INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The proteolytic processing of viral precursor proteins is a crucial step in the life cycle of a majority of viruses that infect eucaryotic cells, and virus-encoded proteases are generally associated with these events. Each viral family may use different strategies for the generation of the cleavage products, which does occur in time and in a specific cellular environment for virus assembly (4). Most of the virus-encoded proteases are related to the chymotrypsin-like cellular proteases with a distinctive double -barrel fold, but only a few, like the Alphavirus capsid protease and the Flaviviridae NS3 proteases, contain the His-Asp-Ser catalytic triad found in the cellular world (14, 31). A large number of viral proteases structurally related (or not) to chymotrypsin-like proteases replace the serine to cysteine as the catalytic nucleophile, as for picornavirus 3C protease or adenovirus protease (1, 9). In addition to this substitution, a number of viral enzymes (also found in adenovirus and some picornaviruses) utilize glutamic acid, instead of the catalytic aspartic acid (9, 22). The human cytomegalovirus (HCMV) protease consists of a single -barrel structure and presents a catalytic triad His-His-Ser (26). Retroviruses, such as human immunodeficiency virus, have an active aspartic protease which is only generated when two identical amino acid domains, each of them bearing a catalytic aspartic acid, join to create a homodimer (35).

Infectious pancreatic necrosis virus (IPNV) is a pathogen that causes an acute, contagious disease of young salmonid fishes (10). IPNV is a member of the birnaviruses, a family of double-stranded RNA viruses with two genomic segments of 2.8 (segment B encoding the RNA polymerase) and 3.1 (segment A) kilobases (11). Translation of the segment A yields a polyprotein and a small protein, VP5. The polyprotein, whose protein order is NH₂-pVP2-VP4-VP3-COOH, is cotranslationally processed and cleaved at the pVP2-VP4 and the VP4-VP3 junctions by the VP4 (also named NS)-associated protease activity (20, 21). The pVP2-to-VP2 conversion involves the cleavage (or cleavages) of pVP2 near the carboxy end as has been shown for infectious bursal disease virus (IBDV) (3), a birnavirus which causes a highly contagious disease of chickens (6). This processing is slow and has been proposed for IPNV to be most likely due to host cell proteases rather than the proteolytic action of the VP4 (11). The peptide bonds cleaved in birnavirus polyprotein are ill defined, but dibasic residue motifs have been proposed to be cleaved in IPNV and IBDV (11). Low sequence homology (<20%) is observed between birnavirus VP4 proteases (Fig. 1).

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Patterns of conserved amino acids in birnavirus VP4s. Arrows indicate approximate positions of the cleavage sites. Three regions, A to C, containing the most conserved patterns of amino acids, are shown. The comparison includes the deduced amino acid sequences of VP4 of IPNV strain SP (accession number U56907), IBDV (accession number P15480), and DSX (accession number U60650). Identical residues are in boldface when present at least in two sequences. Numbers refer to the position of the first amino acid of each domain in the respective VP4s. Residues proposed to be part of the catalytic site of VP4s () are indicated.

In this study, we constructed IPNV VP4 mutants by site-directed mutagenesis and analyzed the processing of the mutated polyprotein in vitro. We identified two VP4 residues (serine 633 and lysine 674), which have properties for active-site residues. The behavior of the polyprotein altered at these positions suggests that VP4 may use a serine-lysine mechanism similar to those proposed for signal peptidases from Escherichia coli and Bacillus subtilis, bacterial LexA-like and Tsp proteases, the class A -lactamases and, in eucaryotes, the two subunits of the mitochondrial inner-membrane protease (characterized in the yeast Saccharomyces cerevisiae) (23, 24, 30, 32, 33). The exact cleavage sites at the pVP2-VP4 and at the VP4-VP3 junctions were determined and are characterized by the Ser/Thr-X-Ala-Ser/Ala sequence, a motif which shares some similarities with bacterial signal peptidase and herpes simplex virus cleavage sites.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of IPNA expression vector SK-IPNA. The pBluescript II SK() phagemid (Stratagene) was cut with SspI and HincII and self-ligated to delete the T7 promoter (plasmid SK). The complete IPNA segment of the IPN Sp strain (French isolate 31-75) (13) was cloned by reverse transcription (RT)-PCR by using standard techniques and procedures (27) by using the oligonucleotides 5'-CGTCGGATCCTAATACGACTCACTATAGGAAAGAGAGTTTCAACGTTAGTGGT and 5'-TGCAGCATCGATGGGGCCCCCTGGGGGGCCGGGGTT and cloned into SK by using restriction sites BamHI and ClaI to generate the plasmid SKIPNA. Thus, the 5' terminus of the IPNA segment (GGAAAGAGA...) was downstream from the T7 promoter. The entire segment was sequenced and compared to the published segment A sequences. The sequence encoding the polyprotein was found to be 98% identical to the published IPNA Sp strain (U56907), resulting in nine amino acid differences (V52I, V152A, K192R, R212S, N234S, Q249R, D252N, K255R, and M883T).

Site-directed mutagenesis of SK-IPNA. The mutations were introduced by using the Pfu DNA polymerase with the QuikChange Site-Directed Mutagenesis Kit (Stratagene) as described by the manufacturer. A novel restriction site was introduced for some mutations (Table 1) and sequence analyses (5'-GGGGTCCTCTTTTGACCACTCGTA) were carried out to confirm amino acid changes and the integrity of reading frames by using the following primers: 5'-GAAAATTCTCCCGAGCCCTCAAG, 5'-CCACAGGGACGATGACTCCTTTTG, 5'-GAACGACATCGAGGACGGAGTTCC, 5'-TACCATCCTTGGAACTCCGTCCTC, and 5'-GGGGTCCTCTTTTGACCACTCGTA.

Construction of vectors expressing truncated IPNA segments. A number of SK-IPNA-derived constructs encoding putative cleavage sites were used to generate a set of T7 expression plasmids from which N- and C-terminally truncated forms of the polyprotein could be expressed in E. coli or in vitro. To obtain a reading frame carrying the putative pVP2-VP4 and VP4-VP3 cleavage sites (construct VP4-VP3), SK-IPNA was digested with NcoI (nucleotide 1596) and SalI (present in the polylinker) and the insert was cloned NcoI-XhoI in a modified pET-22b (Novagen), in which the pelB leader sequence was deleted and the NcoI restriction site was located on the initiation codon. A series of VP4-VP3 0 to 5 constructs were generated by using PCR with the Pfu DNA polymerase and adequate oligonucleotides for cloning into the BamHI restriction site of pET-28b (Novagen). Thus, six histidines and a T7 epitope were present at the N terminus of the open reading frame (ORF) and six histidines were present at the C terminus of VP3, resulting in the addition of 33 and 21 residues at their N and C termini, respectively. In the VP4-VP3 1 AS734-5LE construct, the VP4-VP3 1 construct was subjected to mutagenesis as described above to convert the Ala-Ser dipeptide at positions 734 and 735 to the dipeptide Leu-Glu. Generation of the VP4-His 1 construct was carried out by PCR by using adequate oligonucleotides for cloning into the BamHI-XhoI restriction sites of pET-28b. A silent mutation was introduced in the oligonucleotide in 5' position to destroy the XhoI site in nucleotide position 1628.

In vitro expression and protein labeling. In vitro, T7-derived expression was carried out using the TNT Quick Coupled Transcription/Translation System (Promega) as described by the manufacturer, except that reactions were performed in a final volume of 11 µl. The DNA template (1 µg) was incubated 1.5 h at 30°C. After incubation, aliquots of 2 to 3 µl were submitted to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (15), and gels were dried and exposed for autoradiography.

Expression of 5'-truncated IPNA polyprotein in E. coli, purification, and N-terminal sequence determination of generated polypeptides. Overnight liquid cultures (250 ml) of BL21(DE3) carrying the constructs ipnVP4-VP31, ipnVP4-VP31AS734-5LE, and ipnVP4-His1 were half-diluted in L medium with kanamycin (50 µg/ml) and induced by 1 mM IPTG (isopropyl--D-thiogalactopyranoside) for 4 h at room temperature. Cells were collected by centrifugation at 5,000 × g for 5 min, washed in 50 ml of 50 mM Tris-HCl (pH 8)-2 mM EDTA and resuspended in 15 ml of 50 mM Tris-HCl (pH 8)-60 mM NaCl with a cocktail of protease inhibitors without EDTA (Boehringer Mannheim). Lysozyme was added to a concentration of 800 µg/ml, and the mixture was placed in an ice bath for 30 min. Then, 50 µl of benzonase (Boehringer) and 42 µl of 1 M MgCl₂ were added for 10 min. The lysates were centrifugated at 13,000 rpm for 1 h at 4°C. The supernatants were diluted in 20 mM Tris-HCl (pH 8)-5 mM imidazole-0.5 M NaCl. The pellet was resuspended in the same buffer complemented with 6 M urea and solubilized overnight at 4°C under mild agitation. This material was submitted to centrifugation (13,000 × g, 30 min) for clarification. Aliquots of primary and second supernatants were submitted to SDS-PAGE. For the constructs VP4-VP3 1 and VP4-VP3 1 AS734-5LE, the VP3 polypeptides were found solubilized in the second supernatant, and they were processed for Ni²⁺ affinity chromatography under denaturing conditions as described by Novagen. The VP3 polypeptides were found eluted by the 200 mM imidazole buffer. Approximately 50 pmol was processed for automated N-terminal Edman sequencing by using a Perkin-Elmer/Applied Biosystems Procise 494A sequencer with reagents and methods of the manufacturer. The VP4 polypeptide was found in the primary supernatant of the construct VP4-His1. Supernatants (10 µl) were loaded onto a polyacrylamide gel for separation and transferred to a ProBlott membrane (Applied Biosystems) for N-terminal sequencing. The band was visualized on the membrane by Coomassie blue coloration.

High-performance liquid chromatography coupled with micro-IS-MS. The reversed-phase high-performance liquid chromatography (RPLC) was online coupled with an ion-spray mass spectrometer (IS-MS) (Sciex API100; Perkin-Elmer). RPLC was run using a Perkin-Elmer device (Applied Biosystems 140D pump and 785 UV detector with U-shaped fused silica tubing of 7-mm pathlength) on a 0.5-by-150-mm column packed with 5-µm (300 Å) C₁₈ silica in 4 mM CH₃COONH₄-0.1% HCOOH with a CH₃CN linear gradient (4.5% for 5 min, from 4.5 to 9% in 15 min, from 9 to 51% in 100 min at a flow rate of 5 µl/min). After being monitored for absorbance at 215 nm, the flow was split between the micro-IS source (0.2 µl/min) and the microblotter (Perkin-Elmer/Applied Biosystems 173A) for polyvinylidene difluoride membrane blotting. IS-MS experiments were controlled with the Sample Control 1.3 software by using a positive mode from 400 to 2,100 amu with 0.2 amu steps and a 0.4-ms dwell time. The micro-IS voltage was 5,000 V, and the orifice plate voltage was 40 V. MS data were analyzed with Perkin-Elmer Sciex Bio-Multi-View 1.2 software. The average molar masses were calculated from sequences by using the Perkin-Elmer Sciex Peptide Map 2.2 software.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Candidates for functionally important residues of the VP4 protease. A site-directed mutagenesis approach was used to map VP4 residues that are functionally important for protease activity. Positions for mutagenesis were chosen on the basis of several criteria. First, residues must be conserved in the different sequences of IPNV strains available in sequence databases. Second, each of the (conserved) histidines, the aspartic acids, the seven glutamic acids, and the two serine residues were mutated because these residues are often important components of known proteolytic enzymes (such as the chymotrypsin-like proteases or aspartyl proteases, enzymes encoded by a large number of viruses). Third, glycine 631 and leucines 636 and 638 were mutated because these residues fall in a small region that shows homology with other birnaviruses (Fig. 1, region B) and with an active-site region of chymotrypsin (MGPSA in VP4; MGDSG in chymotrypsin). Finally, alignments between the VP4 of IPNV, IBDV, and Drosophila X virus (DXV) shown in Fig. 1 were used as a guide, since these proteins share about 40% identity in the three regions shown but have little homology in other regions. Altogether, 42 positions were selected, full-length VP2, VP3, and VP4 ORF proteins carrying mutations were constructed (Table 1) and expressed in a rabbit reticulocyte lysate system, and their processing was analyzed by SDS-PAGE.

Activity of VP4 variants. Processing of the wild-type polyprotein yielded the expected cleavage products pVP2 (62 kDa), VP3 (32 kDa), and VP4 (29 kDa) (Fig. 2A) as observed previously (21). In a first set of experiments, all conserved histidine and aspartic acid residues of VP4 were replaced by different residues. As shown in Fig. 2A, H547S and H679L mutants showed an efficient activity in the cleavage of the pVP2-VP4 junction but had a strongly reduced activity of approximately 50 and 20% of the wild-type activity for the VP4-VP3 junction, respectively. Two additional bands of 34 and 27 kDa were visualized with these mutants. Immunoprecipitation with the anti-VP3 4F4 monoclonal antibody (kindly provided by J. Dominguez) showed that the 34-kDa band was a longer VP3, suggesting that cleavage specificity was altered in these mutants with a subsidiary cleavage site present in the carboxyl part of VP4 (not shown). A double mutant, H547S H679L, was constructed. These changes resulted in a nearly complete loss of cleavage activity at both VP4 junctions (Fig. 2A). In contrast, the H704S mutant showed a wild-type activity. The D693L mutant showed band patterns similar to those of H547S and H679L mutants, with a cleavage at the VP2-VP4 junction and a partial cleavage at the VP4-VP3 junction, whereas the seven mutants D573Q, D585I, D595L, D601S, D644I, DD660-1GS, and D672N had a wild-type activity (Fig. 2B). Two additional substitutions (with glutamic acid and asparagine) were carried out on position D693. The two mutants exhibited a wild-type activity (Fig. 2B). To assess the possible role of the glutamic acids at positions 525, 530, 531, 533, 570, 594, and 680, these residues were also replaced by other residues. As shown in Fig. 2B, all of these mutants have an activity similar to that of the wild type. These first data indicated that H547, H679, and D693 residues are important for both cleavage activity and specificity between VP4 and VP3 but are not essential for the cleavage between pVP2 and VP4.

View larger version (35K): [in this window] [in a new window]   FIG. 2.   Mutagenesis of the conserved histidine, aspartic acid, and glutamic acid residues of the IPNV VP4. The autoradiographs show the results obtained with SKIPNA wild type (wt) and a set of mutant SKIPNA-derived constructs encoding proteins with a single (or double) amino acid substitution(s). (A) Substitution of histidine residues. (B) Substitution of aspartic acid and glutamic acid residues. The constructs were expressed with the rabbit reticulocyte expression system (Promega), and expression products were analyzed by SDS-PAGE. p2, pVP2; 3, VP3; 4, VP4; 4-3, VP4-VP3; p2-4-3, uncleaved polyprotein precursor. Note that single mutations can strongly affect the behavior of the VP4 in SDS-PAGE.

View larger version (84K): [in this window] [in a new window]   FIG. 3.   Mutagenesis of the conserved serine residues in the B region of IPNV VP4. A number of substitutions were introduced at the position of serines 633 and 639. Expression was carried out as described for Fig. 2.

View larger version (26K): [in this window] [in a new window]   FIG. 4.   Mutagenesis of conserved residues in the regions A, B, and C of VP4. The autoradiographs show the results obtained with SKIPNA wild type (wt) and a set of mutant SKIPNA-derived constructs encoding proteins with a single amino acid substitution. Region A, left panel; region B, middle panel; region C, right panel. Expression was carried out as described for Fig. 2.

View larger version (76K): [in this window] [in a new window]   FIG. 5.   Mutagenesis of the lysine 674 of IPNV VP4. A number of substitutions were introduced at the position of lysine 674. Expression was carried out as described for Fig. 2.

View larger version (65K): [in this window] [in a new window]   FIG. 6.   Mutagenesis of hydrophobic residues in the region C of the IPNV VP4. Expression was carried out as described for Fig. 2.

Identification and mutagenesis of the cleavage site(s) at the VP4-VP3 junction. Considering that VP4 protease was shown to be functional in E. coli (20), we used this expression system to map the VP4 cleavage sites. We engineered two constructs encoding NH₂-truncated forms of the polyprotein: VP4-VP3 with the methionine 494 as initiation methionine and VP4-VP3 1 with a six-His tag located at both termini of the polyprotein (Fig. 7A). The latter construct VP3 product was expected to be easily purified by nickel affinity chromatography and further analyzed by N-terminal sequencing. Preliminary results indicated that (i) cleavage was efficient at the VP4-VP3 junction with the two constructs expressed in E. coli or in reticulocyte lysates and (ii) the VP3-His polypeptides immunoprecipitated with an anti-VP3 antibody comigrated, suggesting that cleavage occurred at the same position in both expression systems (not shown). E. coli-expressed VP3-His protein was purified, and its first six amino acids were determined through N-terminal sequencing: NH₂-Ser-Gly-Met-Asp-Glu-Glu. The sequence is identical to the IPN sequence starting at serine residue 735. We further tested the effect of the double substitution AS734-5LE on the polyprotein processing in the reticulocyte expression system (Fig. 8B). As expected, this double mutation resulted in the generation of the uncleaved VP4-VP3 polypeptide, but two additional polypeptides of 34 and 27 kDa were also identified. Thus, in this mutant, the cleavage site generating VP3 (32 kDa) and VP4 (29 kDa) was not functional, but part of the VP4-VP3 polypeptide was cleaved at another site.

View larger version (12K): [in this window] [in a new window]   FIG. 7.   Mapping of the cleavage site at the VP4-VP3 junction. (A) Scheme of the set of three 5'-truncated expression products. Numbers refer to the amino acid position on the full-length ORF polyprotein, and additional sequences are indicated in italics. The HisTag-T7Tag (NH2 extremity) and the HisTag (COOH extremity), which derive from pET-28 vector, are made of 33 residues and 21 residues, respectively. Mutated residues are indicated in boldface. (B) Expression in reticulocyte lysates of the polyprotein mutant carrying a double substitution at the P1-P'1 position of the cleavage site between the VP4 and VP3 proteins. 3*, 34-kDa band; 4*, 27-kDa band.

View larger version (39K): [in this window] [in a new window]   FIG. 8.   Mapping of the cleavage site at the pVP2-VP4 junction. (A) Schematic representation of the set of 5'-truncated expression products. Numbers refer to the amino acid position on the full-length ORF polyprotein, and additional sequences indicated in italics are as described in Fig. 8. (B) Expression was carried out as described for Fig. 2. (C) Mutagenesis of the cleavage site. Expression in reticulocyte lysates of polyprotein mutants carrying double (at the P1-P'1 position) or single (at the P1, P'1, and P'2 positions) substitutions of the cleavage site between the pVP2 and VP4 proteins. 2*, short pVP2.

Identification and mutagenesis of the cleavage site at the pVP2-VP4 junction. The sizes of the VP4s generated by the proteolytic processing of either the VP4-VP3 construct or the wild-type polyprotein were first compared (Fig. 8). Samples were run side by side in the same gel (Fig. 8B). As a result, the VP4s comigrated, suggesting that VP4 expressed from VP4-VP3 contained the natural N terminus of VP4. A series of NH₂-truncated forms of the wild-type polyprotein with an increment of 10 residues and six-His tagged at both termini were constructed (Fig. 8A). The VP4 expressed from VP4-VP3 0 and VP4-VP3 1 comigrated with VP4, suggesting that the N terminus of VP4 was present in these constructs (Fig. 8B). In contrast, the VP4 expressed from VP4-VP3 2 and 3 migrated more slowly than the natural VP4, suggesting that the cleavage site at the pVP2-VP4 junction was not present in these constructs. Thus, the domain located between valine 503 and glycine 514 contains determinants associated with (or close to) the cleavage site between pVP2 and VP4. These results also suggest that the residues which are deleted in constructs 4 and 5 are necessary for the cleavage at the VP3-VP4 junction.

Mutagenesis of potential additional cleavage site(s) in the carboxyl domain of pVP2. Figure 9 shows a sequence comparison between the pVP2-VP4 and the VP4-VP3 cleavage site regions. The Ala-Ser dipeptide was found as the conserved scissible peptidyl group bound in both cleavage sites. A serine at the P3 position and a glycine at the P'2 position also appeared to be conserved within the two cleavage sites. Two other putative cleavage sites corresponding to the (Ser/Thr)-X-Ala-Ala-Gly motif were also found in the COOH terminus of pVP2. To determine whether these sites are possible targets for VP4, the effect of Ala-Ala to Gln-Leu substitutions at positions 486 and 487 and positions 495 and 496 was analyzed. Substitutions present simultaneously on the same mutant in positions 486 and 487, 495 and 496, and 508 and 509 indeed completely inhibited the generation of pVP2 and shorter pVP2 forms, suggesting that all potential cleavage sites at the pVP2-VP4 junction domain are blocked in this construct (Fig. 10). The two cumulated substitutions at positions 495 and 496 and positions 508 and 509 generated a shorter pVP2 and a larger VP4 than when the residues at positions 508 and 509 were only substituted, suggesting that the Ala-Ala dipeptide at positions 495 and 496 is cleaved in the mutant AS508-9QL and that the Ala-Ala dipeptide 486-487 is efficiently used in the mutant with the cumulated substitutions AA495-6QL and AS508-9QL. These data indicate that, in addition to the cleavage site at the pVP2-VP4 junction identified between residues 508 and 509, two subsidiary cleavage sites at positions 495 and 496 and positions 486 and 487 in the C-terminal part of pVP2 were found as potential targets of the VP4 protease.

View larger version (17K): [in this window] [in a new window]   FIG. 9.   Cleavage sites (and putative cleavage sites) for the IPNV VP4 in the polyprotein. (Top) Schematic representation of the IPNA polyprotein. The amino acids of the VP2-VP4 and VP4-VP3 domains are indicated. The P1-P'1 cleavage site positions are underlined. (Bottom) The sequences relative to cleavage sites identified by N-terminal sequencing are in boldface. The sequences in regular typeface were identified by sequence homology and probed by mutagenesis.

View larger version (46K): [in this window] [in a new window]   FIG. 10.   Mutagenesis of potential cleavage sites for the VP4 protease in the COOH part of pVP2. Expression in reticulocyte lysates of the polyprotein mutants carrying double substitution(s) at the P1-P'1 position of the potential cleavage sites (positions 486 and 487 and positions 495 and 496). p2-4, VP2-VP4; 2*, short pVP2.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The initial objective of this study was to identify amino acid residues that play an essential role in the catalytic activity of IPNV VP4 protease and to characterize its substrate cleavage sites. The relative importance of residues belonging to conserved domains in the birnavirus VP4 homologs was also addressed.

Active-site mutations. To identify critical residues for the VP4 protease function, we first postulated that VP4 is either a serine or an aspartyl protease, and we mutated conserved residues possibly associated with the catalytic site. Thus, histidines, aspartic acids, glutamic acids, and serines found in the glycine-X-serine signature of serine hydrolases were substituted in the VP4 domain. Only changes at the serine 633 completely eliminated polyprotein processing at both pVP2-VP4 and VP4-VP3 junctions (Table 2). Replacing serine 633 by a cysteine resulted in a partially active protease, while replacing it by a threonine, alanine, or glycine completely blocked protease activity. The sulfhydryl group of the cysteine is expected to react chemically in a way similar to the serine hydroxyl group and is known to act as a nucleophile in cysteine proteases. These findings make serine 633 the strongest candidate active-site nucleophile so far identified and lead us to postulate that the IPNV VP4 is a member of the serine proteinase superfamily. The possibility that the VP4 belongs to aspartic or metallo superfamilies is unlikely. First, unlike aspartic proteases, VP4 was shown to be insensitive to Pepstatin-A and H-261 (11). In addition, none of the mutations of the nine conserved aspartic acids leads to an inactive protease. Second, unlike metalloproteases, VP4 activity was not inhibited by the metal chelator, EDTA (11). Moreover, none of the substitutions of a conserved histidine completely eliminated protease activity.

Cleavage sites. Two IPNV VP4 cleavage sites, located at the pVP2-VP4 and VP4-VP3 junctions, were identified by N-terminal sequence analysis of cleavage products produced in E. coli and probed by site-directed mutagenesis. They are characterized by the Ser-X-AlaSer-Gly motif. Two other additional cleavage sites in the carboxyl part of pVP2 (P1 and P'1 positions 486 and 487 and positions 495 and 496) were first identified by sequence comparison (Fig. 10). The P1 and the P'2 residues appeared to be conserved as an alanine and a glycine, respectively. The P3 serine residue can substitute to a threonine (cleavage site positions 486 and 487), whereas the P'1 serine residue was notably substituted to an alanine. The cleavage between pVP2 and VP4 was abolished only when the Ala-Ser 508-509 and the two Ala-Ala 495-496 and 486-487 pairs were mutated together. These observations do not prove definitively the existence (and the functionality) of these two additional cleavage sites. However, these data correlate well with the fact that pVP2 is slowly processed at its carboxy end (3) and strongly suggest that conversion of pVP2 to VP2 is associated with these VP4 substrate cleavage sites. But these data also raise a question. Why does expression of the polyprotein in in vitro translation experiments (11; this study) or in a baculovirus-derived system (19) only produce pVP2 without further processing of pVP2 to VP2, as observed in IPNV-infected cells? A possible explanation may be that the pVP2 product, which results from the self-cleavage of the polyprotein, is an intramolecular process, in contrast to the pVP2-to-VP2 conversion that involves a bimolecular reaction. Thus, the local concentration of pVP2 and VP4 may dramatically influence the rate of conversion to VP2, and it is likely that, in infected cells, the concentration of viral proteins is higher in the capsid assembly site than with recombinant expression systems. It cannot be excluded that host cell proteins may also contribute to the VP2 conversion. It was also remarkable that, compared to the pVP2-VP4 cleavage, processing of the VP4-VP3 junction was much more sensitive to substitutions in the VP4 (positions H547 and H679 for instance). This possibly indicates that the two cleavage sites have different structural properties and/or the association of the protease with as-yet-unidentified cofactors.