- Email: [email protected]
Processing of the tobacco etch virus 49K protease requires autoproteolysis
VIROLOGY
160, 355-362
(1987)
Processing
of the Tobacco
Etch Virus 49K Protease
JAMES C. CARRINGTON Department
of Microbiology, Received
Oregon April
AND
WILLIAM
State
28, 1987;
Autoproteolysis
G. DOUGHERTY’
University,
accepted
Requires
Corvallis.
Oregon
97331
June 2, 1987
The final products encoded by the tobacco etch virus genome arise by proteolytic cleavage of a single large polyprotein precursor. Processing of the polyprotein at several sites requires the activity of a viral protease of 49,000 molecular weight (49K). We have examined the excision of the 49K protease from polyproteins translated from defined RNA transcripts. Polyproteins containing an intact 49K protein were efficiently processed after synthesis in a rabbit reticulocyte lysate to yield the 49K product. Introduction of a single amino acid substitution (cysteine to alanine) at the putative active site of the 49K protease abolished processing, indicating that the protease was excised from the polyprotein via an autocatalytic mechanism. Release of the 49K protease was determined to require autoproteolysis, since synthetic polyproteins which contained either or both 49K cleavage sites were processed poorly, if at all, in trans reactions. Protein microsequence analysis revealed that processing in vitro occurred between a glutamine-glycine dipeptide to generate the 49K amino terminus. o 1987 Academic PWSS, hc.
INTRODUCTION
which contains conserved cysteine and histidine residues, represent the active site of these proteases (Argos et a/., 1984). Interestingly, the entire amino-terminal half of the 49K protein can be deleted without loss of proteolytic activity, indicating that the proteolytically active moiety resides within a 27K domain near the carboxy terminus (Carrington and Dougherty, 1987). We have proposed that the 49K protease functions autocatalytically to release from the polyprotein; this is in addition to functioning in Vans at several points in the TEV polyprotein substrate (Carrington and Dougherty, 1987). Autoprocessing of the 49K protease presumably would occur as a primary event in the proteolytic cascade. In this paper, we have examined processing of the 49K protease using cell-free translation in conjuction with in vitro transcription of cloned cDNA. Cleavage of the 49K protease from the polyprotein was found to occur via an autocatalytic mechanism. Additionally, processing in the cell-free system was demonstrated biochemically to occur at authentic and predicted sites.
Tobacco etch virus (TEV) belongs to the plant potyvirus group. Members of this group are characterized by long, flexuous rod morphology and a single-component, single-stranded RNA genome of approximately 10,000 nucleotides (Hollings and Brunt, 1981). The genome of TEV (9495 nucleotides) is expressed as a large polyprotein precursor (346,000 mol wt) which undergoes extensive proteolytic processing to yield mature viral proteins (Dougherty, 1983; Allison et al., 1985b, 1986; Carrington and Dougherty, 1987). The protease responsible for many TEV polyprotein cleavage events has been demonstrated to be a 49,000 mol wt (49K), viral-encoded product (Carrington and Dougherty, 1987). Within the nuclei of TEV-infected plant cells, this protease coaggregates with a distinct 54,000 mol wt (54K), viral-encoded protein to form plate-like inclusion bodies (Knuhtsen et a/., 1974; Dougherty and Hiebert, 1980b). The activity of this protease has been characterized using 49K protein products synthesized from defined SP6 transcripts in vitro, as well as from authentic 49K protein present within nuclear inclusion bodies extracted from infected plants (Carrington and Dougherty, 1987). A segment of amino acid sequence near the carboxy terminus of the 49K protein exhibits a high degree of homology with sequences of other thiol proteases encoded by several animal picornaviruses and cowpea mosaic virus (Argos et al., 1984; Allison et al., 1986). It has been hypothesized that amino acid residues within this cluster, ’ To whom
requests
for reprints
should
MATERIALS Virus,
bacteria,
AND
METHODS
and plasmids
Complementary DNA (cDNA) used for cloning was synthesized from RNA of the highly aphid transmissible isolate of TEV (Pirone and Thornbury, 1983). A nonaphid transmissible isolate (Simons, 1976) was used for preparation of nuclear inclusion bodies. Except where noted, Escherichia co/i strain HBl 01 was used as host for all DNA cloning.
be addressed. 355
0042.6822/87
$3.00
CopyrIght 0 1987 by Academic Press. Inc. All rtghts of reproduction in any form reserved.
356
CARRINGTON
The in vitro expression vector pTL-8 has been described (Carrington and Dougherty, 1987). Briefly, this plasmid contains cDNA representing the 5’ noncoding and initial coding regions of the TEV genome downstream from an SP6 transcriptional promoter. A multiple cloning site downstream from this TEV cDNA sequence facilitates insertion of DNA. If the inserted DNA contains an open reading frame which is in-phase with the initial TEV coding region, translation of the SP6 transcripts will yield a fusion product consisting of 20 TEV-encoded amino acids and the gene product of interest. Plasmids pTL-5468 and pTL-5473, which consist of cDNA representing TEV genome nucleotides 54126759 and 5412-7285, respectively, inserted into pTL8, have been described (Carrington and Dougherty, 1987). Plasmid pTL-5473/C-A contains the same DNAfragment insert as pTL-5473 except that the sequence TGT representing genome nucleotides 6706-6708 was changed to GCC by site-directed mutagenesis. This alteration results in the incorporation of alanine in place of cysteine at amino acid position 339 of the 49K protease (amino acid 2 188 on the TEV 346K polyprotein). In vitro mutagenesis Plasmid pTL-5473/C-A was derived by site-directed mutagenesis of pTL-5473 (see above) using the method of Taylor et al, (1985a,b). All system components were purchased from Amersham and used according to their recommendations. The cDNA in pTL-5473 was subcloned into pUC1 18 in E. co/i MVl 190. Single-stranded DNA production was induced by transfection with the helper phage M 13K07 and incubation at 37” for 16 hr. Ten micrograms of ssDNA was annealed with 50 ng of the oligonucleotide pGGACTGCCGGCCTGC which contained the 3-nucleotide substitution (underlined). Mutant-strand DNA synthesis was carried out with Klenow fragment in the presence of dATP, dGTP, dlTP, and the thionucleotide dCTPaS, after which the strand was circularized with T4 DNA ligase. Residual ssDNA which was not converted to the double-stranded form was removed by passage of the DNA through a nitrocellulose filter in the presence of 0.5 M NaCI. The dsDNA was nicked at several positions on the “nonmutant” strand by digestion with the restriction endonuclease /WI; this enzyme will not cleave phosphorothioate-modified DNA strands (i.e., the mutant strand). The nicked, nonmutant strand was removed by digestion with 50 units of exonuclease Ill at 37” for 30 min. After inactivation of the exonuclease III, the “degraded” strand was resynthesized and circularized using DNA polymerase I and T4 DNA ligase during a 3-hr incubation at 16”. The dsDNA was used to transform E. co/i strain TGl. Since the desired mutation re-
AND
DOUGHERTY
sulted in creation of a Nael restriction site, plasmid clones were screened with Nael. The sequence at the predicted site of mutation on one plasmid, pTL-54731 C-A, was confirmed by chemical sequence analysis (Maxam and Gilbert, 1980). In vitro transcription,
translation,
and processing
Uncapped transcripts were synthesized from linearized plasmid DNA using SP6 polymerase (Promega Biotec) by the method of Melton et a/. (1984) as modified previously (Carrington and Dougherty, 1987). SP6 transcripts (“2 pg) were translated in a rabbit reticulocyte lysate (Green Hectares) system (30 ~1) in the presence of radiolabeled amino acids as described (Dougherty and Hiebert, 1980a; Carrington and Dougherty, 1987). Radiolabeled translation products (15 ~1) which served as proteolytic substrates were reacted with 7 pg of purified TEV nuclear inclusion protein in the presence of 3 mM phenylmethylsulfonyl fluoride (PMSF) at 30” for 2 hr. Nuclear inclusion bodies were purified from TEV-infected Datura strarnonium plants (Dougherty and Hiebert, 1980b). 35S- and 3H-labeled cell-free translation and processing products were analyzed by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) (Laemmli, 1970) and fluorography using En3Hance (New England Nuclear). Protein
microsequencing
Partial amino-terminal amino acid sequence analyses were performed on proteins which were generated by cell-free translation of SP6 transcripts followed by proteolytic processing. Approximately 25 pg of transcripts were translated in the rabbit reticulocyte lysate (300 ~1) in the presence of 100 &i of [35S]methionine, 100 &i of [3H]leucine, or 100 PCi of [3H]alanine (New England Nuclear). In some instances, the radiolabeled product was reacted with 35 pg of nuclear inclusion protein as described above. The translation products were resolved by preparative SDS-PAGE. After localization of the bands by autoradiography of the dried gel, the proteins of interest were rehydrated, electroeluted, and concentrated essentially by the method of Kelly et al. (1983). The eluted protein (in 100 ~1 H20) was subjected to partial sequence analysis using an Applied Biosystern gas-phase sequencer; radioactivity released during each cycle was monitored using a Beckman scintillation counter. RESULTS Effect of a mutation putative active site
within
the 49K protease
The position of the 49K protease coding sequence on the TEV genome is shown in Fig. 1A. To investigate the processes required for release of the 49K protease
SELF-PROCESSING
--A-
-__
_ _ _ _ _ _ - - - ---
- -
pTL-!+#73,C..A
B. cffcs::c 3’
I
pTL-5473
pTL-5473/C-A
FIG. 1. (A) Diagrammatic representation of the TEV genome (top) and structure of cDNA in clones pTL-5473 and pTL-5473/C-A (enlarged figure). The genomic RNA (9495 nucleotides) is linked at the 5’ end to a small protein, termed VPg (Hari, 1981). The 5’ and 3’ noncoding regions are represented by straight lines, while the open reading frame for the large polyprotein is designated by the boxed area. Except for the 30K capsid and 49K protease, the functions and precise coding sequence locations for the proteins are based on amino acid sequence homologies with proteins of known function, gene order conservation between TEV and several other viruses, and predicted cleavage sites (see Carrington and Dougherty, 1987). The genome region represented as cDNA in plasmids pTL-5473 and pTL-5473/C-A is indicated by the stippled area. This cDNA has been inserted downstream from the TEV “leader” fragment in pTL-8, and is presented below the genetic map. The Q-G symbols above the box represent the positions of probable cleavage sites on the polyprotein which delineate the 49K protease boundaries. Transcription of this cDNA is afforded by the SP6 promoter upstream from the TEV “leader”. A portion of the amino acid sequence (single-letter code) encoded by the DNA of wild-type and mutant constructs is shown, where dashed lines denote identical residues. The asterisks indicate those residues which were proposed to function directly at the active center of the protease (see text). Cl, cytoplasmic cylindrical inclusion; NI, nuclear inclusion; Cap, capsid; Pro, protease; Pal. putative polymerase. (B) Nucleotide sequence of pTL-5473 and pTL5473/C-A at the site of mutagenesis and the deduced amino acid sequence of the proteins expressed from each. Plasmid DNA was digested with Ndel (corresponding to genome nucleotide 6580) 5’ end-labeled with [Y-~‘P]ATP using phage T4 polynucleotide kinase. and subjected to base-specific cleavage reactions as described (Maxam and Gilbert, 1980). DNA fragments were separated on an 8% polyacrylamide gel containing 8 M urea and were detected by autoradiography.
OF
TEV
357
PROTEASE
from the viral polyprotein, we expressed (using in vitro transcription and cell-free translation) segments of TEV cDNA which were cloned in the vector pTL-8. pTL5473 harbors cDNA (representing TEV genome nucleotides 5414-7285) which encompasses the entire 49K coding region as well as flanking coding regions (Fig. 1A). Previously, we have shown that translation of pTL-5473 transcripts results in efficient synthesis of a 75,000 mol wt precursor and processing to a 49,000 mol wt product with immunological and biochemical properties similar to authentic 49K nuclear inclusion protein (Carrington and Dougherty, 1987). To determine if processing at the 49K borders occurs by autoproteolysis or by another viral or reticulocyte lysate component, an amino acid substitution at the proposed (Argos et al., 1984; Allison et a/., 1986) active site of the 49K protease was introduced. The triplet (TGT) specifying cysteine at residue position 339 of the 49K protease was converted, by site-directed mutagenesis of cDNA in pTL-5473, to a triplet (GCC) specifying alanine (Fig. 1A). Nucleotide sequencing of wild-type (pTL-5473) and mutagenized (pTL-5473/C-A) cDNA in the vicinity of the predicted alteration confirmed that a coding change from cysteine to alanine had been incorporated (Fig. 1 B). Transcripts from pTL-5473 and two pTL-5473/C-A clones, each possessing the coding potential for a 75,000 mol wt product, were translated in the reticulocyte lysate system in the presence of [35S]methionine. Translation of pTL-5473 transcripts yielded the 49K product (Fig. 2, lane 1) as shown previously(Carrington and Dougherty, 1987). However, translation of the pTL5473/C-A transcripts resulted in accumulation of a
1 2 3
FIG. 2. Expression of pTL-5473 and two pTL-5473/C-A isolates. Plasmid DNA was linearized with Pvull (which cleaves within the vector sequence) and transcribed with SP6 polymerase. The uncapped transcripts were translated in a rabbit reticulocyte lysate in the presence of [35S]methionine. The products were separated on a SDS-polyacrylamide (5% stacking, 12.5% resolving) gel and detected by fluorography. Translation products were synthesized from transcripts of pTL-5473 (lane 1) and two pTL-5473/C-A plasmids (lanes 2 and 3). Apparent molecular weights (Xl O-3) are shown on the left,
358
CARRINGTON
75,000 mol wt product which was not processed to the 49K form (Fig. 2, lanes 2 and 3; these lanes show translation products of transcripts from two apparently identical plasmids which were selected during the mutagenesis procedure). The faint, 35S-labeled polypeptides underneath the 75,000 mol wt product are believed to represent premature translational termination products since further incubation (3 hr) did not result in their accumulation (data not shown). Processing containing
in tm~s of synthetic 49K cleavage sites
AND
DOUGHERTY
A VPg Pro
VP4,
87 K
2
50 K
.
70 K
c’
~~:~:;:::::~~~:~~q(~)
O-S O-G
jljjlllllijiijjljjjlIlii.ii:iitilllllili
O-S O-G
pTL-5468
P;G
:::::::::::y.:.::.:.::. .:.‘,j::::::::::::::.:::.::::::::::I::: ..:..........:. :,.::.:.:.:..
pTL-5473(C-A)
polyproteins
In an earlier report (Carrington and Dougherty, 19871, we have shown that processing in trans was inefficient at the amino-terminal cleavage site of the 49K protein in the synthetic polyprotein encoded by pTL-5468 transcripts (see diagram in Fig. 3A). This polyprotein harbors amino acid sequences for TEV protein totalling approximately 12,000 mol wt upstream from the cleavage junction, followed by a 41,000 mol wt polypeptide representing the amino-terminal segment of the 49K protease. The 53,000 mol wt translation product does not undergo autoproteolysis since the carboxy terminus of the protease, shown to be essential for enzymatic activity (Carrington and Dougherty, 1987), has been deleted. This poor substrate activity using 49K protease from purified nuclear inclusions could result from aberrant structural conformation due to deletion of the carboxy terminus or from an inherent property of the cleavage site. To test the hypothesis that the entire 49K protein is required for processing in Vans at its borders, cleavage reactions were performed on synthetic polyprotein substrates produced by translation of transcripts from pTL-5468, pTL-5473/C-A, and pTL6895. The 75,000 mol wt product of pTL-5473/C-A transcripts serves to test both 49K cleavage sites on a polyprotein which contains the complete 49K sequence with the exception of the Cys to Ala substitution at position 339. The transcripts synthesized from pTL6895 encode a polyprotein of approximately 95,000 mol wt consisting of an -6000 mol wt fragment from the carboxy terminus of the 49K protease and the entire 54K and 30K (capsid) sequences (Fig. 3A). The 95,000 mol wt polyprotein (from pTL-6895) serves as a substrate to analyze cleavage at the carboxy terminus of the 49K protein as well as the 54K-30K cleavage junction [which is “efficiently” processed in tram (Carrington and Dougherty, 1987)l. Transcripts from each plasmid were translated in vitro in the presence of [35S]methionine and tested for their substrate activity using the 49K protease from nuclear inclusion bodies. As in the prior study (Carrington and Dougherty, 1987), the 53,000 mol wt product of pTL5468 transcripts was processed poorly in Vans, yielding
6
1234567
FIG. 3. Trans processing of three polyproteins containing the aminoand/or carboxy-terminal cleavage sites for the 49K protease. (A) Schematic representation of cDNAs expressed by in vitro transcription and translation. The shadings and symbols are the same as described in Fig. 1. The asterisk below the pTL-5473 diagram indicates the position of the Cys to Ala amino acid substitution in pTL-5473/C-A. (6) Substrate activities of synthetic polyproteins, encoded by transcripts of cDNAs shown in (A), using the processing activity of the 49K protease within purified nuclear inclusion bodies. The %-labeled polyproteins synthesized in the reticulocyte lysate system (30 pl) were divided into equal portions: one-half was incubated with 49K protease (as detailed under Materials and Methods), while the other half was incubated with H20. Products were analyzed by SDS-polyacrylamide gel electrophoresis and fluorography. Lane I, pTL-5473 derived proteins; lanes 2 and 3. pTL-5473/C-Aderived products reacted without (lane 2) or with (lane 3) protease; lanes 4 and 5, pTL-5468 derived products reacted without (lane 4) or with (lane 5) protease; lanes 6 and 7, pTL-6895 derived products reacted without (lane 6) or with (lane 7) protease. The arrows indicate the products (60,000 mol wt in lane 3; 41,000 mol wt in lane 5) resulting from inefficient ffans cleavage at the amino terminus of the 49K protease (see Results). The molecular weights (Xl Om3) and positions of Y-labeled protein markers are shown at the left.
minor amounts of a 41,000 mol wt product (Fig. 3B, lanes 4 and 5), whereas the polyprotein encoded by pTL-5473 transcripts underwent efficient autoproteolysis (lane 1). Microsequence analysis of the amino terminus of the 41 K product revealed that (inefficient) processing in trans occurred at the same site as (efficient) autoproteolysis (see below).
SELF-PROCESSING
Under identical processing conditions as with the pTL-5468 derived product, the 75,000 mol wt polypeptide encoded by pTL-5473/C-A transcripts also was processed with poor efficiency, yielding only minor mol wt product and little or amounts of an -60,000 no accumulation of a 49,000 mol wt species (Fig. 3B, lanes 2 and 3). Although amino acid sequence analysis was not performed on the -60,000 mol wt cleavage product, we assumed that this protein represented an inefficient proteolytic event at the amino terminus of the 49K sequence, as with the pTL-5468 cleavage product. Little or no proteolysis was detected at the carboxy-terminal cleavage site of the 49K protein using the pTL-5473/C-A derived polyprotein substrate. This cleavage site also was tested using the synthetic substrate polyprotein encoded by pTL-6895 transcripts. Processing in Vans of the 95,000 mol wt polyprotein resulted in complete disappearance of the substrate and accumulation of a 30,000 mol wt protein (capsid) and an -65,000 mol wt product (lanes 6 and 7). Significantly, processing failed to yield a 54K product which would have been the result of cleavage at the 49K54K junction. The positive control for 49K protease processing activity is shown by the “efficient” release of 30K capsid protein from the 95,000 mol wt precursor encoded by pTL-6895 transcripts (lanes 6 and 7). In time course analyses, cleavage in trans at the 30K capsid amino terminus occurred at a rate at least 50 times faster than at the 49K protease amino terminus (data not shown). Thus, excision of the 49K protease from the TEV polyprotein appears to require an autocatalytic mechanism in vitro. Amino acid sequence synthesized in vitro
analysis
of 49K protease
To identify the site at the amino terminus of the protease where autoprocessing takes place, radiolabeled 49K protein synthesized in vitro from pTL-5473 transcripts was analyzed by microsequence techniques. The amino terminus of authentic 49K protein isolated from nuclear inclusion bodies is blocked (unpublished observations), thereby preventing direct amino acid sequence analysis. The 49K translation product synthesized from pTL-5473 was labeled by incorporation of [35S]methionine or [3H]leucine, purified by electroelution from a SDS-polyacrylamide gel, and applied to an automated sequencer. The radioactivity released during cycles l-20 for each labeled sample are shown in Fig. 4. An unambiguous peak of [35S]methionine was released at position 1 1, whereas a substantial yield of [3H]leucine was identified at residue position 9. The tripeptide Leu-Xaa-Met exists at only one position within the deduced amino acid sequence encoded by pTL-5473 transcripts, corresponding to residues 1858-
OF
TEV
359
PROTEASE
1860 on the 346K TEV polyprotein. Therefore, autoproteolysis at the amino terminus of the 49K protease is predicted to occur between the dipeptide glutamineglycine at positions 1849-l 850, located nine residues prior to the leucine identified by sequence analysis. To determine if the low degree of trans processing observed when the pTL-5468 derived protein was reacted with the viral protease (yielding a 41 K polypeptide) was in fact occurring at the actual amino terminus of the 49K sequence (Fig. 3B, lanes 4 and 5) the 41K product was subjected to microsequence analysis as described above. Based on the positions (relative to the amino terminus) at which either [35S]methionine or [3H]leucine was released during the automated sequencing run (Fig. 4) the limited cleavage in trans was occurring at the same dipeptide (Gln-Gly) as cleavage during autocatalysis of intact 49K protease precursor. To confirm the accuracy of processing mediated by the 49K protease in vitro, a partial amino acid sequence was determined for the 30K capsid protein generated during a cell-free processing reaction. The sequence at the amino terminus of authentic TEV capsid protein has been determined (Allison et a/., 1985a,b). A 34,000 mol wt protein encoded by transcipts from pTL-8595 (cDNA representing genome nucleotides 8462-9495), which contains coding sequences for a short segment of the 54K protein and the entire 30K capsid coding region (see Fig. 1A), serves as an useful substrate to examine 49K protease-mediated cleavage in a trans assay. In the presence of active 49K protease, the 34,000 mol wt substrate is converted to a 30,000 mol wt product of the same size and immunological reactivity as authentic capsid protein (Carrington and Dougherty, 1987). The 34,000 mol wt product was synthesized in vitro in the presence of [3H]alanine, reacted with protease present within TEV nuclear inclusion bodies, and subjected to preparative SDS-gel electrophoresis. The 30,000 mol wt cleavage product was electroeluted from the gel and applied to the automated sequencer. Peaks of [3H]alanine were released during cycles 6, 8, and 10 (Fig. 4). This combination of alanine residues exists at only one position within the pTL-8595 coding region (corresponding to amino acid residues 2797, 2799, and 2801 in the TEV polyprotein), therefore indicating cleavage occurred between the glutamine-serine dipeptide at positions 279 l-2792. Indeed, this site is the same as that predicted from sequence analysis of authentic capsid protein (Allison et al., 1985a,b). DISCUSSION Thiol proteases use a free sulfhydryl group from cysteine as a nucleophile during catalysis of peptide bond cleavage (Lowe, 1976). Based on studies involving in-
360
CARRINGTON 49K (Protease) GKKNPKHKLKHREARGARGQ
AND
DOUGHERTY
4 I K (Partial
49K)
4Fj
30K (Capsid)
Residue
Residue
Residue
FIG. 4. Partial amino acid sequence analysis of TEV products derived by translation and processing reactions in vitro. The 49K (protease) and 30K (capsid) proteins were labeled individually with [%]methionine, [3H]leucine, or [3H]alanine, while the 41 K (partial 49K) product was colabeled with [36S]methionine and [3H]leucine. Radiolabeled proteins were eluted from SDS-polyacrylamide gels and applied to an automated sequencer (Applied Biosystems). Radioactivity released at each cycle was quantitated using a Beckman scintillation counter with channels set to discriminate between % and 3H. The amino acid sequence presented above each plot was deduced from the genome nucleotide sequence (Allison et a/., 1986). Radioactivity applied to the automated sequencer were as follows: %-49K (protease), 453,000 cpm; 3H-49K (protease). 850,000 cpm; 3H-30K (capsid), 128,000 cpm; 35S-41 K (partial 49K), 103,000 cpm; 3H-41 K (partial 49K), 24,000 cpm.
hibitors, chemical requirements for activity, and amino acid sequence homologies, most groups of singlestranded RNA viruses which utilize proteolytic processing as the predominant mode of gene expression are thought to encode thiol proteases (Pelham, 1978, 1979; Gorbalenya and Svitkin, 1983; Peng and Shih, 1984; Argos et al,, 1984; Yeh and Gonsalves, 1985; Lloyd et al., 1986; Werner et al., 1986). To test the importance of the cysteine residue (339) hypothesized to function at the active site of the TEV 49K protease (Allison et al., 1986), alanine was introduced at position 339 in place of cysteine. Synthetic polyproteins expressed from transcripts derived from DNA harboring this point mutation failed to process; thus, we conclude (i) the 49K protease is excised from the polyprotein via an autocatalytic mechanism, and (ii) the hypothesis that cysteine (339) exists at the active center of the protease is supported, although not proven. If the 49K cleavage sites were processed by a reticulocyte lysate or cellular component, as proposed for the primary cleavage event of the cowpea mosaic virus B-RNA translation product (Tian and Shih, 1986) cleavage would most likely have been unaffected by a single amino acid substitution distant to the processing sites. It has been shown or suggested that intramolecular or autoproteolytic cleavages to release a protease occur at the early stages of viral polyprotein processing
for many viruses. In addition, it is often suggested that free viral protease, excised from the polyprotein precursor by autocatalysis, then functions by an intermolecular mechanism to release additional protease molecules from subsequently synthesized polyproteins (for example, Pelham, 1979; Palmenberg and Reuckert, 1982; Hanecak et a/., 1984; Toyoda et a/., 1986). Our data suggest that this is probably not the case for TEV, since cleavage in Vans at the 49K protease boundaries occurred poorly in the reactions studied here. Since autoprocessing at each 49K protease border occurs with high efficiency in vitro (see Figs. 2 and 3) and presumably in v&o, there may be little need for trans processing to function. It is interesting, however, to consider why neither junction functions efficiently as a substrate for exogenous protease. Although clues are available through analysis of the amino acid sequence (see below), the precise sequence or protein folding requirements for the TEV substrate sites are not yet known. It is possible that cleavage sites destined for autoprocessing adopt an orientation different from those destined for Vans processing. Presumably, the autoproteolytic orientation results in efficient presentation of the cleavage site to the active center of the protease. If true, conformational switching (prior to proteolysis) at the 49K protease (or other TEV protein) boundaries might be possible such that timing of pro-
SELF-PROCESSING
tein maturation throughout the infection cycle could be regulated. Although most of the known potyviral-encoded proteins have been isolated from infected plants (Knuhtsen et a/., 1974; Dougherty and Hiebert, 1980b; Hari, 1981; Thornbury et a/., 1985; de Mejia et a/., 1985), little information is available regarding in vivo synthesis and maturation of the polyprotein or regulation of these events. The autoprocessing cleavage site which defines the amino terminus of the 49K protease was determined using protein derived in cell-free systems (Fig. 4). The validity of this approach was confirmed by demonstrating accurate cell-free processing by the 49K protease at the known 54K-30K protein cleavage junction. The location of the 49K amino-terminal cleavage site has been predicted correctly in a previous study based on approximate size of the protein and on the presence of an amino acid sequence motif shared with the known cleavage site (at the 54K-30K junction) (Carrington and Dougherty, 1987). At positions -6, -3, -2, and -1 relative to the cleaved peptide bond, the cleavage site at the amino terminus of the 49K protease and 30K capsid share Glu, Tyr, Phe, and Gln, respectively. When this comparison is extended to three additional predicted cleavage sites (at the boundaries separating 50K-70K, 70K-6K, and 49K-54K proteins, see Fig. 1A) which we hypothesize are processed by the 49K protease, the Glu (-6), Tyr (-3), and Gln (-1) residues are strictly conserved (Carrington and Dougherty, 1987). Additionally, hydrophobicity profiles in the immediate vicinity of the cleavage sites are strikingly similar (data not shown), suggesting they may share a similar local three-dimensional structure as well (Sweet and Eisenberg, 1983). Therefore, information critical for su’bstrate activity and specificity may reside within a cluster of amino acid residues in a specific orientation surrounding the cleavage site. Through site-directed mutagenesis, the relative contribution of these residues may be evaluated. ACKNOWLEDGMENTS We thank Mr. Ronnie Johnson for consistent technical assistance, MS. Mary Kelly for conducting the protein sequencing runs, and Ms. Dawn Parks and Dr. Dennis Hruby for their critical reviews of this manuscript. This research was supported by a Postdoctoral Fellowship from the North Carolina Biotechnology Center (to J.C.C.) and Grant DBM-8601939 from the National Science Foundation (to W.G.D.). Oregon Agricultural Experiment Station Technical Paper No. 8254.
REFERENCES ALLISON, R. F.. DOUGHERTY, W. G.. PARKS, T. D., WILLIS, L., JOHNSTON, R. E., KELLY, M.. and ARMSTRONG, F. B. (1985a). Biochemical analysis of the capsid protein gene and the capsid protein of tobacco etch virus: N-terminal amino acids are located on the virion’s surface. Virology 147, 309-3 16.
OF
TEV
361
PROTEASE
ALLISON, R. F., JOHNSTON, R. E., and DOUGHERP/, W. G. (1986). The nucleotide sequence of the coding region of tobacco etch virus genomic RNA: Evidence for the synthesis of a single polyprotein. Virology 154, 9-20. ALLISON, R. F., SORENSON, J. G., KELLY, M. E., ARMSTRONG, F. B., and DOUGHERIY, W. G. (1985b). Sequence determination of the capsid protein gene and flanking regions of tobacco etch virus: Evidence for the synthesis and processing of a polyprotein in potyvirus genome expression. Proc. Natl. Acad. Sci. USA 82, 3969-3972. ARGOS, P., KAMER, P., NICKLIN, M. J. H., and WIMMER, E. (1984). Similarity in gene organization and homology between proteins of animal picornaviruses and a plant comovirus suggest common ancestry of these virus families. Nucleic Acids Res. 12, 7251-7267. CARRINGTON, J. C., and DOUGHERTY, W. G. (1987). Small nuclear inclusion protein encoded by a plant potyvirus genome is a protease. J. Viral., 81, in press. DE MEJIA. M. V. G., HIEBERT, E., and PURCIFULL, D. E. (1985). Isolation and partial characterization of the amorphous cytoplasmic inclusions associated with infections caused by two potyviruses. Virology 142, 24-33. DOUGHERTY, W. G. (1983). Analysis of viral RNA isolated from tobacco leaf tissue infected with tobacco etch virus. Virology 131, 473481. DOUGHERN, W. G., and HIEBERT, E. (1980a). Translation of potyvirus RNA in a reticulocyte lysate: Reaction conditions and identification of capsid protein as one of the products of in vitro translation of tobacco etch and pepper mottle viral RNAs. Virology 101, 466474. DOUGHERTY, W. G., and HIEBERT, E. (1980b). Translation of potyvirus RNA in a reticulocyte lysate: Identification of nuclear inclusion proteins as products of tobacco etch virus RNA translation and cylindrical inclusion protein as a product of the potyvirus genome. Virology 104, 174-l 82. DOUGHERPI, W. G., and HIEBERT, E. (198Oc). Translation of potyvirus RNA in a reticulocyte lysate: Cell-free translation strategy and a genetic map of the potyviral genome. Virology 104, 183-l 94. GORBALENYA, A. E., and SVITKIN, Y. V. (1983). Protease of encephalomyocarditis virus: Purification and role of sulfhydryl groups in processing of the structural proteins precursor. Biokhimiya 48, 442-453. HANECAK, R., SEMLER, B. L., ARIGA, H., ANDERSON, C. W., and WIMMER, E. (1984). Expression of a cloned gene segment of poliovirus in E. co/i: Evidence for autocatalytic production of the viral proteinase. Cell 37, 1063-l 073. HARI, V. (1981). The RNA of tobacco etch virus: Further characterization and detection of protein linked to RNA. Virology 112, 391399. HOLLINGS, M., and BRUNT, A. A. (1981). Potyviruses. In “Handbook of Plant Virus Infection and Comparative Diagnosis” (E. Kurstak. Ed.), pp. 731-807. Elsevier/North Holland Biomedical, New York. KELLY, C., Tom, N. F., WATERFIELD. M. D.. and CRUMPTON, M. J. (1983). Amino acid sequencing of polypeptides eluted from 2D polyacrylamide gels. Biochem. lnt. 6, 535-544. KNUHTSEN, H., HIEBERT, E., and PURCIFULL. D. E. (1974). Partial purification and some properties of tobacco etch virus intranuclear inclusions. Virology 61, 200-209. LAEMMLI, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227, 680-685. LLOYD, R. E., TOYODA, H., ETCHISON, D., WIMMER, E., and EHRENFELD, E. (1986). Cleavage of the cap binding protein complex polypeptide ~220 is not effected by the second poliovirus protease 2A. Virology 150,299-303. LOWE, G. (1976). 302.
The cysteine
proteinases.
Tetrahedron
32, 291-
362
CARRINGTON
AND
MAXAM, A. M., and GILBERT, W. (1980). Sequencing end-labeled DNA with base-specific chemical cleavages. ln “Methods in Enzymology” (L. Grossman and K. Moldave, Eds.), Vol. 65, pp. 499-560. Academic Press, New York. MELTON, D. A., KRIEG, P. A., REBAGLIATI, M. R., MANIATIS, T., ZINN, K., and GREEN, M. R. (1984). Efficient in vitro synthesis of biologically active RNA and RNA hybridization probes from plasmids containing a bacteriophage SP6 promoter. Nucleic Acids Res. 12,7035-7056. PALMENBERG, A. C., and RUECKERT. R. R. (1982). Evidence for molecular self-cleavage of picornaviral replicase precursors. J. 41,244-249. PELHAM, H. R. B. (1978). Translation of encephalomyocarditis RNA in v&o yields an active proteolytic processing enzyme. J. Biochem. 85, 457-462. PELHAM, H. R. B. (1979). Synthesis and proteolytic processing cowpea mosaic virus proteins in reticulocyte lysates. Virology 463-477. PENG, X. X., and SHIH. D. S. (1984). Proteolytic proteins translated from the bottom component mosaic virus. f. Biol. Chem. 259, 3197-3201.
intraViral. virus Eur. of 96,
processing of the RNA of cowpea
PIRONE, T. P., and THORNBURY, D. W. (1983). Role of virion and helper component in regulating aphid transmission of tobacco etch virus. Phytopathology 73, 872-875. SIMONS, J. N. (1976). Aphid transmission of a non-aphid transmissible strain of tobacco etch virus. Phytopathology 66, 652-654. SWEET, R. M., and EISENBERG, D. (1983). Correlation of sequence
DOUGHERTY hydrophobicities measures similarity in three-dimensional protein structure. 1. Mol. Biol. 171, 479-488. TAYLOR, J. W., OTT, J., and ECKSTEIN, F. (1985b). The rapid generation of oligonucleotide-directed mutations at high frequency using phosphorothioate-modified DNA. Nucleic Acids Res. 13, 87658785. TAYLOR, J. W., SCHMIDT, W., COSSTICK, R., OKRUSZEK, A., and ECKSTEIN, F. (1985a). The use of phosphorothioate-modified DNA in restriction enzyme reactions to prepare nicked DNA. Nucleic Acids Res. 13, 8749-8764. THORNBURY, D. W., HELLMANN, G. M., RHOADS, R. E., and PIRONE, T. P. (1985). Purification and characterization of potyvirus helper component. Virology 144, 260-267. TIAN, Y. C., and SHIH, D. S. (1986). Cleavage of a viral polyprotein by a cellular proteolytic activity. I. Viral. 57, 547-55 1. TOYODA, H., NICKLIN, M. J. H., MURRAY, M. G.,ANDERSON. C. W., DUNN, J. J., STUDIER, F. W.. and WIMMER, E. (1986). A second virus-encoded proteinase involved in proteolytic processing of poliovirus polyprotein. Cell 45, 761-770. WERNER, G., ROSENWIRTH, B., BAUER, E., SEIFERT, J.-M., WERNER, F.-J., and BESEMER, J. (1986). Molecular cloning and sequence determination of the genomic regions encoding protease and genomelinked protein of three picornaviruses. J. Viral. 57, 1084-l 093. YEH, S.-P., and GONSALVES, D. (1985). Translation of papaya ringspot virus RNA in vitro: Detection of a possible polyprotein that is processed for capsid protein, cylindrical-inclusion protein and amorphous-inclusion protein. Virology 143, 260-271.
160, 355-362
(1987)
Processing
of the Tobacco
Etch Virus 49K Protease
JAMES C. CARRINGTON Department
of Microbiology, Received
Oregon April
AND
WILLIAM
State
28, 1987;
Autoproteolysis
G. DOUGHERTY’
University,
accepted
Requires
Corvallis.
Oregon
97331
June 2, 1987
The final products encoded by the tobacco etch virus genome arise by proteolytic cleavage of a single large polyprotein precursor. Processing of the polyprotein at several sites requires the activity of a viral protease of 49,000 molecular weight (49K). We have examined the excision of the 49K protease from polyproteins translated from defined RNA transcripts. Polyproteins containing an intact 49K protein were efficiently processed after synthesis in a rabbit reticulocyte lysate to yield the 49K product. Introduction of a single amino acid substitution (cysteine to alanine) at the putative active site of the 49K protease abolished processing, indicating that the protease was excised from the polyprotein via an autocatalytic mechanism. Release of the 49K protease was determined to require autoproteolysis, since synthetic polyproteins which contained either or both 49K cleavage sites were processed poorly, if at all, in trans reactions. Protein microsequence analysis revealed that processing in vitro occurred between a glutamine-glycine dipeptide to generate the 49K amino terminus. o 1987 Academic PWSS, hc.
INTRODUCTION
which contains conserved cysteine and histidine residues, represent the active site of these proteases (Argos et a/., 1984). Interestingly, the entire amino-terminal half of the 49K protein can be deleted without loss of proteolytic activity, indicating that the proteolytically active moiety resides within a 27K domain near the carboxy terminus (Carrington and Dougherty, 1987). We have proposed that the 49K protease functions autocatalytically to release from the polyprotein; this is in addition to functioning in Vans at several points in the TEV polyprotein substrate (Carrington and Dougherty, 1987). Autoprocessing of the 49K protease presumably would occur as a primary event in the proteolytic cascade. In this paper, we have examined processing of the 49K protease using cell-free translation in conjuction with in vitro transcription of cloned cDNA. Cleavage of the 49K protease from the polyprotein was found to occur via an autocatalytic mechanism. Additionally, processing in the cell-free system was demonstrated biochemically to occur at authentic and predicted sites.
Tobacco etch virus (TEV) belongs to the plant potyvirus group. Members of this group are characterized by long, flexuous rod morphology and a single-component, single-stranded RNA genome of approximately 10,000 nucleotides (Hollings and Brunt, 1981). The genome of TEV (9495 nucleotides) is expressed as a large polyprotein precursor (346,000 mol wt) which undergoes extensive proteolytic processing to yield mature viral proteins (Dougherty, 1983; Allison et al., 1985b, 1986; Carrington and Dougherty, 1987). The protease responsible for many TEV polyprotein cleavage events has been demonstrated to be a 49,000 mol wt (49K), viral-encoded product (Carrington and Dougherty, 1987). Within the nuclei of TEV-infected plant cells, this protease coaggregates with a distinct 54,000 mol wt (54K), viral-encoded protein to form plate-like inclusion bodies (Knuhtsen et a/., 1974; Dougherty and Hiebert, 1980b). The activity of this protease has been characterized using 49K protein products synthesized from defined SP6 transcripts in vitro, as well as from authentic 49K protein present within nuclear inclusion bodies extracted from infected plants (Carrington and Dougherty, 1987). A segment of amino acid sequence near the carboxy terminus of the 49K protein exhibits a high degree of homology with sequences of other thiol proteases encoded by several animal picornaviruses and cowpea mosaic virus (Argos et al., 1984; Allison et al., 1986). It has been hypothesized that amino acid residues within this cluster, ’ To whom
requests
for reprints
should
MATERIALS Virus,
bacteria,
AND
METHODS
and plasmids
Complementary DNA (cDNA) used for cloning was synthesized from RNA of the highly aphid transmissible isolate of TEV (Pirone and Thornbury, 1983). A nonaphid transmissible isolate (Simons, 1976) was used for preparation of nuclear inclusion bodies. Except where noted, Escherichia co/i strain HBl 01 was used as host for all DNA cloning.
be addressed. 355
0042.6822/87
$3.00
CopyrIght 0 1987 by Academic Press. Inc. All rtghts of reproduction in any form reserved.
356
CARRINGTON
The in vitro expression vector pTL-8 has been described (Carrington and Dougherty, 1987). Briefly, this plasmid contains cDNA representing the 5’ noncoding and initial coding regions of the TEV genome downstream from an SP6 transcriptional promoter. A multiple cloning site downstream from this TEV cDNA sequence facilitates insertion of DNA. If the inserted DNA contains an open reading frame which is in-phase with the initial TEV coding region, translation of the SP6 transcripts will yield a fusion product consisting of 20 TEV-encoded amino acids and the gene product of interest. Plasmids pTL-5468 and pTL-5473, which consist of cDNA representing TEV genome nucleotides 54126759 and 5412-7285, respectively, inserted into pTL8, have been described (Carrington and Dougherty, 1987). Plasmid pTL-5473/C-A contains the same DNAfragment insert as pTL-5473 except that the sequence TGT representing genome nucleotides 6706-6708 was changed to GCC by site-directed mutagenesis. This alteration results in the incorporation of alanine in place of cysteine at amino acid position 339 of the 49K protease (amino acid 2 188 on the TEV 346K polyprotein). In vitro mutagenesis Plasmid pTL-5473/C-A was derived by site-directed mutagenesis of pTL-5473 (see above) using the method of Taylor et al, (1985a,b). All system components were purchased from Amersham and used according to their recommendations. The cDNA in pTL-5473 was subcloned into pUC1 18 in E. co/i MVl 190. Single-stranded DNA production was induced by transfection with the helper phage M 13K07 and incubation at 37” for 16 hr. Ten micrograms of ssDNA was annealed with 50 ng of the oligonucleotide pGGACTGCCGGCCTGC which contained the 3-nucleotide substitution (underlined). Mutant-strand DNA synthesis was carried out with Klenow fragment in the presence of dATP, dGTP, dlTP, and the thionucleotide dCTPaS, after which the strand was circularized with T4 DNA ligase. Residual ssDNA which was not converted to the double-stranded form was removed by passage of the DNA through a nitrocellulose filter in the presence of 0.5 M NaCI. The dsDNA was nicked at several positions on the “nonmutant” strand by digestion with the restriction endonuclease /WI; this enzyme will not cleave phosphorothioate-modified DNA strands (i.e., the mutant strand). The nicked, nonmutant strand was removed by digestion with 50 units of exonuclease Ill at 37” for 30 min. After inactivation of the exonuclease III, the “degraded” strand was resynthesized and circularized using DNA polymerase I and T4 DNA ligase during a 3-hr incubation at 16”. The dsDNA was used to transform E. co/i strain TGl. Since the desired mutation re-
AND
DOUGHERTY
sulted in creation of a Nael restriction site, plasmid clones were screened with Nael. The sequence at the predicted site of mutation on one plasmid, pTL-54731 C-A, was confirmed by chemical sequence analysis (Maxam and Gilbert, 1980). In vitro transcription,
translation,
and processing
Uncapped transcripts were synthesized from linearized plasmid DNA using SP6 polymerase (Promega Biotec) by the method of Melton et a/. (1984) as modified previously (Carrington and Dougherty, 1987). SP6 transcripts (“2 pg) were translated in a rabbit reticulocyte lysate (Green Hectares) system (30 ~1) in the presence of radiolabeled amino acids as described (Dougherty and Hiebert, 1980a; Carrington and Dougherty, 1987). Radiolabeled translation products (15 ~1) which served as proteolytic substrates were reacted with 7 pg of purified TEV nuclear inclusion protein in the presence of 3 mM phenylmethylsulfonyl fluoride (PMSF) at 30” for 2 hr. Nuclear inclusion bodies were purified from TEV-infected Datura strarnonium plants (Dougherty and Hiebert, 1980b). 35S- and 3H-labeled cell-free translation and processing products were analyzed by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) (Laemmli, 1970) and fluorography using En3Hance (New England Nuclear). Protein
microsequencing
Partial amino-terminal amino acid sequence analyses were performed on proteins which were generated by cell-free translation of SP6 transcripts followed by proteolytic processing. Approximately 25 pg of transcripts were translated in the rabbit reticulocyte lysate (300 ~1) in the presence of 100 &i of [35S]methionine, 100 &i of [3H]leucine, or 100 PCi of [3H]alanine (New England Nuclear). In some instances, the radiolabeled product was reacted with 35 pg of nuclear inclusion protein as described above. The translation products were resolved by preparative SDS-PAGE. After localization of the bands by autoradiography of the dried gel, the proteins of interest were rehydrated, electroeluted, and concentrated essentially by the method of Kelly et al. (1983). The eluted protein (in 100 ~1 H20) was subjected to partial sequence analysis using an Applied Biosystern gas-phase sequencer; radioactivity released during each cycle was monitored using a Beckman scintillation counter. RESULTS Effect of a mutation putative active site
within
the 49K protease
The position of the 49K protease coding sequence on the TEV genome is shown in Fig. 1A. To investigate the processes required for release of the 49K protease
SELF-PROCESSING
--A-
-__
_ _ _ _ _ _ - - - ---
- -
pTL-!+#73,C..A
B. cffcs::c 3’
I
pTL-5473
pTL-5473/C-A
FIG. 1. (A) Diagrammatic representation of the TEV genome (top) and structure of cDNA in clones pTL-5473 and pTL-5473/C-A (enlarged figure). The genomic RNA (9495 nucleotides) is linked at the 5’ end to a small protein, termed VPg (Hari, 1981). The 5’ and 3’ noncoding regions are represented by straight lines, while the open reading frame for the large polyprotein is designated by the boxed area. Except for the 30K capsid and 49K protease, the functions and precise coding sequence locations for the proteins are based on amino acid sequence homologies with proteins of known function, gene order conservation between TEV and several other viruses, and predicted cleavage sites (see Carrington and Dougherty, 1987). The genome region represented as cDNA in plasmids pTL-5473 and pTL-5473/C-A is indicated by the stippled area. This cDNA has been inserted downstream from the TEV “leader” fragment in pTL-8, and is presented below the genetic map. The Q-G symbols above the box represent the positions of probable cleavage sites on the polyprotein which delineate the 49K protease boundaries. Transcription of this cDNA is afforded by the SP6 promoter upstream from the TEV “leader”. A portion of the amino acid sequence (single-letter code) encoded by the DNA of wild-type and mutant constructs is shown, where dashed lines denote identical residues. The asterisks indicate those residues which were proposed to function directly at the active center of the protease (see text). Cl, cytoplasmic cylindrical inclusion; NI, nuclear inclusion; Cap, capsid; Pro, protease; Pal. putative polymerase. (B) Nucleotide sequence of pTL-5473 and pTL5473/C-A at the site of mutagenesis and the deduced amino acid sequence of the proteins expressed from each. Plasmid DNA was digested with Ndel (corresponding to genome nucleotide 6580) 5’ end-labeled with [Y-~‘P]ATP using phage T4 polynucleotide kinase. and subjected to base-specific cleavage reactions as described (Maxam and Gilbert, 1980). DNA fragments were separated on an 8% polyacrylamide gel containing 8 M urea and were detected by autoradiography.
OF
TEV
357
PROTEASE
from the viral polyprotein, we expressed (using in vitro transcription and cell-free translation) segments of TEV cDNA which were cloned in the vector pTL-8. pTL5473 harbors cDNA (representing TEV genome nucleotides 5414-7285) which encompasses the entire 49K coding region as well as flanking coding regions (Fig. 1A). Previously, we have shown that translation of pTL-5473 transcripts results in efficient synthesis of a 75,000 mol wt precursor and processing to a 49,000 mol wt product with immunological and biochemical properties similar to authentic 49K nuclear inclusion protein (Carrington and Dougherty, 1987). To determine if processing at the 49K borders occurs by autoproteolysis or by another viral or reticulocyte lysate component, an amino acid substitution at the proposed (Argos et al., 1984; Allison et a/., 1986) active site of the 49K protease was introduced. The triplet (TGT) specifying cysteine at residue position 339 of the 49K protease was converted, by site-directed mutagenesis of cDNA in pTL-5473, to a triplet (GCC) specifying alanine (Fig. 1A). Nucleotide sequencing of wild-type (pTL-5473) and mutagenized (pTL-5473/C-A) cDNA in the vicinity of the predicted alteration confirmed that a coding change from cysteine to alanine had been incorporated (Fig. 1 B). Transcripts from pTL-5473 and two pTL-5473/C-A clones, each possessing the coding potential for a 75,000 mol wt product, were translated in the reticulocyte lysate system in the presence of [35S]methionine. Translation of pTL-5473 transcripts yielded the 49K product (Fig. 2, lane 1) as shown previously(Carrington and Dougherty, 1987). However, translation of the pTL5473/C-A transcripts resulted in accumulation of a
1 2 3
FIG. 2. Expression of pTL-5473 and two pTL-5473/C-A isolates. Plasmid DNA was linearized with Pvull (which cleaves within the vector sequence) and transcribed with SP6 polymerase. The uncapped transcripts were translated in a rabbit reticulocyte lysate in the presence of [35S]methionine. The products were separated on a SDS-polyacrylamide (5% stacking, 12.5% resolving) gel and detected by fluorography. Translation products were synthesized from transcripts of pTL-5473 (lane 1) and two pTL-5473/C-A plasmids (lanes 2 and 3). Apparent molecular weights (Xl O-3) are shown on the left,
358
CARRINGTON
75,000 mol wt product which was not processed to the 49K form (Fig. 2, lanes 2 and 3; these lanes show translation products of transcripts from two apparently identical plasmids which were selected during the mutagenesis procedure). The faint, 35S-labeled polypeptides underneath the 75,000 mol wt product are believed to represent premature translational termination products since further incubation (3 hr) did not result in their accumulation (data not shown). Processing containing
in tm~s of synthetic 49K cleavage sites
AND
DOUGHERTY
A VPg Pro
VP4,
87 K
2
50 K
.
70 K
c’
~~:~:;:::::~~~:~~q(~)
O-S O-G
jljjlllllijiijjljjjlIlii.ii:iitilllllili
O-S O-G
pTL-5468
P;G
:::::::::::y.:.::.:.::. .:.‘,j::::::::::::::.:::.::::::::::I::: ..:..........:. :,.::.:.:.:..
pTL-5473(C-A)
polyproteins
In an earlier report (Carrington and Dougherty, 19871, we have shown that processing in trans was inefficient at the amino-terminal cleavage site of the 49K protein in the synthetic polyprotein encoded by pTL-5468 transcripts (see diagram in Fig. 3A). This polyprotein harbors amino acid sequences for TEV protein totalling approximately 12,000 mol wt upstream from the cleavage junction, followed by a 41,000 mol wt polypeptide representing the amino-terminal segment of the 49K protease. The 53,000 mol wt translation product does not undergo autoproteolysis since the carboxy terminus of the protease, shown to be essential for enzymatic activity (Carrington and Dougherty, 1987), has been deleted. This poor substrate activity using 49K protease from purified nuclear inclusions could result from aberrant structural conformation due to deletion of the carboxy terminus or from an inherent property of the cleavage site. To test the hypothesis that the entire 49K protein is required for processing in Vans at its borders, cleavage reactions were performed on synthetic polyprotein substrates produced by translation of transcripts from pTL-5468, pTL-5473/C-A, and pTL6895. The 75,000 mol wt product of pTL-5473/C-A transcripts serves to test both 49K cleavage sites on a polyprotein which contains the complete 49K sequence with the exception of the Cys to Ala substitution at position 339. The transcripts synthesized from pTL6895 encode a polyprotein of approximately 95,000 mol wt consisting of an -6000 mol wt fragment from the carboxy terminus of the 49K protease and the entire 54K and 30K (capsid) sequences (Fig. 3A). The 95,000 mol wt polyprotein (from pTL-6895) serves as a substrate to analyze cleavage at the carboxy terminus of the 49K protein as well as the 54K-30K cleavage junction [which is “efficiently” processed in tram (Carrington and Dougherty, 1987)l. Transcripts from each plasmid were translated in vitro in the presence of [35S]methionine and tested for their substrate activity using the 49K protease from nuclear inclusion bodies. As in the prior study (Carrington and Dougherty, 1987), the 53,000 mol wt product of pTL5468 transcripts was processed poorly in Vans, yielding
6
1234567
FIG. 3. Trans processing of three polyproteins containing the aminoand/or carboxy-terminal cleavage sites for the 49K protease. (A) Schematic representation of cDNAs expressed by in vitro transcription and translation. The shadings and symbols are the same as described in Fig. 1. The asterisk below the pTL-5473 diagram indicates the position of the Cys to Ala amino acid substitution in pTL-5473/C-A. (6) Substrate activities of synthetic polyproteins, encoded by transcripts of cDNAs shown in (A), using the processing activity of the 49K protease within purified nuclear inclusion bodies. The %-labeled polyproteins synthesized in the reticulocyte lysate system (30 pl) were divided into equal portions: one-half was incubated with 49K protease (as detailed under Materials and Methods), while the other half was incubated with H20. Products were analyzed by SDS-polyacrylamide gel electrophoresis and fluorography. Lane I, pTL-5473 derived proteins; lanes 2 and 3. pTL-5473/C-Aderived products reacted without (lane 2) or with (lane 3) protease; lanes 4 and 5, pTL-5468 derived products reacted without (lane 4) or with (lane 5) protease; lanes 6 and 7, pTL-6895 derived products reacted without (lane 6) or with (lane 7) protease. The arrows indicate the products (60,000 mol wt in lane 3; 41,000 mol wt in lane 5) resulting from inefficient ffans cleavage at the amino terminus of the 49K protease (see Results). The molecular weights (Xl Om3) and positions of Y-labeled protein markers are shown at the left.
minor amounts of a 41,000 mol wt product (Fig. 3B, lanes 4 and 5), whereas the polyprotein encoded by pTL-5473 transcripts underwent efficient autoproteolysis (lane 1). Microsequence analysis of the amino terminus of the 41 K product revealed that (inefficient) processing in trans occurred at the same site as (efficient) autoproteolysis (see below).
SELF-PROCESSING
Under identical processing conditions as with the pTL-5468 derived product, the 75,000 mol wt polypeptide encoded by pTL-5473/C-A transcripts also was processed with poor efficiency, yielding only minor mol wt product and little or amounts of an -60,000 no accumulation of a 49,000 mol wt species (Fig. 3B, lanes 2 and 3). Although amino acid sequence analysis was not performed on the -60,000 mol wt cleavage product, we assumed that this protein represented an inefficient proteolytic event at the amino terminus of the 49K sequence, as with the pTL-5468 cleavage product. Little or no proteolysis was detected at the carboxy-terminal cleavage site of the 49K protein using the pTL-5473/C-A derived polyprotein substrate. This cleavage site also was tested using the synthetic substrate polyprotein encoded by pTL-6895 transcripts. Processing in Vans of the 95,000 mol wt polyprotein resulted in complete disappearance of the substrate and accumulation of a 30,000 mol wt protein (capsid) and an -65,000 mol wt product (lanes 6 and 7). Significantly, processing failed to yield a 54K product which would have been the result of cleavage at the 49K54K junction. The positive control for 49K protease processing activity is shown by the “efficient” release of 30K capsid protein from the 95,000 mol wt precursor encoded by pTL-6895 transcripts (lanes 6 and 7). In time course analyses, cleavage in trans at the 30K capsid amino terminus occurred at a rate at least 50 times faster than at the 49K protease amino terminus (data not shown). Thus, excision of the 49K protease from the TEV polyprotein appears to require an autocatalytic mechanism in vitro. Amino acid sequence synthesized in vitro
analysis
of 49K protease
To identify the site at the amino terminus of the protease where autoprocessing takes place, radiolabeled 49K protein synthesized in vitro from pTL-5473 transcripts was analyzed by microsequence techniques. The amino terminus of authentic 49K protein isolated from nuclear inclusion bodies is blocked (unpublished observations), thereby preventing direct amino acid sequence analysis. The 49K translation product synthesized from pTL-5473 was labeled by incorporation of [35S]methionine or [3H]leucine, purified by electroelution from a SDS-polyacrylamide gel, and applied to an automated sequencer. The radioactivity released during cycles l-20 for each labeled sample are shown in Fig. 4. An unambiguous peak of [35S]methionine was released at position 1 1, whereas a substantial yield of [3H]leucine was identified at residue position 9. The tripeptide Leu-Xaa-Met exists at only one position within the deduced amino acid sequence encoded by pTL-5473 transcripts, corresponding to residues 1858-
OF
TEV
359
PROTEASE
1860 on the 346K TEV polyprotein. Therefore, autoproteolysis at the amino terminus of the 49K protease is predicted to occur between the dipeptide glutamineglycine at positions 1849-l 850, located nine residues prior to the leucine identified by sequence analysis. To determine if the low degree of trans processing observed when the pTL-5468 derived protein was reacted with the viral protease (yielding a 41 K polypeptide) was in fact occurring at the actual amino terminus of the 49K sequence (Fig. 3B, lanes 4 and 5) the 41K product was subjected to microsequence analysis as described above. Based on the positions (relative to the amino terminus) at which either [35S]methionine or [3H]leucine was released during the automated sequencing run (Fig. 4) the limited cleavage in trans was occurring at the same dipeptide (Gln-Gly) as cleavage during autocatalysis of intact 49K protease precursor. To confirm the accuracy of processing mediated by the 49K protease in vitro, a partial amino acid sequence was determined for the 30K capsid protein generated during a cell-free processing reaction. The sequence at the amino terminus of authentic TEV capsid protein has been determined (Allison et a/., 1985a,b). A 34,000 mol wt protein encoded by transcipts from pTL-8595 (cDNA representing genome nucleotides 8462-9495), which contains coding sequences for a short segment of the 54K protein and the entire 30K capsid coding region (see Fig. 1A), serves as an useful substrate to examine 49K protease-mediated cleavage in a trans assay. In the presence of active 49K protease, the 34,000 mol wt substrate is converted to a 30,000 mol wt product of the same size and immunological reactivity as authentic capsid protein (Carrington and Dougherty, 1987). The 34,000 mol wt product was synthesized in vitro in the presence of [3H]alanine, reacted with protease present within TEV nuclear inclusion bodies, and subjected to preparative SDS-gel electrophoresis. The 30,000 mol wt cleavage product was electroeluted from the gel and applied to the automated sequencer. Peaks of [3H]alanine were released during cycles 6, 8, and 10 (Fig. 4). This combination of alanine residues exists at only one position within the pTL-8595 coding region (corresponding to amino acid residues 2797, 2799, and 2801 in the TEV polyprotein), therefore indicating cleavage occurred between the glutamine-serine dipeptide at positions 279 l-2792. Indeed, this site is the same as that predicted from sequence analysis of authentic capsid protein (Allison et al., 1985a,b). DISCUSSION Thiol proteases use a free sulfhydryl group from cysteine as a nucleophile during catalysis of peptide bond cleavage (Lowe, 1976). Based on studies involving in-
360
CARRINGTON 49K (Protease) GKKNPKHKLKHREARGARGQ
AND
DOUGHERTY
4 I K (Partial
49K)
4Fj
30K (Capsid)
Residue
Residue
Residue
FIG. 4. Partial amino acid sequence analysis of TEV products derived by translation and processing reactions in vitro. The 49K (protease) and 30K (capsid) proteins were labeled individually with [%]methionine, [3H]leucine, or [3H]alanine, while the 41 K (partial 49K) product was colabeled with [36S]methionine and [3H]leucine. Radiolabeled proteins were eluted from SDS-polyacrylamide gels and applied to an automated sequencer (Applied Biosystems). Radioactivity released at each cycle was quantitated using a Beckman scintillation counter with channels set to discriminate between % and 3H. The amino acid sequence presented above each plot was deduced from the genome nucleotide sequence (Allison et a/., 1986). Radioactivity applied to the automated sequencer were as follows: %-49K (protease), 453,000 cpm; 3H-49K (protease). 850,000 cpm; 3H-30K (capsid), 128,000 cpm; 35S-41 K (partial 49K), 103,000 cpm; 3H-41 K (partial 49K), 24,000 cpm.
hibitors, chemical requirements for activity, and amino acid sequence homologies, most groups of singlestranded RNA viruses which utilize proteolytic processing as the predominant mode of gene expression are thought to encode thiol proteases (Pelham, 1978, 1979; Gorbalenya and Svitkin, 1983; Peng and Shih, 1984; Argos et al,, 1984; Yeh and Gonsalves, 1985; Lloyd et al., 1986; Werner et al., 1986). To test the importance of the cysteine residue (339) hypothesized to function at the active site of the TEV 49K protease (Allison et al., 1986), alanine was introduced at position 339 in place of cysteine. Synthetic polyproteins expressed from transcripts derived from DNA harboring this point mutation failed to process; thus, we conclude (i) the 49K protease is excised from the polyprotein via an autocatalytic mechanism, and (ii) the hypothesis that cysteine (339) exists at the active center of the protease is supported, although not proven. If the 49K cleavage sites were processed by a reticulocyte lysate or cellular component, as proposed for the primary cleavage event of the cowpea mosaic virus B-RNA translation product (Tian and Shih, 1986) cleavage would most likely have been unaffected by a single amino acid substitution distant to the processing sites. It has been shown or suggested that intramolecular or autoproteolytic cleavages to release a protease occur at the early stages of viral polyprotein processing
for many viruses. In addition, it is often suggested that free viral protease, excised from the polyprotein precursor by autocatalysis, then functions by an intermolecular mechanism to release additional protease molecules from subsequently synthesized polyproteins (for example, Pelham, 1979; Palmenberg and Reuckert, 1982; Hanecak et a/., 1984; Toyoda et a/., 1986). Our data suggest that this is probably not the case for TEV, since cleavage in Vans at the 49K protease boundaries occurred poorly in the reactions studied here. Since autoprocessing at each 49K protease border occurs with high efficiency in vitro (see Figs. 2 and 3) and presumably in v&o, there may be little need for trans processing to function. It is interesting, however, to consider why neither junction functions efficiently as a substrate for exogenous protease. Although clues are available through analysis of the amino acid sequence (see below), the precise sequence or protein folding requirements for the TEV substrate sites are not yet known. It is possible that cleavage sites destined for autoprocessing adopt an orientation different from those destined for Vans processing. Presumably, the autoproteolytic orientation results in efficient presentation of the cleavage site to the active center of the protease. If true, conformational switching (prior to proteolysis) at the 49K protease (or other TEV protein) boundaries might be possible such that timing of pro-
SELF-PROCESSING
tein maturation throughout the infection cycle could be regulated. Although most of the known potyviral-encoded proteins have been isolated from infected plants (Knuhtsen et a/., 1974; Dougherty and Hiebert, 1980b; Hari, 1981; Thornbury et a/., 1985; de Mejia et a/., 1985), little information is available regarding in vivo synthesis and maturation of the polyprotein or regulation of these events. The autoprocessing cleavage site which defines the amino terminus of the 49K protease was determined using protein derived in cell-free systems (Fig. 4). The validity of this approach was confirmed by demonstrating accurate cell-free processing by the 49K protease at the known 54K-30K protein cleavage junction. The location of the 49K amino-terminal cleavage site has been predicted correctly in a previous study based on approximate size of the protein and on the presence of an amino acid sequence motif shared with the known cleavage site (at the 54K-30K junction) (Carrington and Dougherty, 1987). At positions -6, -3, -2, and -1 relative to the cleaved peptide bond, the cleavage site at the amino terminus of the 49K protease and 30K capsid share Glu, Tyr, Phe, and Gln, respectively. When this comparison is extended to three additional predicted cleavage sites (at the boundaries separating 50K-70K, 70K-6K, and 49K-54K proteins, see Fig. 1A) which we hypothesize are processed by the 49K protease, the Glu (-6), Tyr (-3), and Gln (-1) residues are strictly conserved (Carrington and Dougherty, 1987). Additionally, hydrophobicity profiles in the immediate vicinity of the cleavage sites are strikingly similar (data not shown), suggesting they may share a similar local three-dimensional structure as well (Sweet and Eisenberg, 1983). Therefore, information critical for su’bstrate activity and specificity may reside within a cluster of amino acid residues in a specific orientation surrounding the cleavage site. Through site-directed mutagenesis, the relative contribution of these residues may be evaluated. ACKNOWLEDGMENTS We thank Mr. Ronnie Johnson for consistent technical assistance, MS. Mary Kelly for conducting the protein sequencing runs, and Ms. Dawn Parks and Dr. Dennis Hruby for their critical reviews of this manuscript. This research was supported by a Postdoctoral Fellowship from the North Carolina Biotechnology Center (to J.C.C.) and Grant DBM-8601939 from the National Science Foundation (to W.G.D.). Oregon Agricultural Experiment Station Technical Paper No. 8254.
REFERENCES ALLISON, R. F.. DOUGHERTY, W. G.. PARKS, T. D., WILLIS, L., JOHNSTON, R. E., KELLY, M.. and ARMSTRONG, F. B. (1985a). Biochemical analysis of the capsid protein gene and the capsid protein of tobacco etch virus: N-terminal amino acids are located on the virion’s surface. Virology 147, 309-3 16.
OF
TEV
361
PROTEASE
ALLISON, R. F., JOHNSTON, R. E., and DOUGHERP/, W. G. (1986). The nucleotide sequence of the coding region of tobacco etch virus genomic RNA: Evidence for the synthesis of a single polyprotein. Virology 154, 9-20. ALLISON, R. F., SORENSON, J. G., KELLY, M. E., ARMSTRONG, F. B., and DOUGHERIY, W. G. (1985b). Sequence determination of the capsid protein gene and flanking regions of tobacco etch virus: Evidence for the synthesis and processing of a polyprotein in potyvirus genome expression. Proc. Natl. Acad. Sci. USA 82, 3969-3972. ARGOS, P., KAMER, P., NICKLIN, M. J. H., and WIMMER, E. (1984). Similarity in gene organization and homology between proteins of animal picornaviruses and a plant comovirus suggest common ancestry of these virus families. Nucleic Acids Res. 12, 7251-7267. CARRINGTON, J. C., and DOUGHERTY, W. G. (1987). Small nuclear inclusion protein encoded by a plant potyvirus genome is a protease. J. Viral., 81, in press. DE MEJIA. M. V. G., HIEBERT, E., and PURCIFULL, D. E. (1985). Isolation and partial characterization of the amorphous cytoplasmic inclusions associated with infections caused by two potyviruses. Virology 142, 24-33. DOUGHERTY, W. G. (1983). Analysis of viral RNA isolated from tobacco leaf tissue infected with tobacco etch virus. Virology 131, 473481. DOUGHERN, W. G., and HIEBERT, E. (1980a). Translation of potyvirus RNA in a reticulocyte lysate: Reaction conditions and identification of capsid protein as one of the products of in vitro translation of tobacco etch and pepper mottle viral RNAs. Virology 101, 466474. DOUGHERTY, W. G., and HIEBERT, E. (1980b). Translation of potyvirus RNA in a reticulocyte lysate: Identification of nuclear inclusion proteins as products of tobacco etch virus RNA translation and cylindrical inclusion protein as a product of the potyvirus genome. Virology 104, 174-l 82. DOUGHERPI, W. G., and HIEBERT, E. (198Oc). Translation of potyvirus RNA in a reticulocyte lysate: Cell-free translation strategy and a genetic map of the potyviral genome. Virology 104, 183-l 94. GORBALENYA, A. E., and SVITKIN, Y. V. (1983). Protease of encephalomyocarditis virus: Purification and role of sulfhydryl groups in processing of the structural proteins precursor. Biokhimiya 48, 442-453. HANECAK, R., SEMLER, B. L., ARIGA, H., ANDERSON, C. W., and WIMMER, E. (1984). Expression of a cloned gene segment of poliovirus in E. co/i: Evidence for autocatalytic production of the viral proteinase. Cell 37, 1063-l 073. HARI, V. (1981). The RNA of tobacco etch virus: Further characterization and detection of protein linked to RNA. Virology 112, 391399. HOLLINGS, M., and BRUNT, A. A. (1981). Potyviruses. In “Handbook of Plant Virus Infection and Comparative Diagnosis” (E. Kurstak. Ed.), pp. 731-807. Elsevier/North Holland Biomedical, New York. KELLY, C., Tom, N. F., WATERFIELD. M. D.. and CRUMPTON, M. J. (1983). Amino acid sequencing of polypeptides eluted from 2D polyacrylamide gels. Biochem. lnt. 6, 535-544. KNUHTSEN, H., HIEBERT, E., and PURCIFULL. D. E. (1974). Partial purification and some properties of tobacco etch virus intranuclear inclusions. Virology 61, 200-209. LAEMMLI, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227, 680-685. LLOYD, R. E., TOYODA, H., ETCHISON, D., WIMMER, E., and EHRENFELD, E. (1986). Cleavage of the cap binding protein complex polypeptide ~220 is not effected by the second poliovirus protease 2A. Virology 150,299-303. LOWE, G. (1976). 302.
The cysteine
proteinases.
Tetrahedron
32, 291-
362
CARRINGTON
AND
MAXAM, A. M., and GILBERT, W. (1980). Sequencing end-labeled DNA with base-specific chemical cleavages. ln “Methods in Enzymology” (L. Grossman and K. Moldave, Eds.), Vol. 65, pp. 499-560. Academic Press, New York. MELTON, D. A., KRIEG, P. A., REBAGLIATI, M. R., MANIATIS, T., ZINN, K., and GREEN, M. R. (1984). Efficient in vitro synthesis of biologically active RNA and RNA hybridization probes from plasmids containing a bacteriophage SP6 promoter. Nucleic Acids Res. 12,7035-7056. PALMENBERG, A. C., and RUECKERT. R. R. (1982). Evidence for molecular self-cleavage of picornaviral replicase precursors. J. 41,244-249. PELHAM, H. R. B. (1978). Translation of encephalomyocarditis RNA in v&o yields an active proteolytic processing enzyme. J. Biochem. 85, 457-462. PELHAM, H. R. B. (1979). Synthesis and proteolytic processing cowpea mosaic virus proteins in reticulocyte lysates. Virology 463-477. PENG, X. X., and SHIH. D. S. (1984). Proteolytic proteins translated from the bottom component mosaic virus. f. Biol. Chem. 259, 3197-3201.
intraViral. virus Eur. of 96,
processing of the RNA of cowpea
PIRONE, T. P., and THORNBURY, D. W. (1983). Role of virion and helper component in regulating aphid transmission of tobacco etch virus. Phytopathology 73, 872-875. SIMONS, J. N. (1976). Aphid transmission of a non-aphid transmissible strain of tobacco etch virus. Phytopathology 66, 652-654. SWEET, R. M., and EISENBERG, D. (1983). Correlation of sequence
DOUGHERTY hydrophobicities measures similarity in three-dimensional protein structure. 1. Mol. Biol. 171, 479-488. TAYLOR, J. W., OTT, J., and ECKSTEIN, F. (1985b). The rapid generation of oligonucleotide-directed mutations at high frequency using phosphorothioate-modified DNA. Nucleic Acids Res. 13, 87658785. TAYLOR, J. W., SCHMIDT, W., COSSTICK, R., OKRUSZEK, A., and ECKSTEIN, F. (1985a). The use of phosphorothioate-modified DNA in restriction enzyme reactions to prepare nicked DNA. Nucleic Acids Res. 13, 8749-8764. THORNBURY, D. W., HELLMANN, G. M., RHOADS, R. E., and PIRONE, T. P. (1985). Purification and characterization of potyvirus helper component. Virology 144, 260-267. TIAN, Y. C., and SHIH, D. S. (1986). Cleavage of a viral polyprotein by a cellular proteolytic activity. I. Viral. 57, 547-55 1. TOYODA, H., NICKLIN, M. J. H., MURRAY, M. G.,ANDERSON. C. W., DUNN, J. J., STUDIER, F. W.. and WIMMER, E. (1986). A second virus-encoded proteinase involved in proteolytic processing of poliovirus polyprotein. Cell 45, 761-770. WERNER, G., ROSENWIRTH, B., BAUER, E., SEIFERT, J.-M., WERNER, F.-J., and BESEMER, J. (1986). Molecular cloning and sequence determination of the genomic regions encoding protease and genomelinked protein of three picornaviruses. J. Viral. 57, 1084-l 093. YEH, S.-P., and GONSALVES, D. (1985). Translation of papaya ringspot virus RNA in vitro: Detection of a possible polyprotein that is processed for capsid protein, cylindrical-inclusion protein and amorphous-inclusion protein. Virology 143, 260-271.