Analysis of in vitro activities of herpes simplex virus type 1 UL42 mutant proteins: correlation with in vivo function.

The DNA polymerase (pol) catalytic subunit of herpes simplex virus type 1, encoded by UL30, and its accessory factor, UL42 protein, are both essential for the replication of the virus. Because the stable interaction between UL42 and pol renders the pol fully processive for replicative DNA synthesis, disruption of this interaction represents a potential goal in the development of novel antiviral compounds. To better compare the effects of mutations in UL42 protein on its known in vitro functions, mutations were expressed as glutathione-S-transferase (GST)-fusions and the fusion proteins used in affinity chromatography. In this report, we demonstrate the relationship between the abilities of mutant UL42 fusion proteins to bind pol and to stimulate pol activity in vitro, and the abilities of nonfusion mutant proteins to function in viral replication. The pol stimulation assay using GST fusion proteins was found to be a more accurate and sensitive measure of the ability of the UL42 protein to function in vitro than the pol binding assay using the fusion proteins linked to a solid matrix. We also found an excellent correlation between the ability of purified GST fusion proteins to stimulate pol activity in vitro and the ability of full-length nonfusion UL42 mutant genes to support DNA replication in infected cells. Our results demonstrate that two noncontiguous stretches of amino acids, from 137 to 142 and from 274 to 282, are essential for UL42 function in vivo and in vitro. Although mutant d241-261 exhibited close to wild-type abilities to stimulate pol activity in vitro, it was not capable of complementing the replication of a UL42 null mutant virus. The region of UL42 protein within or close to 241-261 may serve to hinge the essential regions within the N- and C-terminal portions of the protein which are thought to interdigitate. It is hypothesized that reduction in the length of the hinge region could alter the ability of UL42, and/or its complex with pol, to function with one or more of the other proteins present in the DNA replisome within infected cells.

Structure-function analysis of the herpes simplex virus type 1 UL12 gene: correlation of deoxyribonuclease activity in vitro with replication function.

Although the product of the UL12 gene of herpes simplex virus type 1 (HSV-1) has been shown to possess both exonuclease and endonuclease activities in vitro, and deletion of most of the gene within the viral genome results in inefficient production and maturation of infectious virions, the function of the deoxyribonuclease (DNase) activity per se in virus replication remains unclear. In order to correlate the in vitro and in vivo activities of the protein encoded by UL12, mutant proteins were tested for nuclease activity in vitro by a novel hypersensitivity cleavage assay and for their ability to complement the replication of a DNase null mutant, AN-1. Rabbit reticulocyte lysates programmed with wild-type UL12 RNA cleaved at the same sites cleaved by purified HSV-1 DNase, but distinct from those cleaved by DNase 1 or micrococcal nuclease. All mutants which lacked DNase activity in vitro also failed to complement the replication of AN-1 in nonpermissive cells. Likewise, all mutants which contained HSV-1 DNase activity, as detected by the hypersensitivity cleavage assay, were capable of complementing the replication of the DNase null mutant, though to varying extents. Of particular note was the d1-126 mutant protein, which, despite having the same specific activity as the wild-type enzyme in vitro, complemented the replication of AN-1 significantly less than the wild-type protein. The results suggest that DNase activity per se is required for efficient replication of HSV-1 in vivo. However, residues, including the N-terminal 126 amino acids, which are dispensable for enzymatic activity in vitro may facilitate the accessibility or activity of the protein in vivo.

Interaction between the herpes simplex virus type 1 origin-binding and DNA polymerase accessory proteins.

Interactions between the herpes simplex virus type 1 (HSV-1) origin (ori)-binding protein (UL9) and two other components of the functional DNA replication complex have been observed. However, to date, no interaction between UL9 and a component of the DNA polymerase holoenzyme has been demonstrated. In this report, we demonstrate that UL9 and the DNA polymerase accessory protein (UL42) can form a stable complex in vitro as determined by coimmunoprecipitation with specific antibodies to each protein and by affinity chromatography using glutathione S-transferase (GST) fusion proteins. Complex formation does not require the presence of other viral proteins and occurs in the presence of ethidium bromide, indicating that UL9-UL42 interaction is DNA independent. Affinity beads charged with increasing concentrations of GST-42 fusion protein up to 5 microM bound increasing amounts of UL9 expressed by in vitro transcription/translation in rabbit reticulocyte lysates. Binding of N- and C-terminal portions of UL9 to GST affinity matrices revealed that the N-terminal 533 amino acids were sufficient for binding to GST-42, albeit at approximately a four- to six-fold reduced affinity compared to the full-length protein. No binding of a polypeptide containing the remainder of the UL9 C-terminal residues was observed. Thus the ori-binding protein, UL9, can physically associate with at least one member of each of the complexes (helicase/primase, DNA polymerase holoenzyme, single-stranded DNA-binding protein) required for origin-dependent DNA replication. These specific interactions provide a means by which the ordered assembly of HSV-1 DNA replication proteins at origins of replication can occur in the infected cell for initiation of viral DNA synthesis.

Cloning, sequencing, and functional characterization of the two subunits of the pseudorabies virus DNA polymerase holoenzyme: evidence for specificity of interaction.

The pseudorabies virus (PRV) genes encoding the two subunits of the DNA polymerase were located on the genome by hybridization to their herpes simplex virus type 1 (HSV-1) homologs, pol and UL42, and subsequently were sequenced. Like the HSV-1 homologs, in vitro translation products of the PRV gene encoding the catalytic subunit (pol) possessed activity in the absence of the Pol accessory protein (PAP). However, the PRV PAP stimulated the activity of Pol fourfold in the presence of 150 mM KCl, using an activated calf thymus DNA template. The stimulation of Pol activity by PAP under high-salt conditions and the inhibition of Pol activity by PAP when assayed in low salt (0 mM KCl) together were used to determine the specificity with which PAP interacted with Pol. Despite functional similarity, HSV-1 UL42 and PRV PAP could neither stimulate the noncognate Pols at high salt nor inhibit them at low salt. Furthermore, a PRV Pol mutant lacking the 30 C-terminal amino acids retained basal Pol activity but could be neither stimulated nor inhibited by the PRV PAP. Sequence comparisons of the Pol proteins of the alphaherpesviruses reveal a conserved domain in the C terminus which terminates immediately before the last 41 residues of both PRV and HSV-1 proteins. These results indicate that the ability and specificity for interaction of the PRV Pol with PAP most likely resides predominantly in the extreme Pol C terminus.

The essential in vivo function of the herpes simplex virus UL42 protein correlates with its ability to stimulate the viral DNA polymerase in vitro.

The product of the UL42 gene of herpes simplex virus type 1 (HSV-1) is an essential protein required for viral DNA synthesis in both transient origin of replication-dependent DNA replication assays and in virus-infected cells. In vitro, UL42 has been shown to form a heterodimeric complex with the 140-kDa protein product of the viral DNA polymerase (pol) gene. Although the pol gene possesses catalytic activity in vitro in the absence of UL42, UL42 stimulates pol activity presumably by increasing its processivity. In order to investigate whether the essential in vivo function for UL42 is related to its ability to associate with and modify pol activity, we have examined the ability of a UL42 null mutant, Cgal delta 42, to induce pol activity in nonpermissive Vero cells or permissive V9 cells. No detectable high salt-resistant pol activity was observed in Vero cells, although substantial activity was induced in V9 cells. Use of temperature-sensitive and host range mutants with defects in other genes revealed that failure to induce pol activity was due to neither direct nor indirect effects caused by lack of viral DNA synthesis. Furthermore, pol protein accumulated in Cgal delta 42 virus-infected nonpermissive cells with similar kinetics and to approximately the same level as in cells infected with wild-type virus. These results suggest a direct dependence on UL42 for pol activity. We also examined whether the same domains of UL42 affected the ability of the protein to stimulate pol activity in vitro and to complement the replication of Cgal delta 42. The excellent correlation between the activities of the mutant UL42 proteins in the in vitro pol stimulation assays and in the in vivo transient complementation assay indicates that the predominant in vivo role of UL42 is to provide pol accessory function, although additional essential functions for UL42 cannot be ruled out.

Association between the p170 form of human topoisomerase II and progeny viral DNA in cells infected with herpes simplex virus type 1.

Endogenous host topoisomerase II acts upon herpes simplex virus type 1 (HSV-1) DNA in infected cells (S.N. Ebert, S.S. Shtrom, and M.T. Muller, J. Virol. 56:4059-4066, 1990), and cleavage is directed exclusively at progeny viral DNA while parental DNA is resistant. To evaluate the possibility that HSV-1 induces topoisomerase II activity which could account for the preferential cleavage of progeny viral DNA, we assessed topoisomerase II cleavage activity on cellular and viral DNA substrates before and after the initiation of viral DNA replication. We show that cleavage of a host gene in mock-infected cells was similar to that observed in HSV-1-infected cells, regardless of whether viral DNA replication had occurred. In addition, quantitative measurements revealed comparable amounts of topoisomerase II activity in infected and mock-infected cells; thus, HSV-1 neither induces nor encodes its own type II topoisomerase and cleavages in vivo are due to a preexisting host topoisomerase. Human cells contain two isozymes of topoisomerase II (p170 and p180), encoded by separate genes. Through the use of isozyme-specific antibodies, we demonstrate that only p170 was found to be cross-linked to HSV-1 DNA even though both forms were present at nearly constant levels in HSV-1-infected cells. Immunofluorescence revealed that by 6 h postinfection, p170 becomes redistributed and localized to sites of active viral DNA synthesis. The data suggest that p170 gains preferential access to replicated viral DNA molecules, which explains why topoisomerase II activity is concentrated on progeny DNA.

Two regions of the herpes simplex virus type 1 UL42 protein are required for its functional interaction with the viral DNA polymerase.

Two essential gene products of herpes simplex virus type 1, the viral DNA polymerase (pol) and UL42, its accessory protein, physically and functionally interact to form the core of the viral DNA replication complex. Understanding this essential interaction would provide a basis from which to develop novel anti-herpesvirus agents. We previously have shown that when coexpressed in an in vitro transcription-translation system, UL42 stimulates pol activity (M. L. Gallo, D. I. Dorsky, C. S. Crumpacker, and D. S. Parris, J. Virol. 63:5023-5029, 1989). By analyzing various insertion, deletion, and frameshift mutations of UL42 in this system, we found the C-terminal 149 amino acids to be dispensable for the ability of the protein to stimulate pol activity. In addition, two distinct internal regions of UL42 were found to be required for pol stimulation. Regions I and II were defined to lie between amino acid residues 129 and 163 and between residues 202 and 337, respectively. When physical association was examined with antibody to UL42, pol was found to coimmunoprecipitate to the same level when expressed with a UL42 mutant protein lacking region I as that with wild-type UL42. Thus, mere physical association is insufficient for stimulation of pol activity. Deletion of region II reduced or eliminated coimmunoprecipitation with pol. Interestingly, an antibody to pol specific for residues 1216 to 1224 coimmunoprecipitated UL42 when both proteins were synthesized in a baculovirus expression system but not in rabbit reticulocyte lysates. These results indicate that (i) at least a portion of the region recognized by the pol antiserum may be accessible in the pol-UL42 heterodimer and (ii) immunoprecipitation results for products made in different expression systems may vary. Thus, at least two distinct regions of UL42 are essential for functional interaction with pol. Moreover, these results point to a UL42 region I function other than physical association with pol.

Isolation of a herpes simplex virus type 1 mutant deleted for the essential UL42 gene and characterization of its null phenotype.

We isolated a cell line, designated V9, stably transformed with the herpes simplex virus type 1 (HSV-1) UL42 gene, which is one of seven genes required in trans for the replication of plasmids containing an HSV origin of replication (C. A. Wu, N. J. Nelson, D. J. McGeoch, and M. D. Challberg, J. Virol. 62:435-443, 1988). V9 cells inducibly expressed the product of the UL42 gene, the 65-kDa DNA-binding protein (65KDBP), and were used as a permissive host to construct a mutant virus deleted for this essential gene. The UL42 deletion mutant, designated Cgal delta 42, displayed a tight early phenotype in nonpermissive Vero cells producing no infectious progeny, viral DNA, or late gene products but accumulated selected immediate-early and early transcripts with kinetics similar to those of wild-type virus. Wild-type levels of viral DNA and infectious progeny were produced in permissive V9 cells, despite the fact that V9 cells infected with Cgal delta 42 accumulated less than 1% of the UL42 RNA and protein found in Cgal+ virus-infected V9 or Vero cells. These results indicate that only small quantities of the 65KDBP are required for the synthesis of HSV DNA and the production of infectious virus. Although we could find no evidence that the superinduction of the 65KDBP in V9 cells infected with Cgal+ repressed expression of HSV-1 genes as observed in cells expressing another DNA-binding protein, ICP8 (P. K. Orberg and P. A. Schaffer, J. Virol. 61:1136-1146, 1987), the induction of the 65KDBP in V9 cells correlated with an approximately 2-h-earlier shift in the expression of genes from all three kinetic classes. The availability of the UL42 mutant should facilitate the construction of more subtle UL42 mutants which will be useful in the elucidation of the interrelationship between the 65KDBP and other DNA replication proteins as well as in the characterization of additional important functional domains.

Localization of the herpes simplex virus type 1 65-kilodalton DNA-binding protein and DNA polymerase in the presence and absence of viral DNA synthesis.

Using indirect immunofluorescence, well-characterized monoclonal and polyclonal antibodies, and temperature-sensitive (ts) mutants of herpes simplex virus type 1, we demonstrated that the 65-kilodalton DNA-binding protein (65KDBP), the major DNA-binding protein (infected cell polypeptide 8 [ICP8]), and the viral DNA polymerase (Pol) colocalize to replication compartments in the nuclei of infected cells under conditions which permit viral DNA synthesis. When viral DNA synthesis was blocked by incubation of the wild-type virus with phosphonoacetic acid, the 65KDBP, Pol, and ICP8 failed to localize to replication compartments. Instead, ICP8 accumulated nearly exclusively to prereplication sites, while the 65KDBP was only diffusely localized within the nuclei. Although some of the Pol accumulated in prereplication sites occupied by ICP8 in the presence of phosphonoacetic acid, a significant amount of Pol also was distributed throughout the nuclei. Examination by double-labeling immunofluorescence of DNA- ts mutant virus-infected cells revealed that the 65KDBP also did not colocalize with ICP8 to prereplication sites at temperatures nonpermissive for virus replication. These results are in disagreement with the hypothesis that ICP8 is the major organizational protein responsible for attracting other replication protein to prereplication sites in preparation for viral DNA synthesis (A. de Bruyn Kops and D. M. Knipe, Cell 55:857-868, 1988), and they suggest that other viral proteins, perhaps in addition to ICP8, or replication fork progression per se are required to organize the 65KDBP.

The essential 65-kilodalton DNA-binding protein of herpes simplex virus stimulates the virus-encoded DNA polymerase.

The 65-kilodalton DNA-binding protein (65KDBP) of herpes simplex virus type 1 (HSV-1), the product of the UL42 gene, is required for DNA replication both in vitro and in vivo, yet its actual function is unknown. By two independent methods, it was shown that the 65KDBP stimulates the activity of the HSV-1-encoded DNA polymerase (Pol). When Pol, purified from HSV-1-infected cells, was separated from the 65KDBP, much of its activity was lost. However, addition of the 65KDBP, purified from infected cells, stimulated the activity of Pol 4- to 10-fold. The ability of a monoclonal antibody to the 65KDBP to remove the Pol-stimulating activity from preparations of the 65KDBP confirmed that the activity was not due to a trace contaminant. Furthermore, the 65KDBP did not stimulate the activity of other DNA polymerases derived from T4, T7, or Escherichia coli. The 65KDBP gene transcribed in vitro from cloned DNA and translated in vitro in rabbit reticulocyte lysates also was capable of stimulating the product of the pol gene when the RNAs were cotranslated. The product of a mutant 65KDBP gene missing the carboxy-terminal 28 amino acids exhibited wild-type levels of Pol stimulation, while the products of two large deletion mutants of the gene could not stimulate Pol activity. These experiments suggest that the 65KDBP may be an accessory protein for the HSV-1 Pol.

Kinetics of expression of the gene encoding the 65-kilodalton DNA-binding protein of herpes simplex virus type 1.

The 65-kilodalton DNA-binding protein (65KDBP) of herpes simplex virus type 1, encoded by gene UL42, is required for herpes simplex virus origin-dependent DNA replication (C.A. Wu, N.J. Nelson, D.J. McGeoch, and M.D. Challberg, J. Virol. 62:435-443, 1988). We found by indirect immunofluorescence with monoclonal antibody to 65KDBP that the protein was first detectable at 3 h postinfection. It localized first to the inner periphery of the nucleus, but accumulated in large globular compartments within the nucleus by 6 h postinfection in a pattern similar to that displayed by the major DNA-binding protein ICP8. Immune electron microscopy revealed that 65KDBP was associated with the marginated heterochromatin at the early times, but migrated further into the nucleus at late times when the only discernible areas devoid of 65KDBP were the nucleoli and heterochromatin. The 65KDBP gene is a member of the beta kinetic class as determined by the ability of the mRNA to be expressed at significant levels even in the absence of viral DNA synthesis. Furthermore, in the presence or absence of the DNA polymerase inhibitor phosphonoacetic acid, the patterns of accumulation of protein as well as mRNA were virtually indistinguishable from those displayed by the model beta genes encoding ICP8 and thymidine kinase. Nuclear run-on experiments demonstrated that maximum rates of 65KDBP gene transcription occurred prior to the maximum rate of progeny viral DNA synthesis and confirmed that the expression of the 65KDBP gene is regulated at the level of transcriptional initiation.

Purification of the herpes simplex virus type 1 65-kilodalton DNA-binding protein: properties of the protein and evidence of its association with the virus-encoded DNA polymerase.

Using a combination of conventional column chromatography and velocity sedimentation, we have purified the 65-kilodalton DNA-binding protein (65KDBP) encoded by herpes simplex virus (HSV) greater than 625-fold. The HSV type 1 (HSV-1)-encoded DNA polymerase (pol) cofractionated with 65KDBP through DEAE-Sephacel, Blue Sepharose, and Mono Q columns and was only separated from 65KDBP by sedimentation through a glycerol gradient. Immunoaffinity columns containing monoclonal antibody (MAb) 6898 immunoglobulin effectively bound most of the HSV-1 pol activity which coeluted with 65KDBP. The pattern of reactivities of HSV-1/HSV-2 recombinants with MAbs specific for HSV-1 65KDBP or the HSV-2-infected cell-specific protein ICSP34,35 strongly suggests that these two species are serotype equivalents of the same protein. Taken together, all these data indicate that 65KDBP is a pol-associated protein and the HSV-1 counterpart of HSV-2 ICSP34,35 previously reported to have similar properties (P. J. Vaughan, D. J. M. Purifoy, and K. L. Powell, J. Virol. 53:501-508, 1985). Purified preparations of 65KDBP were capable of binding to double-stranded DNA, as determined by filter retention and mobility shift assays. The protein-DNA complex formed with 65KDBP was distinct from that produced by pol and could be further shifted by the addition of immunoglobulin specific for 65KDBP. These results demonstrate that 65KDBP has been purified substantially free from pol and indicate that DNA binding is an inherent property of the protein.

Identification of the gene encoding the 65-kilodalton DNA-binding protein of herpes simplex virus type 1.

Hybrid arrest of in vitro translation was used to localize the region of the herpes simplex virus type 1 genome encoding the 65-kilodalton DNA-binding protein (65KDBP) to between genome coordinates 0.592 and 0.649. Knowledge of the DNA sequence of this region allowed us to identify three open reading frames as likely candidates for the gene encoding 65KDBP. Two independent approaches were used to determine which of these three open reading frames encoded the protein. For the first approach a monoclonal antibody, MAb 6898, which reacted specifically with 65KDBP, was isolated. This antibody was used, with the techniques of hybrid arrest of in vitro translation and in vitro translation of selected mRNA, to identify the gene encoding 65KDBP. The second approach involved preparation of antisera directed against oligopeptides corresponding to regions of the predicted amino acid sequence of this gene. These antisera reacted specifically with 65KDBP, thus confirming the gene assignment.

The 65,000-Mr DNA-binding and virion trans-inducing proteins of herpes simplex virus type 1.

The possible identity of the herpes simplex virus type 1 (HSV-1) 65K (65,000-Mr) virion protein which stimulates transcription from immediate-early genes with the HSV-1 65K DNA-binding protein was investigated. The two proteins were found to be distinct by the three separate criteria of immunological reactivity, tryptic peptide fingerprinting, and mobility in two-dimensional gels. Using HSV-1/HSV-2 intertypic recombinants and a serotype-specific antiserum, we located the gene encoding the 65K DNA-binding protein between coordinates 0.574 and 0.682 on the HSV-1 genome. The protein is posttranslationally modified by phosphorylation. In crude extracts of HSV-1-infected cells the 65K trans-inducing protein did not detectably bind to double-stranded calf thymus DNA under the conditions of our assay.

Characterization of a herpes simplex virus type 2 deoxyuridine triphosphate nucleotidohydrolase and mapping of a gene conferring type specificity for the enzyme.

The herpes simplex virus type 2 (HSV-2)-induced deoxyuridine triphosphate nucleotidohydrolase (dUTPase) was purified approximately 600 +/- 43-fold using a combination of affinity, hydrophobic, absorption, and ion-exchange chromatography techniques. The only substrate for the dUTPase was dUTP with a Km of 3.6 +/- 1.1 microM. There was no apparent divalent cation requirement, but the HSV-2-induced dUTPase was inhibited by EDTA (0.1 mM) and this inhibition was reversed by either Co2+ (0.5 mM) or Mg2+ (0.5 mM). The HSV-2-induced dUTPase was distinguished from the HSV-1-induced and cellular dUTPases based upon differences in sensitivity to substrate inhibition, thermostability, and electrophoretic migration in nondenaturing polyacrylamide gels. Analysis of HSV-1 temperature-sensitive (ts) mutants demonstrated that ts A15 and ts K13 did not induce significant amounts of dUTPase activity at the permissive or nonpermissive temperatures. Mutants with defects in HSV-induced DNA polymerase or in the major DNA binding protein induced dUTPase at both temperatures. In contrast ts mutants defective in the alpha polypeptide VP175 (ICP4) did not induce normal levels of dUTPase at the nonpermissive temperature. The location of a gene encoding for the type specificity of the HSV induced dUTPase was mapped to the left 20% of the genome in Us in the region 0.060 to 0.100 or from 0.148 to 0.204.

Association of type I DNA topoisomerase with herpes simplex virus.

A topoisomerase activity is associated with herpes simplex virus type 1. The enzyme was recovered from purified virions which were disrupted with 6 M-guanidine-HCl followed by renaturation of extracted proteins. Based upon the following observations, the virion activity is classified as a type I topoisomerase: (i) the linking number of a unique DNA topoisomer is altered in steps of one; (ii) ATP and MgCl2 are not required for activity; (iii) the enzyme can be trapped in a covalent complex with DNA; (iv) the covalent linkage to DNA is through a 3' phosphoryl bond. A number of lines of evidence strongly indicate that the topoisomerase is external to the nucleocapsid. For example, the activity was released by treatment of intact virions with NP40, and subsequent washing steps extracted most residual activity. When guanidine extracts were prepared from nucleocapsids, topoisomerase activity was not detectable. Finally, DNA within the virion did not appear to contain covalently attached proteins with properties similar to topoisomerases. Thus, the enzyme appears to be a component of the envelope or tegument structure of the virion.

Detection of herpes simplex virus intertypic recombinant genomes in infected cell DNA.

Intertypic recombination between herpes simplex virus types 1 and 2 (HSV-1 and HSV-2) was detected using DNA from mixedly infected cells. Because HSV-1 and HSV-2 share a 50% base sequence homology along the genome but have markedly different DNA restriction enzyme cleavage patterns, recombination events can be detected and quantified by analysis of restriction endonuclease digests for the presence of novel DNA fragments. We have used this technique to quantify the degree of interference by HSV-2 on HSV-1 replication as well as the effect of limiting the availability of one genome on the frequency of intertypic recombination. Because this technique does not require production of viable progeny virions, it should also be useful for studying early recombination events.

Herpes simplex virus variants restraint to high concentrations of acyclovir exist in clinical isolates.

Acyclovir (ACV) has been shown to inhibit the replication of herpes simplex virus (HSV) in vitro. We examined a wide variety of HSV clinical isolates for the presence of naturally occurring ACV-resistant (ACVr) variants. Although the ACV doses that inhibited 50% of these isolates were within the range of doses inhibiting 50% of the ACV-susceptible wild-type strains, we successfully isolated variants resistant to high ACV concentrations (25 to 75 microM) from each virion population even in the absence of prior drug exposure. Furthermore, we demonstrated, by fluctuation analysis of two encephalitis strains, that the ACVr variants were clonally distributed in the virus populations before exposure to ACV and did not result from rapid adaptation to ACV. All variants isolated after a single exposure to a high dose of ACV were true ACVr variants, as demonstrated by their plating efficiencies in the presence of ACV. We found that 36 and 50% of the ACVr variants of the two strains examined in detail displayed plating efficiencies in phosphonoacetic acid of greater than 0.1, possibly indicating that many of the ACVr variants contained alterations in the DNA polymerase gene locus. Because the distribution of ACVr variants in natural populations is relatively high (10(-4), these results suggest that selection of ACVr strains during ACV therapy is possible.