Pyrimidine Dimers as Premutagenic Lesions: a Study of Targeted vs. Untargeted Mutagenesis in T H E Laci Gene of Escherichia Coli

We have employed conjugal transfer of an F’ lac episome to examine targeted and untargeted mutagenesis in the lacl gene of Escherichia coli and to determine the relative importance of pyrimidine dimers as premutational UV lesions compared to (6-4) photoproducts that also may have a mutational role. This conjugal system allowed us to assess the premutagenic role of UV lesions independently from any role as inducers of SOS functions. F’ DNA was transferred to an SOS-induced recipient strain from: unirradiated donor cells, UVtreated donor cells or donor cells that were irradiated and then exposed to photoreactivating light. The results indicate that SOS-related, untargeted events may account for as much as one-third of the nonsense mutations (i.e., base substitutions) recovered after undamaged F‘ DNA is transferred to UVirradiated recipients. When the donor strain also is irradiated, in excess of 90% of the mutations detected following conjugation appear to be targeted. Photoreactivation of the UV-treated donor cells, prior to F’ transfer to the SOSinduced recipient strain, demonstrated that in this experimental system virtually all recovered UV-induced mutations are targeted by photoreactivable lesions. We presume that these lesions are pyrimidine dimers because (6-4) photoproducts are not photoreactivable. LTRAVIOLET light (UV)-induced mutagenesis has been studied in EschU erichia coli for more than a third of a century. Much has been learned about the kinetics and host requirements for mutagenesis and about the macromolecular basis of repair (for reviews, see WITKIN 1976; KIMBALL 1978; HALL and MOUNT 1981; LITTLE and MOUNT 1982; HASELTINE 1983). Yet, the precise premutagenic lesion responsible for UV mutagenesis continues to be the subject of speculation, and our knowledge of the molecular processes involved in mutation fixation remains fragmentary. On the basis of a number of observations (see HASELTINE 1983), it has long been assumed that pyrimidinepyrimidine cyclobutane dimers (henceforth referred to as pyrimidine dimers) represent the major premutagenic lesion induced by UV. However, recent findings suggest that pyrimidine-pyrimidine (6-4) UV photoproducts [hereafter Present address: Department of Biology, York University, 4700 Keele Street, Toronto, Ontario, Canada M3J 1P3. Genetics 106: 347-364 March, 1984 348 B. A. KUNZ AND B. W. GLICKMAN designated (6-4) photoproducts] may play an important role in UV mutagenesis (LIPPKE et al. 1981; BRASH and HASELTINE 1982; HASELTINE 1983). Pyrimidine dimers, but not (6-4) photoproducts, are photoreactivable (D. E. BRASH and W. A. HASELTINE, cited in HASELTINE 1983). Thus, the question of whether pyrimidine dimers or (6-4) photoproducts are the major premutagenic lesions might be investigated by examining UV mutational spectra following photoreactivation. Unfortunately, photoreactivation of pyrimidine dimers prevents the induction of error-prone SOS repair functions (RADMAN 1974; WITKIN 1976) as evidenced by the concomitant loss of expression of genes or processes coregulated with the SOS response (D. E. BRASH and W. A. HASELTINE, cited in HASELTINE 1983; G. C. WALKER, personal communication; B. W. GLICKMAN, unpublished results). Since SOS induction is a prerequisite for mutation induction by UV (WITKIN 1976; KATO and NAKANO 198 I) , reduction in UV-induced mutagenesis by photoreactivation might not reflect loss of pyrimidine dimers as potential premutagenic lesions, but rather elimination of the SOS-inducing signal. In an attempt to examine the role of pyrimidine dimers in UV-induced mutagenesis, we have employed a system in which the photoreactivation of pyrimidine dimers in a target gene can be accomplished without loss of the SOS-inducing signal. The system involves UV treatment of a donor cell carrying a target gene (in this case l a d ) on an F’ episome and subsequent transfer of the F’, via conjugation, to a recipient cell UV induced for SOS functions. In this manner, irradiation of the donor cell followed by photoreactivation prior to conjugation permits the production of UV photoproducts and the subsequent specific reversal of the pyrimidine dimers. Thus, by transferring the UV-irradiated and then photoreactivated F’ DNA to recipient cells in which SOS repair has been elicited by UV treatment, we can assess the significance of pyrimidine dimers as premutagenic lesions independently of their role as inducers of SOS functions. This experimental system also has allowed us to study what has been termed “direct” or “targeted” us. “indirect” or ‘“targeted” mutagenesis (DRAKE 1974; WITKIN and WERMUNDSEN 1978). It has been suggested that hot spots for UVinduced forward mutations in the lacl gene of E. coli might reflect hot spots for UV-induced DNA base damage (BRASH and HASELTINE 1982). On the basis of this finding and the mutational specificities of several agents, it has been argued that most mutations induced in E. coli by UV and certain chemical carcinogens occur at the sites of DNA damage, i.e., are targeted at specific DNA lesions (EISENSTADT et al. 1982; FOSTER, EISENSTADT and CAIRNS 1982; MILLER 1982; FOSTER and EISENSTADT 1983; FOSTER, EISENSTADT and MILLER 1983). On the other hand, UV irradiation of E. coli, prior to infection with bacteriophage A, leads to increased mutation of the X DNA by rec’4-dependent SOS processes (JACOB 1954; DEVORET 1965; DEFAIS et al. 1971; ICHIKAWARYO and KONDO 1975). Similarly, when tif44 (recA441) or dnaB,, strains of E. coli, thermoinducible for SOS functions, are incubated under restrictive conditions, mutagenesis is enhanced markedly (WITKIN 1974, 1976; WITKIN and WERMUNDSEN 1978). Such findings have been taken to indicate that there is a TARGETED AND UNTARGETED MUTAGENESIS 349 mutator activity associated with the inducible SOS system (RADMAN 1974; WITKIN 1976). If so, a component of UV mutagenesis may be untargeted, i.e., occur at undamaged sites in DNA. However, it has been difficult to determine quantitatively the degrees to which targeted and untargeted events may contribute to mutagenesis at specific sites within a single gene. T o probe this question, targeted mutagenesis was studied by irradiating donor cells and selecting l a d mutations in the recipient strain after F’ transfer. T o test for untargeted mutagenesis, recipient cells were UV irradiated to induce SOS functions, and then undamaged F’ DNA was introduced via conjugation. The controls for these experiments involve the transfer of unirradiated DNA to unirradiated recipient cells. We (KUNZ and GLICKMAN 1983) found this transfer to be considerably less accurate than vegetative replication. This decrease in replication fidelity was independent of recA gene function in either or both the donor and recipient strain. The results presented here indicate that untargeted mutagenesis may contribute significantly to the production of nonsense mutations in undamaged F ’ DNA transferred into UV-irradiated recipients. This untargeted component of mutagenesis is quantitatively much less important when the donor strain also is irradiated. In the latter situation, targeted events gave rise to in excess of 90% of the total nonsense mutations detected. Finally, photoreactivation of UV-irradiated cells revealed that, in the conjugal system employed here, virtually all UV-induced mutations were targeted at photoreactivable lesions, presumably pyrimidine dimers. EXPERIMENTAL RATIONALE The E. coli l a d system allows the detection of nonsense mutations at a large number of sites within the l a d gene; specific mutations can be identified by analysis of suppression pattern and deletion mapping (COULONDRE and MILLER 1977a,b; MILLER et al. 1977; SCHMEISSNER, GANEM and MILLER 1977). As the DNA sequence of the nonsense mutations has been determined, each mutation can be correlated with a specific base change (FARABAUGH 1978; MILLER, COULONDRE and FARABAUCH 1978). An essential feature of this system is that the l a d gene is situated on the F’ episome and, hence, can be transferred. Moreover, since the lac operon is deleted from the chromosomal DNA of both the donor and recipient, transmission of UV-induced DNA lesions to the F’ copy of the l a d gene via recombination cannot occur. Therefore, any mutations arising within the l a d gene must be the result of DNA synthesis or repair processes acting directly on F’ DNA. During conjugal transfer of the F’, a single DNA strand is passed from donor to recipient, and this strand is the one having its 5’ end at the origin of transfer (OHKI and TOMIZAWA 1968; RUPP and IHLER 1968; IHLER and RUPP 1969). The transferred strand is replaced in the donor cell by conjugal DNA synthesis. A complement to the donor strand is synthesized in the recipient (OHKI and TOMIZAWA 1968; VAPNEK and RUPP 1970, 1971). For quantitative measurements of untargeted events giving rise to l a d mutations, unirradiated, i . e . , undamaged F‘ DNA is transferred to recipients in350 B. A. KUNZ AND B. W. GLICKMAN duced for SOS functions. Any increase in mutagenesis greater than that detected using uninduced recipients is ascribed to untargeted mutation. For quantitative measurements of l a d mutations due to targeted events, UV-irradiated F' DNA is transferred to recipient cells. Since the particular F' DNA strand transmitted during conjugation and its sequence are known, it is possible to correlate specific target sites for UV damage with the induction of certain l a d nonsense mutations. To determine whether these mutations arise as a consequence of UV-induced pyrimidine dimers or (6-4) photoproducts, UV-irradiated donor cells are exposed to photoreactivating light prior to conjugation with recipients in which SOS functions have been induced by UV treatment. If the premutagenic lesions are mainly pyrimidine dimers, photoreactivation of the irradiated donor strain is expected to result in a mutational spectrum similar to that for F' transfer from unirradiated donors to UV-treated recipients. However, if (6-4) photoproducts constitute the bulk of the premutagenic damage, the resultant mutational spectrum is expected to resemble that observed for conjugation involving UV-irradiated donor cells. MATERIALS AND METHODS The methods of bacterial conjugation, l a d mutant selection, media and buffers have been described in detail (see KUNZ and GLICKMAN 1983). Briefly, exponentially growing donor and recipient cells were concentrated by centrifugation and mated at 37" for 60 min in nutrient broth (donor to recipients = 1 : 10). Selection for F' transfer depended upon the transfer of pro' carried on the F', whereas Inclmutants were selected by their growth on phenyl-@+-galactoside. Growth of the recipient was prevented by the absence of proline in the selection plates; growth of the donor was prevented by the addition of streptomycin to the selection plates. The results presented are from at least three independent experiments; in the cases in which the donor strain was unirradiated (and thus the mutation frequency low), the experiments were repeated at least nine times. Data concerning the frequency and specificity of EacI mutations arising in crosses without irradiation in the wild-type strain and in the presence of recA mutations have been presented previously (KUNZ and GLICKMAN 1983). Struius otid i n~d io : Unless otherwise stated, bacterial strains, media and techniques for the l a d system were the same as those described by COULONDRE and MILLER (1977a,b) and TODD and GLICKMAN (1982). The wild-type strains KMBL3835 [F' pro-lac/ara, A (pro-lac), thi, trpE97771 and S9OcN [FA (prodor), oro, thi, strA, ~ l ~ ] have been described previously (GLICKMAN 1979; KUNZ and GLICKMAN 1983). CV irrudiotrott uud plzotorPnctizJntio,,: A 30-watt GE germicidal lamp emitting mainly 254 nm light was used for UV irradiation. The incident dose rate was adjusted to 1 Jm-'sec-' as measured with an International Light IL570 germicidal radiometer. Experiments were carried out under yellow light to avoid unintentional photoreactivation. Exponential phase cells were washed twice with and resuspended in cold Vogel-Bonner buffer (KUNZ and GLICKMAN 1983) at 3-4 X 10' cells/ml. Cell suspensions were agitated during UV exposure and were kept on ice following UV treatment. The UV doses resulted in surviving fractions of approximately 70 and 60% for the donor and recipient strains, respectively. A lamp containing two Sylvania Blacklite Blue F15T8-BLB bulbs emitting mainly 380 nm light was used for photoreactivation. Photoreactivation was carried out immediately following UV irradiation in covered 10-cm plastic Petri dishes (Falcon) containing 10ml suspensions of irradiated cells. Exposure was at 37" for 10 min, 20 cm from the light source.