Streptococcus pyogenes - PNAS

Proc. Nat!. Acad. Sci. USA Vol. 88, pp. 3190-3194, April 1991 Genetics

Assembly and analysis of a functional vaccinia virus "amplicon" containing the C-repeat region from the M protein of Streptococcus pyogenes (homologous recombination/epitope duplication/antlgenlc variation)

DENNIS E. HRUBY*, OLAF SCHNEEWIND , ELIZABETH M. WILSON*, AND VINCENT A. FISCHETTIt *Center for Gene Research and Biotechnology, Department of Microbiology, Oregon State University, Corvallis, OR 97331-3804; and tThe Rockefeller University, 1230 York Avenue, New York, NY 10021-6339

Communicated by K. E. van Holde, December 20, 1990

that >70%6 of the protein is composed of separate blocks of tandem-sequence repeats (2). Development of a vaccine to prevent group A streptococcal infections, particularly pharyngitis, has been hampered by the facts that opsonic antibodies are type specific and that >80 serotypes have thus far been identified. Type-specific antibodies bind to the hypervariable amino-terminal end of the molecule distal to the cell surface; however, the carboxylterminal C-repeat region (CRR) located proximal to the cell surface is highly conserved among streptococci of many distinct serotypes (2). To determine whether antibodies to the conserved exposed epitopes of M protein influence the course of nasopharyngeal colonization by group A streptococci, peptides corresponding to these regions were used as immunogens in a mouse model (3). It was found that mice immunized intranasally with a cholera toxin-peptide complex showed a significant reduction in colonization compared with mice immunized with cholera toxin alone, and the protection was not type specific (4). As an alternative approach, genetically engineered recombinant vaccinia virus (VV) strains expressing all or part of the streptococcal M protein have been constructed for testing as potential live-virus vaccines (5, 6). In initial studies, sequences encoding the entire open reading frame of the M protein from serotype 6 S. pyogenes (M6) were inserted by recombination into the VV genome in a transcriptionally active configuration. The derived VV recombinant (VV:M6) was capable of expressing high levels of full-length antigenically authentic M protein in either infected tissue culture cells or experimental animals (5). Because the VV:M6 recombinant expressed the entire M protein, including the immunodominant epitopes and hypervariable regions found within the amino-terminal half of the M6 molecule (7), this formulation was not a suitable vaccine candidate. A second-generation VV recombinant (VV:M6') was, therefore, constructed that expressed only the region corresponding to sequences found within the conserved carboxyl-terminal half of the M6 protein. In animal trials, pharyngeal colonization by streptococci after intranasal challenge with these organisms was significantly reduced in mice immunized intranasally with the VV:M6' virus. M-proteinspecific serum IgG was markedly elevated in vaccinated animals and absent from controls. Most significantly, the protective immunity induced by the VV:M6' was also protective against challenge by a heterologous M14 streptococcal serotype (6). Because the VV:M6' recombinant virus exhibits obvious potential as an effective anti-streptococcal immunization vehicle, further development of this vector system to enhance antigenicity of the protective M protein epitopes seemed in order. In this communication we report on the

Previous studies have shown that when inocABSTRACT ulated intranasally into nmice, vaccinia virus (W) recombinants expressing the carboxyl half of the Streptococcus pyegenes M protein [which contains the C-repeat region (CRR)] could elicit a protective immune response against subsequent challenge by both homologous and heterologous serotypes of pathogenic group A streptococci. In the present study, an insertion plasmid was constructed that contained three tandem in-frame repeats of a 310-base-pair DNA sequence encoding the CRR from streptococcal M6 protein under control of a constitutive viral promoter. The plasmid was used to introduce the bacterial sequences into the W genome by homologous recombination. Surprisingly, the recombinant W:CRR3X virus that was isolated appeared to represent not an individual recombinant virus but a complex mixture of variants that contained from 1 to >20 tandem copies of the CRR region at the insertion site. This genomic complexity was mirrored at the transcriptional level in that a nested set of coterminal transcripts was detected in W:CRR3X-infected cells, which increased in size from 1400 to 6600 bases by increments of ==300 bases. All transcripts containing two or more CRR inserts appeared functional, as Western (immuno) blot analyses of W:CRR3X-infected cell extracts revealed a family of CRR-related proteins with apparent molecular masses that increased from 30 kDa upward in increments of 10 kDa. All data are consistent with the hypothesis that variation in the W:CRR3X recombinants is from random crossover events that occur within the CRR region during viral DNA replication. These results suggest that the genomic diversity generated by the "recombinogenic" properties of vaccinia recombinants containing tandem foreign inserts could be used to facilitate induction of a broadly protective immune response against antigenically diverse pathogenic agents.

Human diseases from infection by Streptococcus pyogenes (group A) remain a significant health problem. In the United States alone, 25 to 35 million cases of group A streptococcal infections, which primarily afflict school-age children, are reported annually (1). The high incidence and potential severity of streptococcal infections provide impetus for development of an effective and safe vaccine to prevent streptococcal-related diseases. Although the surface of the group-A streptococcus represents a complex antigenic mosaic, the ability of these organisms to cause infection is attributed to the M protein, a coiled-coil fibrillar molecule on the cell-wall surface that gives the organism the ability to resist phagocytosis. The translated nucleotide sequence of the M molecules reveals The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviations: VV, vaccinia virus; CRR, M protein C-repeat region.

3190

Genetics:

Hruby et al.

construction and molecular genetic analysis of a vaccinia recombinant that presents multiple copies of the CRR of the M6 molecule (VV:CRR3X). In examining the characteristics of the VV:CRR3X virus for use in our animal model we found that the virus exhibited some surprising and unusual "recombinogenic" properties that may enhance the effectiveness of the VV as a vector for similarly constructed antigens.

MATERIALS AND METHODS Cells and Virus. VV (WR strain) was grown and titered on BSC-40 cells as described (8). Construction of pW3:CRR3X. Plasmid pVV3:M6 (5) was digested with HincII/Pvu II and the 992-base-pair (bp) fragment of the emm6.1 gene that contained the CRR was subcloned into the Sma I site of the pMac5-8 vector (provided by H.-J. Fritz and K. Fiedrich, Max-Planck Institute for Biochemistry, Martinsried, F.R.G.). We used this template and two oligonucleotides (GAAACTTTGTTGAATTCATCTTTTTTAGC and CTTCAAGTTTGAATTCTAGCTCAGCT) in concert with the site-directed mutagenesis (9) to mutagenize a 310-bp fragment encoding the CRR region (amino acid residues 224-335) of the M6 protein flanked by two EcoRI sites. The 310-bp DNA fragment was sequenced, purified, and subjected to partial-ligation conditions. Multimers of the CRR were separated by preparative agarose gel electrophoresis. The 930-bp DNA fragment was ligated to the EcoRI site of expression vector pINIIIompA2 and transformed into E. coli DH1. Recombination of the coding sequences was confirmed by restriction analyses and by SDS/PAGE analysis of the expressed polypeptide followed by immunoblotting. The CRR trimer insert was released from the vector with Xba I/BamHI cuts and ligated into the BamHI site of pVV3 (10) after filling all overlapping ends with Klenow polymerase to yield pVV3:CRR3X. This designed protein was expressed from the VV 7.5-kDa constitutive promotor imbedded within the coding sequences of the VV thymidine kinase gene (tk). The ribosome-binding site, start codon, and outer membrane protein A (ompA) signal sequence come from the expression vector pINIIIompA2. The polypeptide ends with a 12-amino acid tail of the pVV3 polylinker region. Marker Transfer. The 7.5-kDa:CRR3X chimeric construction was introduced into the VV by homologous recombination, essentially as described (10) by using calcium phosphate-mediated transfection procedures (11) in concert with conditional-lethal VV mutants (12, 13). Potential recombinants were selected using 5-bromodeoxyuridine (14) and plaque-hybridization procedures (15). Analysis of Recombinant Virus. Viral DNA was isolated from wild-type or VV:CRR3X-infected cells as described (10), digested with the appropriate restriction endonucleases, and subjected to Southern blot hybridization procedures (16) by using a CRR-specific DNA fragment that had been labeled by the random-primed labeling procedure (17). After exposure to film, the blot was stripped and rehybridized with a radioactive probe fragment prepared similarly that was specific for the 5' half of the VV tk gene. Total cytoplasmic RNA was isolated from cells infected with either wild-type or VV:CRR3X virus in the presence of cycloheximide at 100 ,g/ml to amplify the expression of viral early mRNA species (18). The RNA was isolated from infected cells and purified by pelleting through CsCI gradients containing 1% N-lauroylsarkosine (19). The different RNA species were separated according to size by denaturing agarose/formaldehyde gel electrophoresis (20). After electrophoresis, the RNA was transferred to nitrocellulose and subjected to dual-probe hybridization, as described above. Monolayers of cells were infected with either wild-type or recombinant VV:CRR3X VV at a multiplicity of 20 plaque-

Proc. Natl. Acad. Sci. USA 88 (1991)

3191

forming units per cell and incubated at 370C for 18 hr. The infected cells were then harvested, washed twice with phosphate-buffered saline, and lysed in 150 /.l of SDS/gel loading buffer. The DNA was sheared by passing the extract through a 22-gauge needle five times. Ten microliters of each sample was separated by SDS/PAGE on mini-slab gels (21). After electrophoresis the proteins were electroblotted onto nitrocellulose and immunoblotted, as described (22) by using a monoclonal antibody directed against the CRR region of the M6 protein (23).

RESULTS AND DISCUSSION Construction of Recombinant W:CRR3X. From synthetic peptide studies (4) in conjunction with the VV:M6' results (6), the region responsible for cross-protection against streptococcal pharyngeal colonization was localized to the CRR of the M protein, which is located within the carboxyl-terminal half of the molecule. The amino acid sequence corresponding to the CRR is shown in Fig. lA. This 102-amino acid sequence is derived from residues 234-335 of the native M6 protein. Molecular cloning and site-directed mutagenesis-based procedures were used to construct an artificial gene in which an ATG codon in an appropriate mammalian translational context abutted an insert consisting of a 310-bp DNA fragment encoding the CRR, which had been triplicated in-frame, followed by a stop codon. This gene was inserted into the VV genome by DNA-mediated recombination using standard marker transfer techniques to produce the VV:CRR3X recombinant (Fig. 1B). Plaques arising from the VV:CRR3X recombinants were detected by plaque hybridization using a CRR-specific probe. The virus was plaque purified, and the viral DNA was extracted and analyzed by Southern blotting procedures. Genomic Structure of W:CRR3X Recombinant. Fig. 1B shows that a HindlII site is contained within each CRR unit, such that digestion with this enzyme should release the 310-bp CRR insert plus the chimeric 5' and 3'-flanking DNA fragments that would be expected to be 1.35- and 0.1-kilobase pairs (kbp), respectively. In contrast, Pst I endonuclease should cut external to the entire chimeric transcriptional unit and release a single 1.3-kbp fragment that can be detected by using a CRR-specific probe. Fig. 2 shows the results of digesting the starting pVV:CRR3X plasmid DNA, wild-type VV DNA, and VV:CRR3X DNA with either HindIII or Pst I restriction endonucleases. The agarose gel stained with ethidium bromide shows the DNA-fragment pattern (Fig. 2 Left). The 4.8-kbp HindIII J DNA fragment, which contains the target recombination site for the CRR:3X insert, is absent in the DNA from the VV:CRR3X recombinant. When this gel was transferred to nitrocellulose and analyzed with a tkspecific probe, DNA fragments of the predicted sizes (Fig. 1) were detected in all digests with HindIII and Pst I enzymes (Fig. 2 Center). In contrast, although the pattern of fragments detected in the VV:CRR3X HindIII digest was as predicted when using a CRR-specific probe, the pattern seen in the VV:CRR3X Pst I digest was highly aberrant (Fig. 2 Right). Instead of liberating an insert with an expected size of 1.3 kbp, a series of inserts that increased upward from 700 bp, in 300-bp increments, was apparent. In darker exposures, fragments as large as 6.6 kbp could be seen (data not shown). The higher-molecular mass bands seen are incomplete digestion products that disappear upon prolonged incubation (data not shown). Assuming the size variation is from recombination within the CRR unit, this variation would correspond to a range of 1-21 inserts. Thus, the VV:CRR3X recombinant is apparently not a single entity but rather represents a family of recombinants with variable numbers of CRR inserts. The apparent structure of the VV:CRR3X recombinant genome is shown at the bottom of Fig. 1B. Although the sequence of events responsible for generating the CRR insert heteroge-

3192


Genetics: Hruby et al.

A

XKKTAIAIAVALAGFATVAQ

AA

EFNKVSEASRRGLRRDLDASRBAKKQVEKDLANLTAELDKVKBBKQISDASR QGLRRDLDASREAKKQVEKALEEANSKLAALEKLNKELEESKLTESKEKAL

EFNKVSEASRKGLRRDLDASREAKKQVEKDLANLTAELDKVKBEEQISDASR QGLRRDLDASREAKKQVEKALHEANSKLAALBKLNKELEESKKLTEKEKAEL EFNKVSEABRKQLRRDLDASREAKKQVEKDLANLTAELDKVKEEKQISDASR

QGLRRDLDASRZAKKQVEKALBEANSKLAALEKLNKELEESKKLTEKEKAEL EFQAWIDPSISF

H

P

H

0

786

1355

H

H

1666

H

H

1978

2109

4549

2086

FIG. 1. Predicted genome structure of recombinant VV:CRR3X in the region containing the M protein CRR insert. (A) Translated amino acid sequence (in one-letter code) of the artificial gene from VV:CRR3X. The outer membrane protein A (ompA) signal sequence-cleavage site is indicated by the arrow. Each 104-residue M6 protein CRR segment is separately grouped and starts with the amino acid residues EF derived from the EcoRI linker. The protein ends with a 12-residue tail coded by the polylinker region. (B) Predicted structure of HindIll J region of recombinant VV:CRR3X viral genome. Letters above line indicate locations of HindIll (H) and Pst I (P) restriction endonuclease-cleavage sites. Numbers below line indicate rightward distance in base pairs from the leftmost HindIll-cleavage site, which corresponds to thejunction between the HindIII L and J fragments on the viral genome. Bold line on left and bold arrow at right indicate positions of 5' and 3' halves, respectively, of the viral tk gene serving as the genomic insertion site for CRR repeats. Position and orientation of the VV 7.5-kDa promoter element is shown by open box enclosing an arrow. CRR repeats are indicated by black boxes. Grey regions between CRR repeats correspond to EcoRI linkers that provide in-frame ligations. Hatched box represents a short region derived from the parental CRR plasmid that provides an in-frame ATG codon. Positions of the early (cross-hatched circle) and late (black circle) viral transcriptional start sites are shown as well as the transcriptional stop signal used during the early phase of infection (o). Positions of the initiator methionine (*) and translational stop codon (O) are similarly indicated. The genome below indicates structure of the VV:CRR3X recombinant genome (where n equals 1-20) as derived from molecular genetic analyses.

neity is unknown, the mechanism is probably similar to the unequal crossover mechanism proposed to account for the size diversity generated with the VV terminal repeats during normal viral replication (24). V

HindlIl

PstI

HindIIl P

R

P

V

Expression of CRR Sequences. To determine whether each member of the VV:CRR3X recombinant family was transcriptionally active, RNA-blot hybridization analyses were done on cytoplasmic RNA isolated from recombinant or

R

P

V

R

P

Pstl V

Psti

Hindill R

23.1--

P

-

V

R

P

V

R

31:1.1

*;

-9t

9.4-

6.7-

-h.

_

4.4-

-4.4

2.2-

.

2.0

2.0

f I-0.5b 0.56-

Agarose Gel

VV tk

CRR

FIG. 2. Southern blot analyses of genomic DNA from wild-type and recombinant VV:CRR3X VV. The parental insertion plasmid pVV3:CRR3X (P), as well as viral DNA isolated from either purified wild-type VV (V) or the recombinant VV:CRR:3X (R) were digested with HindI11 or Pst I; the resulting DNA fragments were resolved by agarose gel electrophoresis. After being visualized by staining with ethidium bromide, the DNA fragments were transferred to a nylon membrane and sequentially hybridized with 32P-labeled probes corresponding to an internal portion of the VV tk gene or the M protein CRR. Sizes in kbp and relative migration of bacteriophage A HindI11 DNA fragments included as markers are indicated.

Genetics:

Hruby et al.


wild-type VV-infected cells (Fig. 3). The filter was sequentially hybridized with radioactive probes specific for either VV tk or streptococcal CRR sequences to detect both the native viral tk gene transcript as well as any chimeric CRR:tk transcripts. Results shown in Fig. 3 Left indicate that, as expected (25), the tk-specific probe detected a 700-base transcript present in the RNA isolated from VV-infected cells. In contrast, when the RNA isolated from VV:CRR3Xinfected cells was analyzed, a collection of transcripts was detected that began at 300 bases in length and increased upward in size by increments of -300 bases. As shown in Fig. 3 Right, a CRR-specific probe detected the same pattern of transcripts in recombinant-infected cells but did not hybridize to any sequences in wild-type VV-infected cells. These results substantiate the conclusions reached above regarding structure of the VV:CRR3X recombinants and suggest that most, if not all, recombinants are transcriptionally active. Ability of the chimeric CRR:3X transcripts detected in Fig. 3 to be translated in vivo into protein was determined by immunoblotting infected cell extracts with a monoclonal antibody directed against the CRR. Fig. 4 indicates that no CRR-related proteins were detected in extracts from either mock-infected or wild-type VV-infected cells, whereas, as reported (6), the VV:M6' recombinant expresses a 29-kDa protein corresponding to the carboxyl half of the M6 protein containing the CRR. Proteins containing CRR epitopes are evident; these had apparent molecular masses of 30, 40, 50, 60, 70, and 80 kDa in the extracts from VV:CRR3X-infected cells. Sizes of these protein products are consistent with those predicted for VV:CRR3X recombinants containing two or more CRR inserts. By using longer periods of electrotransfer and/or lower dilutions of monoclonal antibody, CRR-related proteins with molecular masses >80 kDa could be detected (data not shown). These results suggest that each of the nested sets of CRR:3X transcripts has resulted from an V

R

R

v

- 9.5 -

7.5

-

4.4 -

-

- 2.4 - 1.77 - 1.52 -

1.40

1.28

=

- 0.78 - 0.53 - 0.40 - 0.24 -

tk

CRR

FIG. 3. Northern (RNA) blot analysis of the CRR insert-derived transcripts. Total cytoplasmic RNA was isolated from BSC-40 cells infected with either wild-type VV (V) or recombinant VV:CRR3X (R) virus with cycloheximide to amplify the concentration of early viral mRNA species. RNAs were separated according to size by denaturing agarose/formaldehyde gel electrophoresis. RNA was then transferred to a nylon membrane and sequentially hybridized with 32P-labeled probes corresponding to an internal portion of the VV tk gene or the M protein CRR. Numbers indicate sizes in kb and relative migration of RNA standards included in analysis as markers.

3193

x P

-

H

E

>

>

kDO 92 . 56 9-

4 6-

30-

21. 5-

FIG. 4. Western (immuno) blot analysis of CRR-related proteins expressed in VV:CRR3X-infected cells. Total detergent-soluble cytoplasmic extracts were prepared from mock-infected BSC-40 cells (MI) or cells infected with either wild-type VV (WT VV), a VV recombinant expressing the carboxyl halfof the M protein (VV:M6'), or the VV:CRR3X recombinant (VV:CRR3X). Extracts were separated by SDS/PAGE and then transferred by electroblotting onto a nitrocellulose filter membrane. To detect any CRR-related proteins the filter was immunoblotted by using a monoclonal antibody directed against the CRR region of the M protein as the primary antibody, and then horseradish peroxidase-conjugated anti-mouse serum was used as the secondary antibody. Numbers at left indicate sizes and relative migration of protein standards.

in-frame-crossover event that produces a functional message that is translated within the infected cell into a polyepitope protein containing immunoreactive CRR sequences. Interestingly, this hypothesis explains the previously observed size heterogeneity in the proteins expressed by VV recombinants expressing a plasmodial S-antigen containing tandem repetitive sequences (26). Stability of W:CRR3X Recombinant. Given that the VV:CRR3X isolate represents a complex mixture of CRRcontaining recombinants, whether the recombinant population was stable or in dynamic equilibrium was an interesting question. To address this question, individual VV:CRR3X plaques, which should have arisen from a single infectious viral particle containing a single DNA molecule with a fixed number of CRR inserts, were picked and grown under conditions that should allow only a single round of viral replication to amplify the viral DNA. The genomic DNA from nine individual VV:CRR3X plaques amplified in this manner was extracted and analyzed by Southern blot hybridization with a CRR-specific probe (Fig. 5). Relative to the parental VV:CRR3X DNA shown at right, a spectrum of different insert banding patterns was evident in the plaque isolates, ranging from the VV:CRR3X-G isolate, which had predominantly a monomer insert, to the VV:CRR3X-B isolate, which displayed a full range of CRR inserts from 1 to >20, with six repeats being the most prevalent. These results confirm that, although the CRR insert content of individual VV:CRR3X plaques differs, the recombination process continues to occur during each round of replication to expand the pool of CRR genomic diversity. This communication summarizes the construction and analysis of a VV recombinant (VVCRR:3X) that expresses multiple copies of the CRR from the streptococcal M protein. Although the virus displays unusual recombinogenic proper-

3194

Genetics: Hruby et al.


Plaque Isolates A B C D E F G H

I

A*

# 0

I R

'WI

23.1

40

a

-9.4

native molecule undergoes in vivo, although without selective pressure. Thus, VV recombinants such as VV:CRR3X may offer an opportunity to present the host with an antigenic mosaic for the induction of a broad array ofantibodies against a variable epitope of a pathogenic agent and, hence, provide more complete protection. Use of the recombinogenic properties of VV vectors may be of particular relevance with regard to developing effective vaccination strategies against infection by serotypically diverse pathogens such as human

immunodeficiency virus.

K,;

-

_

4.4 67

0

0-2.2 _

2.0

ILL FIG. 5. Genomic comparisons of individual VV:CRR3X plaque isolates. An appropriate dilution of VV:CRR3X was inoculated onto a BSC-40 cell monolayer under agar. After 48-hr incubation, progeny of individual plaques were picked and inoculated onto cells of a 24-well tissue-culture dish. Infections were allowed to proceed for 12 hr. Cytoplasmic DNA was then prepared and analyzed by Southern blot hybridization. Lanes A-I correspond to DNA derived from nine different plaque isolates. DNA from the parental VV:CRR3X recombinant (R) was included as reference. Sizes and relative migration of bacteriophage A HindIII DNA fragments included in the analysis as size markers are at right.

ties, this ability does not affect its ability to express immunoreactive CRR-containing proteins in infected cells. Therefore, this recombinant virus should prove useful in addressing the questions of (t) how the CRR contributes to the induction of cross-protective immunity and (ii) whether increasing epitope dosage augments or possibly inhibits the development of protective antibodies. In addition, these VV:CRR3X results may have more general implications for the use of VV as a recombinant vaccine. Our hypothesis is that polyepitope protein expression will simply require introduction of tandem copies of the desired coding sequence into the VV genome, after which the recombinogenic properties of the virus should operate on the "amplicon" to generate the level of diversity seen in our experiments. An alternative hypothesis, is that under certain conditions, the recombinant virus will delete the extra copies of the insert. In any case, high levels of recombination may be anticipated to introduce random mutations into the target sequence at a reasonable frequency. Although some of these mutations will undoubtedly be silent or introduce stop codons, others will introduce missense mutations. As such, the VV recombinant population should generate antigenic diversity, mirroring the process that the

This research was supported by Grants Al-00666 (D.E.H.) and AI-11822 (V.A.F.) from the National Institutes of Health and a grant from the Mallinckrodt Foundation to V.A.F. 1. Wannamaker, L. W. (1973) Circulation 48, 9-18. 2. Fischetti, V. (1989) Clin. Microbiol. Rev. 2, 285-314. 3. Bessen, D. & Fischetti, V. (1988) Infect. Immun. 56, 26662672. 4. Bessen, D. & Fischetti, V. (1990) J. Immunol. 145, 1251-1256. 5. Hruby, D. E., Hodges, W. M., Wilson, E. M., Franke, C. A. & Fischetti, V. A. (1988) Proc. Natl. Acad. Sci. USA 85, 5714-5717. 6. Fischetti, V. A., Hodges, W. M. & Hruby, D. E. (1989) Science 244, 1487-1490. 7. Fischetti, V. A. & Windels, M. (1988) J. Immunol. 141, 35923599. 8. Hruby, D. E., Guarino, L. A. & Kates, J. R. (1979)J. Virol. 29, 705-715. 9. Kramer, W., Drutsa, V., Jansen, H.-W., Kramer, B., Pflugfelder, M. & Fritz, H.-J. (1984) Nucleic Acids Res. 12, 9441-9456. 10. Rice, C. M., Franke, C. A., Strauss, J. H. & Hruby, D. E. (1985) J. Virol. 56, 227-239. 11. Graham, F. L. & van der Eb, A. J. (1973) J. Virol. 54, 536-539. 12. Condit, R. & Motyczka, A. (1981) Virology 113, 224-241. 13. Fahti, Z., Sridhar, P., Pacha, R. F. & Condit, R. C. (1986) Virology 155, 97-105. 14. Mackett, M., Smith, G. & Moss, B. (1984)J. Virol. 49, 857-864. 15. Villarreal, L. P. & Berg, P. (1977) Science 196, 183-186. 16. Southern, E. M. (1975) J. Mol. Biol. 98, 503-517. 17. Feinberg, A. P. & Vogelstein, B. (1983) Anal. Biochem. 132, 6-13. 18. Hruby, D. E. & Ball, L. A. (1981) J. Virol. 40, 456-464. 19. Glisin, V., Crkvenjakov, R. & Byus, C. (1974) Biochemistry 13, 2633-2637. 20. Lehrach, H., Diamond, D., Wozney, J. M. & Boedtker, H. (1977) Biochemistry 16, 4743-4751. 21. Studier, F. W. (1973) J. Mol. Biol. 79, 237-248. 22. Miner, J. N. & Hruby, D. E. (1989) Virology 170, 227-237. 23. Jones, K. F., Manjula, B. N., Johnston, K. H., Hollingshead, S. K., Scott, J. R. & Fischetti, V. A. (1985) J. Exp. Med. 161, 623-628. 24. Baroudy, B. M. & Moss, B. (1982) Nucleic Acids Res. 10, 5673-5679. 25. Hruby, D. E., Maki, R. A., Miller, D. B. & Ball, L. A. (1983) Proc. NatI. Acad. Sci. USA 80, 3411-3415. 26. Langford, C. J., Edwards, S. J., Smith, G. L., Mitchell, G. F., Moss, B., Kemp, J. & Anders, R. F. (1986) Mol. Cell. Biol. 6, 3191-3199.