Fidelity of reverse transcriptase of the simian ... - Europe PMC

0 downloads 0 Views 935KB Size Report
Andreas Manns, Herbert Konig', Michael Baierl, Reinhard Kurth1 and Frank Grosse*. Abteilung .... After 22 h at 20°C, 10 Al of the assay mixture was withdrawn ..... Morfeldt-Manson, L., Asjo, B. and Wain-Hobson, S. (1989) Cell 58,. 901-910.
Nucleic Acids Research, Vol. 19, No. 3 533

.=) 1991 Oxford University Press

Fidelity of reverse transcriptase of the simian immunodeficiency virus from African green monkey Andreas Manns, Herbert Konig', Michael Baierl, Reinhard Kurth1 and Frank Grosse* Abteilung Chemie, Max-Planck-Institut fOr experimentelle Medizin, D-3400 Gbttingen and 1Paul-Ehrlich-lnstitut, D-6070 Langen, FRG Received October 30, 1990; Revised and Accepted December 13, 1990

ABSTRACT The in vitro fidelity of highly purified recombinant reverse transcriptase from simian immunodeficiency virus of African green monkeys (SlVagm) was determined. By using the 4X174am16 reversion assay an overall error rate of 1/19,000 was determined. This is 2.4-fold higher than the overall accuracy of purified recombinant HIV-1 reverse transcriptase, measured in parallel. The evaluation of error frequencies from nucleotide pool bias studies suggest an even higher accuracy for the SlVagm-derived reverse transcriptase. T:dGMP mismatches were formed most frequently with an error rate of 1/155,000, followed by G:dGMP (1/230,000), A:dGMP (1/315,000), G:dAMP (1/340,000), T:dCMP (1/540,000), T:dTMP (1/790,000), and A:dCMP (1/1,050,000) mispairs. Thus, according to pool bias effects and depending on the mismatch under consideration SlVagm reverse transcriptase appears to be 2 to 20-fold more accurate than the homologous enzyme from the human immunodeficiency virus type 1. This higher accuracy is not due to a co-purifying exonuclaease activity. Like the enzyme from HIV-1, the simian monkey-derived enzyme was found to be devoid of a proofreading 3' to 5' exonuclease. INTRODUCTION Acquired immune deficiency syndrome (AIDS) in humans is caused by a lentivirus, the human immunodeficiency virus type-I (HIV-1) (1, 2). Closely related lentiviruses, designated as simian immunodeficiency viruses, have been isolated so far from five species of monkey: African green monkey (SIVagm) (3, 4, 5), mandrill (SIV,nd) (6), macaque (SIV,,.) (7), sooty mangabey (SIVsm) (8, 9, 10) and, recently chimpanzee (SIVcpz) (11). Despite the close relationship to HIV, many of these viruses presumably do not cause AIDS-like syndromes in their natural hosts. Nevertheless, some SIVagm isolates are able to infect rhesus monkeys and induce pathological changes like persistent peripheral lymphadenopathy in the infected monkeys (12). Hence, the infection of monkeys with SIVagm might lead to the development of an animal model for studies on vaccines and antiviral therapy relevant for AIDS. As a member of the group *

To whom correspondence should be addressed

of retroviruses, the SIV genome codes for a reverse transcriptase (EC 2.7.7.49) which is essential for viral growth in both ape and human hosts. Furthermore, this enzyme is the main target for antiviral therapies (13). Therefore, attempts have been undertaken to study the enzymological features of reverse transcriptases from simian monkeys and to compare these with the characteristics of the human-derived enzyme (14, 15). A special feature of reverse transcriptases is their low accuracy of synthesizing DNA. This might give a rational for the frequently observed high variability of the retroviral genomes (16). The reverse transcriptase from HIV-1 has been shown to be particularly error prone: In direct comparisons of several reverse transcriptases from various origins it was shown that HIV-1 RT was 3 to 10-fold less accurate than reverse transcriptases from avian and murine species (17, 18). Therefore, it was of interest to study how accurate the reverse transcriptase from a closely related simian host is. In the present paper it is shown that SIVagm RT, like HIV-1 RT, replicates singly-primed singlestranded DNA efficiently in vitro. This allowed the evaluation of the fidelities of seven base substitutions at the 4'X174 amber 16 codon. A comparison of the accuracies of RTs from the closely related SIVagm and HIV-1 might help to a better understanding of retroviral genomic variations. Moreover, it might give a clue on whether the SIVagm-infected rhesus monkey is a valuable model for studying HIV-1 infections in human beings.

MATERIALS AND METHODS Materials Reverse transcriptases derived from SIVagm 3 (5) and HIV-1 were purified to near homogeneity from overexpressing E. coli cells as described elsewhere (19, 20). Bacterial indicator strains E. coli C (wild type, sup' ), E. coli CQ2 (supE), E. coli C600 (supF, for the preparation of spheroplasts) and phage 4X174am16 were as described (21). The 14-base oligonucleotides 5'-CAGTAACTTTTCCC-3' and 5'-CCCAGCCTGAATCT-3' were synthesized on an Applied Biosystems model 380B oligonucleotide synthesizer, deprotected, 5'-phosphorylated and hybridized in a 5:1 ratio (primer/template) to 4X174wnm16 singlestranded DNA. All other chemicals sources.

were from commercial

534 Nucleic Acids Research, Vol. 19, No. 3 Methods RNA-dependent DNA Polymerase Assay SIV RT activity was measured as described elsewhere for the RT from HIV-1 (21). One unit of reverse transcriptase activity is defined as the amount of enzyme that catalyzes the incorporation of 1 nmol dTMP into poly(rA):oligo(dT) in 10 min at 37°C. The specific activity of purified HIV-1 RT was 9,500 U/mg, the specific activity of purified SIVagm RT was 16,700 U/mg. DNA-dependent DNA polymerase assay For the replication of singly-primed single-stranded DNA a 14-mer oligonucleotide, complementary to region AB5282 to AB5296 of the 4X174 genome (22), was annealed in a 5-fold molar excess to the 41X174 DNA (+) strand DNA. The assay mixture (100 1l) contained 50 mM Tris/HCl (pH 8.0), 50 mM NaCl, 10 mM MgCl2, 60 ,uM (nucleotide) singly-primed 4X174 (+) strand DNA, 30 yM (each) dNTPs and 1 yg RT. When pool biases were applied, the concentration of one of the four dNTPs was raised to 100 1iM, 300 ,tM, and 1 mM, respectively. After 22 h at 20°C, 10 Al of the assay mixture was withdrawn, heated to 70°C, and then analyzed on a 1 % agarose gel. Within this time > 95% of the DNA was replicated; when replication was completed the resulting double-stranded DNA was used for transfection. Fidelity assay Fidelity measurements were performed as described (21). Spheroplasts were produced from E. coli C600 (supF) by the method of Hussy et al. (23). The use of serum albumin from Sigma (RIA grade, lot 28F-02821) resulted in a 10-fold higher competence for the uptake of DNA than that observed before (21). Therefore, approximately tenfold more progeny plaques were characterized than in our previous measurement on the accuracy of HIV RT (21). This was necessary to avoid statistic variations because of the higher accuracy of SIV-RT compared to HIV-1 RT. Marker expression due to mismatch repair events inside the E. coli spheroplast was determined as described (24). A minus strand expression of 0.29 was found for a G:T mispair when unbiased nucleotide pools were used for DNA replication. An expression rate of 0.25 was found when dATP was present at a 30-fold higher concentration than the other three dNTPs. Therefore, the accuracy data were corrected for minus strand expression by an averaged factor of 0.27. Exonuclease assay 3' to 5' exonuclease activity was measured by incubating 1.2 yg purified SIV RT with 0.25 pmol (labeled 3' ends) of poly(dA) oligo(dT)20-[3H](dA), in a reaction volume of 50 Itl for 1, 3, 10, 30, 60, and 120 min, and analyzing the loss of radioactivity, exactly as described elsewhere (25).

RESULTS SIV RT replicates natural single-stranded DNA to the doublestranded form The DNA-directed DNA synthesizing ability of SIVa,n, RT was analyzed on singly-primed 4X 174 DNA. A synthetic oligonucleotide primer, which was complementary to the sequence position AB5282 to AB5296 (22) and contained a 3'-terminus three nucleotides in front of the amber 16 codon,

was hybridized to 4Xl74am16 DNA and subsequently elongated by SIVagm RT. The primer was elongated at 20°C with a linear kinetics for about 20 h (data not shown). An analysis of the replication products on agarose gels revealed that during this time replication was driven to near completeness, resulting in > 95 % of replicative form II.

Fidelity of simian monkey reverse transcriptase under conditions of unbiased nucleotide pools 4X174am16 DNA replicated in an assay mixture containing 30 ,uM of each of the four dNTP shows a reversion frequency of 5.4 + 1.1 10-5 (1/19,000) (six measurements). The resulting plaques were classified according to their size and their growth characteristics at different temperatures as described (26, 27). Wild-type (wt) and pseudo-wild type (,wt) phages were formed via T:dCTP and T:dGTP mispairs with a frequency of 1.2- 10-5 (1/85,000), ts42 was generated via an A:dGTP mispair with a frequency of 5.3 * 10-6 (1/190.000), ts38 was created via a G:dATP mispair with a frequency of 1.33 l0-s (1/75,000). The small sized plaques, representing the sum of ts35/ts34 (formed via G:dGTP and A:dCTP mispairs, respectively) occurred with a frequency of 8.85 10-6 (1/1 13,000). Thus, reverse transcriptase from SIVagm seems to be considerably more accurate than the homologous enzyme from HIV- 1, which was measured earlier with the same assay system (21). This was also true in parallel comparisons of the accuracies of HIV-1 RT and SIVagm RT. DNA was replicated in parallel with the simian- and the human-derived enzyme and then transfected to the same batch of spheroplasts. Reproducible overall accuracies, comprising 1/20,000 for SIV-RT and 1/8,000 for HIV-RT were measured at the amber 16 codon.

SIVn, RT is devoid of a 3'

to 5' exonuclease activity

The rather high accuracy of SIV RT prompted us to check whether exonucleolytic proofreading might contribute to the fidelity of this enzyme. After incubation of 1.2 yg SIV RT with 0.25 pmol (labeled 3' ends) of poly(dA) * oligo(dT)20-[3H](dA) for 2 hrs at 37°C, no detectable loss of the radioactive label was observed. From this it can be calculated that purified SIV RT removes less than 1 nucleotide per 106 nucleotides incorporated. In comparison, the exonuclease to polymerase ratio for DNA polymerase I from E. coli is about 1/200 (25). Thus, exonucleolytic proofreading is not responsible for the rather high accuracy of the SIVagm-enzyme.

Fidelity of SIV-RT under conditions of unbalanced nucleotide concentrations (pool bias) Reversions at the amber 16 codon of 4X 174 DNA can be forced into distinct pathways by biasing the concentrations of the four dNTPs during in vitro replication. Increasing dGTP over the other three dNTPs will favour the occurrence of T:dGMP, A:dGMP, and G:dGMP revertants, which are able to grow at 2 44°C (i wt), c 380C (ts38, and s 35°C (tS35), respectively. This protocol has been performed for all of the seven evaluable mismatches. All of the biases studied gave a linear response in the occurrence of the relevant phenotypes (Table I). The error frequencies for the occurrence of individual mismatches under unbiased conditions were derived from the slopes of the straight lines obtained from a graph of the measured reversion frequency over the pool bias applied upon in vitro replication of (X174am16 DNA (24). As has been also observed earlier with different enzymes (25, 28), a 30-fold excess of dGTP over the other three dNTPs led to a saturation behavior: there were less (ts35) or only

Nucleic Acids Research, Vol. 19, No. 3 535 Table I: Effect of pool biases on the fidelity of SIVagm RT.

Mispair induced

Phenotype evaluated

1:1 3:1 10:1 30:1

T:dGTP

4Wt

1:1 3:1 10:1 30:1

A:dGTP

1:1 3:1 10:1 30:1

G:dGTP

1:1 3:1 10:1 30:1

T:dCTP

1:1 3:1 10:1 30:1

A:dCTP

1:1 3:1 10:1 30:1

T:dTTP

1:1 3:1 10:1 30:1

G:dATP

Pool Bias applied

Ratio

GI

Ct

TI

At

c

c

tS42

tS35

Wt

ts34

ts43

tS38

Reversion frequency a

efor rate b

R

[ x 10-6] 2.9 6.8 18.8 (24.8)

1/155,000 (1/385,000)

0.999 (0.829)

1.3 3.4 9.2 (12.6)

1/315,000 (1/750,000)

0.998 (0.847)

1/230,000

0.999

1/540,000

0.978

2.6 5.3 13.1 no revertants 3.2 3.2 7.2 17.3 2.1 3.2 3.7 9.8 10.7 16.1 16.4 22.7 10.1 13.3 10.3 33.3

SIVagm

1/1,050,000

1/790,000

1/340,000

0.978

0.845

0.882

a The reversion frequency is given by (number of the evaluated mutant) + (number of total plaques) . 0.27. b The rate laws were calculated from the slopes of the straight lines of plots of the reversion frequency against the nucleotide pool bias applied. R2 values of these plots are as indicated in the rightmost column. c Because an increase of dATP and an increase of dTTP over the other three nucleotides leads unambiguously to G:dATP and T:dTTP mispairs respectively, the total increase in plaques with the increase of these two biases was evaluated without any phenotypic characterization of the resulting phages.

a few more revertant phages (4'wt, ts42) produced at a 30-fold bias of dGTP compared to the corresponding 10-fold bias. In these cases, error rates were calculated from the 1 to 10-fold nucleotide pool bias (Table I). Error rates that include the evaluation of the 30-fold biases are given in parentheses (Table I). The error frequencies derived from pool bias studies are in a reasonable agreement with error rates determined at equal concentrations of all the four nucleoside triphosphates and subsequent phenotypic characterization of the revertants (see above).

DISCUSSION The formation of complementary DNA from genomic RNA is a crucial step in the life cycle of retroviruses. Errors made by reverse transcriptase, which catalyzes DNA formation from RNA, are at least partially responsible for the high variability of the genomic information. The RT from HIV-1 has been shown to be extremely error prone. Base substitution errors of as high as 1/2,000 to 1/4,000 were observed with a 4XI74am3-based reversion assay (29). Frame shift errors were even more prominent: in an M13mp2-based forward mutation assay these kinds of errors were as high as 1/70 at particular sites of the

heterologous lacZ indicator gene (18). Moreover, in a direct comparison of various RTs from several different origins it was shown that HIV-1 RT was approximately 3 to 10-fold less accurate in vitro than homologous RTs from murine and avian retroviruses (17, 18). This raised the question of how accurate the RT from a HIV-1-related retrovirus, i.e. SIVagm, might be. In order to determine the fidelity of SIVagm RT, reversion mutations at the amber 16 codon were scored after single-stranded DNA of bacteriophage cIX174 has been replicated in vitro with purified RT from SIVagm. To exclude variations in the biological assay system as well as in the biochemical reactions leading to the formation of double-stranded DNA, we repeated several of the experiments done in the earlier study (21) in order to get a direct comparison of the accuracies of the HIV-1 and the

SIVagm-derived enzymes.

The overall nucleotide substitution rate of SIVagm RT at the amber 16 codon of 4)X174 DNA was measured to be 5.4 4 1.1 10-5 (1/19,000). Similar error frequencies have been measured for reverse transcriptases from murine and avian origins (18, 30, 31). In contrast, error rates of 1/2,000 to 1/7,000 have been determined for HIV-1 RT (21, 29). This rather low accuracy of the HIV-enzyme was confirmed in this study. A two- to threefold higher accuracy of the simian-derived enzyme was

536 Nucleic Acids Research, Vol. 19, No. 3 measured in a direct comparison of the two RTs from SIVagm and HIV-1. The higher accuracy of the SIVagm enzyme was even more pronounced upon studying the rate laws for base substitution errors by using unbalanced nucleotide concentrations. Such an analysis revealed that SIVagm RT requires between 2-fold (in the case of G:dGTP errors) and 20-fold (in the case T:dGTP errors) higher concentrations of the unbalanced nucleotide to become as erroneous as the homologous enzyme from HIV-1. All the errors observed with SIV RT occurred within one order of magnitude (10-5 and 10-6), whereas errors produced by HIV RT varied much more between the evaluated mispairs (l0-4 to 10-6). Although T:dGTP errors were most frequent with both RTs, this type of error was approximately 20-fold less prevalent with the enzyme encoded by SIVagm as compared to HIV-1. A comparison of the fidelities of AMV RT and HIV- 1 RT by using an M13-based forward mutational assay revealed that the difference between these two enzymes was also most prominent for G:T mispairs; T:dGTP mispairs were formed 20-fold more frequently by HIV-1 RT than by AMV RT, G:dTTP errors occurred about 5-fold more frequent with the avian-derived enzyme (31). A higher accuracy for SIVag, RT compared to HIV- 1 RT was unexpected because SIVagm has been reported to display an unsual high genetic diversity (32). The reasons for genetic diversity may be manifold, but it is likely that the error rate of the replicative reverse transcriptase plays a major role in this process. Therefore, finding differences in the in vitro accuracy between two closely related RTs from closely related viruses raises the question as to whether these differences are real or may be due to artifacts of the in vitro assay system. To avoid experimental artifacts that might arise from the rather complicated biological assay system, many experiments have been done in parallel with the RTs from both HIV-1 and SIVagm. These experiments clearly revealed a higher accuracy of the SIV polymerase. This does not exclude, however, that there are other explanations for the more accurate behavior of the SIVagmderived enzyme compared to the enzyme from HIV-1. One is that SIVagm RT is contaminated by a co-purifying 3' to 5' exonuclease from E. coli. Such an explanation was excluded by using a highly sensitive assay for measuring a potential proofreading activity. Another possibility is that the HIV-1 enzyme, determined earlier and used in this study, has lost fidelity e.g. because of heterologous expression or purification from the E. coli host. This seems unlikely because, like the HIV-derived enzyme, the simian enzyme was heterologously expressed and purified from E. coli. Moreover, a direct comparison of the fidelities of authentic (not recombinant) HIV RT with RTs from various origins also revealed a lower accuracy for the HIV-1 enzyme compared to enzymes isolated from murine and avian organisms (17). Thus, it appears that the fidelity of the SIVagm enzyme seems to be similar to the fidelities of the murine and avian-derived enzymes, whereas HIV-1 RT seems to belong to a less accurate class of reverse transcriptases. Clearly, several aspects of in vivo replication, such as the influence of nucleic acids binding proteins, is not being represented in in vitro assays. Therefore, it could not be solved by this study whether the remarkable difference in the accuracy of two highly related polymerases from closely related viruses is of physiological significance. Furthermore, it is well known that different locations of a given DNA sequence give rise to different degrees of mutations depending on the enzymes used (33) as well as on the assay system (18, 21, 29) and even on the conditions of the assay by using one and the same enzyme and one and the same assay

system (34, 35). Moreover, in a recent study it was shown that the sites of mutations as observed in vitro were not reflected by the same stretch of target sequence under in vivo conditions (36) (it must be pointed out, however, that in this particular study in vivo mutation sites of spleen necrosis virus were compared with in vitro mutation sites of AMV and HIV-1 RT, obtained in a different laboratory (31)). Nevertheless, it should be kept in mind that two highly related enzymes have been studied in parallel under comparable conditions and that there is at least a chance for a biological significance of the different accuracies observed in vitro. If nucleotide selection by HIV-RT should be also less accurate in the infected cell, this might give a rational for the observed high selectivity of 1-[(2-hydroxy-ethoxy) methyl]-6-phenylthiothymine (HEPT) derivatives (37) as well as tetrahydro-imidazo[4 ,5, 1 -jk] [1 ,4]-benzodiazepin-2( 1 H)-one (TIBO) derivatives (38) to HIV-1 RT only. Furthermore, it has been suggested that a high error rate of HIV-1-RT maximizes the probability of producing mutants which are able to escape the counteracting immune system (39). The observed rapid evolution of retroviral quasispecies in infected patients (40) apparently supports this speculation. Again, a higher in vivo error rate of HIV- 1 RT might be a conditio sine qua non for an enhanced evolutionary rate of HIV-RT and the observed fatal effects of this retrovirus to infected individuals. A further substantiation of these intriguing speculations probably requires further studies to clarify the in vivo relevance of our findings, e.g. by using manipulated HIV-1 and SIVagm retroviruses containing a forward mutatable target sequence (36, 41).

ACKNOWLEDGEMENTS We are grateful to F. Benseler for preparing the oligonucleotides. We thank M. Kruh0ffer for his generous gift of HIV- 1 RT. The expert technical assistance of T. Weski is gratefully acknowledged. We thank D. Williams, J. Taylor and F. Eckstein for critically reading the manuscript and many stimulating discussions. This work was supported by grants 111-003-89 and 11-097-89 from the Bundesministerium far Forschung und

Technologie.

REFERENCES 1. Barre-Sinoussi, F., Cherman, J. C., Rey, R., Nugeyre, M. T., Chamaret, S., Gruest, J., Dauguet, C., Axler-Blin, C., Vezinet-Brun, F., Rouzioux, C., Rozenbaum, W. and Montagnier, L. (1983) Science 220, 868-871. 2. Gallo, R. C., Safin, P. S., Gelmann, E. P., Robert-Guroff, M., Richardson, E., Kalyanaraman, V. S., Mann, D. L., Sidhu, G., Stahl, R. E., ZollaPazner, S., Leibowitch, J. and Popovic, M. (1983) Science 220, 865-867. 3. Ohta, Y., Masuda, T., Tsujimoto, H., Ishikawa, K., Kodama, T., Morikawa, S., Nakai, M., Honjo, S. and Hayami, M. (1988) Int. J. Cancer 41, 115-122. 4. Daniel, M. D., Li, Y., Naidu, Y. M., Durda, P. J., Schmidt, D. K., Troup, C. D., Silva, D. P., MacKey, J. J., Kestler III, H. W., Sehgal, P. K., King, N. W., Ohta, Y., Hayami, M. and Desrosiers, R. C. (1988) J. Virol. 62, 4123-4128. 5. Kraus, G., Werner, A., Baier, M., Binninger, D., Ferdinand, F. J., Norley, S. and Kurth, R. (1989) Proc. Natl. Acad. Sci. USA 86, 2892-2896. 6. Tsujimoto, H., Cooper, R. W., Kodama, T., Fukasawa, M., Miura, T., Ohta, Y., Ishikawa, K., Nakai, M., Frost, E., Roelands, G. E., Roffi, J. and Hayami, M. (1988) J. Virol. 62, 4044-4050. 7. Daniel, M. D., Letvin, N. L., King, N. W., Kannagi, M., Sehgal, P. K., Hunt, R. D., Kunki, P. J., Essex, M. and Desrosiers, R. C. (1985) Science 228, 1201-1204. 8. Fultz, P. N., McClure, H. M., Anderson, D. C., Swenson, R. B., Anand, R. and Srinivasan, A. (1986) Proc. Natl. Acad. Sci. USA 83, 5286-5290. 9. Lowenstine, L. J., Pederson, N. C., Higgins, J., Pallis, K. C., Uyeda, A., Marx, P., Lerche, N. W., Munn, R. J. and Gardner, M. B. (1986) Int .J. Cancer 38, 563-574.

Nucleic Acids Research, Vol. 19, No. 3 537 10. Murphey-Corb, M., Martin, L. N., Rangan, S. R. S., Baskin, G., Gormus, B. J., Wolf, R. H., Andes, W. A., West, M. and Montelaro, R. C. (1986) Nature 321, 435-437. 11. Huet, T., Cheynier, R., Meyerhans, A., Roelants, G. and Wain-Hobson, S. (1990) Nature 345, 356-359. 12. Herchenr6der, O., Stahl-Hennig, C., Like, W., Schneider, J., Schulze, G., Hartmann, H., Schmidt, H., Tenner-Racz, K., Racz, P., Hayami, M., Kelliher, J. C. and Hunsmann, G. (1989) Intervirology 30 (Suppl. 1), 66-72. 13. De Clercq, E. (1990) Trends Pharmac. Sci. 11, 198-205. 14. Like, W., Hoefer, K., Moosmayer, D., Nickel, P., Hunsmann, G. and Jentsch, K.-D. (1990) Biochemistry 29, 1764-1769. 15. Kraus, G., Behr, E., Baier, M., K6nig, H. and Kurth, R. (1990) Eur. J. Biochem. 192, 207-213. 16. Starcich, B. R., Hahn, B. H., Shaw, G. M., McNeely, P. D., Modrow, S., Wolf, H., Parks, E. S., Parks, W. P., Josephs, S. F., Gallo, R. C. and Wong-Staal, F. (1986) Cell 45, 637-648. 17. Takeuchi, Y., Nagumo, T. and Hoshino, H. (1988) J. Virol. 62, 3900-3902. 18. Roberts, J. D., Bebenek, K. and Kunkel, T. A. (1988) Science 242, 1171-1173. 19. Baier, M., Behr, E., Kurth, R. and Konig, H. (1990) Arzneim.-Forsch./Drug Res. 40 (II), 1284-1287. 20. Heidenreich, O., Kruh0ffer, M., Grosse, F. and Eckstein, F. (1990) Eur. J. Biochem. 192, 621-625. 21. Weber, J. and Grosse, F. (1989) Nucleic Acids Res. 17, 1379-1394. 22. Sanger, F., Air, G. M., Barrell, B. G., Brown, N. L., Coulson, A. R., Fiddes, J. C., Hutchinson, C. A., HI, Slocombe, P. M., Smith, M., Drouin, J., Friedman, T. and Smith, A. J. H. (1978) In The Single Strand Phages. (Denhardt eds.) pp. 659-669, Cold Spring Harbor Press, Cold Spring Harbor, N. Y. 23. Hussy, P., Tiimmler, B., Maass, G., Grosse, F. and Schomburg, U. (1986) Antimicrob. Agents Chemother. 6, 1073-1078. 24. Grosse, F., Krauss, G., Knill-Jones, J. W. and Fersht, A. R. (1983) EMBO J. 2, 1515-1519. 25. Bialek, G., Nasheuer, H.-P., Goetz, H. and Grosse, F. (1989) EMBO J. 8, 1833-1839. 26. Fersht, A. R. and Knill-Jones, J. W. (1983) J. Mol. Biol. 165, 633-654. 27. Fersht, A. R. and Knill-Jones, J. W. (1981) Proc. Natl. Acad. Sci. USA 78, 4251-4255. 28. Rienitz, A., Grosse, F., BlUcker, H., Frank, R. and Krauss, G. (1985) Nucleic Acids Res. 13, 5685-5695. 29. Preston, B. D., Poiesz, B. J. and Loeb, L. A. (1988) Science 242, 1168-1171. 30. Roberts, J. D., Preston, B. D., Johnston, L. A., Soni, A., Loeb, L. A. and Kunkel, T. A. (1989) Mol. Cell. Biol. 9, 469-476. 31. Bebenek, K., Abbotts, J., Roberts, J. D., Wilson, S. H. and Kunkel, T. A. (1989) J. Biol. Chem. 264, 16948-16956. 32. Johnson, P. R., Fomsgaard, A., Allan, J., Gravell, M., London, W. T., Olmsted, R. A. and Hirsch, V. M. (1990) J. Virol. 64, 1086-1092. 33. Ricchetti, M. and Buc, H. (1990) EMBO J. 9, 1583-1593. 34. Eckert, K. A. and Kunkel, T. A. (1990) Nucl. Acids Res. 18, 3739-3743. 35. Tindall, K. R. and Kunkel, T. A. (1988) Biochemistry 27, 6008-6013. 36. Pathak, V. K. and Temin, H. M. (1990) Proc. Natl. Acad. Sci. 87, 6019-6023. 37. Baba, M., Tanaka, H., De Clercq, E., Pauwels, R., Balzarini, J., Schols, D., Nakashima, H., Perno, C.-F., Walker, R. T. and Miyasaka, T. (1989) Biophys. Biochem. Res. Commun. 165, 1375-1381. 38. Pauwels, R., Andries, K., Desmyter, J., Schols, D., Kukla, M. J., Breslin, H. J., Raeymaeckers, A., Van Gelder, J., Woestenborghs, R., Heykants, J., Schellekens, K., Janssen, M. A. C., De Clercq, E. and Janssen, P. A. J. (1990) Nature 343, 470-474. 39. Nowak, M. (1990) Nature 347, 522-522. 40. Meyerhans, A., Cheynier, R., Albert, J., Seth, M., Kwok, S., Snisky, J., Morfeldt-Manson, L., Asjo, B. and Wain-Hobson, S. (1989) Cell 58, 901-910. 41. Pathak, V. K. and Temin, H. M., (1990) Proc. Natl. Acad. Sci. 87, 6024-6028.