Reaction of N-(2-Chloroethyl)-N-nitrosoureas with ... - ACS Publications

N-(2-Chloroethyl)nitrosoureas (CNU) are clinically used anticancer drugs whose cytotoxicity is associated with the generation of DNA interstrand cross...
0 downloads 0 Views 338KB Size
Download PDF
208

Chem. Res. Toxicol. 1996, 9, 208-214

Reaction of N-(2-Chloroethyl)-N-nitrosoureas with DNA: Effect of Buffers on DNA Adduction, Cross-Linking, and Cytotoxicity Fa-Xian Chen,† William J. Bodell,‡ Gangning Liang,†,§ and Barry Gold*,†,§ Eppley Institute for Research in Cancer and Department of Pharmaceutical Sciences, University of Nebraska Medical Center, Omaha, Nebraska 68198-6805, and Brain Tumor Research Center, Department of Neurological Surgery, University of California, San Francisco, San Francisco, California 94143-0806 Received June 7, 1995X

N-(2-Chloroethyl)nitrosoureas (CNU) are clinically used anticancer drugs whose cytotoxicity is associated with the generation of DNA interstrand cross-links. While studying the sequence selectivity for a series of CNU, a dramatic increase in the formation of N7-alkyldeoxyguanosine was observed when Tris buffer was used rather than phosphate or cacodylate buffers. Moreover, the formation of N7-alkyldeoxyguanine lesions continues in Tris long after all of the CNU has hydrolyzed. These effects are not seen with the monofunctional alkylating analogues, e.g., N-methyl- and N-(2-hydroxyethyl)-N-nitrosourea. In order to determine if the nature of the CNU-mediated DNA damage was altered by Tris, studies were initiated on the following: (1) alkylation of N7-G in end-labeled DNA restriction fragments; (2) covalent modification of DNA with [ethyl-3H]-N-(2-chloroethyl)-N-nitrosourea; and (3) cytotoxicity in L1210 cells. The data presented demonstrate that Tris increases the yield of the “normal” CNU monofunctional crosslinked adducts, i.e., N7-(2-hydroxyethyl)deoxyguanosine, N7-(2-chloroethyl)deoxyguanosine, O6-(2-chloroethyl)deoxyguanosine, and bifunctional adducts, i.e., 1-(deoxycytid-3-yl)-2-(deoxyguanosin-1-yl)ethane and 1,2-bis(deoxyguanosin-7-yl)ethane. In addition, CNU appears to react with Tris to give a long-lived alkylating intermediate that affords large amounts of DNA adducts not seen with CNU in the absence of Tris. However, in vivo toxicity of CNU in L1210 cells is not affected by the presence of Tris, indicating that the reaction pathway(s) responsible for cross-linking is not significantly sensitive to the nature of the buffer.

Introduction 1

N-(2-Chloroethyl)-N-nitrosoureas (CNU; see Figure 1 for structures) are clinically useful antineoplastic agents employed in the treatment of brain tumors (1, 2). It is generally accepted that CNU exert their cytotoxicity by the generation of DNA interstrand cross-links that are initiated by the formation of an O6-(2-chloroethyl)deoxyguanosine (O6-ClEt-dG) lesion (Figure 2) (3-10). The O6-ClEt-dG then cyclizes to N1,O6-ethanodeoxyguanosine, which in turn undergoes nucleophilic attack by the N3-position of the cytosine on the complementary strand. Although the O6-ClEt-dG adduct, as the precursor of the 1-(2′-deoxycytid-3-yl)-2-(2′-deoxyguanosin-1-yl)ethane (dC-dG) interstrand cross-link, is central to the * To whom all correspondence should be addressed (FAX, 402-5594651; Email address, [email protected]). † Eppley Institute for Research in Cancer, University of Nebraska Medical Center. ‡ University of California, San Francisco. § Department of Pharmaceutical Sciences, University of Nebraska Medical Center. X Abstract published in Advance ACS Abstracts, December 1, 1995. 1 Abbreviations: ANU, N-alkyl-N-nitrosourea; N7-bisdG, 1,2-bis(deoxyguanosin-7-yl)ethane; BCNU, N,N′-bis(2-chloroethyl)-Nnitrosourea; CCNU, N-(2-chloroethyl)-N′-cyclohexyl-N-nitrosourea; ClENU, N-(2-chloroethyl)-N-nitrosourea; CNU, N-(2-chloroethyl)-N′alkyl-N-nitrosourea; dC-dG, 1-(2′-deoxycytid-3-yl)-2-(2′-deoxyguanosin1-yl)ethane; d-s, double-stranded; HO-ENU, N-(2-hydroxyethyl)-Nnitrosourea; MCNU, N-methyl-N′-cyclohexyl-N-nitrosourea; MNU, N-methyl-N-nitrosourea; N7-alk-dG, N7-alkyldeoxyguanosine; N7-ClEt-dG, N7-(2-chloroethyl)deoxyguanosine; N7-Me-dG, N7-methyldeoxyguanosine; O6-alk-dG, O6-alkyldeoxyguanosine; O6-Cl-Et-dG, O6-(2chloroethyl)deoxyguanosine; O6-HO-Et-dG, O6-(2-hydroxyethyl)deoxyguanosine; N1-HO-Et-dG, N1-(2-hydroxyethyl)deoxyguanosine; N7-HO-Et-dG, N7-(2-hydroxyethyl)deoxyguanosine; s-s, single-stranded; Tris, tris(hydroxymethyl)aminomethane.

0893-228x/96/2709-0208$12.00/0

toxicity of CNU, it is quantitatively a minor product (3%); the other major lesions (and their % yields) are N7-(2hydroxyethyl)deoxyguanosine (N7-HO-Et-dG) (36%), N7(2-chloroethyl)deoxyguanosine (N7-Cl-Et-dG) (15%), 1,2bis(deoxyguanosin-7-yl)ethane (N7-bisG) (3%), N1-(2hydroxyethyl)deoxyguanosine (N1-HO-Et-dG) (3%), O6(2-hydroxyethyl)deoxyguanosine (O6-HO-Et-dG) (2%), and phosphotriesters (26%) (11). In the process of studying the sequence selectivity for CNU, we noticed significant differences in the yield and the time course for the formation of N7-alkyldeoxyguanosine (N7-alk-dG) lesions in different buffers (12). There have been previous reports that the rate of CNU degradation is slower in tris(hydroxymethyl)aminomethane (Tris) than in other buffers (13-17), and based on kinetic studies, it has been proposed that the hydrolysis of CNU in Tris buffer occurs with the formation a stable complex (16, 17). Importantly, it was suggested that the CNU-Tris complex could be used to stabilize CNU for clinical use (17). In order to understand the influence of Tris on the rate and type of DNA damage induced by the CNU, we initiated in vitro DNA alkylation and in vivo cytotoxicity studies.

Experimental Section All chemical reagents were purchased from Aldrich Chemical (Milwaukee, WI) or Sigma Chemicals (St. Louis, MO) and used without any further purification. Enzymes and restriction endonucleases were obtained from New England Biolabs (Beverly, MA) or Bethesda Research Laboratory (Bethesda, MD). The

© 1996 American Chemical Society

CNU-Tris

Chem. Res. Toxicol., Vol. 9, No. 1, 1996 209

Figure 1. Structures of CNU compounds, buffers, and solutes.

Figure 2. Hydrolysis scheme for CNU and the pathway for formation of the dC-dG interstrand cross-link. [ethyl-3H]Cl-ENU was custom synthesized by Movarek Biochemicals (Brea, CA). Caution: All nitrosoureas are considered toxic and carcinogenic and must be handled with the appropriate safety precautions. Sequencing Reactions with 32P-End-Labeled DNA Fragments. An 85 base pair 5′-[32P]-labeled DNA restriction fragment was prepared as previously described (18) from a 3220 base pair DNA clone containing the promotor region of the coat protein gene of the canine parvovirus (19) by initial endonuclease restriction with NcoI, followed by sequential dephosphorylation with calf intestine alkaline phosphatase, phosphorylation with T4 kinase in the presence of [γ-32P]ATP, and HindIII digestion. The labeled fragment was purified by gel electrophoresis on a 5% polyacrylamide gel and isolated by electroelution. The reactions were performed as follows: the restriction fragment (80 000-100 000 cpm) and sonicated calf thymus DNA (final concentration, 100 µM nucleotide) were dissolved in 10 mM buffer (buffer and pH defined in figure legends) containing

the desired concentration of solute. This DNA solution was incubated at 37 °C with a freshly prepared solution of CNU or N-alkyl-N-nitrosourea (ANU) for varying lengths of time. When noted, the CNU or ANU was incubated in the buffer in the absence of DNA for a specified time and then the DNA added and the incubation continued. In control experiments, no alkylating agent was added. The reactions were terminated by cooling in ice and precipitation of the DNA with NaOAc and EtOH. The DNA was washed with cold 70% EtOH and dried in vacuo. Strand breaks in the reacted DNA were generated by treatment with 1 M piperidine at 90 °C for 25 min to preferentially convert N7-alk-dG into s-s breaks (20). The piperidine was removed in vacuo. The dried DNA was then suspended in loading buffer and denatured at 90 °C for 1.5 min and cooled in ice. The DNA was electrophoresed using 12% polyacrylamide (7.8 M urea) denaturing gels run at 65 W. The standard G and/or G+A reaction lanes (20) were included as sequence markers. Either the gels were exposed to Kodak X-OMAT AR film at -70 °C and the resulting autoradiogram

210 Chem. Res. Toxicol., Vol. 9, No. 1, 1996

Figure 3. Effect of 10 mM buffer (pH 7.8) and preincubation times on the alkylation of the 5′-[32P]-85 bp restriction fragment at N7-G by 1 mM CCNU: lane a, G; lane b, G+A; lanes c-h (Tris), i-n (cacodylate), o-t (triethanolamine), and u-z (phosphate), preincubation of CCNU in buffer prior to adding endlabeled DNA for 0, 25, 50, 75, 100, and 150 min, respectively for each buffer; C, control lane (no CNU). was quantitated using a Shimadzu CS-9000 scanning densitometer, or they were analyzed on a Molecular Dynamics PhosphorImager. DNA Alkylation Studies. The alkylation of calf thymus DNA (250 µg) by [ethyl-3H]-N-(2-chloroethyl)-N-nitrosourea (ClENU) (250 µCi, specific activity 5.8 Ci/mmol) was performed at 37 °C in a final volume of 1 mL in either 10 mM Tris, 1 mM EDTA (pH 7.6), or 10 mM sodium cacodylate buffer, 1 mM EDTA (pH 7.6). The Cl-ENU was preincubated in the buffer for 0, 15, 30, or 45 min prior to adding the DNA. After continuing the incubation for 6 h, the DNA was repeatedly precipitated with EtOH until a constant specific activity was obtained. The DNA was then digested and analyzed by HPLC and scintillation counting as previously described (11). Cell Survival Studies. The toxicity of N-(2-chloroethyl)N′-cyclohexyl-N-nitrosourea (CCNU) in the presence and absence of Tris was determined using L1210 cells. Incubations were performed using 123 000 cells/T25 flask containing 8 mL of 5% fetal bovine serum medium. Cells were incubated at 37 °C for 48 h with 20 µL of MeOH solvent, CCNU in 20 µL of MeOH, or CCNU in 20 µL of MeOH and 80 µL of 1 M Tris (pH 7.2). Afterward, 4 mL of 5% fetal bovine serum medium and 120 µL of HEPES buffer (pH 10.0) were added and the incubation was continued for an additional 24 h at 37 °C. In experiments requiring preincubation, the CCNU, or CCNU with Tris, and medium were kept at 37 °C for 2 h before being added to the cells. In all cases the number of cells was determined using a Coulter counter at the end of the incubation. The data are presented as percent relative to solvent treated cells.

Results Sequencing Gels. Figure 3 shows the results obtained from the incubation of 1 mM CCNU in Tris (lanes c-h), cacodylate (lanes i-n), triethanolamine (lanes o-t), or phosphate buffer (lanes u-z) for 0, 25, 50, 75, 100, and 150 min prior to adding the 32P-end-labeled DNA. After the specified preincubation period, all reactions were then continued for 4 h in the presence of the endlabeled DNA. The results for all buffers except Tris show a time-dependent decrease in cleavage at G with longer preincubation periods. Analysis of the intensities of these bands indicates that the CCNU hydrolyzes in the ca-

Chen et al.

codylate buffer with a t1/2 of 1 mM) of potential CCNU hydrolysis products, including cyclohexyl isocyanate, ethylene oxide, 2-chloroethylamine, and 2-chloroethanol, did not yield DNA strand breaks using the same incubation conditions (data not shown). DNA Adducts. [ethyl-3H]Cl-ENU was preincubated with calf thymus DNA in 10 mM Tris (pH 7.6) or 10 mM cacodylate buffer (pH 7.6) for 0, 15, 30, or 45 min before the DNA was added. The incubations were then continued for an additional 6 h. The levels of DNA alkylation are given in Table 1. With cacodylate buffer, there was a significant decrease in the extent of DNA alkylation with preincubation. Comparison of the levels with no preincubation with those found after a 15 min preincubation demonstrates that this preincubation decreased the extent of DNA alkylation by 86%. Preincubation of [ethyl-3H]Cl-ENU in cacodylate buffer for longer times produced similar results. The influence of Tris buffer on the extent of DNA alkylation by [ethyl-3H]Cl-ENU products was also examined. Pairwise analysis of the levels of DNA alkylation in Tris and cacodylate buffers with no preincubation shows that the level of DNA modification in Tris is 2.8-fold higher than that in cacodylate buffer. Preincubation of [ethyl-3H]Cl-ENU in Tris buffer for 15 min resulted in a 40% reduction in the level of

Chem. Res. Toxicol., Vol. 9, No. 1, 1996 211

Figure 5. Effect of preincubation times and different solutes (10 mM) on DNA alkylation of the 5′-[32P]-85 bp restriction fragment at N7-G by 1 mM CCNU at pH 8.0: lane 1, G; lane 2, G+A; lane 3, control; lanes 4-31, 1 mM CCNU; lanes 4-7 (10 mM cacodylate buffer), lanes 8-11 (10 mM Tris buffer), lanes 12-15 (Tris in cacodylate buffer), lanes 16-19 (tricine in cacodylate buffer), lanes 20-23 (serinol in cacodylate buffer), lanes 24-27 (ethanolamine in cacodylate buffer), and lanes 2831 (3-amino-1,2-propanediol in cacodylate buffer), preincubation of CCNU in buffer prior to adding end-labeled DNA for 0, 25, 50, 75, and 150 min, respectively, for each buffer set.

DNA alkylation as compared with no preincubation. Similar results were obtained after longer periods of preincubation in Tris. Interestingly, the level of DNA alkylation found in Tris buffer with no preincubation is approximately equal to the sum of the levels of adducts formed in cacodylate with no preincubation and in Tris with 15 min preincubation. The influence of buffer on the formation of the individual DNA alkylation products was examined by enzymatic digestion of the radiolabeled DNA and chromatographic separation of the products. In Figure 6A, the results with HPLC separation of the DNA alkylation products formed in cacodylate buffer are presented. Seven products designated A-G were quantitated, and their amounts are given in Table 1. Peaks A and G are the principal alkylation products detected, and they represent 30.3% and 25.3%, respectively, of the alkylation. Peaks B-F are identified in Table 1, and they represent 3.4%, 2.1%, 13.0%, 3.0%, and 1.2% of the alkylation, respectively. The HPLC pattern of DNA alkylation products observed when the reaction occurred in the presence of Tris buffer is more complex (Figure 6B). In addition to the products found with cacodylate buffer (A-G), at least 5 Tris-dependent products were detected. These adducts are identified in Figure 6B by numbers and their levels are given in Table 1. Analysis of the results in Table 1 reveals that when the reaction of [ethyl-3H]Cl-ENU with DNA occurs in Tris buffer the principal alkylation product formed is peak 3 followed by peaks A and G. Comparison of the levels of the individual adducts A-G produced in cacodylate buffer with their levels in Tris buffer shows that they have been increased on an average of 80% in Tris buffer.

212 Chem. Res. Toxicol., Vol. 9, No. 1, 1996

Chen et al.

Table 1. HPLC Analysis of DNA Adducts Formed from the Reaction of [ethyl-3H]Cl-ENU and Calf Thymus DNA in Cacodylate or Tris Buffer (pH 7.6) yields

adducte preincubation time (min)a buffer µmol of adducts/mol of DNAd

0 cacb 30.1

0 Trisc 85.2

15 cac 4.2

15 Tris 49.2

30 cac 2.8

30 Tris 47.3

45 cac 2.6

45 Tris 46.8

N7-HO-Et-dG (A) N1-HO-Et-dG (B) N7-bisG (C) N7-Cl-Et-dG (D) dC-dG (E) O6-HO-Et-dG (F) triesters (G) peak 2 peak 3 peak 4 peak 5 peak 9 totalg

9.14 1.03 0.64 3.94 0.91 0.37 7.62 nd f nd nd nd nd 23.65

15.80 1.32 2.04 5.29 1.66 0.97 9.37 3.62 26.60 2.19 3.61 0.96 73.43

nd nd nd nd nd nd nd nd nd nd nd nd

nd nd nd nd nd nd nd 2.90 28.30 1.81 1.76 1.13 35.9

nd nd nd nd nd nd nd nd nd nd nd nd

nd nd nd nd nd nd nd 2.71 28.40 1.55 1.45 1.03 35.14

nd nd nd nd nd nd nd nd nd nd nd nd

nd nd nd nd nd nd nd 3.01 27.20 1.31 1.43 1.07 34.02

a Time that [3H]CNU is incubated in buffer at 37 °C prior to adding DNA. b 10 mM sodium cacodylate buffer (pH 7.6). c 10 mM TrisHCl buffer (pH 7.6). d Based on radioactivity associated with DNA prior to digestion. e Identity of adduct determined by coinjection with authentic standard. f nd ) not determined because of insufficient radioactivity. g Total adducts in identified peaks.

The influence of preincubation of [ethyl-3H]Cl-ENU in Tris buffer for 45 min on the formation of individual products was analyzed (Figure 6C). Under these conditions, only the Tris-dependent adducts were detected. The levels of the Tris-dependent adducts 2, 3, 4, 5, and 9 are very similar with that seen with either no preincubation or with preincubation for up to 45 min (Table 1). Toxicity. The effect of Tris on the cytotoxicity of CCNU in L1210 cells is shown in Figure 7. CCNU alone and CCNU in the presence of 10 mM Tris cause the same dose-dependent inhibition of cell growth. When CCNU is preincubated in the medium for 2 h, with or without 10 mM Tris, the toxicity response is shifted to the right, but there is no significant difference between the two curves.

Discussion CCNU and related CNU compounds hydrolyze fairly rapidly at physiological pH (13) and even faster in plasma (2). For example, the t1/2 of CCNU in phosphate buffer at pH 7.4 and 37 °C is ∼50 min, and in plasma the t1/2 drops to ∼5 min. This instability presents a clinical problem since much of the drug decomposes in the blood during iv administration. It is assumed that the drug must hydrolyze in the target cells because of the transient nature of the DNA alkylating intermediates generated from CNU. Therefore, the observations that Tris stabilized CNU in aqueous buffer through formation of a complex represented a potential way to improve the efficacy of the drug (16, 17). The results of our study show that Tris enhances the overall extent of DNA alkylation by CNU. An ∼100% increase in adduct formation is observed in the sequencing gels that provide information on all N7-alk-dG lesions. The more detailed HPLC analysis of the reaction of [3H]Cl-ENU with DNA shows that the average increase for each individual “normal” N7-alk-dG adduct (A, C, D) is also ∼100%, with an overall increase in these adducts (A + C + D) of ∼70%. When the non-N7-alkdG adducts (B, E, F, G) are included in the comparison, the increase of the normal adducts (A-G) in Tris buffer drops to ∼50%. The origin of the increased efficiency in DNA modification by CNU in the presence of Tris is not understood, but other buffers, albeit to a different extent, also can affect the yield of adducts as detected in the

sequencing gels. The yield of N7-alk-dG adducts in the different buffers follows the order: phosphate , cacodylate < Tris ≈ triethanolamine. None of the Tris analogues tested simulated the results obtained with Tris, and tricine, serinol and ethanolamine actually caused a decrease in N7-alk-dG formation. While the yield of the “normal” Cl-ENU adducts (A-G) is buffer dependent, there does not appear to be any significant qualitative change in the array of adducts, i.e., products A-G are seen in the same proportions in caocdylate and Tris buffers. There also appears to be no change in stability of the “normal” reactive intermediates formed from Cl-ENU in the Tris buffer as the yield of the “normal” adducts (AG) precipitously drops off when Cl-ENU is preincubated with buffer for 15 min prior to the addition of the DNA (Table 1). The same effect is found in the cacodylate buffer. This decrease in products A-G was anticipated since Cl-ENU undergoes rapid base-catalyzed hydrolysis at near neutral pH, and this decomposition is required for the production of the reactive intermediates that modify DNA and any other nucleophilic target present. We have also demonstrated that the extent of DNA alkylation in all buffers is decreased as the ionic strength of the incubation increases. The same salt effect has previously been reported for N7-alk-dG formation by other ANU and nitrogen mustards (18, 21, 22). If Tris affected the partition of CNU into alkanediazonium ion, oxadiazolium ion, and 1,2,3-oxadiazole intermediates, there should be a difference in the ratio of 2-chloroethylated vs 2-hydroxyethylated adducts. Since the yield of chloroethyl and hydroxyethyl adducts at all the sites (N1-G, N7-G, O6-G, and phosphate backbone) is similarly altered, it appears that Tris does not change the mechanism of CNU hydrolysis or the types of reactive intermediates involved in the formation of adducts A-G. Therefore, it is most likely that the fate of the reactive intermediates (formed from the hydrolysis of CNU) in the different buffers may be responsible for the modulation of the yield of “normal” products. The most important results of this study are the extended time course for N7-alk-dG formation in Tris buffer, and the extensive formation of Tris-dependent adducts. Based on the observation that DNA alkylation occurs even after the CNU has passed through as many as four half-lives before the DNA was added, and that

CNU-Tris

Chem. Res. Toxicol., Vol. 9, No. 1, 1996 213

Figure 7. Effect of Tris solute, preincubation times, and CCNU concentration on the toxicity of CCNU in L1210 cells: O, control incubation with no Tris and no preincubation period; b, 10 mM Tris added to cells prior to adding CCNU and no preincubation time; 3, 10 mM Tris preincubation with CCNU in medium for 2 h prior to adding cells; 1, no Tris added but CCNU was preincubated in medium for 2 h prior to adding cells.

Figure 6. Effect of preincubation and different solutes on the formation of DNA alkylation products by [3H]CNU: (A) 10 mM cacodylate buffer (pH 7.6) with no preincubation; (B) 10 mM Tris buffer (pH 7.6) with no preincubation; and (C) 10 mM Tris buffer (pH 7.6) with 45 min preincubation. After enzymatic digestion, the alkylation products were separated by HPLC and fractions (30 s) were collected and analyzed for radioactivity. The identity of the individual peaks was determined by coinjection with authentic standards (11). Quantitation of the individual products is shown in Table 1.

this new set of adducts (peaks 2-5, 9) is seen only with Tris, it is hypothesized that CNU reacts with Tris to form a nitrogen half mustard. The involvement of this type of alkylating agent would do the following: (i) afford monofunctional and not cross-linked adducts; (ii) show DNA alkylation rates that differ from CNU hydrolysis; (iii) be independent of the structure of the CNU precursor; and (iv) not be observed with monofunctional analogues of CNU, e.g., MNU, MCNU, or HO-ENU. All of these speculations are confirmed. What is not clear is why the coaddition of Tris analogues (Figure 5) does not cause a similar effect since they could also generate nitrogen half mustards. It may be that the yield of a Tris half mustard is high because Tris uniquely complexes with CNU prior to its hydrolysis (see below).

Figure 8. Scheme for Tris-CNU complexation (17), and possible pathway for formation of nitrogen half mustard.

The initiative for the current study was the previous kinetic investigation that suggested that Tris and CNU formed a “stable” complex, which could either collapse back to CNU or decompose to products (Figure 8). At near physiological pH in carbonate buffer containing Tris, the pseudo-first-order rate constants ko and kc (Figure 8) are 23.2 × 102 and 12.6 × 102 min-1, respectively, and the stability constant Kc was reported to be 34.5 M-1 (17). There was no speculation on the nature of this CNUTris complex, but it was reported that the entropy term for complex formation is very negative (-217 J mol-1 deg-1), suggesting a highly ordered state (17). Our results are generally consistent with this scheme. The

214 Chem. Res. Toxicol., Vol. 9, No. 1, 1996

partially reversible formation of a CNU-Tris complex would explain the reported increased stability of CNU in this buffer (13-17). The generation of the Tris related adducts could correspond to the efficient collapse of a Tris-CNU complex (kC) to a nitrogen half mustard (Figure 8). Finally, the Tris-CNU complex may have an enhanced affinity for DNA, possibly via an electrostatic attraction, which would explain the increase in the “normal” CNU adducts. This last scenario would require that the Tris-CNU complex dissociate to CNU and Tris in the environment of DNA. Work is currently underway to elucidate the structure of the DNA adducts associated with Tris. In conclusion, Tris retards the decomposition of CNU and increases the yield of adducts normally seen in the CNU-mediated alkylation of DNA. However, CNU in the presence of Tris also results in the formation of a very significant amount of adducts that do not appear to enhance cytotoxicity. Based on these results, the use of Tris to stabilize CNU in order to improve their chemotherapeutic efficacy may not be prudent since Tris appears to enhance the formation of DNA lesions that are only weakly cytotoxic but which may be mutagenic. The results also indicate that the reactions of CNU with DNA are highly buffer and solute dependent and that care must be taken in relating experiments done in different buffer systems.

Acknowledgment. This work was supported by Public Health Service Grants CA29088 (B.G.) and CA13525 (W.J.B.), NCI Core Grant CA36727, and American Cancer Society Core Grant SIG-16.

Chen et al.

(7)

(8)

(9)

(10)

(11)

(12) (13)

(14)

(15)

(16) (17)

(18)

References (1) Levin, V. A., and Wilson, C. B. (1976) Nitrosourea chemotherapy for primary malignant gliomas. Cancer Treat. Rep. 60, 719-724. (2) Prestayko, A. W., Crooke, S. T., Baker, L. H., Carter, S. K., and Schein, P. S., Eds. (1981) Nitrosoureas: Current Status and New Developments, Academic Press, New York. (3) Erickson, L. C., Laurent, G., Ducore, J., Sharkey, N., and Kohn, K. W. (1980) DNA crosslinking and monoadduct repair in nitrosourea-treated human tumour cells. Nature (London) 288, 727-729. (4) Gombar, C. T., Tong, W. P., and Ludlum, D. B. (1980) Mechanism of action of the nitrosoureassIV. Reactions of BCNU and CCNU with DNA. Biochem. Pharmacol. 22, 2639-2643. (5) Tong, W. P., Kohn, K. W., and Ludlum, D. B. (1982) Modifications of DNA by different haloethylnitrosoureas. Cancer Res. 42, 44604464. (6) Tong, W. P., Kirk, M. C., and Ludlum, D. B. (1982) Formation of the cross-link 1-[N3-deoxycytidyl],2[N1-deoxyguanosinyl]ethane

(19) (20)

(21)

(22)

in DNA treated with N,N′-bis(2-chloroethyl)-N-nitrosourea. Cancer Res. 42, 3102-3105. Brent, T. (1984) Suppression of cross-link formation in chloroethylnitrosourea-treated DNA by an activity in extracts of human leukemic lymphoblasts. Cancer Res. 44, 1887-1892. Robins, P., Harris, A. L., Goldsmith, I., and Lindahl, T. (1983) Cross-linking of DNA induced by chloroethylnitrosourea is prevented by O6-methylguanine-DNA methyltransferase. Nucleic Acids Res. 11, 7743-7758. Ludlum, D. B., Mehta, J. R., and Tong, W. P. (1986) Prevention of 1-(3-deoxycytidyl), 2-(1-deoxyguanosinyl)-ethane cross-link formation in DNA by O6-methylguanine-DNA methyltransferase. Cancer Res. 46, 3353-3357. Bodell, W. J., Aida, T., Berger, M. S., and Rosenblum, M. L. (1986) Increased repair of O6-alkylguanine adducts in glioma-derived human cells resistant to the cytotoxic and cytogenetic effects of 1,3 bis-(2-chloroethyl)-1-nitrosourea. Carcinogenesis 7, 879-883. Bodell, W. J., and Pongracz, K. (1993) Chemical synthesis and detection of the cross-link 1-[N3-(2-deoxycytidyl)]-2-(N1-(2-deoxyguanosinyl)]ethane in DNA reacted with 1-(2-chloroethyl)-1nitrosourea. Chem. Res. Toxicol. 6, 434-438. Chen, F.-X., and Gold, B. (1990) DNA alkylation by Cl-ENU. Proc. Am. Assoc. Cancer Res. 31,160. Reed, D. J., May, H. E., Boose, R. B., Gregory, K. M., and Beilstein, M. A. (1975) 2-Chloroethanol formation as evidence for a 2-chloroethyl alkylating intermediate during chemical degradation of 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea and 1-(2-chloroethyl3-trans-4-methylcyclohexyl)-1-nitrosourea. Cancer Res. 35, 568576. Chatterji, D. C., Greene, R. F., and Gallelli, J. F. (1978) Mechanism of hydrolysis of halogenated nitrosoureas. J. Pharm. Sci. 67, 1527-1532. Betteridge, R. F., Culverwell, A. L., and Bosanquet, A. G. (1989) Stability of tauromustine (TCNU) in aqueous solutions during preparation and storage. Int. J. Pharm. 56, 37-41. Loftsson, T., and Fridriksdottir, H. (1990) Degradation of lomustine in aqueous solution. Int. J. Pharm. 62, 243-247. Loftsson, T., and Fridriksdottir, H. (1992) Stabilizing effect of tris(hydroxymethyl)aminomethane on N-nitrosoureas in aqueous solution. J. Pharm. Sci. 81, 197-198. Wurdeman, R. L., Church, K. M., and Gold, B. (1989) DNA methylation by N-methyl-N-nitrosourea, N-methyl-N′-nitro-Nnitrosoguanidine, N-nitroso(1-acetoxyethyl)methylamine, and diazomethane: mechanism for the formation of N7-methylguanine in sequence-characterized 5′-32P-end-labeled DNA. J. Am. Chem. Soc. 111, 6408-6412. Rhode, S. L., III (1985) Nucleotide sequence of the coat protein gene of the canine parvovirus. J. Virol. 54, 630-633. Maxam, A. M., and Gilbert, W. (1980) Sequencing end-labeled DNA with base-specific chemical cleavages. Methods Enzymol. 65, 499-560. Wurdeman, R. L., and Gold, B. (1988) The effect of DNA sequence, ionic strength, and cationic DNA affinity binders on the methylation of DNA by N-methyl-N-nitrosourea. Chem. Res. Toxicol. 1, 146-147. Hartley, J. A., Gibson, N. W., Kohn, K. W., and Mattes, W. B. (1986) DNA sequence selectivity of guanine-N7 alkylation by three antitumor chloroethylating agents. Cancer Res. 46, 1943-1947.

TX950097G