Insertional mutagenesis of the vaccinia virus gene encoding a type I DNA topoisomerase: Evidence that the gene is essential for virus growth

VIROLOGY

170,302-306

Insertional

(1989)

Mutagenesis of the Vaccinia Virus Gene Encoding a Type I DNA Topoisomerase: Evidence that the Gene Is Essential for Virus Growth STEWART SHUMAN,’ MICHAEL GOLDER, AND BERNARD Moss

Laboratory

of Viral Diseases, National Institute of Allergy and infectious

Diseases, National Institutes of Health, Bethesda, Maryland 20892

Received November 23, 1988; accepted January 2, 1989 Vaccinia virus encodes a type I DNA topoisomerase whose function in virus replication is not known. To determine whether topoisomerase is required for growth of vaccinia in cell culture, we attempted to isolate null mutations in the topoisomerase gene through insertional mutagenesis. Plasmids containing mutant topoisomerase alleles were constructed by intragenic insertion of the Escherichia coligpt gene. Recombinant viruses containing the gpt insertion were isolated by selection for growth in the presence of mycophenolic acid. Analysis of the genome structures of drugresistant viruses revealed that in every case (n = 22) both the wild-type and thegpf-inserted allele were present in viral DNA. We interpret the retention of the wild-type allele as indicative of the essential nature of the topoisomerase gene for VaCCinia Virus growth. 0 1999Academic Press,inc.

Vaccinia virus encapsidates within the infectious virion a DNA topoisomerase (I-4). The vaccinia enzyme resembles eukaryotic cellular type I enzymes (5) in its biochemical properties, including (i) relaxation of both positive and negative supercoils, (ii) stimulation by divalent cations, and (iii) formation of a cleavable complex with DNA consisting of enzyme covalently linked to the 3’ phosphoryl end of DNA. Vaccinia topoisomerase is distinguished from its cellular counterpart by its small size (n/l, 32,000) and its apparent resistance to camptothecin (S), a specific inhibitor of cellular type I topoisomerases (7). We have recently assigned the gene encoding the vaccinia DNA topoisomerase to the H7r open reading frame (ORF) of the viral genome (4), and have confirmed this assignment by expression of the vaccinia topoisomerase in active form in a heterologous system (6). The vaccinia topoisomerase gene encodes a polypeptide of 3 14 amino acids (4, 8) that contains a region homologous to the type I topoisomerase of yeast (9). Appreciation of the role of vaccinia topoisomerase in virus growth would be enhanced by a genetic approach, first by determining whether the topoisomerase gene is essential for viability, and, if so, then by isolating virus mutants affected specifically in topoisomerase function. With respect to the former point, it is not a foregone conclusion that the topoisomerase gene is required for viral replication. Vaccinia, like yeast (I&12), might not require any type I topoisomerase function for viability. Alternatively, even if a type I topo-

isomerase is essential, vaccinia may be able to utilize the topoisomerase I activity of the host cell in place of the viral enzyme (as is the case for another viral gene product, thymidine kinase), or it might even encode more than one DNA topoisomerase. The most straightforward proof that the topoisomerase gene that we have identified is essential would be to identify conditional lethal mutations that map to the H7r ORF among the available collection of temperature sensitive vaccinia mutants; the collection does include a complementation group that maps by marker rescue to the rightward end of the HindIll H genomic fragment (13), but, we have been unable to demonstrate marker rescue of this mutant by the topoisomerase gene (unpublished data). In the present study, we have tested the alternate hypothesis-that the topoisomerase gene is nonessential for growth in cell culture-by attempting to isolate null mutations through insertional mutagenesis. Vaccinia, by virtue of its relatively high frequency of homologous recombination (14- 16), is well-suited to targeted mutation of individual genes. The object is to replace the wild-type gene with a copy of the gene into which has been inserted a dominant selectable marker, so as to both inactivate the gene and provide a means of selecting recombinant viruses containing the mutant allele. We have used a selection scheme based on expression of the gpt gene of Escherichia co/i that was developed by Mulligan and Berg (77) and adapted to vaccinia by Falkner and Moss (18). Expression of gpt (encoding xanthine-guanine phosphoribosyltransferase) renders vaccinia resistant to the drug mycophenolic acid (an inhibitor of purine metabolism that blocks the replication of wild-type vaccinia in cell

’ To whom requests for reprints should be addressed at present address: Program in Molecular Biology, Sloan-Kettering Institute New York, NY 1002 1. 0042-6822189

$3.00

Copyright 0 1999 by Academic Press. Inc. All rights of reproduction in any form reserved

302

SHORT COMMUNICATIONS

ptopo$ptH2

(5.9kb)

FIG. 1. Construction of insertionally mutated topoisomerase alleles Plasmid ~1940 containing the entire topoisomerase structural gene in a pUCl9 vector (6) was linearized by cleavage at a unique Hincll site within the gene. The gpt cassette (containing the f. co/i gpt gene under the control of the vaccinia P7.5 promoter) was excised from plasmid pTK61 -gpt (18) and inserted into the topoisomerase gene to yield plasmids ptopo/gptH 1 and ptopo/gptH2 that differ in the orientation of the gpt transcription unit with respect to that of the topoisomerase gene. The direction of transcription is indicated by the arrows. The topoisomerase coding sequences (solid fill) flanking the inserted got cassette (hatched fill) are referred to as topo L (left) and topo R (right).

culture in the presence of xanthine). Two plasmids containing mutant alleles of the topoisomerase gene were constructed by inserting a DNA fragment containing the gpt gene (under the control of the vaccinia 7.5K promoter) at a unique Hincll restriction site within the topoisomerase coding sequence (Fig. 1). The mutant alleles differ in the orientation of the inserted gpt cassette with respect to the topoisomerase gene. These plasmids were transfected into virus-infected CV-1 cells as described (18) and recombinant viruses (n = 22) were isolated by two rounds of plaque-purification in medium containing mycophenolic acid, hypoxanthine, and xanthine. All of the MPA-resistant viruses obtained in this way should have a copy of the mutant topoisomerase allele containing the gpr insertion if, as expected, the crossover points for homologous recombination between the transfected plasmid and the viral genome lie within the topoisomerase sequences flanking the gpt gene in the plasmid. If the gene encoding the topoisomerase is nonessential then at least some of the MPA-resistant isolates should have a genotype where the wild-type gene has been replaced by the insertionally inactivated gene. As shown in Fig. 2, allelic replacement may be the consequence of two successive single recombinations, double crossover recombination, or gene con-

303

version events. We can predict, based on previous studies of vaccinia recombination by Spyropoulos era/. (15), that half of the MPA-resistant progeny should have the wild-type topoisomerase gene replaced by the mutated one if the topoisomerase gene is dispensable. If, however, the topoisomerase gene is essential, all of the nondefective recombinants should retain the wild-type allele (for viability) as well as the mutant allele (for drug resistance). The genomes of such viruses may have a nontandem repeat of the wild-type and mutant alleles separated by the plasmid vector DNA, a configuration resulting from a single crossover recombination event (Figs. 2c and d). Such nontandem repeats are not stable in the vaccinia genome, but should be selected for in the present case by use of MPA. Even with selection there will be a tendency (at the level of individual viral genomes within the infected cell) to segregate the single recombinant configuration to yield double recombinant and wild-type genomes (15). However, at the whole-population DNA level, both wild-type and mutant alleles may be present. We can readily determine the genome configuration of the recombinant viruses by Southern blotting of appropriately restricted DNA from virus-infected cells using labeled probes for the topoisomerase and gpt genes. Inserted and wildtype alleles are distinguished based on the size of the restriction fragments, as are wild-type, single crossover, and double crossover (or gene conversion) genomes. The predicted sizes of Spel and EcoRI restriction fragments of the topoisomerase loci are shown in Fig. 2. Twenty-two twice plaque-purified MPA-resistant virus isolates were amplified by growth in selective medium on BSC-1 monolayers and total cellular DNA was obtained for hybridization analysis. Figure 3 shows a Southern blot of EcoRI-digested DNA hybridized with a labeled topoisomerase gene probe. fcoRl cleaves wild-type vaccinia DNA upstream of the topoisomerase gene and at a single site within the gene (Fig. 2). The probe use in these experiments extends from nucleotide +12 in the topoisomerase gene to the intragenic fcoRI cleavage site. Wild-type viral DNA is expected to yield a single 2.6-kb fcoRl fragment capable of hybridizing with the probe, and indeed this was found to be the case (Fig. 3). The hybridization patterns of the MPA-resistant recombinants (eight of which are shown in the figure) are notable for the retention of the wildtype fcoRl fragment, as well as the appearance of several new fragments. The sizes of the new fragments (4.7, 3.1, and 1 .O kb) are consistent with those predicted to arise from recombination between the insertionally mutated plasmid and the genomic copy of the topoisomerase gene. Twenty-one out of twenty-two of the DNA samples analyzed contained all four predicted

304

SHORT COMMUNICATIONS

(b) transfected

plasmid

-

(d) sing*

e

4.7 -

cmssow

In tcqo A

Ecofil

EcoRl

Eat71

E&l

spel

Spel

u

+A*

w -

-

2.6 I

1.0 +

8.0

3.1

b

FIG. 2. Insertional mutagenesis of vaccinia topoisomerase: predicted genome structure of mycophenolic acid resistant recombinants. In this proposed recombination scheme, the wild-type vaccinia genome (a) recombines at the topoisomerase locus with homologous sequences in the transfected plasmid to generate direct nontandem duplications of the topoisomerase alleles, separated by plasmid (pUC) DNA. The product of a single crossover in topo L is depicted in (c); the product of a single crossover in topo R is shown in (d). Gene conversion or double crossover events resulting directly in allelic substitution will yield the structure shown in (e). The (e) genome configuration can arise also via intramolecular recombination of single recombinant genomes (c) or(d), so as to segregate double recombinant and wild-type genomes. The location of restriction sites for EcoRl and Spel are indicated by arrows above the DNA structure. The sizes (in kilobase pairs) of anticipated restriction products are indicated below the DNA molecules. More complex recombinant molecules (containing greater than two copies of topoisomerase alleles) can arise by intermolecular recombination events in products (c) and (d); such structures (not shown) would not contain any novel EcoRl restriction products beyond those depicted above.

fragments (4.7, 3.1, 2.6, and 1.O kb). The lone exception (Fig. 3, lane 4) containing only the 4.7- and 2.6kDa fragments is discussed below. The 4.7-kb EcoRl fragment can arise either from single recombinant genomes crossing over in topo L or from double recombinant genomes; similarly, the 2.6-kb fragment can derive from single recombinant genomes crossing over in topo R or from wild-type genomes. Digestion with EcoRl does not permit distinction between single recombinant and double recombinant molecules. The 1 .O- and 3.1 -kb fragments arise from genomes containing more than one topoisomerase allele, generated respectively, by crossing over in the left or right flanking sequences. The finding that there is often a greater degree of hybridization to the 1.O-kb fragment than to the 3.1 -kb fragment may reflect a higher rate of crossover in the left flank, owing to its larger size (topo L = 840 bp vs topo R = 247 bp). The occurrence of both 1.Oand 3.1-kb fragments in the progeny DNA population of plaque-purified viruses can be attributed to intermolecular crossovers between single recombinant genomes, rather than a failure to obtain a “pure” plaque. Such a randomization of genomic configurations has been noted previously by Spyropoulos et al. ( 75).

In order to examine more thoroughly the structure of the recombinant genomes, the DNAs were digested with Spel, an enzyme that cleaves outside the topoisomerase gene, and also fails to cleave either the got cassette or the plasmid vector DNA. Hybridization of a topoisomerase probe to Spel-treated wild-type DNA labels a single band of 2.1 kb (not shown), of mobility identical to that of the fragment designated as wt in Fig. 4. Once again, all MPA-resistant recombinants give rise to additional bands, the most prominent of which correspond in sizes to those predicted for single and double recombinant genomes (8.0 and 4.2 kb, respectively). Hybridization of Spel-digested DNAs with a probe specific for the got cassette labeled the 8.0- and 4.2-kb fragments, but not the 2.1-kb fragment, confirming that the genomes designated as single and double recombinants actually contain the insertionally mutated topoisomerase allele (not shown). Taken together, the data of Figs. 3 and 4 can reveal both the distribution of genomes (single recombinant, double, or wild-type) and the location of the primary crossover site, e.g., the plaque isolate in lane 5 can be construed to contain a mixture of wild-type DNAs and genomes derived from a single crossover in topo L. The excep-

305

SHORT COMMUNICATIONS wt

1

2

3

4

5

6

7

8

-5.9 -3.8

FIG. 3. Southern blot analysis of topo/gpt recombinants: EcoRl digest of infected cell DNA. CV-1 cell monolayers (5 X lo’/25 cm* plate) were inoculated with vaccinia virus strain WR at a multiplicity of 2. At 2 hr postinfection. cells were transfected with 5 pg plasmid DNA(ptopo/gptH 1 or ptopo/gptH2) and 2 rg wild-type vaccinia DNA, essentially as described (78). Recombinants carrying the gpt gene were selected by plating infected cell lysates on BSC-1 monolayers under soft agar overlay in the presence of 25 wglml mycophenolic acid (MPA), 250 pglml xanthine, and 15 @g/ml hypoxanthine. The yields of MPA-resistant virus due to transfection with the ptopolgpt plasmids was comparable to that obtained with the same amount of pTK50-gpt, a plasmid that targets insertion of the gpt cassette into the viral thymidine kinase gene. No MPA-resistant progeny were observed in mock-transfected controls. Individual plaques were picked and replated on BSC-1 cells under selective conditions. Twenty-two twice-plaque-purified isolates were chosen for genomic DNA analysis after further amplification by passage on BSC-1 monolayers. This group included 5 isolates derived from transfection with ptopo/gptH 1 and 17 recombinants arising from transfection with ptopo/gptH2. Plaques were suspended in 0.5 ml selective medium, frozen and thawed three times, sonicated, and then plated on BSC-1 monolayers (12-well plates) in liquid selective medium. Cells were harvested after 48-72 hr and concentrated by centrifugation. Cell lysis was achieved by treatment with hypotonic buffer containing 0.75% SDS. Lysates were digested with proteinase K, then extracted sequentially with phenol, phenol:chloroform, and chloroform. Nucleic acid was recovered by ethanol precipitation. Samples containing recombinant viral DNA or wild-type DNA purified from virus particles were digested with EcoRl and electrophoresed through an agarose gel. The structures of the topo gene in recombinant viral DNA were determined by DNA hybridization (19) after transfer of the nucleic acid to a membrane (Gene Screen). The labeled topoisomerase gene hybridization probe spanned the region from the intragenic Bsp1286 restriction site to the EcoRl site. An autoradiograph of the hybridized membrane is shown in the figure. The lane designated wt contains EcoRl digested wild-type DNA. Lanes l-3 contain DNA from cells infected with distinct MPA-resistant isolates derived from transfection with ptopo/gptHl, Lanes 4-8 contain DNA from cells infected with distinct MPA-resistant isolates arising from transfection with ptopoIgptH2. The positions and sizes of marker DNAs (in kilobase pairs) are indicated by the arrows.

tional virus isolate shown in lane 4 of Figs. 3 and 4 appears to contain only wild-type and double recombinant genomes, with little or no single recombinants.

This situation might have arisen by the picking of a mixed virus plaque, or by efficient segregation of a virus containing a single recombinant genome at an early stage during plaque formation. The key feature of this analysis is that none of the twenty-two recombinant viruses has lost the wild-type allele because of allelic substitution. This cannot be attributed to a failure to recombine appropriately, since double recombinant genomes are generated as part of the population of progeny DNA molecules. The present data on gpt insertion into the topoisomerase gene are in marked contrast with the findings of Falkner and Moss (18) that insertional mutagenesis of the nonessential vaccinia thymidine kinase gene readily yielded progeny containing exclusively double recombinant genomes. We therefore interpret the retention of the wild-type allele in the case of the topoisomerase as indicative of the essential nature of the topoisomerase gene for vaccinia virus growth. The exact role(s) served by the DNA topoisomerase in the vaccinia life cycle remains uncertain. The encapsidation of the enzyme within the virion has suggested a role in the transcription of early viral genes, a process that occurs in the virus particle and is mediated by a virus encoded RNA polymerase. Alternatively, the presence of the topo in virions may reflect its involvement in packaging the genome into the virion core during morphogenesis. Other potential roles include action as a swivel during DNA replication and/or participation in recombination and telomere resolution. Our finding that the topoisomerase gene is essential provides the impetus for further genetic analysis, including the gen-

12345678

single

-

y11.3 -8.6 L5.9

double -

wt -

-3.8

-1.9

FIG. 4. Southern blot analysis of topo/gpt recombinants: Spel digest of infected cell DNA. Samples containing amplified recombinant viral DNAs were digested with Spel and analyzed as described in the legend to Fig. 3. An autoradiogram of the membrane hybridized with a topoisomerase-specific probe is shown in the figure. The DNAs in lanes l-8 are derived from the samples described in Fig. 3. The positions and sizes of marker DNAs (in kilobase pairs) are indicated by the arrows.

306

SHORT COMMUNICATIONS

eration of conditional topoisomerase mutants by targeted mutagenesis. REFERENCES 1. BAUER, W. R., RESSNER,E. C.. KATE% J., and PATZKE,J. V., Proc. Nat/. Acad. Sci. USA74, 1841-1845 (1977). 2. FOGELSONG,P. D., and BAUER,W. R., J, Viral. 49, l-8 (1984). 3. SHAFFER,R., and TRAKTMAN, P., J. Biol. Chem. 262, 9309-9315 (1987). 4. SHUMAN, S., and Moss, B., Proc. Nat/. Acad. Sci, USA 84, 74787482 (1987). (1985). 5. WANG, J. C., Annu. Rev. Biochem. 54,665-697 6. SHUMAN, S., GOLDER, M., and Moss, B., J. Biol. Chem. 263, 16,401-16,407 (1988). 7. HSIANG, Y-H., HERTZBERG,R., HECHT, S., and LIU, L. F., 1. Biol. Chem. 27, 14,873-14,878 (1985). 8 ROSEL, J. L., EARL, P. L.. WEIR, J. P., and Moss, B., J. Viral. 60, 436-449 (1986).

9. THRASH, C., BANKIER,A. T., BARRELL,B. G., and STERNGLANZ,R., Proc. Natl. Acad. Sci. USA 78, 2747-2751 (1985). 10. THRASH, C., VOELKEL,K., DINARDO, S., and STERNGLANZ,R., J. Biol. Chem. 259,1375-1377(1984). 11. UEMURA, T., and MITSUHIRO,Y., EMBOJ. 3, 1737-l 744 (1984). 12. GOTO, T., and WANG, J. C., Proc. Natl. Acad. Sci. USA 82,717871 82 (1985). 13. THOMPSON, C. L., and CONDIT, R. C., Virology 150, 1O-20 (1986). 14. BALL, L. A., J. Viral. 61, 1788-1795 (1987). 15. SPYROPOULOS,D. D., ROBERTS,B. E., PANICALI, D. L., and COHEN, L. K., J. Viral. 62, 1046-1054 (1988). 16. EVANS, D., STUART, D., and MCFADDEN, G., J. viral. 62,367-375 (1988). 17. MULLIGAN, R., and BERG, P., Proc. Nat/. Acad. Sci. USA 78, 2072-2076 (1981). 18. FALKNER,F., and Moss, B., J. Viral. 62, 1849-l 854 (1988). 19. MANIATIS, T., FRITSCH,E. F., and SAMBROOK,J., “Molecular Cloning: A Laboratory Manual.” Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.