THE JOURNAL OF BIOLOGICAL CHEMISTRY LC‘, 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol ,268, No. 3, Issue of January pp. 204&2051,1993 Printed in U.S.A.

Identification of the Herpes Simplex Virus-1 Protease Cleavage Sites by Direct Sequence Analysis of Autoproteolytic Cleavage Products*

(Received for publication, August 31, 1992)

Carolyn L. DiIanniSS, Diana A. Drier$, Ingrid C. Deckman$, Patrick J. McCann III$, Fenyong Liun, Bernard Roizmanll, Richard J. Colonno$, and Michael G. Cordingleyz From the $Department of Virology, Bristol-Myers Squibb Pharmaceutical Research Institute, Princeton, New Jersey 08543-4000 and the ’1TMurjorie B. Kouler Viral Oncology Laboratories, University of Chicago, Chicago, Illinois 60637

Herpes simplex virus type- 1 (HSV- 1) encodes a pro- tease responsible for proteolytic processing of the virus assembly protein, ICP35 (infected cell protein 35). The coding region of ICP35 is contained within the gene that encodes the protease, and ICP35 shares amino acid identity with the carboxyl-terminal 329 amino acids of the protease. The HSV-1 protease was ex- pressed in Escherichia coli as a fusion protein contain- ing a unique epitope and the protein A Fc binding domain at its carboxyl terminus. The fusion protease underwent autoproteolytic cleavage at two distinct sites. The size of the cleavage products containing the carboxyl-terminal epitope mapped one cleavage site near the carboxyl terminus of the protease correspond- ing to the proteolytic processing site of ICP35, and the second site proximal to the amino terminus consistent with previous data. The carboxyl-terminal autoproteo- lytic cleavage products were partially purified on an IgG affinity column by virtue of the protein A Fc binding domain and subjected to direct amino-terminal sequence analysis. Protein sequencing revealed that cleavage occurs between the Ala and Ser residues at amino acids 610161 1 and 2471248 of the HSV-1 pro- tease. The flanking sequences share homology with each other and are highly conserved in homologous proteases of other herpes viruses.

Infected cell protein 35 (ICP35)’ of HSV-1 is a family of proteins that is a major component of viral B-capsids found in the nucleus of infected cells but is absent from mature virions (1-3). Consequently, ICP35 and its homologues in other herpesviruses have been called virus assembly proteins and are proposed to play a role analogous to that of the scaffolding protein of bacteriophages (4). ICP35c,d is the product of the UL26.5 gene of HSV-1 and undergoes proteo- lytic processing to the lower molecular weight species, ICP35e,f, that lack approximately 20 amino acids at the carboxyl terminus (5). An HSV-1 ts mutant that is defective in the processing of ICP35 at the nonpermissive temperature

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom reprints should be addressed. Tel.: 609-252-4135; Fax: 609-252-6058,

The abbreviations used are: ICP35, infected cell protein 35; CMV, cytomegalovirus; C,F, carboxyl-terminal product of the carboxyl- t.erminal cleavage of the fusion protease; HSV-1, herpes simplex virus type 1; PAGE, polyacrylamide gel electrophoresis; PraF, protease fusion protein; NaF, carboxyl-terminal product of the amino-terminal cleavage; HPLC, high pressure liquid chromatography.

also fails to package viral DNA, and only aberrant empty capsids accumulate (6). The proteolytic processing of the virus assembly protein therefore appears to be critical for virus particle maturation. The protease responsible for these pro- cessing events is encoded in the 635-amino acid open reading frame of UL26, which is 3’ coterminal with UL26.5, the gene encoding ICP35 (7). The 329-amino acid open reading frame encoding ICP35 is in frame with that of the protease (7) and as a consequence, the protease shares its carboxyl-terminal domain with ICP35 and is capable of autoproteolytic process- ing of its own carboxyl terminus at a site identical to that of its substrate. We have recently demonstrated that Escherichia coli-expressed HSV-1 protease undergoes autoproteolytic processing at the carboxyl-terminal site that it shares with ICP35 and at an additional site, proximal to the amino terminus, that is unique to the protease (8).

In this report we identify the two proteolytic processing sites within the protease by direct amino acid sequencing of the isolated proteolytic products. This was achieved by con- structing a HSV-1 protease fusion gene (PraF) encoding the complete HSV-1 protease fused at its carboxyl terminus with a 20-amino acid epitope that is recognized by monoclonal antibody CH28-2 (5) and 253 amino acids from the Fc binding domain of staphylococcal protein A. The resulting carboxyl- terminal products are easily identified by the unique epitope and purified by affinity chromatography exploiting the Fc binding domain of protein A. This construct has been used previously in studies to characterize the proteolytic processing of the UL26 gene product expressed in reticulocyte lysates (5). In addition, a mutant HSV-1 protease containing substi- tution of the His at position 148 with an Ala (H148A), which results in a protein that does not undergo proteolytic process- ing (9), was utilized to verify that the cleavage products generated in this construct arise from autoproteolytic proc- essing.

MATERIALS AND METHODS

Constructs-The protease fusion protein was expressed in E. coli in the vector p E T l l d (Novagen) under the control of the inducible T7 promoter using established protocols (10). The construct was created by inserting a 2745-base pair XbaI-EcoRI fragment of pRB4214 (5) into the NheI-EcoRI sites of the vector. The resulting vector, pT74214PRT, encodes a fusion protein containing 2 amino acids of the vector, the upstream 7 amino acids of the UL26 gene, followed by the start methionine and coding region of HSV-1 protease, the 20-amino acid epitope recognized by monoclonal antibody CH28- 2 (5), and the Fc binding domain of staphylococcal protein A. The mutant fusion protein H148A was cloned from pRB8077 (9) utilizing an identical strategy.

Antibodies-Monoclonal antibody CH28-2 reacts specifically with a 20-amino acid CMV epitope and was obtained from B. Roizman and L. Pereira (5).

Protein Purification-Bacterial cells from a 15-liter culture ex-

2048

Identification of the HSV-1 Protease Cleavage Sites 2049

Full Length Fusion Protease (RaF)

1 1

WT H148A r 0 1 3 5 " 0 1 3 5 'Hours

PraF- i

-1 06

1-m Na Fusion (NaF)

FIG. 1. Schematic of the HSV- 1 protease fusion protein a n d location of predicted autoproteolytic cleavage products. Pro- tease sequences are represented by the unshaded rectangle; the black rectangle indicates the location of the 20-amino acid epitope recog- nized by monoclonal antibody CH28-2; and the checked rectangLe represents 253 amino acids of the protein A Fc binding domain. The arrows indicate proteolytic processing sites within the protease pro- tein (8). Two carboxyl-coterminal cleavage products, NaF and C,F, are illustrated.

pressing the fusion protein were suspended in 500 ml of 10 mM sodium phosphate (pH 7.2), 500 mM NaCl and lysed by pulsed sonication on ice. Cell debris was pelleted by centrifugation a t 28,400 X g a t 4 "C for 20 min. The resulting supernatant was incubated with 15 ml of rabbit IgG-agarose (Sigma) by rocking for 18 h a t 4 "C. Elution of the NaF and C,F species was achieved by washing the column with 100 mM glycine (pH 2.4) and 150 mM NaC1. Concentra- tion of this material was achieved by dialysis against 5 mM Hepes (pH 7.0), 0.5 mM EDTA, 1 mM dithiothreitol and subsequent lyoph- ilization. Further purification of NaF was accomplished by HPLC on a reverse phase C, column (4.6 mm (inner diameter) X 25 cm, Vydac) eluting with a 0-100%/100-min gradient of acetonitrile in 0.1% trifluoroacetic acid.

Sequence Analysis-Concentrated fractions containing C,F and NaF were resolved by SDS-PAGE and transferred to Immobilon (Millipore) for direct sequence analysis. In order to minimize chemical blocking of the amino terminus, 12.5% acrylamide resolving gels were polymerized for 18 h a t 4 "C and prerun using SDS-PAGE running buffer containing 0.1 mM thioglycolic acid (11) for 1 h a t 8 mA a t 25 "C. Samples were electrophoresed a t 30 mA (constant current) for 8 h a t room temperature in SDS-PAGE running buffer without thioglycolic acid and then transferred (50 V) to Immobilon (Waters) in Laemmli buffer with 20% methanol for 18 h a t 30 V, 4 "C. Sequence analyses were performed by the Edman degradation method.

RESULTS

DNA encoding the wild-type HSV-1 fusion protease or the H148A mutant fusion protease was cloned into a T7 vector for expression in E. coli (see "Materials and Methods"). The HSV-1 protease fusion protein and its predicted cleavage products are illustrated in Fig. 1. The expression of PraF and H148A proteins was monitored by immunoblotting with monoclonal antibody directed against the unique carboxyl- terminal epitope (Fig. 2). In each case, full-length protein was detected within 1 h of induction (Fig. 2); however, only the wild-type protease underwent autoproteolysis to generate 80- and 33-kDa species, NaF and C,F, respectively, that accu- mulate to high levels a t later time points. While it cannot be ruled out that some NaF is generated by the mutant protease, i t is most likely that these minor species represent degradation products consistently observed with this mutant protein.* Additional immunoreactive species only accumulated to a low level following induction of either form of the protease (Fig. 2).

The carboxyl-coterminal cleavage products, NaF and C,F, were generated by autoproteolysis of the full-length precursor protein at each of the two predicted cleavage sites within the protease domain of the fusion protein, and each contains the

* I. C. Deckman and P. McCann, unpublished data.

NaF-

CoF-

-80

-50

-33

-28

FIG. 2. The t ime course of the induction of E. coli expressing the wi ld type and H148A mutant fusion proteins. Expression of the wild type ( W T ) and H148A fusion proteins was analyzed by gel electrophoresis (12.5% acrylamide/bis gel) and probing immuno- blots with the monoclonal antibody CH28-2. Aliquots of culture were removed a t 0, 1, 3, and 5 h postinduction as indicated.

Fc binding domain of protein A at the carboxyl terminus (Fig. 1). This domain facilitated simple purification of the carboxyl- terminal cleavage products by affinity chromatography on an IgG column. E. coli cells expressing the fusion protein were harvested 2-5 h after induction to ensure the presence of the carboxyl-terminal cleavage products. The cells were lysed by sonication, cell debris was removed by centrifugation, and the resulting soluble supernatant was subjected to affinity chro- matography on an IgG-agarose column. NaF and C,F were identified in the column fractions by SDS-PAGE, immuno- blotting with the monoclonal antibody directed to the unique carboxyl-terminal epitope, and staining with Coomassie Bril- liant Blue. Ninety-seven percent of the total protein flowed through the column while NaF and C,F were bound, and eluted a t low pH (see "Materials and Methods"). Coomassie staining of the proteins in eluted fractions (Fig. 3A, lane 2 ) revealed that C,F could be sufficiently resolved by gel electro- phoresis to enable direct sequencing. In contrast, the lower yield of NaF and existence of several contaminating species with similar molecular weights made further purification of this species imperative. IgG-purified material containing NaF (Fig. 3A, lune 2 ) was therefore subjected to reverse phase HPLC (Fig. 3B). Fractions containing NaF were effectively resolved from contaminating species as a discrete peak of protein (Fig. 3B, fraction 58) in which NaF was the major staining species (Fig. 3A, lane 3 ) . This material served as a source of NaF for sequence analysis.

The preparations of C,F and NaF were concentrated, sub- jected to electrophoresis, and then transferred to Immobilon. C,F and NaF were identified by Coomassie staining, excised, and subjected to amino-terminal sequence analysis. The re- sulting amino-terminal sequence of the C,F cleavage fragment was compared with the predicted amino acid sequence of the protease as deduced from the nucleotide sequence of the UL26 gene (12) and found to align with sequences from amino acids 611 to 622 (Table I). Only 25 pmol of Ser was detected in

2050 Identification of the HSV-1 Protease Cleavage Sites

A

1 2 3

28-

-

1 -

I -

i 0.0 "y-y 10 20 30 40 50 SO 70 80 90 100

Fnctkn N u m b

B

I

FIG. 3. Purification of C,F and NaF. A, Coomassie-stained proteins following purification by affinity chromatography on IgG-agarose (lane 2 ) and subsequently by C, reverse phase HPLC column ( l a n e 3); lane I contains molecular weight markers. B, chromatogram of absorbance at 280 nm of the C, reverse phase HPLC purification of NaF. Fraction 58 from the C, purification, which contains primarily NaF, is shown in lane 3.

TABLE I Amino-terminal sequencing of CoF and NaF by Edman degradation

Each sequence is the result of two determinations. CoF NaF

Cycle Amino acid Predicted Amino acid Predicted seauence seauence

1 Ser Ser-611 Ser Ser-248 2 Ser Ser-612 Glu Glu-249 3 Ala Ala-613 LY s Lys-250 4 Ala Ala-614 Phe Phe-251 5 His His-615 LYS Lys-252 6 Val and Tyr Val-616 Met Met-253 7 Asp Asp-617 Trp Trp-254 8 Val andTyr Val-618 GlY Gly-255 9 Asp Asp-619 Ala Ala-256

10 Ala Thr-620 Glu Glu-257 11 Ala Ala-621 Pro Pro-258 12 Arg Arg-622 Val Val-259

cycles 1 and 2, whereas 400 pmol of Ala was detected in cycles 3 and 4. This inconsistency is most likely due to the low recovery of the phenylthiohydantoin derivative of Ser in the sequenator. Both Val and Tyr were present in cycles 6 and 8, although Val is the predicted amino acid at both positions. DNA sequencing has indicated that Val is the encoded amino acid at these positions in the fusion protease gene con~ t ruc t .~ The reason for this discrepency remains unclear. In cycle 10, Ala was observed instead of the expected Thr; however, only low amounts of material were available at this point during the sequencing. From these data we conclude that the car- boxyl-terminal cleavage that generates C,F occurs between Ala-610 and Ser-611.

The sequence determined for the amino-terminal 12 amino acids of NaF unambiguously matched the predicted amino acid sequence of the protease from amino acid 248 to 260

' I. C. Deckman and P. J. McCann 111, unpublished data.

(Table I), indicating conclusively that the amino-terminal cleavage occurred between Ala-247 and Ser-248.

DISCUSSION

The HSV-1 protease has been expressed in E. coli as a fusion protein that undergoes autoproteolytic cleavage to gen- erate characteristic cleavage products. A mutant protease in which His-148 was changed to Ala failed to undergo proteo- lytic processing, confirming that the protease itself is respon- sible for the observed cleavages. The size and immunoreactiv- ity of the cleavage products that accumulated are consistent with the previous observation (8) that two processing sites exist in the HSV-1 protease, one close to the carboxyl termi- nus and the other proximal to the amino terminus (Fig. 1). Amino-terminal sequence analysis of the carboxyl-terminal proteolytic cleavage products, C,F and NaF, mapped the location of the carboxyl- and amino-terminal cleavages, re- spectively. The carboxyl-terminal cleavage occurred between Ala-610 and Ser-611, 25 amino acids from the carboxyl ter- minus of the protease, and the amino-terminal cleavage oc- curred between Ala-247 and Ser-248. The sequences sur- rounding the two cleavage sites share homology with one another and are conserved a t analogous positions within the homologues of the protease in other herpesviruses (Table 11).

Examination of the amino acid sequences flanking the processing sites of herpesvirus proteases reveals that a con- sensus sequence can be derived (Table 11) in which P3, P1, and P1' are strongly conserved in both the amino- and car- boxyl-terminal cleavage sites. These data strongly suggest that the cleavage sites for all of the herpesvirus proteases will occur between the conserved Ala and Ser residues (Pl/Pl'). Downstream residues do not appear to be conserved except that P2' is usually small. Interestingly, the P4 position at the carboxyl-terminal site is conserved as a small aliphatic resi- due, while the equivalent position at the amino-terminal site is occupied by a highly conserved Tyr in all of the sequences

Identification of the HSV-1 Protease Cleavage Sites 2051

TABLE I1 Sequence homology between herpes virus proteases

Sequences are from GenBank and Refs. 12-14. Amino acids in boldface type are highly conserved. VZV, varicella zoster virus; CMV, human cytomegalovirus strain AD 169; EBV, Epstein-Barr virus; ILTV, infectious laryngotracheitis virus.

Carboxyl-terminal cleavage

604 Ala Gly Ala Leu Val Asn Ala Ser Ser Ala Ala H i s Val Asp 617 572 Asp Val Asn Ala Val Glu Ala Ser Ser Lys Ala Pro Leu I l e 585 675 Gln Ala Gly Val Val Asn Ala Ser Cys Arg Leu Ala Thr Ala 688 562 Gly Lys Lys Leu Val Gln Ala Ser Ala Ser Gly Val Ala Gln 575 582 Ala Arg Glu Thr Val Asp Ala Ser Met Pro Lys Arg Leu Lys 595

P7 P6 P 5 P4 P3 P2 P1 P1' P2' P3' P4' P5' P6' P7'

Amino-terminal cleavage

241 Gly His Thr Tyr Leu Gln Ala Ser Glu Lys Phe Lys Met Trp 254 230 Gly H i s Val Tyr Leu Gln Ala Ser Thr Gly Tyr Gly Leu Ala 243 288 Arg Glu Ser Tyr Val Lys Ala Ser Val Ser Pro Glu Ala Arg 301 229 Ala Glu Ser Tyr Leu Lys Ala Ser Asp Ala Pro Asp Leu Gln 242 263 Asn Pro Lys Tyr Leu Gln Ala Asn Glu Val I l e Thr I l e Gly 276

HSV - 1 vzv CMV EBV ILTV

HSV-1 vzv CMV EBV ILTV

examined. This most likely reflects differences in the struc- tural requirements for the two cleavage sites. Structural fea- tures outside P1 and P1' are clearly important for cleavage site recognition since cleavage is not observed at the five other naturally occurring Ala/Ser sites within the HSV-1 protease.

The carboxyl-terminal cleavage site of the HSV-1 protease is also present at the carboxyl terminus of ICP35, which shares sequence with the protease and is a substrate in in- fected cells (5, 7). Proteolytic processing of the ICP35 family of virus assembly proteins (ICP35c,d) to lower molecular weight forms (ICP35e,f) is essential for successful virus mat- uration (6). The carboxyl-terminal cleavage event itself or the newly cleaved ICP35 must play a critical role in scaffold assembly or subsequent steps in virus maturation. Interest- ingly, the CMV assembly protein (15), which is the homologue of ICP35, undergoes an analogous processing event near its carboxyl terminus (16, 17). Recent studies utilizing mass spectroscopy of the fragmented assembly protein, purified from immature CMV particles, have located the cleavage event in this protein between Ala-557 and Ser-558 (18). As indicated in Table 11, this region shares a high degree of homology with the carboxyl-terminal HSV-1 protease cleav- age site. Based upon this sequence homology, Welch and colleagues (18) proposed the existence of an amino-terminal cleavage site within the CMV protease. Our data provides direct experimental evidence that the HSV-1 protease is pro- teolytically cleaved at this site and suggests that this is a common feature in the maturation processing of herpesvirus proteases. The role of the amino-terminal cleavage, which occurs within sequences of the protease not shared with ICP35, is not clear. It has been suggested that the amino- terminal product of cleavage at an equivalent site in the CMV protease results in release of active protease (18). This species does indeed appear to retain proteolytic activity (18); however, a requirement for this event in activation of the protease has not been shown. We have obtained data that the amino- terminal fragment of the HSV-1 protease (amino acids 1-247) and the corresponding carboxyl-terminal fragment processed

at the carboxyl-terminal cleavage site (amino acids 248-610) are incorporated into viral B-capsids along with processed forms of the assembly protein, ICP35e,f.4 A role for this second cleavage in the assembly or maturation of B-capsids and subsequent encapsidation of viral DNA therefore cannot be excluded.

Additional studies of the protease and its activities both in vivo and in vitro will lead to elucidation of its precise role in virus maturation. The obligate requirement for proteolytic processing of the herpesvirus assembly proteins and the iden- tification of the protease as virus-encoded reveal that this protease is an exciting new antiviral target.

Acknowledgments-We thank Dr. M. Flocco at Princeton Univer- sity for the sequence analysis. We thank Dr. S. P. Weinheimer for many helpful discussions.

REFERENCES 1. Gibson, W., and Roizman, B. (1972) J. Virol. 10, 1044-1052 2. Gibson, W., and Roizman, B. (1974) J. Virol. 13, 155-165 3. Perdue. M. L.. Kemo. M. C.. Randall. C. C.. and O'Callaehan. D. J. (1974)

Virology 59', 201-*216 '

. . I I . ,

4. Showe, M. K., and Black, L. W. (1973) Nat. New Biol. 242.70-75 5. Liu, F., and Roizman, B. (1991) J. Virol. 6 5 , 5149-5156 6. Preston, V. G., Coates, J. A. V., and Rixon, F. J. (1983) J. Virol. 4 5 , 1056-

1 nGA 7. Li;;F, and Roizman, B. (1991) J. Virol. 6 5 , 206-212 8. Deckman, I. C., Hagen, M., and McCann, P. J., 111 (1992) J. Virol. 66,

9. Liu, F., and Roizman, B. (1992) Proc. Natl. Acad. Sci. U. S. A. 8 9 , 2076- 7362-7367

mm 10. Stidre-r, F. W., Rosenberg, A. H., Dunn, J. J., and Dubendorff, J. W. (1990)

11. Moos, M., Jr., Nguyen, N. Y., and Liu, T. Y. (1988) J. Biol. Chem. 263 , Methods Enzymol. 185,60-89

finnR-Gnnn 12. McGeoch, D. J., Dalrymple, M. A,, Davison, A. J., Dolan, A,, Frame, M. C.,

McNab, D., Perry, L. J., Scott, J. E., and Taylor, P. (1988) J. Gen. Virol. 69.1521-1574

"" ""

", "" - 13. Davison, A. J., and Scott, J . E. (1986) J. Gen. Virol. 67, 1759-1816 14. Griffin, A. M. (1990) Nucleic Acids Res. 18, 3664 15. Irmiere, A., and Gibson, W. (1985) J. Virol. 5 6 , 277-283 16. Gibson, W., Marcy, A. I., Comolli, J. C., and Lee, J. (1990) J. Virol. 6 4 ,

17. Schenk, P., Woods, A. S., and Gibson, W. (1991) Nat. New Biol. 242 , 70-

18. Welch, A. R., Woods, A. S., McNally, L. M., Cotter, R. J., and Gibson, W.

1241-1249

75

(1991) Proc. Natl. Acad. Scr. U. S. A. 88,10792-10796

S. P. Weinheimer and D. O'Boyle, personal communication.