(Acetoxymethyl)nitrosamino - American Chemical Society

American Health Foundation, 1 Dana Road, Valhalla, New York 10595. Received April 22, 1996X. Pyridyloxobutylation of DNA yields adducts that react wit...
0 downloads 0 Views 179KB Size
Chem. Res. Toxicol. 1996, 9, 949-953

949

Pyridyloxobutylation of Guanine Residues by 4-[(Acetoxymethyl)nitrosamino]-1-(3-pyridyl)-1-butanone Generates Substrates of O6-Alkylguanine-DNA Alkyltransferase Xiao-Keng Liu, Thomas E. Spratt, Sharon E. Murphy, and Lisa A. Peterson* Division of Chemical Carcinogenesis and Division of Carcinogenesis and Molecular Epidemiology, American Health Foundation, 1 Dana Road, Valhalla, New York 10595 Received April 22, 1996X

Pyridyloxobutylation of DNA yields adducts that react with O6-alkylguanine-DNA alkyltransferase (AGT) to prevent the repair of O6-methylguanine (O6-mG). The chemical characterization of pyridyloxobutyl adducts has been confounded by their instability under DNA hydrolysis conditions. They decompose to 4-hydroxy-1-(3-pyridyl)-1-butanone (HPB) during the chemical or enzymatic hydrolysis of DNA. The goal of these studies was to determine which bases are pyridyloxobutylated to form AGT-reactive adducts. The model pyridyloxobutylating agent, 4-[(acetoxymethyl)nitrosamino]-1-(3-pyridyl)-1-butanone (NNKOAc), was reacted with either poly(dAdT) or poly(dGdC) to generate DNA substrates for reaction with AGT. Only the pyridyloxobutylated poly(dGdC) was able to prevent the ability of partially purified rat liver AGT to repair O6-mG. These results paralleled those obtained for the corresponding methylated substrates. These studies are consistent with the pyridyloxobutylation of GC base pairs and not AT base pairs in the DNA to generate a substrate for AGT. In order to distinguish between the formation of reactive adducts at C residues versus G residues, two oligomers were designed that were complementary to one another. One oligomer contained A, T, and G residues, whereas its complement contained T, A, and C residues. Only the dG-containing oligomer reacted with NNKOAc to generate an AGT-reactive adduct, again paralleling the results obtained for a methylating agent. These results demonstrate that pyridyloxobutylation of only guanine residues produces adducts that react with AGT. These AGT-reactive guanine adducts are relatively stable within DNA, with a half-life of 1-2 weeks at 37 °C. They represent up to 70% of the total HPB-releasing adducts in the NNKOAc-treated DNA. We postulate that a potential AGT-reactive adduct is an O6-(pyridyloxobutyl)guanine adduct.

Introduction The tobacco-specific nitrosamine, 4-(methylnitrosamino)1-(3-pyridyl)-1-butanone (NNK),1 is a potent lung carcinogen in experimental animals (1-3). It is activated to DNA-reactive species via R-hydroxylation (1). Since NNK is an asymmetric nitrosamine, it has two activation pathways (Figure 1). Oxidation of the methylene position leads to DNA methylation (4, 5). Methyl hydroxylation, on the other hand, results in the formation of pyridyloxobutyl adducts (5, 6). In the A/J mouse lung, the formation and persistence of O6-methylguanine (O6-mG) is the critical step in NNK-induced tumor formation (5). However, the pyridyloxobutylation pathway may be important in the persistence of this promutagenic adduct. The model pyridyloxobutylating agent, 4-[(acetoxymethyl)nitrosamino]-1-(3-pyridyl)-1-butanone (NNKOAc), increased the tumorigenic activity of the methylating * To whom correspondence and reprint requests should be addressed. X Abstract published in Advance ACS Abstracts, July 15, 1996. 1 Abbreviations: AAF, 2-(acetylamino)fluorene; AGT, O6-alkylguanine-DNA alkyltransferase; AMMN, (acetoxymethyl)methylnitrosamine; BSA, bovine serum albumin; DTT, dithiothreitol; HEPES, 4-(2hydroxyethyl)-1-piperazineethanesulfonic acid; HPB, 4-hydroxy-1-(3pyridyl)-1-butanone; MNU, methylnitrosourea; [3H]MeDNA, [3HMe]methylated DNA; NNK, 4-(methylnitrosamino)-1-(3-pyridyl)-1butanone; NNKOAc, 4-[(acetoxymethyl)nitrosamino]-1-(3-pyridyl)-1butanone; O6-mG, O6-methylguanine; TCA, trichloroacetic acid; TDEGP, 20 mM Tris (pH 7.4) containing 5 mM dithiothreitol, 1 mM EDTA, 5% glycerol, and 5 mM phenylmethanesulfonyl fluoride or 0.1 mM Pefabloc SC.

S0893-228x(96)00067-7 CCC: $12.00

Figure 1. Proposed activation pathways for NNK.

agent, (acetoxymethyl)methylnitrosamine (AMMN), apparently by increasing O6-mG levels (5). The promutagenic O6-mG is repaired by O6-alkylguanine-DNA alkyltransferase (AGT) in a reaction where the methyl group is transferred from the O6-position of guanine to a cysteinyl residue at the protein’s active site (7). The methylation of this cysteine renders the protein inactive. We proposed that the increased O6-mG levels observed when NNKOAc was administered with AMMN were a © 1996 American Chemical Society

950 Chem. Res. Toxicol., Vol. 9, No. 6, 1996

result of pyridyloxobutyl adducts interfering with the repair of O6-mG via inhibition of AGT. Subsequently, calf thymus DNA treated with NNKOAc in the presence of esterase was shown to reduce the ability of rat liver AGT to react with O6-mG (8). Based on these results, a cocarcinogenic role for pyridyloxobutylation in NNKinduced lung tumorigenesis was proposed in which the pyridyloxobutyl DNA adduct(s) compete with O6-mG for reaction with AGT. This results in enhanced persistence of O6-mG, which can lead to increased tumor yield. Attempts to characterize the pyridyloxobutyl DNA adducts have been unsuccessful as a result of their instability under chemical or enzymatic hydrolysis conditions (6). The majority of the adducts are released as 4-hydroxy-1-(3-pyridyl)-1-butanone (HPB). The multiphasic decomposition of HPB-releasing DNA adducts under physiological conditions indicates that several different DNA adducts are formed (9). Previously, we observed that a subset of the thermally labile HPBreleasing adducts are capable of reacting with rat liver AGT (8). Since mammalian AGT specifically repairs O6alkylguanine residues (10), an O6-(pyridyloxobutyl)guanine derivative is a likely candidate for the AGT-reactive adduct. In order to circumvent the difficulties associated with the chemical characterization of the pyridyloxobutyl adducts, the studies described in this report were designed to gain more information about the chemical structure of the AGT-reactive pyridyloxobutyl adduct(s). The data support the involvement of a guanine pyridyloxobutyl adduct in the reaction of AGT with NNKOActreated DNA.

Experimental Procedures Warning: NNKOAc, AMMN, and AAF are mutagens as well as carcinogens in experimental animals and should be used with caution. Materials. NNKOAc (11) and [3H-Me]methylated DNA (MeDNA; 8) were prepared as previously reported. Calf thymus DNA, poly(dAdT), poly(dGdC), and porcine liver esterase were purchased from Sigma Chemical Co. (St. Louis, MO). The oligomers were custom synthesized by Oligos Etc. (Wilsonville, OR). AMMN was obtained from the NCI Chemical Carcinogen Repository (Midwest Research Institute, Kansas City, MO). [3HMe]Acetic acid was purchased from Amersham Life Science (Arlington Heights, IL). Pefabloc SC was obtained from Boehringer Mannheim (Indianapolis, IN). [3H-Me]Methylamine Hydrochloride. Polyphosphoric acid was prepared by heating 3 g of P2O5 with 3 g of H3PO4 at 90 °C overnight. One mL of this solution was added to a 10 mL flask containing [3H-Me]acetic acid (34 µmol, 100 mCi) and 3.5 mg of NaN3 (54 µmol). This solution was stirred at 45 °C for 60 h. The solution was cooled on ice, diluted with 20 mL of ice-cold H2O, and made strongly basic by the addition of solid KOH while keeping the temperature below 10 °C. The methylamine solution was then distilled into a receiving flask containing 10 mL of 0.1 N HCl which was cooled in an ice bath. This was carried out with a stream of nitrogen going from the distillation flask to the receiving flask and then to two bubblers filled with 0.1 N HCl and cooled at 0 °C. The distillate and traps were combined and evaporated in a rotary evaporator to afford [3HMe]methylamine hydrochloride (60 mCi) in a 60% yield. N-[3H-Me]Methyl-N-nitrosourea. KNCO (3.3 mg, 42 µmol) was added to [3H-Me]methylamine hydrochloride (21 µmol, 60 mCi) dissolved in 0.1 mL of H2O. H2SO4 (20 µL, 0.1 N) was added, and the flask was sealed, heated at 110 °C for 12 min, and cooled in an ice bath. The solution was acidified by the addition of 0.13 mL of 5 N H2SO4. NaNO2 (2.9 mg, 42 µmol) dissolved in 50 µL was added in five portions over 25 min. After 5 min the mixture was extracted with ether (8 × 0.5 mL) and dried over MgSO4 in the dark. The solution was filtered,

Liu et al. concentrated, and applied to a 0.5 × 10 cm column of Florisil and eluted with ether. The radioactive fractions were combined to afford a colorless product in 67% yield (40 mCi). The identity of the product was checked by TLC against standard. AGT Preparation. Rat liver AGT was obtained using previously published procedures (8, 12, 13). Briefly, male Wistar rats (200 g, Charles River, Kingston, NY) were injected with AAF (ip, 60 mg/kg) and sacrificed 48 h later (14). The livers were removed and homogenized in ice-cold 20 mM Tris (pH 7.4) containing 5 mM dithiothreitol, 1 mM EDTA, 5% glycerol, and 5 mM phenylmethanesulfonyl fluoride or 0.1 mM Pefabloc SC (TDEGP buffer) containing 0.5 M NaCl (3 mL/g of tissue). After centrifugation at 150000g for 70 min at 4 °C, the supernatant was removed and ammonium sulfate was added to a final concentration of 1 M ammonium sulfate. After centrifugation at 23000g for 30 min at 4 °C, the supernatant was applied to a phenyl Sepharose CL-4B (Pharmacia LKB, Uppsala, Sweden) column (10 cm × 16 mm) equilibrated with TDEGP buffer containing 1 M ammonium sulfate. The column was eluted with TDEGP buffer with a linear gradient from 1 to 0 M ammonium sulfate over 60 min. Active fractions were combined, dialyzed against 50 mM Tris-HCl (pH 8.3) containing 5 mM dithiothreitol, 0.1 mM EDTA, and 5% glycerol overnight, and stored at -80 °C until used. The AGT activity of this fraction was determined by measuring the number of [3H]methyl groups transferred to the protein per milligram of protein using the AGT assay described below. AGT activity of this fraction remained stable for several months. The protein concentration was measured using Coomassie Plus protein assay reagent (Pierce, Rockford, IL). Alkylated DNA Substrates. Calf thymus DNA, poly(dAdT), or poly(dGdC) (2 mg/mL) was incubated with NNKOAc (0-5 mM) or AMMN (0.01, 1, or 5 mM) and esterase (0.022 mg/mL) in 15 mM sodium citrate (pH 7) containing 1 mM EDTA at room temperature. After 1 h, 2 M NaCl was added (final concentration, 0.33 M), and the DNA was precipitated with ethanol, washed extensively with ethanol, and dried under a stream of N2. Control DNA substrates were similarly prepared but in the absence of alkylating agent. The NNKOAc-treated calf thymus DNA and control calf thymus DNA were then redissolved in 100 mM sodium phosphate buffer (pH 7.4) containing 1 mM EDTA (3 mg/mL). A portion of this DNA was incubated at room temperature or 37 °C for 8 days or 2 weeks. The DNA that was incubated for 8 days was extracted with CHCl3-isoamyl alcohol (24:1) and then precipitated with ice-cold ethanol following the addition of 1/5 volume 2 M NaCl. The DNA that was incubated for 2 weeks was directly precipitated from the solution without any organic extraction. The control and NNKOAc-treated calf thymus DNA samples were redissolved in deionized water (approximately 1.5 mg/mL) and dialyzed overnight against 2 × 2 L of water at 4 °C. The DNA was then lyophilized and stored at -20 °C. HPB-releasing adduct levels in the DNA were determined using the established GC-MS assay (15). The stability of the AGT-reactive adducts was tested by dissolving the alkylated DNA in 50 mM HEPES buffer (pH 7.8) containing 1 mM EDTA and heating it at 37, 40, or 56 °C for 1 h. The mixtures were allowed to cool slowly to room temperature and used in the AGT assay as described below. Levels of AGT-reactive adducts were determined by using various amounts of alkylated DNA in the AGT assay described below. The reaction between NNKOAc-treated DNA and AGT was complete within the 30 min preincubation period. The degree of AGT inhibited was determined by subtracting the amount of [3H]methyl transferred to the protein in the presence of the alkylated substrate from the amount of [3H]methyl transferred in the full activity control performed with esterasetreated DNA. This was then divided by the amount of alkylated DNA used in the incubation to determine the pmol of [3H]methyl transfer inhibition per mg of alkylated DNA.

Pyridyloxobutyl dG Adducts Are AGT Substrates Effects of DNA Substrates on AGT Activity. Alkylated DNA substrates (0-100 µg) were incubated with rat liver AGT (0.5-1 pmol) for 30 min at 37 °C prior to the addition of [3H]MeDNA (sp act. 2900 mCi/mmol) containing 1.2 or 1.7 pmol of O6-[3H-Me]mG in 50 mM HEPES buffer (pH 7.8) containing 1 mM EDTA and 1 mM DTT (total volume: 1 mL). Then the incubations were continued for another 30 min. Blanks in which BSA was substituted for AGT or 1 mM AMMN-treated DNA used as the DNA substrate were run as no activity controls in each experiment. Each experiment was run in triplicate. Esterase-treated calf thymus DNA was used to bring the total amount of DNA in each sample to 100 µg. The reactions were stopped by the addition of BSA (0.6 mg) and 50% TCA (200 µL). The mixture was heated at 80 °C for 30 min to release purines (including unreacted O6-mG) from the DNA substrate. The precipitate, containing apurinic DNA and [3H]methylated AGT, was pelleted by centrifugation for 30 min. The supernatant was removed, and the pellet was washed twice with cold 5% TCA. The pellet was dissolved in 0.3 mL of 0.1 N NaOH and transferred to a scintillation vial. After combination with 0.3 mL of water and 0.6 mL of 0.2 M Tris-HCl washes, scintillation fluid (Picofluor; Packard, Meriden, CT) was added (10 mL) and the amount of [3H]methyl transferred to the protein was determined by scintillation counting. Oligomer Experiment. The 19mers (2 mg/mL) were reacted individually with 5 mM NNKOAc or 5 mM AMMN in the presence of esterase (0.022 mg/mL) for 1 h in 15 mM sodium citrate (pH 7) containing 1 mM EDTA at room temperature. Control oligomers were prepared in the absence of alkylating agent. The oligomers were then loaded onto a Sephadex G15 (Sigma Chemical Co., St. Louis, MO) column (30 cm × 1.5 cm) equilibrated in 10 mM Tris (pH 7.4) containing 0.1 mM EDTA and eluted with the same buffer. The oligomer eluted in the void volume whereas the hydrolysate products of NNKOAc eluted in 3 times the void volume. The fractions containing DNA were combined and concentrated by lyophilization. Each oligomer was combined with its complement in 50 mM HEPES buffer (pH 7.8) containing 1 mM EDTA the day prior to the experiment and stored overnight at 4 °C. The annealed oligomers (20 µg) were incubated with AGT (0.6 pmol, 174 µg of protein) in 50 mM HEPES buffer (pH 7.8) containing 1 mM DTT and 1 mM EDTA for 30 min at 37 °C prior to the addition of [3H]MeDNA containing 1.2 pmol of O6-[3H-Me]mG (total volume: 1 mL). Incubations were stopped 30 min later by the addition of BSA (0.6 mg) and 50% TCA (200 µL). The amount of [3H]methyl transferred was determined as described above.

Results and Discussion Previously, we reported that 2-4% of the total HPBreleasing adducts were responsible for AGT inactivation by NNKOAc-treated DNA (8). These data were obtained using estimated levels of total HPB-releasing adducts. In order to more clearly define the percentage of the total HPB-releasing adducts that inhibit the ability of AGT to repair O6-mG, the levels of HPB-releasing adducts were measured directly in the NNKOAc-treated DNA used in the in vitro AGT assay using the established GCMS assay (15). The levels of AGT-reactive adducts were determined by the in vitro AGT assay. The amount of AGT-reactive adducts did not significantly change between alkylation experiments I and II (8.0 and 7.0 pmol/ y; Table 1). However, a large difference in the total HPBreleasing adduct levels was observed in these two experiments, consistent with the reported instability of these adducts (9). The difference in stability between the total HPBreleasing adducts and the AGT-reactive pyridyloxobutyl adducts allowed us to prepare DNA with varying percentages of AGT-reactive adducts. Incubation of 2.5 mM NNKOAc-treated DNA for 1 or 2 weeks at room temper-

Chem. Res. Toxicol., Vol. 9, No. 6, 1996 951 Table 1. Levels of AGT-Reactive Adducts in NNKOAc-Treated DNA

DNAa

incubation conditions

total HPB-releasing adductsb (pmol/mg of DNA)

AGT-reactive adductsc (pmol/mg of DNA)

% reactive adducts

I I II II II

none 2 wks, 37 °C none 8 days, RT 8 days, 37 °C

26.8 5.1 108 15.0 8.9

8.0 ( 1.2 3.6 ( 0.9 7.0 ( 0.9 7.2 ( 0.8 3.5 ( 1.0

30 71 6 48 39

a 2.5 mM NNKOAc-treated DNA was generated in two different experiments (I and II). These DNA substrates were incubated under the specified conditions and dialyzed overnight against water prior to the measurement of adduct levels. b HPB-releasing adduct levels were measured using a previously described GCMS method (15). c AGT-reactive adducts were determined by incubating NNKOAc-treated DNA (0-100 µg) with AGT (0.5-0.9 pmol) for 30 min prior to the addition of [3H]MeDNA containing 1.2 or 1.7 pmol of O6-[3H-Me]mG.

ature or 37 °C decreased substantially the total levels of HPB-releasing adducts observed by GC-MS (Table 1). The AGT reactivity of this DNA was not affected by incubation at room temperature. The half-life of the AGT-reactive adduct(s) at 37 °C was 1-2 weeks. These results are consistent with the presence of a relatively stable pyridyloxobutyl adduct in DNA that is responsible for the reactivity toward AGT. The effect of temperature on the AGT reactivity of the DNA was determined by incubating 2.5 mM NNKOActreated DNA for 1 h at room temperature, 37, 40, and 56 °C prior to incubation with AGT in the in vitro assay. The pyridyloxobutyl adducts responsible for the inhibition of [3H]methyl group transfer are stable in the DNA up to 56 °C for 1 h. Previously, the majority of the pyridyloxobutylated DNA that was AGT-reactive could be removed from the DNA by heating at 100 °C for 30 min (8), conditions that also release the majority of pyridyloxobutyl DNA adducts as HPB (6). In order to determine which base pair is involved in the AGT reactivity of NNKOAc-treated DNA, poly(dAdT) and poly(dGdC) were reacted with NNKOAc in the presence of esterase. In reactivity studies, NNKOAc was found to pyridyloxobutylate each of the residues in the following order: dC > dG > dT = dA.2 As positive controls, poly(dAdT) and poly(dGdC) were reacted with AMMN in the presence of esterase. In this case, the major AGT-reactive adduct formed by AMMN is O6-mG (10). Therefore, only AMMN-treated poly(dGdC), not AMMN-treated poly(dAdT), is expected to contain AGTreactive properties. Following isolation, these DNA substrates were incubated with the partially purified rat liver AGT prior to the addition of [3H]MeDNA. Only the poly(dGdC) substrates generated by these two alkylating agents dramatically interfered with the transfer of [3H]methyl groups to the protein (Table 2). There was a slight inhibition of methyl transfer in the AMMN-treated poly(dAdT)-treated DNA. Consistently, mammalian AGT can repair O4-methylthymidine, albeit at a slow rate (16, 17). Since NNKOAc-treated poly(dGdC) but not NNKOAc-treated poly(dAdT) is able to interfere with the ability of AGT to repair O6-[3H-Me]mG, the AGT-reactive pyridyloxobutyl adducts are either guanine or cytosine adducts. To determine whether pyridyloxobutylation of dG or dC residues yields AGT-reactive adducts, NNKOAc was 2

T. E. Spratt and S. S. Hecht, unpublished results.

952 Chem. Res. Toxicol., Vol. 9, No. 6, 1996

Liu et al.

Table 2. Inhibitory Activity of Poly(dAdT) or Poly(dGdC) Treated with NNKOAc or AMMN relative AGT act.b (%)

DNA treatmenta

poly(dAdT)

poly(dGdC)

control 5 mM NNKOAc 5 mM AMMN

100 98 85

100 0 0

a Poly(dAdT) and poly(dGdC) were incubated with esterase in the presence or absence of 5 mM NNKOAc or 5 mM AMMN for 1 h in sodium citrate (pH 7) containing 1 mM EDTA. b Treated DNA (50 µg) was incubated with AGT (0.7-1 pmol) 30 min prior to the addition of [3H]MeDNA containing 1.7 pmol of O6-mG. The results are expressed as the percent of the full activity control and are an average of two experiments (varied by less than 6%).

Table 3. Inhibitory Activity of dG- or dC-Containing Oligomers Treated with NNKOAc or AMMNa

oligomersa

AGT act. remainingb (pmol)

[3H]methyl transfer inhibited (%)

dG:dC AMMN-dG:dC dG:AMMN-dC NNKOAc-dG:dC dG:NNKOAc-dC

0.64 ( 0.04 (3)c 0.03 (2) 0.57 ( 0.07 (3) 0.10 ( 0.01 (3) 0.57 ( 0.05 (3)

0 95 11 84 11

a The dG oligomer (5′ AGAGTGGGAATGGGAGAT) and the dC oligomer (3′ TCTCACCCTTACCCTCTA) were incubated with esterase in the presence or absence of 5 mM NNKOAc or 5 mM AMMN for 1 h in 15 mM sodium citrate (pH 7) containing 1 mM EDTA at room temperature. b Oligomers (20 µg) were incubated with AGT (0.6 pmol, 174 µg of protein) for 30 min at 37 °C prior to the addition of [3H]Me-DNA containing 1.2 pmol of O6-mG. Incubations were stopped 30 min later by the addition of BSA (0.6 mg) and 50% TCA (200 µL). c The numbers in parentheses represent the number of incubations.

incubated with oligomers containing only dG or dC in the presence of esterase, and these substrates were used in our in vitro assay. Again, AMMN-treated oligomers were used as positive controls. Initially, NNKOAc or AMMN was reacted with poly(dG)10-14 and poly(dC)10-14. Even the methylated substrates did not result in inhibition of methyl transfer, presumably a result of tetraplex formation by poly(dG)10-14 (18, 19). To avoid this possibility, we designed two complementary oligomers, one containing dG, dT, and dA and the other dC, dT, and dA with no more than three dG or dC in a row:

These oligomers were alkylated with either NNKOAc or AMMN in the presence of esterase and isolated by Sephadex G15 chromatography. Then each oligomer was annealed with its complement that had only been incubated with esterase. These DNA substrates were then reacted with AGT for 30 min prior to the addition of [3H]MeDNA to determine residual AGT activity. The oligomer containing dG residues reacted with NNKOAc or AMMN to generate an AGT-reactive adduct (Table 3). A small but not significant inhibition was observed with the alkylated dC oligomer. This oligomer did not lose its ability to inhibit the transfer of [3H]methyl groups to DNA when heated at 100 °C for 1 h (data not shown). The difference in stability of the AGT-reactive nature of the alkylated oligomer versus alkylated DNA is unknown at this time. These data are consistent with the proposal that pyridyloxobutylation of guanine residues generates AGT-reactive adducts.

Figure 2. Proposed mechanisms for AGT inactivation by pyridyloxobutyl DNA adducts.

There are at least three possible mechanisms by which pyridyloxobutylated DNA can interfere with the repair of O6-mG by AGT. The first possibility involves the reaction of AGT with a specific pyridyloxobutyl DNA adduct, such as an O6-(pyridyloxobutyl)guanine adduct. This adduct would inactivate AGT as a result of transfer of the alkyl group to active site of AGT, preventing the repair of O6-mG. A second option involves the protein becoming tightly bound to an adduct, so that the protein is not irreversibly inactivated but still unable to repair O6-mG. A third potential mechanism is that these bulky adducts could inhibit the repair activity of AGT via a nonspecific transfer of the alkyl group from DNA to AGT, as a result of adduct decomposition to a reactive electrophile (oxonium ion) that then alkylates the protein. This mechanism does not require a specific adduct and might result in nonspecific binding of the alkyl group to AGT, reacting with amino acid residues different than the active site cysteine. Our data support the involvement of a specific adduct in the inactivation of AGT. First, not all of the HPBreleasing adducts are reactive toward AGT (Table 1; 8). Second, the AGT-reactive pyridyloxobutyl adduct(s) are more stable at 37 °C than the majority of HPB-releasing adducts (Table 1). The half-life of the reactive adduct is 1-2 weeks at 37 °C. Furthermore, only alkylation of guanine residues yields AGT-reactive adducts (Table 3). Since AGT primarily reacts with O6-alkylguanine residues in DNA (10), we believe that an O6-(pyridyloxobutyl)guanine adduct is the AGT-reactive adduct (Figure 2). NNKOAc hydrolyzes to 4-oxo-4-(3-pyridyl)-1-butanediazonium ion that can react directly with nucleophiles to form straight chain adducts or indirectly via a cyclic oxonium ion to form cyclic adducts (11, 20). Therefore, there are two possible structures for an O6-(pyridyloxobutyl)guanine derivative, an open chain adduct, 4-(O6guanyl)-1-(3-pyridyl)-1-butanone (1, Figure 2), and a cyclic adduct, 2-(O6-guanyl)-2-(3-pyridyl)-2,3,4,5-tetrahydrofuran (2, Figure 2). The relative stability of the AGTreactive adduct in the DNA argues for an open chain adduct as the most likely candidate for the AGT-reactive adduct since the cyclic derivative would be substantially less stable. This is supported by the lower stability of the corresponding cysteinyl adduct (20). Since the mechanism by which AGT repairs O6alkylguanine residues is believed to involve an SN2 reaction between the alkyl group and a cysteine residue at AGT’s active site, the reaction rate decreases with increasing alkyl chain length. Consistently, studies using rat liver AGT have established that the relative rate of repair of O6-alkylguanine residues in DNA by AGT is

Pyridyloxobutyl dG Adducts Are AGT Substrates

methyl > ethyl, n-propyl > n-butyl > isopropyl, isobutyl > 2-hydroxyethyl (21). An exception to this trend is O6benzylguanine which is a better substrate for AGT than O6-mG (22). This adduct inactivates AGT by the transfer of the benzyl group (23). If an O6-(pyridyloxobutyl)guanine derivative is responsible for the inactivation of AGT’s ability to repair O6-mG, the transfer of this bulky adduct to the active site cysteine of AGT may be facilitated by the neighboring group participation of the carbonyl oxygen in the case of the open chain adduct or the tetrahydrofuran oxygen of the cyclic adduct. Experiments are ongoing in our laboratory to understand the mechanism by which pyridyloxobutyl adducts interfere with the repair of O6-mG by AGT.

Acknowledgment. We would like to thank Ms. Yin Liu and Mr. Stuart Coleman for the GC-MS analyses for HPB-releasing adducts and the AHF Research Animal Facility for the treatment of the animals and tissue isolation. The Research Animal Facility is partially supported by National Cancer Institute Cancer Center Support Grant CA-17613. This study was supported by Grant CA-59887 from the National Cancer Institute.

References (1) Hecht, S. S., and Hoffmann, D. (1988) Tobacco-specific nitrosamines: an important group of carcinogens in tobacco and tobacco smoke. Carcinogenesis 9, 875-884. (2) Hecht, S. S., Morse, M. A., Amin, S. G., Stoner, G. D., Jordan, K. G., Choi, C.-I., and Chung, F.-L. (1989) Rapid single dose model for lung tumor induction in A/J mice by 4-(methylnitrosamino)1-(3-pyridyl)-1-butanone and the effect of diet. Carcinogenesis 10, 1901-1904. (3) Hecht, S. S., Chen, C. B., Ohmori, T., and Hoffmann, D. (1980) Comparative carcinogenicity in F344 rats of the tobacco specific nitrosamines, N′-nitrosonornicotine and 4-(N-methyl-N-nitrosamino)-1-(3-pyridyl)-1-butanone. Cancer Res. 40, 298-302. (4) Hecht, S. S., Trushin, N., Castonguay, A., and Rivenson, A. (1986) Comparative tumorigenicity and DNA methylation in F344 rats by 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone and N-nitrosodimethylamine. Cancer Res. 46, 498-502. (5) Peterson, L. A., and Hecht, S. S. (1991) O6-Methylguanine is a critical determinant of 4-(methylnitrosamino)-1-(3-pyridyl)-1butanone tumorigenesis in A/J mouse lung. Cancer Res. 51, 55575564. (6) Hecht, S. S., Spratt, T. E., and Trushin, N. (1988) Evidence for 4-(3-pyridyl)-4-oxobutylation of DNA in F344 rats treated with the tobacco specific nitrosamines 4-(methylnitrosamino)-1-(3pyridyl)-1-butanone and N′-nitrosonornicotine. Carcinogenesis 9, 161-165. (7) Pegg, A. E. (1990) Mammalian O6-alkylguanine-DNA alkyltransferase: Regulation and importance in response to alkylating carcinogenic and therapeutic agents. Cancer Res. 50, 6119-6129.

Chem. Res. Toxicol., Vol. 9, No. 6, 1996 953 (8) Peterson, L. A., Liu, X.-K., and Hecht, S. S. (1993) Pyridyloxobutyl DNA adducts inhibit the repair of O6-methylguanine. Cancer Res. 53, 2780-2785. (9) Peterson, L. A., Mathew, R., Murphy, S. E., Trushin, N., and Hecht, S. S. (1991) In vivo and in vitro persistence of pyridyloxobutyl DNA adducts from 4-(methylnitrosamino)-1-(3-pyridyl)-1butanone. Carcinogenesis 12, 2069-2072. (10) Pegg, A. E. (1990) Properties of mammalian O6-alkylguanineDNA transferases. Mutat. Res. 233, 165-175. (11) Spratt, T. E., Peterson, L. A., Confer, W. L., and Hecht, S. S. (1990) Solvolysis of model compounds for R-hydroxylation of N′-nitrosonornicotine and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone: Evidence for a cyclic oxonium ion intermediate in the alkylation of nucleophiles. Chem. Res. Toxicol. 3, 350-356. (12) Chan, C. L., Wu, Z., Eastman, A., and Bresnick, E. (1990) Induction and purification of O6-methylguanine-DNA-methyltransferase from rat liver. Carcinogenesis 11, 1217-1221. (13) Pegg, A. E., Wiest, L., Foote, R. S., Mitra, S., and Perry, W. (1983) Purification and properties of O6-methylguanine-DNA transmethylase from rat liver. J. Biol. Chem. 258, 2327-2333. (14) Wilkinson, M. C., Cooper, D. P., Southan, C., Potter, P. M., and Margison, G. P. (1990) Purification to apparent homogeneity and partial amino acid sequence of rat liver O6-alkylguanine-DNAalkyltransferase. Nucleic Acids Res. 18, 13-16. (15) Foiles, P. G., Akerkar, S. A., Carmella, S. G., Kagan, M., Stoner, G. D., Resau, J. H., and Hecht, S. S. (1991) Mass spectrometric analysis of tobacco-specific nitrosamine-DNA adducts in smokers and nonsmokers. Chem. Res. Toxicol. 4, 364-368. (16) Becker, R. A., and Montesano, R. (1985) Repair of O4-methyldeoxythymidine residues in DNA by mammalian liver extracts. Carcinogenesis 6, 313-317. (17) O’Toole, S. M., Pegg, A. E., and Swenberg, J. A. (1993) Repair of O6-methylguanine and O4-methylthymidine in F344 rat liver following treatment with 1,2,-dimethylhydrazine and O6-benzylguanine. Cancer Res. 53, 3895-3898. (18) Sen, D., and Gilbert, W. (1988) Formation of parallel four-stranded complexes by guanine-rich motifs in DNA and its implications for meiosis. Nature 334, 364-366. (19) Kim, J., Cheong, C., and Moore, P. B. (1991) Tetramerization of an RNA oligonucleotide containing a GGGG sequence. Nature 351, 331-332. (20) Carmella, S. G., Kagan, S. S., Spratt, T. E., and Hecht, S. S. (1990) Evaluation of cysteine adduct formation in rat hemoglobin by 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone and related compounds. Cancer Res. 50, 5453-5459. (21) Chae, M.-Y., McDougall, M. G., Dolan, M. E., Swenn, K., Pegg, A. E., and Moschel, R. C. (1994) Substituted O6-benzylguanine derivatives and their inactivation of human O6-alkylguanineDNA alkyltransferase. J. Med. Chem. 37, 342-347. (22) Dolan, M. E., Moschel, R. C., and Pegg, A. E. (1990) Depletion of mammalian O6-alkylguanine-DNA alkyltransferase activity by O6-benzylguanine provides a means to evaluate the role of this protein in protection against carcinogenic and therapeutic alkylating agents. Proc. Natl. Acad. Sci. U.S.A. 87, 5368-5372. (23) Pegg, A. E., Boosalis, M., Samson, L., Moschel, R. C., Byers, T. L., Swenn, K., and Dolan, M. E. (1993) Mechanisms of inactivation of human O6-alkylguanine-DNA alkyltransferase by O6-benzylguanine. Biochemisty 32, 11998-12006.

TX960067T