VIROLOGY

190, 884-888

Effects

(1992)

of Site-Directed

Mutagenesis on the Presumed Catalytic Grapevine Fanleaf Nepovirus 24-kDa ROG~RIO

MARGIS*+

AND

LOTHAIRE

*Institut de Biologie Moltkulaire des Plantes du CNRS et Universik? Louis Pasteur, Strasbourg, France; and tEscola Tkcnica federal de Qui’mica do Rio de Janeiro, Received

May

13, 1992; accepted

Triad and Substrate-Binding Proteinase

Pocket

of

PINCK*~’

Laboratoire de Virologie. 12 rue du G&&al Zimmer, Rua Senador Furtado 12 1, 20270 Rio de Janeiro-R/, June 22,

67084 Brazil

1992

Grapevine fanleaf nepovirus (GFLV) has a bipartite plus-sense RNA genome. Its structural and functional proteins originate from polyprotein maturation by at least one virus-encoded proteinase. Here we describe the cloning of the 24-kDa proteinase cistron located between the virus-linked protein (VPg) and the RNA-dependent RNA polymerase cistron in GFLV RNA1 (nucleotides 3966 to 4622). Proteinase expressed from this clone is able to cleave GFLV polyprotein P2 in order to produce the coat protein and a 66-kDa protein which is further processed to the 38-kDa presumed movement protein. The GFLV 24-kDa proteinase sequence contains sequence similarities with other nepovirus and comovirus proteinases, particularly at the level of the conserved domains corresponding to the hypothetical catalytic triad and to the substrate-binding pocket (amino acids 192 to 200). Site-directed mutagenesis of residues His43, Glu”, and Leuig7 abolished proteinase activity. Inactivation of the enzyme is also observed if the catalytic residue cys”9 was substituted by isoleucine, but replacement by a serine at the same position produced a mutant with an activity identical to that of native proteinase. All our data show that GFLV cysteine proteinase presents structure similarities to the proteinases of cowpea mosaic virus and potyviruses but is most closely related to trypsin. o 1992 Academic

Press, Inc.

Virus-encoded proteinases play a fundamental role in the replication of many animal and plant viruses. The synthesis and processing of polyproteins is a characteristic of positive-strand RNA viruses and retroviruses (1). At least five plant virus groups with single-stranded positive-sense RNA use polyproteins as a strategy of genome expression. Three of these, the potyviruses, the comoviruses, and the nepoviruses, are grouped in the picorna-like superfamily of plant viruses (2) because of their resemblance with the animal picornaviruses. They contain certain common structural features including a virus-linked protein (VPg) covalently attached to the 5’ terminus and a poly(A) tract at the 3’ terminus (3). The viral cysteine proteinases from the picornavirus family and the plants coma- and potyvirus group are best studied with data available from sequence”comparison analysis (4-8), inhibitor studies (9, IO), and site-directed mutagenesis (1 l- 16). The picornaviridae and the related groups of plant viruses encode proteinases of approximately 24 kDa. Histidine, aspartic or glutamic acid, and cysteine are the amino acid residues present at the putative catalytic site (6, 8, 1). Grapevine fanleaf virus (GFLV), a member of the nepovirus group, possesses a bipartite positive-strand RNA genome (17). GFLV RNA-l encodes a polyprotein

’ To whom 0042-6822/92

reprint

requests

should

$5.00

CopyrIght 0 1992 by Academic Press, Inc. All rlghts of reproduction WI any form reserved.

PI of 253 kDa, containing the putative viral helicase, the VPg (genome-linked viral protein), a 24-kDa proteinase, and a putative RNA-dependent RNA polymerase (18, 19). GFLV RNA-2 codes for polyprotein P2 of 122 kDa (20), composed of three proteins: the coat protein at its carboxy-terminus, a 38-kDa putative movement protein, and a 28.kDa amino terminal protein of unknown function (Margis et al,, manuscript in preparation). A sequence alignment between the GFLV 24-kDa proteinase and the proteinases of two other nepoviruses, tomato black ring virus (TBRV) and grapevine chrome mosaic virus (GCMV), and of a comovirus, cowpea mosaic virus (CPMV), is shown in Fig. 1. This alignment resulted from a Bestfit analysis between each pair of sequences and from a multiple Pileup analysis of the four sequences (21). The identity (I) and similarity (S) index between the GFLV proteinase and those of the two nepoviruses and a comovirus, calculated by the Bestfit analysis were very similar: GFLV versus TBRV (I = 23.50/o, S = 48.0%); GFLV versus GCMV (I = 25.90/o, S = 49.7%); GFLVversus CPMV(I = 23.60/o, S = 45.5%). Nevertheless, TBRV and GCMV are more closely related to one another (I = 67.8%, S = 81.3%) than to GFLV. It should be noted that these values may have to undergo slight revisions once the exact limits of the TBRV and GCMV proteinases have been determined. The alignments previously reported (6, 8) concerned only TBRV as representative of the

be addressed. 884

SHORT

COMMUNICATIONS

885

GFW: ..... TBR” : Lxwf: CPHV :

. . . . . . . . . .

HSLDQSS”AIWSKCR...ANL”.....

GCMl:HAYH..

.

FIG. 1. Sequence comparison between GFLV 24.kDa proteinase and that of two nepoviruses, TBRV and GCMV, and a comovirus, CPMV. The amino- and carboxy-ends of the TBRV and GCMV proteinase sequence were arbitrarily chosen since the cleavage sites producing these proteinase are unknown. The aligned sequences are numbered according to GFLV proteinase, with the glutamic acid (residue 1242 of GFLV polyprotein Pl) taken as the first residue. All the amino acid residues from one of the three other viruses presenting identity with the GFLV are boxed. Amino acid residues (H, E, and C) corresponding to the presumed GFLV proteinase catalytic triad and to the catalytic triad of CPMV are in bold lettering. GFLV proteinase mutageneized residues are labeled (#)

nepoviruses. The present alignment shows that the amino acid residues of the proteinase catalytic triad proposed by Gorbalenya et al. (8) for the CPMV have their counterpart in GFLV, i.e., His43, ASHES, and Cys17’ (Fig. 1). Bazan and Fletterick (7) proposed an aspartate as a putative component of the TBRV proteinase catalytic triad, which in our alignment corresponds to GFLV Lys’O6. GFLV-F13 proteinase activity was previously characterized using the clone pVP7. This clone encompassed GFLV RNA1 from nucleotides 3894 to 4789, cloned in the BarnHI site of the BlueScribe (PBS+) phagemid. Translation of pVP7 transcripts in the rabbit reticulocyte lysate produced an active proteinase (19), but the protein synthesized corresponds to a set of three genes (VPg, proteinase, and the amino terminus of the RNA polymerase) plus extraviral nucleotides until the stop codon. The complete amino acid sequence of the GFLV VPg allowed us to identify Gly’24’/GIu”42 (polyprotein Pl numbering) as the cleavage site between the VPg and the proteinase (22). The last amino acid residue present at the proteinase carboxy extremity is either Arg1460 or GIY’“~’ as deduced from maturation studies with pVP7 and its partial transcripts, with Arg 1460/Gly’46’ or Gly’46’/GIu’462 as the two possible cleavage sites between the proteinase and the polymerase (19). From these data, a new GFLV-F13 clone (pPro7) was constructed so as to contain exclusively the proteinase gene ending with Arg1460. The region corresponding precisely to the proteinase cistron (GFLV-F13 RNA1 nt 3966 to 4022) was amplified and mutagenized at its extremities by PCR mutagenesis as previously described (19). The GFLV-F13 RNA1 full-

length clone pFL148 was used as template. The mutagenic primer P3494 (5’CCCGGGATCCTATGAAAAATTTCTCCATGGAGGGACCTTCCAAAGG3’)which hybridizes to nt 3966 to 3982 of RNAl, introduced an in-phase start codon, the leader sequence of GFLV satellite RNA3, and three cloning sites (Smal, BarnHI, and /Vcol) upstream of the proteinase cistron. Primer P3496 (5’CCCGGGATCCI I I I I I I I I I I I I I I I I I I I ITATCTAACTACCTAATAAAAGAAGAACTG3’) which hybridizes to nt 4004 to 4022 of RNA1 , introduced a stop codon after the Arg’460 codon, a 22nucleotidelong poly-A tail and two cloning sites (BarnHI and Smal). The PCR-amplified fragment, corresponding to the mutagenized proteinase gene, was digested by BarnHI and cloned into the BarnHI site of PBS’. Proteinase transcripts were obtained by bacteriophage T7 RNA polymerase transcription of HindIll-digested pPro7. The wild-type proteinase produced by in vitro translation of the pPro7 clone was able to cleave GFLV P2 polyprotein in a similar way to that observed for the pVP7 clone when processing translation products of total GFLV RNAs (19). The major cleavage products of polyprotein P2 are the 66-kDa intermediate, the 56kDa coat protein (CP), and a 38-kDa protein (Fig. 2, lane 12). To investigate the mechanism of action of GFLV 24kDa proteinase and to confirm the predicted catalytic site obtained by the sequence alignment (Fig. l), various mutations were introduced into the pPro7 clone. Mutagenesis was conducted by the method of Kunkel (23). Positive clones were subsequently amplified and screened by endonuclease digestion when a new restriction enzyme site was created, and by di-

886

SHORT

116

COMMUNICATIONS

P2

85

66 kDa

55

CP

39

38 kDa 12

3 4 5 6 7

8 9 10111213

FIG. 2. Proteinase activity of different pPro7 mutants measured by their ability to process GFLV polyprotein P2 in trans. Each lane was loaded with the same amount of radioactive mixture. Autoradiography of a 10% polyacylamide gel containing SDS after 24 hr exposure at room temperature. Positions of molecular weight markers (in kDa) are indicated to the left. The major translation products from GFLV RNA-2, the polyprotein P2 and its maturation products, the 66.kDa protein, the capsid protein (CP), and the 38.kDa protein are indicated to the right. TR2 corresponds to the translation product of the GFLVF13 RNA2 full-length clone.

deoxynucleotide sequencing with a T7 sequencing kit (Pharmacia) across the proteinase region flanking the target site to confirm the presence of the specific mutation. Table 1 shows the sequence of the mutagenic primers used, the region of their hybridization in GFLVF13 RNA-l, the nature of the mutation and the proteinase activity of each mutant estimated according to the amount of 66-kDa, 38-kDa, and coat protein produced. Mutations were carried out on the candidate residues of the putative catalytic triad: His43, GluE7, Asplog, TABLE MUTATIONS

Primer number P3763: P4084: P4073: P3764: P4745: P3531: P4083: P4074: P3762: P4075:

Sequence

of mutagenic (5’ + 3’)

INTRODUCED

IN pPro7

Asp”‘, and Cys17’. Furthermore, the hypothetical glycosylation site Asn213 was replaced by a tryptophan and two point mutations were also performed in the putative substrate-binding pocket: Leulg7 and Ser”‘. The ability of the aforesaid proteinase mutants to process the P2 polyprotein and to produce the mature products was compared for each construct (Fig. 2). The tryptophan substitution in pProN213W mutant did not change the proteinase activity (Fig. 2, lane 11). However, pProLlg7H in which a residue of the putative substrate-binding pocket was altered displayed no activity(Fig. 2, lane 9). In mutant pProS”‘A, the serine-toalanine substitution at position 199 creates in GFLV the Val-Ala-Gly consensus present in the proteinases of TBRV, GCMV, and some other picornaviruses (6). The resulting proteinase had slightly reduced activity compared to the native form and the ratio between the 66- and 38-kDa maturation products was altered (Fig. 2, lane 10). The mutants pProH43D and pProH431containing alterations in the putative histidine residue of the catalytic triad were unable to process P2 polyprotein (Fig. 2, lanes 2 and 3). Three residues with acidic lateral chains (GIua7, Asp”‘, and Asp”O), which are present in the central region of the proteinase and are candidates to be part of the catalytic triad, were mutated in order to test their importance in the enzyme activity. The pProDIOgN mutant (Fig. 2, lane 5) was as active as the wild-type Pro7 proteinase (Fig. 2, lane 12). ProD1loI mutant displays a very low activity as observed by the presence of the 66-kDa cleavage product (Fig. 2, lane 6). Residue GluE7seems to be indispensable for proteinase activity because mutation ProE87A produced a nonfunctional form of the protein1 BY SITE-DIRECTED

Position of primer on GFLV-F13 RNA-l

primer

CTCACTCTTGATGACCCGGGATCAAGCCCTGGC CACTCTTGATGACGAGAA-JCAAGCCCTGGCCG CGAATTCCCTGAAAATGCAT’I-AGlTGTCTTl-GAGCATCC TGAGAAGTAClTTG‘TT&GATGCCGATAGACA GAGAAGTACTfTGTTGATflGCCGATAGACAGATTTC GAAGCAAAGAAGTATGATTTGGAGCCTTGGCTG GAAGCAAAGAAGTATGATaGGAGCCTTGGCTG CCCAAAAGTCATCGCAATGCACGTGTCTGGGAATAGAGGTG CGCAATGCTTGTTGCAGGGAATAGAGGTGTG CCTACTClTCTGTGATTCCA~TACAGlTCTTCTTlTATT

a Codons affected by mutagenesis are underlined. * Aal”Aa2 means that residue Aal at position “n” of GFLV-F13 ‘As function of the amount of 66-kDa, 38-kDa, and coat protein

MUTAGENESIS

4073-4105 4075-4107 4208-4246 4273-4306 6580-6617 4482-4515 4482-4515 4535-4575 6852-6882 4579-4619

Pro7 proteinase produced.

has been

Amino acid substitution’

Mutant code

His?Asp His?le GluE7Ala Asp’O’Asn Asp”?le Cys’7sSer Cys17911e Leu”‘His Ser’ssAla Asn213Trp

pProH43D pProH431 pProEa7A pProD”‘N pProD”? pProC’% pProC’7sl pProL”‘H pProS’99A pProN2’3W

changed

into a residue

Aa

Proteinase activity +++ L!z +++ f ++

SHORT

COMMUNICATIONS

ase (Fig. 2, lane 4). The cysteine at position 179 is almost certainly the catalytic residue of the GFLV 24kDa cysteine proteinase since this cysteine residue is found near the carboxy terminus as part of the sequence Asp-Cys-Gly (Fig. l), similar to the Asp-SerGly consensus of members of chymotrypsin super-family of serine proteinases (6). Based upon the mechanism of action of such enzymes, cysteine 179 should be involved in peptide bond cleavage by forming an acyl-enzyme intermediate with tetrahedral geometry, as proposed for the serine residue of the catalytic triad of trypsin (24). If Cys17’ was replaced by an unreactive residue such as isoleucine (mutant ProC17gl), enzyme activity is totally abolished (Fig. 2, lane 7). But if it is replaced by a reactive nucleophile such as serine (mutant ProC’7gS), the enzyme remains fully active (Fig. 2, lane 8). A particularity of the GFLV cysteine proteinase is its ability to recognize a typical trypsin cleavage site, as typified by the maturation process of P2 polyprotein, where cleavage at an Arg605/Gly606 site yields the 66kDa protein and the coat protein (19). This finding suggests the existence of determinants shared between trypsin and the GFLV proteinase. However, the site selection by GFLV proteinase is evidently more complex than that of trypsin, since not all Arg/Gly or Arg/X (with X being any amino acid other than Gly) sites are cleaved. It has been proposed that substrate specificity was due to a complex multidomain structure located at the carboxy-terminus rather than to a ‘short segment of amino acids (16). In our experiments, two single amino acid substitutions were produced in the putative substrate binding region: Leulg7 was substituted by His and Set-“’ by Ala. Leulg7 is a conserved residue in all the three nepovirus sequences presented here (Fig. l), as is His356 rn potyviruses (9) and in picornaviruses (6). The His356 residue has been proposed to play an important role in recognition of a Gln residue at position Pl of the substrate (6, 9). Nepoviruses have basic amino acids residues at the substrate Pl position: arginine in the case of GFLV and GCMV (20, 25), and lysine in the case of TBRV (26). GFLV proteinase is inactivated if Leulg7 is replaced by the histidine residue present in the poty- and picornavirus proteinases. Presumably the histidine residue introduced in the ProLlg7H mutant destabilizes the enzyme-substrate interaction in contrast to the situation with potyviruses (9). The role of acidic residues in the catalytic action of serine proteinases is controversial. The Asplo residue in trypsin is present in the catalytic triad and has a critical role in the stabilization of the appropriate His57 tautomer required for catalysis (27). A second aspar-

887

tate (Asp”‘), at the bottom of the Sl binding pocket, is implicated in the preference of trypsin for Lys or Arg residues (28). Our mutations of the GIua7 and Asp”’ residues produced, respectively, an inactive and a poorly active form of the 24-kDa proteinase, indicating that the GIrY7 residue may in fact participate in the catalytic triad, perhaps to promote the orientation and stabilization of the viral proteinase His43 in a manner similar to that of the Asplo residue of trypsin. The second acidic residue (Asp”‘) could be implicated in the binding of Mg2+ ions by coordination, in a similar fashion to the aspartic and glutamic acid residues present in the loop region of the helix-loop-helix motif characteristic of calcium-binding proteins (29) since the GFLV proteinase has an absolute requirement for Mg2+ ions (19). The existence of a loop placed between two ,6 strands (D-E) and potentially able to bind Ca2+ ions has previously been proposed (6). In contrast to other viral cysteine proteinases, the cysteine-to-serine substitution produced no reduction of activity when compared to the native enzyme (Fig. 2, lane 8). The possibility of a replacement of the cysteine by a serine at the enzyme active site, presumably reflects a random evolutionary event. The results presented here reinforce the model of viral cysteine proteinases as a class of trypsin-like enzymes. In the plant virus groups, mutations of the cysteine residue corresponding to position 147 of poliovirus type 1 (PVl) 3C proteinase produced different results depending on whether cleavage was studied in trans or in cis (30). This shows that mutation analysis of viral proteinase should take into account the existence of inter- as well as intramolecular processing. Here, as a first approach, we have chosen to study mutant 24kDa proteinase activity in a trans cleavage model, due the simplicity of the GFLV P2 polyprotein cleavage products compared to the Pl polyprotein pattern. Furthermore, study of proteinase mutants in Pl polyprotein maturation is rendered difficult because not all cleavage products have been identified so far. It is evident that the importance of these mutations should be analyzed in viva, particularly the viability of the GFLV mutant containing a serine (position 179) as the catalytic residue of the 24-kDa proteinase. Concerning the pProN213W mutant in which the putative N-glycosylation site was removed, no difference was observed in the in vans cleavage of P2 polyprotein between mutant and wild-type proteinase. The in viva action of this mutant merits investigation, since precise Pl polyprotein maturation might be dependent on a glycosylated form of the proteinase. Our next steps in investigation of the GFLV 24-kDa proteinase will be characterization of polyprotein cleavage sites, studies with artificial substrates and enzymological kinetic analysis,

888

SHORT

COMMUNICATIONS

ACKNOWLEDGMENTS Rogerio Margis was supported by a grant from the National Council of Research and Technology (CNPq, Brazil). We thank Dr. M. Pinck for helpful discussion and Dr. K. Richards for improving the manuscript. The EMBL Databank accession number for grapevine fanleaf virus RNA 1 is DO091 5:GFVRNAl.

REFERENCES 1. HELLEN, C. U. T., KRAUSSLICH, H., and WIMMER, E., Biochemistry 28, 9881-9888 (1989). 2. GOLDBACH, R., EGGEN, R., DE JAGER, C., and VAN KAMMEN, A., Ceil Biol. 41, 147-l 62 (1990). 3. MA-HEWS, R. E. F., “Plant Virology,” (3rd ed.) Academic Press, San Diego, 1991. 4. FRANSSEN, H., LEUNISSEN, J., GOLDBACH, R.. LOMONOSSOFF, G., and ZIMMERN, D., EMBO 1. 3, 855-861 (1984). 5. DOMIER. L. L., SHAW, J. G., and RHOADS, R. E., virology 158, 20-27 (1987). 6. BAZAN, J. F., and FLETTERICK, R. J., Proc. /Vat/. Acad. Sci. USA 85, 7872-7876 (1988). 7. BAZAN, J. F., and FLETTERICK, R. J., FEBS Left. 249, 5-7 (1989). 8. GORBALENYA, A. E., DONCHENKO, A. P., BLINOV, V. M., and KOONIN, E. V., FfBS Lett. 243, 103-l 14 (1989). 9. DOUGHERTY, W. G., PARKS, T. D., CARY, S. M., BAZAN, J. F., and FLE~ERICK, R. J., Virology 172, 302-310 (1989). 10. Yu, S. F., and LLOYD, R. E., virology 182, 615-625 (1991). 11. IVANOFF, L. A., TOWATARI, T., RAY, J., KORANT, B. D., and PETEWAY, S. R., Proc. Natl. Acad. Sci. USA 83, 5392-5396 (1986). 12. CARRINGTON, J. C., and DOUGHERTY, W. G., Virology 160, 355362 (1987). 13. BAUM, E. Z., BEBERNITZ, G. A., PALANT, O., MUELLER, T., and PLOTCH, S. J., virology 185, 140-l 50 (1991).

14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

28. 29. 30.

DESSENS, J. T., and LOMONOSSOFF, G. P., virology 184,738-746 (1991). KEAN, K. M., TETERINA. N. L., MARC, D., and GIRARD, M.. Virology 181,609-619 (1991). PARKS, T. D.. and DOUGHERTY, W. G., Virology 182, 17-27 (1991). PINCK, L., FUCHS, M., PINCK, M., RAVELONANDRO, M., and WALTER, B., J. Gen. Viral. 69, 233-239 (1988). RITZENTHALER, C., VIRY, M., PINCK, M., MARGIS, R., FUCHS, M., and PINCK, L., /. Gen. Viral. 72, 2357-2365 (1991). MARGIS, R., VIRY, M., PINCK, M., and PINCK, L., Virology 185, 779-787 (1991). SERGHINI, M. A., FUCHS, M., PINCK, M., REINBOLT, J., WALTER, B., and PINCK, L., /. Gen. Viral. 71, 1433-1441 (1990). DEVEREUX, J., HAEBERLI, P., and SMITHIES, O., Nucleic Acids Res. 12, 387-395 (1984). PINCK, M., REINBOLT, J., LOUDES, A. M., LE RET, M.. and PINCK, L., FEBSLett. 284, 117-119(1991). KUNKEL, T. A., Proc. Nat/. Acad. Sci. USA 82, 488-492 (1985). KRAUT, J., Annu. Rev. Biochem. 46, 331-358 (1977). BRAULT, V., HIBRAND, L., CANDRESSE, T.. LE GALL, O., and DUNEZ, J., Nucleic Acids Res. 17, 7809-7819 (1989). DEMANGEAT, G., HEMMER, O., FRITSCH, C., LE GALL, O., and CANDRESSE, T., /. Gen. Viral. 72, 247-252 (1991). SPRANG, S., STANDING, T., FLETTERICK, R. J., STROUD, R. M., FINERMOORE, J., XUONG, N.-H., HAMLIN, R., RUTTER, W. J., and CRAIK, C. S., Science 237, 905-909 (1987). SHOTHON, D. M., and WATSON, H. C., Narure 225, 811-814 (1970). STRYNADKA, N. C. J., and JAMES, M. N. G., Annu. Rev. Biochem. 58, 951-998 (1989). LAWSON, M. A., and SEMLER, B. L., Proc. Nat/. Acad. Sci. USA 88, 9919-9923 (1991).