.u < c ; x . 53a... 2 : Luna...) flab”? ”runway: . is. 5.. x-.. :1». :0 1.. $wa “.3. .: figw, p. q 1 v .IYI . I ”but. 1 .... mfin$fiflmflm » ml‘fi3p r “Awkwafimkn .. @953... ~ El- '6 7 ‘. I .0 $403.3. v 1.1:]: a 1. 6 . vinwhvflv...‘ an , 10 ilmflnqx ‘2... . ,3 ”I“! t: \l; .. .syx .i .. XIv..~.v;f.. a. .. , s. I. I a t y A! ‘ J: -v\. Vx u u... v. 39. .3 u kuVy: ‘tvavl Alt 2“..an .C ‘ )bfl lu‘ 1.1.3.55 PI v‘.‘ ‘1': ~wl‘ll.\l i. l. . I“ .rle‘... \ ..\. . \1 a: .i‘:... L. h! .1... 1.. It! 5.5%. ‘ .3 :12 .u. : rw.$ 5...? 5 .5 cl \\ usiflns HIGAN STATE U lllllllllllllllll lllllllllll‘ll - 31293 - .. rum 1020 4851 .-. -Vg This is to certify that the dissertation entitled EVIDENCE FROM THE FREQUENCY AND SPECTRUM OF MUTATIONS THAT HUMAN FIBROBLASTS CAN REMOVE POTENTIALLY MUTAGENIC LESIONS INDUCED BY N-ETHYL-N-NITROSOUREA USING EITHER NUCLEOTIDE EXCISION REPAIR OR 06-ALKYLGUANINE-DNA ALKYLTRANSFERASE OR BOTH KINDS OF REPAIR presented by LISA R. ORTQUIST has been accepted towards fulfillment of the requirements for M.S. degree in BIOCHEMISTRY Major professor VERONICA M. MAHER, Ph.D. Date OCTOBER 1‘5 199‘} MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 LIBRARY Mlchlgan State Unlverslty rues-anew!“ Boxmmwomhchodtouflunywmd. TO AVOID FINES return on or Mon duo duo. DATE DUE DATE DUE DATE DUE MSU loAn Affirmative Wen-l Oppommlly Intuition WM! EVIDENCE FROM THE FREQUENCY AND SPECTRUM OF MUTATIONS THAT HUMAN FIBROBLASTS CAN REMOVE POTENTIALLY MUTAGENIC LESIONS INDUCED BY N-ETHYL-N-NITROSOUREA USING NUCLEOTIDE EXCISION REPAIR OR 06-ALKYLGUANINE-DNA ALKYLTRANSFERASE OR BOTH KINDS OF REPAIR By Lisa R. Ortquist A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Biochemistry 1994 ABSTRACT EVIDENCE FROM THE FREQUENCY AND SPECTRUM OF MUTATIONS THAT HUMAN FIBROBLASTS CAN REMOVE POTENTIALLY MUTAGENIC LESIONS INDUCED BY N-ETHYL-N-NITROSOUREA USING EITHER NUCLEOTIDE EXCISION REPAIR OR 06-ALKYLGUANINE-DNA ALKYLTRANSFERASE OR BOTH KINDS OF REPAIR By Lisa R. Ortquist N-ethyl-N-nitrosourea (ENU) alkylates 12 sites in DNA and is an efficient mutagen and carcinogen. Nucleotide excision repair (NER) and Cf—alkylguanine-DNA alkyltransferase (AGT) are cellular DNA repair systems that can remove ENU-induced alkyl adducts from DNA in vitro and in Escherichia coli. The function of NER and AGT in repairing alkylation damage in human cells is unclear. To investigate the role of NER and AGT in repairing such damage in human cells, I treated four populations of diploid human fibroblasts. differing in AGT and NER capacities. with ENU. I assayed the treated cells for survival and frequency of HPRT'mutations. and sequenced mutants from each group. The results are consistent with the interpretation that ENU-induced lesions are removed in human cells by either NER or AGT, and that if both repair systems are active. NER and AGT compete for repair of these lesions. ACKNOWLEDGMENTS I would like to express my gratitude and appreciation of the invaluable encouragement and input supplied by nu/Inajor professor. Dr. Veronica M. Maher. Thanks are also due to Dr. J. Justin McCormick for his excellent advice and support. I would like to thank the other members of my graduate committee. Drs. Margaret Jones and Steve Triezenberg. for their valuable time and guidance. I would especially like to acknowledge the counsel and encouragement given to me by Dr. Triezenberg. Any success I have as a research scientist I owe to the excellent training in good science practice that I received as a technician in his lab. Thanks, Steve! I owe thanks to Clarissa Stropp. Terry McManus. and Dong Wei for their many helpful discussions regarding this project and for their technical advice and encouragement. Thanks are also due to others in the Carcinogenesis Laboratory for their encouragement and technical advice. They include Suzanne Kohler. Dr. Rebecca Ddenwaller, Dr. Jeanette Scheid. Dr. Yi-Ching Wang. Scott Boley. Cindy Wilson. and Lonnie Milam. Finally, I would like to thank my husband. Benjamin Mosher, for his support during the course of this work. TABLE OF CONTENTS Page LIST OF TABLES ........................... vi LIST OF FIGURES .......................... vii ABBREVIATIONS .......................... viii INTRODUCTION ........................... 1 CHAPTER 1. LITERATURE REVIEW A. DNA Alkylating Agents ................... 3 B. DNA Repair ........................ 7 1. Nucleotide Excision Repair .............. 8 2. ()6 -alkylguanine DNA Alkyltransferase ......... 12 3. DNA glycosylase ................... 14 C. Mutation Induction by Alkylating Agents .......... 15 I. Mutations Caused by Ethyl Adducts in vitro and in E. coli .................... 18 2. How Eukaryotic Cells Respond to Alkylation Damage . . 23 REFERENCES ........................... 30 CHAPTER II. EVIDENCE FROM THE FREQUENCY AND SPECTRUM OF MUTATIONS THAT HUMAN FIBROBLASTS CAN REMOVE POTENTIALLY MUTAGENIC LESIONS INDUCED BY N-ETHYL-N-NITROSOUREA USING EITHER NUCLEOTIDE EXCISION REPAIR OR 05-ALKYLGUANINE-DNA ALKYLTRANSFERASE OR BOTH KINDS OF REPAIR ....... 37 Summary ........................... 38 Introduction .......................... 40 iv Materials and Methods ..................... 43 Cells and Media ...................... 43 Treatment of cells with ENU in the presence or absence of 06-benzylguanine ................. 43 Assay for cell survival .................. 44 Assay for frequency of mutants .............. 45 Synthesis of cDNA directly from mRNA in cell lysates . . . 45 Amplification of HPRT cDNA and DNA sequencing ....... 45 Results ........................... 48 Effect of repair on the cytotoxicity and frequency of mutations induced by ENU in diploid human fibroblasts ..................... 49 Spectra of mutations induced by ENU in diploid human fibroblasts .................. 51 Effect of DNA repair on the types of base pair substitutions induced by ENU in human fibroblasts . . 59 Discussion ........................... 66 REFERENCES ........................... 71 LIST OF TABLES TABLE PAGE CHAPTER I I. Initial in vivo alkylation .................. 5 CHAPTER II I. Kinds and locations of mutations induced by ENU in the coding II. III. IV. region of the HPRT gene in the absence of nucleotide excision repair and alkyltransferase activity ............. 52 Kinds and locations of mutations induced by ENU in the coding region of the HPRT gene in the absence of nucleotide excision repair but in the presence of alkyltransferase activity . . . . 54 Kinds and locations of mutations induced by ENU in the coding region of the HPRT gene in the presence of nucleotide excision repair but in the absence of alkyltransferase activity . . . . 55 Kinds and locations of mutations induced by ENU in the coding region of the HPRT gene in the presence of nucleotide excision repair and alkyltransferase activity ............. 57 Summary of base substitution mutations observed in the coding region of the HPRT gene in ENU«treated diploid human fibroblasts differing in NER and AGT repair capacities .......... 60 vi LIST OF FIGURES FIGURE PAGE CHAPTER I 1. Mechanisms of alkylation by N-nitroso compounds ........ 4 2. Chemical formulas of the N-nitroso compounds Diethylnitrosourea (DEN). N-ethyl-N-nitrosourea (ENU). and N-methyl-N-nitro- N-nitrosoguanidine (MNNG) ................... 4 CHAPTER II 1. Sequence and location of primers used in cycle sequencing of HPRT cDNA ......................... 47 2. Survival and induced mutations frequency in ENU-treated diploid human fibroblasts differing in NER and AGT capacities ..... 50 3. Strand distribution of guanines involved in ENU-induced base substitutions ...................... 62 4. Strand distribution of thymines involved in ENU-induced base substitutions ...................... 64 vii AGT DHFR ENU HPRT MNNG NER NER'AGT‘ NER'AGT+ NERTAGT' NERIAGTT O6-alkG O6-BzG OG-EtG 06-MeG OZ-EtC OZ-EtT O4-alkl' O4-EtT 04-MeT TG TGR ABBREVIATIONS CP-alkylguanine-DNA alkyltransferase Dihydrofolate reductase N-ethyl-N-nitrosourea Hypoxanthine (guanine) phosphoribosyltransferase N-methyl-N'-nitro-N-nitrosoguanidine Nucleotide excision repair Deficient in NER and depleted of AGT Deficient in NER but proficient in AGT Proficient in NER but depleted of AGT Proficient in both NER and AGT Oé-alkylguanine CP-benzylguanine (P-ethylguanine (f—methylguanine CP—ethylcytosine CE—ethylthymine (Y-alkylthymine CP-ethylthymine (P-methylthymine 6-thioguanine 6-thioguanine-resistant viii INTRODUCTION DNA damage can lead to mutations. which often result in altered gene expression or changes in protein structure and function. In this way. DNA damage can cause cell death or changes in cell growth patterns that, in a eukaryotic organism. provoke the formation of tumors. Continuous exposure to DNA-damaging agents has permitted the evolution. in the cell. of numerous DNA repair mechanisms to insure the preservation of the genetic code. One example of DNA repair is the removal of alkyl adducts from bases iri DNA by line repair protein, alkylguanine-DNA alkyltransferase (AGT). Another example of the cellular response to DNA damage is the repair of UV light-induced damage. and other types of damage causing distortions in the DNA helix. via the nucleotide excision repair pathway (NER). AGT and NER have been studied extensively in Escherichia coli. and are now being studied in mammalian cells and organisms. By studying DNA repair. researchers are acquiring a better understanding of the origins and prevention of tumor-forming mutations in DNA. The purpose of my research project has been to study the repair. via AGT and NER. of alkyl-DNA adducts induced by N-ethyl-N-nitrosourea (ENU) in human cells. Repair of the alkyl adducts induced in DNA by ENU has been studied in E. coli. and in transformed human cell lines. The results reported in this thesis. however. are the first to show an analysis of the repair of these lesions by AGT and NER in diploid human fibroblasts. 2 Chapter One of this thesis is a review of literature relevant to this project. Chapter Two is presented as a manuscript. currently in preparation for submission to the journal Carcinogenesis. describing the results of my research project. The CE-benzylguanine used in depleting cells of AGT was supplied by Dr. Anthony Pegg of the Milton S. Hershey Medical Center at Pennsylvania State University. The ENU treatment of the cell strains used. the isolation of the resultant thioguanine-resistant mutants. and the purification of HPRT cDNA was performed by me and my colleague Lubov Lukash. DNA sequence analysis of the HPRT mutants was accomplished by me. L. Lukash. Dr. M. Chia-Mia Mah. Dr. Yi-Ching Wang. Dr. Janet Boldt. and Krisztina Nadas. CHAPTER I LITERATURE REVIEW A. DNA Alkylating Agents DNA alkylating agents are efficient mutagens and carcinogens (see Saffhill et al.. 1985 for review) and are present in the environment. Endogenous formation of N-nitroso compounds can occur at various sites in the human body. including the gastrointestinal tract and the lungs (Bartsch et al.. 1990). For example. diethylnitrosamine (DEN) and dimethylnitrosamine (DMN) are alkylating agents found in trace amounts in human blood following ingestion of food containing nitrates (Fine et al.. 1977). Nitrosoamines require metabolic activation. with enzymes in the cell converting the nitrosamine into reactive metabolites. which go on to generate alkyl adducts in DNA. Other alkylating agents. like N-ethyl-N- nitrosourea (ENU) andqumethyl-N-nitro-N-nitrosoguanidine~(MNNG). directly alkylate DNA tn/ forming reactive intermediates spontaneously at physiological pH (Figures 1 and 2)(Montesano. 1981: Singer 1985). MNNG and ENU are direct-acting alkylating agents that are used in chemical laboratories to generate diazo compounds (Beranek. 1990). These simple alkylating agents are also useful as model compounds for the reactive forms of the metabolically-activated nitrosamines (reviewed in Saffhill et al.. 1985). and have been used as such in the Carcinogenesis Laboratory to study repair of alkylation damage in the DNA of human cells. Exposure to N-nitroso alkylating agents. including MNNG and ENU. results in the formation of a variety of alkyl-DNA adducts (Table I). but Figure 1. Mechanisms of alkylation by N-nitroso compounds. (from Magee. 1971: Montesano. 1981; and Singer. 1985) N-Nitrosamines N-Nitrosamides DEN ENU Enzymes H+ NADPH, 02 H20 4. [Ham-120 N2] DNA, RNA, Proteins CH3CH2R Figure 2. Chemical formulas of the N-nitroso compounds Diethylnitroso- urea (DEN). N—ethyl-N-nitrosourea (ENU). and N-methyl-N'-nitro-N- nitrosoguanidine (MNNG) (from Singer. 1985). <:i—12-c:+-i3 / O=N—N\ CHz-CHs DEN CH2—CH3 / CH3 / O=N—N \ O=N—N \ I H C—NH C—N ll 2 II \ NO2 O NH ENU MNNG Table I. Initial in vivo alkylation (expressed as percent of total alkylation) Site of Base alkylation MNNG" M Adenine N1 -- 0.1 N3 8.6 4 N7 -- 0.6 Cytosine O2 -- 2 N3 -- 0.3 Guanine O6 9.2 8 N3 -- 1.5 N7 82.2 12 Thymine 02 -- 7 O4 -- 2.5 N3 -- 0.4 Phosphate -- Triester 58 a. From Beranek. 1990 D. From Singer and Dosanjh. 1990 6 whether each of these adducts is mutagenic is still being determined. The extent of DNA alkylation observed in vitro parallels the initial alkylation of DNA in cultured cells and in animals exposed to these alkylating agents (Montesano. 1981). The alkyl adducts at oxygens in DNA bases are more chemically stable in vitro. in the absence of DNA repair mechanisms. than those at nitrogens under physiological conditions (see Singer. 1979. for review). N-3 and N-7 alkyl purines are easily depurinated (Mme to lability cyf their glycosyl bonds. Alkylphospho- triesters are the nmst stable of the alkylation adducts formed by AL nitroso compounds. These lesions are reported to be cytotoxic lesions. but there is no evidence to suggest that they are premutagenic lesions. The more potent carcinogenic alkylating agents show a greater tendency to react at oxygen sites in DNA 'than do 'the less carcinogenic agents (reviewed in Saffhill et al.. 1985). Thus. it seems likely that 06~alkG. (Y—alkT. (E-alkT. and CE-alkC are the most reasonable candidates for the lesions in DNA resulting in the mutations observed in MNNG- and ENU- treated cells. Methylating agents are more reactive with DNA than ethylating agents (Singer and Grunberger. 1983). but produce adducts that are more efficiently repaired in the cell. making methylating agents less mutagenic than ethylating agents at comparable doses in vivo. (Singer. 1985). ENU and MNNG are thought to react with oxygens in DNA by an SNI mechanism. The reaction is dependent on the formation of an electrophilic carbocation intermediate. which is trapped by the nucleophilic oxygen in DNA. forming a covalently bound adduct (reviewed in Beranek. 1990). At least one of the alkyl adducts generated by N-nitroso compounds has been directly implicated in carcinogenesis. The formation of the 06- 7 methylguanirma(CP-MeG) adduct in DNA following metabolic activation of the tobacco-specific nitrosoamine 4-(N-methyl-N-nitrosoamino)—1-(3-pyridyl)- l-butanone (NKK) appears to be a major factor in the induction of lung tumors in rats and mice. and in the activation of the K-ras proto-oncogene in lung tumors in mice (Belinsky et al.. 1986. 1990). Using monoclonal antibodies against.(E-MeG. Umbenhauer et al. (1985) found elevated CP-MeG levels in DNA from normal esophageal tissue taken from human cancer patients living in a region in China with a documented high incidence of esophageal cancer. Goth and Rajewsky (1974) have demonstrated that the lack of repair of'CP-alkylguanine (CP—alkG) lesions is related to tumor formation in the brains of rat neonates treated with ENU. At various times following treatment with 1“C-ENU. DNA was isolated from various tissues in the neonatal rats and the amount of adducts remaining in the DNA was measured. Although other adducts were rapidly lost from the DNA in brain tissue. CP-ethylguanine was not removed rapidly from this DNA in the treated rats. and the rats eventually developed tumors of the brain. Mutations involving thymine bases are also biologically relevant. Perantoni et al. (1987) exposed pregnant rats to ENU and discovered nervous system tumors in the progeny of the treated rats. These tumors were found to contain a T-A +.A T activating mutation in the neu oncogene. Popp et al.(1983) also found a change in the amino acid sequence of B- globin in the progeny of an ENU-treated mouse that can be attributed to a T-A e A-T base substitution in the B-globin gene. 8. DNA Repair Defects in DNA repair have been linked to an increased risk of cancer in humans. For example. many patients with xeroderma pigmentosum 8 (XP). a disease resulting from a deficiency in the nucleotide excision repair of UV-light—induced DNA damage. die at an early age of complications arising from neoplasia (Cleaver. 1990). implying that the defect in repair of UV-induced DNA damage promotes tumor formation. Extracts from cultured fibroblasts from patients with lung cancer reportedly have lower alkyltransferase activity than that observed in fibroblast cell extracts from healthy controls (Rudiger et al.. 1989). suggesting that a reduced capacity to repair alkylation damage in DNA may be a risk factor for lung cancer. Furthermore, transgenic mice expressing high concentrations of the human homolog for AGT in the thymus. an organ that in mice ordinarily has low levels of alkyltransferase. are protected from the development of thymic lymphomas after exposure to the DNA alkylating agent N-methyl-N-nitrosourea (MNU) (Dumenco et al.. 1993). demonstrating that DNA repair via AGT is instrumental in protecting against tumorigenic alkyl-DNA lesions. 1. Nucleotide Excision Repair Nucleotide excision repair is one of the mechanisms by which a cell preserves the integrity of its genome. NER has been studied at the molecular level in E. coli. and has been shown to involve the removal of an oligonucleotide containing the damaged base(s) and repair of the resulting gap by DNA polymerase and DNA ligase (see Sancar and Sancar. 1988. for review). In E. coli the NER mechanism requires at least six proteins: UvrA. UvrB. UvrC. Uer. DNA polymerase I. and DNA ligase. The UvrA protein binds single-stranded DNA or damaged double-stranded DNA and has ATPase functional domains. The UvrB protein alone does not bind DNA or have 9 ATPase activity. However. when associated with the UvrA protein. UvrB increases the stability of the UvrA-DNA complex. and DNA and UvrB together stimulate the ATPase activity of UvrA” UvrC is a DNA-binding protein that has been reported to be capable of exonuclease activity. The UvrB subunit combines with two UvrA subunits in the presence of ATP. and the A281 complex binds to the DNA containing the damaged nucleotide(s). UvrC then binds to the Agfi-DNA complex and two incisions take place - hydrolyzing the 8th phosphodiester bond 5' and the 4th or 5th phosphodiester bond 3' to the damaged base to remove a 12-13 base oligomer from the DNA. Uer is a helicase. and works with DNA pol I and DNA ligase to release the damage- containing oligomer and fill in the resulting gap. The UvrABC complex excises UV—light-induced DNA damage from DNA as well as the base adducts formed in DNA by many chemicals. and is considered to recognize the distortion of the DNA helix caused by these types of damage. The molecular mechanisms of nucleotide excision repair in eukaryotes have not been studied in as much detail as NER in E. coli. In the yeast S. cerevisiae. five genes essential for NER have been cloned and sequenced. One of these. RAD3. produces a DNA-dependent ATPase and a helicase. suggesting similarities between the mechanisms of NER in prokaryotes and eukaryotes. Cells from patients with XP have been assigned to seven complementation groups. A - G. based on their ability to perform UV-induced unscheduled DNA synthesis when fused to cells from different complementation groups (Timme and Moses. 1988). The number of different complementation groups implies that. as in E. coli and S. cerevisiae. NER in human cells involves multiple gene products. One human gene that has been cloned and shown to correct a NER defect in a rodent cell line is ERCCI. The ERCCI gene product has sequence homology with 10 RADIO. a yeast NER protein. and also some homology to the E. coli UvrC protein. Interestingly. the ERCCI gene does not correct the NER defect in cells from any of the XP complementation groups. Another cloned human gene. ERCCZ. shows an amino acid sequence similar to that of the yeast RAD3. a DNA helicase. ERCC2 has been implicated as the defective gene in XP patients from complementation group D. ERCCB is a cloned human gene that corrects the NER defect in cells from XP group 8 patients. and also appears to be a DNA helicase. The NER defect demonstrated by cells from XP patients from group A. C. or G can be corrected by the human genes XPA. XPC. and ERCCS. respectively (See Bootsma and Hoeijmakers. 1994 for review). The structure of DNA in eukaryotes is more complex than that in E. coli. and the NER mechanisms in eukaryotes are probably more complicated than in bacteria. However. the generally accepted mechanism for NER in human cells is based on the NER pathway in E. coli. The lesion in DNA is recognized and a dual incision is made in the damaged DNA 27-29 nucleotides apart. Excision of the lesion-containing oligomer and subsequent repair involves DNA.helicase. single-stranded binding proteins. and the activity of DNA polymerase 6 or 6 (NER in eukaryotic cells is reviewed in Sancar and Sancar. 1988; Cleaver. 199D: and Bootsma & Hoeijmakers. 1994). Evidence suggests that there is a distinction between repair of the genome overall and repair of actively transcribed genes. It has been shown. for example. that UV-induced pyrimidine dimers are removed faster from the amplified DHFR gene in the human cell line 6A3 than from the genome overall (Mellon et al.. 1986). Moreover. the transcribed strand of the amplified DHFR gene is repaired faster than the non-transcribed strand 11 in this same human cell line and in a rodent cell line (Mellon et al.. 1987). Vrieling et al. (1989) compared the spectrum of UV-induced mutations in the HPRT gene of Chinese hamster cell lines proficient or deficient iri repair of LNlinduced lesions. and reported preferential repair of UV-induced lesions in the transcribed strand of the HPRT gene. Fibroblast cells from a patient of the XP complementation group C can repair UV-induced lesions in actively transcribed genes but not those in inactive regions of the genome (Venema et al.. 1990). These results suggest that the defect in NER in these cells involves one or more gene products that allow the repair of damage in the inactive regions of the overall genome. implying separate mechanisms for the repair of actively transcribed genes and of transcriptionally silent regions of the genome. The association between the NER machinery and transcription has recently been clarified (reviewed in Friedberg et al.. 1994 and Bootsma and Hoeijmakers. 1994). The ERCC3 gene product appears to be one of the components of “the human basal transcription factor TFIIH. which is required for a late step in the initiation of transcription by RNA polymerase II. The human ERCCZ gene product shares homology with the yeast RAD3 gene product. which has been shown to be part of the yeast RNA Pol II preinitiation complex. implying a similar role for ERCCZ. ERCCZ and ERCC3 gene products are reported to interact with each other. At least two . then. of the human genes directly involved in NER are believed to also associate directly with the transcription apparatus. Friedberg et al.. (1994) suggest that NER proteins are coupled to the transcription apparatus in order to recognize base damage in the template strand during transcription and act as nucleation sites for the assembly of the NER machinery. They further suggest that if these NER proteins have a higher 12 affinity for sites of DNA damage than for the transcription initiation complex. this may serve to limit transcription initiation when cells are exposed to DNA damage. Finally. they hypothesize that there are two distinct repair complexes. one operating in transcription and repair of active genes. and the other working on DNA damage in the genome overall. 2. (P-Alkylguanine-DNA Alkyltransferase Another important means by which the cell rids the genome of potentially mutagenic DNA lesions is by using the repair protein 06- alkylguanine-DNA alkyltransferase (AGT). (For a review of'AGT in bacteria and mammalian cells. see Yarosh. 1985; Laval. 199D: Pegg. 1990a.b; and Sassanfar et al.. 1991.) In both prokaryotes and eukaryotes. this protein repairs alkyl adducts in DNA. particularly at the CF position of guanine bases. by transferring the alkyl group from the DNA to a cysteine in its own amino acid sequence. A single AGT protein can act only once. AGT activity is depleted in cells treated with alkylating agents by the reaction of the protein with alkylated DNA. Once the AGT activity has been depleted from a cell in this manner. only synthesis of new AGT proteins can restore the original level of activity. In this respect AGT is not a true enzyme. as a single reaction with an alkyl substrate leaves the protein inactive. AGT binds to double-stranded DNA. and requires no other protein or cofactor to affect the removal of an alkyl group from DNA. Ethyl groups are removed from DNA by AGT more slowly than methyl groups. A mechanism has been proposed for the removal of alkyl groups from DNA by AGT: A basic residue in the protein interacts with the acceptor cysteine residue. gaining a proton from the cysteine and generating a thiolate anion. The anion then attacks the alkyl group in 13 the damaged base and removes it. forming S-alkylcysteine in the protein and restoring the substrate base in DNA. AGT repair in E. coli has been well characterized. There are apparently two genes responsible for the AGT activity in E. coli. The ada gene product contains two domains separated by a hinge region. and has two alkyltransferase functions. At the amino terminus the protein can act on alkylphosphotriesters. The carboxy terminus acts on CP-alkG and CP-alkT. Binding of an alkyl adduct at the amino terminus converts the ada gene product into a strong activator of the ada gene and of the alhA gene. which codes for another DNA repair enzyme. N-3 methyladenine-DNA glycosylase II. Thus. E. coli initially exposed to an alkylating agent subsequently become resistant to a second exposure to such alkylation damage. responding with increased levels of AGT and glycosylase in what is known as the adaptive response. Another AGT protein is the product of the ogt gene. This protein repairs Os-alkG and O4-alkT lesions. It shows homology to the C—terminal domain of the ada gene product. but unlike ada. expression of ogt is not induced by exposure to alkylating agents. The ogt gene appears to be responsible for protecting E. coli from low levels of alkylation damage in the absence of induction of the ada gene. The ada gene product demonstrates a higher affinity for CP-alkG lesions than the ogt gene product. and the ogt gene product has a higher affinity for O‘- alkT lesions than does the ada protein (Sassanfar et al.. 1991). indicating that. although similar in activity. the ada protein and the ogt protein may have different repair specificities in the cell. Although the eukaryotic alkyltransferase has not been defined in as much detail as AGT in E. coli. the same mechanism is apparently employed in eukaryotic cells as that demonstrated by the bacterial AGT. Transfer 14 of an alkyl group from the 06 position of guanine produces an S-methyl cysteine residue in the protein, and only RNA and protein synthesis can restore AGT activity in mammalian cells depleted of AGT by exposure to alkylating agents (Pegg. 1990a. 1990b). Unlike the bacterial AGT. mammalian cell extracts with AGT activity exhibit no repair of alkylphosphotriesters and CP-alkT (Yarosh et al.. 1985). However. CP-MeT lesions in DNA oligomers have recently been shown to inhibit. with low affinity. the activity of purified human AGT. suggesting that the human AGT may repair CP—MeT. but with very poor efficiency (Sassanfar et al.. 1991). Cells from human tissues have a higher alkyltransferase activity than cells from the corresponding tissues in rats or nnce. hi human tissues. liver has the highest activity. followed by the GI tract. then the lung. Brain tissues have the lowest levels of activity (Grafstrom et al., 1984). Induction of'AGT activity following treatment with alkylating agents has been observed in rat liver cells. and in human tumor cell lines. but not in normal rat or human fibroblasts (Laval. 1990). 3. DNA Glycosylase The DNA glycosylase enzymes are another cellular defense against alkylating agents. Glycosylases remove alkylated bases from DNA by catalyzing the cleavage of the sugar-base bond. In general, the E. coli glycosylases can remove N3 and N7 alkylpurines from DNA. and remove ethylated bases at these positions more slowly than methylated bases. The tag gene in E. coli codes for N-3 methyladenine-DNA glycosylase I. which excises N3-methyladenine from DNA. The E. coli alk gene. which is upregulated by the methylated ada gene product in the adaptive response. codes for the N-3 methyladenine-DNA glycosylase II enzyme. The alk gene 15 product excises N3-methylpurines and. more slowly. N-7 methylpurines and (Y-methylpyrimidines. Cell extracts from E. coli and from rat tissues removed 05-MeG from DNA treated with 3H-MNU. but 3H-O‘S-MeG was not detected as a free base. indicating that glycosylases are not involved in the repair of this lesion. Unlike the bacterial glycosylase. mammalian DNA-3- methyladenine glycosylase is not active on CF-MeC or CF~MeT in vitro. (Reviewed in Saffhill et al.. 1985: Yarosh. 1985: Brent et al.. 1988: Lindahl 8 Sedgwick. 1988; and Samson et al.. 1988). C. Mutation Induction by Alkylating Agents Of the alkyl adducts formed in DNA by ENU and MNNG. the lesions at oxygens in DNA bases are considered to be the most mutagenic lesions. resulting in miscoding bases (Singer. 1979: Montesano. 1981; Larson et al.. 1985). Alkylation at the N-3 position of pyrimidines and the N-l position of purines leads to errors in transcription and inhibition of DNA replication that may generate the cytotoxic effects of the alkylating agent. These lesions have not been implicated in mutagenesis. however. except when loss of the alkylated base. either spontaneously or due to glycosylase activity. results in an apurinic/apyrimidinic site (reviewed in Saffhill. et al.. 1985). More than 20 years ago. Loveless (1969) first implicated OswalkG as a potentially mutagenic lesion. Since then. in vitro site-specific studies have shown that.CE-MeG very frequently pairs with T instead of C during both DNA replication and transcription using Klenow and E. coli RNA polymerase I. resulting in G-C + A-T base substitution mutations (Snow et al.. 1984; Toorchen and Topal. 1983: Singer and Dosanjh. 1990). DNA polymerase may mistake the alkylated G for A. since O6walkG resembles A in 16 bond angles and bond lengths (reviewed in Swann. 1990). In 1990. Singer and Dosanjh reported that prokaryotic and eukaryotic DNA polymerases. using a template containing (P—MeG. preferentially inserted a deoxythymidine opposite (F-MeG. Insertnmi of the correct nucleotide. deoxycytidine. opposite(T¥MeG results hia decrease in extension past the site of the lesion relative to the extension observed following incorporation of deoxythymidine. From two-dimensional NMR assays the 06- MeG:C base pair appears to distort the DNA helix. while‘Ume(P-MeG:T base pair does not (reviewed in Swann. 1990: Basu and Essigmann. 1990). and this preservation of the helix alignment may hide the erroneous base pair from any repair mechanism. (P-MeG in an M13 bacteriophage vector transfected into E. coli gives rise to G'C 9'A T base substitutions in the progeny phage DNA (Loechler et al.. 1984). consistent with the 06-MeG miscoding observed in vitro. Treatment of the bacteria with MNNG prior to insertion of the Cf-MeG- containing vector depletes the cells of AGT. The mutation frequency in the progeny phage from the bacteria not depleted of AGT is reduced relative to the frequency in phage obtained from bacteria depleted of AGT by MNNG treatment. This is indicative of repair of the CP-MeG lesion in the vector by AGTL CP-MeG is the most frequent adduct formed at oxygens in DNA bases by treatment of DNA with MNNG in vitro (Table 1). The majority of mutations induced by MNNG in E. coli are G-C a A-T transitions. (Richardson et al., 1987a: Gordon et al.. 1990). AGT repairs (P-MeG lesions in vitro. The presence of AGT reduces the frequency of G-C + A-T mutations in the gpt gene of MNNG-treated E. coli. relative to the frequency observed in AGT' bacteria (Richardson et al.. 1987a). strengthening the hypothesis that in E. coli. MNNG-induced G-C -> A~T 17 mutations are most likely caused by(T1MeG lesions and demonstrating that. as seen in the in vitro studies. AGT works on CP-MeG in bacteria. In vitro. methylated DNA can also serve as a substrate for uvrABC excision repair (Van Houten and Sancar. 1987). The NER pathway repairs 06- M63 in E. coli in the absence of the adaptive response. Samson and colleagues (1988) treated strains of E. coli differing in AGT and NER capacities with MNU. and compared the rate of removal of‘CP-MeG from the DNA of the treated cells. They found that the rate of removal of CP-MeG from the DNA of the treated cells was similar in strains normal in NER capability. regardless of AGT capacity. for the first hour after treatment. After the first hour. i.e. the time required for the induction of the adaptive response. cells normal in expression of the ada gene repaired 05-MeG faster than ada’ cells. Cells normal in ada and NER repaired these lesions faster than cells normal in ada but lacking NER at all time points. These authors conclude that following transient exposure to an alkylating agent that saturates the AGT proteins in the cell. NER will play a major role in the repair of'CP-MeG lesions. Richardson et al. (1987a) have reported a strand bias for MNNG- induced 06-MeG premutagenic lesions in treated E. coli. They analyzed the MNNG-induced mutations in the gpt gene of E. coli proficient or deficient iri‘Uma ada gene. Both cell strains showed statistically significant preference for induction of premutagenic lesions on the non-transcribed strand. This could be due to less efficient repair of the'CF-MeG lesions in the nontranscribed strand by AGT. which prefers double stranded DNA. Roldna-Arjona and colleagues (1994). however. recently reported a similar bias in MNU-induced mutations in a shuttle vector gene in treated E. coli completely lacking AGT activity (ada‘. ogt). This suggests that the 18 strand bias observed could be due to a bias in induction of lesions on the non-transcribed strand. rather than to some property of AGT repair. 1. Mutations Caused by Ethyl Adducts in vitro and in E. coli Unlike the pattern of methyl-DNA adduct formation caused by the methylating agent MNNG. the distribution of ethyl adducts at oxygens in ENU-treated DNA is not dominated by lesions at the~CP-position of guanine (See Table 1). Instead.(Y-EtT'and Ofiiflfl'together account for roughly an equal percentage of the total alkylation adducts asifiEtG. This predicts that mutations at T's will be responsible for a percentage of the ENU- induced mutations. Although occurring at a much lower frequency than O6-EtG. Oz-EtC may also be mutagenic. It was observed in vitro that CP-EtG miscodes as A. resulting in G-C a A'T base substitutions. and that AGT works on CP-EtG. and that the frequency of G-C +1A T transitions observed in vivo decreases dramatically in cells competent in AGT relative to cells with no AGT. Therefore. OG-EtG has been accepted as the primary cause of the G-C + A-T transition. andifKEtC has been ruled out by a lack of evidence indicating otherwise. However, it is conceivable that the mammalian DNA glycosylase could work on Oz-alkT or Oz-alkC in vivo. leaving apyrimidinic sites. even though this activity has not been demonstrated by mammalian glycosylase in vitro. If this were to occur, and if DNA polymerases preferentially misinserted dATP opposite the apyrimidinic sites (Kunkel. 1984). this would result in T-A -» A’T or C-G -> T-A mutations. A C-G -» T-A base substitution mutation could be misinterpreted as a G'C + A-T mutation and attributed to CF—EtG. In vitro analysis of the mutagenic specificity of (Y-EtC is in progress in the laboratory of 8. Singer. and the results 19 should clarify this point. The mutagenic specificities of'CP-EtT and CE—EtT have been studied in vitro by inserting the modified base into an oligomer and analyzing the effects of DNA polymerases on the modified template (Singer and Dosanjh. 1990; Bhanot et al.. 1992: Grevatt et al.. 1992: Dosanjh et al.. 1993. Menichini et al .. 1994). As was shown for 05-MeG. insertion of the correct nucleotide. deoxyadenosine. opposite (E-Etl' or (Y-EtT. obstructs the progress of the DNA polymerase past the site of the lesion. If. however. the polymerase inserts a dGTP opposite the Oiififlfl or a dTTP opposite the (Y-EtT. replication can continue past the lesion. Two-dimensional NMR studies predict that the CF-MeTzG base pair will not distort the helix. while 04-MeT base-paired with A should result in a distorted helix conformation (Basu and Essigman. 1990). suggesting that. as with the O5- alkG T mispair. the (F-alkT:G base pairing could effectively hide the lesion from DNA repair. (T-Etl. then. is implicated in T-A + C-G base substitutions and Oiififl'in T-A + A-T base substitutions in alkylated DNA in vitro. O4-EtT and OZ-EtT can also be connected with the TA ~> G-C base substitutions observed in the DNA of ENU-treated cells. Drinkwater and colleagues (Eckert et al.. 1989) suggest that l“/\-* G-C mutations can result from the processing of adducts that block DNA replication. i.e. via the SOS response in E. coli. The SOS response in E. coli results in induction (H: NER, recombination. and mutagenic repair mechanisms and inhibits cell division (Sancar and Sancar. 1988). Incorporation of dATP opposite O4-EtT or Oz-EtT has been reported to block DNA replication (Dosanjh et al.. 1993: Grevatt et al.. 1992). Perhaps SOS-type processing of these replication-blocking lesions can account for the observed T-A e 20 G-C mutations. Preston and colleagues (1987) have determined that 04-MeT or O4-EtT lesions iri a phage vector introduced into an If. coli strain that is defective in the ada gene result in T-A + C-G mutations in the progeny phage. confirming the mutagenic specificity of'O4-alkT lesions observed in vitro. In E. coli proficient in the alkyltransferase activity encoded by the ada gene. the frequency of this mutation is not higher than background. (Y—MeT has been shown in vitro to be repaired by the ada gene product. The increase iri the frequency'cyf mutations 'hi ada‘ bacteria relative to the frequency seen in the bacteria normal in the ada gene is consistent with repair of’CP-alkT by this alkyltransferase. In El coli the importance of NER increases and the importance of AGT repair decreases. as the size of the alkyl adduct increases (reviewed in Yarosh. 1985). In the lac] gene of normal and NER-deficient E. coli. treatment with ENU results primarily in G-C 6‘ AT and TA -> C-G base substitutions mutations. probably due to persistent O‘-EtT and OG-EtG lesions. The frequency of these mutations in the normal E. coli was 5- fold lower 'than ‘HI the~ NER—deficient, cells. indicating ‘that NER is involved in the repair of OG-EtG and O4-EtT (Burns et al .. 1988). Samson et al.. (1988) propose that NER is the primary pathway for the repair of O4-EtT and O6-EtG. and repairs 06-MeG in the absence of the adaptive response. They treated various strains of E. coli. differing in NER and ada phenotypes. with ENU and MNU. Using monoclonal antibodies specific for (Y-EtT. CP-EtG. and (F-MeG, they then analyzed the time-dependent removal of these adducts from the DNA of the treated cells. Initially following treatment. 05-MeG. O6-EtG and O4-EtT were removed from the DNA at the same rate in aab‘ and ao'a+ cells. Cells expressing abnormally high 21 levels of the ada gene product (adac). however. showed a more rapid rate of removal than that of the ada+ or the ada' cells. Approximately 1 hour after treatment. the rate of removal of 06-MeG increased in the ada‘ cells relative to the rate in the ada’ cells. indicating the induction of the adaptive response. This increase in rate of removal was not apparent for O4-EtT or Os—EtG adducts. In NER‘ derivatives of the ada' and ma cell strains. removal of 06-MeG. O6-EtG and O4-EtT was slower than in the NER proficient cell strains. They conclude that in wild-type E. coli in the absence of the adaptive response. NER is the primary pathway for the initial repair of O4-EtT. O6-EtG. and 06-MeG. and that the inducible ada protein can repair ethyl adducts (Samson et al.. 1988). One would expect that cells functional both in NER and AGT would be better at surviving alkylation damage and maintaining DNA free of alkylation-induced mutations than cells with only one of these repair systems operating. However. some studies in E. coli have shown that NER and AGT interfere with one another. For example, Rossi and colleagues (1989) found that an 06-MeG lesion in an M13mp18 vector transfected into E. coli was more mutagenic in cells competent in NER and AGT than in cells competent only in AGT repair. They suggest that binding of the uvrA protein to the damaged DNA inhibits repair by alkyltransferase. Their results agree with those of Chambers et al.. (1985). Chambers and colleagues placed an 05-MeG lesion in a ¢XI74 vector and transfected this vector into uvrA' cells (i .e. lacking NER but normal in AGT) and into uvrA+ cells (normal in both NER and AGT). They found that the mutation frequency in the progeny phage DNA was 40-fold higher in the uvrA+ cells than in the uvrA' cells. indicating that the NER apparatus interfered with the repair of 06-MeG via AGT. 22 In E. coli treated with ENU. miscoding lesions are more likely to occur at 5 -Pu-G-3' or 5'-Pu-T-3' sites in the gpt gene (Richardson et al.. 1987). This specificity for induction of mutations at sites with a 5' purine is characteristic of SNI methylating agents (Horsfall et al.. 1990). and is also seen in MNNG-treated cells (Gordon et al.. 1990). As reviewed by Horsfall et al. (1990). the observed site specificity may result from reduced efficiency of repair of the adduct with a 5'-flanking purine or from increased initial formation of the adduct at such a site. Burns et al.. (1988) treated NER-proficient E. coli and NER-deficient E. coli with ENU. They found a bias in the formation of G-C 9 A-T and T-A a C-G substitution mutations in the lacI gene at guanines and thymines with G:C base pairs 5' and 3’ to the mutated base in the NER-proficient cells. This bias was not present in the mutations observed in the NER-deficient cells. They conclude that. in the lad gene of ENU-treated E. coli. excision repair is less efficient at removing CP-EtG and CP-EtT if that guanine or thymine is flanked by G:C base pairs. Wong et al. (1992) found in in vitro studies. using oligomers contairnrNJCF-MeG residues. that there are localized. strand specific disruptions in the chemical structure of DNA containing CP-MeG that make DNA containing this lesion more sensitive to the restriction enzyme MBeII. Differences in the 5’ neighboring base and the base-pairing pyrimidine alter the disturbance in the helix caused by this lesion. These perturbations affect the activity of the cloned human AGT. They conclude that it is likely that the stacking interactions between 05-MeG and its neighboring bases determines the kinetics of formation and/or repair of this. and other alkyl lesions. 23 2. How Eukaryotic Cells Respond to Alkylation Damage (E-alkG. (T4alkT. and (Y—alkl' are also relevant mutagenic and carcinogenic lesions in eukaryotic organisms. Lung tumors induced by ENU in mice were found to have G-C + A-T base substitutions activating the K- ras protooncogene. consistent with the ENU-induced formation of persistent OG-EtG lesions (You et al.. 1992). T-A a G-C and T-A a A'T mutations have also been found. In ENU-treated rats. activation of the neu oncogene has been shown to be due to a T-A 9 A-T base substitution in tumors of the peripheral nervous system (Perantoni et al., 1987). In granuloma pouch skin fibroblasts of ENU-treated rats, ENU-induced mutations in the hprt gene were predominantly T-A -> AT and TA -> G'C base substitutions (Jansen et al.. 1994). Since it has been shown that the repair specificities of the eukaryotic AGT and NER differ from those of the prokaryotic AGT's and NER in vitro. researchers have been studying the response of cultured eukaryotic cells to DNA damage in order to investigate NER and AGT in these cells. There are also differences in repair efficiency among eukaryotes. Human cells show greater efficiency in repairing UV-induced damage in the entire genome using NER. whereas NER in rodent cells concentrates more on the DNA in actively transcribed genes (Cleaver. 1990). Likewise. human cells have higher AGT activities than the corresponding rodent cells. Therefore. in order to understand DNA repair in the human organism. it is important to study repair in human cells. One approach to studying DNA repair in human cells in culture is by making use of shuttle vector systems. Such systems use recombinant plasmids able to replicate in both mammalian and bacterial cells. with 24 genes that can be expressed in bacteria for detection and analysis of mutations. For example. Klein and colleagues (1990) used an SV40 viral- based shuttle vector carrying a singleiOI-EtT lesion at a defined position. When replicated in human HeLa cells. this vector gave rise to T-A 9 C-G mutations. in agreement with the mutagenic specificity of this lesion observed in vitro and in E. coli. In a related approach. Drinkwater and colleagues (Eckert et al.. 1988) used a herpes simplex virus thymidine kinase gene carried on an EBV—based shuttle vector to analyze the induction of mutations by ENU in HeLa cells. They found that. in addition to the G-C 9 A'T and T-A 9 C-G mutations that had been observed in ENU- treated E. coli. ENU induced T-A 9 A-T and T-A 9 G-C mutations in human cells. They did not observe any strand or sequence bias in the formation of the 8C 9 AT mutations, as had been reported for the BC 9 AT mutations induced by MNNG. These investigators introduced the hypothesis that OZ-EtT is the lesion responsible for the observed T A-9iA Tinutations. Although vector systems are useful, there may be differences between repair of the shuttle vector DNA and repair of the DNA in an actively transcribed endogenous gene. Therefore. many investigators are analyzing DNA repair in endogenous genes in mammalian cells. A particularly well- studied gene is the gene that codes for the enzyme hypoxanthine (guanine) phosphoribosyltransferase (HPRT). which is involved in the purine salvage pathway. a crucial process by which a cell recycles nucleotide bases (for reviews on HPRT. see Stout and Caskey. 1985: Chinault and Caskey. 1984; Morrow. 1983). HPRT catalyzes the formation of inosine monophosphate and guanosine monophosphate via the condensation of 5-phosphoribosyl diphosphate and hypoxanthine or guanine bases. The HPRT gene is X-linked in mammals. making it hemizygous in males and functionally hemizygous. via 25 inactivation of one of the two X chromosomes. in females. The sequence of the gene is known. with an open reading frame of 654 bases coding for a protein of 217 amino acids. The enzyme has been shown to be intolerant of slight structural changes. such as those that might be caused by a single base substitution. HPRT is nonessential in cultured cells that are normal in the ability to synthesize purines de novo. HPRT' mutants can be selected for by growing cells in media containing 6-thioguanine, which adversely affects DNA replication in cells competent in HPRT. but has no measurable effect on cells with no functioning HPRT. Maher and colleagues have spent several years studying the effects of DNA repair on the induction of HPRT'ntmants by alkylating agents in a number of human fibroblast cell strains that differ in NER and AGT capacities (Simon et al.. 1981; Domoradzki et al.. 1984. 1985: Maher et al.. 1990). With methylating agents. the presence or absence of NER was found to make little difference in the toxicity of MNNG or the frequency of induced HPRT” mutants. Diploid fibroblast cell strain XP128E from a patient with xeroderma pigmentosum. complementation group A. which has negligible levels of NER repair (Cleaver and Bootsma. 1975). was not more sensitive than diploid human fibroblasts. designated SL68. from an apparently normal neonate to the cytotoxic and mutagenic effects of MNNG. However. the capability for AGT repair makes a large difference in the response of human fibroblasts to this carcinogen. Cells proficient in AGT activity demonstrate higher survival and lower mutation frequencies when treated with MNNG compared to the high cytotoxicity and mutagenicity of MNNG in cells that have a very low level of AGT activity. In contrast to the results shown for MNNG. the XP12BE cells are a great deal more sensitive than the SL68 cells to the cytotoxic and 26 mutagenic effects of ENU. This indicates that. as in bacteria. NER plays a larger role in the repair of ethyl adducts than does AGT (Simon et al.. 1981; Maher et al., 1990). One potential problem with studying repair of alkyl—DNA lesions by AGT or NER in cell strains of different repair phenotypes that are not isogenic is the possibility that the unidentified mutation that has rendered the cell incompetent in AGT (or NER) affects other pathways that contribute tangentially to NER (or AGT). For example. GM0011 cells have an uncharacterized, spontaneous mutation resulting iri the loss (if AGT activity (Middlestadt et al.. 1984). These cells also have a slightly decreased capacity for NER (Maher et al.. 1990). It could be that these two phenomena are related. It has been shown. for example. that conversion of AGT-competent human lymphoblastoid or fibroblast cell lines to AGT-incompetent cell lines can be accompanied by simultaneous loss of expression of thymidine kinase and galactokinase (Karran et al.. 1990). perhaps as a result of the loss of some common regulatory mechanism. Dolan et al. (1990) recently reported that CP-benzylguanine, as a free base in the media at a concentration of'SDM. could completely titrate out the AGT activity in AGT-proficient cultured human tumor cells. 06- benzylguanine was shown to be non-toxic at this dose in these cells. Thus. 06-benzylguanine can be used to study the repair of alkyl adducts in DNA in the presence or absence of AGT activity in cells that are isogenic. Maher and colleagues (Lukash et al.. 1991) have used CP-benzylguanine to study the affect of AGT repair on MNNG-induced cytotoxicity and mutagenicity iri the diploid human cell strain SL68. (F-benzylguanine depletes the cell of AGT to less than 1% of the normal level of activity. The AGT activity remains at this low level for up to 48 hours following 27 removal of the (E-benzylguanine. Using this inhibitor. Maher and colleagues showed that the cytotoxicity and mutagenicity of MNNG is much lower in the cells with normal levels of AGT compared to the cells depleted of AGT. This confirmed that AGT is directly involved in the repair of MNNG-induced cytotoxic and mutagenic lesions in diploid human fibroblasts. When Lukash et al. (1990) analyzed the HPRT mutations induced by MNNG in the AGT proficient and AGT deficient SL68 cells by sequencing the coding region of the HPRT'gene. they found that MNNG treatment results in predominantly G-C 9 A-T transition mutations in either phenotype. There are also a small percentage of mutations at T's. possibly from methyl lesions at T's. If one can assume that the G-C 9 A-T mutations arose as the result of a methylated purine. e.g. CP—MeG. then the majority of the G-C 9 A-T mutations (67%) occurred in the nontranscribed strand in the AGT-proficient cells. and 70% occurred in the nontranscribed strand in the AGT-depleted cells. This suggests that MNNG-induced premutagenic lesions in the HPRT gene occur preferentially in the nontranscribed strand. and that AGT repair of’CP-MeG in human fibroblasts is not strand-specific. Skopek and colleagues (Bronstein et al.. 1991) have treated EBV- transformed human lymphoblastoid cells with ENU and sequenced the coding region of the resulting HPRT-deficient mutants. The three cell lines they employed. designated X, A. and N. differed in repair capacity.1 The X cells. from a patient with XP. are completely deficient in NER capability. but are normal in AGT repair. The A cells are from a patient with hereditary spherocytosis and are devoid of AGT repair. but normal in NER. The N cells are normal in both kinds of repair. They found that the N cells. with NER and AGT both functioning. demonstrate a decreased 28 cytotoxic and mutagenic response to ENU compared to either the X or the A cells. with only AGT or NER working on the ENU-induced lesions. Interestingly. the X cells and the A cells are equally sensitive to both the cytotoxic and the mutagenic effects of ENU. Upon sequencing the HPRT gene of the resulting mutants. they found that all three cell lines show transition and transversion mutations at G's and T's following ENU treatment. There was no significant neighboring-base sequence bias or strand bias in the distribution of lesions. The most significant difference they found among the three cell lines was an 8-fold increase in the frequency of G-C 9 A-T mutations in the A cells. and a 3-fold increase in frequency of this mutation in the X cells. compared to the frequency observed in the N cells. They attribute this increase to a lack of repairtnyF-EtG by AGT. ‘They conclude that NER and AGT are both important in the repair of ENU-induced ethyl adducts in human cells. Skopek and colleagues continued in their investigation of the repair of ENU-induced alkylation damage in the X. A. and N cells (Bronstein et al.. 1992). They quantitated the removal of'CE-EtG. CF-EtT and CE-EtT adducts from the DNA of treated cells by measuring the binding of monoclonal antibodies to these lesions in DNA isolated from treated cells. They found that theiTKEtG lesions are removed from the DNA of the treated cells only when NER and AGT are both present in the cell. i.e in the N cells. The O‘-EtT and Oz-EtT lesions are not removed from the DNA of any of the three cell lines. Skopek and colleagues conclude that only cells expressing both AGT and NER are able to efficiently remove OS-EtG from DNA. and that neither NER nor AGT act on O“-EtT or Oz-EtT lesions. The results reported by Skopek and colleagues could be complicated 29 by the differences in the genetic background of the lymphoblastoid cell lines used. As discussed above. the different repair phenotypes observed in the lymphoblastoid cell lines could result from genetic alterations that affect multiple pathways in the cell. In our present study. we have employed 05-benzylguanine to specifically deplete diploid human fibroblasts of AGT. This has allowed us to study DNA repair in human cells. while limiting the number of genetically different cell strains used. We have used the XP128E cell strain. which is deficient in NER. These cells demonstrate an equal or slightly elevated capacity for AGT as that of the other cell strain used. SL68. which is normal in NER. Thus. the defect in NER of the XP12BE cells is not altering the capacity for AGT in this cell strain. By analyzing the induction and repair of alkyl adducts in normal. diploid human cells. we can obtain a better understanding of the initiation and processing of DNA lesions that lead to tumorigenic mutations in the human organism. REFERENCES BARTSCH. H.. OHSHIMA. H.. SHUKER. D.E.G.. PIGNATELLI. 8.. AND CALMELS. S. (1990) Exposure of humans to endogenous N-nitroso compounds: implications in cancer etiology. MDtat. Res.. 238. 255-267. BASU. A.K. AND ESSIGMANN. J.M. (1990) Site-specifically alkylated oligodeoxynucleotides: Probes for inutagenesis. DNA repair, and the structural effects of DNA damage. MDtat. Res.. 233. 189 - 201. BELINSKY. S.A.. WHITE. C M.. BOUCHERON. J.A.. RICHARDSON. F.C.. SWENBERG. J.A . AND ANDERSON, M. (1986) Accumulation and persistence of DNA adducts in respiratory tissue of rats following multiple administrations of the tobacco specific carcinogen 4-(N-methyl-N-nitrosoamino)-1-(3-pyridyl)-1- butanone. Cancer Res.. 46. 1280-1284. BELINSKY. S A.. DEVEREUX. T R.. AND ANDERSON, M.W. (1990) Role of DNA methylation in the activation of proto-oncogenes and the induction of pulmonary neoplasia by nitrosamines. MDtat. Res.. 233. 105-116. BERANEK. D.T. (1990) Distribution of methyl and ethyl adducts following alkylation with monofunctional alkylating agents..MDtat. Res.. 231. 11-30. BHANOT. O. S. GREVATT. P. C DONAHUE. J. M.. GABRIELIDES C. N. AND SOLOMON. J. J. (1991) In vitro DNA replication implicates C)2 —ethyldeoxythymidine in Eganggirsion mutagenesis by ethylating agents. NDcleic Acids Res 20. 7- . BOOTSMA. D. AND HOEIJMAKERS. J.H.J. (1994) The molecular basis of nucleotide excision repair syndromes. MDtat. Res.. 307. 15-23. BRENT. T.P.. DOLAN. M.E.. FRAENKEL-CONRAT, H.. HALL. J.. KARRAN. P.. LAVAL. F.. MARGISON, G.P.. MONTESANO, R.. PEGG. A.E.. POTTER. P.M.. SINGER. 8.. SWENBERG. J.A.. AND YAROSH. 0.8. (1988) Repair of O- alkylpyrimidines in mammalian cells: a present consensus. Proc. Aetl. Acad. Sci. U S.A. 85. 1759-1762. BRONSTEIN. S.M.. COCHRANE. J.E.. CRAFT, T.R.. SWENBERG. J.A.. AND SKOPEK. T.R. (1991) Toxicity. mutagenicity. and mutational spectra of N-Ethyl-N- Nitrosourea in human cell lines with different DNA repair phenotypes. Cancer Res.. 51. 5188-5197. BRON65TEIN S. M. SKOPEK. T. R. AND SWENBERG. J2 A. (1992) Efficient repair of O6 -ethylguanine. but not 0 -ethylthymine or 02 -ethylthymine. is dependent upon C)6 -alkylguanine- DNA alkyltransferase and nucleotide excision repair activities in human cells. Cancer Res.. 52. 2008- 2011. BURNS. P.A.. GORDON. A.J.E.. KUNSMANN. K.. AND CNICKMAN, B.W. (1988) Influence of Neighboring base sequence on the distribution and repair of N-ethyl-N-nitrosourea-induced lesions in E. coli. Cancer Res.. 48. 4455- 4458. 30 31 CHAMBERS. R.W. . SLEDZIEWSKA- GOJSKA. E.. HIRANI- HOJATTI. 3.. AND BOROWY- BOROWSKI. H. (1985) uvrA and recA mutations inhibit a site- specific transition produced by a single 06 -methylguanine in gene G of bacteriophage ¢X174. Proc. Natl Acad. Sci U S A. 82. 7173- 7177. CHINAULT. A.C.. AND CASKEY. C.T. (1984) The hypoxanthine phosphoribosyltransferase gene: a model for the study of mutation in mammalian cells. Prog. Nucleic Acid Res. Mb]. Biol.. 31. 295-313. CLEAVER. J.E.. AND BOOTSMA. D. (1975) Xeroderma pigmentosum: biochemical and genetic characteristics. Annu. Rev. Genet.. 9. 19-38. CLEAVER. J.E. (1990) Do we know the cause of xeroderma pigmentosum? Carcinogenesis. 11. 875-882. DOLAN. M. E. MOSCHEL. R. C. AND PEGG. A. E. (1990)6 Depletion of mammalian O6 -alkylguanine- DNA.alkyltransferase activity byO6 -benzylguanine provides a means to evaluate the role of this protein in protection against carcinogenic and therapeutic alkylating agents. Proc. Natl. Acad. Sci. U S A. 87. 5368-5372. DOMORADZKI. J. PEGG. A E. GDOLAN. M E.. MAHER. V M. AND MCCORMICK. J. J. (1984) Correlation between D6 -methylguanine- DNA methyltransferase activity and resistance of human cells to the cytotoxic and mutagenic effect of N- methyl-N -nitro-N-nitrosoguanidine. Carcinogenesis. 5. 1641-1647. DOMORADZKI. J.. PEGG. A.E.. DOLAN. M.E.. MAHER. V.M.. MCCORMICK. J.J. (1985) Depletion of (f-methylguanine-DNA methyltransferase in human fibroblasts increases the mutagenic response to N-methyl-N'-nitro-N- nitrosoguanidine, Carcinogenesis, 6. 1823-1826. DOSANJH. M K.. MENICHINI. P.. ERITJA. R.. AND SINGER. 8. (1993) Both 04- methylthymine and 044ethylthymine preferentially form alkyl T:G pairs that d: no; glogkgin vitro replication in a defined sequence. Carcinogenesis. 1 . 1 1 - 1 . DUMENCO. L.L ALLAY. E. NORTON K. GERSON. S. L. (1993) The prevention of ‘thymic lymphomas in transgenic mice: by human 06 -alkylguanine- DNA alkyltransferase. Science 259. 219- 222. ECKERT. K.A.. INGLE. C.A.. KLINEDINST. D.K.. AND DRINKWATER. N.R. (1988) Molecular analysis of mutations induced in human cells by N-ethyl-N- nitrosourea, Mb]. Carcinog.. 1. 50-56. ECKERT. K A.. INGLE. C.A.. AND DRINKWATER. N.R. (1989) N-ethyl-N- nitrosourea induces A:T to C:G transversion mutations as well as transition mutations in SOS induced E. coli. Carcinogenesis. 10. 2261- 2267. FINE. D.H.. ROUNBEHLER. D P.. FAN. T.. AND ROSS. R. (1977) Human exposure to N-nitroso compounds in the environment. in Origins of'HUman Cancer. 4. H.H. Hiatt. J. 0. Watson. J. A. Winsten. eds. 32 FRIEDBERG. E.C.. BARDWELL. A J.. BARDWELL. L.. WANG, Z.. DIANOV. G. (1994) Transcription and nucleotide excision repair - reflections. considerations and recent biochemical insights. MDtat. Res.. 307. 5-14. GORDON. A.J E.. BURNS. P.A.. AND GLICKMAN. B.W. (I990) N-Methyl-N'-nitro- N-nitrosoguanidine induced DNA sequence alteration: non-random components in alkylation mutagenesis. MDtat. Res.. 233. 95-103. GOTH. R. AND RAJEWSKY. M.F. (1974) Persistence CG NT AAA ITC TTT PHE to LEU XBE1 16 1.15 241 3 TA - CG NT CAT IAC ATC TYR to HIS X8E107 1.15 258 3 TA -. cc NT CTG AAI AGA No change“ X862 0.5 308 3 TA -v CG T CTG AAG AGC LYS to ARG X851 28 1.0 475 6 TA -’ CG T GTC AAG GTC LYS to GLU XBE100 1 .15 542 8 TA - CG NT GGA TIT GAA PHE to SER XBE127 1.0 602 8 TA .. CG T AGG GAT TTG ASPto GLY XBE34 1.0 466 6 TA - AT T CCA AAG ATG LYS to Stop X86125 1 .0 548 8 TA - AT NT GAA AIT CCA ILE to ASN XBE18 0.5 590 8 TA —. AT T AAT GAA TAC GLU to VAL XBE129 1.0 49 2 TA » GC NT GGT IAT GAC TYR to ASP XBE126 1.0 49 2 TA -. GC NT GGT IAT GAC TYR to ASP XBEZ4° 1.5 410 6 TA .. GC NT ATA AIT GAC ILE to SER XBE8° 0.5 494 7 TA -» GC NT CTG GIG AAA VALto GLY XBE8° 0.5 595 8 TA - GC NT TAC ITC AGG PHE to VAL XBE14c 1.0 625 9 TA -» GC T ATT AGT GAA SER tn ARG 3T, transcribed; NT, nontranscribed. bThe sequence written 5' to 3' is that of the nontranscribed coding strand. cThis mutant contained more than one mutation. dNo other mutation was detected. 53 Table I. (Continued) ENU Missing Mutant (mM) Exon X85103 X8531 X85104 X8519 X85130 X85131 X85132 X85133 X85111 X8517 X8523 do—b—I—I-J—loo—I—I U'I 'o'mL-b'o'o'o'm'm'o'o 7 bp missing from the first part of exon 9 54 Table II. Kinds and locations of mutations induced by ENU in the coding region of the HPRT gene in the absence of nucleotide excision repair, but in the presence of alkyltransferase activity Strand with Affected Neighboring Amino acid Mutant Site Exon Mutation G or Ta basesb Change X5124° 1.0 172 3 GO -v TA NT ATG GGA GGC GLY to Stop X525 1.0 197 3 GC -. TA NT CTC T_C_5_T GTG CYS to PHE X54 0.4 202 3 GC -> TA T GTG QTC AAG LEU to ILE X543° 0.9 400 5 GC -’ TA NT GTG GAA GAT VAL to PHE X5128 1.0 419 6 GC .. TA NT ACT G_C_SC AAA GLY to VAL X5121 1.0 197 3 GO -' AT NT CTC TQT GTG CYS to TYR X5126 1.0 197 3 GC - AT NT CTC T_G_T GTG CYS to TYFI X544 0.9 464 6 GC -. AT T ATT CQA AAG PRO to LEU X52 0.5 110 2 TA -> CG NT TTT AIT CCT ILE to THR X5103 1 .1 219 3 TA + CG T TAT AA_A_ TTC PHE to LEU X527 1.0 290 3 TA -’ CG NT ACT GIA GAT VAL to ALA X523° 0.5 290 3 TA - CG NT ACT GIA GAT VAL to ALA XE124° 1.0 383 4 TA -o CG T GGA AAG AAT LEU to TRP X5125 1.0 499 7 TA - CG T AAA AGG ACC ARG to GLY X548 1.1 530 7 TA -> CG T CCA G_A_C TTT ASP to GLY X529 1.5 533 8 TA -. CG NT GAC TIT GTT PHE to SER X56 1 .0 592 8 TA -+ CG NT GAA IAC TTC TYR to HIS X523° 0.5 605 8 TA » CG NT GAT TIG AAT LEU to SER X5122 0.75 95 2 TA .. AT NT GAT TIG GAA LEU to Stop X530 .5 104 2 TA -. AT NT AGG GIG TTT VAL to GLU X547 .15 215 3 TA - AT T GGC TAT AAA TYR to PHE X540 .65 247 3 TA - AT T ATC AAA GCA LYS to Stop X531 .5 290 3 TA - GC NT ACT GIA GAT VAL to GLY X5123 .75 290 3 TA .. GC NT ACT GIA GAT VAL to GLY X542 .9 473 6 TA -o GC NT ATG GIC AAG VAL to GLY X5127 .75 473 6 TA -v GC NT ATG GIC AAG VAL to GLY XE43° .9 541 8 TA - GC NT GGA ITT GAA PHE to VAL X55 .5 547 8 TA - GC T GAA ATT CCA ILE to LEU X526 .0 547 8 TA -. GC T GAA ATT CCA ILE to LEU ENU Missing Mutant (mM) Exon X5101 1.15 4 X51 1 0.5 6 X524 1.0 7 X536 0.65 8 aT, transcribed; NT, nontranscribed. bThe sequence written 5' to 3' is that of the nontranscribed coding strand. °This mutant contained more than one mutation. 55 Table III. Kinds and locations of mutations induced by ENU in the coding region of the HPRT gene in the presence of nucleotide excision repair, but in the absence of alkyltransferase activity Strand with ENU Affected Neighboring Amino acid Mutant (mM) Site Exon Mutation G or T3 basesb Change BEN122 2.5 6 1 GC —. TA NT ATG GCQ ACC No changec 85N77 1.5 634 9 GC 4 TA NT ACT GGA AAA GLY to Stop BEN117 2.5 197 3 GC -’ AT NT CTC TGT GTG CYS to TYR BEN11 1.0 209 3 GC » AT NT AAG G_GG GGC GLY to GLU 85N29° 2.5 336 4 GC -. AT NT ACA GGQ GAC No changed BEN103 1.8 464 6 GC -» AT T ATT CQA AAG PRO to LEU BEN137 3.5 539 8 GC - AT NT GTT G_G_A TTT GLY to GLU BEN141 3.5 539 8 GC _. AT NT GTT GQA TTT GLY to GLU BEN46 2.5 635 9 GC _. AT NT ACT GQA AAA GLY to GLU 85N108 1.8 65 2 TA -- CG NT TTA TIT TGC PHE to CYS BEN112 2.5 122 2 TA - CG NT GGA CIA ATT LEU to PRO BEN114 2.6 125 2 TA _. CG NT CTA AIT ATG ILE to THR BEN120 2.5 203 3 TA -> CG NT GTG CIC AAG LEU to PRO BEN138 3.5 236 3 TA -. CG NT CTG CIG GAT LEU to PRO BEN1 10 2.5 530 7 TA - CG T CCA GAC TTT ASP to GLY 85N29° 2.5 66 2 TA -. cc NT TTA TTI TGC PHE to LEU BEN48 2.5 64 2 TA - AT NT TTA ITT TGC PHE to |L5 85N76 1.5 1 10 2 TA -. AT NT TTT AIT CCT ILE to ASN BEN35 1.5 216 3 TA -* AT NT GGC TAI AAA TYR to Stop BEN66 1.5 245 3 TA .. AT NT TAC AIC AAA ILE to ASN BEN139 3.5 284 3 TA —. AT NT CCT AIG ACT MET to LYS BEN107 1 .8 284 3 TA —v AT NT CCT AIG ACT MET to LYS BEN15 1.0 573 8 TA _. AT NT GGA TAI GCC TYR to LEU 85N55 1.0 614 9 TA -v AT NT CAT GIT TGT VAL to ASP BEN39 1. 299 3 TA -o GC NT TTT AIC AGA ILE to SER BEN1 13 2. 330 4 TA -» 60 T CAG TCA ACA No change BEN140 3.5 395 5 TA .. GC NT TTG AIT GTG |L5 to SER BEN101 2.6 473 6 TA - GC NT ATG GIC AAG VAL to GLY 85N4 0.5 522 7 TA -. GC NT GGA TAI AAG TYR to Stop BEN37 1.5 542 8 TA .. GC NT GGA TIT GAA PHE to CYS BEN102 2.6 543 8 TA - GC NT GGA TTI GAA PHE to LEU 85N96 2.5 614 9 TA - GC NT CAT GIT TGT VAL to GLY 3T, transcribed; NT, nontranscribed. bThe sequence written 5' to 3' is that of the nontranscribed coding strand. °No other mutation was detected. dThis mutant contained more than one mutation. 56 Table III. (Continued) ENU Missing Mutant (mM) Exon 85N51 2.5 4 BEN68 2.5 6 85N44 2.5 8 BEN100 1.8 8 85N8 0.5 8 85N104 1 .8 8 85N4O 2.5 14 bp missing from the first part of exon 9 (610-623) BEN64 2.5 17 bp missing from the first part of exon 9 (610-626) 57 Table IV. Kinds and locations of mutations induced by ENU in the coding region of the HPRT gene in the presence of nucleotide excision repair and alkyltransferase activity Strand with ENU Affected Neighboring Amino acid Mutant (mM) Site Exon Mutation G or T“ basesb Change 5N105 1.8 96 2 GC -» TA NT GAT TT§ GAA LEU to PHE 5N58° 1.5 1 12° 2 GO - TA T ATT QCT CAT PRO toTHR 5N42° 1.5 130 2 GC -» TA NT ATG G_AC AGG ASP to TYR 5N48 1.0 197 3 GC - TA NT CTC T_GT GTG CYS to PHE EN41 1.5 539 8 GC -> TA NT GTT GQA TTT GLY to VAL 5N44 1.5 575 8 GC -> TA T TAT GCC CTT ALA to ASP 5N11 1.5 575 8 GC —. TA T TAT GQC CTT ALA to ASP 5N80 3.5 119 2 GC -> AT NT CAT GGA CTA GLY to GLU EN55 1.5 149 3 GC —. AT T CTT GQT CGA ALA to VAL 5N83 3.0 151 3 GC -> AT T GCT QGA GAT ARG to Stop EN50 1.0 202 3 GC _. AT T GTG QTC AAG LEU to PHE EN102 2.6 334 4 GC -’ AT NT ACA QGG GAC GLY to ARG 5N63° 2.5 429 6 GC _. AT NT ACA ATG CAG MET to ILE 5N63° 2.5 601 8 GC -. AT NT AGG GAT TTG ASP to ASN 5N1 10 1.8 125 2 TA —. CG NT CTA AIT ATG ILE to THFl EN51 1.0 170 3 TA _. CG NT GAG AIG GGA MET to THR 5N100 1.8 214 3 TA _. CG NT GGC IAT AAA TYR to HIS 5N37 1.5 1 10 2 TA -> AT NT TTT AIT CCT ILE to ASN EN82 3.5 194 3 TA -' AT NT GCC CIC TGT LEU to HIS 5N2 1.0 198 3 TA - AT NT CTC TGI GTG CYS to Stop EN85 3.5 290 3 TA - AT NT ACT GIA GAT VAL to GLU 5N22° 2.5 537 8 TA -> AT NT TTT GTI GGA No change° EN112 2.6 541 8 TA a AT NT GGA ITT GAA PHE to ILE 5N22° 2.5 542 8 TA _. AT NT GGA TIT GAA PHE to TYR 5N3 1.0 566 8 TA —- AT NT GTT GIA GGA VAL to GLU EN62 2.5 566 8 TA —» AT NT GTT GIA GGA VAL to GLU EN84 3.5 64 2 TA # GC NT TTA ITT TGC PHE to VAL 5N58° 1.5 1 1 1° 2 TA » GC NT TTT ATI CCT (LE to MET EN43 1.5 136 3 TA 4 GC T AGG ACT GAA THR to PRO EN66 2.5 185 3 TA -> GC NT CAC AIT GTA ILE to SER 5N81 3.5 299 3 TA -v GC NT TTT AIC AGA ILE to SER EN53 1.5 449 6 TA -. GC NT TTG GIC AGG VAL to GLY aT, transcribed; NT, nontranscribed. bThe sequence written 5' to 3' is that of the nontranscribed coding strand. cThis mutant contained more than one mutation. dThis mutation is next to a second mutation in the same mutant. If the lesions involved a G and a T, these were on opposite strands. 58 Table IV. (Continued) ENU Missing Mutant (mM) Exon 5N60 2.5 4 5N1 1 1 1.8 4 5N17 1.5 7 5N13 1 .5 8 5N108 2.6 8 5N42° 1.5 8 5N32 1.0 8 5N39 1.5 8 5N21 2.5 21 bp missing from the first part of exon 8 (533-553) 5N106 1.8 8 bp from the last part of intron 2 have been inserted between exon 1 & 2, probably due to a change in a splice site 59 positions 197 and 290. Iri ENU-treated E. ammno mcowpmpss cowpzpwpmpzm mmmn Co Accessm .mmPpWUmqmu LCQOL Hu< use amz c? mcwgm54_u mummPQOany cuss; u_o_a_c umpmmcu .> mpnmh 61 the distribution of substitutions seen in the other three ENU-treated populations. As shown in Table V. 63% involve T-A base pairs and 37% involve G-C base pairs. In the NER/AGT+ cells. the frequency of mutants decreased by ~35% relative to that observed in the NER/AGT‘ cells (Figure 2). If one assigns the fractions of mutations in this populathmi per mM dose as described above. one can analyze the distribution for decreases in particular classes of mutations to gain insight into the effect of AGT repair (see Table V). The majority of the decrease in mutations involved lesions producing G-C 9'A T base substitutions. This could be interpreted to indicate that the lesion produced in this class of base substitution. i.e. D°-EtG. was removed by AGT. There was also a decrease in the frequency of T-A 9 C-G transitions compared to the frequency in the NER‘ /AGT“ cells. This suggests that AGT can also remove ethyl groups from 0‘- EtT to some extent. There was little or no decrease in the contribution of classes other than G-C 9 A-T and T-A 9 C-G. In particular. there was no decrease in the contribution to the overall mutant frequency of lesions resulting in T A-9.A T transversions compared to that seen in the NERf/AGT’ cells. The decrease in lesions responsible for 6C 9 AT transitions. presumably D°-EtG, does not show a bias for repair of the transcribed strand. Rather it shows greater loss from the nontranscribed strand. the strand initially containing the majority of these lesions (compare Tables I and II). The ratio of this lesion in the transcribed strand relative to the nontranscribed strand is 12:88 in the NER/AGT’ cells and 33:67 in the NERf/AGT+ cells. This decrease in strand bias is reflected in Figure 3. When normal cells proficient in NER but depleted of AGT by 0°-BZG are ._. + Ampv m 0< mmz ._.0< mmz +. I .+ Aev P AmmVMw s3 #2 Amwvmw snmez Amwv m ._.0< mmz ._..O< mmz + u n . 2UNV m sea, A Ammv m smmpz 2mm; o smfihz maven mmmn u.u um mcowgzuvumnsm mmmn Co amass: n ummcpcmLmq Co muvmuzo amass: .UCmLum umawcumcmgp u .mcoruzprumnzm.mmmn umuzuc_-:zm :_ um>_o>:_ mo:_:m3m Co cowuzpwapm_c ucaeum omen um>cmmno Co amass: u -coz n 42 .A> m_nm» mmmv cowpm_:aoa emacmp asp cw;u_z mwmmcpcmcmq :_ amass: .mLqu mmma u a pm mco_p:u4umn:m A .ucmcam umn_58mcmcp .m mesm_m 63 compared with the cells lacking both kinds of repair. the overall frequency of mutants decreases ~62% (Figure 2). A similar analysis of the decreases in classes of base substitutions (Table V) strongly suggests that in these cells. NER is able to repair lesions resulting in G-C 91A T. G-C 9 T-A. T-A 9 C-G, and T-A 9 G-C base substitutions. There was no decrease in T-A 9.A T base substitutions. suggesting little or no excision repair of the lesions resulting in this class of base substitution, presumably Oz-EtT. NER appears to be as efficient as AGT alone in removing lesions leading to (SC 9 AT transitions. i.e. 0°-EtG. but far more efficient than AGT in removing lesions that lead to T-A 9 C-6 transitions (probably'Cf-EtT) and to T-A 9 G-C transversions (lesion unknown). There was no evidence in NERTAAGT' cells of a preferential removal of O°-EtG from the nontranscribed strand. as there was no change in strand bias from that seen in the NBTVAGT“ cells. Analysis of the change in strand distribution of the lesions resulting in T-A 9 C-G substitutions. presumably Cf-EtT. supports the hypothesis that strand-specific excision repair occurred. (See Tables I and III.) The ratio of these lesions in the transcribed vs. nontranscribed strand is 40:60 in the NER'/AGT' cells and 14:86 in the NERl/AGT' cells. The bias is reflected in Figure 4. Evidence that NER and AGT acting in the same cell do not have an additive effect. In the cells proficient in both NER and AGT. the overall frequency of mutants decreased by ~67% relative to the NER /AGT' cells (Figure 2). This overall decrease is not significantly greater than that observed in NERl/AGT' cells (Figure 2). However. analysis as above (Table V) indicates that the decrease in the class of G-C 91A T transitions is very similar to +HG< mwz FG<.mmz HO< mwz HG< mwz + I + + I I I Amvr Aevm Ammvm AOKVK Aonvrm Avmvmw AmFFVNF Ammvnw srmez smmez smopz ewvm pa; .A> cramp wmmv co_pmrsaoq meLm_ m5“ crguwz m5_ca mmmn <.H um mcorpzpwpmnzm omen Co amass: n m_mmcpcmcma cw amass: ”memQ amen <.H pm mcovpspwumnzm amen um>5mmno Co 5625:: u mwmmcpcmcma Co mu_mpzo amass: ”ucmcpm umnvcumcmge u H uucmcpm cmnwLUWCmLp -coz u Hz .m502pzuwpmnsm omen umozvcr-:zm cw um>Fo>:_ mQCAExcp Co covpsnwapm_u ccccum .4 mgzm?d 65 what was seen in NER/AGT+ cells. as if AGT repaired these lesions. Consistent with this interpretation is the change in strand distribution of these lesions (Compare Tables I. II. and IV). The data show a bias in the loss of such lesions from the nontranscribed strand. as was seen when AGT'was acting in the absence of NER. This bias is reflected in Figure 3. In contrast. the NERVAGT+ cells reduced the contribution of lesions leading to TA 9 CG transitions (i.e. 04-EtT) significantly. to a frequency lower than was observed in the NERI/AGT‘ cells. This loss of D‘- EtT lesions showed a strong preference for the transcribed strand (ratio of transcribed strand nontranscribed strand. 40 60 in NERf/AGT'. 0:100 in NERT/AGTT). This supports the hypothesis that excision repair is responsible for the decrease in this class of substitutions. This change in bias is reflected in Figure 4. Once again. there was no evidence that either AGT or NER decreased the contribution of lesions giving rise to T-A 9 AT (i.e. DZ-EtT). DISCUSSION We found that both NER and AGT are important in repairing ENU- induced lesions in the HPRT gene in human fibroblasts. Alkyltransferase acting independent of NER was shown to be operative in the repair of ENU- induced lesions. as we saw that the cytotoxicity and mutagenicity of ENU decreased in the NERf/AGTT cells relative to that in the NERf/AGT‘ cells. NER acting independent of AGT was shown to be instrumental in repair of ENU-induced premutagenic lesions. because in the NER proficient SL68 cells the cytotoxicity and mutagenicity of ENU were greatly reduced compared to that observed in the NER deficient XP cells. Our results in diploid human fibroblasts are not similar to those reported by Bronstein. et al. (17. 25). They reported that the constitutive absence of AGT in lymphoblasts normal in NER significantly increases the cytotoxic and mutagenic effects of ENU relative to cells normal in NER and AGT. Yang et al. also report a similar finding in fibroblasts (15). In our NER proficient cells. depletion of AGT did not significantly increase the ENU-induced cytotoxicity or mutation frequency. One explanation for the differences between our observations and theirs could be that.Cf-benzylguanine did not adequately deplete the SL68 cells of AGT. Dolan et al. (27) have shown. however. that treatment of cells with as low as 2.5 IWIIF-BZG results in a rapid decrease in AGT activity. It was shown previously in our laboratory that a 2 hour pre- treatment of SL68 cells with 25;Ai(f-BZG titrates out the AGT activity in these cells to less than 2% of the original activity. and an additional 24 hours of'Cf-BzG treatment leaves the levels of AGT at <1% of the control for up to 28 hours after removal of inhibition (28). Ne incubated ENU— 66 67 treated cells in the presence oflTLBzG for 48 hours. allowing a total of 96 hours with no appreciable .AGT‘ activity following ENU treatment. Therefore. even though cells can rapidly regenerate AGT (37). it is unlikely that regeneration or inadequate depletion of AGT can explain the similar toxic and mutagenic effect of ENU observed in the NERl/AGTT and NERl/AGT“ cells. Another possible explanation for the difference between our observations and those of Bronstein et al. and Yang et al. of the effect of AGT on ENU toxicity and mutagenicity is that the genetic change resulting in a constitutive lack of AGT activity also affects other pathways of DNA repair. Karran et al. (38). for example. demonstrated that the loss of AGT activity in human lymphoblastoid or fibroblast cell lines could be accompanied by simultaneous loss of expression of thymidine kinase and galactokinase. possibly due to loss of some common regulatory mechanism. GM0011. the human fibroblast cell strain used by Yang et al. (15). is constitutively devoid of AGT. and has also been shown in our laboratory to be somewhat lacking in NER capacity compared to the NER proficient SL68 cells used in our present study (29). It may be that the small difference in NER capability between SL68 and GM0011. as measured by survival of cloning ability following UV treatment. becomes a larger disadvantage in the presence of alkylation damage. Our results show that NER/AGT cells were able to repair the cytotoxic and mutagenic lesions caused by ENU. indicating that cooperation between NER and AGT was not absolutely required in the repair of these lesions. It could be argued that the SL68 cells and XP128E cells differ in more than just NER capacity. The SL68 and XP12BE cells. however. showed nearly the same cytotoxic and mutagenic response to MNNG (28). indicating that these cells respond in a similar manner to alkylation when 68 NER is nonessential for repair of this damage. Although one could also argue that there is some residual NER activity in our NERf/AGTlluells. this capacity for repair would be shared by the NER/AGT“ cells. Therefore. the most likely explanation for the decrease in mutant frequency and increase in survival in the NER/AGT+ cells compared to the NER/AGT‘ cells is that AGT is repairing ENU-induced lesions independent of NER activity. The human AGT, like the bacterial AGT. repairs O°-EtG by transferring the alkyl group irreversibly to a cysteine in its own amino acid sequence (reviewed in 23). The bacterial AGT is able to repair'CY-alkle lesions (19). Recent studies have indicated that human AGT can also repair 0“- alkle in vitro. but much less efficiently than Cf—alkylG (39.40). The sequencing data in our present study (Table V) indicate that there is a decrease in the frequency of T-A 9 C-G mutations in the NER/AGT+ cells relative to the frequency in the NER/AGT' cells. suggesting repair of the (Y-EtT lesions by AGT. Again. these results in fibroblasts conflict with those reported by Skopek and colleagues. who observed that lymphoblasts deficient in NER but proficient in AGT are not able to repair O°-EtG or 0°- EtT lesions. In bacteria. UvrABC can repair 0°-EtG and O‘-EtT (19). Skopek and colleagues reported recently that in human lymphoblasts. NER is not able to efficiently repair‘Cf-EtG lesions in genomic DNA in the absence of AGT (17.25). Our results indicate. however. that in human fibroblasts NER can effectively repair'Cf-EtG lesions in the HPRT gene in the absence of AGT. Skopek and colleagues reported that lymphoblasts were unable to repair 0‘- EtT lesions regardless of NER or AGT (25). and that the difference in spectra between NERVAGT+ lymphoblasts and lymphoblasts deficient in NER or AGT was primarily due to G~C 9 A-T transitions (17). However. we show 69 that NER. in the absence of AGT activity. significantly reduced the frequency of the TA 9 CG transitions likely resulting from 04-EtT lesions. and also that NER reduced the T-A 9 G-C transversions caused by an unknown lesion compared to the frequency of these mutations in the NER' /AGT“ cells. In vitro studies have indicated that DNA polymerases can incorporate deoxythymidine opposite Oz-EtT. resulting in TA 9 AT transversions (18). This transversion mutation is believed to account for mutational activation of the neu proto-oncogene in ENU—treated rats (3). and to be the causative mutation in the mouse a- and B-globin genes in the progeny of the ENU-treated mouse (4,41). In ENU-treated rats. the majority of the mutations observed in pouch skin fibroblasts were T A-9.A T transversions, with G-C 9 A-T. T-A 9 C-G. and T-A 9 G-C substitution mutations occurring at low frequency (42). This is the expected result if repair of OZ-EtT is slower than that of the other potentially mutagenic lesions induced by ENU. There appears to be some removal of these lesions. as measured by decay of Oz-EtT adducts over time, in human lymphoblasts (25). Human fibroblasts in our study did not demonstrate reduction in frequency of T-A 9 A-T transversions regardless of NER or AGT repair capacity compared to the frequency in the NER/AGT' cells. indicating that NER and AGT were not involved in the repair of the OZ-EtT lesions. If this is a very persistent lesion. it may be that fibroblasts in the rat skin "pouch" (42) had time to repair the other ENU-induced lesions before replication. leaving the T-A 9 A-T transversion as the most frequently observed mutation. It has been demonstrated that in eukaryotes. components of the NER machinery are coupled to the transcription machinery (reviewed in 43). and NER has been shown to work more efficiently on UV-induced damage in the 70 transcribed strand vs. the non-transcribed strand (44-46). In both prokaryotic and eukaryotic systems there have also been reports indicating preferential repair of the transcribed strand by AGT (42.47.48). Our results suggest that NER preferentially removed ENU-induced damage in the transcribed strand of the HPRT gene of diploid human fibroblasts. AGT appeared to concentrate on removal of the O°-EtG lesions in the nontranscribed strand. most likely because that was the strand initially containing the most O°-EtG. ‘This is in agreement with the results seen for MNNG-induced O°-MeG premutagenic lesions by Lukash et al. (28) showing that AGT exhibits no bias for repair of lesions at 6's in the txanscribed strand. Interference. rather than cooperation. between NER and AGT has been observed in bacteria (49.50). If, in human cells. NER and AGT were working cooperatively on ethyl lesions at 6'5 as suggested by Skopek and colleagues (17,25). then we would expect to see the frequency of mutations at 6’5 in the NERl/AGT+ cells decrease relative to the frequencies observed in NER/AGT+ or NERT/AGT‘ cells. Instead. we saw that the frequency remained nearly the same. We would also expect that in the cells competent in NER and AGT. the lesions remaining at 6’5 would be found predominately on the nontranscribed strand. since NER would be working diligently on the transcribed strand and AGT would be working randomly on either strand. The strand distribution we saw. however. did not fit this pattern. but showed an unexpected inversion in strand distribution of the lesions remaining at 6's in the NERTLAGTT cells compared to the NERl/AGT’ cells (see Figure 3). We speculate that NER and AGT. when both present in the cells. compete for the repair of premutagenic lesions at 6'5. 10. 11. REFERENCES Sancar. A. and Sancar. 6.8. (1988) DNA repair enzymes. Annu. Rev. Biochem.. 57. 29 - 67. Pegg, A.E. (1990) Properties of mammalian O°—alkylguanine-DNA transferases. MDtat. Res. 233. 165 - 175. Perantoni. A.O.. Rice. J.M.. Reed. C.D.. Watatani, M.. and Nenk M.L. (1987) Activated neu oncogene sequences in primary tumors of the peripheral nervous system induced in rats by transplacental exposure to ethylnitrosourea. Proc. Natl..Acad. Sci. U.S.A, 84. 6317 - 6321. Lewis. S.E.. Johnson. F.M.. Skow. L.C.. Popp. D.. Barnett. L.B.. and Popp. R.A. (1985) A mutation in the B-globin gene detected in the progeny of a female mouse treated with ethylnitrosourea. Proc. Natl. Acad. Sci. U S A. 82. 5829-5831. Goth. R.. and Rajewsky, M.F. (1974) Persistence of'Cf-ethylguanine in rat-brain DNA: Correlation with nervous system-specific gircgggggggsis by ethylnitrosourea. Proc. Natl. Acad. Sci. U S A. Bartsch. H.. Ohshima, IL. Shuker. D.E.G.. Pignatelli. 8.. and Calmels, S. (1990) Exposure of humans to endogenous N-nitroso cggpounds: implications in cancer etiology: Mbtatu Res. 238. 255 - Singer. 8.. Bodell. W.J.. Cleaver. J.E.. Thomas. G.H.. Rajewsky. M.F., and Thon. w. (1978) Oxygens in DNA are main targets for ethylnitrosourea in normal and xeroderma pigmentosum fibroblasts and fetal rat brain cells. Nature. 276. 85-88. Singer. 8. (1985) In vivo formation and persistence of modified nucleosides resulting from alkylating agents. Environ. Health Perspect. 62. 41-48. Singer. 8. and Dosanjh. M.K. (1990) Site-directed mutagenesis for quantitation of base-base interactions at defined sites. Amwat. Res.. 233. 45-51. Singer. 8. (1979) N-nitroso alkylating agents: formation and persistence of alkyl derivatives in mammalian nucleic acids as contributing factors in carcinogenesis. J. Natl. Cancer Inst.. 62. 1329-1338. Basu. A.K.. and Essigmann. J.M. (1990) Site-specifically alkylated oligodeoxynucleotides: Probes for mutagenesis. DNA repair. and the structural effects of DNA damage. Murat. Res.. 233. 189 - 201. 71 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 72 Klein. J.C.. Bleeker. M.J.. Lutgerink. J.T.. van Diuk. W.J.. Brugghe. J.F.. van den Elst. H.. van der Marel. G.A.. van Boom. J.H.. Nestra. J.G.. Berns. A.J.M.. and Kriek. E. (1990) Use of shuttle vectors to study the molecular processing of defined carcinogen-induced DNA damage: mutagenicity of single 0‘- itgylthymine adducts in HeLa cells. Nucleic Acids Res.. 18. 4131 - Richardson. K.K.. Richardson. F.C.. Crosby. R.M.. Swenberg. J.A.. and Skopek. T.R. (1987) DNA base changes and alkylation following in vivo exposure of E. coli to N-methyl-N-nitrosourea or N-ethyl-N- nitrosourea. Proc. thl. Acad. Sci. U S A. 84. 344 - 348. Eckert. K.A.. Ingle. C A.. Klinedinst. D.K.. and Drinkwater. N.R. (1988) Molecular analysis of mutations induced in human cells by N- ethyl-N-nitrosourea. Mbl. Carcinog.. 1. 50 - 56. Yang. J-L.. Lee. P-C.. Lin. S-R.. and Lin. 3-6. (1994) Comparison of the mutation spectra induced by N-ethyl-N-nitrosourea in the hprt generof Mer’and Mer diploid human fibroblasts. Carcinogenesis. 15. 939-945. Jansen, J.G.. deGrootw A.J.L., van ‘Teijlingen, C.M.M.. Lohman. P.H M . Mohn, G.R.. Vrieling. H.. and van ZEeland. A.A. (1994) Formation and persistance of DNA adducts in pouch skin fibroblasts and liver tissue of rats exposed in vivo to the monofunctional alkylating agents N-methyl-N-nitrosourea or N-ethyl-N-nitrosourea. MDtat. Res.. 307. 95-105. Bronstein. S.M.. Cochrane. J.E.. Craft. T.R.. Swenberg. J.A.. and Skopek. T.R. (1991) Toxicity. mutagenicity. and mutational spectra of N-Ethyl-N-Nitrosourea in human cell lines with different DNA repair phenotypes. Cancer Res.. 51. 5188 - 5197. Grevatt, P.C.. Solomon. 3.3.. and Bhanot. OS. (1992) In vitro mispairggg specificity of OZ-ethylthymidine. Biochemistry. 31. 4181-41 . Samson. L.. Thomale. J.. and Rajewsky. M.F. (1988) Alternative pathways for the in vivo repair of (f9alkylguanine and O4- alkylthymine 'hi Escherichia coli: the adaptive response and nucleotide excision repair. EMBO J.. 7. 2261-2267. Burns, P.A.. Gordon. A.J.E.. Kunsmann. K.. and Glickman. B.W. (1988) Influence of neighboring base sequence on the distribution and repair of N-ethyl-N-nitrosourea-induced lesions in E. coli. Cancer Res.. 48. 4455-4458. Domoradzki. J.. Pegg. A.E.. Dolan. M.E.. Maher. V.M.. and McCormick. 3.3. (1984) Correlation between (f-methylguanine-DNA methyltransferase activity and resistance of human cells to the cytotoxic and mutagenic effect of N-methyl-N’-nitro-N- nitrosoguanidine. Carcinogenesis. vol. 5. no. 12. pp. 1641 - 1647. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 73 Pegg. A.E. (1990) Mammalian Cf—alkylguanine-DNA alkyltransferase: Regulation and importanceiin response to alkylating carcinogenic and therapeutic agents. cancer Res.. 50. 6119-6129. Brent, T.P.. Dolan. M.E.. Fraenkel-Conrat. H.. Hall. J.. Karran. P.. Laval. F.. Margison. G.P.. Montesano. R.. Pegg. A.E.. Potter. P.M.. Singer. 8.. Swenberg. J.A.. and Yarosh. 0.8. (1988) Repair of O- alkylpyrimidines in mammalian cells: a present consensus. Proc. Natl. Acad. Sci. U S'A. 85. 1759 - 1762. Simon. L.. Hazard. R.M.. Maher. V M.. and McCormick. J.J. (1981) Enhanced cell killing and mutagenesis by ethylnitrosourea in xeroderma pigmentosum cells. Carcinogenesis. 2. 567 - 570. Bronstein. S. M6. Skopek. T. R. . and Swenberg J A. (1992) Efficient repair of 0° -ethylguanine. b6ut not 0 -ethylthymine or 02_ ethylthymine. lS dependent upon0°- a-lkylguanine- DNA alkyltransferase and nucleotide excision repair activities in lunmni cells. cancer Res.. 52. 2008-2011. Petinga. R.A.. Andrews, A.O.. Tarone. R.E.. and Robbins, J.H. (1977) Typical xeroderma pigmentosum complementation group A fibroblasts have detectable ultraviolet light-induced unscheduled DNA synthesis. Biochim. Biophys. Acta. 479. 400-410. Dolan. M. E. Moschel. R. C. and Pegg. A..E. (1990) Depletion of mammalian C)° -alkylguanine- DNA alkyltransferase activity by 0° - benzylguanine provides a means to evaluate the role of this protein in protection against carcinogenic and therapeutic alkylating agents. Proc. Natl. Acad. Sci. U S A. vol. 87. pp. 5368 - 5372. Lukash. L.L.. Boldt. J.. Pegg. A.E., Dolan M. E. Maher. V. M. and McCormick. J.J. (1991) Effect of (1° -alkylguanine- DNA alkyltransferase on the frequency and spectrum of mutations induced by N-methyl-N’-nitro-N-nitrosoguanidine in the hprt gene of diploid human fibroblasts. MUtat. Res.. 250. 397 - 409. Maher. V.M.. Domoradzki. J.. Bhattacharyya. N P.. Tsujimura. T.. Corner. R.C.. and McCormick. J.J. (1990) Alkylation damage. DNA Eegair. and mutagenesis in human cells. Mbtat. Res.. 233. 235 - 4 . Wang. Y.. Maher. V M.. Liskay. R M.. and McCormick. J.J. (1988) Carcinogens can induce homologous recombination between duplicated chromosomal sequences in mouse L cells. Mbl. Cell. Biol.. 8. 196- 202. Ryan. P. A Maher. V M . and McCormick. J. J. (1987) Modification of MCDB 110 medium to support prolonged growth and consistant high cloning egficiency of diploid human fibroblasts. Exp. Cell Res. 172. 318- 28. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 74 Yang. J- L. Maher. V. M. McCormick. J J. (1989) Amplification and direct nucleotide sequencing of cDNA from the lysate of low numbers of diploid human cells. Gene 83. 347- 354. Montesano. R. (1981) Alkylation of DNA and tissue specificity in gitrgggmggg carcinogenesis. J. Supramol. Struct. Cell. Biochem.. Saffhill, R.. Margison. G.P.. and O'Connor. P.J. (1985) Mechanisms of carcinogenesis induced by alkylating agents: Biochhn. Biophys. Acta. 823. 111-145. Dosanjh, M K.. Menichini. P. Eritja R. . and Singer. 8. (1993) Both C) - methylthymine and C) -ethylthymine preferentially form alkyl T: G pairs that do not block in vitro replication in a defined sequence. Carcinogenesis. 14. 1915-1919. Bhanot. O S.. Grevatt. P.C., Donahue, J.M.. Gabrielides. C.N.. and Solomon. J.J. (1991) In vitro DNA replication implicates 02- ethyldeoxythymidine in transversion mutagenesis by ethylating agents. Nocleic Acids Res.. 20. 587-594. Dolan. M...E Pegg A...E Hora. MK. and Erickson. LC (1988) Effect of'Cl° -methylguanine on DNA interstrand cross- -link formation by chloroethylnitrosoureas and 2- chloroethyl- (methylsulfonyl) methanesulfonate. Cancer Res.. 48. 3603-3606. Karran. P.. Stephenson. C.. Cairns-Smith. S.. and MacPherson. P. (1990) Regulation of’ Cf-methylguanine-DNA methyltransferase Egpggssion in Burkitt's lymphoma cell line Raji. Mbtat. Res.. 233. Sassanfar. M.. Dosanjh. M.K.. Essigmann. J.M. . and Samson. L. (1991) Relative efficiencies of the bacterial. yeast. and human DNA methyltransferases for the repair of (1° -methylguanine and (J- methylthymine. J. Biol. Chem.. 266. 2767- 2771. Koike. G. Maki. H.. Takeya. H.. Hayakawa. H.. and Sekiguchi M. (1990) Purification. structure. and biochemical properties of human 04-7m6ezthylguanine DNA methyltransferase. J. Biol. Chem” 265 14754- 1 Popp. R.A.. Bailiff. E.G.. Skow. L.C.. Johnson. F M.. and Lewis. S.E. (1983) Analysis of a mouse a-globin gene mutation induced by ethylnitrosourea. Genetics. 105. 157-167. Jansen. J.G.. Mohn, G.R.. Vrieling. H.. van Teijlingen. C.M.M.. Lohman. P H M.. and van Zeeland. A.A. (1994) Molecular analysis of hprt gene mutations in skin fibroblasts of rats exposed in vivo to gimgtgygéN-nitrosourea or N-ethyl-N-nitrosourea. Cancer Res.. 54. - 4 . 43. 44. 45. 46. 47. 48. 49. 50. 75 Friedberg, E. C.. Bardwell. A.J.. Bardwell. L.. Wang. Z.. and Dianov. G. (1994) Transcription and nucleotide excision repair - Eeflecgggnsé considerations and recent biochemical insights. MUtat. es Mellon. I.. Bohr V. A. Smith. C. A.. and Hanawalt. P. C. (1986) Preferential DNA repair of an active gene in human cells Proc. Natl. Acad. Sci. U S.A. 83. 8878- 8882. Mellon. I.. Spivak. G.. and Hanawalt. P.C. (1987) Selective removal of transcription-blocking DNA damage from the transcribed strand of the mammalian DHFR gene. Cell. 51. 241-249. Vrieling. H.. Van Rooijen. M.L.. Groen. N.A.. Zdzienick. M.Z.. Simons. J.W.I.M.. Lohman. P.H.M.. and van Zeeland. A.A. (1989) DNA strand specificity for UV-induced mutations in mammalian cells. Mbl. Cell. Biol.. 9. 1277-1283. Roldna-Arjona. T.. Luque-romero. F.L.. Ariza. R.R.. Jurado. J.. and Pueyo C. (1994) Influence of DNA repair by ada and ogt alkyltransferases on the mutational specificity of alkylating agents. Mbl. Carcinog.. 9. 200-209. Yang. J-L . Hu. M-C.. and Wu. c-w. (1991) Novel mutational spectrum induced by N-methyl-N-nitro-N-nitrosoguanidine in the coding region of the hypoxanthine (guanine) phosphoribosyltransferase gene in diploid human fibroblasts. J..Mbl. Biol.. 221. 421-430. Chambers. R. N.. Sledziewska- Gojska. E.. Hirani- -Hojatti. S.. and Borowy- -Borowski, H. (1985) uvrA and recA mutations inhibit a site- specific transition produced by a single 0° -methylguanine in gene G of bacteriophage ¢X174 Proc. Natl. Acad. Sci. USA 82. 7173- 7177. Rossi. S. C. Conrad. M. Voigt. J. M. and Topal. M. D. (1989) Excision repair