key: cord-259095-mfptcw8t
authors: Lu, Yiqi; Denison, Mark R.
title: Determinants of Mouse Hepatitis Virus 3C-like Proteinase Activity
date: 1997-04-14
journal: Virology
DOI: 10.1006/viro.1997.8479
sha: 
doc_id: 259095
cord_uid: mfptcw8t

Abstract The coronavirus, mouse hepatitis virus strain A59 (MHV), expresses a chymotrypsin-like cysteine proteinase (3CLpro) within the gene 1 polyprotein. The MHV 3CLpro is similar to the picornavirus 3C proteinases in the relative location of confirmed catalytic histidine and cysteine residues and in the predicted use of Q/(S, A, G) dipeptide cleavage sites. However, less is known concerning the participation of aspartic acid or glutamic acid residues in catalysis by the coronavirus 3C-like proteinases or of the precise coding sequence of 3CLpro within the gene 1 polyprotein. In this study, aspartic acid residues in MHV 3CLpro were mutated and the mutant proteinases were tested for activity in anin vitro transcleavage assay. MHV 3CLpro was not inactivated by substitutions at Asp3386(D53) or Asp3398(D65), demonstrating that they were not catalytic residues. MHV 3CLpro was able to cleave at a glutamine–glycine (QG3607-8) dipeptide within the 3CLpro domain upstream from the predicted carboxy-terminal QS3635-6cleavage site of 3CLpro. The predicted full-length 3CLpro (S3334to Q3635) had an apparent mass of 27 kDa, identical to the p27 3CLpro in cells, whereas the truncated proteinase (S3334to Q3607) had an apparent mass of 24 kDa. This 28-amino-acid carboxy-terminal truncation of 3CLpro rendered it inactive in atranscleavage assay. Thus, MHV 3CLpro was able to cleave at a site within the putative full-length proteinase, but the entire predicted 3CLpro domain was required for activity. These studies suggest that the coronavirus 3CL-proteinases may have a substantially different structure and catalytic mechanism than other 3C-like proteinases.

The role of aspartic acid or glutamic acid residues in 3CLpro activity is less well understood. Mutagenesis of The coronavirus, mouse hepatitis virus, strain A59

Asp/Glu residues of 3C and 3C-like proteinases suggests (MHV-A59), contains a chymotrypsin-like proteinase that they might not participate as catalytic residues in all within the 750-kDa gene 1 polyprotein ( Fig. 1) (Lu et al., cases. Aspartic/glutamic acid residues have been shown 1995). The 3C-like proteinases of the coronaviruses to be essential for proteinase activity of tobacco etch MHV-A59, infectious bronchitis virus (IBV), and the huvirus (TEV), poliovirus, and human rhinovirus 14 (HRVman coronavirus 229E (HCV-229E) are encoded in a con-14) (Gorbalenya and Koonin, 1993) . The positioning of served region of ORF 1a (Boursnell et al., 1987; Gorba- His, Cys, and Glu residues in the HRV-14 3Cpro has been lenya et al., 1989; Lee et al., 1991; Herold et al., 1993). shown to be very similar to that of cellular trypsin by MHV, HCV-229E, and IBV encode 3CLpro molecules with analysis of the HRV crystal structure (Matthews et al., apparent masses of 27, 34, and 35 kDa, respectively, 1994) . In contrast, analysis of the crystal structure of as determined by SDS-PAGE analysis (Lu et al., 1995;  hepatitis A virus 3Cpro indicates that Asp84 most likely Ziebuhr et al., 1995; Tibbles et al., 1996) . The classificadoes not participate directly in catalysis (Allaire et al., tion of the coronavirus proteinases as ''3C-like'' is sup-1994) . More directly relevant to MHV, it has been shown ported by mutagenesis studies of predicted catalytic cysthat mutagenesis at Glu residues in IBV 3CLpro does teine or histidine residues. We have demonstrated that not abolish activity of the expressed proteinase (Liu and mutations at His 41 or Cys 145 of MHV 3CLpro abolish pro- Brown, 1995) . teolytic activity. Similar results have been obtained for Several 3CLpro cleavage sites within the gene 1 3CLpro of IBV and HCV-229E, confirming the essential polyprotein recently have been defined for MHV, IBV, nature of these residues and demonstrating that His and and 229E. The experimentally confirmed coronavirus Cys residues are in positions similar to those of the picor-3CLpro cleavage sites have a leucine, isoleucine, or navirus 3C proteinases (Lu et al., 1995; Liu and Brown, valine at position P2, glutamine at position P1, and 1995; Ziebuhr et al., 1995; Tibbles et al., 1996) .

serine or alanine at position P1 (Liu and Brown, 1995; Lu et al., 1995 Lu et al., , 1996 Ziebuhr et al., 1995;  Tibbles et

The amino acid sequences of the coronaviruses were obtained from GenBank; MHV (X73559) (Bonilla et al., 1994) , IBV (M94356) (Boursnell et al., 1987) , HCV-229E (X69721) (Herold et al., 1993) , and TGEV (Z34093) (Eleouet et al., 1995) . Deduced amino acid sequences of coronavirus 3CLpro domains were compared using a pam 250 scoring matrix and a word size of 2 (MacVector 4.5.3, IBI-Kodak) (Fig. 2) . Numbering of MHV amino acid sequences was from the beginning of ORF 1a. Numbering of amino acid residues within MHV 3CLpro is based The linear schematic of the MHV genome shows the organization of Ser 1 (Lu et al., 1995) Site-directed mutagenesis of pGpro with bovine chymotrypsin (Gorbalenya and Koonin, 1993) , 3Cpro of human rhinovirus 14 (Matthews et al., 1994) , and hepatitis A virus Asp53 and Asp65 were mutagenized by the Chame-3Cpro (Allaire et al., 1994) . Black bars reflect the number of amino acid leon double-stranded, site-directed mutagenesis kit per residues in the proteinase. The proteins are aligned at the confirmed the manufacturer's instruction (Stratagene). Two primers Cys or Ser catalytic residues to show the carboxy-terminal extent of the protein. Catalytic Asp residues of HRV-14 and chymotrypsin have were simultaneously annealed to the template. One sebeen confirmed by crystallography. Asp53 and Asp65 of MHV are prelection primer changed one nonessential unique restricdicted catalytic residues. LQS indicates cleavage sites at amino termition site AlwNI on the pGpro vector to a new restriction nus (confirmed) and carboxy terminus (putative) of 3CLpro.

site. The other primer encoded for a specific mutation. After annealing and extension, all new plasmid DNA was incubated with the restriction enzyme AlwNI. The dialanine at P2 and glycine at P1 (Lee et al., 1991) . The gested mixtures were then transformed into repair-defiprecise determinants of 3CLpro cleavage site seleccient XLmutS cells, and the resultant colonies were isotion remain to be determined. Finally, comparison of lated. The mutant plasmids were then purified and the coronavirus 3CLpro sequences with those of other digested with AlwNI again, and the resultant DNA digesproteinases suggests that they may differ from other tion was transformed into XL1-Blue cells. All specific muviral and cellular proteinases in their size and structations were confirmed by bidirectional sequencing ture. The coronavirus 3CLpro domains contain signifi-(Sequenase II, U.S. Biochemicals, per the manufacturer's cantly more amino acids downstream from the putative instructions). substrate binding site than other 3C or 3C-like proteinases ( Fig. 1 ) (Gorbalenya and Koonin, 1993) . The role

Constructs expressing full-length and truncated of this additional region of polypeptide in structure or versions of 3CLpro activity of the coronavirus proteinases is not known.

In this study, we demonstrate that aspartic acid resi-The Ser3334 to Gln 3607(pG-S/FQ) and Ser3334 to dues in MHV 3CLpro are not required for catalytic activity.

Gln3635(pG-S/LQ) fragments were obtained by PCR am-In addition, we show that the MHV 3CLpro is able to plification of the region between nt 10212 and 11034 and cleave at a Gln-Gly cleavage site upstream from the prethat between nt 10212 and 11117, respectively (Fig. 5 ). dicted carboxy terminal Gln-Ser cleavage site and within

The pG-S/FQ left primer with an added on XbaI restricthe proteinase itself. The 3CLpro extending from the contion site (5-ATTCTAGATGTCTGGTATAGTGAAGATGG firmed amino-terminal serine to the internal glutamine/ TGTCG-3) and pG-S/FQ right primer with an added on glycine cleavage site is inactive in vitro. In contrast, the HindIII restriction site (5-TAAAATAAGCTT TCACTGGAA-''full-length'' protein extending to the predicted glutamine/ TCCAGAATGCAGCCT-3) were used to prime DNA synserine cleavage site is identical in size to 3CLpro exthesis from the pGpro construct. The PCR products were pressed in vitro and in virus-infected cells and is an digested by XbaI and HindIII for 2 hr and then run on an active proteinase. Thus, it appears that the entire pre-0.8% low-melting-point agarose gel. The product band was excised and ligated into the EcoRI and HindIII sites dicted coding region is required for 3CLpro activity. 3. with a pam 250 scoring matrix. The MHV-A59 His41 and Cys145 residues, and the corresponding residues of IBV, TGEV, and 229E are shown in boldface letters. The locations of aspartic acid residues (Asp53 and Asp65) of MHV-A59 are shown by asterisks. Other conserved asparagine (N95) and aspartic/glutamic acid residues (D110) are indicated by a dot. Residues predicted to be involved in substrate binding (Thr135 and His163) are indicated by a diamond. The solid arrowhead indicates the experimentally confirmed amino-terminal cleavage site of the MHV and 229E 3CLpro (Lu et al., 1995; Ziebuhr et al., 1995) . The open arrowhead indicates the predicted carboxyl terminal LQ_S/A cleavage sites of the proteinases (Gorbalenya et al., 1989; Lee et al., 1991; Gorbalenya and Koonin, 1993) . The location of the FQ_G sequence in MHV is indicated as an underlined arrowhead. Numbering of MHV His41, Cys145, Asp53, Asp65, Asn95, and Glu110 residues is based on identifying Ser 3334 of the ORF 1a polyprotein as Ser1 of 3CLpro. MHV-A59 amino acid numbers were derived from the submitted nucleotide sequence of Bonilla, et al. (1994) . ''iisvkes'' is a seven-amino-acid region present only in the IBV sequence. of pGEM-3Zf(0) (Promega) behind the T7 promoter which

In vitro transcription and translation constituted pG-S/FQ. pGopt-S/FQ was similarly con-Recombinant plasmids were transcribed and transstructed using a left primer with an optimal ATG (5lated using a coupled in vitro transcription/translation GGGCGAATTCGCCACCATGAGTGGTATAGTGAAGAT rabbit reticulocyte lysate system (TnT, Promega), as pre-GGTGTCG-3). pC-S/FQ was constructed by using a left viously described (Lu et al., 1995 (Lu et al., , 1996 . Approximately primer with an added NcoI restriction site (5-TCATCC-0.5 mg of plasmid DNA was incubated at 30Њ with 12.5 ATGGCCTCTGGTATAGTGAAGAT G-3) and a right ml TnT lysate, 1 ml TnT reaction buffer, 0.5 ml T7 RNA primer with an added EcoRI restriction site (5-AATTTpolymerase, 20 units RNasin, 0.5 ml 1 mM methionine-GAATTCACTGGAATCCAGAATGCAGCCT-3). The fragfree amino acid mixture, and 20 mCi [ 35 S]methionine in ment was then subcloned into pCITE (Novagen). pC-S/ a final volume of 25 ml. Samples were taken at various LQ was similarly constructed by using a left primer with time points and electrophoresed on an SDS 5-18% gradian added NcoI restriction site (5-TCATCCATGGCCTCTent polyacrylamide gel (SDS-PAGE). GGTATAGTGAAGATG-3) and a right primer with an added EcoRI restriction site (5-TGTGCG AATTCACTG-Trans cleavage assay TAGCTTGACACCAGCTA-3). The fragment was then subcloned into the NcoI and EcoRI sites of pCITE (Nova-Inactive site-directed mutants of pGpro (pGproH41G or pGproH41Q) were translated in the presence of [ 35 S]-gen) behind the T7 promoter. methionine. The parental pGpro construct, plasmids en-Asp 3386 (D53) was conserved as either Asp or Glu among the four viruses. Glu 3443 (E110) of MHV was conserved coding the predicted full-length 3CLpro (pC-S/LQ), and plasmids encoding the truncated forms of 3CLpro (pG-as Glu or Asp, and Asn 3428 (N95) of MHV was identical in all four coronavirus sequences; however, the location S/FQ, pGopt-S/FQ, and pC-S/FQ) were transcribed and translated in the presence of nonradiolabeled L-methio-of these residues relative to the essential His and Cys residues makes them less appealing as potential cata-nine. After 40 min, transcription and translation were terminated by the addition of RNase (10 mg/ml) and cyclo-lytic residues. Comparison of the coronavirus 3CLpro amino acid se-heximide (5 mg/ml) for 5 min. Following termination of transcription and translation, labeled mutant and unla-quences with chymotrypsin confirmed the additional amino acids between the putative substrate binding resi-beled 3CLpro reaction lysates were mixed 1:1 and incubated for an additional 135 min. The reaction mixtures due His 3496 (H163) and the probable carboxy-terminal QS 3635-6 cleavage site of 3CLpro. The comparison of the were checked for residual expression and processing from the pGpro construct by the addition of [ 35 S]-four coronavirus 3CLpro sequences revealed two potential cleavage sites present only in MHV, QS 3554-5 , and methionine to an aliquot of the unlabeled reaction mixture after treatment with RNase and cycloheximide and QG 3607-8 . Overall, the comparison of the coronavirus sequences indicated that there was variation among the incubation for an additional 135 min. All products were analyzed by electrophoresis by SDS gradient PAGE, fol-proteinases in the location of potential catalytic residues and cleavage sites. lowed by fluorography.

Mutagenesis of aspartic acid residues

In vitro transcription and translation were performed Based on the analysis of the protein alignments, we in a total volume of 200 ml with 8.0 mg pGpro DNA in the chose Asp53 and Asp65 residues for mutagenesis studpresence of 160 mCi [ 35 S]methionine (DuPont NEN), 400

ies. Asp65 has been considered the most likely candi-mCi [ 3 H]valine (Amersham), or 160 mCi [ 3 H]leucine (Amerdate for a third residue to be involved in catalysis. Asp53 sham) for 120 min at 30Њ. The products were separated was in a less favorable position relative to the His, but on 5-18% gradient polyacrylamide gels, transferred to a was conserved among the coronaviruses and provided polyvinylidene difluoride (PVDF) membrane at 50 V at 4Њ a good control. In addition, studies of other viruses have for 6 hr in transfer buffer containing 25 mM Tris-base, demonstrated that deviation from predictions of active 192 mM glycine, and 10% (v/v) methanol. After transfer, residues is not uncommon. The construct used for these the PVDF membrane was air dried and exposed to X-ray studies (pGpro) encoded amino acids 3239-3687 of film. Radiolabeled proteins were identified by autoradiog-MHV gene 1, including 3CLpro (3334-3635) and portions raphy, and the corresponding bands were excised from of the flanking domains (Fig. 3A) . We have previously the PVDF membrane and subjected to amino-terminal shown that translation of pGpro in vitro results in a presequencing on an ABI 470 sequencer. The amino acid cursor polypeptide from which active 3CLpro is autoprofraction from each cycle was quantitated in a Beckman teolytically cleaved and that 3CLpro has an apparent scintillation counter. mass of 27-29 kDa (p27) following SDS-PAGE (Lu et al., 1995) . Liberation of p27 3CLpro was therefore used RESULTS as a marker of proteolytic activity of proteins expressed from different constructs in vitro.

The wild-type proteinase construct (pGpro) and mutant proteinase domains proteinase constructs were transcribed and translated in a rabbit reticulocyte lysate (Fig. 3B) . The proteinase Predictions of catalytic residues of the coronavirus 3Clike proteinases have not strongly predicted aspartic or expressed from pGpro was able to process p27 3CLpro (Fig. 3B, lane 1) , whereas the proteinase with the His41 glutamic acid residues. Comparison of the deduced amino acid sequences of 3CLpro from the coronaviruses to Gln mutation (H41Q) did not cleave p27 (Fig. 3B , lane 2). Mutation of Asp65 to Pro or Ala (D65P and D65A) MHV-A59, IBV, HCV-229E, and TGEV revealed no completely conserved Asp or Glu residues at positions analo-resulted in a proteinase with activity comparable to that expressed from wild-type pGpro (Fig. 3B, lanes 3 and 4) . gous to catalytic Asp or Glu residues of other 3C or 3Clike proteinases (Fig. 2) . There was relative conservation Substitution of Asp53 by Glu (D53E) did not affect 3CLpro activity (Fig. 3B, lane 5) , whereas the substitution of of Asp, Glu, or Asn among the coronaviruses at the residue analogous to Asp 3398 (D65) of MHV. It has been Asp53 by Pro (D53P) impaired processing of p27 approximately 70% relative to pGpro (Fig. 3B, lane 6) . The D53P shown that the analogous residues within the IBV 3CLpro at Asp 2841 or Asp 2843 (D62 or D64) are not required for change might be expected to cause a change in the proteinase structure with a concomitant alteration of ac-proteolytic activity (Liu and Brown, 1995). The MHV pGpro in the presence of leupeptin blocked cleavage of p27 and also completely blocked processing of the small polypeptide fragment (Fig. 4, lanes 4-5) . The small cleavage fragment indicated by the arrow was consistently seen when pGpro or proteolytically active mutants were translated, but not when proteolytically inactive mutants were expressed (Fig. 4B) . The cleavage fragment was detected following translation of pGpro, H127Q, H127M, and C142R (Fig. 4B, lanes 1, 2, 3, and 4, respectively) , all of which also processed p27. In contrast, no small fragment was seen after translation of C145G or H41Q (Fig. 4B, lanes 5 and 6) , both of which are inactive in p27 processing. Together these results indicated that this cleavage fragment was processed by products expressed from the proteinase constructs in vitro, rather than by proteinases in the reticulocyte lysate.

The smallest proteolytic fragment was used for amino terminus radiosequencing since it was the most discrete and abundant. The pGpro construct was transcribed and translated in the presence of [ 3 H]leucine, the peaks of radioactivity were consistent with leucine and Asp65 were expressed in a combined transcription and translation at residues 5 and 10, cysteine at residue 8, and valine lysate as previously described (Lu et al., 1995) . Samples were taken at residue 9. The only cleavage site within the pGpro at 120 min for analysis by 5-18% SDS gradient PAGE. The wild-type pGpro construct and the His41 to Gln mutant (H41Q) were used as expression product that could result in a product with controls. D65A refers to an Ala substitution at Asp53; other constructs this pattern was Gln-Gly 3607-8 . are similarly labeled. Mass markers are to the right of the gel and the The QG 3607-8 was not conserved in any of the other location of p27 is shown to the left of the gel. Processing of p27 coronaviruses and previously had not been predicted as by 3CLpro expressed from pGpro was considered as 100%, and the a cleavage site for 3CLpro. Radiosequencing with three percentage of proteinase activity of each expressed protein is shown beneath the lane markers. different amino acids confirmed specific cleavage between glutamine 3607 and glycine 3608 by the in vitro translated proteinase. We could not define the presumed cartivity; however, the D53P substitution diminished but did boxy-terminal fragment containing the predicted QS 3635-6 not abolish activity, indicating that D53 was not an indiscleavage site, possibly due to the compression of propensable residue. It was interesting that the D65P teins in this region (5.8 kDa) of the gel by the nonlabeled change did not diminish proteolytic activity, suggesting globin protein from the lysate. Additionally, the protein that even a major change at this location was inconsefrom the construct may have been targeted for rapid degquential for liberation of p27 3CLpro in the in vitro system. radation. We also did not detect any prominent alterna-Deletion of Asp53 (D53del) resulted in complete loss of tive form of p27 3CLpro. Since we do know the order of proteinase activity (lane 7). This was not surprising since cleavages or pattern of precursors expressed from pGpro such a change might be expected to significantly alter it has not been possible to determine when the FQ/G the structure of the proteinase. These experiments demsite is cleaved. Direct comparison of these cleavage sites onstrated that neither Asp53 nor Asp65 was directly inwill require constructs expressing single cleavage sites volved in catalysis with His41 or Cys145.

to determine specificity.

Identification of a 3CLpro cleavage site Truncation of 3CLpro and trans cleavage activity During in vitro translation of pGpro several proteins in vitro with apparent masses of less than 14.3 kDa were seen along with p27 (Fig. 4) . The pulse-label expression (Fig. Since 3CLpro was able to cleave upstream of the predicted QS 3635-6 cleavage site, we determined whether 4A) showed that the smallest of these polypeptides appeared concurrently with p27 3CLpro but then decreased Ser 3334 to Gln 3607 was the entire coding region for the active p27 3CLpro protein detected in virus-infected cells over a 4-hr period (Fig. 4, lanes 1-3) . Translation of and during in vitro translation of pGpro (Fig. 5) . The amino proteins by incubating the nonradiolabeled translation products of these constructs with radiolabeled sub-acid sequence extending from Ser 3334 to Gln 3635 would predict a protein of with a calculated mass of 33 kDa, strate expressed from the inactive proteinase mutant pGproH41G (His to Gly) (Fig. 5B, lane 6) . The ''full-whereas cleavage at Gln 3607 would predict a protein of 30 kDa in mass, somewhat closer in size to the apparent length'' 27-kDa 3CLpro expressed from the Ser 3334 -Gln 3635 construct was able to cleave the pGproH41G mass of p27 (Fig. 5A ). We constructed a panel of plasmids containing cDNAs encoding amino acids from expressed protein in trans (Fig. 5B, lane 8) , whereas the 24-kDa truncated 3CLpro expressed from the S 3334 -Gln 3607 or S 3334 -Gln 3635 (Fig. 5A) . The cDNAs were expressed in a variety of plasmids, using either the first Ser 3334 -Gln 3607 constructs did not process the mutant protein (Fig. 5B , lanes 9, 10, and 11). This result demon-natural AUG (pG-S/FQ) or an optimized AUG before Ser 3334 (pGopt-S/FQ). We also used vectors containing strated that a 28-amino-acid carboxy terminal truncation of 3CLpro abolished proteolytic activity. EMCV IRES elements to ensure that translation initiated before the Ser 3334 (pC-S/FQ and pC-S/LQ).

The constructs were used to direct translation in vitro DISCUSSION and the proteins either were radiolabled or were translated in nonlabeled medium and used in a trans cleavage MHV 3CLpro is postulated to mediate the majority of assay of the inactive mutant pGproH41G (Fig. 5B ). Transcleavages in the gene 1 polyprotein during virus replicalation of pC-S/LQ, encoding Ser 3334 to Gln 3635 , resulted in tion. We have shown that aspartic acid residues of MHV a single 27-kDa protein, the same migration pattern as 3CLpro in locations analogous to essential Asp/Glu resi-p27 3CLpro detected after expression of pGpro (Fig. 5B , dues of other 3C and 3C-like proteinases are not neceslanes 1 and 2). In contrast, translation of three different sary for processing of substrate by 3CLpro in vitro. Reconstructs encoding the truncated 3CLpro domain from sults similar to ours have been reported for infectious Ser 3334 to Gln 3607 resulted in a single 24-kDa protein (Fig. bronchitis virus (IBV) (Liu and Brown, 1995) . Our study 5B, lanes 3, 4, and 5). These data indicated that proteins demonstrates that conservation of Asp/Glu in this region expressed from the 3CLpro domain differed in their calof the coronavirus 3C-like proteinases is not due to an culated and apparent masses by 6 to 10 kDa. The results indispensable catalytic role. The mechanism of the MHV also supported the conclusion that the active 3CLpro in 3CLpro may be more similar to that of hepatitis A virus, MHV-infected cells and from in vitro translation products in which the Asp is on an external motif and not directly incorporated Ser 3334 to Gln 3635 .

involved in the catalytic unit (Allaire et al., 1994) . It is We assessed the in vitro cleavage activity of the possible that this variation in the use of a third residue may have coevolved with the specificity for cleavage 24-kDa Ser 3334 -Gln 3607 and the 27-kDa Ser 3334 -Gln 3635 sites. There is a precedent in other virus systems for a consensus substrate binding residues, whereas the MHV sequence extends an additional 137 residues to its contribution of Asp residues to proteinase specificity even though they may not be involved in catalysis. For carboxy terminus (Lee et al., 1991) . Our study demonstrates that a small deletion of this part of the proteinase example, poliovirus contains an FRD (D85) sequence that was initially thought to be involved in catalytic activity abolishes its ability to cleave new molecules of p27 3CLpro in trans, demonstrating that the entire carboxy-but subsequently was found to be in a flanking turn domain and to be involved in autocatalytic cleavage of 3CD terminal region is essential for 3CLpro activity. Analysis of confirmed and predicted cleavage sites in (Hammerle et al., 1992) . Despite the lack of use of an Asp residue, the MHV 3CLpro should still be classified the gene 1 polyproteins of MHV-A59, HCV-229E, IBV, and TGEV has revealed a preference for a Gln at P1 and Leu, as a chymotrypsin-like enzyme because of the localization of histidine and cysteine residues as well as flanking Ile, Val, or less often Phe or Met at P2 (Boursnell et al., 1987; Breedenbeek et al., 1990; Lee et al., 1991 ; Herold residues considered to be important in protein structure and substrate binding (Gorbalenya and Koonin, 1993). et al., 1993; Bonilla et al., 1994; Eleouet et al., 1995) . Although the Phe-Gln-Gly (FQ/G) 3CLpro cleavage site Analysis of the full-length 3CLpro domain (Ser 3334 to Gln 3635 ) reveals several possible differences between the we identified within 3CLpro has similarities to other predicted 3CLpro sites from P4 to P1, it is not present in MHV 3CLpro and other viral proteinases in the group of cysteine-containing enzymes (Fig. 1) . First, most of these the other sequenced coronaviruses. We have not determined if the FQ/G 3608 site can be cleaved in virus-infected enzymes terminate within 30 amino acids following the

Intracellular and in vitro translated 27-kDa proteins contain the 3C-like proteinase activity of the

Picornaviral 3C cysteine proteinases have a fold similar to chymocoronavirus MHV-A59

Identification and characterizatrypsin-like serine proteinases

Mouse hepatitis tion of a serine-like proteinase of the murine coronavirus MHV-A59

Effect of expression of the aphthovirus protease 3C on viral infection and gene expression

Completion of the sequence of the Virology

genome of the coronavirus avian infectious bronchitis virus

The primary structure and ase reveals a trypsin-like polypeptide fold, RNA-binding site, and means for cleaving precursor polyprotein

Characterization of a human coronavirus (strain 229E) 3C-like proteinase activity

Eleouet, J. F., Rasschaert, D., Lambert, P., Levy, L., Vende, P., and Laude, cells. It is possible that the fragment of gene 1 used in H. (1995) . Complete sequence (20 kilobases) of the polyprotein-enthese studies allows presentation of this site in a manner G cleavage site is used by the proteinase in cells, it Gorbalenya, A., and Koonin, E. (1993) . Comparative analysis of aminoacid sequences of key enzymes of replication and expression of might represent a pathway for regulation of the protein- In conclusion, we have identified several unique fea- ase gene polyprotein processing and virus replication. Lawson, M. R., and Semler, B. L. (1992) . Alternate poliovirus nonstructural protein processing cascades generated by primary sites of 3C