Chem. Res. Toxicol. 1989,2, 375-378
375
MNDO Calculations of Highly Mutagenic (Chloromethyl)benzo[ a Ipyrenes James C. Ball* and Irving T. Salmeen Scientific Research Laboratory, Ford Motor Company, Room S3083, P.O. Box 2053, Dearborn, Michigan 48121 Received April 19, 1989
(Chloromethyl)benzo[a]pyrenes (ClMeB[a]P) are among the most potent direct-acting mutagens (i.e., not requiring metabolic activation) detected by the Ames assay using strain TA98 and TA100. We wanted to determine if molecular quantum mechanics would yield insights into the reactivity and mutagenicity of these compounds. Calculations using the MNDO semiempirical method (MOPAC program) were carried out on a series of ClMeB[a]P. The experimental rates of solvolysis of ClMeB[a]P in 50% aqueous acetone correlated well (r = -0.95) with the differences between the calculated heats of formation of ClMeB[a]P and those of their respective carbocations, supporting the hypothesis that ClMeB[a]P react via carbocations. None of the calculated parameters correlated with mutagenicity. Thus, the prediction of mutagenicity of ClMeB[a]P in the Ames assay is more complex than the simple formation of carbocation intermediates.
Introduction Most chemicals that cause cancer or mutations do so through the reaction of the chemical itself, if the compound is a reactive electrophile, or through the reaction of a reactive intermediate (formed as the result of metabolic activation of the compound) with DNA. We are interested in the properties of the intermediates that ultimately react with DNA to yield mutations. Chloromethyl arenes are a class of compounds that can react directly with DNA, thereby avoiding the complications of metabolic activation. The simplest member of this class is benzyl chloride. Benzyl chloride has been reported to be carcinogenic (1) and weakly mutagenic, without metabolic activation, in Salmonella typhimurium strain TAlOO (2). We and others, however, have found that benzyl chloride is toxic but not mutagenic in either strain TAlOO or strain TA98 (3, 4). In addition, we have carried out Ames assays (strains TA98 and TAlOO without metabolic activation) on a wide variety of para-substituted benzyl chlorides (3),bifunctional (dhalomethyl)benzenes, and a-methylbenzyl halides (5). Of the compounds found to be mutagenic, two were nitrobenzyl derivatives which were mutagenic because of reduction of the nitro group by endogenous bacterial nitroreductases. One compound,p-acetoxybenzyl chloride, was mutagenic due to simple hydrolysis of the ester to yield a reactive intermediate, p-hydroxybenzyl chloride. The ortho-, meta-, and para-substituted (dichloromethy1)benzenes were weakly mutagenic. These experiments suggested that simple (chloromethy1)benzenes were only weakly mutagenic even when substituents imparted solvolytic reactivities ranging from extremely reactive to fairly stable (3, 5). Following our studies with benzyl chlorides, we then tested a series of (chloromethyl)benzo[a]pyrenes (ClMeB[a]P, chloromethyl group at positions 1,4-6, and 10-12) for mutagenicity in the Ames assay (without the addition of rat liver S9 mix) using strains TA98 and TA100. In contrast to the benzyl chlorides, we found that all of the ClMeB[a]P tested were mutagenic, and several were among the most potent direct-acting mutagens ever reported in the Ames assay (6, 7). In addition, 11- and 12-ClMeB[a]P were potent mutagens in Chinese hamster
V79 cells (8). Since the corresponding (hydroxymethy1)benzo[a]pyrenes were not mutagenic, we concluded that these ClMeB[a]P can react directly with DNA without the need for metabolic activation. Little is known, however, about the reactivity of these ClMeB[a]P compounds. There is good evidence that 6-C1MeB[a]P reacts with nucleophiles via a carbocation intermediate (9),and there are data from mass spectrometry experiments suggesting that the other six isomers also react via carbocation intermediates (7). Since ClMeB[a]P are substituted polycyclic aromatic hydrocarbons that can stabilize a carbocation through electronic delocalization of the charge, we wanted to investigate the possibility that molecular quantum mechanics could rationalize the reactivity of these compounds and yield insights into the mechanisms of mutagenicity of ClMeB[a]P.
Methods Calculations were carried out by using the MNDO (modified neglect of diatomic overlap) semiempirical method (10) as part of the MOPAC program (Program 455 MOPAC,Quantum Chemical Program Exchange Center, Indiana University). The program was modified to run on a Floating Point Systems computer (FPS-264). Complete geometry optimization was carried out by using standard bond angles and lengths for the initial structures. The aromatic rings were initially confined to a plane. The structure of the carbocation derived from l-ClMeB[a]P is shown in Figure 1to illustrate the numbering system used in this report.
Results and Discussion Heats of Formation. The calculated heats of formation for the twelve ClMeB[a]P (Table I) show that sterically crowded isomers have larger heats of formation than less crowded isomers. For example, the optimized structures for the crowded bay region (positions 10 and 11)ClMeB[a]P show that the aromatic rings are no longer planar but are twisted away from each other (Figures 2 and 3); these two isomers have the largest heats of formation. In contrast, the calculated heats of formation for the wbocations do not depend on steric crowding (Table I). These results are reasonable because the relative energies of the carbocations should depend more on the electronic delocalization of the resulting positive charge than on the strained
0 1989 American Chemical Society 0893-228~/89/2702-Q375~Q1.5Q/Q
Ball and Salmeen
376 Chem. Res. Toxicol., Vol. 2, No. 6, 1989 ~
Table I. Solvolvsisn and MNDO-Calculated Heats of Formation of (Chloromethylbenzola I ~ y r e n e s re1 energy difference, log rate constant of mf,CIMeB[nlP solvolysis of isomer mf,cIMeB[oPt mf,MeB[a]P cation* mf,MeB[n]P cation, ClMeB[a]P in 50% position kcal/mo/ kcal/mol kcal/mol aqueous acetonen 1 67.60 258.9 191.3 -1.72 268.3 203.9 64.40 2 192.1 259.8 67.73 3 265.5 197.5 -5.18 67.97 4 197.1 -5.55 265.4 68.26 5 257.8 184.5 -0.013 73.34 6 263.2 194.9 68.32 7 203.2 267.4 64.16 8 261.7 197.3 64.39 9 271.3 196.7 -5.57 74.64 10 197.0 -5.68 272.0 74.95 11 262.4 194.4 -5.38 68.04 12 Solvolysis data from Leon et al. (7).
A
B Figure 1. Structure and numbering of the carbocation derived from 1-(chloromethy1)benzo [a]pyrene.
Figure 3. Three-dimensional perspective display (Dissplay graphics software, Issco, CA) of the MNDO-calculatedoptimized structure for (A) 11-(chloromethyl)benzo[uJpyreneand (B)11methylbenzo[a]pyrene carbocation. The solid circles represent the positions of carbon atoms, open circles represent the positions of hydrogen atoms, and the concentric circles represent the position of the chlorine atom.
d
Figure 2. Three-dimensional perspective display (Dissplay graphics software, Issco, CA) of the MNDO-calculated optimized structure for (A) 10-(chloromethyl)benzo[a]pyreneand (B) 10methylbenzo[a]pyrene carbocation. The solid circles represent the positions of carbon atoms, open circles represent the p i t i o n s of hydrogen atoms, and the concentric circles represent the position of the chlorine atom. geometries. Thus, the heat of formation of the carbocation located at the S p i t i o n has the lowest energy even though the location of the exocyclic carbon is sterically crowded because of adjacent hydrogen atoms at positions 5 and 7. Silverman and Lowe have used the relative energies of carbocations as an indication of the relative ease of formation of carbocations (11). However, this approximation neglects effects of steric crowding on the heat of formation. As a result of the steric crowding observed with 6-, lo-, and ll-ClMeB[a]P, we could not use the heats of formation of t h e carbocations alone to estimate t h e ease of formation
8
Figure 4. Comparison of the rates of hydrolysis of ClMeB[aIP in 50% aqueous acetone (data from ref 7) with the calculated difference in energy between the CMeB[a]P and its respective carbocation (correlation coefficient = -0.95). The numbers adjacent to the boxes refer to the position of the chloromethyl group. of these ions. Therefore, t h e heat of formation of the parent ClMeB[a]P was also calculated, and the difference i n energy between t h e carbocations a n d t h e ClMeB[a]P
MNDO Calculations of (Chloromethyl)benzo[u]pyrenes
isomer position 1
2 3 4 5 6 7 8 9 10 11 12
Chem. Res. Toxicol., Vol. 2, No. 6,1989 377
Table 11. Mutagenicity and MNDO-Calculated Parameters for ClMeB[a IP Carbocations mutagenicity: charge on LUMO coeff of bond order revertanb/nmol P, orbital on energy of of exocyclic exocyclic TA98 TAl00 carbon exocyclic carbon LUMO, eV carbon-to-ring bond 120 75 +0.14 0.34 -6.00 1.74 +0.24 0.54 -6.25 1.64 +0.15 0.37 1.73 -6.02 19OOO 7200 +0.14 0.40 -6.15 1.69 0.40 -6.16 1.70 1800 1000 +0.15 0.22 39 48 +0.12 -6.04 1.75 0.36 -5.98 1.73 +0.15 0.52 -6.24 1.65 +0.22 0.39 -6.01 1.71 +0.17 0.30 -6.18 1.71 2000 3300 +0.17 69000 12000 +0.17 -6.21 0.38 1.69 0.37 1.72 43000 17000 +0.14 -6.05
"Mutagenicity data from Ball et al. (8).
was used to estimate the ease of formation of carbocations. For seven of the ClMeB[a]P, solvolytic reactivity (log of solvolysis rate constants) in 50% aqueous acetone correlates linearly (r = -0.95) with the difference between the calculated heats of formation of the ClMeB[a]P and those of their respective carbocations (Figure 4). These results give theoretical support for the hypothesis that the seven ClMeB[a]P tested react with nucleophiles via carbocation intermediates (7).The standard error in the slope of the line is 14% of the slope. The most likely source of error in this plot is the MNDO method and the parameters used to develop this method (10). Errors in the measurement of the rate constants should have a negligible effect on the slope because the rate constants were reproducible to with 10% (7). Geometry Optimization. Except for 6-, lo-, and 11ClMeB[a]P and their respective carbocations, the optimized structures for ClMeB[a]P and their respective carbocations show that the aromatic rings are planar to within 0.2', and the chlorine bond angles and distances are 113' and 1.7 A, respectively. The optimized structure of 6ClMeB[a]P shows that the exocyclic carbon is bent slightly out of the plane of the aromatic rings (approximately ll'), while the exocyclic carbon of the carbocation is bent about 33O out of the plane of the rings. The optimized structures for the bay-region ClMeB[a]P isomers and carbocations (Figures 2 and 3) show that the carbocations of these two ClMeB[a]P are twisted almost to the same extent as the parent molecule. Therefore, the loss of the chlorine atom would not result in the relief of a significant amount of steric strain, and we would not expect these ClMeB[a]P to be more reactive than other ClMeB[a]P isomers (7). Summary of Mutagenicity. The log of the rate constant for solvolysis of these ClMeB[a]P compounds showed a very poor linear correlation with mutagenicity when strain TA98 (r = 0.49) or strain TAlOO (r = 0.57) was used. In addition, none of the following calculated parameters showed a correlation with mutagenicity: the charge on the exocyclic carbon atom, the magnitudes of the LUMO coefficients of the P, orbital on the exocyclic carbon, or the bond orders of the exocyclic carbon-to-ring bonds (Table 11). In addition, there was no correlation with the ionization potentials (IP) of the parent ClMeB[a]P (IP ranged from 7.98 to 8.05 eV). The lack of correlation between calculated parameters of the reacting carbocation and mutagenicity suggests that the prediction of Ames assay mutagenicity of a compound is more complex than simply understanding the properties of the reactive intermediates. One possibility is that these compounds interact with DNA prior to reaction (12),and this preliminary binding to DNA must be an important step in the
mechanism by which these chemicals induce DNA damage.
Acknowledgment. We are grateful to Brent Besler for modifying MOPAC to run on the FPS-264 and for his helpful discussions. Registry No. 1-ClMeB[a]P, 100757-60-6; 2-C1MeB[a]P, 123358-36-1; 3-ClMeB[a]P, 123358-37-2; 4ClMeB[a]P, 94500-46-6; 5-C1MeB[a]P, 29852-26-4; 6-C1MeB[a]P, 49852-84-8; 7-ClMeB[alp, 123358-38-3; 8-ClMeB[a]P, 123358-39-4; 9-ClMeB[a]P, 123358-40-7; lO-ClMeB[a]P, 86803-19-2; 11-ClMeB[a]P, 9450054-6; 12-ClMeB[a]P, 94500-44-4; ClMeB[a]P, 123358-53-2; 1MeB[a]P cation, 123358-41-8; 2-MeB[a]P cation, 123358-42-9; 3-MeB[a]P cation, 123358-43-0; 4-MeB[a]P cation, 123358-44-1; 5-MeB[a]P cation, 123358-45-2; 6-MeB[a]P cation, 123358-46-3; 7-MeB[a]P cation, 123358-47-4;8-MeB[a]P cation, 123358-48-5; 9-MeB[a]P cation, 123358-49-6; lO-MeB[a]P cation, 12335850-9; 11-MeB[a]P cation, 123358-51-0; la-MeB[a]P cation, 123358-52-1; MeB[a]P cation, 123358-54-3.
References (1) Druckrey, H., Kruse, H., Preussmann, R., Ivankovic, S., and
Landschutz, C. (1970) Cancerogene alkyleirende substanzen, 3. Alkyl-halogenide, -sulfate, -sulfonate und ringgespannte heterocyclen. Z . Krebsforsch. 74, 241-273. (2) McCann, J., Spingarn, N. E., Korboi, J., and Ames, B. N. (1975) Detection of carcinogens as mutagens: bacterial tester strains with R factor plasmids. Proc. Natl. Acad. Sci. U.S.A. 72, 979-983. (3) Ball, J. C., Foxall-VanAken, S., and Jensen, T. E. (1984) Mutagenicity Studies of p-substituted benzyl derivatives in the Ames Salmonella plate-incorporation assay. Mutat. Res. 138,145-151. (4) Zeiger, E.; Anderson, B., Haworth, S., Lawlor, T., Mortelman, K., and Speck, W. (1987) Salmonella mutagenicity test: 111. Results from testing of 255 Chemicals. Environ. Mutagen. 9 (Suppl. 91, 1-109. (5) Ball, J. C., and Young, W. C. (1987) Mutagenicity studies of cup'-dihalomethylbenzenes, a-methylbenzyl halides, and 1chloromethylpyrene in the Ames assay. Mutat. Res. 191,79-84. (6) Ball, J. C., Foxall-VanAken, S., Leon, A. A., Vander Jagt, D. L., and Daub, G. H. (1984) 11-and 12-Chloromethylbenzo[a]pyrenes are potent direct acting- mutagens - in the Ames assav. Mutat. Res. 142; 141-144. (7) Leon, A. A., Ball, J. C., Foxall-VanAken, S., Daub, G. H., and Vander Ja& D. L. (1985) Chemical Reactivities and mutaeenicities of a series of chloromethylbenzo[a]pyrenes. Chem.-Biol. Znteract. 56, 101-111. (8) Ball, J. C., Leon, A. A., Foxall-VanAken, S., Zacmanidis, P., and Vander Jagt, D. L. (1988) Mutagenicity of chloromethylbenzo[alpyrenes in the Ames assay and in Chinese hamster V79 cells. In Polynuclear Aromatic Hydrocarbons: A Decade of Progress (Cooke, M., and Dennis, A. J., Eds.) pp 41-58, Battelle press, Columbus, OH. (9) Royer, R. E., Daub, G. H., and Vander Jagt, D. L. (1979) Solvolytic reactivity of 6-chloromethylbenzo[a]pyreneand selectivity of trapping of the arylmethyl cation by added nucleophiles. J. Org. Chem. 44,3196-3201. (10) Dewar, M. J. S., and Thiel, W. (1977) Ground states of molecules. 38. The MNDO method. Approximations and Parameters.
378 Chem. Res. Toxicol., Vol. 2, No. 6, 1989 J. Am. Chem. SOC.99,4899-4907. (11) Silverman, B. D., and Lowe, J. P. (1988) Calculated reactivities of the ultimate carcinogens of polycyclic aromatic hydrocarbons: A useful too in predicting structure-activity relationships. In Polycyclic Aromatic Hydrocarbon Carcinogenesis: Structure-
Ball and Salmeen Actiuity Relationships (Yang, S. K., and Silverman, B. D., Eds.) Vol. 11, pp 89-109, CRC Press, Boca Raton, FL. (12) Geacintov, N. E. (1986) Is intercalation a critical factor in the covalent binding of mutagenic and tumorigenic polycyclic aromatic diol epoxides to DNA? Carcinogenesis 7, 759-766.
