- Email: [email protected]
Molecular Interactions between the HSV-1 Capsid Proteins as Measured by the Yeast Two-Hybrid System
VIROLOGY
220, 516–521 (1996) 0341
ARTICLE NO.
SHORT COMMUNICATION Molecular Interactions between the HSV-1 Capsid Proteins as Measured by the Yeast Two-Hybrid System PRASHANT DESAI and STANLEY PERSON1 Department of Molecular Genetics and Biochemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15261 Received January 26, 1996; accepted April 12, 1996 HSV-1 B capsids are composed of seven major proteins, designated VP5, VP19C, 21, 22a, VP23, VP24, and VP26. VP indicates that the capsid protein is also a component of the infectious virion. Capsid proteins 21, 22a, and VP24 are specified by a single open reading frame (UL26) that encodes 635 amino acids. An objective of the work in our laboratory is to identify and map interactions among and between capsid proteins. In the present studies we employed the yeast GAL4 two-hybrid system developed by Fields and his colleagues (Nature 240, 245–246 (1989)) for this purpose. DNA corresponding to the capsid open reading frames was derived as a PCR product and fused to sequences of the GAL4 activation and DNA binding domains. Using this system each of the capsid proteins has been tested for interactions with all of the other capsid proteins. Three interactions have been identified: a relatively strong self-interaction between 22a molecules (residues 307– 635 of UL26), bimolecular interactions between 22a and VP5, and another between VP19C and VP23. The interactions were detected by the expression of b-galactosidase enzyme activity, and yielded 289, 86, and 63 units of enzyme activity, respectively. For the 22a self-interaction, elimination of residues 611–635 resulted in an approximately twofold decrease in enzyme activity. The C-terminal 25 amino acids of 22a were also essential for the bimolecular interaction between 22a and VP5. q 1996 Academic Press, Inc.
Three types of capsids have been identified in herpes simplex virus type 1 (HSV-1)-infected cells designated A, B, and C corresponding to increasing distance sedimented through sucrose gradients. B capsids contain a centrally located (scaffold) protein that is replaced with viral DNA in the more rapidly sedimenting C capsids. The slowest sedimenting A capsids are devoid of the scaffold protein and DNA, and may result from abortive packaging events. The capsid shells of HSV-1 are comprised of four proteins, designated VP5, VP19C, VP23, and VP26 (1–3). The major capsid protein, VP5, comprises the hexon and penton subunits of the icosahedral capsid shell, VP19C and VP23 connect the subunits, and VP26 is found on the tips of the hexons (4–6). The remaining components of B capsids, 21, 22a, and VP24, are encoded by a single open reading frame (ORF), UL26, whose organization is depicted at the top of Fig. 1. The UL26 gene product encodes an autocatalytic protease (7, 8) which cleaves the UL26 gene product between residues 247 and 248 to generate VP24 (1–247) and 21 (248–635) (9–11). The more abundant capsid protein,
22a (307–635) is translated in the same reading frame from a second, shorter transcript (7, 12). For HSV-1 a second cleavage site between amino acids 610 and 611 of UL26 generates products referred to as the mature forms of the proteins, as opposed to the full-length, or precursor forms, representative of the primary translation products (8–10). 22a and to a lesser extent 21 occupy the inner capsid space and act to form a scaffold upon which the icosahedral shell is constructed (1–3, 13). The human and simian-cytomegalovirus homologues of UL26 appear to have the same biological activities as those of HSV-1 and have been extensively characterized (see for example, 14–16). Gibson et al. (17) were the first to suggest that this protein may serve as a scaffold for capsid shell assembly in a manner analogous to bacteriophage capsid assembly. The proteins of the UL26 ORF may be thought of as maturational as well as structural capsid proteins, as opposed to the four proteins that comprise the capsid shell that seem to have only a structural role in assembly. There are probably a number of other viral-specified proteins with roles in capsid maturation that may be present in capsids in low copy numbers and may not be components of completed virions. By analogy to bacteriophage assembly these include DNA packaging and associated docking proteins, proteins that bend DNA and proteins that plug the hole used for DNA packaging. Recently the
1 To whom correspondence and reprint requests should be addressed at current address: Virology Laboratories, Department of Pharmacology and Molecular Sciences, Johns Hopkins University School of Medicine, 725 N. Wolfe Street, Baltimore, MD 21205. Fax: (410) 9553023.
0042-6822/96 $18.00
516
Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
AID
VY 7959
/
6a18$$$101
05-16-96 19:14:14
vira
AP: Virology
SHORT COMMUNICATION
517
FIG. 1. Characterization of capsid protein interactions. The UL26 ORF is shown at the top of A. The position of the autoproteolytic cleavages are indicated between residues 247 and 248 and between 610 and 611. Cleavage after residue 247 generates VP24 and 21 from the UL26 translation product. 22a is translated from a second transcript whose initiation codon is at position 307 of the UL26 ORF. The protein coding regions are shown as horizontal lines and relevant terminal amino acid residues are marked at the ends of the lines. Replacement cloning and complete nucleotide sequence analysis were not performed for the 22a mutant constructs, although the termini were shown to be correct by DNA sequencing. In B, the VP5 coding region was cloned into pGBT9 and 22a into pGAD424. In C, VP19C was cloned in pGBT9 and VP23 in pGAD424. DNA amplification and cloning was as indicated in the legend to Table 1. The transformation of Y526 and subsequent measurement of enzyme activity were also as described in the legend to Table 1. Enzyme units { one standard deviation are shown at the left of each construct and are the averages of three determinations.
UL6 gene product was shown to be a capsid protein and may be representative of one of these proteins (18). The purpose of the present experiments was to detect molecular interactions among the capsid proteins of HSV-1 using the yeast two-hybrid system (19–21), which employs the yeast transcriptional activator, GAL4, characterized by Ptashne and his colleagues (see for example, 22–24). The GAL4 protein consists of a DNA binding domain (residues 1–147) and a transcriptional activation domain (residues 768–881), which are distinct and separable (22). Transcriptional activation of a gene by GAL4
requires that both domains be in close proximity in order to interact with the basal transcriptional complex. Normally the two domains are joined by covalent bonds but Fields and his colleagues showed that weak interactions between proteins fused to each domain also resulted in transcription. Transcriptional activation was measured using a lacZ reporter gene fused to a GAL4 responsive promoter (GAL1). Two plasmids were employed; one, pGBT9, encodes the DNA binding domain and the other, pGAD424, specifies the GAL4 activation domain. Polylinkers in these plasmids allow the insertion of test pro-
AID
vira
VY 7959
/
6a18$$$102
05-16-96 19:14:14
AP: Virology
518
SHORT COMMUNICATION
tein sequences in phase with, and downstream from, the respective GAL4 domains. pGBT9 also encodes the TRP1 gene and pGAD424 encodes the LEU2 gene so that following transformation of appropriate yeast cells, colonies may be selected that grow in the absence of leucine and tryptophan. Colonies can be assayed directly for enzyme activity by the appearance of blue color using X-GAL as a substrate (19). Sequences encoding each capsid protein of HSV-1 were obtained as PCR products and cloned into both plasmids. Constructs representing the entire UL26 ORF (residues 1–635), and 22a (residues 307–635), were used in this assay. Wherever possible the capsid proteins were tested when present in both vectors. Most experiments utilized Y526 yeast cells. The Y526 cell contains the lacZ reporter sequence fused to the upstream activating and TATA sequences of the GAL1 promoter (19). Following transformation, colonies that developed on plates containing selective growth media were assayed for the development of blue color. The results are summarized in Table 1. A total of three interactions, designated by a /, were detected; a bimolecular interaction between VP5 and 22a (and with UL26), another between VP23 and VP19C, and a self-interaction between 22a molecules. Blue color, in all of the colonies, was detectable after 20 min of incubation for the 22a self-interaction and after approximately 40 min for the other interactions. No interactions were detected for VP24 and VP26, either with themselves or with the other capsid proteins. The enzyme units of b-galactosidase expressed as nanomoles of ONPG hydrolyzed per minute per milligram of total protein were determined for each interaction and are listed in the shaded squares of the Table. An interaction between cellular p53 and SV40 T-antigen produced approximately 150 units of enzyme activity when employed in the two-hybrid system (25), a value that we also obtained when these constructs were used as a positive control. The self interaction between 22a scaffold molecules resulted in a similar level of enzyme activity (289 units). The two vectors without inserts resulted in 0.4 { 0.3 units of activity (standard deviations are not listed in Table 1 but are given in Fig. 1). Therefore, values greater than this value, or approximately 1 unit of enzyme activity, represent a positive response, as was noted by Li and Fields (25). VP19C and 22a inserted into the DNA binding domain of pGBT9 each produced a false positive response (enzyme activity in the presence or absence of an activation domain) of 3 and 8 units of activity, respectively (Table 1, column 2). Interactions given in the Table that employ these constructs have not been corrected for the small false positive activity. The signal of the false positive response of 22a and VP19C, as judged by enzyme activity, did not increase in the presence of another noninteracting capsid protein in pGAD424. The false positive interaction noted above for pGBT9-22a was detected when the precursor (307–635), but not the ma-
ture (307–610), form of 22a was employed. The interaction between VP5 and 22a (and UL26) was only obtained when VP5 was expressed from pGBT9. A second yeast cell, HF7c (26), was employed to determine if the same interactions would be detected when the b-galactosidase reporter gene was fused to a different promoter. The lacZ gene in HF7c is fused to the TATA portion of the CYC1 promoter into which three copies of a GAL4 responsive (17-mer) DNA binding site were inserted. Leucine- and tryptophan-independent transformants were obtained as described above and assayed for enzyme activity using X-GAL screening. The three capsid protein interactions detected using Y526 were also detected using the GAL4 responsive CYC1 promoter in HF7c. In addition, the upstream activating and TATA sequences of the GAL1 promoter are fused to the HIS3 gene in HF7c, so that growth in the absence of histidine requires GAL4-dependent transcription. Transformed cells were prepared as described above and aliquots were spread onto the surface of selection plates that did not contain histidine. Colonies were formed when plasmids were used that gave rise to the three capsid protein interactions observed by X-GAL screening. All of the colonies formed gave a positive response in the presence of X-GAL and at times that were similar to those for the leucine- and tryptophan-independent colonies. No additional interactions were detected using HF7c cells. In summary the same capsid protein interactions found with Y526 were also detected with HF7c using a geneticscreen assay and a phenotypic-selection assay for colony formation. Additional constructs were prepared and tested in order to localize portions of the capsid proteins responsible for the molecular interactions. Results are shown in Fig. 1. Mapping experiments for the self-interaction of 22a are shown in Fig. 1A. The C-terminal residue for each construct is indicated to the right of that construct and the N-terminal residue to the left, when the number is different from that of the authentic start position (residue 307 of UL26). The self-interaction of the precursor form of 22a gave the highest value for b-galactosidase enzyme activity (289 enzyme units). By contrast, 138 units were obtained when the mature form of 22a (22a-610) was used in the two vectors. Shortening of the C-terminus of 22a-610 to residue 504 reduced the activity to 88 units, while a further shortening to residue 457 resulted in an almost complete loss of b-galactosidase activity (1.6 units). Similar enzyme activities were obtained when the C-terminal mutant constructs were tested against the fulllength molecule. A polypeptide that terminated translation at amino acid 418 abolished enzyme activity completely, as did deletion of N-terminal residues up to 375 in the 22a-504 construct. We conclude that the N-terminal residues up to approximately amino acid 504 are essential for the 22a self-interaction and that the 25 C-terminal
AID
vira
VY 7959
/
6a18$$$102
05-16-96 19:14:14
AP: Virology
SHORT COMMUNICATION
519
TABLE 1 HSV-1 Capsid Protein Interactions
Note. Plasmids encoding each of the capsid proteins were used in standard PCR reactions. These were: for VP23 (UL18), pKKI; for VP5 (UL19), pKBgN (37); for the UL26-specified proteins, pKNotI (1); for VP26 (UL35), pKEL; and for VP19C (UL38), a 3.8-kb BamHI/SalI fragment derived from pKBH (unpublished). Oligonucleotide primers were used that annealed to DNA encoding the first and last six amino acids for the coding and noncoding strand, respectively. The 5*-end of the coding strand primer terminated with an EcoRI site, and the 5*-end of the noncoding strand primer terminated in a BamHI site for UL18, UL35, and UL38, and a BglII site for UL19 and UL26. Agarose gel-purified fragments were cloned into the EcoRI/BamHI sites of pGBT9 and pGAD424 (19) (Obtained from Dr. S. Fields). For VP5, VP19C, VP23, 22a, 22a-610, and UL26 most of the DNA obtained by PCR was replaced with authentic sequences derived from the clones noted above that had previously been shown to express functional proteins. The remaining 5* and 3* terminal regions of each gene were sequenced and found to be the same as the corresponding sequences reported by McGeoch et al. (38) for HSV-1 strain 17. DNAs were prepared using Qiagen kits (Chatsworth, CA). The yeast cell, Y526 (also obtained from Dr. S. Fields) was grown in YEPD medium, and competent cells were prepared and transformed for leucine and tryptophan essentially as described by Bartel et al. (19). Early log phase cells were pelleted, washed with water, and resuspended in 1/30 original volume in TE (10 mM Tris, pH Å 7.5, 1 mM EDTA) containing 100 mM lithium acetate (pH Å 7.5). Each transformation reaction contained 100 ml of cells, 20 mg salmon sperm carrier DNA, and 0.35 mg of each plasmid. The reactions were incubated at room temperature for 30 min (inverted approximately every 5 min), prior to the addition of dimethylsulfoxide (70 ml) and exposure to a heat shock (427 for 15 min). Cells were pelleted and resuspended in 200 ml TE and 125 ml was spread directly onto wetted Whatman No. 1 filter papers placed on the surfaces of selection plates composed of SD media lacking leucine, trytophan, and uracil. Cells were incubated for 3 days at 307 and approximately 300 colonies grew on the filters. Filters were lifted from the plate and assayed for the appearance of blue color using X-GAL as a substrate for b-galactosidase (filter assay). Subsequently, b-galactosidase enzyme activity was measured using o-nitrophenyl-b-D-galactopyranoside (ONPG) as substrate (liquid culture assay). For the latter 5 ml of SD media without leucine, tryptophan and uracil were added to the culture dishes containing transformants and the yeast cells in the colonies were scraped into the media. Assays were generally done directly on the resulting cell suspension. Aliquots of the crude lysate were assayed for b-galactosidase activity as described by Rose et al. (39). However, the protein concentration was determined by the Lowry (40) rather than the Bradford method. The development of color produced by the metabolism of ONPG was measured as absorbance units at 420 nm and protein determinations were at 750 nm, following sedimentation for 3 min in a microfuge to remove debris. The units of enzyme activity are nanomoles of ONPG cleaved per minute per milligram of protein, and the absorbance of 1 nm of ONPG at 420 nm was taken as 0.0045 (41). Pairs of capsid proteins that resulted in interactions are indicated by shaded squares, and the number of units of enzyme activity are given within the shaded squares. Results indicated by w gave enzyme activities that were not significantly greater than the false positive values produced by pGBT9-19C or pGBT9-22a.
amino acids increase the interaction by approximately twofold. The C-terminal 25 residues of 22a were also essential for its interaction with VP5. When the precursor form of 22a was expressed from pGAD424 and VP5 from pGBT9, 86 units of enzyme activity were observed (Fig. 1B). However, using the mature form of 22a (22a-610) the background level of enzyme activity was detected. The same result was obtained if UL26 and UL26-610 were substi-
tuted for 22a and 22a-610, respectively. However, for the pGBT9-VP5/pGAD424-UL26 combination 20 units of bgalactosidase activity were obtained. This is the first report of a self-interaction between 22a scaffold molecules. The relative strength of the interaction is similar to that reported previously for p53 and the SV40 T-antigen. This relatively strong interaction was somewhat of a surprise because monomer formation is likely required for scaffold exit from fully formed B-cap-
AID
vira
VY 7959
/
6a18$$$102
05-16-96 19:14:14
AP: Virology
520
SHORT COMMUNICATION
sids. Monomers of 22a have a similar mass to the phage P22 scaffold protein which has an axial ratio of 9:1 (27). For the same axial ratio, the monomer diameter of 22a would be approximately 20 A which is small enough to exit the 40 A penton pores in the capsid (6). The cleavage of the scaffold protein for HSV-1 may serve to terminate the scaffold/major capsid protein interaction and to promote subsequent monomer formation. The increase in the 22a self-interaction, due to the presence of the Cterminal 25 residues, may promote oligomerization of scaffold molecules and help to minimize protease cleavage until the interaction with VP5 occurs. If the C-terminal residues are essential for scaffold formation, then the scaffold-like structures observed using constructs that express precursor and mature forms of 22a in insect cells may not be representative of viral-infected mammalian cells (28, 29). The data presented here for the VP5-22a interaction are consistent with those from other experiments. Amino acid residues 611–635 of HSV-1 22a have been shown to be required for capsid formation in HSV-1-infected Vero cells (30) and insect cells infected with recombinant baculoviruses (28, 31). Hong et al. (32) showed that as few as 12 C-terminal amino acids of 22a, expressed and purified from Escherichia coli, were sufficient to bind VP5 obtained from virus-infected mammalian cells. In transfection experiments, the transport of VP5 to the nucleus was enhanced in the presence of 22a (33, 34). The nuclear colocalization of VP5 and 22a was detected using the precursor (307–635) and the mature (307–610) form of 22a (28, 34). Colocalization of VP5 and 22a-610 implies a molecular interaction which was not detected using the two-hybrid assay. Similar data obtained using the two-hybrid system and the human- and simian-cytomegalo viruses scaffold and major capsid proteins are presented in the manuscript by Wood et al. (35). The scaffold protein was shown to participate in a self-interaction, and the C-terminal residues (precursor form) were essential for interaction with the major capsid protein. The self-interaction was decreased by the presence of the C-terminal residues, which is opposite to that for HSV-1. However, the HSV1 self-interaction is stronger than that for CMV so that monomer formation for CMV following capsid assembly may occur without an involvement of the C-terminal residues. The VP19C/VP23 interaction was detected when either ORF was expressed from pGAD424 or pGBT9, despite the fact that VP19C gave a false positive response when present in the DNA binding domain vector. pGBT9VP19C/pGAD424-VP23 gave 63 units of enzyme activity compared with 3 units for pGBT9-VP19C/pGAD424. The pGBT9-VP23/pGAD424-VP19C combination produced 11 units of b-galactosidase activity. Attempts to map the VP19C/VP23 interaction were not successful. Constructs that encode the N- and C-terminal one-half and two-
AID
VY 7959
/
6a18$$$102
05-16-96 19:14:14
thirds portions of each molecule were negative when tested with a full-length copy of the other member of the interacting pair. This may mean that lacZ expression is due to interaction of noncontiguous portions of each molecule, and mapping the interactions will require an approach that reflects their 3D nature. Using nonreducing and reducing two-dimensional gel analysis of proteins prepared from HSV-2 capsids, Zweig et al. (36) demonstrated the existence of disulfide crosslinks between the VP5 and VP19C counterparts of HSV2. Chemical cross-linking experiments were reported recently for HSV-1 B capsids (1). Cross-links were detected between VP5, VP19C, and VP23 and another between VP5, VP19C, and 22a. It is somewhat surprising that the yeast two-hybrid system did not detect an interaction between VP5 and VP19C. However, as is often the case the two-hybrid system detects most, but not all, of the interactions indicated by a variety of other techniques. The future goal of our work is to use the two-hybrid system to identify and characterize the domains involved in the interactions reported here, and to use other techniques to uncover additional capsid protein interactions. ACKNOWLEDGMENTS This work was supported by Public Health Service Grant AI33077 from the National Institutes of Health. It is a pleasure to acknowledge discussion of data with Neal A. DeLuca.
REFERENCES 1. Desai, P., Watkins, S. C., and Person, S., J. Virol. 68, 5365–5374 (1994). 2. Tatman, J. D., Preston, V. G., Nicholson, P., Elliot, R. M., and Rixon, F. J., J. Gen. Virol. 75, 1101–1113 (1994). 3. Thomsen, D. R., Roof, L. L., and Homa, F. L., J. Virol. 68, 2442– 2457 (1994). 4. Booy, F. P., Trus, B. L., Newcomb, W. W., Brown, J. C., Conway, F. F., and Steven, A. C., Proc. Natl. Acad. Sci. USA 91, 5652– 5656 (1994). 5. Newcomb, W. W., Trus, B. L., Booy, F. P., Steven, A. C., Wall, J. S., and Brown, J. C., J. Mol. Biol. 232, 499–511 (1993). 6. Zhou, Z., Prasad, B. V. V., Jakana, J., Rixon, F. J., and Chiu, W., J. Mol. Biol. 242, 456–469 (1994). 7. Liu, F., and Roizman, B., J. Virol. 67, 1300–1309 (1993). 8. Preston, V. G., Rixon, F. J., McDougall, I. M., McGregor, M., and AlKobaisi, M. F., Virology 186, 87–98 (1992). 9. Dilanni, C. L., Drier, D. A., Deckman, I. C., McCann, P. J., Liu, F., Roizman, B., Colonno, R. J., and Cordingly, M. G., J. Biol. Chem. 268, 2048–2051 (1993). 10. Liu, F., and Roizman, B., J. Virol. 65, 5149–5156 (1991). 11. Weinheimer, S. P., McCann, P. J., O’Boyle, D. R., Stevens, J. T., Boyd, B. A., Drier, D. A., Yamanaka, G. A., Dilanni, C. L., Deckman, I. C., and Cordingly, M. G., J. Virol. 67, 5813–5822 (1993). 12. Liu, F., and Roizman, B., J. Virol. 65, 206–212 (1991). 13. Newcomb, W. W., and Brown, J. C., J. Virol. 65, 613–620 (1991). 14. Baum, E. Z., Bebernitz, G. A., Hulmes, J. D., Muzithras, V. P., Jones, T. R., and Gluzman, Y., J. Virol. 67, 497–506 (1993). 15. Welch, A. R., Woods, A. S., McNally, L. M., Cotter, R. J., and Gibson, W., Proc. Natl. Acad. Sci. USA 88, 10792–10796 (1991). 16. Welch, A. R., McNally, L. M., Hall, M. R. T., and Gibson, W., J. Virol. 67, 7360–7372 (1993).
vira
AP: Virology
SHORT COMMUNICATION 17. Gibson, W., Marcy, A. I., Comolli, J. C., and Lee, J., J. Virol. 64, 1241– 1249 (1990). 18. Patel, A. H., and Maclean, J. B., Virology 206, 465–478 (1995). 19. Bartel, P. L., Chien, C-T., Sternglanz, R., and Fields, S., In ‘‘Cellular Interactions in Development: A Practical Approach’’ (D. A. Hartley, Ed.), pp. 153–179. Oxford Univ. Press, Oxford, UK, 1993. 20. Chien, C-T., Bartel, P. L., Sternglanz, R., and Fields, S., Proc. Natl. Acad. Sci. USA 88, 9578–9582 (1991). 21. Fields, S., and Song, O-K., Nature 340, 245–246 (1989). 22. Keegan, L., Gill, G., and Ptashne, M., Science 231, 699–704 (1986). 23. Ma, J., and Ptashne, M., Cell 48, 847–853 (1987). 24. Ma, J., and Ptashne, M., Cell 55, 443–446 (1988). 25. Li, B., and Fields, S., FASEB 7, 957–963 (1993). 26. Feilotter, H. E., Hannon, G. J., Rudell, C. J., and Beach, D., Nucleic Acids Res. 22, 1502–1503 (1994). 27. Prassad, B. V. V., Prevelige, P. E., Marietta, E., Chen, R. O., Thomas, D., King, J., and Chiu, W., J. Mol. Biol. 231, 65–74 (1993). 28. Kennard, J., Rixon, F. J., McDougall, I. M., Tatman, J. D., and Preston, V. G., J. Gen. Virol. 76, 1611–1621 (1995). 29. Preston, V. G., Al-Kobaisi, M. F., McDougall, I. M., and Rixon, F. J., J. Gen. Virol. 75, 2355–2366 (1994). 30. Matusick-Kumar, L., Newcomb, W. W., Brown, J. C., McCann III,
AID
VY 7959
/
6a18$$$103
05-16-96 19:14:14
31. 32. 33. 34. 35. 36. 37. 38.
39. 40. 41.
521
P. J., Hurlburt, W., Weinheimer, S. P., and Gao, M., J. Virol. 69, 4347–4356 (1995). Thomsen, D. R., Newcomb, W. W., Brown, J. C., and Homa, F. L., J. Virol. 69, 3690–3703 (1995). Hong, Z., Beaudet-Miller, M., Durkin, J., Zhang, R., and Kwong, A. D., J. Virol. 70, 533–540 (1996). Matusick-Kumar, L., Hurlburt, W., Weinheimer, S. P., Newcomb, W. W., Brown, J. C., and Gao, M., J. Virol. 68, 5384–5394 (1994). Nicholson, P., Addison, C., Cross, A. M., Kennard, J., Preston, V. G., and Rixon, F. J., J. Gen. Virol. 75, 1091–1099 (1994). Wood, L. J., Baxter, M. K., and Gibson, W., submitted for publication. Zweig, M., Heilman, C. J., and Hampar, B., Virology 94, 442–450 (1979). Desai, P., DeLuca, N. A., Glorioso, J. C., and Person, S., J. Virol. 67, 1357–1364 (1993). 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., J. Gen. Virol. 69, 1531–1574 (1988). Rose, M., Casadaban, M. J., and Botstein, D., Proc. Natl. Acad. Sci. USA 78, 2460–2464 (1981). Schacterle, G. R., and Pollack, R. L., Anal. Biochem. 51, 654–655 (1973). Miller, J. H., ‘‘Experiments in Molecular Genetics.’’ Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1972.
vira
AP: Virology
220, 516–521 (1996) 0341
ARTICLE NO.
SHORT COMMUNICATION Molecular Interactions between the HSV-1 Capsid Proteins as Measured by the Yeast Two-Hybrid System PRASHANT DESAI and STANLEY PERSON1 Department of Molecular Genetics and Biochemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15261 Received January 26, 1996; accepted April 12, 1996 HSV-1 B capsids are composed of seven major proteins, designated VP5, VP19C, 21, 22a, VP23, VP24, and VP26. VP indicates that the capsid protein is also a component of the infectious virion. Capsid proteins 21, 22a, and VP24 are specified by a single open reading frame (UL26) that encodes 635 amino acids. An objective of the work in our laboratory is to identify and map interactions among and between capsid proteins. In the present studies we employed the yeast GAL4 two-hybrid system developed by Fields and his colleagues (Nature 240, 245–246 (1989)) for this purpose. DNA corresponding to the capsid open reading frames was derived as a PCR product and fused to sequences of the GAL4 activation and DNA binding domains. Using this system each of the capsid proteins has been tested for interactions with all of the other capsid proteins. Three interactions have been identified: a relatively strong self-interaction between 22a molecules (residues 307– 635 of UL26), bimolecular interactions between 22a and VP5, and another between VP19C and VP23. The interactions were detected by the expression of b-galactosidase enzyme activity, and yielded 289, 86, and 63 units of enzyme activity, respectively. For the 22a self-interaction, elimination of residues 611–635 resulted in an approximately twofold decrease in enzyme activity. The C-terminal 25 amino acids of 22a were also essential for the bimolecular interaction between 22a and VP5. q 1996 Academic Press, Inc.
Three types of capsids have been identified in herpes simplex virus type 1 (HSV-1)-infected cells designated A, B, and C corresponding to increasing distance sedimented through sucrose gradients. B capsids contain a centrally located (scaffold) protein that is replaced with viral DNA in the more rapidly sedimenting C capsids. The slowest sedimenting A capsids are devoid of the scaffold protein and DNA, and may result from abortive packaging events. The capsid shells of HSV-1 are comprised of four proteins, designated VP5, VP19C, VP23, and VP26 (1–3). The major capsid protein, VP5, comprises the hexon and penton subunits of the icosahedral capsid shell, VP19C and VP23 connect the subunits, and VP26 is found on the tips of the hexons (4–6). The remaining components of B capsids, 21, 22a, and VP24, are encoded by a single open reading frame (ORF), UL26, whose organization is depicted at the top of Fig. 1. The UL26 gene product encodes an autocatalytic protease (7, 8) which cleaves the UL26 gene product between residues 247 and 248 to generate VP24 (1–247) and 21 (248–635) (9–11). The more abundant capsid protein,
22a (307–635) is translated in the same reading frame from a second, shorter transcript (7, 12). For HSV-1 a second cleavage site between amino acids 610 and 611 of UL26 generates products referred to as the mature forms of the proteins, as opposed to the full-length, or precursor forms, representative of the primary translation products (8–10). 22a and to a lesser extent 21 occupy the inner capsid space and act to form a scaffold upon which the icosahedral shell is constructed (1–3, 13). The human and simian-cytomegalovirus homologues of UL26 appear to have the same biological activities as those of HSV-1 and have been extensively characterized (see for example, 14–16). Gibson et al. (17) were the first to suggest that this protein may serve as a scaffold for capsid shell assembly in a manner analogous to bacteriophage capsid assembly. The proteins of the UL26 ORF may be thought of as maturational as well as structural capsid proteins, as opposed to the four proteins that comprise the capsid shell that seem to have only a structural role in assembly. There are probably a number of other viral-specified proteins with roles in capsid maturation that may be present in capsids in low copy numbers and may not be components of completed virions. By analogy to bacteriophage assembly these include DNA packaging and associated docking proteins, proteins that bend DNA and proteins that plug the hole used for DNA packaging. Recently the
1 To whom correspondence and reprint requests should be addressed at current address: Virology Laboratories, Department of Pharmacology and Molecular Sciences, Johns Hopkins University School of Medicine, 725 N. Wolfe Street, Baltimore, MD 21205. Fax: (410) 9553023.
0042-6822/96 $18.00
516
Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
AID
VY 7959
/
6a18$$$101
05-16-96 19:14:14
vira
AP: Virology
SHORT COMMUNICATION
517
FIG. 1. Characterization of capsid protein interactions. The UL26 ORF is shown at the top of A. The position of the autoproteolytic cleavages are indicated between residues 247 and 248 and between 610 and 611. Cleavage after residue 247 generates VP24 and 21 from the UL26 translation product. 22a is translated from a second transcript whose initiation codon is at position 307 of the UL26 ORF. The protein coding regions are shown as horizontal lines and relevant terminal amino acid residues are marked at the ends of the lines. Replacement cloning and complete nucleotide sequence analysis were not performed for the 22a mutant constructs, although the termini were shown to be correct by DNA sequencing. In B, the VP5 coding region was cloned into pGBT9 and 22a into pGAD424. In C, VP19C was cloned in pGBT9 and VP23 in pGAD424. DNA amplification and cloning was as indicated in the legend to Table 1. The transformation of Y526 and subsequent measurement of enzyme activity were also as described in the legend to Table 1. Enzyme units { one standard deviation are shown at the left of each construct and are the averages of three determinations.
UL6 gene product was shown to be a capsid protein and may be representative of one of these proteins (18). The purpose of the present experiments was to detect molecular interactions among the capsid proteins of HSV-1 using the yeast two-hybrid system (19–21), which employs the yeast transcriptional activator, GAL4, characterized by Ptashne and his colleagues (see for example, 22–24). The GAL4 protein consists of a DNA binding domain (residues 1–147) and a transcriptional activation domain (residues 768–881), which are distinct and separable (22). Transcriptional activation of a gene by GAL4
requires that both domains be in close proximity in order to interact with the basal transcriptional complex. Normally the two domains are joined by covalent bonds but Fields and his colleagues showed that weak interactions between proteins fused to each domain also resulted in transcription. Transcriptional activation was measured using a lacZ reporter gene fused to a GAL4 responsive promoter (GAL1). Two plasmids were employed; one, pGBT9, encodes the DNA binding domain and the other, pGAD424, specifies the GAL4 activation domain. Polylinkers in these plasmids allow the insertion of test pro-
AID
vira
VY 7959
/
6a18$$$102
05-16-96 19:14:14
AP: Virology
518
SHORT COMMUNICATION
tein sequences in phase with, and downstream from, the respective GAL4 domains. pGBT9 also encodes the TRP1 gene and pGAD424 encodes the LEU2 gene so that following transformation of appropriate yeast cells, colonies may be selected that grow in the absence of leucine and tryptophan. Colonies can be assayed directly for enzyme activity by the appearance of blue color using X-GAL as a substrate (19). Sequences encoding each capsid protein of HSV-1 were obtained as PCR products and cloned into both plasmids. Constructs representing the entire UL26 ORF (residues 1–635), and 22a (residues 307–635), were used in this assay. Wherever possible the capsid proteins were tested when present in both vectors. Most experiments utilized Y526 yeast cells. The Y526 cell contains the lacZ reporter sequence fused to the upstream activating and TATA sequences of the GAL1 promoter (19). Following transformation, colonies that developed on plates containing selective growth media were assayed for the development of blue color. The results are summarized in Table 1. A total of three interactions, designated by a /, were detected; a bimolecular interaction between VP5 and 22a (and with UL26), another between VP23 and VP19C, and a self-interaction between 22a molecules. Blue color, in all of the colonies, was detectable after 20 min of incubation for the 22a self-interaction and after approximately 40 min for the other interactions. No interactions were detected for VP24 and VP26, either with themselves or with the other capsid proteins. The enzyme units of b-galactosidase expressed as nanomoles of ONPG hydrolyzed per minute per milligram of total protein were determined for each interaction and are listed in the shaded squares of the Table. An interaction between cellular p53 and SV40 T-antigen produced approximately 150 units of enzyme activity when employed in the two-hybrid system (25), a value that we also obtained when these constructs were used as a positive control. The self interaction between 22a scaffold molecules resulted in a similar level of enzyme activity (289 units). The two vectors without inserts resulted in 0.4 { 0.3 units of activity (standard deviations are not listed in Table 1 but are given in Fig. 1). Therefore, values greater than this value, or approximately 1 unit of enzyme activity, represent a positive response, as was noted by Li and Fields (25). VP19C and 22a inserted into the DNA binding domain of pGBT9 each produced a false positive response (enzyme activity in the presence or absence of an activation domain) of 3 and 8 units of activity, respectively (Table 1, column 2). Interactions given in the Table that employ these constructs have not been corrected for the small false positive activity. The signal of the false positive response of 22a and VP19C, as judged by enzyme activity, did not increase in the presence of another noninteracting capsid protein in pGAD424. The false positive interaction noted above for pGBT9-22a was detected when the precursor (307–635), but not the ma-
ture (307–610), form of 22a was employed. The interaction between VP5 and 22a (and UL26) was only obtained when VP5 was expressed from pGBT9. A second yeast cell, HF7c (26), was employed to determine if the same interactions would be detected when the b-galactosidase reporter gene was fused to a different promoter. The lacZ gene in HF7c is fused to the TATA portion of the CYC1 promoter into which three copies of a GAL4 responsive (17-mer) DNA binding site were inserted. Leucine- and tryptophan-independent transformants were obtained as described above and assayed for enzyme activity using X-GAL screening. The three capsid protein interactions detected using Y526 were also detected using the GAL4 responsive CYC1 promoter in HF7c. In addition, the upstream activating and TATA sequences of the GAL1 promoter are fused to the HIS3 gene in HF7c, so that growth in the absence of histidine requires GAL4-dependent transcription. Transformed cells were prepared as described above and aliquots were spread onto the surface of selection plates that did not contain histidine. Colonies were formed when plasmids were used that gave rise to the three capsid protein interactions observed by X-GAL screening. All of the colonies formed gave a positive response in the presence of X-GAL and at times that were similar to those for the leucine- and tryptophan-independent colonies. No additional interactions were detected using HF7c cells. In summary the same capsid protein interactions found with Y526 were also detected with HF7c using a geneticscreen assay and a phenotypic-selection assay for colony formation. Additional constructs were prepared and tested in order to localize portions of the capsid proteins responsible for the molecular interactions. Results are shown in Fig. 1. Mapping experiments for the self-interaction of 22a are shown in Fig. 1A. The C-terminal residue for each construct is indicated to the right of that construct and the N-terminal residue to the left, when the number is different from that of the authentic start position (residue 307 of UL26). The self-interaction of the precursor form of 22a gave the highest value for b-galactosidase enzyme activity (289 enzyme units). By contrast, 138 units were obtained when the mature form of 22a (22a-610) was used in the two vectors. Shortening of the C-terminus of 22a-610 to residue 504 reduced the activity to 88 units, while a further shortening to residue 457 resulted in an almost complete loss of b-galactosidase activity (1.6 units). Similar enzyme activities were obtained when the C-terminal mutant constructs were tested against the fulllength molecule. A polypeptide that terminated translation at amino acid 418 abolished enzyme activity completely, as did deletion of N-terminal residues up to 375 in the 22a-504 construct. We conclude that the N-terminal residues up to approximately amino acid 504 are essential for the 22a self-interaction and that the 25 C-terminal
AID
vira
VY 7959
/
6a18$$$102
05-16-96 19:14:14
AP: Virology
SHORT COMMUNICATION
519
TABLE 1 HSV-1 Capsid Protein Interactions
Note. Plasmids encoding each of the capsid proteins were used in standard PCR reactions. These were: for VP23 (UL18), pKKI; for VP5 (UL19), pKBgN (37); for the UL26-specified proteins, pKNotI (1); for VP26 (UL35), pKEL; and for VP19C (UL38), a 3.8-kb BamHI/SalI fragment derived from pKBH (unpublished). Oligonucleotide primers were used that annealed to DNA encoding the first and last six amino acids for the coding and noncoding strand, respectively. The 5*-end of the coding strand primer terminated with an EcoRI site, and the 5*-end of the noncoding strand primer terminated in a BamHI site for UL18, UL35, and UL38, and a BglII site for UL19 and UL26. Agarose gel-purified fragments were cloned into the EcoRI/BamHI sites of pGBT9 and pGAD424 (19) (Obtained from Dr. S. Fields). For VP5, VP19C, VP23, 22a, 22a-610, and UL26 most of the DNA obtained by PCR was replaced with authentic sequences derived from the clones noted above that had previously been shown to express functional proteins. The remaining 5* and 3* terminal regions of each gene were sequenced and found to be the same as the corresponding sequences reported by McGeoch et al. (38) for HSV-1 strain 17. DNAs were prepared using Qiagen kits (Chatsworth, CA). The yeast cell, Y526 (also obtained from Dr. S. Fields) was grown in YEPD medium, and competent cells were prepared and transformed for leucine and tryptophan essentially as described by Bartel et al. (19). Early log phase cells were pelleted, washed with water, and resuspended in 1/30 original volume in TE (10 mM Tris, pH Å 7.5, 1 mM EDTA) containing 100 mM lithium acetate (pH Å 7.5). Each transformation reaction contained 100 ml of cells, 20 mg salmon sperm carrier DNA, and 0.35 mg of each plasmid. The reactions were incubated at room temperature for 30 min (inverted approximately every 5 min), prior to the addition of dimethylsulfoxide (70 ml) and exposure to a heat shock (427 for 15 min). Cells were pelleted and resuspended in 200 ml TE and 125 ml was spread directly onto wetted Whatman No. 1 filter papers placed on the surfaces of selection plates composed of SD media lacking leucine, trytophan, and uracil. Cells were incubated for 3 days at 307 and approximately 300 colonies grew on the filters. Filters were lifted from the plate and assayed for the appearance of blue color using X-GAL as a substrate for b-galactosidase (filter assay). Subsequently, b-galactosidase enzyme activity was measured using o-nitrophenyl-b-D-galactopyranoside (ONPG) as substrate (liquid culture assay). For the latter 5 ml of SD media without leucine, tryptophan and uracil were added to the culture dishes containing transformants and the yeast cells in the colonies were scraped into the media. Assays were generally done directly on the resulting cell suspension. Aliquots of the crude lysate were assayed for b-galactosidase activity as described by Rose et al. (39). However, the protein concentration was determined by the Lowry (40) rather than the Bradford method. The development of color produced by the metabolism of ONPG was measured as absorbance units at 420 nm and protein determinations were at 750 nm, following sedimentation for 3 min in a microfuge to remove debris. The units of enzyme activity are nanomoles of ONPG cleaved per minute per milligram of protein, and the absorbance of 1 nm of ONPG at 420 nm was taken as 0.0045 (41). Pairs of capsid proteins that resulted in interactions are indicated by shaded squares, and the number of units of enzyme activity are given within the shaded squares. Results indicated by w gave enzyme activities that were not significantly greater than the false positive values produced by pGBT9-19C or pGBT9-22a.
amino acids increase the interaction by approximately twofold. The C-terminal 25 residues of 22a were also essential for its interaction with VP5. When the precursor form of 22a was expressed from pGAD424 and VP5 from pGBT9, 86 units of enzyme activity were observed (Fig. 1B). However, using the mature form of 22a (22a-610) the background level of enzyme activity was detected. The same result was obtained if UL26 and UL26-610 were substi-
tuted for 22a and 22a-610, respectively. However, for the pGBT9-VP5/pGAD424-UL26 combination 20 units of bgalactosidase activity were obtained. This is the first report of a self-interaction between 22a scaffold molecules. The relative strength of the interaction is similar to that reported previously for p53 and the SV40 T-antigen. This relatively strong interaction was somewhat of a surprise because monomer formation is likely required for scaffold exit from fully formed B-cap-
AID
vira
VY 7959
/
6a18$$$102
05-16-96 19:14:14
AP: Virology
520
SHORT COMMUNICATION
sids. Monomers of 22a have a similar mass to the phage P22 scaffold protein which has an axial ratio of 9:1 (27). For the same axial ratio, the monomer diameter of 22a would be approximately 20 A which is small enough to exit the 40 A penton pores in the capsid (6). The cleavage of the scaffold protein for HSV-1 may serve to terminate the scaffold/major capsid protein interaction and to promote subsequent monomer formation. The increase in the 22a self-interaction, due to the presence of the Cterminal 25 residues, may promote oligomerization of scaffold molecules and help to minimize protease cleavage until the interaction with VP5 occurs. If the C-terminal residues are essential for scaffold formation, then the scaffold-like structures observed using constructs that express precursor and mature forms of 22a in insect cells may not be representative of viral-infected mammalian cells (28, 29). The data presented here for the VP5-22a interaction are consistent with those from other experiments. Amino acid residues 611–635 of HSV-1 22a have been shown to be required for capsid formation in HSV-1-infected Vero cells (30) and insect cells infected with recombinant baculoviruses (28, 31). Hong et al. (32) showed that as few as 12 C-terminal amino acids of 22a, expressed and purified from Escherichia coli, were sufficient to bind VP5 obtained from virus-infected mammalian cells. In transfection experiments, the transport of VP5 to the nucleus was enhanced in the presence of 22a (33, 34). The nuclear colocalization of VP5 and 22a was detected using the precursor (307–635) and the mature (307–610) form of 22a (28, 34). Colocalization of VP5 and 22a-610 implies a molecular interaction which was not detected using the two-hybrid assay. Similar data obtained using the two-hybrid system and the human- and simian-cytomegalo viruses scaffold and major capsid proteins are presented in the manuscript by Wood et al. (35). The scaffold protein was shown to participate in a self-interaction, and the C-terminal residues (precursor form) were essential for interaction with the major capsid protein. The self-interaction was decreased by the presence of the C-terminal residues, which is opposite to that for HSV-1. However, the HSV1 self-interaction is stronger than that for CMV so that monomer formation for CMV following capsid assembly may occur without an involvement of the C-terminal residues. The VP19C/VP23 interaction was detected when either ORF was expressed from pGAD424 or pGBT9, despite the fact that VP19C gave a false positive response when present in the DNA binding domain vector. pGBT9VP19C/pGAD424-VP23 gave 63 units of enzyme activity compared with 3 units for pGBT9-VP19C/pGAD424. The pGBT9-VP23/pGAD424-VP19C combination produced 11 units of b-galactosidase activity. Attempts to map the VP19C/VP23 interaction were not successful. Constructs that encode the N- and C-terminal one-half and two-
AID
VY 7959
/
6a18$$$102
05-16-96 19:14:14
thirds portions of each molecule were negative when tested with a full-length copy of the other member of the interacting pair. This may mean that lacZ expression is due to interaction of noncontiguous portions of each molecule, and mapping the interactions will require an approach that reflects their 3D nature. Using nonreducing and reducing two-dimensional gel analysis of proteins prepared from HSV-2 capsids, Zweig et al. (36) demonstrated the existence of disulfide crosslinks between the VP5 and VP19C counterparts of HSV2. Chemical cross-linking experiments were reported recently for HSV-1 B capsids (1). Cross-links were detected between VP5, VP19C, and VP23 and another between VP5, VP19C, and 22a. It is somewhat surprising that the yeast two-hybrid system did not detect an interaction between VP5 and VP19C. However, as is often the case the two-hybrid system detects most, but not all, of the interactions indicated by a variety of other techniques. The future goal of our work is to use the two-hybrid system to identify and characterize the domains involved in the interactions reported here, and to use other techniques to uncover additional capsid protein interactions. ACKNOWLEDGMENTS This work was supported by Public Health Service Grant AI33077 from the National Institutes of Health. It is a pleasure to acknowledge discussion of data with Neal A. DeLuca.
REFERENCES 1. Desai, P., Watkins, S. C., and Person, S., J. Virol. 68, 5365–5374 (1994). 2. Tatman, J. D., Preston, V. G., Nicholson, P., Elliot, R. M., and Rixon, F. J., J. Gen. Virol. 75, 1101–1113 (1994). 3. Thomsen, D. R., Roof, L. L., and Homa, F. L., J. Virol. 68, 2442– 2457 (1994). 4. Booy, F. P., Trus, B. L., Newcomb, W. W., Brown, J. C., Conway, F. F., and Steven, A. C., Proc. Natl. Acad. Sci. USA 91, 5652– 5656 (1994). 5. Newcomb, W. W., Trus, B. L., Booy, F. P., Steven, A. C., Wall, J. S., and Brown, J. C., J. Mol. Biol. 232, 499–511 (1993). 6. Zhou, Z., Prasad, B. V. V., Jakana, J., Rixon, F. J., and Chiu, W., J. Mol. Biol. 242, 456–469 (1994). 7. Liu, F., and Roizman, B., J. Virol. 67, 1300–1309 (1993). 8. Preston, V. G., Rixon, F. J., McDougall, I. M., McGregor, M., and AlKobaisi, M. F., Virology 186, 87–98 (1992). 9. Dilanni, C. L., Drier, D. A., Deckman, I. C., McCann, P. J., Liu, F., Roizman, B., Colonno, R. J., and Cordingly, M. G., J. Biol. Chem. 268, 2048–2051 (1993). 10. Liu, F., and Roizman, B., J. Virol. 65, 5149–5156 (1991). 11. Weinheimer, S. P., McCann, P. J., O’Boyle, D. R., Stevens, J. T., Boyd, B. A., Drier, D. A., Yamanaka, G. A., Dilanni, C. L., Deckman, I. C., and Cordingly, M. G., J. Virol. 67, 5813–5822 (1993). 12. Liu, F., and Roizman, B., J. Virol. 65, 206–212 (1991). 13. Newcomb, W. W., and Brown, J. C., J. Virol. 65, 613–620 (1991). 14. Baum, E. Z., Bebernitz, G. A., Hulmes, J. D., Muzithras, V. P., Jones, T. R., and Gluzman, Y., J. Virol. 67, 497–506 (1993). 15. Welch, A. R., Woods, A. S., McNally, L. M., Cotter, R. J., and Gibson, W., Proc. Natl. Acad. Sci. USA 88, 10792–10796 (1991). 16. Welch, A. R., McNally, L. M., Hall, M. R. T., and Gibson, W., J. Virol. 67, 7360–7372 (1993).
vira
AP: Virology
SHORT COMMUNICATION 17. Gibson, W., Marcy, A. I., Comolli, J. C., and Lee, J., J. Virol. 64, 1241– 1249 (1990). 18. Patel, A. H., and Maclean, J. B., Virology 206, 465–478 (1995). 19. Bartel, P. L., Chien, C-T., Sternglanz, R., and Fields, S., In ‘‘Cellular Interactions in Development: A Practical Approach’’ (D. A. Hartley, Ed.), pp. 153–179. Oxford Univ. Press, Oxford, UK, 1993. 20. Chien, C-T., Bartel, P. L., Sternglanz, R., and Fields, S., Proc. Natl. Acad. Sci. USA 88, 9578–9582 (1991). 21. Fields, S., and Song, O-K., Nature 340, 245–246 (1989). 22. Keegan, L., Gill, G., and Ptashne, M., Science 231, 699–704 (1986). 23. Ma, J., and Ptashne, M., Cell 48, 847–853 (1987). 24. Ma, J., and Ptashne, M., Cell 55, 443–446 (1988). 25. Li, B., and Fields, S., FASEB 7, 957–963 (1993). 26. Feilotter, H. E., Hannon, G. J., Rudell, C. J., and Beach, D., Nucleic Acids Res. 22, 1502–1503 (1994). 27. Prassad, B. V. V., Prevelige, P. E., Marietta, E., Chen, R. O., Thomas, D., King, J., and Chiu, W., J. Mol. Biol. 231, 65–74 (1993). 28. Kennard, J., Rixon, F. J., McDougall, I. M., Tatman, J. D., and Preston, V. G., J. Gen. Virol. 76, 1611–1621 (1995). 29. Preston, V. G., Al-Kobaisi, M. F., McDougall, I. M., and Rixon, F. J., J. Gen. Virol. 75, 2355–2366 (1994). 30. Matusick-Kumar, L., Newcomb, W. W., Brown, J. C., McCann III,
AID
VY 7959
/
6a18$$$103
05-16-96 19:14:14
31. 32. 33. 34. 35. 36. 37. 38.
39. 40. 41.
521
P. J., Hurlburt, W., Weinheimer, S. P., and Gao, M., J. Virol. 69, 4347–4356 (1995). Thomsen, D. R., Newcomb, W. W., Brown, J. C., and Homa, F. L., J. Virol. 69, 3690–3703 (1995). Hong, Z., Beaudet-Miller, M., Durkin, J., Zhang, R., and Kwong, A. D., J. Virol. 70, 533–540 (1996). Matusick-Kumar, L., Hurlburt, W., Weinheimer, S. P., Newcomb, W. W., Brown, J. C., and Gao, M., J. Virol. 68, 5384–5394 (1994). Nicholson, P., Addison, C., Cross, A. M., Kennard, J., Preston, V. G., and Rixon, F. J., J. Gen. Virol. 75, 1091–1099 (1994). Wood, L. J., Baxter, M. K., and Gibson, W., submitted for publication. Zweig, M., Heilman, C. J., and Hampar, B., Virology 94, 442–450 (1979). Desai, P., DeLuca, N. A., Glorioso, J. C., and Person, S., J. Virol. 67, 1357–1364 (1993). 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., J. Gen. Virol. 69, 1531–1574 (1988). Rose, M., Casadaban, M. J., and Botstein, D., Proc. Natl. Acad. Sci. USA 78, 2460–2464 (1981). Schacterle, G. R., and Pollack, R. L., Anal. Biochem. 51, 654–655 (1973). Miller, J. H., ‘‘Experiments in Molecular Genetics.’’ Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1972.
vira
AP: Virology