Interstrand Cross-Linking of DNA by 1,3-Bis(2-Chloroethyl)- 1-Nitrosourea and Other 1-(2-Haloethyl)-1-Nitrosoureas

[CANCER RESEARCH 37, 1450-1454, May 1977] Interstrand Cross-linking of DNA by 1,3-Bis(2-chloroethyl)- 1-nitrosourea and Other 1-(2-haloethyl)-1-nitrosoureas

Kurt W. Kohn

Laboratory of Molecular Pharmacology, Division of Cancer Treatment, National Cancer Institute, NIH, Bethesda, Maryland 20014

SUMMARY shows that chloroethylnitrosoumeas beaning a single alkylat ing function are active producers of DNA interstrand cross Bifunctional alkylating agents are known to cross-link links, and the mechanism of this effect is investigated. An DNA by simultaneously alkylating two guanine residues lo abstract describing this work has appeared (11). cated on opposite strands. Despite this apparent require ment for bifunctionality, 1-(2-chlonoethyl)-1 -nitrosoureas beaminga single alkylating function were found to cross-link MATERIALS AND METHODS DNA in vitro. Cross-linking was demonstrated by showing inhibition of alkali-induced strand separation. Extensive Nitrosoureas were supplied by Drug Research and Devel cross-linking was observed in DNA treated with 1-(2-chloro opment, Division of Cancer Treatment, National Cancer ethyl)-1 -nitmosourea, 1,3-bis-(2-chlonoethyi)-1 -nitrosoumea, Institute, NIH. The compounds were dissolved in 95% and 1-(2-chlonoethyl)-3-cyclohexyl-1 -nitrosounea. The reac ethanol immediately before use. DNA was isolated from tion occurs in two steps, an initial binding followed by a Escherichia coli by standard procedures using Pronase, second step which can proceed after removal of unbound RNase, and alcohol precipitation. drug. It is suggested thatthe first step is chloroethylation of Reaction mixtures contained DNA (20 @g/ml),0.08 M a nucleophilic site on one strand and that the second step NaCI, 0.01 M NaH2PO4,0.03 M Na2HPO4,0.1 mM EDTA, 5% involves displacement of CI by a nucleophilic site on the ethanol, and the specified concentration of drug. The reac opposite strand, resulting in an ethyl bridge between the tions were carried out at 37Â°and at pH 7.1. strands. Consistent with this possiblity, 1-(2-fluoroethyl)-3- In postincubation experiments, drug was removed by pre cyclohexyl-1 -nitrosourea produced much less cross-link cipitating the DNA with 2 volumes of ethanol. For facilitation ing, as expected from the known low activity of F, corn of precipitation, the solution was made 0.3 M in NaCI before pared with C1 , as leaving group. 1-Methyl-1-nitrosourea, adding ethanol. The DNA was centrifuged and the pellet which is known to depuninate DNA, produced no detectable was washed with 70% ethanol and medissolved in 1 mM cross-linking. tnisodium EDTA at 0Â°.Abuffer concentrate was then added to give the same composition used in the initial reaction, and incubation at 37Â°was resumed. INTRODUCTION For alkali denaturation, 0.1 ml of reaction mixture or postincubation mixture was mixed with 0.2 ml of 0.085 N The 1-alkyl-1 -nitnosouneas, including BCNU,1 CCNU, NaOH-1 mM EDTA. After 3 mm at 23Â°,the solution was GNU, FCNU, and MNU, can alkylate nucleophilic groups neutralized (to approximately pH 8) by adding 0.2 ml of 0.1 N (reviewed by Wheeler) (25), and are frequently considered citric acid-0.03 M Tnis. After several mm, the solutions were to be a kind of biological alkylating agent. Although stan frozen in dry ice-alcohol and stored at â€”20Â°.Undenatured dard alkylating agents such as nitrogen mustards require 2 controls were prepared as above, except that the NaOH and functional groups for antitumon activity, many haloethylni citric acid solutions were mixed before DNA was added, so trosoureas bearing a single alkylating function are ex that the DNA was not exposed to alkali. tremely active antitumon agents (9, 10, 19). The requirement DNA cross-linking was measured by determining the frac for bifunctionality suggests that the cytotoxic lesion pro tion of the DNA that is denaturable by the above alkali duced by alkylating agents such as nitrogen mustards in treatment (12). One cross-link per molecule is sufficient to volves alkylations at 2 neighboring sites and that it perhaps prevent denaturation (12). The fraction of the DNA dena stems from the cross-linking of these sites (21). In particu tured was determined by equilibrium centnifugation in CsCI. Ian,the ability of these bifunctional alkylating agents to form Sample solutions (0.5 ml) were mixed with 0.1 ml of 0.5 M cross-links between opposite DNA strands (7, 8, 12, 14) is an [email protected] Na@CO3-Sarkosyl(0.4 mg/mI), and 0.8 g of attractive possibility for the origin of cytotoxicity. This work CsCI. Clostridium perfringens DNA was added for buoyant density reference. Solutions were adjusted to a refractive

I The abbreviations used are: BCNU, 1 ,3-bis(2-chloroethyl)-1-nitnosourea index of 1.3995 by adding small volumes of water. The (NSC 409962); CCNU, 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea (NSC sodium carbonate buffer increased the separation between 79037); CNU, 1-(2-chlonoethyl)-1 -nitrosourea (NSC 47547); FCNU, 1-(2-fluo native and denatured bands (12, 24), and the detergent roethyl)-3-cyclohexyl-1-nitnosourea (NSC 87974); MNU, 1-methyl-1-nitrosou rea (NSC 23909). prevented selective loss of denatured DNA (12). Samples Received November 23, 1976; accepted February 7, 1977. were placed in 12-mm double-sector analytical cells; the

1450 CANCER RESEARCH VOL. 37

reference sectors were filled with solutions containing CsCI only. Samples were spun at 44,000 rpm for 20 hr at 23Â°ina A Beckman Model E ultracentnifuge equipped with mono Ref.

chrometer and electronic scanner. The areas of the native %@ J L No Drug and denatured bands were measured with a planimeten. @JN RESULTS B HJ\@c LCH?CH, 0 The banding patterns of control native and denatured N-C-NH

DNA are shown in Chart 1. It is seen that DNA not exposed N@ 0 to alkali was totally native; it contained no material at dena 2mM CCNU tured density (Chart 1A), whereas DNA exposed to alkali Ji@@ was completely denatured (Chart 1B). A mixture of equal parts of native and alkali-denatured DNA produced clearly C@F CH)CH, 0 \ II separated bands of nearly equal areas (Chart 1C). N-C-NH Q Chart 2 shows the effects of incubation of DNA with N-0 nitrosoureas for 9 hr at 37Â°.Extensive cross-linking, mdi 2mM FCNU cated by the persistence of DNA banding at native density after alkali treatment, was produced by 2 mM CCNU, BCNU, or GNU; only the result with CCNU is shown (Chart 2B). The DI CH@0 fluoro analog of CCNU, however, produced little cross \ p linking at this concentration and incubation time (Chart 2C). Nâ€”Câ€”NH, MNU at 5 mM produced no cross-linking(Chart2D);the N=0 denatured DNA band was measurably broadened, however, 5mM MNU showing that the compound did react with the DNA but produced mainly chain breaks rather than cross-links. Chart 2. Effect of nitrosouneas on DNA denaturability. E. coli DNA was incubated with or without drug for 9 hr at 37Â°andthen treated with alkali. A, The questionofcross-linkingbyFCNU was pursuedfur nodrug;B,2mMCCNU;C,2mMFCNU;D,5mMMNU. then by extending the reaction time to 21 hr (Chart 3), which yielded small but distinct amounts of cross-linked DNA @ A \\\,@@@_NoDrug (Chart 3, B and C). With MNU, however, the longer reaction time produced further band broadening, but still no detect able cross-linking (Chart 3D). The formation of a covalent connection between 2 DNA strands requires 2 successive reactions: (a) an alkylation or other modification of 1 strand; (b) a reaction of the modified strand with the complementary DNA strand. Since experi B 2mMFCNU ments showed that the extent of cross-linking continued to increase for time periods that exceeded the expected stabil

A Ref. Native C

B: @ Denatured @ D 5mMMNU

@ C Na@ve Chart 3. Effects of prolonged incubation with FCNU or MNU. DNA was : Denatured incubated with or without drug for 21 hr at 37Â°andthen treated with alkali. A, no drug; B, 2 mM FCNU; C, 5 m@ FCNU; D, 5 mM MNU.

ity of the nitrosourea compounds, the possibility was con Chart 1. Equilibrium centnifugation in CsCI-sodium carbonate-Sankosyl. sidered that the rate of cross-link formation was limited by a A, native E. coli DNA; B, alkali-denatured DNA; C, equal parts native and denatured DNA. C. perfringens DNA added for buoyant density reference. slow 2nd step in the reaction sequence. The 2nd step in the Buoyant density increases from left to right. reaction can be isolated by first reacting DNA with drug for a

MAY 1977 1451

relatively short time and then removing the drug and contin on incubation in the absence of drug after an initial treat uing the incubation of the DNA by itself. Continued cross ment with either CCNU, BCNU, or GNU. Chart 5 shows the link formation during this postincubation would be due to same for FCNU, although cross-linking by this compound completion of the 2nd reaction step. was much smaller in extent and required a longer time to In a test of this idea, DNA was incubated with CCNU, become apparent. BCNU, CNU, or FCNU for 1 hr; alcohol precipitated to Since the cross-linking activity of FCNU is so low com remove drug;redissolved;andincubatedfurtherintheab pared with that of CCNU, the possibility was considered that sence of drug. Chart 4 shows that cross-linking increased the activity of FCNU was due to a trace contamination by a chloroethylnitrosounea. This possibility was excluded by Dr.

CI CH,CH, 0 CICH,CH,0 CI CH,CH, 0 Peter Lim, Stanford Research Institute, who reviewed the N_C_NHO Nâ€”Câ€”NHCH,CH,CI Nâ€”Câ€”NH, method of synthesis and reanalyzed the FCNU preparation N-0 @â€˜=0 N=0 (Lot CC2O-21-1, Stanks) used in these experiments. The CCNU BCNU CNU preparation was estimated to be 96% pure and contained 4 trace chromatognaphic components but no detectable CCNU within a detection limit of 0.1%. Elemental analysis failed to detect any chlorine on chloride. The kinetics of DNA cross-linking by BCNU, with or with out drug removal after 1 hr, is shown in Chart 6. The final extent of cross-linking after 9 hr of total incubation was 78% without drug removal and 67% with drug removal at 1 hr. Hence most of the initial reaction, presumably alkylation of 1 DNA strand, occurred during the 1st hr. A more detailed kinetic analysis was not attempted in these experiments because of the molecular weight heterogeneity of bacterial DNA preparations. The kinetics of the 2nd step of cross-linking following a 1- hr reaction with CCNU, BCNU, or CNU was examined as a function of time after removal of drug (Chart 7). The kinetics is seen to be similar for all 3 compounds, and this is espe cially well demonstrated by the similarity between the 9hj\\@J\J@J'\ curves for 2 mM BCNU and 1 mM GNU. The similarity sug gests that the chemical nature of the initial alkylation Chart 4. Delayed cross-linking after reaction of DNA with chloroethylnitro product is the same for all of these compounds and per soureas. DNA was incubated with drug for 1 hr at 37Â°,precipitated with haps for all chloroethylnitrosoureas. ethanol, redissolved, and then incubated in the absence of drug for 0, 1.25, @ on9 hrat 37Â°.A,B, C, 1 mM CCNU; D, E, F, 1 BCNU; G, H. I, 1 mM CNU. Incubation times after DNA was redissolved: A, D, G, 0 hn;B, E, H, 1.25 hr; C, @ F, I, 9 hr. The DNA reference bands have been omitted.

w -J

I- z w 0 L z 0 U

U-

HOURS AT 37Â° Chart 6. Kinetics of DNA cross-linking by BCNU. DNA was reacted with 2 Chart 5. Delayed cross-linking after reaction of DNA with 5 m@FCNU for 1 mM BCNU for various times ( 0), on drug was removed after 1 hr (arrow). and hr. Procedure similar to that in legend to Chart 4. A, immediately after DNA incubation was continued in the absence of drug (â€¢).Fractiondenatunable is was redissolved. B, after further incubation for 21 hr at 37Â°. the fraction of the DNA remaining uncross-linked.

1452 CANCER RESEARCHVOL. 37

(0.2%). Most ofthese sites of reaction have also been identi fied for the case of methylnitrosourea, although the relative extents of reaction at the various sites are different [Lawley et at. (15)]. The sites of reaction of chloroethynitrosoureas

UJ with DNA have not yet been determined. Kramer et at. (13), -J however, studied the reaction of BCNU with polycytidylic acid and identified, as products, 3-hydroxyethylcytidyiate and 3-4-ethanocytidylate. These products are consistent with z w 0 an initial chloroethylation of cytosine, followed by a displace z 0 ment of Cl, in the 1st case by hydrolysis, and in the 2nd

0 case by a 2nd alkylation of the cytosine moiety. All of this evidence suggests that chloroethylnitrosoureas can chloro ethylate a variety of DNA sites and that there is a possibility of a 2nd alkylation at another DNA site, coupled with the displacement of CI. If the 2 reaction sites are on opposite strands of DNA, an intenstrand cross-link would result (Chart 8). In order to accomplish this, however, it would be neces sary for the 2 opposed reaction sites to be spanned by a PO5T-INCUBATION(Hours) bridge of only 2 carbon atoms. Unless the DNA helix is Chart 7. Kinetics of the delayed step in DNA cross-linking by chloroethyl grossly distorted by the initial chloroethylation, the possibil nitrosoureas. DNA was reacted for 1 hr with 1 [email protected],1 mM BCNU, 2 [email protected] ities for a cross-bridge only 2 carbons long are limited to the BCNU, or 1 mM CNU. The DNA was then alcohol precipitated, redissolved, sites normally involved in hydrogen bonding between the and incubated in the absence of drug for various times at 37Â°. bases. The most attractive possibility of this type is a cross bridge between guanine 0-6 and cytosine N-4. This possibil DISCUSSION ity is supported by the active ethylation of guanine 0-6 in helical DNA by ethylnitrosourea (20, 22) and by the involve lnterstrand DNA cross-links are produced by a variety of ment of the cytosine N-4 position in an intramolecular cross bifunctional alkylating agents, primarily through reactions bridge formed in the reaction of BCNU with polycytidylic acid at guanine N-7 (1, 12, 14). The formation of such cross-links (13). by chlomoethylnitnosouneas may at first seem surprising be The findings in this work are in accord with the above cause these compounds are not bifunctional in the same proposal. The observed inhibition of strand separation in sense as the usual alkylating agents. Recent insights into drug-reacted DNA is a clear indication that covalent inter the chemistry of these compounds, however, suggest a strand cross-links have been produced (12). A notable fea chemical mechanism by which interstrand cross-bridges tune of the cross-linking by chloroethylnitnosouneas is the may be produced. slowness of the 2nd step of the reaction, in which the The reactivity of nitrosoureas derives from 2 types of alkylated DNA becomes increasingly cross-linked after re highly reactive products into which the nitrosourea mole moval of drug. It is reasonable to suppose that some of the cule tends to split; an alkyldiazohydroxide and an isocya initial alkylation may occur at guanine 0-6, since a variety of nate(17,25). alkylating agents, and in particular ethylnitrosourea and OH methyinitrosounea, are known to alkylate this site. The proposed mechanism requires only that part of the initial R1â€”N--Câ€”Nâ€”R2 -@R,â€”N=Nâ€”OH + O==C=N-.--R2 alkylation occur at this site; it may well be a minor site. The N=O 2nd step,which would entailareactionatcytosineN-4, would be slow because this position is a relatively weak Recent evidence indicates that chloroethylnitrosouneas nucleophile, as indicated by the small extents of alkylation such as CCNU follow this same scheme and are capable of of cytosine N-4 in model compounds (20). Although alkyla transferring the CICH2CH2group in alkylation reactions (4, tion at cytosine N-4 has not been reported to occur in native 5, 18). Although both the alkyldiazohydroxide and the iso DNA, this may be due to a reduced reactivity of the hydro cyanate products are highly reactive, only the former reacts measurably with nucleic acids; this was demonstrated by CICH2CH2 0 Cheng et a!. (3), who found that ethylene-'4C label from \ II C) CCNU binds to purified DNA, RNA, and various proteins, Nâ€”Câ€”NHR CH2@CH2 CH2â€”CH2 whereas cyclohexyl-14C label binds to proteins but not to r@J=0 DNA on RNA. The reaction with nucleic acids is one of alkylation, in which the A1 group is transferred to various nucleophilic sites. The sites of reaction of ethylnitrosourea (A, = C2H5; R2 = H) with purified mammalian DNA have recently been reported by Sun and Singer (22). Ethylation ICICH2CH,@]DNA was reported at phosphates (70%), guanine 0-6 (12%), guanine N-7 (11%), adenine N-3 (4.2%), thymine N-3 (0.8%), Chart 8. Proposed mechanism of DNA cross-linking by chlonoethylnitro cytosine N-3 (0.2%), adenine N-i (0.2%), and adenine N-7 soureas. X and Y, nucleophilic sites on opposite DNA strands.

MAY 1977 1453

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1977 American Association for Cancer Research. K. W. Kohn gen-bonded group. An initial alkylation at guanine 0-6 Action: Linking of Complementary DNA Strands. Proc. NatI. Acad. Sci. U. S., 50: 355-362, 1963. would eliminate this hydrogen bond and would promote the 9. Johnston, T. P., McCaleb, G. S., Opligen, P. S., Laster, W. R., and possibility of a reaction with the released cytosine amino Montgomery, J. A. Synthesis of Potential Anticancer Agents. 38. N- group. The kinetics of this reaction, in which C1 is me Nitrosoureas. 4. Further Synthesis and Evaluation of Haloethyl Deriva tives. J. Med. Chem., 14: 600-614, 1971. leased, should be the same for all chlonoethylnitrosoureas, 10. Johnston, T. P., McCaleb, G. 5., Opliger, P. S., and Montgomery, J. A. since the initial alkylation would be the same. Accordingly, Synthesis of Potential Anticancer Agents. 36. N-Nitrosoureas. II. Haloal the time course of cross-linking after drug removal was the kyl Derivatives. J. Med. Chem., 9: 892-911, 1966. 11. Kohn, K. W. Inter-strand Cross-linking of DNA by Chloroethylnitrosou same for BCNU as for GNU. The corresponding reaction reas. Federation Proc., 35: 1708, 1976. should be much slower, however, for fluoroethylnitrosou 12. Kohn, K. W., Spears, C. L., and Doty, P. Inter-strand Crosslinking of DNA by Nitrogen Mustard. J. Mol. Biol., 19: 266-288, 1966. reas, because F- is a much weaker leaving group than is Cl 13. Kramer, B. S., Fenselau, C. C., and Ludlum, D. B. Reaction of BCNU with (23). This would explain the greatly reduced cross-linking Polycytidylic Acid. Substitution of the Cytosine Ring. Biochem. Biophys. efficiency of FCNU. Res. Commun., 56: 783-788, 1974. 14. Lawley, P. D., and Brookes, P. D. Intenstrand Cross-linking of DNA by Di An alternative to the proposed mechanism would be that functional Alkylating Agents. J. Mol. Biol., 25: 143-160, 1967. the observed cross-linking was due to depunination of alkyl 15. Lawley, P. D., Orn, D. J., and Jarman, M. Isolation and Identification of ated bases, followed by reaction of the resulting sugar Products from Alkylation of Nucleic Acids: Ethyl- and Isopropyl-punines. Biochem. J., 145: 73-84, 1975. aldehyde groups with amino groups of the opposite strand 16. Lipsett, M. N., and Weissbach, A. The Site of Alkylation of Nucleic Acids (2, 6). Arguing against this possibility is the finding that by Mitomycin. Biochemistry, 4: 206-21 1â€¢1965. 17. Montgomery, J. A., James, R., McCaleb, G. S., and Johnston, T. P. The MNU, which is known to alkylate and depurinate guanines Modes of Decomposition of BCNU and Related Compounds. J. Med. and adenines, produced no observable cross-linking under Chem., 10: 668-674, 1967. the conditions of our experiments. 18. Reed,D.J.,May,H.E.,Boose,R.B.,Gregory,K.M.,andBeilstein,M.A. 2-Chlonoethanol Formation during Chemical Degradation of CCNU and Trans-Methyl CCNU; Evidence for a 2-Chlonoethyl Alkylating Intermedi ate. Cancer Res., 35: 568-576, 1975. 19. Schabel, F. M., Jr., Johnston, T. P., McCaleb, G. S., Montgomery, J. A., REFERENCES Laster, W. R., and Skipper, H. E. Experimental Evaluation of Potential Anticancer Agents. VIII. Effects of Certain Nitnosouneasand Intracerebral 1. Brookes, P., and Lawley, P. D. The Reaction of Mono- and Di-functional L1210 Leukemia. Cancer Res., 23: 725-733, 1963. Alkylating Agents with Nucleic Acids. Biochem. J., 80: 496â€”503,1961. 20. Singer, B. The Chemical Effects of Nucleic Acid Alkylation and Their 2. Bunnotte, J., and Verly, W. G. Crosslinking of Methylated DNA by Moder Relations to Mutagenesis and Carcinogenesis. Progn. Nucleic Acid Res. ate Heating at Neutral pH. Biochim. Biophys. Acta,262: 449â€”452,1972. Mol. Biol., 15: 219â€”284,1975. 3. Cheng, C. J. . Fujimuna, S., Grunbenger, D., and Weinstein, I. B. Intenac 21. Stacey, K. A., Cobb, M., Cousens, S. F., and Alexander, P. Reactions of tion of CCNU with Nucleic Acids and Proteins in Vivo and in Vitro. Cancer the â€œRadiomimeticâ€•AlkylatingAgents with Macromolecules in Vitro. Res., 32: 22-27, 1972. Ann. N. Y. Acad. Sci., 68: 682-701 , 1958. 4. Colvin, M., Bnundnett, R. B., Cowens, W., Jardine, I., and Ludlum, D. B. 22. Sun, L., and Singer, B. The Specificity of Different Classes of Ethylating A Chemical Basis for the Antitumor Activity of Chlonoethylnitrosoureas. Agents toward Various Sites of HeLa Cell DNA and in Wvo. Biochemistry, Biochem. Pharmacol., 25: 695-699, 1976. 14: 1795-1802,1975. 5. Colvin, M., Cowens, J. W., Brundrett, R. B., Kramer, B. 5. , and Ludlum, 23. Thornton, E. R. Solvolysis Mechanisms, p. 165. New York: Ronald Press, D. B. Decomposition of BCNU in Aqueous Solution. Biochem. Biophys. 1964. Res. Commun., 60: 515-520, 1974. 24. Vinograd, J., Morris, J., Davidson, N., and Dove, W. F., Jr. Buoyant 6. Freese, E., and Cashel, M. Crosslinking of DNA by Exposure to Low pH. Behavior of Viral and Bacterial DNA in Alkaline CsCI. Proc. NatI. Acad. Biochim. Biophys. Acta, 91: 67-77, 1964. Sci. U. S., 49: 12-17, 1963. 7. Geiduschek, E. P. â€œReversibleâ€•DNA.Proc. Nat). Acad. Sci. U. S., 47: 25. Wheeler, G. P. Mechanism of Action of Alkylating Agents: Nitrosouneas. 950-955, 1961. In: A. C. Sartorelli, and D. G. Johns (eds.),Handbookof Experimental 8. Iyer, V. N., and Szybalski, W. A Molecular Mechanism of Mitomycin Pharmacology, Vol. 38, pp. 65-84. Berlin: Springer-Verlag, 1975.

1454 CANCER RESEARCHVOL. 37

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1977 American Association for Cancer Research. Interstrand Cross-linking of DNA by 1,3-Bis(2-chloroethyl)-1-nitrosourea and Other 1-(2-haloethyl)-1-nitrosoureas

Kurt W. Kohn

Cancer Res 1977;37:1450-1454.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/37/5/1450

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/37/5/1450. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.