The Gene Encoding the Nucleocapsid Protein of Gill-Associated Nidovirus of Penaeus monodon Prawns Is Located Upstream of the Glycoprotein Gene

ABSTRACT The ORF2 gene of Gill-associated virus (GAV) of Penaeus monodon prawns resides 93 nucleotides downstream of the ORF1a-ORF1b gene and encodes a 144-amino-acid hydrophilic polypeptide (15,998 Da; pI, 9.75) containing 20 basic (14%) and 13 acidic (9%) residues and 19 prolines (13%). Antiserum to a synthetic ORF2 peptide or an Escherichia coli-expressed glutathione S-transferase-ORF2 fusion protein detected a 20-kDa protein in infected lymphoid organ and gill tissues in Western blots. The GAV ORF2 fusion protein antiserum also cross-reacted with the p20 nucleoprotein in virions of the closely related Yellow head virus. By immuno-gold electron microscopy, it was observed that the ORF2 peptide antibody localized to tubular GAV nucleocapsids, often at the ends or at lateral cross sections. As GAV appears to contain only two structural protein genes (ORF2 and ORF3), these data indicate that GAV differs from vertebrate nidoviruses in that the gene encoding the nucleocapsid protein is located upstream of the gene encoding the virion glycoproteins.

Gill-associated virus (GAV) of Penaeus monodon prawns is a type species of the genus Okavirus in the Roniviridae of the order Nidovirales (5,7,22). Chronic GAV infection, in which replication is restricted to the foci of hypertrophied cells in the lymphoid organ (LO), is ubiquitous in wild and farmed P. monodon prawns on the east coast of Australia (8,33,39). Acute-phase infection, in which GAV spreads to a wide range of tissues, has been linked to farm disease outbreaks since at least 1996 (33,34,35). The tubular helical nucleocapsids and rod-shaped, enveloped virions of GAV are morphologically identical to those of Yellow head virus (YHV), which has caused mass deaths in P. monodon prawns cultured in Asia, and both viruses cause similar cytopathologies (2,3,20,34,35,36). Sequence similarity levels in the ORF1b gene and the ORF3 glycoprotein gene indicate that GAV and YHV are closely related geographic topotypes (6,16,28).
The plasmids pGEX-ORF2, pQE-ORF2, pQE-ORF2-M 11 , and pQE-ORF2-M 61 were transformed into E. coli M15 (pREP4) cells (QIAGEN) to express glutathione S-transferase (GST)-ORF2, His 6 -ORF2, His 6 -ORF2-M 11 , and His 6 -ORF2-M 61 fusion proteins, respectively. Briefly, the overnight cultures were diluted to an A 600 of 0.2 in SuperBroth medium containing 100 g of ampicillin/ml and 25 g of kanamycin/ml and grown at 37°C to an A 600 of 0.6 to 0.8, and protein expression was induced by the addition of 1 mM IPTG (isopropyl-␤-D-thiogalactopyranoside). The cultures were incubated for 2 to 4.5 h and analyzed directly by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) or used to purify GST-ORF2 inclusion bodies. Inclusion bodies were purified by repeated Dounce homogenization and sonication in lysis buffers containing 1 mg of lysozyme/ml and 0.1 to 0.5% Triton X-100 and used directly as an antigen in immunizations (25).
Antiserum was produced in crossbred rabbits injected intramuscularly (two sites; 0.5 ml/site; ϳ0.5 mg of KLH-PN1/site or ϳ0.2 mg of GST-ORF2/site) with GST-ORF2 inclusion bodies or KLH-PN1 peptide emulsified in an equal volume of Montanide ISA70 adjuvant (Seppic, Paris, France). The rabbits were given a booster injection twice at intervals of 2 to 3 weeks, and the specificity and titer of the peptide antiserum were assessed by an enzyme-linked immunosorbent assay (14).
Small pieces (Ͻ1 mm 2 in diameter) of LO tissue were fixed in 4% paraformaldehyde-0.1% glutaraldehyde in 66 mM cacodylate buffer (pH 7.4) for 48 h at 4°C, cryopreserved in polyvinylpyrrolidone-sucrose, snap-frozen in liquid nitrogen, and cryosubstituted with methanol containing 0.5% uranyl acetate for 48 h at Ϫ85°C. The tissue was warmed to Ϫ45°C, embedded in Lowicryl K11 M, and UV polymerized. Ultrathin sections mounted on carbon-and Formvar-coated nickel grids were incubated in a 1:20 dilution of KLH-PN1 or GST-ORF2 antiserum in phosphate-buffered saline (PBS) block solution (PBS, 0.2% bovine serum albumin, 0.2% fish skin gelatin, 20 mM glycine, 0.05% Tween 20) for 2 h. Control sections were incubated in prebleed serum. After being washed in PBS block solution, the sections were incubated for 1 h in goat anti-rabbit immunoglobulin G conjugated with 10-nm-diameter colloidal gold particles (BioCell) diluted 1:40 in PBS block solution and washed twice for 5 min in water. The tissues were stained with 4% uranyl acetate and then with Reynold's lead citrate and examined at 80 kV with a Jeol 1010 transmission electron microscope.
The 435-nt ORF2 gene of GAV encodes a 144-aa polypeptide and is bounded by intergenic sequences of 93 nt to the upstream ORF1a-ORF1b gene and 57 nt to the downstream ORF3 gene (Fig. 1). The ORF2 protein has a deduced molecular mass of 15,998 Da and a pI of 9.75 and is highly hydrophilic, containing 20 basic (14%) and 13 acidic (9%) residues. There is a cluster of 7 arginines in the N-terminal 26 aa, and 5 of the 8 C-terminal residues are acidic. ORF2 also contains no cysteines and 19 prolines (13%), 8 of which are clustered between residues 7 and 22. quences of the GenBank database identified no polypeptides significantly related to ORF2. This is consistent with alignments of pp1ab motifs showing that GAV is evolutionarily very distant from the vertebrate nidoviruses (7). In size, the 144-aa ORF2 is between the N proteins of arteriviruses (110 to 128 aa) (4, 10, 12, 23) and toroviruses (160 to 167 aa) (11,18,30) and significantly smaller than the 377-to 454-aa N proteins of coronaviruses (19). ORF2 displays general structural similarities to torovirus N proteins in the absence of cysteines and a high prevalence of prolines and of basic residues that are likely to facilitate nucleic acid binding (15). It is also noteworthy that the architecture of the helical nucleocapsids and rod-shaped virions of GAV (33,34) and YHV (2, 3, 24, 40, 43) somewhat resemble torovirus nucleocapsid and virion structures (41,42). Moreover, curled virions similar in form to the crescent-shaped torovirus particles (42) have been observed in purified YHV (24), suggesting that okavirus nucleocapsids are also quite flexible. As N proteins play a significant role in defining the structures of nucleocapsids and virions, it is not surprising that the GAV N protein might share more general similarities with the cognate proteins of toroviruses than with those of coronaviruses or arteriviruses, which have distinctly different virion architectures.
Western blotting with antiserum to either the KLH-PN1 peptide or the GST-ORF2 fusion protein detected the ORF2 gene product (molecular mass, ϳ22 kDa) in the LO and gill tissue of a single moribund P. monodon prawn sampled at 6 days after injection (34) with GAV (Fig. 2). In LO, three smaller polypeptides (molecular masses, ϳ21, 20, and 17 kDa) were also detected, although the 17-kDa protein was detected weakly by the GST-ORF2 antiserum (Fig. 2a). Since these initial ORF2 size estimates were based on poorly resolved, prestained protein standards, they were reassessed using biotinylated protein standards. In Western blots with the PN1 peptide antiserum (Fig. 2c and 3a), the native ORF2 protein migrated alongside the 20-kDa biotinylated standard, and the relative migration of the His 6 -ORF2 protein (molecular mass, ϳ21.5 kDa) was consistent with the calculated additional mass (1,639 Da) of its N-terminal His 6 tag. The estimated size of ORF2 was thus revised to 20 kDa, and the sizes of the three smaller ORF2 derivatives were revised to 19, 17, and 14.5 kDa, respectively.
ScanProcite was used to identify potential posttranslational modification motifs that might explain the difference between the calculated (16 kDa) and estimated (20 kDa) masses of the GAV ORF2 protein. Several phosphorylation sites were identified, as is common in RNA-binding proteins, including the Berne torovirus N protein (15). However, phosphorylation alone would not account for the size disparity, which we suspect is due to the electrophoretic mobility of ORF2 being retarded by its intrinsic structure (14% basic and 13% Pro residues). Two smaller forms of the Berne torovirus nucleocapsid (N) protein have been detected in infected cells, although there are conflicting reports about whether these result from proteolysis (15) or internal initiation of translation (30). To establish whether the 14.5-kDa ORF2 derivative detected in LO tissue might be due to internal initiation of translation, His 6 -ORF2 proteins N-terminally truncated upstream of Met 11 and Met 61 were expressed in E. coli and analyzed in Western blots. The PN1 peptide antiserum detected the His 6 -ORF2 (ϳ21.5-kDa), His 6 -ORF2-M 11 (ϳ19-kDa), and His 6 -ORF2-M 61 (ϳ15-kDa) proteins in addition to three minor polypeptides that were proportionally larger and smaller than each primary recombinant protein (Fig. 3a and c). Although bacterial proteins were also obvious, all but the smallest minor His 6 -ORF2-related polypeptides were also detected by the GST-ORF2 antiserum (Fig. 3b). Although the 15-kDa His 6 -ORF2-M 61 protein was similar in size to the 14.5-kDa ORF2 derivative detected in LO tissue (Fig. 2), importantly, unlike the latter, it reacted well with the GST-ORF2 antiserum (compare Fig. 2a and 3b).
We hypothesize, therefore, that the 14.5-kDa ORF2 derivative is more likely the result of proteolysis at a C-terminal position in or beyond the PN1 peptide sequence. However, to be consistent with our data, antibodies to GST-ORF2 would primarily have to target epitopes C terminal to the proteolysis site. In support of this supposition, a protein proportionally smaller than each His 6 -ORF2 fusion protein, suggesting that it was trimmed at a common C-terminal position, was detected by antiserum to the PN1 peptide but not to GST-ORF2 ( Fig.  3b and c). Although the overall homology was poor, a Clustal W multiple sequence alignment (data not shown) identified a trypsin cleavage motif (KR 109 ) in the GAV ORF2 sequence encompassed by the PN1 peptide that was also present at a comparable position in torovirus N proteins (11,18,30). Mutagenesis studies will be required to determine the relevance of this or other potential proteolysis sites in the C-terminal regions of these proteins.
It is of note that the LO tissue used in this analysis was from a moribund prawn acutely infected with GAV and nearing death. Although this fact ensured the presence of high virus protein levels, significant LO structural and cellular degeneration occurs late in infection (34). This degeneration likely involves apoptotic cell death, as has been reported with YHV (17), and explains why ORF2 may have been subjected to proteolysis. Late in infection, nucleocapsids and virions appear in high numbers in gill cells, usually in the absence of major cytopathology, and large arrays of GAV particles accumulate at the external boundaries of epidermal cells and the gill cuticles (34), presumably to be eliminated following molting (21). This accumulation of mature virions likely explains why only intact ORF2 was detected in the gill tissues of the same prawn.
YHV purified from the hemolymph of infected P. monodon prawns by Urografin (Schering) density gradient centrifugation (43) was analyzed by SDS-PAGE and Western blotting to verify that antiserum to the GAV ORF2 gene product would cross-detect the nonglycosylated YHV p20 protein. The three YHV structural proteins (gp116, gp64, and p20) were detected by Coomassie blue staining (Fig. 4a), and the p20 protein, which appeared to be slightly larger than the GAV ORF2 protein synthesized in gill tissue, cross-reacted with the GST-ORF2 antiserum (Fig. 4b).
Ultrathin sections of LO cells from GAV-infected P. monodon prawns were immuno-gold particle labeled using the ORF2 PN1 peptide antiserum and examined by electron microscopy (Fig. 5). Gold particles associated with free, striated (ϳ16-nm-diameter) tubular GAV nucleocapsids within the cell cytoplasm and were often observed at the nucleocapsid ends rather than at positions on their longitudinal surfaces (Fig. 5b). Gold particles also associated with nucleocapsids within mature rod-shaped, enveloped GAV virions (Fig. 5d) and within newly formed virions that had not completely disaggregated into discrete rod-shaped structures and that occurred where nucleocapsids appeared to be actively budding or had recently budded into membranous endoplasmic vesicles (Fig. 5e). Within virions, gold particles were also commonly associated with nucleocapsid ends or with positions where nucleocapsids appeared to have been cross-sectioned transversely. Gold particles were not seen to associate with any specific cellular structures in cytoplasmic regions where GAV nucleocapsids were not evident. Similar results were obtained with the GST-ORF2 antiserum (data not shown), and no association of gold particles with GAV nucleocapsids was observed with preimmune serum (Fig. 5c).
The 6.2-kb region of the GAV genome downstream of the ORF1a-ORF1b gene (GenBank accession no. AY039647) contains no discrete genes, other than ORF2, of sufficient length or structure to encode a viral N protein (5). The 5-kb ORF3 genes of GAV and YHV encode a putative precursor glycoprotein containing six transmembrane domains. In YHV, ORF3 is cleaved after transmembrane domains 3 and 5 to generate the structural glycoproteins gp116 and gp64, respectively (16), the larger of which has so far been shown to protrude from the virion envelope (32). No protein of the size (25.4 kDa) predicted for the ORF3 N-terminal, trimembranespanning cleavage product of YHV (16) and GAV (5) has been detected in virions. Although it has some structural similarities to the triple-membrane-spanning viral membrane (M) proteins of vertebrate nidoviruses, its function remains un- known. The small (9.6-kDa) ORF4 gene coding sequence is unlikely to be translated at sufficient levels to form a major structural component of virions, since no 3Ј-coterminal sgm-RNA for ORF4 is transcribed in abundance (9). Moreover, the ORF4 region in YHV is interrupted by multiple stop codons, indicating that it is not essential for virus replication (29).
Immuno-gold particle labeling of GAV nucleocapsids by antibodies to the ORF2 synthetic peptide and recombinant GST-ORF2 fusion protein offers strong evidence that the 20-kDa ORF2 gene product is the viral N protein. YHV contains a homolog of the GAV ORF2 gene (28,29), and it is clear from data presented here that the ORF2 protein is equivalent to the nonglycosylated p20 structural protein of YHV (16). Moreover, a monoclonal antibody reactive for the YHV p20 protein has also recently been shown to bind to nucleocapsids (32). The finding that the ORF2 gene encodes the GAV N protein conspicuously distinguishes this crustacean okavirus from the vertebrate nidoviruses in which the N protein gene resides in the 3Ј-terminal genome region downstream of genes coding for the virion glycoproteins and membrane protein.
We thank Roger Pearson for help with peptide conjugation and antibody production, Ross Tellam and Tony Vuocolo for useful discussions on bacterial expression, Rick Webb and Malcom Jones, University of Queensland, and Russell McCulloch for assistance with immunoelectron microscopy, and Sasimanas Unajak, Mahidol University, Bangkok, Thailand, for supplying purified YHV.