Direct Determination of the Point Mutation Rate

Transcription

1 JOURNAL OF VIROLOGY, June 1992, p X/92/ $2./ Copyright , American Society for Microbiology Vol. 66, No. 6 Direct Determination of the Point Mutation Rate of a Murine Retrovirus RAYMOND J. MONK, FRANK G. MALIK, DAVID STOKESBERRY, AND LEONARD H. EVANS* Laboratory of Persistent Viral Diseases, Rocky Mountain Laboratories, National Institute ofallergy and Infectious Diseases, Hamilton, Montana 5984 Received 21 October 1991/Accepted 12 March 1992 The point mutation rate of a murine leukemia virus (MuLV) genome (AKV) was determined under conditions in which the number of replicative cycles was carefully controlled and the point mutation rate was determined by direct examination of the RNA genomes of progeny viruses. A clonal cell line infected at a low multiplicity of infection (2 x 1-3) was derived to provide a source of virus with high genetic homogeneity. Virus stocks from this cell line were used to infect cells at a low multiplicity of infection, and the cells were seeded soon after infection to obtain secondary clonal cell lines. RNase TI-oligonucleotide fingerprinting analyses of virion RNAs from 93 secondary lines revealed only 3 base changes in nearly 13, bases analyzed. To obtain an independent assessment of the mutation rate, we directly sequenced virion RNAs by using a series of DNA oligonucleotide primers distributed across the genome. RNA sequencing detected no mutations in over 21, bases analyzed. The combined fingerprinting and sequencing analyses yielded a mutation rate for infectious progeny viruses of one base change per 5, (2 x 1-5) bases per replication cycle. Our results suggest that over 8% of infectious progeny MuLVs may be replicated with complete fidelity and that only a low percentage undergo more than one point mutation during a replication cycle. Previous estimates of retroviral mutation rates suggest that the majority of infectious progeny viruses have undergone one or more point mutations. Recent studies of the mutation rates of marker genes in spleen necrosis virus-based vectors estimate a base substitution rate lower than estimates for infectious avian retroviruses and nearly identical to our determinations with AKV. The differences between mutation rates observed in studies of retroviruses may reflect the imposition of different selective conditions. The rate of evolution is the rate at which mutations accumulate in a genome and is frequently measured in terms of nucleotide substitutions per site per year. RNA viruses exhibit very high rates of evolution compared with those observed in eukaryotic DNA genomes (46, 49). For instance, it has been estimated that 1 to 2% of the nucleotides in the poliovirus genome have been altered by mutation during a period of 1 year of active spread (33). Similarly, the neuraminidase and hemagglutinin genes of influenza virus have been estimated to evolve at rates of.4 to.7% of their nucleotides per year (2, 28, 41, 43). These values compare with an average rate of evolution of approximately 1-7% per year for eukaryotic genes (3, 27). Retroviruses are reportedly among the most rapidly evolving viruses (9, 2-22, 34). The Moloney murine sarcoma virus genome contains a viral oncogene which originated from a homologous gene in the mouse genome, enabling a comparison of the evolution rate of the same gene in different genomic contexts. The rate of evolution of this viral oncogene (v-mos) was found to exceed the rate of evolution of its cellular homolog (c-mos) by greater than 1 millionfold (2, 21). The env gene of the human immunodeficiency virus (HIV) has been estimated to evolve at a rate of.1 to 1% per year (22), which also exceeds the average rate of evolution of eukaryotic DNA genes by about 1 millionfold. While the rate of evolution refers to the accumulation of mutations, usually expressed on an annual basis, the mutation rate refers to the number of mutations which occur during a replication cycle. The mutation rate of a virus, the * Corresponding author rate of replication of the virus, and the ability of the virus to sustain mutations without elimination from the population are all important factors in virus evolution. The mutation rate is determined by the fidelity of the polymerases involved in viral replication, the operation of repair mechanisms for the polymerases, and perhaps the cellular environment in which replication is taking place. Retroviruses are unique among RNA viruses in that they utilize RNA-dependent DNA, DNA-dependent RNA, and DNA-dependent DNA synthesis in their replicative cycle, and it has been suggested that the high rate of evolution of retroviruses may reflect a high rate of mutation during replication. Although nearly all mutation rate estimates of retroviruses have been high, almost all such estimates do not allow a precise determination of the point mutation rate in terms of misincorporation per base copied during a replication cycle. Coffin et al. (9) calculated the mutation rate of the avian Rous sarcoma virus (RSV), based on changes observed in RNase Tl-resistant oligonucleotides after repeated cell-free passage of virus. Although a precise mutation rate was not arrived at from these studies, a rate of 1-' to 1-4 was estimated, suggesting that each progeny virus differed from its parent by 1 to 1 bases. Leider et al. (26) determined a point mutation rate for RSV under conditions in which the number of replication cycles was carefully controlled and several defined regions of the RNAs of progeny viruses were examined. They reported a rate of 1.4 x 1-4 mutations per nucleotide during a single replication cycle for RSV, corroborating the estimate of Coffin et al. (9). Almost all studies of in vivo viral mutation rates have examined the rate in progeny viruses which are viable in terms of infectivity and replication. Temin and coworkers

2 3684 MONK ET AL. J. VIROL. Ag s:2. 'M; AKV I~1 i &' moa ) I \31-/ o_424 b I C- r Q ~23A D-22 - l9~32 isa A B D A 13C J SA 15 Q ' IA t FIG. 1. Fingerprint analysis of the RNase T1-resistant oligonucleotides of virion RNA from the progenitor clone of AKV. RNA extracted from 32P-labeled virions released from the progenitor clonal cell line was digested with RNase T1, and the oligonucleotides were resolved by fingerprinting. (A) Autoradiograph of the fingerprint. (B) Schematic depiction of the fingerprint with the number or letter designations of the oligonucleotides considered in the mutation rate analyses. The oligonucleotides which have numerical designations have been identified in the nucleotide sequence of AKV, while the oligonucleotides which have one-letter designations (A to N) are oligonucleotides which are present in the viral RNA genome but have not been unambiguously identified in the published sequence of AKV. (1, 11, 36, 37) have studied mutation rates of selectable marker or reporter genes contained in spleen necrosis virus (SNV)-based vectors and report base substitution rates near 1-5 per base pair per replication cycle. The selective conditions in the SNV system differ substantially from conditions which select genes obligatory for the generation of viable retroviruses; thus, the significance of differences in mutation rates observed with SNV-based vectors and those observed with viable progeny retroviruses is difficult to assess. The point mutation rate of a mammalian retrovirus in terms of mutations per nucleotide per replication cycle has not been previously determined. In the present study, we determined the point mutation rate of a murine retrovirus by direct examination of RNAs from viruses isolated after a single replication cycle. Our results indicate that the mutation rate for murine retroviruses is quite high compared with that for eukaryotic genes but is somewhat lower than estimates of the mutation rate of infectious progeny RSV. MATERIALS AND METHODS Cells and viruses. Mus dunni cells (25) were obtained originally from M. Cloyd, and NIH 3T3 cells were originally obtained from E. M. Scolnick. Both cell lines were maintained on minimal essential medium containing penicillin and supplemented with 5% fetal calf serum. The ecotropic murine leukemia virus (MuLV) AKR 2A (7) was obtained from M. Cloyd. Following infection, clonal cell lines were obtained by diluting the cells to approximately 2 cells per ml and plating 5 ml of the cell suspension on a 1-mm tissue culture dish. After attachment of the cells, the medium was replaced by fresh medium containing.4% Bactoagar (GIBCO) at 4 C and allowed to solidify at ambient temperature. The dishes were incubated undisturbed at 37 C until colonies of cells became macroscopically visible (ca. 1 mm in diameter). Our initial experiments to obtain a chronically infected cell line at a low multiplicity of infection were done with NIH 3T3 cells. However, during the course of this work, technical difficulties were encountered in establishing NIH 3T3 cells as single-cell colonies. Thus, M. dunni cells were used for the subsequent isolation of progeny viruses in our protocol (see below). Detection of infected clones. Infected colonies of cells were detected by a previously described immunofluorescence assay (45). Briefly, when the colonies became macroscopically visible, the agar-containing medium was removed by inverting the dish and the cells were rinsed with PBBS (5) containing 2% fetal calf serum. The dishes were then incubated at 37 C for 3 min with 2,ul of monoclonal antibody 83A25 (hybridoma culture supernatant), which is reactive with cell-surface MuLV envelope proteins (16). The dishes were rinsed twice and incubated with a fluorescein isothiocyanate-conjugated goat anti-mouse immunoglobulin serum (Sigma) at 37 C for 3 min. Finally, the dishes were rinsed

3 VOL. 66, 1992 MuLV POINT MUTATION RATE *lS 4 4c 1 7 / \ an O48U) 574 N R U5 DNA Primer Ollgonuclootides I. -env 3- U3 R a I I--- a I I I I I-a I kilobases FIG. 2. Genomic locations of RNase TI-resistant oligonucleotides and DNA primers employed in mutation analyses. The RNA genome of AKV is represented by the bar diagram. Gene boundaries and lengths are indicated by arrows and the scale aligned with the bar diagram, respectively. The locations of each of the numbered RNase T1 oligonucleotides and each of the DNA primers are indicated by lines drawn from the RNase T1 oligonucleotide numbers (above the bar diagram) or from the DNA primers which are designated by the base position of their 3' termini (below the bar diagram). three times and examined for fluorescence under a Leitz orthoplan fluorescence microscope. Infected colonies, as evidenced by cell surface fluorescence, were removed from the tissue culture dish and subcultured to obtain infected clonal cell lines. RNA isolation and TI-resistant oligonucleotide fingerprinting. Metabolic labeling of virion RNA with 32P and details of the fingerprinting procedure have been previously described (14). Virions were sedimented from tissue culture media in a Beckman SW 28 at 25, rpm at 4 C for 9 min. The pellet was resuspended in 1 mm Tris-1 mm NaCl-1 mm EDTA, ph 7.4 (TSE), and extracted with phenol-chloroform (9:1) several times until no precipitate was visible at the interface. The resultant aqueous phase was precipitated by the addition of an equal volume of n-propanol after adjusting to 1 mm NaCl. The nucleic acid precipitate was collected by sedimentation and dissolved in TSE containing.1% sodium dodecyl sulfate (SDS), layered onto a 12-ml 15 to 4% glycerol gradient in TSE-.1% SDS, and then sedimented at 39, rpm in a Beckman SW 41 rotor at 2 C for 3.25 h. The gradient was fractionated, and the peak of radioactivity sedimenting at approximately 7S was precipitated with n-propanol as described above. The resultant precipitate was collected by sedimentation, washed two times with 8% ethanol, resuspended in 1 mm Tris-1 mm NaCI-1 mm EDTA, ph 7.4, and stored at -2 C until fingerprinting. Virion RNA for use as a template for sequencing was prepared as described above except, after phenol extraction of the pelleted virions, the aqueous phase was applied to an oligo(dt)-cellulose column to obtain polyadenylated RNA as previously described (14). The polyadenylated RNA was precipitated with n-propanol, washed twice with 8% ethanol, dried, dissolved in water, and stored at -2 C. The RNA was thawed and used directly in the RNA sequencing procedure as described below. RNA sequencing. Retroviral RNAs were sequenced by using the protocol described by Mierendorf and Pfeffer (29) and the GemSeq Transcript Sequencing System (Promega Biotec). Oligonucleotides utilized as sequencing primers were synthesized on the Applied Biosystems Inc. model 38B. The sequencing gels were prepared and processed as described by Williams et al. (51), and the RNA sequencing data were analyzed by using the MBUG (NIH molecular biology users group integrated access system) computer program (4). RESULTS Establishment of a genetically homogeneous virus population and isolation of viruses after a single replicative cycle. NIH 3T3 cells were infected at a low multiplicity of infection with the ecotropic MuLV AKR 2A and seeded as single cells 2 h after infection, prior to the appearance of progeny viruses. The resultant colonies were examined by the immunofluorescence assay for cell surface expression of the viral envelope glycoprotein to identify those infected with MuLV. An MuLV-positive colony was subcultured from an infection in which only one fluorescent colony was observed among over 5 examined. The cell line derived from this colony has been designated the progenitor cell line. Since the integrated provirus in the progenitor cell line replicates via cellular DNA mechanisms, the errors introduced during replication are negligible (1-' to 1-12 per replication cycle) (3, 27) compared with rates reported for retroviruses (1' to 1-') (9, 26). Thus, virus released from the progenitor line should be of the highest obtainable genetic homogeneity. A precise knowledge of the number of replication cycles is necessary to accurately estimate the mutation rate. In this study we chose to analyze numerous progeny viruses isolated after a single round of replication. M. dunni cells were infected with progeny virus from the progenitor cell line, seeded 2 h after infection, and allowed to grow to macroscopic colonies. The colonies were assayed for infection, and the MuLV-positive colonies were subcultured as described above for the derivation of the progenitor cell line. A total of 93 infected clonal cell lines were isolated from cultures infected at a multiplicity of infection of 5 x 1-3. It

4 3686 MONK ET AL. A J% AKV B AKV- M1 AKVFM2 AKV - M3 r... 4~~~~~~~~~~~~~~4 so t~~~'t. FIG. 3. Comparison of RNase Tl-resistant oligonucleotide fingerprints of mutant virion RNAs with virion RNA from the progenitor cell line. The virion RNAs from the progenitor cell line and the three cell lines which harbored mutant viruses (AKV-Ml, AKV-M2, and AKV-M3) were analyzed by fingerprinting as described in the legend to Fig. 1. The oligonucleotides which are missing in the mutant RNAs (no. 75, no. 1, and no. 5 for AKV-Ml, AKV-M2, and AKV-M3, respectively) are circled on the fingerprint of the progenitor wild-type AKV (A). The original position of the missing oligonucleotides are indicated by the empty circles in the fingerprints of the mutants AKV-M1 (B), AKV-M2 (C), and AKV-M3 (D). In the cases of AKV-M2 and AKV-M3, a derivative of the original oligonucleotide was identified and is indicated by an arrow directed from the original location (empty circle) to the location of the derivative (circled oligonucleotide). is likely that each of the secondary lines was infected with an individual virus which had undergone a single replication cycle consisting of transcription by the cellular RNA polymerase in the progenitor cell line and reverse transcription to the proviral DNA by the viral polymerase in the secondary cell line. Mutations which occurred during these processes would be fixed in the provirus and present in the vast majority of virion RNA released from the secondary line. Any subsequent mutations which occurred randomly in the RNA transcribed from the provirus would not be detectable I9I J. VIROL. in analyses of total virion RNA released from the secondary cell line. RNase Tl-resistant oligonucleotide analyses of RNAs from virions released from the progenitor and secondary clonal cell lines. The MuLV AKV was chosen for these analyses because of the detailed knowledge of the genomic structure of the virus, in terms of both the complete nucleotide sequence (12, 23) and the precise genomic location and sequence of a large number of RNase T1-resistant oligonucleotides that are resolved by the fingerprinting procedures (12, 15). Eighty-nine oligonucleotides representing a total of about 1,38 bases were considered in the analyses (Fig. 1). Seventy-five of these oligonucleotides, totaling 1,228 bases, have been unambiguously identified in published AKV RNA sequences (12, 23) and are identified in Fig. 1. These oligonucleotides represent a sampling of many short (8- to 3-base) sequences across the entire genome (Fig. 2). The remaining resolved oligonucleotides (designated by the letters A to N in Fig. 1), representing about 15 bases, are present at stoichiometric levels but could not be unambiguously located in the published sequence. In the oligonucleotide fingerprinting analyses, the positions of oligonucleotides in the fingerprint are very sensitive to base composition. In most cases a single-base change within any of the large oligonucleotides results in an obvious change in the migration of the oligonucleotide. However, A-to-C and C-to-A base substitutions result in small changes in migration and are more difficult to detect, particularly if they occur in one of the larger oligonucleotides. Assuming that each of the 12 possible base exchanges occurs with equal probability and correcting for the average base composition of RNase T1-resistant oligonucleotides (ca. 31% A, 31% U, 31% C, and 7% G), 8% of the point mutations would be easily detected by the fingerprinting analyses even if all A-to-C and C-to-A changes went undetected. The fingerprint of the cloned isolate of AKV from the progenitor cell line was identical to that found in previous analyses of AKR 2A (13) (Fig. 1), indicating that mutations which affect the migration of the large oligonucleotides had not occurred during the initial cloning of the progenitor provirus. The fingerprints of virion RNAs from 9 of 93 secondary clonal cell lines analyzed were identical to that of the progenitor virus, while the fingerprints of 3 isolates were altered (Fig. 3). Each of these mutants differed from the progenitor MuLV by a small sequence difference within a single oligonucleotide, and in each case a different oligonucleotide was altered, indicating that the mutants were unique. Two of the mutations occurred in gag gene sequences, and the third occurred in the pol gene (Fig. 4). DNA primers complementary to sequences in the genome near the mutations were constructed, and the RNA sequences of the regions containing the altered oligonucleotides were determined for each of the mutants. These analyses (Fig. 5) indicated that each altered oligonucleotide differed from those of the progenitor virus by a single-base change which, in each case, resulted in a coding change to a different amino acid. Two of the changes were transition mutations, and the third was a transversion mutation. The observed change of oligonucleotide migration in the fingerprint of each mutant was consistent with the change predicted from the altered base sequence. The fingerprint analyses of 93 isolates represented a total of nearly 13, bases with only three detectable point mutations. Sequencing of RNAs from virions released from secondary clonal cell lines. The fingerprinting analyses suggested a mutation rate substantially lower for the MuLV than esti-

5 Vol. 66, 1992 MuLV POINT MUTATION RATE ED= I R U5 AKV-M2 1 I AKV-M1 AKV-M3 1 env- U3 R 3* AKV (Oligo. # 75) A C- A A A A A IiI AKV M1 A. T T AA A A G L Y AKV (Oligo. # 1) C T CT C C A A A C TC C C C C -C SE R kilobases AKV M2 1AA1 G -^_' FIG. 4. Genomic location of detected mutations. The RNA genome of AKV 2A is represented by the bar diagram. Gene boundaries and lengths are indicated by horizontal arrows and the scale aligned with the bar diagram, respectively. Arrows pointing to the bar diagram indicate the locations of the mutations detected in the three mutant progeny viruses. mates of mutation rates of infectious avian retroviruscs. To obtain an independent assessment of the mutation rate, several DNA oligonucleotides complementary to AKV sequences were synthesized and used to prime RNA sequencing reactions with many of the secondary AKV isolates as well as the progenitor AKV. The oligonucleotides were chosen from regions dispersed across the genome to maximize the probability that the sequences analyzed would reflect the mutation rate of the entire genome, rather than a limited region which may or may not be representative (Fig. 2). Several differences were noted between the sequences generated in our analyses and the published sequence of the molecularly cloned endogenous MuLV of AKR mice (AKV) (12, 23). However, all such differences were reflected in each of the viral RNAs analyzed and indicated differences between our progenitor MuLV and the published sequence of AKV rather than de novo mutations. With the exception of the three oligonucleotide mutations confirmed by sequence determinations, our RNA sequence analyses revealed no de novo mutations in over 21, bases analyzed. This is equivalent to over two complete MuLV RNA genomes without a single mutation, corroborating the data obtained from the fingerprinting analyses. Nearly 15, bases were analyzed by fingerprinting and sequencing in our experiments. The detection of only three mutations yielded an estimated mutation rate of 2. x 1-5. Assuming a Poisson distribution, the detection of three base changes in 15, predicts with 95% confidence that the actual mutation rate for AKV is between 5.85 x 1-5 and 4.13 x 1-6 (ca. between 1 mutation in 17, and 1 in 24, bases) (39). Even at the upper limit this is substantially less than one mutation per progeny virus. DISCUSSION Estimates of the mutation rate of infectious progeny avian retroviruses suggest that the virus may sustain 1 to 1 point mutations during a single replication cycle (9, 26), corresponding to an estimated mutation rate of 1.4 x 1-4. Our estimate for the point mutation rate of AKV (2 x 1-5) is somewhat lower and suggests that nearly 85% of the viable progeny viruses may be replicated with complete fidelity 3%) sustaining more than one point with few viruses (ca. mutation per replicative cycle. Poisson statistics predicted that the actual point mutation rate for AKV is between 1 in 17, and 1 in 24, bases replicated (39). Application of the same Poisson statistics to the mutation rate data obtained for RSV by Leider et al. (26) predicts that the point mutation rate for the avian virus is between 1 in 3,7 and 1 in 16,. Thus, there is a 95% expectation that the differences in mutation rates measured for the avian and murine viruses AKV (Oligo. # 5) C COAT C C A A -A- T x, AKV M3 FIG. 5. RNA sequences of the altered RNase TI-resistant oligonucleotides of the viral mutants. The RNA sequence of the altered oligonucleotide in each of the mutants is compared with the original sequence. Nucleotide identity is indicated by vertical lines between the AKV and mutant AKV sequences, and the altered base of each mutant is shown in bold lettering. The amino acids encoded by the original AKV and the AKV mutants are indicated above and below the sequences, respectively. are not due simply to random variation in the determinations. A number of factors could influence the differences in mutation rates observed for the avian and murine viruses. Since the point mutation rates determined for both retrovirus types are rates for infectious progeny viruses, mutations which result in the loss of the ability to independently replicate would not be detected. The ability to replicate may vary with different selective conditions. Furthermore, the ability to sustain mutations may vary among different rctroviruses or among different genes of a particular retrovirus. Recently, Temin and coworkers have studied mutation rates of genes inserted into SNV-based vectors during a single replication cycle (1, 11, 36, 37). They report a reversion frequency due to base substitution of an amber mutation in the neo gene of 2 x 1-5 (11) and a forward mutation rate resulting from base substitution of the lacz gene of 7 x 16 (36). Although these values are in very close agreement with our present determination of the point mutation rate of AKV, a comparison between the studies is difficult. In our studies, as in nearly all point mutation rate studies of animal viruses, viable progeny viruses were selected. Large genetic alterations such as frameshift mutations or large deletions, which are likely to be catastropic for viral replication or infectivity, would be infrequently detected. Furthermore, the point mutation rate determined for viable progeny viruses is expected to be lower than the actual rate of misincorporation of bases during the replicative cycle (the in vivo polymerase error rate). Under these conditions, the mutation rate largely reflects the incidencc of mutations which are functionally silent with regard to replication and infectivity. The mutation rates determined for SNV were determined at the level of the DNA provirus, thereby eliminating selection for viable progeny retroviruses. In addition to base substitution, those studies detected frameshift mutations, deletions, and hypermutation (an inordinately high level of base substitution on the same polynucleotide). It is somewhat surprising that the basc substitution rates observed in the absence of selection for viability were lower than rates which havc been reported for viable avian retroviruses (9, 26). However, the reversion and forward mutation studies of the SNV vectors involved selection for altered phenotypic functions of the neo and the lacz genes, respectively, and did not examine the incidence

6 3688 MONK ET AL. of functionally silent mutations. The selective stringency for base substitution imposed by selection for viability compared to that imposed by selection for altered functions of genes carried by the SNV vectors is unknown. Such stringeny would likely depend on the characteristics of the particular gene carried by the vector. Pathak and Temin (36) encountered a single provirus containing 15 G-to-A transition mutations corresponding to a mutation rate approximately 1,-fold higher than the rate determined for the virus population as a whole. They term this phenomenon hypermutation and suggest that it may be the result of an aberrant polymerase in the population. Frequent G-to-A hypermutation has also been reported for HIV type 1 (5). Although we did not observe hypermutations (or other extensive genetic alterations), it is possible that they are generated but not frequently selected as infectious progeny virus in our analyses. In this regard, it seems likely that a population of retroviruses which has been allowed to diverge may encode polymerases which vary in their fidelity. In that case, the mutation rate of a virus population would be an average rate and may differ from a mutation rate determined for a clonal isolate. The contribution of polymerase divergence to the population mutation rate is difficult to assess. A high degree of genetic recombination among retroviruses and the formation of viral pseudotypes may facilitate the maintenance of aberrant polymerases in the population. A considerable range of mutation rates have been reported among the RNA viruses. Steinhauer and Holland (48) observed a very high rate of mutation of about 7 x 1' base substitutions per replicative cycle for the vesicular stomatitis virus. Parvin et al. (35) have reported the mutation rate of the NS gene of influenza A virus to be 1.5 x 1-', while the rate observed for the VP1 gene of poliovirus was <2.1 x 1-6. The estimate for MuLV developed in this study appears to be near the middle of this range and close to that of the NS gene of influenza A virus. Parvin et al. (35) have suggested that the higher mutation rate of influenza A virus than of poliovirus may account for the rapid antigenic drift of influenza A virus and the necessity to frequently revise influenza A vaccines. The possible role of immune escape mechanisms has received considerable attention in the case of the lentiviruses such as the visna virus (6, 3, 31), equine infectious anemia virus (4, 24, 38, 44), and HIV (8, 22, 42, 47). Considering the difference between the point mutation rates of MuLV and RSV, it is difficult to extrapolate the data to the lentiviruses. However, in the case of the equine infectious anemia virus, the analyses of two isolates obtained sequentially from an infected horse at a 3-day interval differed by only.25% in their sequences (4). Ignoring in vivo virus replication which occurred, the extensive in vitro isolation procedures required to obtain these isolates (amplification by infectious spread in an in vitro cell line followed by multiple endpoint dilution cloning) suggest a mutation rate nearer that of MuLV than that of RSV. Finally, it is noted that the mutation rate during virus replication in fibroblastic cell lines might not extrapolate to virus replication in various cell types of the infected host. For example, lentiviruses, such as HIV, infect and replicate in macrophages which may serve as a reservoir for the viruses in an infected individual (17-19, 31, 32). Upon appropriate stimulation, macrophages undergo a respiratory burst generating high levels of free radicals which result in a highly mutagenic environment (1). Infection or replication J. VIROL. under these conditions could conceivably result in greatly elevated virus mutation rates. REFERENCES 1. Babior, B. M The respiratory burst of phagocytes. J. Clin. Invest. 73: Both, G. W., M. J. Sleigh, N. J. Cox, and A. P. Kendal Antigenic drift in influenza virus H3 hemagglutinin from 1968 to 198: multiple evolutionary pathways and sequential amino acid changes at key antigenic sites. J. Virol. 48: Britten, R. J Rates of DNA sequence evolution differ between taxonomic groups. Science 231: Carpenter, S., L. H. Evans, M. Sevoian, and B. Chesebro Role of the host immune response in selection of equine infectious anemia virus variants. J. Virol. 61: Chesebro, B., and K. Wehrly Studies on the role of the host immune response in recovery from Friend virus leukemia. I. Antiviral and antileukemia cell antibodies. J. Exp. Med. 14: Clements, J. E., F. S. Pederson,. Narayan, and W. A. Haseltine Genomic changes associated with antigenic variation of visna virus during persistent infection. Proc. Natl. Acad. Sci. USA 77: Cloyd, M. W., J. W. Hartley, and W. P. Rowe Lymphomagenicity of recombinant mink cell focus-inducing murine leukemia viruses. J. Exp. Med. 151: Coffin, J. M Genetic variation in AIDS viruses. Cell 46: Coffin, J. M., P. N. Tsichlis, C. S. Barker, S. Voynow, and H. L. Robinson Variation in avian retrovirus genomes. Ann. N.Y. Acad. Sci. 354: Dougherty, J. P., and H. M. Temin High mutation rate of a spleen necrosis virus-based retrovirus vector. Mol. Cell. Biol. 6: Dougherty, J. P., and H. M. Temin Determination of the rate of base-pair substitution and insertion mutations in retrovirus replication. J. Virol. 62: Etzerodt, M. T., Mikkelsen, F. S. Pedersen, N.. Kjeldgaard, and P. Jorgensen The nucleotide sequence of the Akv murine leukemia virus genome. Virology 134: Evans, L. H Characterization of polytropic MuLVs from three-week-old AKR/J mice. Virology 153: Evans, L. H., and M. W. Cloyd Generation of mink cell focus-forming viruses by Friend murine leukemia virus: recombination with specific endogenous proviral sequences. J. Virol. 49: Evans, L. H., and F. G. Malik Class II polytropic murine leukemia viruses (MuLVs) of AKR/J mice: possible role in the generation of class I oncogenic polytropic MuLVs. J. Virol. 61: Evans, L. H., R. P. Morrison, F. G. Malik, J. Portis, and W. Britt A neutralizable epitope common to the envelope glycoproteins of ecotropic, polytropic, xenotropic, and amphotropic murine leukemia viruses. J. Virol. 64: Gartner, S., P. Markovits, D. M. Markovitz, M. H. Kaplan, R. C. Gallo, and M. Popovic The role of mononuclear phagocytes in HTLV-III/LAV infection. Science 233: Gendelman, H. E.,. Narayan, S. Kennedy-Stoskopf, P. G. E. Kennedy, Z. Ghotbi, J. E. Clements, J. Stanley, and G. Pezeshkpour Tropism of sheep lentiviruses for monocytes: susceptibility to infection and virus gene expression increase during maturation of monocytes to macrophages. J. Virol. 58: Gendelman, H. E.,. Narayan, S. Molineaux, J. E. Clements, and Z. Ghotbi Slow, persistent replication of lentiviruses: role of tissue macrophages and macrophage precursors in bone marrow. Proc. Natl. Acad. Sci. USA 82: Gojobori, T., and S. Yokoyama Rates of evolution of the retroviral oncogene of Moloney murine sarcoma virus and of its cellular homologues. Proc. Natl. Acad. Sci. USA 82: Gojobori, T., and S. Yokoyama Molecular evolutionary

7 VOL. 66, 1992 rates of oncogenes. J. Mol. Evol. 26: Hahn, B. H., G. M. Shaw, M. E. Taylor, R. R. Redfield, P. D. Markham, S. Z. Salahuddin, F. Wong-Staal, R. C. Gallo, E. S. Parks, and W. P. Parks Genetic variation in HTLV-III/ LAV over time in patients with AIDS or at risk for AIDS. Science 232: Herr, W Nucleotide sequence of AKV murine leukemia virus. J. Virol. 49: Kono, Y., K. Kobayashi, and Y. Fukunaga Antigenic drift of equine infectious anemia virus in chronically infected horses. Arch. Gesamte Virusforsch. 41: Lander, M. R., and S. K. Chattopadhyay A Mus dunni cell line that lacks sequences closely related to endogenous murine leukemia viruses and can be infected by ecotropic, amphotropic, xenotropic, and mink cell focus-forming viruses. J. Virol. 52: Leider, J. M., P. Palese, and F. I. Smith Determination of the mutation rate of a retrovirus. J. Virol. 62: Li, W.-H., C.-I. Wu, and C.-C. Luo A new method for estimating synonymous and nonsynonymous rates of nucleotide substitution considering the relative likelihood of nucleotide and codon changes. Mol. Biol. Evol. 2: Martinez, C., L. Del Rio, A. Portela, E. Domingo, and J. Ortin Evolution of the influenza virus neuraminidase gene during drift of the N2 subtype. Virology 13: Mierendorf, R. C., and D. Pfeffer Direct sequencing of denatured plasmid DNA. Methods Enzymol. 152: Narayan, O., J. E. Clements, D. E. Griffin, and J. S. Wolinsky Neutralizing antibody spectrum determines the antigenic profile of emerging mutants of visna virus. Infect. Immun. 32: Narayan, O., S. Kennedy-Stoskopf, D. Sheffer, D. E. Griffin, and J. E. Clements Activation of caprine arthritis-encephalitis virus expression during maturation of monocytes to macrophages. Infect. Immun. 41: Narayan, O., J. S. Wolinsky, J. E. Clements, J. D. Strandberg, D. E. Griffin, and L. C. Cork Slow virus replication: the role of macrophages in the persistence and expression of visna viruses of sheep and goats. J. Gen. Virol. 59: Nottay, B. K.,. M. Kew, M. H. Hatch, J. T. Heyward, and J. F. ObiJeski Molecular variation of type 1 vaccine-related and wild poliovirus during replication in humans. Virology 18: O'Rear, J. J., and H. M. Temin Spontaneous changes in nucleotide sequence in proviruses of spleen necrosis virus, an avian retrovirus. Proc. Natl. Acad. Sci. USA 79: Parvin, J. D., A. Moscona, W. T. Pan, J. M. Leider, and P. Palese Measurement of the mutation rates of animal viruses: influenza A virus and poliovirus type 1. J. Virol. 59: Pathak, V. K., and H. M. Temin Broad spectrum of in vivo forward mutations, hypermutations, and mutational hotspots in a retroviral shuttle vector after a single replication cycle: substitutions, frameshifts, and hypermutations. Proc. Natl. Acad. Sci. USA 87: MuLV POINT MUTATION RATE Pathak, V. K., and H. M. Temin Broad spectrum of in vivo forward mutations, hypermutations, and mutational hotspots in a retroviral shuttle vector after a single replication cycle: deletions and deletions with insertions. Proc. Natl. Acad. Sci. USA 87: Payne, S., B. Parekh, R. C. Montelaro, and C. J. Issel Genomic alterations associated with persistent infections by equine infectious anemia virus, a retrovirus. J. Gen. Virol. 65: Pearson, E. S., and H.. Hartley Biometricka tables for statisticians, 2nd ed., vol. 1, p. 23. Cambridge University Press, New York. 4. Pearson, W. R., and D. J. Lipman Improved tools for biological sequence comparison. Proc. Natl. Acad. Sci. USA 85: Raymond, F. L., A. L. Caton, N. J. Cox, A. P. Kendal, and G. G. Brownlee Antigenicity and evolution amongst recent influenza viruses of HlNl subtype. Nucleic Acids Res. 11: Reitz, M. S., Jr., C. Wilson, C. Naugle, R. C. Gallo, and M. Robert-Guroff Generation of a neutralization-resistant variant of HIV-1 is due to selection for a point mutation in the envelope gene. Cell 54: Saitou, N., and M. Nei Polymorphism and evolution of influenza A virus genes. Mol. Biol. Evol. 3: Salinovich, O., S. L. Payne, R. C. Montelaro, K. A. Hussain, C. J. Issel, and K. L. Schnorr Rapid emergence of novel antigenic and genetic variants of equine infectious anemia virus during persistent infection. J. Virol. 57: Sitbon, M., J. Nishio, K. Wehrly, D. Lodmell, and B. Chesebro Use of a focal immunofluorescence assay on live cells for quantitation of retroviruses: distinction of host range classes in virus mixtures and biological cloning of dual-tropic murine leukemia viruses. Virology 141: Smith, D. B., and S. C. Inglis The mutation rate and variability of eukaryotic viruses: an analytical review. J. Gen. Virol. 68: Starcich, B. R., B. H. Hahn, G. M. Shaw, P. D. McNeely, S. Modrow, H. Wolf, E. S. Parks, W. P. Parks, S. F. Josephs, R. C. Gallo, and F. Wong-Staal Identification and characterication of conserved and variable regions in the envelope gene of HTLV-III/LAV, the retrovirus of AIDS. Cell 45: Steinhauer, D. A., and J. J. Holland Direct method of quantitation of extreme polymerase error frequencies at selected single base sites in viral RNA. J. Virol. 57: Steinhauer, D. A., and J. J. Holland Rapid evolution of RNA viruses. Annu. Rev. Microbiol. 41: Vartanian, J.-P., A. Meyerhans, B. Asjo, and S. Wain-Hobson Selection, recombination, and G-A hypermutation of human immunodeficiency virus type 1 genomes. J. Virol. 65: Williams, S. A., B. E. Slatko, L. S. Moran, and S. M. DeSimone Sequencing in the fast lane: a rapid protocol for [a-35s] datp dideoxy DNA sequencing. BioTechniques 4: