Detection of 1,N6-ethenoadenine in rat urine after chloroethylene oxide exposure

The four etheno adducts of vinyl chloride formed in DNA, 1,N6- ethenoadenine (epsilonA), 3,N4-ethenocytosine, 1,N2-ethenoguanine and N2,3-ethenoguanine were previously reported to be released from DNA by a family of enzymes in the base-excision repair pathway (Dosanjh et al., Proc. Natl Acad. Sci. USA, 91, 1024-1028, 1994; Hang et al., Carcinogenesis, 17, 155-157, 1996; Hang et al., Proc. Natl Acad. Sci. USA, 94, 12869-12874, 1997). Adducts excised from DNA by glycosylases are usually excreted in urine and have been reported to be potential biomarkers of DNA damage in exposed individuals. In this study, we report the detection of epsilonA in the urine of rats exposed to chloroethylene oxide (CEO) using immunoaffinity columns made with specific monoclonal antibodies for enrichment, followed by quantitation by HPLC with fluorescence detection. Chemical analysis of urine samples revealed the presence of a compound chromatographically identical to authentic epsilonA standard. This compound was confirmed by mass spectral analysis. EpsilonA was present in urine of control and CEO- treated rats, with the latter having up to 50-fold greater amounts. The cumulative excretion of epsilonA reached a plateau between 24 and 48 h post-exposure. While it is clear that CEO treatment results in increased excretion of epsilonA, the exact source of the adduct is unknown. When rats were administered epsilonA i.v., approximately 10% of the administered dose was excreted in urine. This research demonstrates that urinary excretion of epsilonA may be a potential biomarker for in vivo alkylation of DNA and nucleotide pools.      

Decreased repair activities of 1,N(6)-ethenoadenine and 3,N(4)-ethenocytosine in lung adenocarcinoma patients

To assess the role of oxidative stress and lipid peroxidation (LPO) in the pathogenesis of lung cancer, we measured the levels of 1,N(6)-ethenoadenine (epsilonA) and 3,N(4)-ethenocytosine (epsilonC) in the DNA by immunoaffinity/(32)P postlabeling (33 cases). We also measured the capacity for epsilonA and epsilonC repair (by the nicking assay) in normal and tumor lung tissues, as well as in blood leukocytes of lung cancer patients (56 cases). Repair activities for epsilonA and epsilonC were also assayed in leukocytes of healthy volunteers, matched with cancer patients for age, sex, and smoking habit (25 individuals). Up to 10-fold variations among individuals were observed both in adducts level and repair activities. No differences in epsilonA and epsilonC levels between tumor and nonaffected lung tissues were recorded. However, leukocytes accumulated a significantly higher number of DNA adducts than the lung tissues. Repair activities for both epsilonA and epsilonC were significantly higher in tumor than in normal lung tissue. No significant differences in epsilonA and epsilonC repair activities were associated with age, sex, or smoking habit. However, a significant difference in repair capacity was observed between two histological types of lung cancer, squamous cell carcinoma (SQ) and adenocarcinoma (AD). In individuals suffering from lung AD, epsilonA- and epsilonC-repair activities in normal lung and blood leukocytes were significantly lower than in SQ patients. Moreover, in nonaffected lung tissue of AD patients, the ratio epsilonA/epsilonC adducts was lower than in SQ patients. Differences have also been found between epsilonA and epsilonC repair activities of cancer patients and healthy volunteers. Repair capacity for epsilonA was significantly lower in blood leukocytes of lung cancer patients than in leukocytes of healthy volunteers (P = 0.012). This difference was even larger between healthy volunteers and patients developing inflammation-related AD (P = 0.00033). Repair activities for epsilonC were the same in leukocytes of healthy controls, all lung cancer patients, and SQ patients. However, individuals with ADs revealed significantly lower epsilonC-repair activity (P = 0.013). These results suggest that oxidative stress-mediated lipid peroxidation might contribute to induction and/or progression of lung cancer. Decreased activity of base excision repair pathway for epsilonA and epsilonC is associated particularly with inflammation-related lung AD.     

Influence of DNA structure on hypoxanthine and 1,N(6)-ethenoadenine removal by murine 3-methyladenine DNA glycosylase

3-Methyladenine DNA glycosylases initiate base excision repair by flipping the nucleotide bearing the target base out of double-stranded DNA into an active site pocket for glycosylic bond cleavage and base release. Substrate bases for the murine 3-methyladenine DNA glycosylase (other than 3-methyladenine) include hypoxanthine and 1,N(6)-ethenoadenine, two mutagenic adducts formed by both endogenous and exogenous agents. Using double-stranded DNA oligonucleotides containing damaged bases at specific sites, we studied the relative removal rates for these two adducts when located in different sequence contexts. One of the sequence contexts was an A:T tract, chosen because DNA secondary structure is known to change along the length of this tract, due to a progressive narrowing of the minor groove. Here we report that removal rates for hypoxanthine, but not for 1,N(6)-ethenoadenine, are dramatically affected by its location within the A:T tract. In addition, the removal rates of hypoxanthine and 1,N(6)-ethenoadenine when paired opposite thymine or cytosine were examined, and in each sequence context hypoxanthine removal decreased by at least 20-fold when paired opposite cytosine versus thymine. In contrast, 1, N(6)-ethenoadenine removal was unaffected by the identity of the opposing pyrimidine. We conclude that the removal of certain bases by the mouse 3-methyladenine DNA glycosylase can be modulated by both adjacent and opposing sequence contexts. The influence of DNA sequence context upon DNA repair rates, such as those described here, may contribute to the creation of mutational hot spots in mammalian cells.     

Increased formation and persistence of 1,N(6)-ethenoadenine in DNA is not associated with higher susceptibility to carcinogenesis in alkylpurine-DNA-N-glycosylase knockout mice treated with vinyl carbamate

Ethenobases are promutagenic DNA adducts formed by some environmental carcinogens and products of endogenous lipid peroxidation. Mutation spectra in tumors induced in mice by urethane or its metabolite vinyl carbamate (Vcar) are compatible with 1,N(6)-ethenoadenine (epsilonA) being an initiating lesion in the development of these tumors. As alkylpurine-DNA-N-glycosylase (APNG) releases epsilonA from DNA in vitro, wild-type and APNG-/- C57Bl/6J mice were treated with Vcar and levels of epsilonA and 3,N(4)-ethenocytosine (epsilonC), which is not a substrate of APNG, were analyzed in liver and lung DNA. At 6 h after the last dose, levels of epsilonA were 1.6-fold higher in DNA from APNG-/- mice and subsequently persisted at higher levels for longer than in DNA from wild-type animals, confirming that epsilonA is released by APNG in vivo. In contrast, approximately 14-fold lower levels of epsilonC were induced by Vcar, and the kinetics of formation and persistence of epsilonC was similar in the two mouse strains. The carcinogenicity of Vcar was compared in APNG-/- and wild-type suckling mice given a single dose of Vcar (30 or 150 nmol/g). After 1 year, only mice in the high-dose group developed hepatocellular carcinoma; however, the incidence was not higher in APNG-/- mice. Although higher levels and increased persistence of epsilonA was observed in hepatic DNA from APNG-/- mice at 150 nmol/g Vcar, apoptosis and cell proliferation levels were similar in both strains of mice. This may explain why differences in epsilonA formation/persistence observed here did not result in higher susceptibility of APNG-/- mice to hepatocarcinogenesis.     

Phase I trial and pharmacokinetics of trimelamol (N2,N4,N6-trihydroxymethyl-N2,N4,N6-trimethylmelamine)

Trimelamol is an analogue of hexamethylmelamine and pentamethylmelamine which does not require metabolic activation and is sufficiently soluble to allow parenteral administration. A Phase I trial has been performed at the Royal Marsden Hospital in which two schedules of administration have been evaluated, a single i.v. infusion repeated every 3 weeks and 3 daily doses repeated every 3 weeks. Pharmacokinetic analysis was performed at all dose levels on both schedules and a linear correlation was demonstrated between dose and area under the curve. Myelosuppression was dose limiting for single dose administration with a maximum tolerated dose of 2400 mg/m2. Median leukocyte nadirs at 1800, 2100, and 2400 mg/m2 were 3.2, 2.6, and 1.5 x 10(9)/liter. Thrombocytopenia and anemia also occurred but were not dose limiting. Doses greater than 1500 mg/m2 caused WHO grade 3 nausea and vomiting but no acute sedation. Three day administration appeared to be less myelosuppressive, giving a maximum tolerated dose of 1000 mg/m2. Median leukocyte nadirs at 800, 900, and 1000 mg/m2 daily for 3 days were 3.0, 2.3, and 1.5 x 10(9)/liter. Nonhematological toxicities were also less marked on the fractionated schedule. Antitumor effects were observed including 1 complete and 9 partial responses. Demonstration of activity in ovarian cancer has led to further evaluation in this disease using the 3-day schedule at a dose of 800 mg/m2 daily for 3 days.     

Evaluation of the mutagenicity of 1,N6-ethenoadenine- and 3,N4-ethenocytosine-nucleosides in Salmonella typhimurium

To provide supporting evidence for the hypothetical involvement of etheno-derivative formation in DNA in chloroacetaldehyde-mediated mutagenesis, etheno nucleosides were examined for their direct-acting mutagenicity in Salmonella typhimurium strain TA100. The results, however, have shown that 1,N6-ethenoadenosine, 1,N6-ethenodeoxyadenosine, 3,N4-ethenocytidine and 3,N4-ethenodeoxycytidine lack mutagenicity in this test system. A lesson learned in this study is that 3,N4-ethenocytosine nucleosides prepared synthetically or obtained from commercial sources can give false positive mutagenicity due to mutagenic contaminants.     

Sequence context is an important determinant in the mutagenic potential of 1,N6-ethenodeoxyadenosine (epsilonA): formation of epsilonA basepairs and elongation in defined templates

Many laboratories have obtained data on mutagenicity of modified bases in naturally occurring DNA sequences. It has often been noted that mutation is favored in certain sequence contexts, sometimes termed 'hot spots'. This approach to the contribution of neighboring sequences does not permit a systematic study of both the qualitative and quantitative mutational frequencies. In the present experiments we have chosen to use the exocyclic adduct, 1,N6-etheno A (epsilonA), site-specifically placed in a defined 25-mer oligonucleotides in which epsilonA is flanked by differing 5' and 3' tandem bases. Mutation was assessed using an in vitro replication assay and five polymerases of varying fidelity. The relevant central sequences were 3' --> 5' -CC-epsilonA-CC- , -GG-epsilonA-GG-, -TT-epsilonA-TT-, -AA-epsilonA-AA-, -GG-epsilonA-TT- , -TT-epsilonA-AA-, -AT-epsilonA-TT- and -TA-epsilonA-TA-. Using the Klenow fragment (Kf) (exo+ or exo-) of E. coli Pol I, it was found the epsilonA is an ambiguous base and, with varying efficiencies, all four dNTPs could be inserted opposite epsilonA in all sequences. However, only 3' --> 5' -TT-epsilonA-TT-, -GG-epsilonA-TT- and -AT-epsilonA-TT- were fully extended to a significant extent. The only sequences essentially blocked at the position of epsilonA were -AA-epsilonA-AA- and -TT-epsilonA-AA-. The others were intermediate. When replication was performed with Sequenase, MMLV RT or HIV RT, different patterns were observed, in which replication terminated one base prior to epsilonA, at epsilonA, or one base after epsilonA without further extension. In favored sequences, using the Klenow fragment, an epsilonA x N pair could be extended to form normal basepairs. No extension could be demonstrated in sequences in which tandem adenines were 5' to epsilonA. Kinetic data showed that two of the epsilonA x N pairs, epsilonA x A and epsilonA x C, could form at 10 microM or less dNTP. Which bases were preferentially inserted opposite epsilonA was a function of the flanking bases. Under the kinetic conditions used, epsilonA x T did not form even at 1 mM dTTP. These results indicate that the chemical structure of an adduct is not the only determinant of mutagenic efficiency. It is likely that the effect of the adduct on replication is due to the changes in the structural environment conferred by the flanking bases.      

Detection of 1,N6-etheno-2'-deoxyadenosine and 3,N4-etheno-2'-deoxycytidine occurring endogenously in DNA

1,N6-Etheno-2'-deoxyadenosine (epsilon dA) and 3,N4-etheno-2'-deoxycytidine (epsilon dC) are DNA adducts formed by a number of genotoxic chemicals, including vinyl chloride. They are also formed endogenously in tissue DNA, probably from a reactive metabolite of lipid peroxidation. Both the qualitative and quantitative detection of endogenous adducts is important in order to place adduct formation by chemicals such as vinyl chloride in the context of this natural background level. Methods with sufficient sensitivity are therefore being developed to measure the natural background of epsilon dA and epsilon dC adducts. We have developed a high-performance liquid chromatography (HPLC)-32P-postlabelling method to measure epsilon dA and epsilon dC at alkylation frequencies of 1 adduct in 10(7)-10(8) nucleotides in 10-microgram samples of DNA. In HPLC-32P-postlabelling analysis of liver DNA from control Wistar rats, epsilon dA and epsilon dC were determined at levels of 1 adduct in 8.1 x 10(7) and 1 adduct in 1.8 x 10(7) nucleotides, respectively. The levels of epsilon dA and epsilon dC measured in liver DNA of animals exposed orally to five daily doses of 50 mg/kg body weight vinyl chloride were found by this method to be 1 adduct in 2.9 x 10(7) and 1 adduct in 1.4 x 10(7) nucleotides, respectively. In contrast, in a direct labelling study, radiolabelled epsilon dA and epsilon dC were not detected in liver DNA of rats exposed for 6 h by nose-only inhalation to [1,2-14C]vinyl chloride at up to 45 ppm v/v. Immunochemical procedures are also being developed for recognizing etheno adducts. Thus, a monoclonal antibody raised to protein conjugates of epsilon dC showed high selectivity in the recognition of this DNA adduct. When the antibody was immobilized on a solid support and used in an immunoenrichment procedure to purify epsilon dC from a large excess of normal nucleotides, one epsilon dC adduct from about 10(8) normal nucleotides could be resolved. Coupling the immunoaffinity enrichment procedure with capillary zone electrophoresis permitted the detection of approximately one epsilon dC adduct in 3 x 10(6) nucleotides.     

SHORT COMMUNICATION: 1, N6-Ethenoadenine and 3, N4-ethenocytosine are excised by separate human DNA glycosylases

We previously reported our finding that human cells containglycosylase activity toward all four etheno bases formed inDNA by chloroacetaldehyde and related bi-functional aldehydes.By enzyme purification, including FPLC, we isolated two separateglycosylase activities for 1,N⁶-ethenoadenine (     

Preclinical toxicology, pharmacokinetics and formulation of N2,N4,N6-trihydroxymethyl-N2,N4,N6-trimethylmelamine (trimelamol), a water-soluble cytotoxic s-triazine which does not require metabolic activation

N2,N4,N6-Trihydroxymethyl-N2,N4,N6-trimethylmelamine (Trimelamol) is a water-soluble synthetic s-triazine which, unlike hexamethylmelamine (HMM) and pentamethylmelamine (PMM), does not require metabolic activation. The physico-chemical characteristics of Trimelamol were studied with the aim of overcoming the problems of chemical instability, low solubility and polymerisation which had hindered the development of the drug for clinical use. Trimelamol had similar activity to PMM against the murine PC6 plasmacytoma, but enhanced activity with respect to PMM against the Walker 256 carcinosarcoma in the rat, a species which metabolizes PMM less efficiently. Pharmacokinetic studies in mouse, rat and man did not show the major species differences characteristic of PMM. The drug exhibited similar toxicity to PMM against rodents, but had virtually no neurotoxicity. The potential advantages of Trimelamol over previously tested melamines are discussed.     

Low central nervous system penetration of N2,N4,N6,-trihydroxymethyl-N2,N4,N6,-trimethylmelamine (Trimelamol): a cytotoxic s-triazine with reduced neurotoxicity

Trimelamol (N2,N4,N6-trihydroxymethyl-N2,N4,N6-trimethylmelamine) is an analogue of pentamethylmelamine (PMM). In early clinical trials PMM failed to show significant anti-tumour activity in man which was attributed to poor metabolic activation. Trimelamol does not require activation and is therefore expected to be more active in man. PMM caused dose-limiting emesis and sedation whereas Trimelamol is much less neurotoxic in rodents. The relative penetration of PMM and Trimelamol into mouse brain has therefore been examined. Mice receiving PMM at 90 mg kg-1 i.p. were found to have high concentrations of the drug in the CNS compared to plasma (mean brain/plasma ratio 1.04) whereas animals receiving Trimelamol had consistently low CNS concentrations (mean brain/plasma ratio 0.08). This difference did not correlate with plasma protein binding which is greater for PMM (68.2%) than for Trimelamol (17.5%). However, it does appear to be related to lipophilicity. In Phase I clinical trial Trimelamol has proved much less emetic than PMM and causes no acute sedation. It is likely that this reduction in toxicity may be explained by the relatively poor ability of Trimelamol to enter the CNS.     

Mutagenesis induced by a single 1,N6-ethenodeoxyadenosine adduct in human cells

To study the genotoxic properties of 1,N6-ethenodeoxyadenosine (epsilondA) in human cells, a novel site-specific mutagenesis approach was developed, in which a single DNA adduct was uniquely placed in either strand of a shuttle plasmid vector. The analysis of progeny plasmid derived from the modified strand shows that epsilondA, when incorporated into the position of the second A of 5'-CAA (codon 61 of the ras gene), is mutagenic in human cells, inducing A-->T, A-->G, and A-->C mutations. The efficient induction of A-->T transversions in experiments using modified double- and singlestranded DNA substrates supports the hypothesis that A:T-->T:A transversions in human and animal tumors induced by vinyl compounds reflect misinsertion of dAMP opposite this adduct. Mutagenic events were similar when the adduct was incorporated into either the leading or the lagging strand. EpsilondA was more mutagenic than 8-oxodeoxyguanosine, which induced targeted G-->T transversions in HeLa cells. In Escherichia coli, epsilondA did not significantly miscode (<0.27%) even in the presence of induced SOS functions.     

1,N6-Ethenodeoxyadenosine and 3,N4-ethenodeoxycytidine in liver DNA from humans and untreated rodents detected by immunoaffinity/32P-postlabelling

The etheno-bridged exocyclic DNA adducts 1,N⁶-ethenodeoxyadenosine(     

Detection of 1,N6-propanodeoxyadenosine in acrolein-modified polydeoxyadenylic acid and DNA by 32P postlabeling

The interaction of acrolein, an alpha,beta-unsaturated aldehyde, with polydeoxyadenylic acid and DNA has been investigated using 32P-postlabeling analysis. In preliminary experiments, polydeoxyadenylic acid was incubated with excess acrolein and then digested to 3' monophosphates prior to transfer of 32P from [gamma-32P]ATP with T4 polynucleotide kinase. The 3',5'-bisphosphates were 3'-dephosphorylated prior to two-dimensional thin layer chromatography on polyethyleneimine-cellulose layers. Autoradiography provided evidence for the formation of one extra spot of radioactivity, compared to the control. To determine the adduct structure, deoxyadenosine-5'-monophosphate was incubated with a 3-fold excess of acrolein. This material was mixed with a 32P-labeled digest of acrolein-polydeoxyadenylic acid, and the sample was analyzed by ion-pair high performance liquid chromatography. The spot of 32P observed by thin layer chromatography co-eluted with the major product of the acrolein nucleotide reaction mixture, which was purified by ion-pair high performance liquid chromatography. Two-dimensional nuclear magnetic resonance spectroscopy and mass spectrometry showed the adduct to be 3-(2'-deoxyribosyl-5'-monophosphatyl)-7,8,9-trihydro-9-hydro xy- pyrimido[2,3-i]purine(1,N6-propanodeoxyadenosine-5'-monophosphate) . High performance liquid chromatography was used to fractionate digests of acrolein-modified DNA prior to detection of this exocyclic adduct by 32P-postlabeling.     

Cytotoxic and mutagenic effects of 1-nitropyrene and 1-nitrosopyrene in diploid human fibroblasts

The cytotoxic and mutagenic effects of 1-nitropyrene (1-NP)and its reduced metabolite 1-nitrosopyrene (1-NOP) were determinedin diploid human fibroblasts. Conditions for the metabolic activationof the parent compound (1-NP) by human cells in culture weredeveloped. The cytotoxic effect of 1-NP in normal cells wascompared with that for repairdeficient xeroderma pigmentosum(XP) cells, and cells from a patient with hereditary cutaneousmalignant melanoma (HCMM), which we have shown earlier are abnormallysensitive to 4-nitroquinoline-1-oxide. The slope of the survivalcurve for XP cells was 2.5 times steeper than that of normalcells; that of HCMM cells was intermediate. When these cellswere exposed to 1-NOP, the slope of the survival curve for theXP cells was also 2.5 times steeper than normal but the HCMMcells showed a normal response, suggesting that their defectis not in repair of DNA adducts, but in activation. 1-NP and1-NOP also proved to be mutagenic in the human cell assay. Whencompared on the basis of concentration, 1-NOP was much moremutagenic than 1-NP. But when the compounds were compared onthe basis of equal cell killing or equal number of DNA adductsinitially bound to DNA, they were very similar. An equal numberof residues covalently bound to DNA caused approximately thesame amount of cell killing for either compound. XP cells werekilled by a 7-fold lower number of bound adducts, suggestingthat the increased survival and decreased mutation inductionin the normal cells reflects their ability to remove potentiallycytotoxic and mutagenic lesions.     

A comparison of two ultrasensitive methods for measuring 1,N6-etheno-2'- deoxyadenosine and 3,N4-etheno-2'-deoxycytidine in cellular DNA

1,N⁶-Ethenodeoxyadenosine (edA) and 3,N⁴-ethenodeoxycytidine(edC) are two mutagenic adducts associated with exposure toethyl carbamate (urethane) and vinyl chloride. We have recentlydeveloped two ultrasensitive methods for determining the moleculardose of these adducts in cellular DNA. In both methods, purifiedDNA was first enzymatically digested to 2c-deoxyribonucleotide3c-monophosphates. Etheno-modified nucleotides were then separatedfrom normal nucleotides in one of two ways: either by reversephase, ion-pair HPLC coupled with 260 nm UV detection, or byimmunoaffinity chromatography using reusable microcolumns containingspecific monoclonal antibodies coupled to Protein A–Sepharose.Fractions enriched for the adducted nucleotides were labeledusing T4 polynucleotide kinase and [³²P]ATP, and individualnucleotides were subsequently resolved by two-dimensional TLC,visualized by autoradiography, and quantified by liquid scintillationcounting. When used to analyze the same sample of etheno-modifiedcalf thymus DNA, both assays produced similar results. However,when both methods were used to analyze rat liver DNA ‘spiked’with known amounts of etheno nucleotide standards, the immunoaffinity/³²PTLC procedure proved to be more sensitive and more reproduciblethan the HPLC/³²P TLC method: while the detection limit of theimmunoaffinity/³²P TLC technique was < 4 etheno adducts/10⁹parent deoxynucleotides, the HPLC/³²P TLC method often failedto detect adducts at concentrations <2/108. In other experiments,the immunoaffinity/³²P TLC method was used to demonstrate formationof edA and edC in cells treated with vinyl chloride monomer.Because of its exquisite sensitivity, the immunoaffinity/³²PTLC method promises to be extremely useful for measuring bothbackground and induced levels of etheno adducts, making it possibleto examine the role of these adducts in inducing mutations and/orcarcinogenesis.     

Activation of 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) to mutagenic metabolites

Metabolism of heterocyclic amines to N-hydroxy intermediates appears critical in the mutagenic and carcinogenic actions of these compounds. We have studied the murine hepatic microsomal and cytosolic activation of 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), a heterocyclic amine mutagen formed in cooked meats. PhIP (25 microM) was almost completely converted to N-hydroxy-PhIP and 4'-hydroxy-PhIP in 30 min by reaction with 3-methylcholanthrene-induced microsomal preparations. Microsomal formation of the active N-hydroxy-PhIP metabolite was slightly favored over the 4'-hydroxy-PhIP detoxification product at all concentrations studied (25-200 microM). Metabolism of PhIP in microsomal preparations derived from control mice was approximately of the induced preparations. Metabolically activated PhIP and synthetic N-hydroxy-PhIP produced concentration-dependent increases in mutagenic activity in both Salmonella strains TA98 and TA98/1,8-DNP6, indicating that acetylated intermediates were not important in the mutagenicity of N-hydroxy-PhIP in these bacteria. Significant stabilization of the N-hydroxy-PhIP intermediate by both microsomal protein and BSA was observed. Addition of cytosol to microsomal incubations with PhIP (25 microM) resulted in an increase in mutagenic activity which could be attributable to stabilization by glutathione. An additional increase in mutagenicity resulted from addition of 3'-phosphoadenosine 5'-phosphosulfate (PAPS), but not acetyl coenzyme A to microsomal preparations containing the cytosolic fraction. Furthermore, addition of PAPS to cytosolic preparations containing synthetic N-hydroxy-PhIP produced a 17% decrease in levels of the arylhydroxylamine relative to controls over 30 min, suggesting that secondary metabolism of N-hydroxy-PhIP to a sulfate conjugate may be relevant to the mutagenic and carcinogenic actions of PhIP.