Regiochemistry of Nucleophilic Attack by the ... - ACS Publications

Derived from Carcinogenic Polycyclic Arylamines and. Nitroarenes: Molecular Orbital Calculations and Simple. Models. George P. Ford* and Jon W. Thomps...
0 downloads 0 Views 176KB Size
Chem. Res. Toxicol. 1999, 12, 53-59

53

Regiochemistry of Nucleophilic Attack by the Guanine 2-Amino Group at the Ring Positions of Nitrenium Ions Derived from Carcinogenic Polycyclic Arylamines and Nitroarenes: Molecular Orbital Calculations and Simple Models George P. Ford* and Jon W. Thompson Department of Chemistry, Southern Methodist University, Dallas, Texas 75275-0314 Received June 19, 1998

Semiempirical AM1 molecular orbital calculations are used to compute the energetics of addition of the guanine 2-amino group to alternative ring positions of aryl nitrenium ions with the general structure ArNH+, where Ar is the phenyl and various positional isomers of the naphthyl, pyrenyl, and benzo[a]pyrenyl groups. The syn or anti orientation of the NH+ group, and factors akin to classical localization energies, are identified as key components of the differential energetics of addition to alternative ring sites. The regiochemistry predicted by the AM1 method can be qualitatively reproduced using simple HMO calculations that require trivial computational effort and, almost as well, using PMO theory that does not require the use of a computer at all. In the latter approach, the most reactive ring positions are predicted to be those where the nonbonding orbital coefficients, a0r, in the analogous odd alternant hydrocarbons are largest. These results are discussed in relation to the available experimental data for the formation of deoxyguanosin-2-yl adducts when DNA is exposed to presumed nitrenium ion precursors.

Introduction Polycyclic aromatic amines and nitro compounds constitute a large class of compounds, many of which are either industrial or environmental carcinogens (1-4). They are converted in vivo to aryl hydroxylamines and aryl hydroxylamine esters which spontaneously decompose to highly reactive aryl nitrenium ions (5, 6) (Figure 1). The latter react with the nucleic acid base sites, leading to a wide variety of adduct types (7-10). The formation of such covalent adducts is believed to be responsible for the carcinogenic and mutagenic effects of these compounds (5). Two types of adducts are most commonly encountered in vivo. In one, the nitrogen atom of the nitrenium ion is bound to the guanine 8-position. In the other, one of the nitrenium ion ring carbons is bound to either an adenine or guanine amino group. However, other examples are known involving C-O, N-O, or N-N linkages as well as a variety of ring-opened and other products (8). The more common adduct types exemplified by 1-3 are found in the in vivo reactions of 2-aminonaphthalene (11), or its N-hydroxy (12, 13) or N-acetoxy derivative (14), with DNA under a variety of conditions. The belief that the genetic consequences of DNA modification will depend on both the kinds and the amounts of adducts formed continues to stimulate efforts to elucidate and rationalize the complex regiochemistry observed in carcinogen-DNA adduct formation (15). While some useful generalizations have emerged (16, 17), even a modestly * To whom correspondence [email protected].

should

be

addressed.

E-mail:

Figure 1. (a) In vivo formation of aryl nitrenium ions from nitroarenes or aminoarenes. (b) Formation of (deoxyguanosinN2-yl)aminoarene adducts via nucleophilic addition followed by proton loss and rearomatization.

predictive capability has so far been achieved only for DNA alkylation (18, 19). The problem is more complex in the case of the aryl nitrenium ions for at least two reasons. First, the relative reactivities of individual base sites toward alkylating agents are qualitatively similar regardless of whether the reactions take place with the deoxynucleoside, with RNA, or with single- or doublestranded DNA (20). This is not true for nitrenium ion precursors where the relative amounts of guanine C8 and

10.1021/tx9801460 CCC: $18.00 © 1999 American Chemical Society Published on Web 12/12/1998

54 Chem. Res. Toxicol., Vol. 12, No. 1, 1999

Ford and Thompson 0.5β0 and 1.0β0, respectively (31). However, the values of Lr were found to be rather insensitive to the exact parametrization.

Results and Discussion

N2 adducts are very sensitive to these factors (21). Second, the reactions of most alkylating agents take place through a single electrophilic center. Thus, the regiochemical outcome reflects only the possibility of reaction at alternative base sites. In the case of aryl nitrenium ions, the problem is multiplied by the possibility of reaction either at the exocyclic nitrogen atom or at one of several alternative ring positions. The relative reactivities of alternative ring positions were considered many years ago by Ford and Scribner (22). They suggested that, in the absence of other factors, nucleophilic attack should occur at those positions where the nonbonding orbital coefficients in the corresponding benzylic cations were largest. If correct, this would be particularly convenient since these quantities are easily obtained using simple pencil-and-paper methods (23). At that time, suitable experimental data for properly testing this idea were unavailable. For example, while the method correctly predicts that ring substitution in the 2-naphthyl nitrenium ion should lead to structures such as 1 and 3, it could be countered that the observed regiochemistry simply reflected the tendency of the nucleophile to add in the vicinity of the formally cationic NH+ group. More recently, several fascinating examples have been discovered by Fu and co-workers (24-26) in which ring substitution takes place at sites remote from the NH+ group. These seemed to provide an ideal framework within which to reexamine the fundamental determinants of this aspect of this complex regiochemical problem. In the work described here, we have used semiempirical molecular orbital calculations (i) to clarify the basic features of the nucleophilic attack of the guanine amino group at the ring positions of representative polycyclic aryl nitrenium ions, (ii) to search for simple relationships between the computed activation barriers and more readily accessible reactivity indices, and (iii) to ascertain the extent to which the predicted regiochemistry of nucleophilic attack coincides with the known reactions of presumed nitrenium ion precursors with guanine residues in DNA.

Computational Methods Semiempirical molecular orbital calculations were carried out at the AM1 level (27) using version 5 of the Spartan package (28) and refer to fully optimized structures. Spartan’s automatic conformational search facility was used extensively to ensure all reported data correspond to the lowest-energy conformers (see below). Transition state geometries were located using Baker’s eigenvector following algorithm (29) and were identified as genuine first-order saddle points by recalculating the force constant matrixes and verifying the presence of a single negative eigenvalue. Simple Hu¨ckel π-localization energies, Lr, were computed using version 2.0.3 of HMO-plus (30) on a Power Macintosh. The Hu¨ckel parameters for nitrogen, hN and βCN, were set to

Syn and Anti Nitrenium Ion Configurations. Although aryl nitrenium ions are formally derivatives of NH2+, extensive delocalization of the cationic charge leads to structures that more closely resemble imino carbenium ions (32). As with imines, alternative orientations of the NH bond give rise to two distinct configurational isomers separated by substantial activation barriers. Recent ab initio calculations on 19 polycyclic ions with the general structure ArNH+ led to values of ∆Hq in the range of 27.8 ( 2.0 kcal mol-1 (33), consistent with the value of 26 kcal mol-1 obtained by Cramer et al. (34) in very high level ab inito and DFT calculations for the degenerate process in PhNH+. If a barrier of approximately this magnitude applies in aqueous solutions, there should be essentially no inversion during the lifetimes of the ions. The configuration of the nitrenium ion that undergoes nucleophilic attack should be the configuration in which it is first formed. We designate the configuration syn if the hydrogen of the NH group is oriented toward the β-ring carbon of higher priority (in the Kahn-Ingold-Prelog sense) and anti otherwise. As discussed in detail elsewhere (33), the relative energies of the two configurations are determined by a combination of steric and electrostatic interactions between the NH groups and the aryl rings. The more stable configurations predicted by ab initio calculations at the HF/6-31G*//HF/3-21G, and in some cases higher, levels are shown in Table 1. Because of the special symmetries of the phenyl and the 2-pyrenyl ions, there is no distinction between the syn and anti configurations. All the other ions have the anti configuration except the 4-pyrenyl. In this work, we adopt the working hypothesis that the lower-energy configuration is also the one that is formed preferentially under biological conditions. This is discussed further below. Nucleophilic Attack by the Guanine Amino Group. It is our intention to eventually address the whole spectrum of nitrenium ion adducts using theoretical methods. Here we focus exclusively on the attack of guanine amino groups at aryl nitrenium ion ring carbons. Limiting the scope of the study to processes involving the formation of a single type of bond ensures the greatest likelihood that systematic errors in the AM1 procedure will cancel each other out. Specifically, we are concerned with the first step in the overall process depicted in Figure 1b. In the initially formed intermediate, the aromatic nucleus is interrupted by the creation of a tetrahedral carbon at the point of attack. If the subsequent rearomatization is rapid, the overall regiochemistry should be determined by the relative activation barriers, leading to alternative tetrahedral intermediates. In the case depicted in Figure 1, nucleophilic attack is shown occurring adjacent to the nitrenium ion nitrogen. However, we consider attack at all active1 nonfused ring 1 These are the positions at which the formal positive charge is located in one of the conventional resonance forms of the nitrenium ion. They correspond to the “starred” positions of simple PMO theory of the analogous benzylic structures. Exploratory calculations involving some of the “inactive” positions confirmed the expectation that adducts at these positions were much higher in energy and, in some cases, were not bound at all.

Regiochemistry of Nucleophilic Attack

Chem. Res. Toxicol., Vol. 12, No. 1, 1999 55

Table 1. AM1 Enthalpies and Hu 1 ckel Localization Energies for the Formation of (Guanin-N2-yl)-Aryl Nitrenium Ion Intermediates

a Structure and numbering for the nitrenium ion. The structure is shown in its lower-energy configuration (shown in italics) as determined by ab initio molecular orbital calculations at the HF/6-31G*//HF/3-21G level or higher. Ab initio data for all but benzo[a]pyrene derivatives are given in ref 33. For the benzo[a]pyrene derivatives, ∆E(antifsyn) values are as follows (positions and values in kilocalories per mole): 1 and 2.99, 2 and 1.70, 3 and 4.61, and 6 and 2.54 (G. P. Ford, unpublished HF/6-31G*//HF/3-21G calculations). b Computed enthalpy change (kilocalories per mole) for the first step in the reaction sequence shown in Figure 1b. ∆H ) ∆Hf°(tetrahedral intermediate) - ∆Hf°(guanine) - ∆Hf°(ArNH+). Values of ∆Hf°(ArNH+) are from ref 44 except when Ar is benzo[a]pyrenyl which are as follows (position, syn value, and anti value): 1, 280.4, and 279.4, 2, 299.3, and 298.2, 3, 282.1, and 280.2, and 6, 277.5, and 275.9. ∆Hf°(guanine) ) 48.8 kcal mol-1. The predicted site of attack on the lower-energy syn or anti nitrenium ion configuration is bold. c Lr ) M+ - Mr, where M+ is the Hu¨ckel π-energy of the nitrenium ion and Mr that of the neutral molecule with the same number of electrons resulting from the exclusion of the carbon at position r from the π-system. d Symmetrical position. Atom numbering with respect to the orientation of the NH bond is as shown on the left.

positions. The AM1 enthalpies for addition to the phenyl nitrenium ion and nine polycyclic nitrenium ions are summarized in Table 1. Results for additions to the syn and anti configurations are recorded separately. In each case, ∆H refers to the formation of the most stable of the nine possible conformers of the tetrahedral intermediate related by rotation about the two 3-fold torsions indicated by the symbols ψ and φ in the 2-naphthyl example shown in Figure 2. In the interest of brevity, we do not give details of all of these geometries. However, this figure illustrates many features common to them all. The guanine and aryl rings occupy approximately parallel planes with the guanine C2-N2 bond approximately

gauche to the two C-C bonds that terminate at the carbon atom of the nitrenium ion undergoing substitution. The orientation of the guanine moiety with respect to the torsion φ was invariably that in which the guanine C6-O dipole was directed toward the region of greatest positive charge concentration. A striking feature of the computed enthalpies is their sensitivity to the syn or anti configuration of the nitrenium ion. This is especially true where the attack occurs adjacent to the NH group. For example, addition to the 1-position of the anti configuration of the 2-naphthyl ion shown in Figure 2 is computed to be 3.5 kcal mol-1 more favorable than the analogous addition to the syn config-

56 Chem. Res. Toxicol., Vol. 12, No. 1, 1999

Ford and Thompson

Figure 2. AM1 geometry of the tetrahedral intermediate formed by nucleophilic attack at the 1-position of the 2-naphthyl nitrenium ion (anti configuration). The structure shown is the most stable of the nine conformers related through rotations about the torsional angles ψ and φ.

uration. For the 12 similar examples in Table 1, these differences varied from 0.5 to 8.9 kcal mol-1. In every case, the addition was predicted to be more favorable in the configuration where the lone pair electrons of the nitrenium ion NH group were directed toward the attacking guanine. This is presumably due to the favorable charge dipole interaction with the now quaternary guanine amino group. For attack at positions remote from the NH group, the differences were much smaller. The median value was only 0.7 kcal mol-1, and in no case did it exceed 3.6 kcal mol-1. We have focused on the enthalpies associated with the formation of the tetrahedral intermediates rather than on the more directly relevant activation quantities (∆Hq), because these require substantially less computational effort. This is permissible only if the two quantities parallel one another. To confirm this, we located the transition states themselves for the representative cases of the 1- and 2-naphthyl nitrenium ions. The expected parallel between ∆Hq and ∆H is demonstrated in Figure 3. Within the limits of reliability of the AM1 procedure, the ∆H data reflect the energetics of interaction between the isolated nitrenium ion and a single guanine monomer in the gas phase. Although these energies fundamentally underlie those in the aqueous phase reactions of nucleic acid polymers, the latter will of course be subtly modified by a myriad of factors yet to be quantified. Nevertheless, it is interesting to compare the regiochemistry predicted by these simple models with what is observed experimentally. As mentioned above, in the absence of more specific information, we assume the nitrenium ions are formed in their most stable syn or anti configuration. This seems most reasonable where steric effects are unimportant, namely, when the NH group lies in neither a peri nor a bay region. Here, the relative energies of the two configurations are largely determined by electrostatic effects. The more stable orientation of the NH group is the one that places the nitrogen lone pair closest (and the hydrogen atom therefore furthest) from the more positive β-ring carbon (33). The same effect should apply even more forcefully in the transition states leading to them where the required N-O heterolysis would be expected to take place along a trajectory passing closer to the more positively charged β-ring atom. The situation is less clear where the greater stability of the anti form

Figure 3. Approximately linear relationship between the calculated (AM1) activation enthalpies (∆Hq) and heats of reaction (∆H) for attack of the guanine 2-amino group at the active positions of the 1- and 2-naphthyl nitrenium ions. Data are as follows (position and ∆Hq): (O) 1-naphthyl (syn), 2 and 5.3; 4 and 5.6; 5 and 16.8; 7 and 16.8; (b) 1-naphthyl (anti), 2 and 8.1; 4 and 5.8; 5 and 16.2; 7 and 15.0; (0) 2-naphthyl (syn), 1 and 4.4; 3 and 13.9; 6 and 13.1; 8 and 15.1; and (9) 2-naphthyl (anti), 1 and 3.0; 3 and 17.0; 6 and 13.6; 8 and 14.7.

is due to the smaller steric requirements of the lone pair versus the hydrogen. This is the case, for example, for the 1-naphthyl ion (33). Orbital symmetry considerations suggest the leaving group must exit approximately perpendicular to the plane of the forming nitrenium ion (35, 36). Therefore, the steric effects of the leaving group may be of less consequence than might at first sight be suspected. These points are currently under scrutiny in our laboratory.

The preferred ring positions for addition of the guanine 2-amino group to the more stable configuration of each ion are highlighted in bold in Table 1. For the 2-naphthyl ion the predictions are clear and unequivocal. Regardless of whether the ion reacts in its syn or anti configuration, the 1-position is predicted to be most reactive. This is in accord with the several experimental observations discussed in the Introduction. The situation for the 1-naph-

Regiochemistry of Nucleophilic Attack

Chem. Res. Toxicol., Vol. 12, No. 1, 1999 57

Figure 4. Nonbonding molecular orbital coefficients for representative odd alternant hydrocarbons. From left to right, aryl moieties are (top row) phenyl, 1-naphthyl, 2-naphthyl, 2-phenanthryl, and 6-chrysenyl, (middle row) 1-pyrenyl, 2-pyrenyl, 4-pyrenyl, and 1-benzo[a]pyrenyl, and (bottom row) 2-benzo[a]pyrenyl, 3-benzo[a]pyrenyl, and 6-benzo[a]pyrenyl.

thyl ion is complicated. In the more stable anti configuration, C4 is predicted to be most reactive, while C2 is predicted to be the most reactive in the syn configuration. The latter is the site at which guanine adducts have been observed. However, they have involved not a guanine N2 linkage but a less common O6-linked adduct 4 (37). That the reactivities of the 2- and 4-positions are very similar is evident from solvolysis data for N-hydroxy-1-aminonaphthalene. Under neutral or acidic conditions, a mixture of the 2- and 4-aminonaphthols in which the former dominates is obtained (37). The most reactive ring positions of the 2- and 4-pyrenyl ions are predicted to be 1, and 5, respectively. So far, no ring-substituted derivatives arising from presumed precursors of these ions have yet been reported. For the 1-pyrenyl ion, the enthalpies of addition to the 6-, 8-, and 9-positions all lie within 0.4 kcal mol-1 of each other which suggests that the reactivities of these positions should be very similar. Indeed, Fu and co-workers (24) have recently found that exposure of DNA to precursors of the 1-pyrenyl nitrenium ion leads to the 6- and 8-substituted derivatives 5 and 6. Among the benzo[a]pyrene derivatives, DNA adducts have been identified for the 1- and 3-nitro derivatives. In both cases, the principal ring-substituted adducts (7 and 8) correspond to attack at the 6-positions (25, 26) of the presumed nitrenium in intermediates. Both are unequivocally predicted by the AM1 results regardless of which nitrenium ion configuration is involved. The 2and 6-nitrobenzo[a]pyrene isomers have been studied (38-40), although adducts arising from the nitrenium ion intermediates have yet to be identified. For the 2-nitro isomer, AM1 predicts the 1-position to be the most reactive while the 6-nitro isomer is expected to yield a

complex mixture of products derived from attack at the 1-, 3-, 4-, and 12-positions. Simple Molecular Orbital Models. The nucleophilic attack at the nitrenium ion ring positions bears an obvious kinship to the analogous process in electrophilic aromatic substitution described by the localization energies, Lr, of simple Hu¨ckel π-electron theory (41). According to these ideas, a smaller value of Lr implies a smaller energetic penalty for removal of that carbon from the aromatic system and, therefore, that the position should be more reactive. The AM1 calculations obviously include subtle steric, electrostatic, and rehybridization effects that are absent from the simple Hu¨ckel indices. Nevertheless, the latter closely parallel the AM1 regiochemical predictions (Table 1). In contrast to the AM1 data,2 they can be obtained with trivial computational effort. Ford and Scribner’s suggestion (22) that the regiochemistry of ring attack can be predicted simply from the magnitudes of the nonbonding molecular orbital coefficients of the analogous benzylic ions does not 2 The central processing unit (CPU) time required for the full AM1 minimization of all conformational/configurational possibilities at each active ring position of a given nitrenium ion varied from approximately 2 h for PhNH+ to approximately 36 h for each of the benzo[a]pyrenyl isomers (Silicon Graphics Inc. O2 Workstation, 180 MHz, R5000 CPU). 3 Benzylic and nitrenium ions differ at the Hu ¨ ckel level only by a change in the Hu¨ckel parameters hN and βCN. Since this involves no first-order change in the wave function, the theorems of PMO theory for alternant hydrocarbons (AHs) should also apply to the nitrenium ions. This general approach has been successfully used to rationalize the relative stabilities of aryl nitrenium ions as a function of the structures of the aryl groups (46) as well as the relative stabilities of the syn and anti configurations of individual nitrenium ions (33). 4 This method also predicts a single preferred site of ring attack for the following nitrenium ions (predicted site in parentheses): 3-pyrenyl (4), 9-pyrenyl (10), 2-chrysenyl (1), 3-chrysenyl (4), 4-BaP (5), 5-BaP (4), 8-BaP (7), 9-BaP (10), 11-BaP (12), and 12-BaP (11).

58 Chem. Res. Toxicol., Vol. 12, No. 1, 1999

require the use of a computer at all. The necessary coefficients, a0r, are easily deduced on inspection using the well-known starring procedure (23). Values for some of the systems discussed in this paper are shown in Figure 4. Their suggestion was based on the obvious analogy between the nitrenium and benzylic ions3 and the recognition that the nonbonding orbital in the latter is also the lowest unoccupied molecular orbital (LUMO). Nucleophilic attack should therefore occur preferentially at the positon where a0r has is largest value since this is also the position where both the LUMO density and the positive charge density are greatest. The same conclusion can be reached in a different way using PMO theory and approximating the localization energies in terms of the nonbonding orbital coefficients. The PMO treatment of localization energies for even alternant hydrocarbons is very familiar and has been widely applied (23). Although the approximations involved are somewhat more severe, it is fairly easy to show (see the Supporting Information) that for an odd alternant hydrocarbon the localization energy at a nonfused ring carbon, r, is ∼2(1 - |a0r|). Again, this implies the most reactive position, i.e., the position at which the localization energy is least, occurs where a0r has its largest value. A survey of the structures in Figure 4 reveals two situations. In one, the largest value of a0r is associated with a unique atom. In every case, this is also the most reactive position predicted by the AM1 calculations as well as by the simple Hu¨ckel localization energies. Where deoxyguanosin-N2-yl adducts have been observed, they have exclusively involved these positions.4 The 2-naphthyl, 2-benzo[a]pyrenyl, and 3-benzo[a]pyrenyl cases have already been discussed. To these can be added the 1-position of the 2-phenanthryl nitrenium ion (42) and the 5-position of the 6-chrysenyl (43) nitrenium ions which are correctly identified as the principal sites of attack by guanine amino groups. In the second situation, the largest value of a0r is common to two or more positions. A striking example is provided by the 1-pyrenyl system where the coefficients are identical at all five active ring positions. As indicated by the AM1 calculations and by the observation of multiple adduct structures, the reactivities of these positions are very similar. Small differences that do exist are likely to depend on factors beyond those inherent in the isolated gaseous reactants. Supporting Information Available: Derivation of the expression for the PMO localization energies of odd alternant hydrocarbons (2 pages). Ordering information is given on any current masthead page.

References (1) Lewtas, J., and Nishioka, M. C. (1990) Nitroarenes: their detection, mutagenicity and occurrence in the environment. In Nitroarenes, Occurrence, Metabolism, and Biological Impact (Howard, P. C., Hecht, S. S., and Beland, F. A., Eds.) pp 61-72, Plenum, New York. (2) Hecht, S. S., and El-Bayoumy, K. (1990) The possible role of nitroarenes in human cancer. In Nitroarenes, Occurrence, Metabolism, and Biological Impact (Howard, P. C., Hecht, S. S., and Beland, F. A., Eds.) pp 309-316, Plenum, New York. (3) Crabtree, H. C., Hart, D., Thomas, M. C., Witham, B. H., McKenzie, I. G., and Smith, C. P. (1991) Carcinogenic ranking of aromatic amines and nitrocompounds. Mutat. Res. 264, 155-162. (4) Later, D. W., Pelroy, R. A., Stewart, D. L., McFall, T., Both, G. M., Lee, M. L., Tedjamulia, M., and Castle, R. N. (1984) Microbial

Ford and Thompson

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16) (17) (18)

(19)

(20)

(21)

(22)

(23) (24)

(25)

(26)

mutagenicity of isomeric two-, three-, and four-ring amino polycyclic aromatic hydrocarbons. Environ. Mutagen. 6, 497-515. Miller, J. A., and Miller, E. C. (1983) Some historical aspects of N-aryl carcinogens and their metabolic activation. Environ. Health Perspect. 49, 3-12. Kadlubar, F. F., and Beland, F. A. (1985) Chemical properties of ultimate carcinogenic metabolites of arylamines and arylamides. In Polycyclic Hydrocarbons and Carcinogenesis (Harvey, R. G., Ed.) ACS Symposium Series 283, pp 341-370, American Chemical Society, Washington, DC. Fu, P. P., Herreno-Saenz, D., Von Tungeln, L. S., Lay, J. O., Wu, Y.-S., Lai, J.-S., and Evans, F. E. (1994) DNA adducts and carcinogenicity of nitro-polycyclic aromatic hydrocarbons. Environ. Health Perspect. 102 (Suppl. 6), 177-183. Beland, F. A., and Kadlubar, F. F. (1990) Metabolic activation and DNA adducts of aromatic amines and nitroaromatic hydrocarbons. In Handbook of Experimental Pharmacology (Cooper, C. S., and Grover, P. L., Eds.) Vol. 94/I, pp 267-325, Springer-Verlag, Berlin. Kadlubar, F. F. (1994) DNA adducts of carcinogenic aromatic amines. In DNA Adducts. Identification and Biological Significance (Hemminki, K., Dipple, A., Shuker, D. E. G., Kadlubar, F. F., Segerba¨ck, D., and Bartsch, H., Eds.) IARC Scientific Publication 125, pp 199-216, IARC, Lyon, France. Delclos, K. B., and Kadlubar, F. F. (1997) Carcinogenic aromatic amines and amides, In Comprehensive Toxicology, Volume 12: Chemical Carcinogens and Anticarcinogens (Bowen, G. T., and Fischer, S. M., Eds.) pp 141-170, Elsevier, Oxford, U.K. Kadlubar, F. F., Anson, J. F., Dooley, K. L., and Beland, F. A. (1981) Formation of urothelial and hepatic DNA adducts from the carcinogen 2-naphthylamine. Carcinogenesis 2, 467-470. Beland, F. A., Beraneck, D. T., Dooley, K. L., Heflich, R. H., and Kadlubar, F. F. (1983) Arylamine-DNA adducts in vitro and in vivo: their role in bacterial mutagenesis and urinary bladder carcinogenesis. Environ. Health Perspect. 49, 125-134. Kadlubar, F. F., Unruh, L. E., Beland, F. A., Straub, K. M., and Evans, F. E. (1980) In vitro reaction of the carcinogen, N-hydroxy2-naphthylamine, with DNA at the C-8 and N2 atoms of guanine and at the N6 atom of adenine. Carcinogenesis 1, 139-150. Famulok, M., Bosold, F., and Boche, G. (1989) Synthesis of N-acetoxy-2-aminonaphthaline, an ultimate carcinogen of the carcinogenic 2-naphthylamine, and its in vitro reactions with (bio)nucleophiles. Tetrahedron Lett. 30, 321-324. Hemminki, K., Dipple, A., Shuker, D. E. G., Kadlubar, F. F., Segerba¨ck, D., and Bartsch, H. (1994) in DNA Adducts. Identification and Biological Significance (Hemminki, K., Dipple, A., Shuker, D. E. G., Kadlubar, F. F., Segerba¨ck, D., and Bartsch, H., Eds.) IARC Scientific Publication 125, IARC, Lyon, France. Dipple, A. (1995) DNA adducts of chemical carcinogens. Carcinogenesis 16, 437-441. Dipple, A., and Moschel, R. C. (1990) Chemistry of DNA alkylation and aralkylation. Prog. Clin. Biol. Res. 340A, 71-80. Ford, G. P. (1997) Semiempirical molecular orbital theory in carcinogenesis research. J. Mol. Struct. (THEOCHEM) 401, 253266. Ford, G. P., and Scribner, J. D. (1990) Prediction of nucleosidecarcinogen reactivity. Alkylation of adenine, cytosine, guanine, and thymine and their deoxynucleosides by alkane diazonium ions. Chem. Res. Toxicol. 3, 219-230. Singer, B. (1975) The chemical effects of nucleic acid alkylation and their relation to mutagenesis. Prog. Nucleic Acid Res. 15, 219-284. Kennedy, S. A., Novak, M., and Kolb, B. A. (1997) Reactions of ester derivatives of carcinogenic N-(4-biphenylyl)hydroxylamine and the corresponding hydroxamic acid with purine nucleosides. J. Am. Chem. Soc. 119, 7654-7664. Ford, G. P., and Scribner, J. D. (1981) MNDO Molecular orbital study of nitrenium ions derived from carcinogenic aromatic amines and amides. J. Am. Chem. Soc. 103, 4281-4291. Dewar, M. J. S., and Dougherty, R. C. (1975) The PMO Theory of Organic Chemistry, pp 78-140, Plenum, New York. Herreno-Saenz, D., Evans, F. E., Beland, F. A., and Fu, P. P. (1995) Identification of two N2-deoxyguanosinyl DNA adducts upon nitroreduction of the environmental mutagen 1-nitropyrene. Chem. Res. Toxicol. 8, 269-277. Herreno-Saenz, D., Evans, F. E., and Fu, P. P. (1994) Nitroreduction of 1-and 3-nitrobenzo[a]pyrene resulting in formation of N2-deoxyguanosinyl adducts through long-range migration.Chem. Res. Toxicol. 7, 806-814. Herreno-Saenz, D., Evans, F. E., and Fu, P. P. (1993) Formation of the adduct 6-(deoxyguanosin-N2-yl)-3-aminobenzo[a]pyrene

Regiochemistry of Nucleophilic Attack

(27) (28) (29) (30)

(31) (32)

(33) (34)

(35) (36)

from the mutagenic environmental contaminant 3-nitrobenzo[a]pyrene. Carcinogenesis 14, 1065-1067. Dewar, M. J. S., Zoebisch, E. G., Healy, E. F., and Stewart, J. J. P. (1985) AM1: A new general purpose quantum mechanical molecular model. J. Am. Chem. Soc. 107, 3902-3909. Spartan, version 5.0 (1997) Wavefunction, Inc., Suite 370, 18401 Von Karman Ave., Irvine, CA 92612. Baker, J. (1986) An algorithm for the location of transition states. J. Comput. Chem. 7, 385-395. Wissner, A. (1990) HMO version 1.1: A Huckel molecular orbital program for the Macintosh. Tetrahedron Comput. Methodol. 3, 63-71 (version 2.0.3 at http://mwn.net/infomac2/science/hmoplus-203.html). Streitwieser, A. J., Jr. (1961) Molecular Orbital Theory for Organic Chemists, John Wiley & Sons, New York. Srivastava, S., Toscano, J. P., Moran, R. J., and Falvey, D. E. (1997) Experimental confirmation of the iminocyclohexadienyl cation-like structure of arylnitrenium ions: Time-resolved IR studies of diphenylnitrenium ion. J. Am. Chem. Soc. 119, 1155211553. Ford, G. P., Herman, P. S., and Thompson, J. W. (1999) Syn and anti aryl nitrenium ions. J. Comput. Chem. (in press). Cramer, C. J., Dulles, F. J., and Falvey, D. E. (1994) Ab initio characterization of phenylnitrenium and phenylcarbene: Remarkably different properties for isoelectronic species. J. Am. Chem. Soc. 116, 9787-9788. Haberfield, P., and DeRosa, M. (1985) σ- and π-nitrenium ions. Chem. Commun., 1716-1717. Ford, G. P., and Herman, P. S. (1991) Conformational preferences and energetics of N-O heterolyses in aryl nitrenium ion precursors: ab initio and semiempirical molecular orbital calculations. J. Chem. Soc., Perkin Trans. 2, 607-616.

Chem. Res. Toxicol., Vol. 12, No. 1, 1999 59 (37) Kadlubar, F. F., Miller, J. A., and Miller, E. C. (1978) Guanyl O6-arylamination and O6-arylation of DNA by the carcinogen N-hydroxy-1-naphthylamine. Cancer Res. 38, 3628-3638. (38) Yu, S., Herreno-Saenz, D., Miller, D. W., Kadlubar, F. F., and Fu, P. P. (1992) Mutagenicity of nitro-polycyclic aromatic hydrocarbons with the nitro substituent situated at the longest molecular axis. Mutat. Res. 283, 45-52. (39) Garner, R. C., Stanton, C. A., Martin, C. N., Harris, C. C., and Grafstrom, R. C. (1985) Rat and human explant metabolism, binding studies, and DNA adduct analysis of benzo[a]pyrene and its 6-nitroderivative. Cancer Res. 45, 6225-6231. (40) Kaneko, M., and Nagata, C. (1983) Covalent binding of 6-nitrobenzo[a]pyrene to DNA in vitro. Gann 74, 5-7. (41) Brown, R. D. (1951) Molecular orbitals and organic reactions. Q. Rev., Chem. Soc., 63-99. (42) Gupta, R. C., Early, L., Fullerton, N. F., and Beland, F. (1989) Formation and removal of DNA adducts in target and nontarget tissues of rats administered multiple doses of 2-acetylaminophenanthrene. Carcinogenesis 10, 2025-2033. (43) Declos, K. B., Miller, D. W., Lay, J. O., Jr., Casciano, D. A., Walker, R. P., Fu, P. P., and Kadlubar, F. F. (1987) Identification of C8-modified deoxyinosine and N2- and C8-modified deoxyguanosine as major products of the in vitro reaction of N-hydroxy6-aminochrysene with DNA and the formation of these adducts in isolated rat hepatocytes treated with 6-nitrochrysene and 6-aminochrysene. Carcinogenesis 8, 1703-1709. (44) Ford, G. P., and Herman, P. S. (1991) Comparison of the relative stabilities of polycyclic aryl nitrenium ions and arylmethyl cations: ab initio and semiempirical molecular orbital calculations. J. Mol. Struct. (THEOCHEM) 236, 269-282.

TX9801460