Chem. Res. Toxicol. 1991,4, 368-372
368
lung with cigarette smoking. Nature 336, 790-792. (14) Randerath, E., Miller, R. H., Mittal, D., Avitta, T. A., Dunaford, H. A., and Randerath, K. (1989) Covalent DNA damage in tissues of cigarette smokers as determined by 32P-postlabelingassay. J. Natl. Cancer Inst. 81, 341-347. (15) van Schooten, F. J., Hillebrand, M. J. X., van Leeuwen, F. E., Lutgerink, J. T., van Zandwijk, N., Jansen, H. M., and Kriek, E. (1990) Polycyclic aromatic hydrocarbon-DNA adducts in lung tissue for lung cancer patients. Carcinogenesis 11, 1677-1681.
(16) Wilson, V. L., Weston, A., Manchester, D. K., Trivers, G. E., Roberta, D. W., Kadlubar, F. F., Wild, C. P., Montesano, R., Willey, J. C., Mann, D. L., and Harris, C. C. (1989) Alkyl and aryl carcinogen adducts detected in human peripheral lung. Carcinogenesis 10, 2149-2153. (17) Shields, P. G., Povey, A. C., Wilson, V. L., Weston, A,, and Harris, C. C. (1990) Combined high-performanceliquid chromat~graphy/~~P-postlabeling assay of N7-methyldeoxyguanosine. Cancer Res. 50, 6580-6584.
Theoretical Structure-Activity Study of Mutagenic Allyl Chlorides Donald W. Boerth,**tErwin Eder,*J Golam Rasul,? and Joseph Moraist Department of Chemistry, Southeastern Massachusetts University, North Dartmouth, Massachusetts 02747,and Institute of Toxicology, University of Wurzburg, 0-8700 Wurzburg, FRG Received April 2, 1990 Ab initio molecular orbital calculations at the STO-3G and 3-21G levels were used t o clarify the mechanism of mutagenic action of various substituted allyl chlorides. Computed molecular properties were compared with experimental mutagenic potentials of these allylic compounds. In agreement with experiment, the computational results suggest that the primary mechanism of action involves the SN1 formation of allylic cations which then react with nucleophilic centers of nucleic acid bases. The usefulness of computed properties in estimating the degree of alkylating activity and mutagenicity was evaluated. In general, stability of the allylic carbocation intermediate and the degree of charge delocalization in the allyl system correlate well with observed mutagenic potentials. Allylic compounds appear in the environment from a wide variety of industrial sources as raw materials, intermediates compounds, and products in the production of agricultural products, food additives, cosmetics, and pharmaceuticals. Mutagenic and carcinogenic properties have been attributed to some of them (1-6). Of these, allyl chloride and substituted allyl chlorides have been systematically investigated with regard to mutagenic and genotoxic potentials and the mechanism of action (4-6). In general, the mutagenic or genotoxic process proceeds from alkylation of DNA or RNA bases at nucleophilic centers which are involved in base pairing. Alkylation produces adducts which no longer pair in the usual manner, with attendant errors in replication or transcription of the base code. Several direct and indirect mechanisms were proposed and considered. These included (a) direct alkylation by nucleophilic substitution (Scheme I), (b) alkylation by formation of allyl free radicals, (c) hydrolysis of the allyl chlorides followed by alcohol dehydrogenase oxidation of the allylic alcohol intermediates (the a,fl-unsaturated carbonyl compounds are known carcinogens), (d) enzymatic epoxidation, followed by rearrangement to reactive a-chloroketones, or (e) metabolic activation by sulfotransferase or acetyltransferase. Specifically in the previous work of Eder, Henschler, and Neudecker (5), a series of substituted allylic compounds were subjected to microbial assay by the procedures of Ames (7). The resulting mutagenic potentials were compared to simple alkylating reactivity of these compounds with I-(p-nitrobenzy1)pyridine (NBP). An overall good correlation was obtained between mutagenicity and alkylating activity for substituted allyl chlorides. The results strongly suggested
* Authors to whom correspondence should be addressed. t Southeastern
Massachueetta University.
* University of Wurzburg.
Scheme I R2
R2 R3
R3
I
I
t
R,HC=C-CHX
I
R,H$$--CH
R2
I
I
-----
R3
I
RtHCC-CH -.
+
+
+ X-
X-
s I
- -.-+-
R2
R3
I
RIHCC-CH -.
-+
:BH
t
I
RlHC=C-CH
Rz
R3
I
I B
I
R3
I
of RlHCC=CH
I
+ H+
0
that alkylation by nucleophilic substitution was the predominant mechanism of action. The patterns apparent in the reactivity-mutagenicity correlation could be attributed to two factors: (1) the structure of the allyl chloride and (2) the nature of the leaving group. Although an SN1process was implicated, the NBP test unfortunately does not permit the distinction between s N 1 and sN2 mechanisms. In this study we have attempted to investigate the relationship of mutagenicity and structure, considering only allylic compounds with the same leaving group. Qualitatively, the experimental correlation of Eder, Henschler, and Neudecker (5) suggested that the activity of the allyl chlorides could be generally related to the anticipated stability of the corresponding allylic carbocations that might be generated as intermediates in these reactions. We have, therefore, pursued a correlation of mutagenicity and alkylating activity with parameters related to these allylic carbocotions, which can be calculated theoretically by application of SCF-MOtheory at the ab initio level. This should provide clarification of the mechanism involved and an understanding of the factors contributing to mutagenic activity. If successful, such an investigation should also serve as a potential vehicle for the development of methods for tentative and preliminary predictions about the activity of suspect compounds.
0893-228x/9~/2~0~-0368~~2.50/0 0 1991 American Chemical Society
Chem. Res. Toxicol., Vol. 4, No. 3, 1991 369
Mutagenic Allyl Chlorides
Experimental Section The experimental methods have been previously described (4-5). In brief, the microbial assay was carried out by incubation of Salmonella typhimurium strain TA 100 with the alkylating compound. The mutagenic potentials (revertanta/pmol)were determined from the slope of the linear portion of doseresponse curves. Alkylating activity was measured by reacting the alkylating compound with 4(pnitrobenyl)pyridine (NBP), yielding a chromophoric product whose molar absorbance can be measured spectrophotometricallyat 560 nm. The molar absorbance has been shown to be related to alkylating potency (4a).
."'i
W
70.0
P 2-21 c9
-C
Computatlonal Methods The bulk of the calculations reported here were carried out with the GAUSSIAN series of programs (8)on VAX 8650 and VAX 8700 mainframes operating with the VMS operating system. The STO-3G basis set (9,lO)and the extended 3-21G basis set (11, 12) were selected for the calculations. Both standard and optimized geometries were utilized. Internal coordinates for the standard geometries are as follows: r(C-C) = 1.54 A, r(C=C) = 1.34 A, r(Cs,,rH) = 1.09 A, r ( C 6 H ) = 1.08 A. Strict tetrahedral and trigonal geometries were maintained for the appropriate carbon centers. Dihedral angles were determined by partial geometry optimization. [Dihedral angles: 2-methylpropenyl cation, ~ c l C 2 C 4 H 488.8'; 2-methyl-2-butenyl cation 0.0';~ H 2-butenyl ~ (trans), ~ c l C 2 C 5 H 5O.O', L H ~ C ~ C cation (trans), LH3C3C4H4 60.0'.] Optimized geometries were obtained a t the 3-21G level, which has been shown to give generally good molecular structures. Total optimizations of all coordinates were carried by the gradient method of Schlegel (13).
-??O.
R2
I
R3
I
RiHC=C-CHX
P';.
+ CH3* t CH3X + R l H C g r - C H +
woi.o= EMd(R+) + EMd(CH3X) - EMd(RX) - EMd(CH;) The results of such calculations are tabulated in Table I. Figure 1 shows moderate agreement between calculated AHoh values for the various substituted allyl chlorides and their corresponding nucleophilic reactivities toward NBP, except for cis- and tram-l,3-dichloropropene. This dem-
40.
50.
60.
70.
-DELTA H ( i s o ) Figure 1. Isodesmic enthalpy and alkylating activity by NBP test.
r
r
Results and Discussion The experimental relationship between mutagenicity and nucleophilic reactivity (5) exhibits a general trend which would seem to be related to the effect of substituents on the stability of the allyl carbocation species derived from the corresponding allylic chloride. Such allylic carbocations would presumably arise from an SN1-typeprocess (Scheme I) involving dissociation through a contact or solvent-separated ion pair. The more stable carbocations are formed more readily and hence display greater alkylating reactivity toward nucleophiles, including nucleic acid bases with potentially mutagenic or genotoxic outcomes. We have, therefore, proceeded to calculate parameters related to cation stability for comparison with the experimental findings. Since the carbocation stabilities or heats of formation cannot be calculated accurately with small basis sets due to the inability to account for bond dissociation energies a t the Hartree-Fock level, we have, instead, used the standard approach of calculating enthalpies of isodesmic reactions (141, where these errors cancel for the most part. For our purposes, the following reaction provides such an isodesmic measure of the relative stabilities of the cations from their precursors.
-
"$0.
A
1
l-cnOID-2-MEfs
'
I
50. 60. -DELTA H ( i s o )
40.
70.
Figure 2. Isodesmic enthalpy and mutagenic potential. onstrates that the reactivities of these two isomers toward ~ The NBP depend predominantly on their S Nreactivity. pattern for the enthalpies and NBP values for methylsubstituted allyl compounds is consistent with stabilization of allylic cations by electron donation by methyl groups (or other alkyl groups). Simple resonance considerations suggest that methyl substitution will have a greater effect in stabilizating an allyl cation at C1 or C3. This is reproduced nicely by the computational results.
Additional alkyl groups are predicted to have an additive effect. A comparison of experimental mutagenicity and Mob values (Figure 2) shows a reasonable linear correlation for a number of the allylic chlorides. The mutagenicity seems generally to increase with the greater stability of the carbocation intermediate. Except for unsubstituted allyl chloride, the compounds which deviate in figure 2 follow a pattern which is strikingly similar to that shown in the study where NBP activity and mutagenic potential were compared (5). In the present and previous studies, 2,3-
370 Chem. Res. Toxicol., Vol. 4, No. 3, 1991
Boerth et al.
Table I. Computed Parameters for Allyl Chlorides mutagenic potential, DHobl(l correspond- revertants/ allyl chloride ing cation pmol kcal/mol ELUMOo (3-21G) Q,+,b 3-chloropropene 9 -50.47 -0.2247 0.1003 -0.1897 trans-l-chloro-2-methyl-2-butene 52 -65.20 0.0505
4
Q,b
QCIH+CIH'
0.5330
0.9282
0.4222
0.6900
3-chloro-2-methyl-1-butene
+
52 58
-63.28
-0.1897
0.0505
0.4222
0.6900
cis-1-chloro-2-butene
p
270
-60.03
-0.2063
0.0648
0.4549
0.7563
trans-1-chloro-2-butene
w
270
-62.88
-0.1988
0.0666
0.4564
0.7413
65
-51.86
-0.2150
0.0844
0.5008
0.8656
19
-42.68
-0.2348
0.0913
0.5135
0.8325
trans-1,3-dichloropropene
28
-44.57
-0.2297
0.0871
0.4961
0.8017
2,3-dichloropropene
26
-38.88
-0.2414
0.1048
0.5346
0.9940
cis-l-chloro-2-methyl-2-butene
3-chloro-2-methyl-1-propene
cis-1,3-dichloropropene
.'+'..
-60.94
-0.1986
0.0477
0.4196
0.7050
&
A
p
CI
3-21G//3-21Gresults. *STO-3G//3-21G results.
dichloropropene exhibits a positive deviation, and 1chloro-2-methyl-2-butene,a negative deviation. A point for 3-chloro-2-methyl-l-butene, which could not be plotted previously due to its inability to give an NBP test, has been added to Figure 2. It falls very near the point for 1chloro-2-methyl-2-butene,which is not surprising since both 3-chloro-2-methyl-1-butene and 1-chloro-2-methyl2-butene would produce the same cation intermediate. Both this theoretical study and the previous experimental study point to other mechanisms or reactions (discussed below) as responsible for the deviations. 2,3-Dichloro-l-propene is more mutagenic than would be expected if only a nucleophilic mechanism was involved. Its activity is attributed to other mechanisms (5). This was demonstrated by a modification of the Ames procedure in which mutagenic potentials are compared with or without a rat liver metabolic activating enzyme, the socalled S-9 mix. The results allow one to distinguish between direct and indirect mutagenic mechanisms. Of all the allylic compounds tested, only 2,3-dichloro-l-propene was enzymatically activated, indicating that a nondirect nucleophilic mechanism is playing an important role. Three enzyme activation pathways were thought plausible: (a) epoxidation followed by rearrangement to a highly reactive, highly alkylating 1,3-dichloroacetone, (b) hydrolysis to an allyl alcohol followed by oxidation to highly alkylating 2-chloroacrolein by an alcohol dehydrogenase, or (c) reaction with glutathione or cysteine to give a reactive mustard-like substance (15). Hence, it is highly unlikely that 2,3-dichloropropene will correlate on any comparison of mutagenicity and nucleophilic activity. In the case of 3-chloro-2-methyl-1-buteneand 1chloro-2-methyl-2-buteneboth experimental reactivity and computed stability values indicate that the resulting common cation should be very reactive. Despite this stability
and the good correlation of mi, with the NBP test, both compounds are much less mutagenic than would be expected. This could be due to several factors: (1)this cation is more sterically hindered than others in the series, and, therefore, may react more slowly with nucleic acid bases, (2) hydrolysis may be more rapid than reaction with DNA bases, and (3) both compounds may also undergo facile elimination to a less genotoxic 2-methyl-l,3-butadienea The foregoing discussion suggests that, excluding steric and enzymatic effects, the stability of a given carbocation is important in determining the reactivity of the parent compound. Charge delocalization has long been considered a very potent factor in stabilization of carbocations. Therefore, charge distribution in the allyl system was also compared with mutagenicity (Table I). We have taken several approaches to the charge-reactivity relationship. The first of these involves the charge (QvH.dyl) at the termini of the allyl system, which are the likely sites for reaction with nucleopilic centers on the DNA/RNA bases. The values for QCH.dyl are determined as the sum of charges at the ends of the allylic cation system including carbons and pendant hydrogens. Since the basis sets make use of attached hydrogens for charge distribution, the QCH values give a better representation of charge at the reacting center. The linear relationship between mutagenic potential and allyl charge QCH.dyl or l/QCHdyl is a reasonable outcome (Figure 3). The correlation is good for a number of compounds, including allyl cation, demonstrating that decreased allyl charge at the reacting centers leads to more stable cations. An SN1interpretation would further suggest that the more stable the cationa, the more easily it will form and hence the more reactive or mutagenic it will become. It is interesting to point out that the cations derived from the three compounds, 2,3-dichloro-l-propene, 3-chloro-2-methyl-l-butene, and l-chloro-2-methyl-2butene, again do not correlate presumably for the same reasons elaborated earlier. Two other approaches to the charge-reactivity analysis make use of the total carbon charge Q and the *-charge, Q,, of the least hindered, and hence more reactive, of the two allyl carbons. Plots of the reciprocal of Q or Q, with
Chem. Res. Toricol., Vol. 4, No. 3, 1991 371
Mutagenic Allyl Chlorides
Standard Geometry
Optimized Geometry
3.0 n
I-C)LmC-.?-MM
p'
00.-
-
1.81.32.02.12.21.8 l/Q(pi)
2.0
1/Q(pi
2.2
2.4
)
Figure 5. Mutagenic potential and a-charge at reacting allyl l/Q(CH-allyl>
(STO-3G//3-21G) Figure 3. Mutagenic potential and charge (C + H at both allyl
3.
centers).
Standard Geometry
0 r
Optimized Geometry
2.
I l--e-mY
1.
?LL 0.08.
10.
12.
14.
l/Q(total)
16. 8. 10. 12. 14. 16. 18. 20. l/Q(total)
Figure 4. Mutagenic potential and ( u + *)-charge at reacting allyl center.
the logarithm of mutagenic potential show again the same type of linear relationship for all compounds (Figures 4 and 5). The only exceptions are consistently 2,3-diand 1chloro-1-propene, 3-chloro-2-methyl-l-butene, chloro-2-methyl-2-butene, which are plagued by the problems described earlier. It is noteworthy that both standard and optimized geometries used for the STO-3G calculations give equally goad correlations. Although the absolute values of Q or Q, are slightly different for some cations, the fact that the general trends are reproducible and are not affected by choice of geometry is a pleasing one and simplifies the process of obtaining computed values for cation properties. Finally, we have employed a HSAB (hard and soft acid-base) approach (16, 17) to further elaborate the chemical nature of the reacting cation species. The use of molecular orbital theory to classify reactions between Lewis acids (or electrophiles) and Lewis bases (or nucleophiles) was developed by Klopmam (17),the details of which are presented elsewhere (18-20). The orbital interactions between the highest occupied molecular orbital (HOMO) of the base and the lowest unoccupied molecular orbital (LUMO) of the acid have been used to describe site selectivity and carcinogenic activities of several species (21-24), in addition to reactivity of simple anionic nucleophiles and cationic electrophiles (17). Since our allyl cations are electron pair acceptors in the reaction process,
0.01:
:';:I
: ( : ; I
E
: : : ; ! : : : :
(Luna)
Figure 6. Mutagenic potential and LUMO energy of the allyl STO-3G//3-21Gresults. cation. ( 0 )3-21G//3-21G results. (0) Key: (1) 3-chloropropene; (2) cis-1,3-dichloropropene;(3) trans-l,3-dichloropropene; (4) 2,3-dichloropropene;(5) l-chloro(7) 3-chloro2-methyl-2-butene; (6) 3-chloro-2-methyl-1-butene; 2-methyl-1-propene; (8) cis-1-chloro-2-butene;(9) trans-lchloro-2-butene.
they are Lewis acids. We have, therefore, considered the energy of the LUMO of the carbocation relative to the mutagenic potential. A generally good linear relationship and mutagenic potential is between the computed ELUMo obtained (Figure 6) with both STO-3G and 3-21G wave functions. The same three compounds mentioned before are consistently displaced from the correlation line in a similar fashion to what has been observed before. Unfortunately, the unsubstituted allyl cation is predicted to have a LUMO energy which is too high for it to correlate well with the other substituted allyl cations.
Conclusions The relationship between mutagenicity of allyl chlorides, experimental alkylating activity, and the stability of incipient carbocation species is consistent with the previously established nucleophilic alkylation mechanism as the primary process for initiation of mutagenic or carcinogenic activity for most of these compounds. The formerly used NBP test gave only information about general alkylating activity and was less specific (or nonspecific) about the mechanism or origin of the activity. In contrast, the direct correlation between mutagenic potential and computed
372 Chem. Res. Toxicol., Vol. 4, No. 3, 1991
properties of the carbocation intermediate, formed in an s N 1 rate-determining step, suggest8 that, for these allylic systems, an s N 1 process is largely responsible for the observed mutagenic activity of the agents. Furthermore, pending further development, this MO methodology has the potential of providing a convenient method of prescreening suspect cancer-causing or mutagenic agents based on computed parameters associated with the carbocation intermediates.
Acknowledgment. We acknowledge the funding of computer hardware for use in this research from the SMU University Research Committee and the support of the Department of Chemistry, SMU, in the acquisition of software for this research. Generous donations of computer time and other services of the Academic Computer Services and Central Computer Services at Southeastern Massachusetts University are also greatly appreciated. Portions of this research were also supported by the Deutsche Forschungsgemeinschaft (SFB 172). Registry No. 3-Chloropropene, 107-05-1; tram-l-chloro-2methyl-&butene, 23009-73-6; cis-l-chloro-2-methyl-2-butene, 23009-74-7; 3-chloro-2-methyl-l-butene, 5166-35-8;ck-l-chloro2-butene, 4628-21-1; trans-l-chloro-2-butene,4894-61-5; 3chloro-2-methyl-l-propene, 563-47-3; cis-1,3-dichloropropene, 10061-01-5; trans-1,3-dichloropropene, 10061-02-6; 2,3-dichloropropene, 78-88-6. Supplementary Material Available: 3-21G optimized molecular structures for the allyl chlorides (Figure l ) and allyl cations (Figure 2) (4 pages). Ordering information is given on any current masthead page.
References (1) McCoy, E. C., Burrows, L., and Rosenkranz, H. S. (1978)Genetic activity of allyl chloride. Mutat. Res. 57, 11-15. (2) Van Duuren, B. L., Goldschmidt, B. M., Loewengart, G., Smith, A. C., Melchionne, S., Seidman, J., and Roth, D. (1979) Carcinogenicity of halogenated olefinic and aliphatic hydrocarbons in mice. J. Natl. Cancer Znst. 63, 1433-1439. (3) Long, E. L., Nelson, A. A,, Fitzhugh, 0. G., and Hansen, W. H. (1963) Liver tumors produced in rata by feeding safrole. Arch. Pathol. 75, 595-604. (4) (a) Eder, E., Neudecker, T., Lutz, D., and Henschler, D. (1980) Mutagenic potential of allyl and allylic compounds, structureactivity relationship as determined by alkylating and direct in vitro mutagenic properties. Biochem. Pharmacol. 29, 993-998. (b) Neudecker, T., Lutz, D., Eder, E., and Henschler, D. (1980) Structure-activity relationship in halogen- and alkyl-substituted allyl and allylic compounds: correlation of alkylating and mutagenic properties. Biochem. Pharmacol. 29, 2611-2617. (5) Eder, E., Henschler, D., and Neudecker, T. (1982) Mutagenic properties of allylic and a,p-unsaturated compounds: consideration of alkylating mechanisms. Xenobiotica 12, 831-848. (6) Lutz, D., Eder, E., Neudecker, T., and Henschler, D. (1982) Structure-mutagenicity relationship in a,&unsaturatd carbonylic compounds and their corresDondine allvlic alcohols. Mutat. Res. 93, 305-315,
Boerth et al. (7) Ames, B. N., McCann, J., and Yamasaki, E. (1975) Methods for detecting carcinogens and mutagens with the Salmonella/mammalian-microsome mutagenicity test. Mutat. Res. 31, 347-364. (8) Frisch, M. J., Binkley, J. S., Schlegel, H. B., Raghavachari, K., Melius, C. F., Martin, R. L., Stewart, J. J. P., Bobrowicz, F. W., Rohlfing, C. M., Kahn, L. R., Defrees, D. J., Seeger, R., Whiteside, R. A., Fox, D. J., Fleuder, E. M., and Pople, J. A. Gaussian 86, Release C, Canegie-MellonQuantum Chemistry Publishing Unit, Pittsburgh, 1984. (9) Hehre, W. J., Sewart, R. F., and Pople, J. A. (1969) Self-consistent molecular orbital methods. I. Use of Gaussian expansions of Slater-type orbitals. J. Chem. Phys. 51, 2657-2664. (10) Hehre, W. J., Ditchfield, R., Stewart, R. F., and Pople, J. A. (1970) Self-consistent molecular orbital methods. IV. Use of Gaussian expansions of Slater-type orbitals. Extension to second-row molecules. J. Chem. Phys. 52, 2769-2773. (11) Binkley, J. S., Pople, J. A., and Hehre, W. J. (1980) Self-consistent molecular orbital methods. 21. Small split-valence basis sets for first-row elements. J. Am. Chem. SOC. 102, 939-947. (12) Gordon, M. S., Binkley, J. A,, Pople, J. A,, Pietro, W. J., and Hehre, W. J. (1982) Self-consistentmolecular orbital methods. 22. Small split-valance basis seta for second-row elements. J. Am. Chem. SOC. 104, 2797-2803. (13) Schlegel, H. B. (1982) Optimization of equilibrium geometries and transition structures. J. Comput. Chem. 3, 214-218. (14) Hehre, W. J., Ditchfield, R., Radom, L., and Pople, J. A. (1970) Molecular orbital theory of the electronic structure of organic compounds. V. Molecular theory of bond separation. J. Am. Chem. SOC. 92,4796-4801. (15) Eder, E., and Dornbusch, K. (1988) Metabolism of 2,3-dichloro-1-propene in the rat. Consideration of bioactivation mechanisms. Drug Metab. Dispos. 16,60-68. (16) Pearson, R. G. (1963) Hard and soft acids and bases. J. Am. Chem. SOC. 85,3533-3539. (17) Klopman, G. (1968) Chemical reactivity and the concept of 90, charge- and frontier-controlled reactions. J. Am. Chem. SOC. 223-234. (18) Fleming, I. (1976) Frontier Orbitals and Organic Chemical Reactions, pp 34-40, Wiley & Sons, London. (19) Salem, L. (1982) Electrons in Chemical Reactions: First Principles, pp 166-169, Wiley-Interscience, New York, NY. (20) Lowry, T. H., and Richardson, K. S. (1987) Mechanism and Theory in Organic Chemistry, pp 318-322, Harper & Row, New York, NY. (21) Lown, J. W., Koganty, R. R., Bhat, U. G., Chauhan, S. M. S., Sapse, A,-M., and Allen, E. B. (1986) Isolation and characterization of electrophiles from 2-haloethylnitrosoureas forming cytotoxic DNA cross-links and cyclic nucleotide adducts and the analysis of base site-selectivity by ab initio calculations. ZARC Sci. h b l . 70, 129-136. (22) Lown, J. W., Chauhan, S. M. S., Koganty, R. R., and Sapse, A,-M. (1984) Alkyldinitrogen species implicated in the carcinogenic, mutagenic and anticancer activities of N-nitroso compounds. Characterization by 16N-NMR of l6N-enriched compounds and analysis of DNA base site-selectivity by ab initio 106, 6401-6408. calculations. J. Am. Chem. SOC. (23) McCoy, E. C., Holloway, M., Frierson, M., Klopman, G., Mermelstein, R., and Rosenkranz, H. S. (1985) Genetic and quantum chemical basis of the mutagenicity of nitroarenes for adeninethymine base pairs Mutat. Res. 149, 311-319. (24) Klopman, G., Tonucci, D. A,, Holloway, M., and Rosenkranz, H. S. (1984) Relationship between polarographic reduction potential and mutagenicity of nitroarenes. Mutat. Res. 126,139-144.