequence similarities with the NV proteins: 22·6 % to IHNV and 25·4 % to VHNV only slightly higher than those generated from random sequences (about 20 %).

To elucidate whether RYSV P6 is a viral protein, SDS-PAGE analysis of purified RYSV virions was performed. Since the P6 protein was not visible in the SDS-PAGE profile of the RYSV virion proteins using Coomassie blue R-250 staining (Fang et al., 1994), an antiserum against glutathione S-transferase (GST)P6 fusion protein was raised and used in immunoblot analysis. ORF 6 was inserted into the BamHI site of the pGEX-3X vector (Amersham Pharmacia) in-frame with the GST gene. Following transformation of E. coli BL21 with the recombinant clone pGEX-3X-6 and induction by IPTG, the GSTP6 fusion protein was expressed and purified with the Bulk GST Purification Module kit (Amersham Pharmacia) (Fig. 3a). Rabbit anti-GSTP6 antiserum was prepared as previously described (Luo et al., 1998). The total proteins from purified RYSV, healthy leafhoppers and viruliferous leafhoppers were extracted, separated on a 16 % Tris/Tricine gel (Schagger & von Jagow, 1987), and electro-transferred onto ImmobilonTM-P PVDF membrane (Millipore). Immunoblots were done as described by Fang et al. (1994) using rabbit anti-GSTP6 antiserum diluted 1 : 3000. P6 protein (10·5 kDa) was detected in purified virions and also in viruliferous leafhoppers, which transmit RYSV among rice plants (Fig. 3b). Such a protein band was not detected in the total protein extracted from RYSV-infected rice plants in a similar immunoblot assay (data not shown), probably due to the very low P6 content in infected rice plants. Thus, unlike the NV proteins of novirhabdoviruses and as the first case in the family Rhabdoviridae, the protein encoded by the small ORF between the G and L genes in the RYSV genome is associated with purified virions and appears to be a viral structural protein.

Although P6 appears to be a virion structural protein and can be phosphorylated, its function remains unknown. P6 has a limited (24 %) sequence similarity with the N-terminal 110 aa of the SYNV L protein. Moreover, P6 contains 32 Asp+Glu (34·4 %) and 4 Lys+Arg (4·3 %) with a large net negative charge, similar to the N-terminal acidic domain of the RYSV L protein (see above), although they share only 18 % sequence similarity. This suggests that P6 may have a close evolutionary relationship with the L protein. Reverse genetic studies have demonstrated that although the NV protein of IHNV was not required for virus replication in cell cultures, it greatly improved virus growth (Biacchesi et al., 2000). Therefore, the small gene preceding the L gene may play an important role in rhabdovirus replication.