Deamination of Single-Stranded DNA Cytosine Residues in Aerobic

P. M. Gross Chemical Laboratory, Duke University, Durham, North Carolina ... Frederick Cancer Research and Development Center, Frederick, Maryland 217...
10 downloads 0 Views 215KB Size
Download PDF
Chem. Res. Toxicol. 1996, 9, 891-896

891

Deamination of Single-Stranded DNA Cytosine Residues in Aerobic Nitric Oxide Solution at Micromolar Total NO Exposures Kunal Merchant,† Hong Chen,† Theodore C. Gonzalez,† Larry K. Keefer,‡ and Barbara Ramsay Shaw*,† P. M. Gross Chemical Laboratory, Duke University, Durham, North Carolina 27708-0346, and Chemistry Section, Laboratory of Comparative Carcinogenesis, National Cancer Institute, Frederick Cancer Research and Development Center, Frederick, Maryland 21702 Received June 13, 1995X

Deamination of cytosine to uracil is a potential source of mutations in DNA. Here we examine the deaminating ability of aerobic nitric oxide (NO) toward single-stranded DNA at very low (micromolar and below) total exposures, using a sensitive genetic method that allows us to study a single deamination event at a specific site in a 7200-nucleotide DNA molecule within a pool of ca. 100 000 other identical DNA molecules. We incubated gapped C141 M13mp2 DNA with the NO-generating compound, Et2N[N(O)NO]Na (DEA/NO), in aerobic buffer for 16 h to ensure complete autoxidation at pH 7.4 and 37 °C. After ultrafiltration to remove small molecules, the DNA was transformed into isogenic Escherichia coli cultures that were either deficient (NR9404, ung-) or proficient (MC1061, ung+) in uracil-DNA glycosylase activity. The gapped DNA was constructed such that the target (CCC) codon was contained in a short single-stranded segment of otherwise double-stranded circular DNA, and the incubation was performed in a closed system to prevent loss of NO to the atmosphere before the reaction was complete. An increase in the reversion frequency in the ung- strain was noted between 0 and 1 µM DEA/NO, and the reversion frequency leveled out between 3 and 30 µM. However, 30 µM “spent” DEA/NO (i.e., that which was similarly incubated for 4 h to complete the autoxidation of NO before the DNA was added) did not increase reversion frequency relative to control. Nearly all (42/43) of the mutations identified after 1 µM DEA/ NO treatment were C f T transitions, and reversion frequency in the isogenic ung+ strain was lower than in the ung- strain. The data are consistent with the hypothesis that total NO exposures in the µmol/L range can lead to C f T mutations via a mechanism most probably involving deamination of DNA cytosine residues.

Introduction Nitric oxide is a highly reactive free radical gas that is a common pollutant present in air, automobile exhaust, and cigarette smoke (1-3). Furthermore, nitric oxide is endogenously produced by macrophages as a physiological response to infection and is also involved in numerous biological functions such as regulation of immune function and as a chemical messenger for neurons (4, 5). Since nitric oxide can diffuse across membranes (6, 7) and may react, in numerous ways, to generate DNAreactive species, the understanding of nitric oxideinduced DNA damage is of vital importance. One of the most frequently occurring forms of DNA damage that leads to mutations is the deamination of cytosine nucleobases to uracil (8, 9). Spontaneous deamination of cytosine at physiological conditions is estimated to occur every 14 min in double-stranded DNA, and as often as every 6.7 s in single-stranded DNA the size of the human genome (∼109 base pairs) (10). Here, we attempt to investigate if cytosine deamination plays an important role in the mutagenic action of nitric oxide on singlestranded DNA. † ‡ X

Duke University. National Cancer Institute. Abstract published in Advance ACS Abstracts, June 15, 1996.

S0893-228x(95)00102-0 CCC: $12.00

In vivo, DNA is mainly in the double-stranded form, where base stacking and hydrogen bonding limit accessibility of nucleobases to reactive species. It is therefore not surprising that the rate of spontaneous hydrolytic deamination in double-stranded DNA has been determined to be approximately 100-fold less than in singlestranded DNA (11-13; reviewed by Shaw, ref 14). However, cytosines in C:C and C:T mismatches are found to deaminate with a 20- to 100-fold rate enhancement; i.e., these rates approach those of single-stranded DNA (12). It has also been observed that there is a direct correlation between the melting of duplex DNA and its rate of deamination (14, 15), indicating that the deamination of cytosine in double-stranded DNA is inversely related to the stability of the duplex. In view of such studies, the process of deamination in double-stranded DNA is thought to occur via a single-stranded intermediate (13, 14). Such a hypothesis becomes more plausible when the transient, local single-stranded character of DNA is considered during processes such as transcription, replication, recombination, and “breathing” (16). In fact, a study by Fix and Glickman (17) in Escherichia coli revealed that spontaneous deamination of cytosine selectively occurred on cytosines flanking AT-rich regions. The authors reasoned that this was due to the increased single-stranded character of the nontranscribed strand and of regions rich in AT base pairs. In eucaryotes, the © 1996 American Chemical Society

892

Chem. Res. Toxicol., Vol. 9, No. 5, 1996

40-fold higher rate of deamination in Saccharomyces cerevisiae over E. coli was attributed to the lower rate of transcription in yeast, as the yeast may keep the DNA in single-stranded form for a longer period of time (18). A DNA repair enzyme, uracil-DNA glycosylase, ubiquitous to procaryotic and eucaryotic cells, shows specificity for efficient excision of uracil from both single- and double-stranded DNA, with preference for the former (19). Excision from single-stranded DNA occurs more rapidly (20, 21), and the enzyme specifically binds uracils that are flipped out from the DNA duplex (22). Thus, using a single-stranded DNA model in a deamination study not only offers the added advantage of experimental ease due to its high rate of deamination and limited number of repair pathways, but it also provides insight into the fate of a physiologically relevant form of DNA. The role of nitric oxide in deamination processes remains elusive. The first reports on exposure of DNA to large amounts of nitric oxide (NO) suggested that NOinduced cytosine deamination might indeed represent an important mechanism of genotoxicity (7). Significant deamination of cytosine residues to uracil was observed on bubbling NO through buffered aerobic solutions of calf thymus DNA at pH 6.8-7.6 (7) (accumulated doses of 0.1-1 mol of NO absorbed per liter of solution). Treatment of Salmonella typhimurium strain TA1535 with SPER/NO, a complex of nitric oxide and spermine that generates NO at low concentrations in solution, induced mutations that were 99% G:C f A:T transitions at the hisG46 (CCC) locus (7). G:C f A:T transitions were also the most abundant mutations seen when plasmid (pSP189) DNA was exposed to millimolar concentrations of SPER/ NO or another NO donor compound, DEA/NO, in neutral buffer and transformed into human Ad293 cells (23). These mutagenicity data are consistent with a mechanism of action involving deamination of cytosine (24) or 5-methylcytosine (25) to uracil or thymine, respectively. Both bases if not repaired will pair with adenine on replication. Quite a different result was observed when NO gas was bubbled through a pSP189 plasmid solution at neutral pH, with total exposures of 100 mmol/L of nitric oxide absorbed. In this case, upon transformation of the DNA the majority of the mutations were A:T f G:C transitions (26). This base substitution is also consistent with a deaminative mechanism, but one involving conversion of adenine to hypoxanthine, which pairs with cytosine (27), rather than deamination of cytosine. A third series of experiments appeared to exclude any role for (methyl)cytosine deamination in the mutagenic action of NO. When Schmutte et al. treated the plasmid pSV2neo with NO gas or SPER/NO at pH 7.4 and transformed it into uracil-DNA glycosylase-deficient (ung-) E. coli (unable to repair deaminated cytosine residues), they did not observe any changes in reversion frequency in the millimolar or micromolar range, respectively (28). Exposure of the transformed bacteria to the NO-generating agent during log-phase growth also failed to produce larger numbers of revertants, and no differences between ung- and ung+ strains were observed. Since the genetic assay used in those experiments was capable of detecting low levels of G:C f A:T transitions, the negative results led the study’s authors to conclude that the mutagenicity of NO was not caused by deamination of cytosine or 5-methylcytosine in double-stranded DNA in E. coli (28). Yet in Salmonella, high concentra-

Merchant et al.

Figure 1. The gapped DNA used in these experiments is constructed from double-stranded C141 M13mp2 DNA (linearized by treating with the restriction enzymes, PvuII and BglI) and single-stranded circular C141 M13mp2 DNA, as shown. The single-stranded gapped region (417 base pairs long) contains the target site. Abbreviations: ds, double-stranded; ss, singlestranded.

tions of SPER/NO produced a large number of revertants, suggesting that NO did produce C f T mutations. Interpretation of these divergent findings becomes even more complicated when one considers that incubation of SPER/NO in aerobic cultures produces a mixture of nitrite ion (the autoxidation product of NO in aqueous solution (29)) with spermine, because nitrite-polyamine mixtures have been shown to be mutagenic in their own right by a mechanism that has yet to be elucidated (3032; reviewed in ref 33). We now report that treatment of a single-stranded DNA target with micromolar concentrations of the nitric oxide-releasing compound, DEA/NO, in a reversion assay first described by Frederico et al. (11) and reviewed by Shaw (14), leads to a significant increase in the number of C f T transitions, which arise from deamination of cytosine to uracil or uracil-like intermediates.

Materials and Methods Preparation of “Gapped” Plasmid. C141 M13mp2 is a mutant of M13mp2 which has a cytosine, in lieu of guanine, at position 141 of the lacZ coding sequence; i.e., the sequence in this region of the mutant gene is 137TTTC141CCCAG146C (34) where the target codon is underlined. Five hundred micrograms of C141 M13mp2 DNA (double-stranded, replicative form) was digested with the restriction enzymes PvuII and BglI, for 1 h at 37 °C in 50 mM Tris (pH 7.9) containing 10 mM MgCl2, 100 mM NaCl, and 1 mM dithiothreitol. These restriction enzymes cut the circular, double-stranded DNA at sites 6377 (BglI) and 5960, 6053, and 6321 (PvuII) such that a 417-bp region encompassing cytosine-141 of the coding strand and its complement are excised (see Figure 1). DNA was then precipitated by addition of 2.5 volumes of ethanol and recovered by centrifugation. Two hundred micrograms was diluted in water to a concentration of 0.5 µg/µL. The diluted DNA was then denatured at 95 °C for 10 min and cooled on ice for 5 min. Fifty micrograms of circular, single-stranded C141 M13mp2 DNA was added to the denatured, linearized DNA, followed by sodium chloride and sodium citrate at final concentrations of 0.3 and 0.03 M, respectively. The mixture was then incubated for 5 min at 60 °C to produce a circular “gapped” hybrid which was doublestranded everywhere except in the region containing cytosine141 of the coding strand. The entire sample was cooled on ice and then desalted using a ChromaSpin-10 + TE (Tris, EDTA) column (Clontech). Incubation Conditions. DEA/NO was synthesized as previously described (35). Its formula is Et2N[N(O)NO]Na

NO-Induced Cytosine Deamination

Chem. Res. Toxicol., Vol. 9, No. 5, 1996 893

Table 1. Time Required To Oxidize 95% of the NO Produced at Each DEA/NO Concentrationa [DEA/NO] (µM)

time (min)

[DEA/NO] (µM)

time (min)

30 10 3

8.7 26.2 87.4

1 0.3 0.1

262.3 874.2 960

a

Calculated using the equation:

time )

0.95 k[O2][NO]0(0.05)

The calculation assumes the following: (1) that each mole of DEA/ NO generates 1.5 mol of NO (35); (2) that the O2 concentration is 230 µM at 37 °C in air-saturated buffer (10 mM HEPES, pH 7.4); (3) that the rate-determining step is the reaction of NO with O2 according to eq 1, where k ) 3.5 × 106 M-2 s-1 (37). A corollary of the latter assumption is that generation of NO from DEA/NO is instantaneous. Since the half-life of DEA/NO is actually 2.1 min at pH 7.4 and 37 °C (35), this is an approximation for high concentrations of DEA/NO, where the rates of generation and oxidation of NO are comparable. However, this approximation does not affect the conclusion that NO oxidation is effectively 95% complete (87% in the case of 0.1 µM DEA/NO) during a 16-h incubation. This has been confirmed in the 30 µM case; analysis for nitrite + nitrate + N-nitrosodiethylamine in duplicate incubations at infinite time yielded values of 57.6 and 54.5 µM (theoretical 60 µM).

(Chemical Abstracts Registry No. 92382-74-6).1 A stock solution was prepared by dissolving it in cold 1 mM sodium hydroxide. Serial dilutions of this stock with 1 mM NaOH were used to produce different concentrations of DEA/NO, and 100-fold dilutions of these concentrations were used for the incubations with DNA. The final sodium hydroxide concentration in each incubation mixture remained the same. A 1.5-mL Eppendorf tube containing 1 µg of gapped DNA in 100 µL of 10 mM HEPES buffer (pH 7.4) was incubated at 37 °C until thermal equilibrium was established, whereupon a 1-µL aliquot of the serially diluted DEA/NO in 1 mM sodium hydroxide was added. The resulting solution was quickly mixed, taken up in a glass reaction tube, and sealed with no headspace within 8 s of adding the DEA/ NO solution. After incubation for 16 h at 37 °C, the DNA solution was diluted 11-fold with ice-cold Milli-Q treated water and concentrated to 100 µL in a Centricon (Amicon, Inc.) with a 100-kDa cutoff by centrifuging at 500g for 25 min. The latter step greatly reduced the concentration of nitrite ion and other low-molecular-weight species prior to transformation. To ensure quantitatively comparable results, we conducted all the reactions in closed vessels with no headspace (36) to avoid variable loss of the highly diffusible NO from the system. In this way, it was calculated that 87% of the NO was converted to a nitrosating species in 16 h when the starting DEA/NO concentration was 0.1 µM, with the higher DEA/NO concentrations giving conversions of 95% or greater during the 16-h incubation (see Table 1). Mutagenicity Assay. Our study utilizes a sensitive genetic assay to determine mutagenicity under well-defined in vitro conditions (11, 14, 34, 38). Deamination of cytosine is a rare event that occurs spontaneously or in the presence of a mutagen. The assay is based on the reversion of the mutant CCC proline codon in the lacZ R gene carried in bacteriophage M13mp2; C f T and C f A substitutions at cytosine-141 (34) as well as any base change at cytosine-142 can be detected. This mutant C141 phage produces a colorless plaque phenotype which, upon deaminating to uracil at either of the first two cytosine residues in the codon, can revert to a dark blue phenotype when transformed into ung- cell strains in the presence of 5-bromo4-chloro-3-indolyl β-D-galactopyranoside (X-Gal). Ung- cell strains are deficient in the DNA repair enzyme, uracil-DNA glycosylase. This enzyme is responsible for the detection and removal of uracil from DNA. Uracil will be mistaken for thymine during DNA replication in cells deficient in this 1

Registry No. was supplied by the author.

Table 2. Reversion Frequency as a Function of DEA/NO Concentration

[DEA/NO] (µM) 0 0.1 0.3 1 3 10 30 30 (“spent”)a

reversion efficiency of total no. total no. frequency transformation of plaques of blue (revertants/ (plaques/pg (in thousands) plaques 105 plaques) of DNA) 162 125 159 94 112 86 109 165

21 22 36 27 40 30 38 24

13.0 17.7 22.6 28.9 35.9 35.0 35.0 14.6

6.8 6.2 9.9 7.8 5.6 5.4 5.4 10.3

a To obtain the spent sample, DEA/NO was incubated for 4 h at 37 °C to accomplish near-complete autoxidation of the NO it released before adding the DNA.

enzyme. The limit of detection of this assay is defined by the background reversion frequency which is ca. 10-4 for the gapped DNA used here, or 1 mutation per 10 000 DNA cytosine target sites. Each reversion frequency represented in Table 2 is obtained from at least three transformations. After incubation, 1-µL volumes containing 4 or 5 ng of DNA aliquots were mixed with 50 µL of concentrated ung- NR9404 hsdR2 hsdM+ hsdS+ araD139 ∆(ara-leu)7697 ∆(lac)x74 galE15 galK16 rpsL (Strr) mcrA mcrB1 ung1::Tn10 or control ung+ MC1061 hsdR2 hsdM+ hsdS+ araD139 ∆(ara-leu)7697 ∆(lac)x74 galE15 galK16 rpsL (Strr) mcrA mcrB1 cells (isogenic strains supplied by R. Schaaper and T. A. Kunkel, NIEHS). Each transformation was accomplished by electroporation using a BTX600 electrocell manipulator set at 2 kV (34). Aliquots of the electroporated cells were mixed with NR9099 cells (host cells for M13mp2) which were obtained from T. A. Kunkel. X-Gal (Biosynth AG) and isopropyl β-D-thiogalactopyranoside (Biosynth AG) were added, and the cell mixture was poured over soft agar plates; overnight incubation revealed predominantly nonrevertant white plaques and any revertant blue plaques (11, 34). The revertant phenotypes of the latter were confirmed by plaque purification, followed by sequencing in the region containing cytosine-141 with the Sanger method.

Results For the experiments described here, a single-stranded DNA target rather than a double-stranded one was used for the following reasons. (1) With single-stranded DNA, the specific strand and base which are damaged by nitric oxide can be clearly identified. The site at which damage occurs is on the strand responsible for mutation, and thus the observed mutation can be directly linked to damage on the same strand. With double-stranded DNA it is difficult to distinguish whether the damage occurs on the cytosine of the target strand or on the guanine of the complementary strand since both strands can serve as templates for DNA replication. (2) Repair mechanisms for double-stranded DNA are more complex (involving excision repair as well as other mismatch repair enzymes), and thus interpretation of results is not straightforward. In a genetic system such as ours that uses isogenic cell strains differing only in uracil-DNA glycosylase repair activity, assessment of cytosine deamination can best be done with a single-stranded or gapped DNA target. (3) Another reason for using single-stranded DNA is that the process of deamination in double-stranded DNA is thought to occur via a single-stranded intermediate (13, 14) and therefore the use of single-stranded DNA is a justifiable step toward our goal to examine whether exposure to low concentrations of aerobic NO causes deamination of cytosine. (4) The rates of deamination for single-stranded DNA are usually orders of magnitude

894

Chem. Res. Toxicol., Vol. 9, No. 5, 1996

Figure 2. The means and standard deviations of the reversion frequencies obtained for DNA, exposed to either 0 or 1 µM DEA/ NO, and transformed into uracil-DNA glycosylase-deficient (ung-) versus -proficient (ung+) E. coli. The DNA was incubated for 16 h at 37 °C in aerobic HEPES buffer, pH 7.4. The asterisk denotes a difference that was significant at the p < 0.001 level relative to the corresponding ung+ value as well as to the untreated ung- control.

faster than with double-stranded DNA (12, 14), making experimental conditions practicable. In order to observe reversion events due to deamination by nitric oxide with a sensitive assay and without excessive loss of signal, it was first necessary to optimize the conditions. Initial experiments were performed by exposing circular single-stranded M13mp2 DNA to concentrations of DEA/NO in the micromolar to millimolar range. Those experiments (data not shown) resulted in low efficiencies of transformation with increasing doses of DEA/NO, accompanied by extensive linearization of the single-stranded DNA; this resulted in a decrease in reversion frequency, because a single break in the singlestranded DNA is sufficient to inactivate it. To increase the probability of detecting induced deamination events, the experiments were repeated, but using instead a gapped DNA template which is double-stranded everywhere except in the target region, making the template more resistant to linearization. It is notable that the efficiencies of transformation for the gapped construct changed very little with increases in NO concentration (Table 2) and were about 10 times higher than those of the completely single-stranded DNA. The gapped DNA construct therefore enabled us to study deamination in a single-stranded target without loss of signal caused by damage elsewhere in the molecule. Using the gapped DNA construct, we demonstrated that DEA/NO is significantly mutagenic under the conditions employed. Reversion frequencies in the ung- strain observed at DEA/NO exposures ranging from 0 to 30 µM are shown in Table 2. An increase was noted with dose to 3 µM, but the reversion frequency remained constant between 3 and 30 µM. Reproducibility was demonstrated by repeating the 1 µM DEA/NO treatment three more times and averaging the four results. The mean reversion frequency ( standard deviation was 25.7 ( 3.2 revertants per 105 plaques, which was significantly different from that of the untreated controls (p ) 0.0009, two-tailed t test). The data are summarized in Figure 2. Transformation of 4- to 5-ng aliquots of DNA, which had been exposed to 1 µM DEA/NO, into the ung+ strain showed that mutagenicity resulting from DEA/NO treatment was substantially reduced. Figure 2 shows the relative reversion frequencies in the isogenic ung+ and ung- strains for DNA treated with 1 µM DEA/NO and for the untreated DNA. The data suggest that the

Merchant et al.

majority of mutations resulted from deamination of cytosine to uracil, which could not be efficiently repaired by the ung- bacteria. Further support for cytosine deamination was found upon sequencing DNA from the mutant clones. All 33 of the mutants isolated from the untreated ung- controls and 42 of the 43 mutants sequenced after 1 µM DEA/ NO treatment had C f T transitions at the target codon. To verify that diethylamine and nitrite ion, the ultimate byproducts of DEA/NO decomposition under these conditions, were not responsible for the observed mutations, 30 µM DEA/NO was incubated at 37 °C for 4 h before mixing with the DNA. Subsequent 16-h exposure of previously untreated DNA to this “spent” DEA/NO and its subsequent transformation into the ung- bacteria as above yielded a reversion frequency of 14.6 mutants per 105 plaques, far fewer than were seen with one-thirtieth that concentration of active DEA/NO (Table 2).

Discussion The genotoxic effects of NO are well-known and have been extensively studied; however, the chemical nature of the interaction of NO or rather its reactive oxidation product(s), NOx, with DNA is still an open question. Since the free radical, nitric oxide, can diffuse through membranes, we have specifically sought to determine whether, at very low micromolar (physiologically relevant) concentrations, NO or NOx is capable of damaging DNA bases via a cytosine deamination mechanism. We have shown, under aerobic conditions, that concentrations as low as 1 µM DEA/NO can lead to C f T mutations in DNA at levels significantly above background. The reversion frequency seems to increase linearly with dose to 1 µM and remains constant between the doses of 3 and 30 µM DEA/NO. Yet, DNA incubated with 30 µM “spent” DEA/NO did not show a significant increase in reversion frequency, indicating that the ultimate byproducts of NO generation, diethylamine and nitrite, are not responsible for the elevated levels of deamination. The majority of mutations in the DNA samples exposed to 1 µM DEA/NO are deaminations of cytosine residues, as indicated by the significant differences in reversion frequencies between ung+ and ungstrains (Figure 2). Further support for a deaminative mechanism was noted in the sequencing results, in that 98% (42 out of 43) of the mutations seen after treating DNA with 1 µM DEA/NO were C f T transitions. Comparisons of the increase in reversion frequency over background between two deaminating agents can provide some quantitative insights on the relative potencies of mutagens. In this study we have observed a 2-fold increase in reversion frequency above background for DNA exposed to 1 µM DEA/NO after a 16-h incubation at 37 °C and pH 7.4 (Table 2). A similar 2-fold increase in reversion frequency (from 9.8 × 10-5 to 20 × 10-5) is achieved in single-stranded DNA incubated with bisulfite under similar conditions (37 °C, pH 7.4; 24-h incubation), but only at a considerably higher bisulfite concentration of 50 mM (39). One can only speculate about the cause of the plateau effect noted for the reversion frequencies from 3 to 30 µM DEA/NO. Considering that the transformation efficiencies are not decreased at this concentration (which would occur if there were single-strand breaks in the single-stranded target), we reasoned that the plateau in reversion frequency between 3 and 30 µM DEA/NO is

NO-Induced Cytosine Deamination

Chem. Res. Toxicol., Vol. 9, No. 5, 1996 895

probably not due to DNA damage. Neither could the plateau be attributed to oxygen depletion, as the oxygen content of air-saturated buffers (ca. 230 µM) is much greater than the total amount of available NO. Since each mole of DEA/NO, with a half-life of 2.1 min at pH 7.4 and 37 °C, generates 1.5 mol of NO (35), we calculated from the rate law (37) for the autoxidation of NO in aqueous media (eq 1, with k ) 3.5 × 106 M-2 s-1 at 37 °C) that 87% of the NO in our 16-h incubation was converted to a nitrosating species when the starting DEA/ NO concentration was 0.1 µM, with the highest DEA/ NO concentrations giving conversions of g95%. If the

rate ) k[O2][NO]2

(1)

reactive nitrosating intermediate partitions between cytosine deamination and its combined other fates in the same ratio for all DEA/NO concentrations used in this study, then a linear relationship is expected between total DEA/NO concentration and increasing reversion frequencies. A reviewer has suggested the interesting possibility that DEA/NO may bind to the DNA at diffusioncontrolled rates, after which the release of NO and the base-modifying reactions occur. If the postulated binding sites are saturable, such a model could account for the observed plateau. However, the exact mechanism for inducing the cytosine deamination observed in our study remains to be established. Regarding the interesting study reported recently by Schmutte et al. (28) using an assay very similar to ours, we speculate that the negative results they found under similar conditions could be explained in a number of ways. (a) Their use of a double-stranded plasmid can cause complications in interpreting results due to repair processes, as explained in the Results section. (b) Their study was conducted using a slow NO-releasing agent, SPER/NO (t1/2 ) 39 min) (35), without confinement in a closed vessel, which would result in a loss of NO to the system since NO is highly diffusible. Further, the authors also indicate that SPER/NO precipitates DNA. (c) Other experimental factors such as their use of 100 mM phosphate buffer cannot be excluded as possible contributors to the observed difference in outcomes, since phosphate is known to inhibit nitrosation by NOx (40). (d) It is also not clear whether the delivery of SPER/NO into E. coli and S. typhimurium cells was equivalent during log-phase growth. Schmutte et al. (28) concluded that the C f T mutations arising due to nitric oxide exposure were not caused by deamination, based on their observation that the already very low reversion frequencies of growing log-phase cultures of E. coli ung+ and ungcells did not change on exposure to SPER/NO; yet the number of C f T mutations increased when growing cultures of S. typhimurium were exposed to SPER/NO. However, their observations could simply be explained in terms of the difference in strains. The S. typhimurium strain TA1535 (41) used by Schmutte et al. (28) contains the rfa mutation that causes partial loss of the lipopolysaccharide barrier on the surface of the bacteria and increases its permeability, which may have allowed SPER/NO to penetrate more freely into this S. typhimurium strain. Results obtained from an experiment conducted with the S. typhimurium strain cannot be compared to those obtained with E. coli strains that have no mutations affecting permeability. In fact, Wink and Laval (42) observed that when wild type E. coli cells were exposed to DEA/NO, DNA repair activity was unaffected,

an observation thought to be related to the impermeability of the cell. They resolved the impermeability by using a special (acr A-) mutant of E. coli which alters the membrane function, sensitizing the strain to various agents (42). It should also be noted that the S. typhimurium strain TA1535 carries another mutation, i.e., the deletion of the uvr B gene, which codes for the DNA excision repair system, and highly sensitizes the strain to mutagens (41). In light of the above facts, it is quite clear that SPER/NO delivery into the cells may not be equal in the two organisms, E. coli and S. typhimurium, used by Schmutte et al. and is probably more extensive in TA1535, causing a higher degree of damage in the latter due to the lack of the nucleotide excision repair system. Our system permits a more straightforward comparison of the effect of NO on DNA damage. We expose the DNA (not cells) directly to NO, allow time for complete autoxidation in a closed system, and then transform the DNA into isogenic ung+ and ung- bacteria. The significant reduction in reversion frequency on transforming DEA/NO-exposed DNA in ung+ bacteria constitutes strong evidence for a mechanism of mutagenicity involving the direct deamination of cytosine to uracil by reactive species produced on autoxidation of NO. The fact that this reduction was substantial, but not complete, suggests that deamination is the major mechanism but that other mutational mechanisms might also be operative. Hartman et al. have proposed several interesting pathways that could account for uracil-DNA glycosylase insensitivity, albeit in double-stranded DNA, in similar experiments with nitrous acid (33). In addition, our experiments cannot rule out other mutation mechanisms that would produce uracil-like intermediates recognized by uracil-DNA glycosylase (43). The extremely high sensitivity of our system permits assessment of DNA damage at low levels of NO in the presence of oxygen. We conducted the NO exposure in vitro in an attempt to accurately control the NO and O2 concentrations and demonstrate that single-stranded DNA is susceptible to damage by even submicromolar concentrations of aerobic NO or NOx. Given the singlestranded character of genomic DNA under certain circumstances, e.g., during transcription and replication, and the likelihood that deamination of bases in doublestranded DNA occurs via a single-stranded intermediate, our findings indicate that the deamination resulting from intracellular exposure to even very low concentrations of NO should not be ignored when considering the toxicological potential of this important bioregulatory species.

Acknowledgment. We thank Drs. R. Schaaper and T. A. Kunkel of the National Institute of Environmental Health Sciences for generously providing the cell lines used in these experiments, Dr. J. E. Saavedra for kindly supplying the DEA/NO, Dr. G. M. Janini and Mr. S. D. Fox for analytical support, Mr. C. W. Riggs for invaluable statistical consultation, and Drs. Lubica Cernakova and Edward Budowsky for helpful discussions. K.M. was supported in part by the R. J. Reynolds-Leon Golberg Fellowship from the Duke University Toxicology Program. Research was supported by a grant to B.R.S. from the National Institutes of Health (5R01-CA-44709).

896

Chem. Res. Toxicol., Vol. 9, No. 5, 1996

References (1) Anonymous (1977) Oxides of Nitrogen. WHO Environmental Health Criteria 4, World Health Organization, Geneva. (2) Lewis, T. R. (1980) Criteria relevant to an occupational health standard for nitrogen dioxide. In Nitrogen Oxides and Their Effects on Health (Lee, S. D., Ed.) pp 361-375, Ann Arbor Science, Ann Arbor. (3) Pryor, W. A., and Stone, K. (1993) Oxidants in cigarette smoke. Radicals, hydrogen peroxide, peroxynitrate, and peroxynitrite. Ann. N.Y. Acad. Sci. 686, 12-28. (4) Marletta, M. A. (1988) Mammalian synthesis of nitrite, nitrate, nitric oxide, and N-nitrosating agents. Chem. Res. Toxicol. 1, 249257. (5) Lowenstein, C. J., and Snyder, S. H. (1992) Nitric oxide, a novel biological messenger. Cell 70, 705-707. (6) Nguyen, T., Brunson, D., Crespi, C. L., Penman, B. W., Wishnok, J. S., and Tannenbaum, S. R. (1992) DNA damage and mutation in human cells exposed to nitric oxide in vitro. Proc. Natl. Acad. Sci. U.S.A. 89, 3030-3034. (7) Wink, D. A., Kasprzak, K. S., Maragos, C. M., Elespuru, R. K., Misra, M., Dunams, T. M., Cebula, T. A., Koch, W. H., Andrews, A. W., Allen, J. S., and Keefer, L. K. (1991) DNA deaminating ability and genotoxicity of nitric oxide and its progenitors. Science 254, 1001-1003. (8) Halliday, J. A., and Glickman, B. W. (1991) Mechanisms of spontaneous mutation in DNA repair-proficient Escherichia coli. Mutat. Res. 250, 55-71. (9) Lindahl, T. (1993) Instability and decay of the primary structure of DNA. Nature 362, 709-715. (10) Shapiro, R. (1981) Damage to DNA caused by hydrolysis. In Chromosome Damage and Repair (Seeberg, E., and Kleppe, K., Eds.) pp 3-18, Plenum Press, New York. (11) Frederico, L. A., Kunkel, T. A., and Shaw, B. R. (1990) A sensitive genetic assay for the detection of cytosine deamination: Determination of rate constants and the activation energy. Biochemistry 29, 2532-2537. (12) Frederico, L. A., Kunkel, T. A., and Shaw, B. R. (1993) Cytosine deamination in mismatched base pairs. Biochemistry 32, 65236530. (13) Lindahl, T., and Nyberg, B. (1974) Heat-induced deamination of cytosine residues in deoxyribonucleic acid. Biochemistry 13, 34053410. (14) Shaw, B. R. (1994) Induced deamination of cytosine: Effects of secondary structure and other factors. In Structural Biology: The State of the Art (Sarma, R. H., and Sarma, M. H., Eds.) Vol. 2, pp 367-378, Adenine Press, Schenectady. (15) Briley, J. D., Moyer, R., and Shaw, B. R. (in preparation). (16) Bjursell, G., Gussander, E., and Lindahl, T. (1979) Long regions of single-stranded DNA in human cells. Nature 280, 420-423. (17) Fix, D. F., and Glickman, B. W. (1987) Asymmetric cytosine deamination revealed by spontaneous mutational specificity in an ung- strain of Escherichia coli. Mol. Gen. Genet. 209, 78-82. (18) Impellizzeri, K. J., Anderson, B., and Burgers, P. M. J. (1991) The spectrum of spontaneous mutations in a Saccharomyces cerevisiae uracil-DNA-glycosylase mutant limits the function of this enzyme to cytosine deamination repair. J. Bacteriol. 173, 6807-6810. (19) Domena, J. D., Timmer, R. T., Dicharry, S. A., and Mosbaugh, D. W. (1988) Purification and properties of mitochondrial uracilDNA glycosylase from rat liver. Biochemistry 27, 6742-6751. (20) Cleaver, J. E., and Layher, S. K. (1995) “If the shoe fits”: Clues on structural recognition of DNA damage. Cell 80, 825-827. (21) Savva, R., McAuley-Hecht, K., Brown, T., and Pearl, L. (1995) The structural basis of specific base-excision repair by uracilDNA glycosylase. Nature 373, 487-493. (22) Mol, C. D., Arvai, A. S., Slupphaug, G., Kavli, B., Alseth, I., Krokan, H. E., and Tainer, J. A. (1995) Crystal structure and mutational analysis of human uracil-DNA glycosylase: structural basis for specificity and catalysis. Cell 80, 869-878. (23) Routledge, M. N., Wink, D. A., Keefer, L. K., and Dipple, A. (1994) DNA sequence changes induced by two nitric oxide donor drugs in the supF assay. Chem. Res. Toxicol. 7, 628-632.

Merchant et al. (24) Duncan, B. K., and Miller, J. H. (1980) Mutagenic deamination of cytosine residues in DNA. Nature 287, 560-561. (25) Coulondre, C., Miller, J. H., Farabaugh, P. J., and Gilbert, W. (1978) Molecular basis of base substitution hotspots in Escherichia coli. Nature 274, 775-780. (26) Routledge, M. N., Wink, D. A., Keefer, L. K., and Dipple, A. (1993) Mutations induced by saturated aqueous nitric oxide in the pSP189 supF gene in human Ad293 and E. coli MBM7070 cells. Carcinogenesis 14, 1251-1254. (27) Kamiya, H., Miura, H., Kato, H., Nishimura, S., and Ohtsuka, E. (1992) Induction of mutation of a synthetic c-Ha-ras gene containing hypoxanthine. Cancer Res. 52, 1836-1839. (28) Schmutte, C., Rideout, W. M., III, Shen, J.-C., and Jones, P. A. (1994) Mutagenicity of nitric oxide is not caused by deamination of cytosine or 5-methylcytosine in double-stranded DNA. Carcinogenesis 15, 2899-2903. (29) Ignarro, L. J., Fukuto, J. M., Griscavage, J. M., Rogers, N. E., and Byrns, R. E. (1993) Oxidation of nitric oxide in aqueous solution to nitrite but not nitrate: Comparison with enzymatically formed nitric oxide from L-arginine. Proc. Natl. Acad. Sci. U.S.A. 90, 8103-8107. (30) Murphey-Corb, M., Kong, H.-L., and Murray, M. L. (1980) Interaction of mutagenic spermidine-nitrous acid reaction products with uvr- and recA-dependent repair systems in Salmonella. J. Bacteriol. 142, 191-195. (31) Murphey-Corb, M., Kong, H.-L., and Murray, M. L. (1983) Mutagenic activity from nitrosation of oligoamines. Environ. Mutagen. 5, 101-109. (32) Correa, P., Kokatnur, M. G., and Murray, M. L. (1978) Spermidine nitrosation and gastric cancer. Lancet 1 (8059), 324. (33) Hartman, Z., Henrikson, E. N., Hartman, P. E., and Cebula, T. A. (1994) Molecular models that may account for nitrous acid mutagenesis in organisms containing double-stranded DNA. Environ. Mol. Mutagen. 24, 168-175. (34) Chen, H., and Shaw, B. R. (1994) Bisulfite induces tandem double CC f TT mutations in double-stranded DNA. 2. Kinetics of cytosine deamination. Biochemistry 33, 4121-4129. (35) Maragos, C. M., Morley, D., Wink, D. A., Dunams, T. M., Saavedra, J. E., Hoffman, A., Bove, A. A., Isaac, L., Hrabie, J. A., and Keefer, L. K. (1991) Complexes of •NO with nucleophiles as agents for the controlled biological release of nitric oxide. Vasorelaxant effects. J. Med. Chem. 34, 3242-3247. (36) Mitchell, J. B., Wink, D. A., DeGraff, W., Gamson, J., Keefer, L. K., and Krishna, M. C. (1993) Hypoxic mammalian cell radiosensitization by nitric oxide. Cancer Res. 53, 5845-5848. (37) Wink, D. A., Darbyshire, J. F., Nims, R. W., Saavedra, J. E., and Ford, P. C. (1993) Reactions of the bioregulatory agent nitric oxide in oxygenated aqueous media: Determination of the kinetics of oxidation and nitrosation by intermediates generated in the NO/ O2 reaction. Chem. Res. Toxicol. 6, 23-27. (38) Moyer, R., Briley, D., Johnsen, A., Stewart, U., and Shaw, B. R. (1993) Echinomycin, a bis-intercalating agent, induces C f T mutations via cytosine deamination. Mutat. Res. 288, 291-300. (39) Chen, H., and Shaw, B. R. (1993) Kinetics of bisulfite-induced cytosine deamination in single-stranded DNA. Biochemistry 32, 3535-3539. (40) Lewis, R. S., Tannenbaum, S. R., and Deen, W. M. (1995) Kinetics of N-nitrosation in oxygenated nitric oxide solutions at physiological pH: Role of nitrous anhydride and effects of phosphate and chloride. J. Am. Chem. Soc. 117, 3933-3939. (41) Maron, D. M., and Ames, B. N. (1983) Revised methods for the Salmonella mutagenicity test. Mutat. Res. 113, 173-215. (42) Wink, D. A., and Laval, J. (1994) The Fpg protein, a DNA repair enzyme, is inhibited by the biomediator nitric oxide in vitro and in vivo. Carcinogenesis 15, 2125-2129. (43) Zastawny, T. H., Doetsch, P. W., and Dizdaroglu, M. (1995) A novel activity of E. coli uracil DNA N-glycosylase excision of isodialuric acid (5,6-dihydroxyuracil), a major product of oxidative DNA damage, from DNA. FEBS Lett. 364, 255-258.

TX950102G