Mutations to azaguanine resistance induced in cultured diploid human fibroblasts by the carcinogen, N-acetoxy-2-acetylaminofluorene

Genetic toxicity of 2-acetylaminofluorene, 2-aminofluorene and some of their metabolites and model metabolites

2-Acetylaminofluorene and 2-aminofluorene are among the most intensively studied of all chemical mutagens and carcinogens. Fundamental research findings concerning the metabolism of 2-acetylaminofluorene to electrophilic derivatives, the interaction of these derivatives with DNA, and the carcinogenic and mutagenic responses that are associated with the resulting DNA damage have formed the foundation upon which much of genetic toxicity testing is based. The parent compounds and their proximate and ultimate mutagenic and carcinogenic derivatives have been evaluated in a variety of prokaryotic and eukaryotic assays for mutagenesis and DNA damage. The reactive derivatives are active in virtually all systems, while 2-acetylaminofluorene and 2-aminofluorene are active in most systems that provide adequate metabolic activation. Knowledge of the structures of the DNA adducts formed by 2-acetylaminofluorene and 2-aminofluorene, the effects of the adducts on DNA conformation and synthesis, adduct distribution in tissues, cells and DNA, and adduct repair have been used to develop hypotheses to understand the genotoxic and carcinogenic effects of these compounds. Molecular analysis of mutations produced in cell-free, bacterial, in vitro mammalian, and intact animal systems have recently been used to extend these hypotheses.

Analysis of large deletions in the HPRT gene of primary human fibroblasts using the polymerase chain reaction

Spontaneous and X-ray-induced mutants of the HPRT gene were isolated from two primary human fibroblast lines. The limited life-span of the mutants restricted the use of methods requiring large quantities of DNA, and the polymerase chain reaction (PCR) was used in particular to check for the presence of multiple genomic sites in mutant analysis. Robust PCR primers were designed to amplify sites of up to 1 kb, mostly with 1-kb spacings between sites, over the entire 56-kb HPRT gene region. Using PCR, large deletions were found in 43% of independent X-ray-induced mutants, and their breakpoints were localized where these fell within the gene. Anonymous DNA sites in the Xq26 chromosomal region containing HPRT (covering > or = 1.5 Mb) were also amplified by PCR to assess codeletion with HPRT; sites up to 1 Mb distal to the gene (DXS86, DXS10) were codeleted in some mutants, but no mutant was found with loss of a proximal site (DXS79).

A role for DNA polymerase in the specificity of nucleotide incorporation opposite N-acetyl-2-aminofluorene adducts

Escherichia coli DNA polymerase I (Klenow fragment), DNA polymerase alpha from both calf thymus and human lymphoma cells and DNA polymerase beta from calf thymus and Novikoff hepatoma cells can incorporate nucleotides opposite N-guanin-8-yl-acetyl-2-aminofluorene in DNA. The polymerases incorporate dCTP opposite some AAF-dG lesions when Mg2+ is the divalent cation. Substitution of Mn2+ for Mg2+ broadens the specificity of insertion: E. coli DNA polymerase I (Klenow fragment) also inserts A, and at specific sites G or T; DNA polymerase alpha inserts any of the four dNTPs with A and C incorporated preferentially to G and T. Polymerase beta is specific, inserting mainly C even in the presence of Mn2+. The Km for addition of dATP opposite a lesion by E. coli polymerase I (Klenow fragment) in the presence of Mn2+ is about 0.5 mM. dNMPs increase the insertion of nucleotides opposite AAF-dG in the presence of Mg2+ and increase both the rate and number of sites at which incorporation occurs in the presence of Mn2+. dNTP alpha S and recA protein increase only the insertion of C. We suppose that the incorporation of dCTP reflects normal base-pairing with the AAF-deoxyguanine in the anti conformation, whereas insertion of the other nucleotides (including some of the C) reflects insertion opposite the AAF adduct in its preferred syn conformation. The fact that the DNA polymerase plays a role in determining the specificity of insertion opposite a lesion terminating DNA synthesis suggests that the spectrum of base substitution mutagenesis seen in vivo may reflect the properties of the protein components, including the polymerase, involved in bypass synthesis.

Two methods to induce 6-thioguanine resistance in human fibroblasts in the presence of rat-liver microsomes

7,12-Dimethylbenz[a]anthracene and 4-aminobiphenyl induced dose-dependent 6-thioguanine-resistant mutations in normal human fibroblasts in the presence of rat S9 microsomes. With each chemical, mutations were induced by two different methods: single treatment of synchronized cells, and multiple treatment of nonsynchronized cells. Multiple treatment of nonsynchronized cells gave higher mutation frequencies at lower doses, and appears to be the more sensitive method.

Genetic effects of N-methyl-N'-nitro-N-nitrosoguanidine and its homologs

Since the discovery of the mutagenic activity of N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) in 1960, this compound has become one of the most widely used chemical mutagens. The present paper gives a survey on the chemistry, metabolism, and mode of interaction of MNNG with DNA and proteins, and of the genotoxic effects of this agent on microorganisms, plants, and animals, including human cells cultured in vitro. Data on the carcinogenicity and teratogenicity of MNNG as well as on the genotoxic effects of homologs of MNNG are also presented.

Use of prolonged treatment and the fluctuation test to detect mutations in human fibroblasts treated with methyl methanesulfonate

The methyl methanesulfonate (MMS) induction of 6-thioguanine-resistant (6TGr) mutants in non-synchronized human fibroblasts has been quantitatively characterized. A 24-h MMS treatment failed to induce mutations in non-synchronized cultures. However, exposure of growing cells to MMS for 3 days caused a significant increase in the mutation frequencies. The mutation induction was dose-dependent. The fluctuation test was also useful in detecting the mutagenic activity of MMS in human fibroblasts. Prolonged treatment of cultures made the human cell mutagenesis testing system more sensitive.

Induction of 6-thioguanine-resistant mutations by rat-liver homogenate (S9)-activated promutagens in human embryonic skin fibroblasts

Most normal human fibroblasts grown in culture do not metabolize promutagens/procarcinogens. Thus screening assays employing normal human fibroblasts have only been successful for direct-acting chemical mutagens and various radiations. In this report we describe a mutation assay (HGPRT locus) employing a normal human embryonic skin fibroblast and a rat-liver homogenate (S9) mixture. 3 model promutagens, benzo[a]pyrene (B[a]P), 3-methylcholanthrene (3MC), and dimethylnitrosamine (DMN) have been utilized in these studies. In addition to discussing conditions for optimizing the response of this assay, our results indicate that at constant amount of S9 protein concentration, there exists a linear correlation between mutagenicity and dose. At 50% survival, the mutant frequencies induced by B[a]P and 3MC (5 micrograms/ml) are 60 and 30 times the background mutant frequency, respectively. Similarly, at 50% survival, DMN (5 mg/ml) induced 6-TGr mutant frequencies are 25-fold over the background frequency. The increase in cytotoxicity resulting from exposure of cells to these 'activated' chemicals is also a linear dose response. At high S9 concentrations a deactivation or detoxification phenomenon occurs. However, the mutagenic efficiency of S9-activated chemicals when plotted as the number of induced mutations versus log survival is unaffected by the deactivating capacity of S9 proteins. This study demonstrates a quantitative mutation assay using an early passage human culture with an exogenous rat-liver microsomal preparation providing activating enzymes.

Induction of 6-thioguanine resistance in synchronized human fibroblast cells treated with methyl methanesulfonate, N-acetoxy-2-acetylaminofluorene and N-methyl-N'-nitro-N-nitrosoguanidine

Chemical induction of 6-thioguanine resistance was studied in synchronized human fibroblast cells. Cells initially grown in a medium lacking arginine and glutamine for 24 h ceased DNA synthesis and failed to enter the S phase. After introduction of complete medium, the cells progressed to the S phase after 16h. DNA synthesis peaked 20 h after removal of nutrient stress and declined. Mutations were induced in S-phase cells by methyl methanesulfonate (MMS), N-acetoxy-2-acetylaminofluorene (NA-AAF) and N-methyl-N'-nitro-N-nitroso-guanidine (MNNG). Chemical treatments resulted in an increase in the absolute number of mutant colonies and in a dose-dependent mutation frequency. In this report, we show that NA-AAF evokes a temporal pattern of mutation in synchronized cells, with such mutations being induced only during the S phase. Evidence indicates that presence of S-phase cells in the treated cultures is a prerequisite for the induction of mutations.

International Commission for Protection against Environmental Mutagens and Carcinogens. ICPEMC working paper 2/5: mutagenesis in mammalian cells

Chemical mutagenesis in animal cells is a complex process. Whereas some chemicals are mutagenic in their original form, others such as the nitrosamines and polycyclic hydrocarbon carcinogens are mutagenic only when enzymatically activated. The active form, or ultimate carcinogen, can interact with proteins and nucleic acids, altering amino acids and producing modified bases in DNA. The modified bases do not usually constitute mutations as produced. Instead they are acted on by the DNA enzymes of the cell, which repair most damaged bases but occasionally insert incorrect base sequences at or near the sites of damage. The frequency at which mutant animal cells are recovered depends upon the selection conditions in culture, upon whether the mutation selected is in a gene present in single or multiple active copies, and upon whether expression is dominant or recessive. Many studies depend on selecting for 8-azaguanine- or 6-thioguanine-resistant mutants, which are due to mutations in the HGPRT locus present in a single active copy on the X-chromosome. Other widely used systems depend on selecting for ouabain resistance, which is dominant and results from a change in the sodium/potassium ATPase activity, or on selecting for thymidine kinase mutants in heterozygous Tk+/Tk- mouse cells. Many other types of mutation including nutritional markers are recessive and express only in cells carrying a single active gene copy, as is sometimes the case in established cell lines. The types of base damage causing mutations have been identified in very few cases only, and little is known about the enzymatic mechanisms of mutagenesis. However, chemical mutagenesis in cultured animal cells provide a practical way of testing chemicals and radiations for mutagenicity directly in animal cells, and much has been learned about the mutagenicity of various carcinogenic substances. To date, there is reasonable qualitative agreement between these results and those obtained in the widely used liver microsome-activated bacterial mutagenesis test systems.

Drug-resistant lymphocytes in man as indicators of somatic cell mutation

Direct in vivo tests of somatic mutation in man may provide realism in assessing the genetic risks of potential environmental mutagens. The autoradiographic determination of purine analogue (8-azaguanine; 6-thioguanine) resistant (AGr; TGr) peripheral blood lymphocytes (PBLs) arising in vivo in man is proposed as a candidate test. PBLs bearing the naturally occurring Lesch-Nyhan (LN) mutation are prototype mutant cells. LN PBLs are AGr and TGr, whereas normal PBLs are AG and TG sensitive. When judged by the inhibition of phytohemagglutinin (PHA) stimulated 3H-thymidine incorporation in vitro, analogue-resistant LN PBLs may be distinguished from analogue-sensitive normal PBLs by several methods. Early studies quantitating PHA stimulation by scintillation spectrometry detected down to 1% of LN PBLs in artificial mixtures with normal PBLs. Although LN heterozygous females could be identified on the basis of lymphocyte mosaicism, scintillation spectrometry was too insensitive to detect rare "LN-like" PBLs in non-LN individuals. Autoradiography, however, detected rare TGr PBLs in normal non-LN individuals. Their frequencies did not increase with age. With this method, TGr PBL frequencies in LN heterozygous females were found to range from 1 x 10(-3) to 5 x 10(-2), whereas blood samples from LN males showed from 23% to 100% TGr cells. Rare LN PBLs could be detected in artificial mixtures with normal cells. Studies in human patients undergoing various potential mutagenic therapies assessed the effects of these therapies on the TGr PBL variant frequencies (Vf) of non-LN individuals. Group TGr PBL Vf values were higher in treated patient groups than in controls. However, some untreated patient groups (cancer and psoriasis) also had elevated values, suggesting that disease itself may affect TGr PBL frequencies. Nonetheless, one patient group (vitiligo) showed elevated Vf values in treated (8-methoxypsoralen and long-range UV light = PUVA) but not in untreated patients, suggesting that treatment was responsible for the TGr PBL elevations. Longitudinal studies over time in cancer patients receiving X-irradiation therapy demonstrated that such exposures also are associated with TGr PBL frequency rises and suggested that longitudinal studies may be necessary to relate TGr PBL Vf elevations to specific environmental influences. Variant TGr PBLs were found at frequencies comparable to those in man in the peripheral blood of rats. They increased in a single study following treatment of the animals with a clinical alkylating agent. Characterization of the TGr PBLs suggests that some of these cells are mutants. Presumably the mutant cells arise in vivo by somatic cell mutation.

Determination of the expression time and the dose--response relationship for mutations at the HGPRT (hypoxanthine-guanine-phosphophoribosyl transferase) locus induced by X-irradiation in human diploid skin fibroblasts

The length of the expression time for mutants resistant to 8-azaguanine or 6-thioguanine induced by X-rays was determined in human diploid skin fibroblasts. The cells were seeded in the selective medium over a period of 14 days after treatment. Direct expression of at least a part of the mutants was observed at day 0, and an increase of the mutant frequency over the entire cultivation period appeared to be due to spontaneous mutation. The dose-response relationship does not appear to deviate from linearity. The mutation rate per R had a mean value of 2.1 x 10(-7) which is about twice the value of the mutation rate found in rodent cells for the same locus.

A comparison of the DNA binding, cytotoxicity and repair synthesis induced in human fibroblasts by reactive derivatives of aromatic amide carcinogens

The cytotoxicity of three structurally-related direct-acting carcinogens, N-acetoxy-2-acetylaminofluorene, N-acetoxy-2-acetylaminophenanthrene and N-acetoxy-4-acetylaminobiphenyl, was compared in normal cells and in excision repair deficient xeroderma pigmentosum cells (XP12BE). All three proved significantly more cytotoxic to the XP cells than to the normal cells. At equicytoxic levels, substantially more residues were initially bound to the DNA of the normal cells than to the XP cells, suggesting that the former are able to remove a large percentage of the DNA bound residues before these can result in cell death. The ability of these cell strains to remove bound residues from DNA, to incorporate thymidine into parental strands of DNA during repair replication, and to recover from potentially lethal damage if held in the non-replicating, density-inhibited G0 state was compared as a function of dose and time. The XP12BE cells proved virtually incapable of excision repair of DNA damage induced by these carcinogens and of recovery. In contrast, normal cells recovered from the potentially lethal effects of these three compounds and did so at a rate comparable to their rate of removal of bound residues and of repair synthesis. In the excision-deficient XP12BE cells, DNA adducts induced by N-acetoxy-2-acetylaminophenanthrene proved 3- to 6-fold more cytotoxic than adducts induced by the other two carcinogens.

6-Thioguanine resistant peripheral blood lymphocytes in humans following psoralen, long-wave ultraviolet light (PUVA) therapy

A recently described method that enumerates variant 6-thioguanine resistant peripheral blood lymphocytes present in vivo in man as a potential marker of somatic cell mutations occurring in vivo was used to study 18 psoriatic patients receiving PUVA therapy, 16 conventinally treated psoriatic patients, 10 vitiligo patients receiving PUVA therapy and 7 untreated individuals with vitiligo. Variant lymphocyte frequencies determined for these individuals were compared with those determined for groups of 10 concurrent and 63 cumulative healthy control individuals. Variant frequencies were elevated in psoriatic and vitiligo patients receiving PUVA therapy and in conventionally-treated psoriatic patients. They were not elevated over control values in untreated vitiligo patients.

Quantification of chemical mutagenesis in diploid human fibroblasts: induction of azaguanine-resistant mutants by N-methyl-N'-nitro-N-nitrosoguanidine

The toxic and mutagenic effects of N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) on cultured diploid human fibroblasts were studied. When 10(5) cells per 60 mm dish were exposed to MNNG for 4 h in Ham's medium F10 containing 0.02 M HEPES buffer at pH 6.8, MNNG concentrations of less than 1 X 10(-6) M resulted in cell survivals near 100%, while the average survival was less than one percent at concentrations greater than 5 X 10(-6) M. After treatment with MNNG, cells were subjected to selection using optimal conditions for the detection of diploid human fibroblasts that are resistant to the guanine-analogs AG and TG because they contain altered or deficient HPRT. The induced mutant frequency was maximized by allowing a 5 to 7 day post-treatment interval for the expression of the mutant phenotype and by replating the cells at the beginning of selection at a population density of less than 450 cells per cm2. Careful attention was given to counting statistically adequate numbers of mutants and to accurately determining cell survival and replating cloning efficiencies. Independent dose-response experiments gave induced mutant frequencies as high as 7.0 X 10(-4) to 8.8 X 10(-4) mutants per viable cell at about 5% survival, compared to a spontaneous mutation rate of 3.7 X 10(-6) to 7.2 X 10(-6) mutants per cell generation. The AGr mutants observed after treatment with MNNG were phenotypically stable and closely resembled prototype AGr cultures derived from humans who have inherited mutant X-chromosomal alleles for HPRT.

Mutagenesis at the ouabain-resistance locus in human diploid fibroblasts

The variables affecting the frequency of ouabain-resistant mutant clones have been studied in a strain of foetal lung fibroblasts. Optimum mutant recovery was obtained when cells were selected in 10(-6) M ouabain at a cell density of 2 X 10(4) cells/cm 2 (10(6) cell per 100-mm dish). The spontaneous mutation rate was estimated to be 4 X 10(-8) per cell generation. Treatment with the mutagens ethyl methanesulfonate (EMS), N-methyl-N' -nitro-N-nitrosoguanidine, and UV light increased the frequency of mutant colonies by an order of magnitude. The maximum number of mutants after mutagenesis with EMS occurred after two population doublings of growth in non-selective medium prior to selection and depended on the dose of EMS. Ouabain-resistance is a useful marker for studies of quantitative mutagenesis in human cells.

Effect of DNA repair on the cytotoxicity and mutagenicity of polycyclic hydrocarbon derivatives in normal and xeroderma pigmentosum human fibroblasts

The cytotoxicity of the "K-region" epoxides as well as several other reactive metabolites or chemical derivatives of polycyclic hydrocarbons was compared in normally-repairing human diploid skin fibroblasts and in fibroblasts from a classical xeroderma pigmentosum (XP) patient (XP2BE) whose cells have been shown to carry out excision repair of damage induced in DNA by ultraviolet (UV) radiation at a rate approx. 20% that of normal cells. Each compound tested exhibited a 2- to 3-fold greater cytotoxicity in this XP strain than in the normal strain. To determine whether this difference in survival reflected a difference in the capacity of the strains to repair DNA damage caused by such hydrocarbon derivatives, we compared the cytotoxic effect of several "K-region" epoxides in two additional XP strains, each with a different capacity for repair of UV damage. The ratio of the slopes of the survival curves for each of the XP strains to that of the normal strain, following exposure to each epoxide, was very similar to that which we had previously determined for their respective UV curves, suggesting that human cells repair damage induced in DNA by exposure to hydrocarbon derivatives with the same system used for UV-induced lesions. To determine whether the deficiency in rate of excision repair in this classical XP strain (XP2BE) causes such cells to be abnormally susceptible to mutations induced by "K-region" epoxides of polycyclic hydrocarbons, we compared them with normal cells for the frequency of induced mutations to 8-azaguanine resistance. The XP cells were two to three times more susceptible to mutations induced by the "K-region" epoxide of benzo(a)pyrene (BP), 7,12-dimethyl-benz(a)anthracene (DMBA), and dibenz(a,h)anthracene (DBA). Evidence also was obtained that cells from an XP variant patient are abnormally susceptible to mutations induced by hydrocarbon epoxides and, as is the case following exposure to UV, are abnormally slow in converting low molecular weight DNA, synthesized from a template following exposure to hydrocarbon epoxides, into large-size DNA.

Liver homogenate-mediated mutagenesis in chinese hamster V79 cells by polycyclic aromatic hydrocarbons and aflatoxins

Several chemical carcinogens that require metabolic activation have been examined for their cytotoxic and mutagenic activity in Chinese hamster V79 cells. Mutagenic activity was measured as the induced frequency of 6-thioguanine-resistant colonies. Metabolic activation was provided by the 9000 g supernatant fraction of rat liver plus cofactors. The cytotoxicity and mutagenicity of aflatoxin B1 and B2, benzo(a)pyrene, 3-methylcholanthrene, 7,12-dimethylbenz(a)anthracene, dibenz(a,h)-anthracene, dibenz(a,c)anthracene, and benz(a)anthracene were examined as functions of concentration. Except for the two isomers of dibenzanthracene, the mutagenic activity in general paralleled the carcinogenic activity. An assay of this type may be useful as a prescreen for environmental chemicals that require metabolic activation.

The fluctuation test as a more sensitive system for determining induced mutation in L5178Y mouse lymphoma cells

Luria-Delbrück fluctuation tests for the determination of the spontaneous mutation rate to ouabain resistance incultured L5178Y mouse lymphoma cells gave values in the range 0.44 to 1.03 X 10(-7) mutants per cell per generation. Addition of very low, non-toxic, levels of EMS (0.1 mM) and MMS (0.012 mM) 30-48 h before plating in selective medium gave a highly significant increase in the number of ouabain-resistant mutants. Methods for the calculation of spontaneous and induced mutation rates are discussed and a method for the computation of induced mutation rates is described. It is suggested that the modified fluctuation test is a simpler and considerably more sensitive assay for mutagens than the conventional experimental design. Some of its limitations are discussed.

X-ray-induced mutation to 6-thioguanine resistance in cultured human diploid fibroblasts

X-ray induced mutation to 6-thioguanine (6TG)-resistance was studied in early passage cultures of human diploid fibroblasts. The appearance of phenotypic induced mutants in irradiated cell populations was linearly related to the number of post-irradiation cell doublings and to the duration of the growth period prior to mutant selection; the maximum yield of X-ray induced mutants was observed when cells surviving radiation had completed 3--4 douplings (6--7 days growth) in non-selective medium. The maximum induced mutation frequency was linearly related to X-ray dose and the mutation rate was estimated to be 3.1-10(-7) mutations per viable cell per rad. The data obtained for X-ray induced mutations in cultured human diploid fibroblasts were compared with (a) similar experimental data obtained with established cell cultures and (b) with theoretical predictions of X-ray mutation rates in human germ cells.

Mutation of human lymphoblasts by methylnitrosourea

The lag in phenotype expression of methylnitrosourea(MNU)-induced mutation to 6-thioguanine (6TG) resistance has been studied in a diploid human lymphoblastoid cell line. We find that a considerable period (8-12 days) elapses before new mutants appear in treated cultures; after 2 weeks, however, a stable maximum fraction is attained, as would be expected for a genetic mutation. We present preliminary data linking this phenotypic lag to the slow degradation rate of hypoxanthine-guanine phosphoribosyl transferase (HGPRT) and to an apparent requirement for very low (less than 0.2% normal) cellular HGPRT content in order for cells to be resistant to 10 mug 6TG/ml. A series of reconstruction experiments are presented, the results of which support the conclusion that selective pressures in the assay procedure do not bias the quantitative estimates of induced mutant fraction.

Concentration-dependent mutation of diploid human lymphoblasts by methylnitronitrosoguanidine: the importance of phenotypic lag

The mutation of diploid human lymphoblasts by methylnitronitrosoguanidine (MNNG) was measured over the range of 0--45 ng of MNNG/ml of of medium. We found a 12-day lag in the phenotypic expression of 6-thioguanine resistance; the occurrence of this lag was independent of MNNG concentration. We hypothesize that the unexpectedly long lag period reflects a requirement for the loss of previously existing molecules of the enzyme hypoxanthine-guanine phosphoribosyltransferase (HGPRT) after mutation at the HGPRT locus.

The isolation and preliminary characterisation of 6-thioguanine-resistant mutants of human diploid fibroblasts

Mutant clones of human diploid fibroblasts deficient in the enzyme, hypoxanthine-guanine phosphoribosyl transferase (HGPRT) were selected by their ability to grow in medium containing the cytotoxic purine analogue, 6-thioguanine (6TG). The optimal condtions for mutant selection were 6TG concentrations between 1 and 5 mug ml1 and cell plating densities approximately 10(3) cells cm-2. Nine spontaneous and four radiation-induced 6TG-resistant mutants had less than 2% of the parental strain HGPRT activity and were unable to grow in medium containing azaserine. These mutants were phenotypically stable during greater than 25 population doublings in non-selective medium and five mutants that were examined showed no gross change from the normal human karyotype. Evidence is presented to show that 6TG is a better selective agent than 8-aza-guanine (8AG) for HGPRT-deficient mutants of human diploid fibroblasts.

Ethyl methanesulphonate mutagenesis with L5178Y mouse lymphoma cells: a comparison of ouabain, thioguanine and excess thymidine resistance

Complete inhibition of growth of sensitive L5178Y mouse lymphoma cells in culture was obtained with 10(-3)M ouabain, 1.65 X 10(-3)M thymidine, 1.8 X 10(-4)M thioguanine and 10(-6)M cytosine arabinoside. The toxicity of methotrexate was dependent upon cell density and this compound was excluded from further study. The expression time before addition of the selective agent was important for detecting EMS induced resistant variants. Ouabain-resistant variants appeared immediately after treatment and were present over a broad time span. No excess thymidine- or thioguanine-resistant variants were seen initially; a peak in variant numbers was seen for excess thymidine resistance at 48-96 h and for thioguanine resistance at 144-192 h. Using induced mutation frequencies at optimum expression times, equal EMS treatments yielded substantially more variants resistant to thioguanine than to ouabain. It is suggested that this difference may have origin in possible constraints in the classes of mutants which are permissible in a vital function, maintenance of the Na+/K+ balance, when compared with a non-vital function, salvage purine biosynthesis. Some data are presented on the stability in culture of resistant variants. A limited number of observations were made following treatment in the peritoneal cavity of the mouse which were in broad agreement with the above results.