Vol. 66, No. 12

JOURNAL OF VIROLOGY, Dec. 1992, p. 7040-7048

0022-538X/92/127040-09$02.00/0 Copyright X 1992, American Society for Microbiology

The 5' End of the Equine Arteritis Virus Replicase Gene Encodes a Papainlike Cysteine Protease ERIC J. SNIJDER,* ALFRED L. M. WASSENAAR, AND WILLY J. M. SPAAN Department of Virology, Institute of Medical Microbiology, Faculty of Medicine, Leiden University, Postbus 320, 2300 AH Leiden, The Netherlands Received 24 July 1992/Accepted 6 September 1992

The presence of a papainlike cysteine protease (PCP) domain in the N-terminal region of the equine arteritis virus (EAV) replicase, which had been postulated on the basis of limited sequence similarities with cellular and viral thiol proteases, was confirmed by in vitro translation and mutagenesis studies. The EAV protease was found to direct an autoproteolytic cleavage at its C terminus which leads to the production of an approximately 30-kDa N-terminal replicase product (nspl) containing the PCP domain. Amino acid residues Cys-164 and His-230 of the EAV replicase polyprotein were identified as the most likely candidates for the role of PCP catalytic residues. By means of N-terminal sequence analysis of a PCP cleavage product, derived from a bacterial expression system, it was shown that cleavage occurs between Gly-260 and Gly-261. No evidence for PCP-directed cleavages at other positions in the EAV replicase was obtained. In cotranslational and posttranslational trans-cleavage assays, neither EAV nspl nor its precursor was able to process the PCP cleavage site in trans.

Equine arteritis virus (EAV) is an enveloped positivestranded RNA virus. Its isometric nucleocapsid core contains a nonsegmented 12.7-kb genome (for a recent review, see reference 32). The morphological characteristics and genome size of EAV are most comparable to those of togaviruses. However, we have recently reported that the EAV replication strategy (14, 18) is similar to that of coronaviruses (for a review, see reference 37) and toroviruses (35, 36), which possess 25- to 31-kb positive-stranded RNA genomes. Among their common features are a polycistronic genome organization, the same basic gene order (5'-replicase gene-envelope protein genes-nucleocapsid protein gene-3'), and the production of a 3'-coterminal nested set of four to seven subgenomic mRNAs, which is used to express the genes in the 3' one-third of the genome. The 5' part of the genomes of these viruses is occupied by two large open reading frames (ORFla and ORFlb) which encode the viral replicase (5, 6, 14, 27, 35). Both ORFla and ORFlb are expressed from the genomic RNA, the latter by means of a ribosomal frameshifting mechanism (8, 14, 35). The predicted ORFlb products of coronaviruses, toroviruses, and arteriviruses contain a number of homologous protein domains (6, 14, 35) which, in addition to the other similarities described above, indicate that these viruses are evolutionarily related. We have therefore proposed these viruses to be members of a coronaviruslike superfamily of positive-stranded RNA viruses (14, 35). Lactate dehydrogenase-elevating virus (LDV), which is structurally similar to EAV (32), is the latest addition to this superfamily; its genome organization, replication strategy, and replicase gene were reported to be coronaviruslike (19, 26). The complete replicase gene sequences of the coronaviruses infectious bronchitis virus (IBV; 5) and mouse hepatitis virus (MHV; 7, 27) and of the arterivirus EAV (14) have been reported. Amino acid sequence comparison has revealed that the ORFla-encoded products of these viruses are much more diverged than the corresponding ORFlb pro*

teins. The latter include the polymerase and helicase motifs common to all positive-stranded RNA viruses. The large replicase gene product (345 kDa for EAV, 741 kDa for IBV, and 810 kDa for MHV) is presumed to be a polyprotein precursor which is posttranslationally cleaved into smaller functional units. In the EAV ORFla amino acid sequence, a putative trypsinlike serine protease has been identified (14). At the same relative position, the IBV (21) and MHV (27) sequences contain a picornavirus 3C-like cysteine protease domain. Both trypsinlike and 3C-like proteases are assumed to belong to the same protease superfamily (3, 20). Additional protease domains, related to the papainlike cysteine protease (PCP) family, were proposed to be located more upstream in the ORFla products of EAV (14) and coronaviruses (27), albeit in different relative positions in the sequence. We have now initiated a study of the coronaviruslike replicase, with the relatively small EAV replicase gene as a model. In this article, we report the cDNA reconstruction and in vitro translation of EAV ORFla. The presence of a PCP domain in the N-terminal part of the EAV ORFla protein was confirmed experimentally. Both its pair of

putative catalytic residues and the residues flanking the PCP cleavage site were identified. MATERIALS AND METHODS General methods, plasmids, enzymes, and sequences. In general, recombinant plasmids were constructed by using standard techniques and procedures (28). Unless stated otherwise, plasmid pBS- (Stratagene) served as a basis for the transcription vectors described below. An XhoI linker (5' CCTCGAGG 3') was inserted into the unique HindIII site of pBS- to create pBSHX. DNA-modifying enzymes from Pharmacia, Promega, or New England Biolabs were used according to the manufacturers' instructions. Nucleotide and amino acid sequence numbers refer to the EAV genomic and protein sequences which we have published previously (14). Restriction sites are indicated by the number of the first nucleotide of their recognition sequences.

Corresponding author. 7040

VOL. 66, 1992

EAV CYSTEINE PROTEASE

7041

TABLE 1. Mutagenesis of the EAV PCP domain and cleavage site Construct

Position of oligonucleotide lin EAV sequence

pCPl pCP2 pCP5 pCP6 pCP7

706-725 706-725 869-892 869-892 869-892 903-926 903-926 903-926 995-1016 995-1016 998-1019

pCP8 pCP9 pCP1O pCPll pCP12 pCP13 pCP14

998-1019

5'

GATGGGTTC GATGGGTTC GGCTGAAATTC GGCTGAAATTC GGCTGAAATTC GOCGAGOCTTGG GCGAGCTTGG GCGAGCTTGG GCAACTAC GCAACTAC ACTACGGC ACTACGGC

In vitro transcription and translation. Plasmid DNA for in vitro transcription was prepared by the alkaline lysis method (28). All constructs were linearized with XhoI. After proteinase K treatment and phenol extraction, the DNA was precipitated with isopropanol and dissolved in water. In vitro transcription was performed by incubating 1 ,ug of linearized plasmid DNA at 37°C in a 20-pl system containing 40 mM Tris-Cl (pH 8.0), 50 mM NaCl, 8 mM MgCl2, 10 mM dithiothreitol, 2 mM spermidine, 0.5 mM (each) ATP, CTP, and UTP, 0.025 mM GTP, 0.25 mM 7mGpppG (cap analog), 100 pg of bovine serum albumin per ml, 1,000 U of RNAsin per ml, and 750 U of T7 RNA polymerase per ml. After 1 h, transcription mixtures were diluted 10-fold in water. One microliter of this RNA preparation was used in a standard 12.5-,ul in vitro translation reaction mixture containing 10 p,l of nuclease-treated methionine-free rabbit reticulocyte lysate (RRL) (Promega), 0.2 ,ul of the amino acid mixture minus methionine supplied by the manufacturer, and 1.3 ,ul of [35S]methionine (New England Nuclear; 10 p,Ci/pl). In vitro translation was carried out for 1 h at 30°C. For direct analysis, the reactions were terminated by adding sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE) sample buffer. trans-cleavage assays. Cotranslational trans cleavage was tested by mixing and translating equal amounts of RNA transcripts. For posttranslational trans-cleavage experiments, labeled substrates were prepared as described above. The proteins to be tested for trans-acting proteolytic activity were synthesized in translation mixtures in which [35S]methionine had been replaced by unlabeled methionine (final concentration, 20 p,M). In vitro synthesis of substrates and enzymes was stopped by adding unlabeled methionine, RNase A, and cycloheximide to final concentrations of 0.5 mM, 10 ng/pl, and 1 p,g/,ul, respectively. Subsequently, substrate and protease preparations were mixed at a 1/1 (vol/vol) ratio and were incubated for an additional 2 to 7 h at 30°C. SDS-PAGE, autoradiography, and densitometry. Translation products were analyzed by SDS-PAGE in SDS-12.5% polyacrylamide gels. Gels were fixed in 40% methanol-10% acetic acid, soaked in 1 M sodium salicylate, dried, and autoradiographed. Densitometry was performed with an LKB Ultroscan-XL laser densitometer. Mutagenesis. Single- or double-base changes were introduced into the PCP-coding region by oligonucleotide-directed mutagenesis as described by Kunkel et al. (25). Mutagenesis was carried out on single-stranded M13 mpl8 or mpl9 DNA which contained restriction fragments from

Sequence 3'

AGC GGC GCC GC GTC GOT GOT GTT GOCC GTA GCC

TGGTTAAA TGGTTAAA

ATTATGAGG ATTATGAGG ATTATGAGGG

ATCACAACACOG ATCACAACACG ATCACAACACG GGCTACAATCC GGCTACAATCC TACAATCCACC

GTA TACAATCCACC

Amino acid substitution

Cys-164- Ser Cys-164--*Gly His-219- >Ala

His-219-*Gly His-219-Val His-230-3Ala

His-230-.Gly

His-230-Val Gly-260-)Ala

Gly-260-Nal

Gly-261--3Ala Gly-261- Val

EAV cDNA clones 586 and 673 (14). M13 mutants were selected by dideoxynucleotide sequencing of single-stranded DNA. Restriction fragments containing the desired mutations were cut from M13 replicative-form DNA and ligated into plaA2-derived vectors (see below). Oligonucleotides and mutations are listed in Table 1. Construction of pEAVla. A cDNA copy of EAV ORFla was reconstructed downstream of the pBS- T7 promoter (pEAVla; Fig. 1A). The generation and sequence analysis of the seven cDNA clones used for this purpose have been described previously (14). After polishing of the sticky ends, the ApaI site (nucleotide [nt] 69) upstream of the ORFla ATG codon (nt 226 to 228) was ligated to the EcoRI site of pBSHX. An XhoI linker (5' CCTCGAGG 3') was inserted into EAV cDNA clone 579 at the SmaI site (nt 5512), which is located in the 5' end of ORFlb. This XhoI site was used to ligate the 3' end of the reconstructed gene to the XhoI site of pBSHX and for linearization prior to in vitro transcription. At the RNA level, this linker insertion destroyed the RNA pseudoknot structure which directs EAV ORFla-ORFlb ribosomal frameshifting (14), thereby preventing frameshiftrelated complications during in vitro translation studies. The following restriction fragments were used to create pEAVla: ApaI (nt 69) to SacI (nt 858) from clone 586; SacI (nt 858) to Kjpnl (nt 1802) from clone 673; KIpnI (nt 1802) to PstI (nt 2455) from clone 696; PstI (nt 2455) to N/teI (nt 2878) from clone 428; N/teI (nt 2878) to EcoRV (nt 4263) from clone 694; EcoRV (nt 4263) to BamHI (nt 5115) from clone 659; and BamHI (nt 5115) to SmaI (nt 5512, containing the XhoI linker) from clone 579. Generation of pEAVia-derived 3'-truncated constructs. To create termination codons for translation at various positions in ORFla, N/teI linkers (5' CTAGCTAGCTAG 3') were inserted into the following pEAVla restriction sites: Sacl (nt

858), HindIII (nt 1501), KpnI (nt 1802), SalI (nt 2608), N/teI (nt 2878), ApaI (nt 3688), EcoRV (nt 4263), and BamHI (nt 5115). Using the same set of restriction sites, an additional X/toI linker was inserted into these plasmids; for each construct, the restriction site immediately downstream of the one containing the NheI linker was used. The untranslated region between the two X/toI sites was deleted byX/AoI digestion and religation. This prevented incidental readthrough translation of sequences far downstream of the inserted termination codons, which could interfere with the analysis of ORFla processing. The resulting deletion constructs were named plaAl through plaA8 (Fig. 1A). In a similar manner, construct plaAP was created, which con-

7042

J. VIROL.

SNIJDER ET AL.

tained the NheI and XhoI linkers in the PstI (nt 2455) and Sall (nt 2608) restriction sites, respectively. Constructs containing PCP and cleavage site mutations. Transcription vector plaA2 (Fig. 1A and 3) was used to test the effect of amino acid substitutions in the EAV PCP domain and its cleavage site. The mutations and the plaA2derived constructs in which they were tested are listed in Table 1. All mutations in the pCPn constructs were confirmed by sequence analysis of the plasmid DNA preparation which was used for in vitro transcription. In pCP14 (Gly261-WVal), oligonucleotide-directed mutagenesis resulted in an AccI restriction site which was used to generate an additional construct. The AccI site was digested, filled, and religated to introduce a 2-nt insertion (pCP14-1). Expression of the ORFla 5' end in Escherichia coli. The bacterial expression vector pGEX-2T (Pharmacia; 34) was used to express the N terminus of the EAV ORFla polypeptide as part of a bacterial fusion protein. Plasmid pGEAV1 was generated by cloning a Sau3A-Sau3A (nt 200 to 1342) fragment into the BamHI site of pGEX-2T. Construct pGEAVla contained the same Sau3A fragment followed by a Sau3A-PvuII fragment (nt 1342 to 1808 of EAV plus 192 bp of pUC18), cloned between the BamHI and filled EcoRI site of pGEX-2T. The 69-kDa GEAV1 fusion protein contained the standard pGEX-derived 26-kDa glutathione S-transferase (GST) moiety, followed by 9 amino acids (aa) encoded by the EAV 5' untranslated region, the 373 N-terminal aa of the ORFla product, and 8 aa encoded by the pGEX multiple cloning region. The 86-kDa GEAVla protein was of similar composition, but in this case the EAV ORFla part was larger (aa 1 to 528) and the 33 C-terminal aa of the protein were derived from a short in-frame vector sequence. Induction of fusion protein synthesis with IPTG (isopro-

pyl-p-D-thiogalactopyranoside)

and lysis of bacteria

were

performed essentially as described by Smith and Johnson (34). Soluble and insoluble bacterial components were separated by centrifugation. The insoluble GEAV fusion proteins were partially purified by extraction with 3 M urea in 10 mM Tris-Cl (pH 7.8)-i mM EDTA-100 mM NaCl. The GEAV proteins remained insoluble under these conditions. After an additional centrifugation step, the proteins from the remaining pellet were separated by SDS-PAGE (12.5% polyacrylamide) (see Fig. 4). Partial protein sequence analysis of the GEAVla protein. A sample containing the GEAVla protein and its two cleavage products was prepared as described above. After SDSPAGE, the proteins were electroblotted onto a polyvinylidene difluoride membrane (29), which was stained with 0.1% Coomassie brilliant blue R-250. The band representing the 32-kDa C-terminal part of the pGEAVla product was cut from the membrane. The N-terminal amino acid sequence of this protein was determined with a protein sequenator (Applied Biosystems model 475A) equipped with on-line phenylthiohydantoin (PTH) analysis using a model 120A PTH analyzer. RESULTS

Reconstruction and in vitro translation of EAV ORFla. Analysis of the posttranslational processing of the EAV replicase polyprotein was started by reconstructing ORFla from seven overlapping cDNA clones (see Materials and Methods). In transcription vector pEAVla, a full-length cDNA copy of ORFla is located downstream of the T7 RNA polymerase promoter. In addition, a set of pEAVla deletion mutants (plaAl through plaA8) was produced (Fig. 1A)

which contained termination codons for translation at various positions in ORFla. Vector pEAVla and the plaAn constructs were used to transcribe a set of 3'-truncated RNAs from which an increasing part of ORFla could be translated. The predicted sizes of the translation products encoded by these RNAs are listed in Fig. 1A. In vitro translation was carried out in an RRL in the presence of [35S]methionine. A direct SDS-PAGE analysis of the translation products of the plaAn series and pEAVla is shown in Fig. 1B. Only plaAl, which contained a termination codon in the center of the putative PCP domain, produced a protein of the predicted size. All other constructs gave rise to a prominent product of about 30 kDa and to accompanying bands which were smaller than predicted from the (partial) ORFla amino acid sequence. These data indicated that an approximately 30-kDa protein was cleaved from the N terminus of the EAV ORFla product. The putative PCP domain, which resided in this cleavage product, could be involved in this proteolytic event. Only small amounts of uncleaved protein were observed after a standard 1-h RRL translation experiment, indicating that cleavage was rapid and efficient under these conditions. Analysis and mutagenesis of the EAV PCP domain. Typical PCPs show a requirement for at least one cysteine and one histidine residue (31). The N-terminal sequence of the EAV ORFla polypeptide (14) is shown in Fig. 2A. On the basis of amino acid sequence comparison of the ORFla sequence and cellular and viral papainlike thiol proteases (14; Fig. 2B), Cys-164 and His-374 had been proposed as active site residues of an EAV PCP (14). However, in vitro translation data from additional ORFla constructs which lacked either the sequence upstream of an MluI site (nt 589) or the sequence downstream of a NarI site (nt 1063) indicated that both the PCP domain and its cleavage site were located between Ala-123 and Gly-280 (data not shown). This implied that His-374 could not be involved in proteolysis and that His-219 and His-230 were more likely candidates for this role. In view of the spacing between the catalytic Cys and His residues in other thiol proteases (Fig. 2B), His-184 was assumed to be too close to Cys-164. To prove that the EAV PCP domain was responsible for the observed proteolytic processing of the ORFla protein and to identify possible active site residues, single amino acid substitutions were introduced into the ORFla sequence. Transcription vector plaA2 (Fig. 1A and 3), encoding a 46-kDa product (seven methionine residues) which is cleaved into 30- and 16-kDa polypeptides (four and three methionine residues, respectively), was used as a basis for these experiments. Derivatives of this construct were used to test the influence of substitutions at the positions of Cys-164, His-219, and His-230 (Table 1). Mutations were tested by in vitro transcription and translation (Fig. 3). Both the rather conservative substitution of Cys-164 by Ser as well as the Cys-164 to Gly mutation completely abolished proteolytic activity. Even after prolonged autoradiography, the 16- and 30-kDa bands could not be detected, indicating that this Cys residue is indeed essential for the protease function. Replacing His-219 by Ala, Gly, or Val did not affect cleavage to a significant extent. In contrast, the same set of substitutions at the position of His-230 exhibited an effect similar to that observed after replacing Cys-164; the His-230 to Val substitution completely inhibited proteolytic activity, and only traces of cleavage products could be detected after translation of the Ala-230 and Gly-230 mutants. These results confirmed that the 5' region of the EAV

VOL. 66, 1992

EAV CYSTEINE PROTEASE

A)

B) pla

SIl

CP ORF)a

gL

1'eplicase C 1

ilb a

P.,- Sc 8t'

A>a -:R

STra

;a

T7

p1I aA I 1 aA3 p1 aA4

24K C>. * C} 46K 57K 86K n >

p1 aA5

96K

p?,- I it

A2

7043

E> p1 1a; A 6 125K E> p I A 7 145K p I A 8 177K E>> p E Ivvi a 187K

200K

Al

A3

A4

A5

lirAV I( A6 A7 A8 la RNA

-

97K 69K -

X

A^

0

*X x *X *

46K

-

30K

-

14K

-

*X

Ox

X

FIG. 1. (A) Schematic representation of the ORFla region of the EAV replicase gene. The positions of the predicted PCP (CP) and trypsinlike serine protease (SP) domains are shown. The indicated restriction sites were used to construct the expression plasmids plaAl through plaA8 and pEAVla, which are depicted in the lower part of the figure. The positions of the translation initiation codon (>), termination codon (*), and XhoI site (x) used for linearization prior to transcription are indicated for each construct. The open asterisk for pEAVla represents the normal ORFla termination codon. The sizes (in kilodaltons) of the predicted translation products were calculated on the basis of the ORFla amino acid sequence. (B) In vitro translation results from EAV ORFla expression constructs. Plasmids plaAl through plaA8 and pEAVla were used for in vitro transcription of a 3'-truncated set of RNAs from which an increasing part of ORFla could be translated. The RNAs were translated in an RRL in the presence of [35S]methionine, and direct analysis of translation products was performed by SDS-PAGE (12.5% gel). The rightmost lane shows the results from a translation to which no exogenous RNA was added.

encodes a proteolytic domain which is responsible for the observed cleavage event. Activity of the EAV PCP domain in E. coli. A number of 3'-truncated fragments from the N terminus of the EAV ORFla protein, which all contained the PCP domain, were expressed as part of bacterial fusion proteins. For this purpose plasmid pGEX was used, which expresses the foreign sequence as the C-terminal domain of a GST fusion protein. Construct pGEAVla, which is shown in Fig. 4, produced a fusion protein which contained about 60 kDa of the N-terminal ORFla protein sequence, including the PCP domain. Remarkably, pGEAVla and three pGEAVla-related 3'-truncated constructs did not only produce fusion proteins of the expected size, they all yielded an additional 55-kDa band. This is illustrated in Fig. 4 for pGEAVla and the somewhat smaller pGEAV1 construct. In view of the data obtained with the plaAn series (see above) and the 26-kDa size of the GST part of the pGEX fusion proteins, it was concluded that the EAV PCP domain was probably functional in E. coli. The 55-kDa protein represented the N-terminal cleavage product consisting of the GST part and the previously observed 30-kDa EAV ORFla product. However, the presence of substantial amounts of uncleaved fusion proteins indicated that the protease was less active than under in vitro translation conditions. The C-terminal cleavage product, which was of variable size, could not genome

always be detected, probably because of the presence of bacterial background bands. To confirm that the observed cleavage was directed by the EAV PCP domain, the Cys-164 to Ser/Gly, and His-230 to Ala/Gly/Val substitutions (see above) were introduced into one of the fusion proteins. As expected, these mutations abolished proteolytic activity; the 55-kDa product was no longer produced, and only uncleaved fusion proteins were detected (data not shown). N-terminal sequence analysis of the C-terminal GEAVla cleavage product. The amount of cleaved fusion protein produced by the pGEX system was sufficient to allow purification of one of the C-terminal cleavage products for N-terminal sequence analysis. For this purpose, the fusion protein derived from construct pGEAVla was selected; its 32-kDa C-terminal cleavage product was readily detected after SDS-PAGE, and no dominant bacterial background bands were present in this region of the gel (compare with the pGEAV1 lane in Fig. 4). After electroblotting, the 32-kDa band was cut from the polyvinylidene difluoride membrane and N-terminal microsequencing was carried out as described in Materials and Methods. The N terminus of the cleavage product was found to consist of the following amino acid residues: Gly-Tyr-Asn-Pro-Pro-Gly-Asp-GlyAla. This sequence is present at aa positions 261 to 269 in the EAV ORFla product, indicating that cleavage takes place

7044

J. VIROL.

SNIJDER ET AL.

A) 001 MATFSATGFGGSFVRDWSLDLPDACEHGAGLCCEVDGSTLCAECFRGCEGMEQCPGLFMG 061 LLKLASPVPVGHKFLIGWYRAAKVTGRYNFLELLQHPAFAQLRVVDARLAIEEASVFIST

121 DHASAKRFPGARFALTPVYANAWVVSPAANSLIVTTDQEQDGF

'-'

"

30K

WLKLLPPDRREAGL

ITTRSC 179 RLYYNHYREQRTGWLSKTGLRLWLGDLGLGINASSGGLKFHIMRGSPQRAW[H A A 237 KLKSYYVCDISEADWSCLPAGNYG [4] GYNPPGDGACGYRCLAFMNGATVVSAGCSSDLW

J

LA]

L..

295 CDDELAYRVFQLSPTFTVTIPGGRVCPNAKYAMICDKQHWRVKRAKGVGLCLDESCFRGI

16K

iSk230

IlS219

cv,S164

4

VimlndlIl

46K

plaA&2 >'"

Str (Gly .la CGIN Val Ala Clv Val

355 CNCQRMSGPPPAPVSAAVLDHILEAATFGNVRVVTPEGQPRPVPAPRVRPSANSSGDVKD A

415 PAPVPPVPKPRTKLATPNPTQAPIPAPRTRLQGASTQEPLASAGVASDSAPKWRVAKTVY

57K

-

40K

-

30K

-

IoK

-N_

B) Papain Bromelain Calpain

pvknQgscGsCWafsa - 126 - kvdHavaavgyn svknQnpcGaCWafaa - 123 - slnHavtaigyg pvknQgacGsCWtf st - 131 - kvnHavlavgyg dic-QgalGdCWllaa - 148 - vkgHaysvtapk

EAV nspl

vttdQeqdGfCWlkll

SIN nsp2 TEV HCpro HAV p29

Cath H

-

58 - rawHittrsckl

anpfscktnvCWakal

-

ekmyianeGyCYmnif

-

69 - pvaHwdnspgtr 65 - ktmHvldsygsr 45 - qcvHivagetfr

*

___m.

_

*

*

gslaQfgqGyCYlsai

_

cellular proteases

- a.-

_

* -

viral proteases

FIG. 2. (A) Amino acid sequence of the N-terminal part of the EAV ORFla product. The sequence analysis of cDNA clones representing this region of the EAV genome was reported previously (14). The ORF1a region which was found to contain both the functional PCP domain and its cleavage site is indicated (0 4; between Ala-123 and Gly-280). The putative PCP active site residues (Cys-164 and His-230) and the cleavage site (arrow between Gly-260 and Gly-261), which were identified in this report, are boxed. Alternative histidine residues are indicated (A). (B) Comparative amino acid sequence analysis of a selection of cellular and viral papainlike proteases. Putative active site cysteine and histidine residues are indicated with asterisks. The alignment of the cellular proteases was based on the data presented by Gorbalenya et al. (22). The Sindbis virus (SIN) nsp2 data were reported by Strauss and Strauss (38). The alignment of the tobacco etch virus (TEV) and hypovirulence-associated virus (HAV) proteases was taken from a publication by Koonin et al. (24).

between Gly-260 and Gly-261 in the EAV sequence (Fig. 2A). The exact molecular size of the resulting N-terminal EAV ORFla product (nonstructural protein 1 [nspl]) could now be calculated at 28.7 kDa. Mutagenesis of the PCP cleavage site. To study the sequence requirements of the PCP cleavage site, amino acid substitutions were introduced into the ORFla protein sequence at positions 260 (P1, according to the nomenclature introduced by Berger and Schechter [4]) and 261 (P1'). Both Gly residues were changed into either Ala or Val (see Table 1), and the effects of these mutations were analyzed in the same system that had been used to map possible active site residues (see above). Replacing Gly-260 with Ala inhibited cleavage only slightly (a reduction of about 5% compared with the plaA2 protein; Fig. 5). The same substitution at position 261 did not result in a significant change in proteolytic activity. The Gly to Val substitutions exerted a more drastic effect; the protein carrying the Val-260-Gly-261 dipeptide was not cleaved at all, whereas cleavage of the Gly-260-Val-261 bond was severely impaired (40 to 50% cleavage). The influence of the residues downstream of Tyr-262 was tested by introducing a frameshift mutation into construct pCP14, which produces the protein carrying the Gly-261 to Val substitution. A pCP14 AccI restriction site, which arose

a 1

_ma.s... 2

5

6 7 8

9 10 RNA

pCIP FIG. 3. Identification of possible active site residues of the EAV PCP. Single amino acid substitutions were introduced into construct plaA2, depicted at the top of the figure, and tested by in vitro translation and SDS-PAGE (12.5% gel). Mutations are indicated at the top of each lane; the name of the corresponding expression plasmid is shown at the bottom of the lane. The 46-kDa full-length translation product and its 30- and 16-kDa cleavage products are indicated. In some lanes, a faint 57-kDa band can be seen, which is interpreted as the product of read-through translation, terminating at the linearization site (an XhoI linker inserted at the KpnI site at nt 1802).

from site-directed mutagenesis, was digested and filled, leading to a 2-nt insertion (construct pCP14-1). As a result, translation downstream of Tyr-262 continues in the + 1 reading frame, terminating at a UGA codon 17 triplets downstream. In the 30.6-kDa pCP14-1 protein, the sequence downstream of Tyr-262 consisted of the following amino acid residues: Thr-Ile-His-Gln-Gly-Thr-Glu-Leu-Ala-Val-ThrGly-Ala-Trp-Pro-Ser. As shown in Fig. 5, the pCP14-1 product cleaved with similar efficiency as the pCP14 protein, even though none of the amino acid residues at the positions P3' through P18' was identical to those in the EAV ORFla sequence (Fig. 2A). The EAV PCP is a cis-acting protease. To test whether the EAV PCP can function in trans, the protease domain and a number of the mutants described above were used in co- and posttranslational trans-cleavage assays. To test posttranslational trans cleavage, PCP-containing proteins were prepared by in vitro translation with unlabeled methionine, whereas labeled (uncleaved) substrates were prepared by translation of mutant constructs in the presence of [3S]methionine. All translations were stopped by a mixture of cycloheximide, RNase A, and unlabeled methionine. Translation reactions which contained substrates were mixed with equal volumes of protease-containing reactions. The 46-kDa translation products derived from pCP1 and pCP10 (carrying the Cys-164--Ser and His-230--3Val substitutions, respectively; see also Fig. 3) were used as sub-

VOL. 66, 1992

EAV CYSTEINE PROTEASE

pGEAV

pGEAV

1

la

fusion protein GEAV I a PcP

86K -

Th.ZUI 26K

..

69K -_

69K liP _ 55K-

O

posttranslational

assay

assay

subst.

pCPl

60K

enz.

U 1

hI.p.t.

86K

]L

32K

_2

51K-

46K

pCPI

pCPIO

F1a |

expression in E. coli W

55K~~~~~~~~ -.'''P.1

cotranslational

CU 1 1

C-

-

1

8

*

_

1

-&*Ie_iS

U 8

pCP

PCPIO C 8

-

-

1

8

U 8

7045

C 8

14

1

8

_.

55K K

N-terminal

30K -

_

sequencing Gly-Tyr-Asn-Pro-ProGly-Asp-Gly-Ala

32K -

FIG. 4. Expression of the EAV PCP domain in E. coli. Semipurified pGEAV1 and pGEAVla expression products were analyzed by SDS-PAGE (12.5% gel). The uncleaved GEAV1 and GEAVla proteins (69 and 86 kDa, respectively) and their cleavage products are indicated. The C-terminal cleavage product of GEAV1 (predicted size, 14 kDa) was not detected. The corresponding GEAVla cleavage product (32 kDa) was purified and used for N-terminal sequence analysis. The right panel shows the composition and processing of the fusion protein encoded by pGEAVla.

strates. They contained normal EAV cleavage sites (Gly260-Gly-261), which remained unprocessed because of mutations in the pCP1 and pCP10 protease domains. In addition to nspl (derived from plaA2), an nspl precursor

Gly260 Ala 46K

Val

Gly261 Ala

_b

30K-

4.

41

pla A2

11 L_

-_

12

FIG. 6. Direct SDS-PAGE analysis (12.5% gel) of cotranslational and posttranslational trans-cleavage assays. Cleaved and uncleaved proteins containing the PCP domain were tested for the ability to process the cleavage sites of mutant constructs pCPl and pCP10. For a cotranslational assay, the RNAs encoding substrate and enzyme were mixed and translated for 1 h in the presence of [35S]methionine. To test posttranslational trans cleavage, translation inhibitors were added to the 35S-labeled substrate translation and the unlabeled protease translation, which were subsequently mixed at a 1/1 ratio and incubated for 7 additional hours. The two rightmost lanes show the results of a similar prolonged incubation of the pCP14 product, which cleaved with reduced efficiency. Abbreviations: subst., substrate; enz., enzyme; U, uncleaved PCP-containing ORFla product (the pCP12 protein); C, cleaved PCPcontaining ORFla product, releasing nspl (derived from plaAP for the cotranslational assay and from plaA2 for the posttranslational assay); -, no enzyme added; h.p.t., hours posttranslation start.

Val

-

16K-

16K -

13 14 14-1

pCP FIG. 5. Mutagenesis of the EAV PCP cleavage site. The P1 and P1' amino acid residues of the Gly-Gly cleavage site were changed into either Ala or Val. Mutations were tested by using the expression construct shown in Fig. 3, in vitro translation, and SDS-PAGE (12.5% gel). Amino acid substitutions are indicated at the top of the lane, the name of the corresponding expression plasmid is at the bottom of the lane. The translation products of construct plaA2 are also shown. The rightmost lane (pCP14-1) shows the results of a pCP14 derivative which contained a 2-nt insertion leading to translation of the + 1 reading frame downstream of Tyr-262 (see Results).

was also tested for its ability to cleave the Gly-260-Gly-261 bond in trans. This precursor, the pCP12 translation product, contained a functional protease domain but remained unprocessed because of its nonfunctional cleavage site (the Gly-260--+Val substitution). Though mixtures of substrates and proteases were incubated at 30°C for up to 7 h, both nspl and its pCP12 precursor were unable to produce detectable amounts of 30and 16-kDa cleavage products (Fig. 6). To exclude the possibility that the absence of trans cleavage was due to the experimental conditions, the pCP14 protein was subjected to a prolonged incubation at 30°C. The pCP14 translation mixture could be considered to contain both a substrate and an enzyme derived from the same construct; because of its cleavage site mutation (Gly-261--Val), a substantial amount of the pCP14 protein remained unprocessed after the standard 1-h translation reaction. After the addition of translation inhibitors and an additional 7 h at 30°C, a limited degree of posttranslational cleavage was observed (Fig. 6). Although posttranslational processing apparently was very slow, these results indicated that at least some PCP activity was possible under these conditions. To examine the option of an exclusively cotranslational trans cleavage, transcripts that encoded substrates and proteases were mixed and translated in the same reaction for 1 h at 30°C. The uncleaved protease (pCP12) was cosynthesized with the uncleaved pCP1 and PCP10 proteins. A different ORFla construct (plaAP) was created to synthesize the cleaved version of the protease in this experiment, since

7046

SNIJDER ET AL.

it would have been impossible to discriminate between the cleavage products (30 and 16 kDa) of construct plaA2 and any products derived from the cleavage in trans of the pCP1 and pCP10 precursor proteins. In addition to the 30-kDa nspl band, plaAP generated a 51-kDa C-terminal cleavage product, thereby permitting detection of the 16-kDa C-terminal product which would result from trans cleavage of the pCP1 or pCP10 protein. However, cotranslation of these precursor proteins with the PCP domain present in a cleaved (nspl produced by plaAP) or uncleaved (pCP12) N-terminal ORFla product did not result in a significant amount of processing of the pCP1 or pCP10 cleavage site (Fig. 6). Only after long exposure of autoradiographs could trace amounts of the predicted cleavage products be detected (data not shown).

DISCUSSION

Although the available coronaviruslike replicase sequences have been searched extensively for possible protease domains (to the extent that their cleavage sites have already been predicted [17, 21, 27]), there is little experimental evidence to support those theoretical analyses. The presence of a number of ORFla- and ORFlb-encoded proteins in MHV-infected cells has been reported (15-17). However, coronavirus proteases have not yet been studied in detail, and replicase cleavage sites and processing pathways remain to be elucidated. Only two MHV replicase cleavage products have been reasonably defined. A 28-kDa protein, which is found in MHV-infected cells, was shown to be autoproteolytically cleaved from the N terminus of the ORFla polypeptide (2, 15, 16). This cleavage event is probably directed by the most upstream of the two postulated papainlike MHV proteases (1, 27). The second MHV cleavage product which has been identified is a 33-kDa protein. It is derived from the 3' end of the MHV ORFlb polypeptide (7, 17) and contains one of the ORFlb domains conserved among coronaviruslike replicases. In this article, the first detailed analysis of a protease domain which is located in a coronaviruslike replicase is reported. The identification of a Cys and a downstream His as putative active site residues indicates that this EAV protease is related to the cellular papainlike enzymes and can therefore be added to the short but growing list of viral PCPs. In a recent review, Gorbalenya et al. (22) discriminated between a group of viral PCPs which mediate the production of a single N-terminal cleavage product (the so-called leader proteases) and PCPs which are thought to be the main proteases involved in multiple processing steps. Of the latter group, only the PCP domain residing in the Sindbis virus nsp2 has actually been shown to possess proteolytic activity (13, 23, 38); it is responsible for production of the nonstructural Sindbis virus proteins from a polyprotein precursor. The EAV PCP clearly belongs to the group of leader proteases which (among others) includes the potyvirus helper component protease (HC-Pro; 9, 10, 30) and the p29 protease of the hypovirulence-associated virus of the chestnut blight fungus (11, 12, 24). Although the overall sequence similarity between the EAV domain and these two wellstudied leader PCPs is very limited, some striking similarities can be observed. The spacing between the putative catalytic Cys and His residues is similar and clearly different from the spacing in cellular papainlike proteases (Fig. 2B). In addition, all three proteases efficiently cleave a Gly-Gly

J. VIROL.

dipeptide which is located 30 to 41 aa downstream of the putative active site His residue. The amino acid sequence requirements of the EAV PCP cleavage site have not yet been studied in detail. However, the fact that all residues downstream of Tyr-262 (P2') appear to be nonessential for proteolytic activity (construct pCP14-1) is another remarkable similarity with the potyviral HC-Pro. Carrington and Herndon (10) recently reported that deletion of the HC-Pro sequences downstream of the P2' position did not affect processing of the cleavage site. In the same report, the effect of a large number of amino acid substitutions between the P5 and P2' positions of the HC-Pro cleavage site was described. Substitutions of the amino acid residues occupying the P4 (Tyr), P2 (Val), P1 (Gly), and P1' (Gly) positions dramatically inhibited proteolysis. The requirements for the EAV P1 and P1' positions appear to be less stringent; the cleavage sites in the pCP11 and pCP13 mutant proteins, Ala-260-Gly-261 and Gly-260-Ala-261, respectively, were processed with efficiencies close to that of the plaA2 protein. Even the drastic Gly-261 to Val substitution in pCP14 only partially inhibited proteolysis. In the case of the potyviral HC-Pro, all these mutations completely abolished proteolytic activity. Under in vitro translation conditions, the release of EAV nspl from its precursor was rapid and efficient. After a 1-h translation reaction, only trace amounts of precursor were detected among the translation products of ORFla constructs (Fig. 1B and 3). Even after short incubation times (5 to 10 min), the bands representing the 16- and 30-kDa plaA2 cleavage products were considerably more prominent than that of their 46-kDa precursor (data not shown). This indicates that the cleavage efficiency may be comparable to that of the potyvirus HC-Pro (>50% in 7 min; 10, 30) and that cleavage may be virtually cotranslational. In fact, the translation of downstream sequences may interfere with posttranslational processing; the mutant pCP14 product, of which about 45% is cleaved during the 1-h translation reaction (Fig. 5 and 6), was processed only very inefficiently during a subsequent 7-h posttranslational incubation (approximately 20% additional cleavage; Fig. 6). Similar observations were made for the pCP11 product (data not shown). We have failed to detect other cleavage events after in vitro translation of ORFla-specific RNAs. This is surprising since, in addition to the PCP domain, the larger constructs from the plaAn series (plaA7 and plaA8) and the full-length ORFla construct contain the coding information for the putative EAV serine protease (aa 1090 to 1210; 14). The alignment of this region with trypsinlike and 3C-like proteases is far more convincing than is the similarity between the PCP domain and papainlike proteases. We therefore expected the appearance of additional cleavage products in translation reactions which expressed the serine protease domain. However, although several additional bands were observed (Fig. 1B), our search for other ORFla cleavage sites has not been successful so far. When the sizes of the additional bands are compared with those of the predicted translation products of the 5'-coterminal nested set of ORFla-encoding RNAs (Fig. 1A), none of the extra bands can be explained as the obvious result of proteolytic processing. Moreover, substitution of the putative catalytic Ser1184 of the serine protease for Cys, Phe, or Tyr did not result in changes in the pattern of pEAVla in vitro translation products (data not shown). The additional bands must therefore be attributed to internal initiation and premature termination of in vitro translation, which are not uncommon in RRL systems. An additional problem is the fact that the

VOL. 66, 1992

EAV CYSTEINE PROTEASE

larger ORFla constructs (starting with plaA5) generate less discrete translation products. This observation could be explained by the increasing size of the translated reading frame, but the presence of protein sequences which induce instability cannot be excluded. Of course our failure to detect serine protease-related processing could simply be due to the absence of additional cleavage sites in the ORFla protein sequence. An alternative explanation would be that the EAV serine protease is not active in the RRL system which was used for our in vitro translation studies. To test the latter possibility, we are currently preparing antibodies specific for ORFla-encoded proteins. They will enable us to study the in vivo processing of the ORFla product in infected cells and transient expression systems. By using an antipeptide rabbit antiserum directed against the N-terminal part of the ORFla product, the 30-kDa EAV nspl was recently identified in infected cells (39). This may enable us to study the function of nspl, which contains a cysteine-rich N-terminal region in addition to the C-terminal PCP domain. Also, EAV nsp2 (starting with Gly-261) will be the subject of future study. We have recently been able to compare the EAV sequence with the N-terminal ORFla sequence of the related LDV (33). Several sequence motifs from the putative EAV nsp2 region could be matched with the LDV ORFla sequence. Among the most interesting sequence similarities is a near-perfect match (10 of 11 residues) directly downstream of the Gly260-Gly-261 cleavage site. The conserved nature of the EAV nsp2 N terminus indicates that the LDV ORFla product may also be cleaved upstream of this motif, at a Tyr-Gly dipeptide. Two possible PCP motifs were detected in the 380-aalong ORFla sequence upstream of this putative cleavage site (33). In contrast to the corresponding ORFlb sequences, the coronavirus and EAV ORFla sequences apparently are too diverged to allow the identification of important domains by searching for conserved amino acid sequences (14). The results from the partial EAV-LDV ORFla comparison described above indicate that the LDV replicase sequences will be more suitable for this purpose. Sequence comparison will therefore remain an important tool during future functional analyses of the EAV replicase. ACKNOWLEDGMENTS We thank Johan den Boon for the construction of plasmid pEAVla, for valuable technical assistance, and for critical reading of the manuscript. We thank Twan de Vries (Department of Virology, Veterinary Faculty, University of Utrecht) for advice on the use of the pGEX system and for the purification of fusion proteins. We are grateful to Alexander E. Gorbalenya for many helpful long-distance discussions and suggestions. We thank R. Amons (Department of Medical Biochemistry, Leiden University) for carrying out the N-terminal sequence analysis of the GEAVla 32-kDa cleavage product. Protein sequencing was carried out at the gasphase sequenator facility which is supported by The Netherlands Foundation for Chemical Research (SON). We thank the Department of Haematology (Leiden University) for allowing us to use their laser densitometer. REFERENCES 1. Baker, S. C., N. La Monica, C.-K. Shieh, and M. M. C. Lai. 1990. Murine coronavirus gene 1 polyprotein contains an auto-

proteolytic activity. Adv. Exp. Med. Biol. 286:283-289. 2. Baker, S. C., C.-K. Shieh, L. H. Soe, M.-F. Chang, D. M. Vannier, and M. M. C. Lai. 1989. Identification of a domain required for autoproteolytic cleavage of murine coronavirus gene A polyprotein. J. Virol. 63:3693-3699.

7047

3. Bazan, J. F., and R. J. Fletterick 1988. Viral cysteine proteases are homologous to the trypsin-like family of serine proteases: structural and functional implications. Proc. Natl. Acad. Sci. USA 85:7872-7876. 4. Berger, A., and I. Schechter. 1970. Mapping of the active site of papain with the aid of peptide substrates and inhibitors. Philos. Trans. R. Soc. Lond. B Biol. Sci. 257:249-264. 5. Boursnell, M. E. G., T. D. K. Brown, I. J. Foulds, P. F. Green, F. M. Tomley, and M. M. Binns. 1987. Completion of the sequence of the genome of the coronavirus avian infectious bronchitis virus. J. Gen. Virol. 68:57-77. 6. Bredenbeek, P. J., C. J. Pachuk, J. F. H. Noten, J. Charite, W. Luytjes, S. R. Weiss, and W. J. M. Spaan. 1990. The primary structure and expression of the second open reading frame of the polymerase gene of the coronavirus MHV-A59. Nucleic Acids Res. 18:1825-1832. 7. Bredenbeek, P. J., E. J. Snider, A. F. H. Noten, J. A. den Boon, W. M. M. Schaper, M. C. Horzinek, and W. J. M. Spaan. 1990. The polymerase gene of corona- and toroviruses: evidence for an evolutionary relationship. Adv. Exp. Med. Biol. 286:307316. 8. Brierley, I., P. Diggard, and S. C. Inglis. 1989. Characterization of an efficient coronavirus ribosomal frameshifting signal: requirement for an RNA pseudoknot. Cell 57:537-547. 9. Carrington, J. C., S. M. Cary, T. D. Parks, and W. G. Dougherty. 1989. A second proteinase encoded by a plant potyvirus genome. EMBO J. 8:365-370. 10. Carrington, J. C., and K. L. Herndon. 1992. Characterization of the potyviral HC-Pro autoproteolytic cleavage site. Virology 187:308-315. 11. Choi, G. H., D. M. Pawlyk, and D. L. Nuss. 1991. The autocatalytic protease p29 encoded by a hypovirulence-associated virus of the chestnut blight fungus resembles the potyvirus-encoded protease HC-Pro. Virology 183:747-752. 12. Choi, G. H., R. Shapira, and D. L. Nuss. 1991. Cotranslational autoproteolysis involved in gene expression from a doublestranded RNA genetic element associated with hypovirulence of the chestnut blight fungus. Proc. Natl. Acad. Sci. USA 88:11671171. 13. De Groot, R. J., W. R. Hardy, Y. Shirako, and J. H. Strauss. 1990. Cleavage site preferences of Sindbis virus polyproteins containing the non-structural proteinase. Evidence for temporal regulation of polyprotein processing in vivo. EMBO J. 9:26312638. 14. Den Boon, J. A., E. J. Snijder, E. D. Chirnside, A. A. F. de Vries, M. C. Horzinek, and W. J. M. Spaan. 1991. Equine arteritis virus is not a togavirus but belongs to the coronaviruslike superfamily. J. Virol. 65:2910-2920. 15. Denison, M. R., and S. Perlman. 1987. Identification of putative polymerase gene product in cells infected with murine coronavirus A59. Virology 157:565-568. 16. Denison, M. R., P. W. Zoltick, S. A. Hughes, B. Giangreco, A. L. Olson, S. Perlman, J. L. Leibowitz, and S. R. Weiss. 1992. Intracellular processing of the N-terminal ORFla proteins of the coronavirus MHV-A59 requires multiple proteolytic events. Virology 189:274-284. 17. Denison, M. R., P. W. Zoltick, J. L. Leibowitz, C. J. Pachuk, and S. R. Weiss. 1991. Identification of polypeptides encoded in open reading frame lb of the putative polymerase gene of the murine coronavirus mouse hepatitis virus A59. J. Virol. 65:

3076-3082. 18. De Vries, A. A. F., E. D. Chirnside, P. J. Bredenbeek, L. A. Gravestein, M. C. Horzinek, and W. J. M. Spaan. 1990. All subgenomic mRNAs of equine arteritis virus contain a common leader sequence. Nucleic Acids Res. 18:3241-3247. 19. Godeny, E. K., D. W. Speicher, and M. A. Brinton. 1990. Map location of lactate dehydrogenase-elevating virus (LDV) capsid protein (Vpl). Virology 177:768-771. 20. Gorbalenya, A. E., A. P. Donchenko, V. M. Blinov, and E. V. Koonin. 1989. Cysteine proteases of positive strand RNA viruses and chymotrypsin-like serine proteases: a distinct protein superfamily with a common structural fold. FEBS Lett. 243: 103-114.

7048

SNIJDER ET AL.

21. Gorbalenya, A. E., E. V. Koonin, A. P. Donchenko, and V. M. Blinov. 1989. Coronavirus genome: prediction of putative functional domains in the nonstructural polyprotein by comparative amino acid sequence analysis. Nucleic Acids Res. 17:48474861. 22. Gorbalenya, A. E., E. V. Koonin, and M. M. C. Lai. 1991. Putative papain-related thiol proteases of positive-strand RNA viruses. FEBS Lett. 288:201-205. 23. Hardy, W. R., and J. H. Strauss. 1989. Processing the nonstructural proteins of Sindbis virus: nonstructural proteinase is in the C-terminal half of nsP2 and functions both in cis and in trans. J. Virol. 63:4653-4664. 24. Koonin, E. V., G. H. Choi, D. L. Nuss, R. Shapira, and J. C. Carrington. 1991. Evidence for common ancestry of a chestnut blight hypovirulence-associated double-stranded RNA and a group of positive-strand RNA plant viruses. Proc. Natl. Acad. Sci. USA 88:10647-10651. 25. Kunkel, T. A., J. D. Roberts, and R. Zakour. 1987. Rapid and efficient site-directed mutagenesis without phenotypic selection. Methods Enzymol. 154:367-382. 26. Kuo, L., J. T. Harty, L. Erickson, G. A. Palmer, and P. G. W. Plagemann. 1991. A nested set of eight RNAs is formed in macrophages infected with lactate dehydrogenase-elevating virus. J. Virol. 65:5118-5123. 27. Lee, H. J., C. K. Shieh, A. E. Gorbalenya, E. V. Koonin, N. Ia Monica, J. Tuler, A. Bagdzhadzhyan, and M. M. C. Lai. 1991. The complete sequence (22 kilobases) of murine coronavirus gene 1 encoding the putative proteases and RNA polymerase. Virology 180:567-582. 28. Maniatis, T., E. F. Fritsch, and J. Sambrook 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

J. VIROL. 29. Matsudaira, P. 1987. Sequence from picomole quantities of proteins electroblotted onto polyvinylidene difluoride membranes. J. Biol. Chem. 262:10035-10038. 30. Oh, C.-S., and J. C. Carrington. 1989. Identification of essential residues in potyviral proteinase HC-Pro by site-directed mutagenesis. Virology 173:692-699. 31. Polgar, L., and P. Halasz. 1982. Current problems in mechanistic studies of serine and cysteine proteinase. Biochem. J. 207:1-10. 32. Plagemann, P. G. W., and V. Moennig. 1991. Lactate dehydrogenase-elevating virus, equine arteritis virus, and simian haemorrhagic fever virus: a new group of positive-stranded RNA viruses. Adv. Virus Res. 41:99-192. 33. Plagemann, P. G. W. Personal communication. 34. Smith, D. B., and K S. Johnson. 1988. Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Gene 67:31-40. 35. Sn"der, E. J., J. A. den Boon, P. J. Bredenbeek, M. C. Horzinek, R. Rijnbrand, and W. J. M. Spaan. 1990. The carboxyl-terminal part of the putative Berne virus polymerase is expressed by ribosomal frameshifting and contains sequence motifs which indicate that toro- and coronaviruses are evolutionary related. Nucleic Acids Res. 18:4535-4542. 36. SnUder, E. J., M. C. Horzinek, and W. J. M. Spaan. 1990. A 3'-coterminal nested set of independently transcribed mRNAs is generated during Berne virus replication. J. Virol. 64:331-338. 37. Spaan, W. J. M., D. Cavanagh, and M. C. Horzinek. 1988. Coronaviruses: structure and genome expression. J. Gen. Virol. 69:2939-2952. 38. Strauss, J. H., and E. G. Strauss. 1991. Alphavirus proteinases. Semin. Virol. 1:347-356. 39. Wassenaar, A. L. M., and E. J. Snider. Unpublished results.