Stereoselectivity in the Microsomal Conversion of N

Rutgcrs Uniuersity, Piscataway, New Jersey 08855, Program Resources, Inc., NCI- ... Medicinal Chemistry, National Cancer Institute, Rethesda, Maryland...
1 downloads 0 Views 728KB Size
540

Chem. Res. Toxicol. 1990, 3, 540-544

Stereoselectivity in the Microsomal Conversion of N-Nitrosodimethylamine to Formaldehyde Larry K. Keefer,*lt Marilyn B. Kroeger-Koepke,l Hiroyuki Ishizaki,§ Christopher J . Michejda,: Joseph E. Saavedra,t Joseph A. Hrabie," C. S. Yang,! and Peter P. RollerL Chemistry Section, Laboratory of Comparative Carcinogenesis, National Cancer Institute, Frederick Cancer Research and Development Center, Frederick, Maryland 21 702, ABL-Basic Research Program, NCI-Frederick Cancer Research and Development Center, Frederick, Marjaland 21 702, Department of Chemical Biologj' and Pharmacognosy, College of Pharmacy, Rutgcrs Uniuersity, Piscataway, N e w Jersey 08855, Program Resources, Inc., NCI-Frederick Cancer Research and Development Center, Frederick, Maryland 21 702, and Laboratory of Medicinal Chemistry, National Cancer Institute, Rethesda, Maryland 20892 Received J u n e 18, 1990

T h e possibility t h a t N-nitrosodimethylamine (NDMA) might be metabolized preferentially a t either the syn (relative t o the nitroso oxygen) or the anti methyl group has been examined by comparing the rates of formaldehyde production when unlabeled NDMA, its fully deuteriated (NDMA-d3) were inanalogue (NDMA-d6), and (2)-or (E)-N-nitrosomethyl(methy1-d3)amine cubated in turn a t concentrations of 0-2.4 mM with acetone-induced rat liver microsomes. The K , values for the conversion of (2)-and (E)-NDMA-d3t o formaldehyde were identical t o each other within experimental error (32 f 2 and 35 f 1 pM, respectively) but different from those for NDMA (24 f 6 pM) and NDMA-d, (116 f 3 pM); similar V,, values were observed for the four isotopic variants [7.5-8.1 nmol/ (mg of protein.min)]. T h e observed similarity of kinetic parameters for (2)-and (E)-NDMA-d, suggested t h a t the isotopic composition of the methyl group is an energetically more important determinant of its rate of oxidation a t the NDMA demethylase active site than is its orientation relative t o the nitroso oxygen atom. The absence of syn vs anti stereospecificity was confirmed via product isolation studies, in which the formaldehyde generated from each of the four isotopomers was trapped as the dimedone adduct and assayed for deuterium content by mass spectrometry; again, a strong preference for metabolism a t CH, vs CD, regardless of stereochemistry was observed, though the data on CH20 generation suggested t h a t there may be a slight net excess of anti attack. T h e results indicate t h a t the microsomal enzymes employed display little regioselectivity in metabolizing the syn vs anti methyl groups of NDMA.

Introduction Elucidating the stereochemical course of a metabolic transformation can provide important insight into the enzymology of the process. In addition to the theoretical interest inherent in such knowledge, critical information concerning the multiplicity of the enzyme system responsible for the reaction can be obtained ( I ) , and the data derived can be useful for inferring mechanistic details ( 2 ) as well as mapping the topography of substrate interaction with the active site(s) involved ( 3 ) . There is considerable evidence that NDMA' and certain analogous nitrosamines are activated to ultimate carcinogens by enzymatic oxidation at an cu-carbon atom ( 4 ) ;in liver microsomes, this reaction is mainly catalyzed by cytochrome P45OTIE1, also known as P450LM3,, P450a,, and P450j (Fj-7). Because of restricted rotation about the nitrogen-nitrogen bond, the two cu-carbons of NDMA are stereochemically nonequivalent, with one being syn to the 'Chemistry Section. National Cancer Institute. ABL-Babic Research Program. 5 Rutgers IJniversity. Program Resources, Inc. Laboratory of Medicinal Chemistry, National Cancer Institute.

I I

I

Y

*C

N-Nitrosodimethylamine (NDMA)

oxygen atom while the other is anti (8). In the present work, we examine the hypothesis that metabolic oxidation of this symmetrical nitrosamine might

f

'

Abbreviations: NDMA, N-nitrosodimethylamine; NDMA-d,, N-

nitrosomethyl(methyl-d3)amine;NDMA-d6, N-nitrosobis(methyl-d&

amine.

oa93-2zaX/go/ ~703-0540~02.50/00 1990 American Chemical Society

Stereochemistry of Nitrosamine Metabolism

occur preferentially at one or the other of the stereochemically distinguishable methyl groups.

Experimental Procedures Warning! The N-nitroso compounds described here are potent carcinogens that should be handled, stored, and discarded with due respect for their toxic properties. Chemicals. (E)-NDMA-d, was prepared as follows. A slurry of 2.26 g (0.023 mol) of potassium (2)methanediazotate [prepared by the method of Muller et al. (9)] in 60 mL of anhydrous ether was ctmled to 0 "C under nitrogen. To the cold mixture was added dropwise 4.35 g (0.03 mol) of iodomethane-d,. The mixture was stirred at 0 "C for 15 h and vacuum filtered into a cold (0 "C) receiver. The clear filtrate was evaporated under nitrogen. The residue was taken up in cold water and filtered through glass wool. The filtrate was collected in a test tube which was immersed in ice. The nitrosamine product was extracted into dichloromethane. The combined organic layer was dried over potassium carbonate, filtered, and evaporated under a stream of nitrogen. All operations were carried out a t 0 "C. The yield was 225 mg (13%) of NDMA-d3 having a Z:E ratio of 6:94, as reflected in the relative dominance of the 'H NMR singlet a t 6 3.02. NDMA-d6 was prepared by base-catalyzed exchange of NDMA with deuterium oxide, as previously described (IO). (Zl-NDMA-d, having a Z E ratio of 19:l was synthesized by alkylation of thallium (E)-methanediazotate with iodomethane-d3 according to a published procedure (11). Deuterium oxide was purchased from Wilmad Glass Co. (99.8% D minimum). 5,5-Dimethylcyclohexane-1,3-dione ("dimedone"), acetic acid-d,, and 40% sodium deuterioxide in D 2 0 (99% D) were obtained from Aldrich Chemical Co. Para(forma1dehyde-d,) was the product of Merck Isotopes, and hexane was Fisher HPLC grade. Nitrosamines were purified as needed before use by low-temperature high-performance liquid chromatography (HPLC), as previously described ( I I). Conformational purity of the NDMA-d3 preparations was measured by using a Varian XL-200 NMR spectrometer to examine the chloroform-d extracts of the solution under study. Nitrosamine concentrations were measured by ultraviolet spectrophotometry assuming an extinction coefficient of 95.5 M-I cm-' a t the 334-nm maximum in aqueous solution. Mass spectral data were collected on a VG 7070E-HF instrument operating in the positive ion electron impact mode a t 70 eV with sample introduction via the solid probe a t room temperature. I n c o r p o r a t i o n of D e u t e r i u m i n t o 5,5-Dimethylcyclohexane-1,3-dione (Dimedone). A mixture containing 2.14 g (15.3 mmol) of dimedone reagent, 2.4 mL of 40% sodium deuterioxide in D20, and 25 mL of deuterium oxide was placed in a 50-mL flask equipped with a condenser and drying trap filled with potassium hydroxide pellets, refluxed for 9 h. allowed to sit a t 25 "C for 60 h, and neutralized to pH 5-6 with about 3 mL of acetic acid-d,. A heavy precipitate formed that was extracted into five 20-mL portions of dry dichloromethane. The combined organic extracts were dried with anhydrous sodium sulfate and evaporated to provide 1.66 g of deuteriated dimedone: mass spectrum (molecular ion region) m / z (relative intensity) 142 (0.51, 143 (9.1), 144 (64.4), 145 (100, d 5 ) , 146 (4$5.6),147 (4.5). P r e p a r a t i o n of M a s s S p e c t r a l S t a n d a r d s . The decadeuteriated internal standard used as a reference for quantifying metabolically produced formaldehyde was synthesized as follows. Para(formaldehyde-dl) was depolymerized by mixing 104 mg with 4 mL of deuterium oxide and heating to 60 "C for 1 h. This solution was added to a solution of the deuteriated dimedone prepared in the previous paragraph (140 mg in 5 mL of deuterium oxide). After standing overnight, the precipitate that formed (55 mg) was collected and recrystallized from 3 mL of methanol-0-d to yield 38 mg of crystalline product: mass spectrum (molecular ion region) n i / z (relative intensity) 299 (0.4), 300 (4.4), 301 (30.8), 302 (100, d l o ) ,303 (20.9), 304 (3.21, 305 (0.4). The authentic dimedone adduct of CD20 was similarly prepared by using para(forma1dehyde-d,), undeuteriatsd dimedone, and nondeuteriated solvents: mass spectrum (molecular ion region) m / z (relative intensitv) 292 (O.B), 293 (1.2), 294 (100, d2),295 (19.9), 296 (2.7),297 (0.3). The authentic CH20 adduct was also prepared: mass spectrum (molecular ion region) m /t (relative intensity) 291 (0.5), 292 (loo), 293 (19.7), 294 (2.6).

Chem. Res. Toxicol., Vol. 3, No. 6, 1990 541 Measurement of Metabolic Rate. Microsome preparation was begun by sacrificing young male Sprague-Dawley rats (body weights 80-90 g) which had been treated with an oral dose of acetone (4 mL/kg of body weight) 20 h earlier (22). The livers were removed to ice-cold saline. To each 10-g portion of wet liver was added 40 mL, of 0.05 M Tris-HC1 containing 1.15% potassium chloride (pH 7.4). Liver tissues were minced and homogenized, and the homogenates were centrifuged a t 9OOOg for 20 min a t 4 "C. The supernatant was removed and further centrifuged a t 100OOg for 90 min. The pellet deposited in this second centrifugation was washed once with 10 mM EDTA, then resuspended in 0.25 M sucrose, and stored frozen a t -70 "C for future metabolism studies. Protein concentration was determined by using the Lowry method (13). Incubation mixtures consisting of 0.064 mg of microsomal protein and an NADPH generating system (14) were equilibrated for 2 min a t 37 "C. The reaction was begun by adding 0.025 mL of the nitrosamine solution in water. Total volume of each incubation mixture was 0.25 mL. Concentrations of NDMA and the NDMA-d3 analogues ranged from 0.02 to 0.6 mM, while NDMA-d6 was added a t levels of 0.09-2.4 mM. When either (2)or (E)-NDMA-d3was used, addition to the incubation mixture was performed as quickly as possible after thawing the frozen stock solution; the maximum time between thawing and the beginning of incubation was 50 min, during which time the solutions were kept a t 0-5 "C on ice. After 10 min of incubation, the reaction was stopped by adding 10% trichloroacetic acid and cooling the flask in an ice bath. After centrifuging the resulting mixture for 10 min, an aliquot of the supernatant was mixed with Nash reagent (15) and incubated for 20 min a t 60 "C. Formaldehyde concentration was measured by comparing the absorbance a t 412 nm (minus blank values obtained for nitrosamine-free control incubations) with those for standards of known concentration. The increase in concentration was shown to be linear with time over the entire incubation period employed. Formaldehyde concentration data were converted to reciprocals of reaction velocities and plotted against the reciprocals of the mean substrate concentrations to obtain V,,, and K , values using a least-squares curve-fitting procedure in the Lineweaver-Burk analysis. Measurement of S u b s t r a t e Equilibration Rate. To determine the rate of interconversion of NDMA-d3 conformers under the conditions of the microsomal reaction, 2 mM (Z)-NDMA-d3 was incubated as above with uninduced microsomes and cofactors in a total volume of 3 mL a t 37 "C for 3 h. Without precipitation of the protein [trichloroacetic acid led to rapid equilibration of the conformers (IS)], periodic aliquots of the reaction mixture were extracted with cold chloroform-d. Layer separation was completed by centrifuging the two-phase mixtures for a total of 10 min a t 4 "C. The aqueous layers were partially removed by aspiration, and the residual water was precipitated by freezing in dry ice. The chloroform-d solutions were analyzed by NMR spectrometry as described above. The observed half-life for 2 -E interconversion was 132 min under these conditions. A final sample taken after 98 h of equilibration revealed a Z E ratio of 50.2:49.8. P r o d u c t Analysis Studies. Trapping of the formaldehyde produced in the microsomal metabolism of each NDMA isotopomer for mass spectral analysis of its deuterium content was performed as follows. A dimedone solution (416 mg in 100 mL of water) was extracted once with hexane to remove any preformed dimedone-formaldehyde adduct. Exposure of this and all other solutions to rubber and plastics was avoided wherever possible to minimize contamination by phthalate plasticizers, which preliminary experiments showed would interfere with the mass spectrometry of the dimedone adduct. Incubations of NDMA, (Z)-NDMA-d,, (E)-NDMA-d3,NDMA-d6, and a nitrosamine-free control were then performed as described above for the kinetic experiments except for the following: the incubation volumes were 0.5 mL; the substrate concentrations were consistently 0.6 mM; 0.124 mg of microsomal protein was used in each incubation; only 30 min was allowed to elapse between thawing the NDMA-d3 stock solutions and adding them to the preincubated reaction mixtures. Z E ratios for the NDMA-d3 stock samples were measured before and after incubation, that of the E conformer being 7:93 throughout while values of 94:6 and 9O:lO were observed for the Z isomer before and after the incubation, respectively. Reaction

542 Chem. Res. Toxicol., Vol. 3, No. 6, 1990 was stopped after 10 min of‘ incubation by adding 0.5 mL of dimedone solution, allowing the mixtures to stand for 30 min, and extracting three times with 1-mL portions of hexane. The com-

bined organic extracts were evaporated in 5-mL glass centrifuge tubes under a gentle stream of argon. The decadeuteriated internal standard prepared as described above was dissolved in hexane (5.1 mg in 2.55 mL), and 7 fiL (46.3 nmol) was added to each tube. Additional hexane was added to dissolve the formaldehyde adduct that had been extracted from the incubation mixture, and the resulting solutions were introduced into the mass spectrometer. An equimolar mixture of the decadeuteriated internal standard ( m / z 302) with the authentic CHzO adduct (m/z 292) was found to give a relative intensity for the two molecular ions of 100:132.59,while an equimolar mixture of the internal standard with authentic CDzO adduct gave an intensity ratio of 100:132.99 for the 302 vs 294 peaks. By use of these ratios as normalization factors, the intensities of the peaks at n / z 292,294, and 302 in the spectra of the incubation mixtures were compared to calculate the amounts of CHPOand CDzO found in each after derivatizing, extracting,and spiking them with internal standard. Multiple scans were collected for each sample, and the single scan with the least intrusive background signals was used to quantify intensities of the relevant ions.

Keefer et al. Scheme I. Possible Stereochemical Results in the a-Hydroxylation of NDMA-dS CH-0

anti

I

-

(t,,2=

N

*O

132 min)

E-NDMA-d,

Results In contrast to the experience with “unsymmetrical” nitrosamines (eq 1, R’ # R) in which the presence of different N-substituents allows the recognition (and

I

slow

7

o+N

Z-NDMA-d3

I o//N

\\IY CD20

sometimes isolation) of discrete E and Z conformers, rotation about the N-N bond of a symmetrical nitrosamine such as NDMA (eq 1,R = R’ = Me) is a degenerate process in which the product is identical with the starting material. Thus it is not possible to determine the stereochemistry of NDMA metabolism by separating the E and Z isomers to measure differences in metabolic rate between them, as Farrelly and his co-workers have done for N-nitrosomethyl(2-oxopropy1)amine( I 7) and N-nitrosomethyl(npentyllamine (181, because E and 2 isomers do not exist for NDMA. By making the two methyl groups isotopically distinguishable, however, this problem can be overcome. If R’ = CH, and R = CD,, for example, the equilibration reaction represented in eq 1 is no longer degenerate. While the individual conformers cannot be separated from one another in this case by chromatography or fractional crystallization, it is possible to prepare them stereoselectively by alkylation of the (E)-or (Z)-methanediazotates with iodomethane-c13 (11). Pure (E)-NDMA-d3thus produced would accordingly be oxidized a t a deuteriated methyl if enzymatic attack occurred anti to the nitroso oxygen but a t the CH3 group if attack is syn to the oxygen, as shown a t the left of Scheme I. For the Z-conformer, the opposite stereochemistry would prevail, as illustrated a t the right of Scheme I. Since numerous previous investigations have demonstrated that deuteriation of NDMA significantly decreases the rate of its metabolism (14, 19-23), we sought to determine the stereochemistry of the process by kinetic measurements utilizing specifically deuteriated substrates. It was reasoned that if cu-hydroxylation occurs exclusively a t the anti carbon of (2)-NDMA-d, or a t the syn methyl of the E-isomer, metabolism should proceed at a high rate similar to that of unsubstituted NDMA. Conversely, syn attack on (Z)-NDMA-d3or anti hydroxylation of the Econformer should be similar in rate to the slower metab-

Table I. Kinetic Parameters for the Conversion of NDMA Isotopomers to Formaldehyde by Liver Microsomes from Acetone-Induced Rats no. of replicate

substrate NDMA (Z)-NDMA-d3

(E)-NDMA-d3 NDMA-d,

incubations 3 3 3 3

Vm, f SD,

nmol/(mg of protein. Km f SD, min) MM 7.47 f 0.26 24 f 6 8.10 f 0.78 32 f 2 7.81 f 0.65 35 f 1 7.49 f 0.30 116 f 3

VmmIKm f SE, mL/ (mg of proteinamin) 0.310 f 0.045 0.250 f 0.017 0.223 f 0.011 0.065 f 0.002

olism of NDMA-d6. Accordingly, we synthesized the Eand Z-isomers of NDMA-d, and measured their rates of conversion to formaldehyde in the presence of rat liver microsomes. Kinetic Studies. Initial work was aimed at confirming the existence of kinetic deuterium isotope effects in the metabolism of NDMA by acetone-induced microsomes, as well as a t measuring their magnitudes in this system. Rates of formaldehyde production were determined a t several different substrate concentrations for NDMA and NDMA-d6 between 0 and 2.4 mM according to established procedures (14). The observed kinetic constants were found to be in good agreement with values from the litVmaxH/VmaxD,and ( VmaX/ erature. Thus the KmH/KmD, Km)H/( Vmax/Km)D isotope effects that characterized the action of the acetone-induced microsomes (0.21, 1.0, and 4.8, respectively) (Table I) were similar to those determined earlier for microsomes of this type (14). The (Z)- and (E)-NDMA-d3preparations, however, instead of behaving like either NDMA or NDMA-d6, were kinetically indistinguishable from each other. As shown in Table I, the K,’s for the (2)-and (E)-NDMA-d3samples were identical within the limit of experimental precision but substantially different from those for NDMA and NDMA-d,; the V,,, values for the four nitrosamine preparations were similar in size in this system. Formaldehyde Trapping Experiments. As an independent check on the conclusions from the kinetic studies,

Stereochemistry of Nitrosamine Metabolism Table 11. Normalized Mass Spectral Ion Intensities for the CH20-and CDzO-Dimedone Adducts Isolated after Incubating NDMA Isotopomers with Acetone-Induced Rat Liver MicrosomesR m / z 292 m l z 294 (CHZO (CD20 substrate microsomes adduct) adduct) 2.27 f 0.23 1.09 f 0.22 none active 0.98 f 0.06 0.61 f 0.07 boiled 14.78 f 0.98 1.01 f 0.16 NDMA active 0.69 f 0.15 boiled 1.39 f 0.17 13.66 f 0.82 1.17 f 0.20 (Z)-NDMA-d3 active 1.15 f 0.14 0.78 f 0.37 boiled 1.79 f 0.14 (E)-NDMA-d3 active 12.13 f 0.22 0.85 f 0.34 boiled 1.29 f 0.17 10.63 f 1.66 NDMA-dfi active 1.73 f 0.29 boiled 1.23 f 0.18 0.60 f 0.13 Values are means f SD of triplicate incubations. Ion intensities are normalized to that of the m / z 302 molecular ion (taken as 100 units) otlserved for decadeuteriated internal standard, 46.3 nmol of which was added to the evaporated hexane extracts of each incuhation mixture before introduction into the mass spectrometer.

product analysis experiments were also conducted. Additional incubations of NDMA, (2)-NDMA-d,, ( E ) NDMA-d3, and NDMA-d6were performed, and the formaldehyde produced was isolated as the dimedone adduct for mass spectral determination of its isotopic composition. The results are shown in Tables I1 and 111. Here too, the (2)-and (E)-NDMA-d3were metabolized preferentially a t the undeuteriated methyl group. The product isolation data thus confirm the conclusions based on the rate studies. Stereochemical Integrity of Substrate. To prove that rapid equilibration of (2)-and (E)-NDMA-d, was not responsible for the failure to observe regiospecificity of enzymatic attack, we measured the rate of equilibration of (2)-NL>MA-d3under the conditions of the incubation. At pH 7.2-7.4 and 37 "C in the presence of uninduced microsomes and cofactors, the mole fraction of 2-conformer dropped from an initial value of 0.95 to around 0.90 during the course of a 20-min incubation, as expected from the previously published half-life of the equilibration in neutral buffer solution (11 ) . Essentially identical results would have been expected if the E-conformer had been used in this experiment, since the equilibrium constant for the 2-E interconversion is K = 1 = kZ-E/kE-z. We conclude that the rate of formaldehyde generation is that of the major NDMA-d3 conformer and that each conformationally biased substrate effectively retains its stereochemical integrity throughout the incubation.

Discussion Our failure in the kinetic studies to observe a significant difference in metabolic rate between the NDMA-d, preparations regardless of the mole fraction of 2-conformer suggested that the cytochrome P450IIE1 in acetone-induced rat liver microsomes attacks the syn and anti methyl groups of NDMA with little or no stereoselectivity. This conclusion is reflected not only in the congruence of rate data for the Z- and E-conformers (Table I) but also in the reasonable correspondence between the magnitudes of the kinetic parameters actually observed and those expected assuming total randomness of metabolic attack. To obtain the expected values, three assumptions were made: metabolic conversion of a CD, group to formaldehyde occurs a t exactly the same rate in NDMA-d, as in NDMA-d6, with the rate of CH, metabolism being the same in NDMA-d3as in NDMA; the CH, and CD, groups of NDMA-d, are equally likely to be directed toward the

Chem. Res. Toxicol., Vol. 3, No. 6, 1990 543 Table 111. Quantities of CHzO and CDzO Produced on Incubating NDMA Isotopomers with Acetone-Induced Rat Liver Microsomeso CHzO produced CDzO produced substrate f SE, nmol f SE, nmol none 0.45 f 0.05 0.16 f 0.05 NDMA 4.67 f 0.20b -0.01 f 0.04 (Z)-NDMA-dS 4.36 f 0.17b 0.02 f 0.08 (E)-NDMA-d, 3.78 f 0.06b1' 0.23 f 0.07 NDMA-dc 0.18 f 0.07 3.49 f 0.34* RAmounts of apparent CH20 and CD20 produced during the 10-min incubations of either 0 or 0.3 pmol of nitrosamine with 0.5 mL of' an active microsomal preparation. The CHzO quantities were obtained by multiplying each value in the mlz 292 (for the molecular ion of CI7Hz4O1)column of Table I1 by 0.349 nmol/intensity unit (a conversion factor obtained by running the mass spectrum of an equimolar mixture of authentic CHzO-dimedone adduct with the decadeuteriated internal standard and dividing the normalized intensity of the m l r 292 peak into the amount of internal standard used) and then subtracting each "boiled" control from its paired "active" value to calculate the net apparent amount of CHzO actually produced metabolically during incubation. CDzO quantities were similarly determined, except that the means in the mlz 294 column of Table I1 were corrected for the 1zC1513C2H2404 contribution at this mass by subtracting 2.6% of the m / t 292 ion intensity before multiplying by the conversion factor, which was 0.348 nmol/intensity unit for the CDzO adduct. bSignificantly greater than the value for the nitrosamine-free control by Student's t test, p 5 0.0001. cSignificantly less than the value for (2)-NDMA-d, by the Student t criterion, p < 0.02.

enzyme's active site, such that on average half the metabolizing loci are kinetically engaged with CH, groups and the other half with CD,'s; metabolism is governed by simple Michaelis-Menten kinetics, so that the total rate of formaldehyde generation a t any given substrate concentration is given by u = 0.5Vm,,H[S]/(K,H

+ [SI) + 0.5V,,,D[S1/(KmD

+ [SI) (2)

By substituting the V,, and K , values determined in parallel experiments with NDMA and NDMA-d6 (Hand D superscripts, respectively), the rate a t which formaldehyde should become colorimetrically observable a t each substrate concentration ([SI)used in the NDMA-d3 experiments was calculated from eq 2. The reciprocals of these calculated velocities were plotted against those of the corresponding [SI values, and the expected V,,, and K , values were estimated by linear regression analysis of the resulting curves. For the three replicate incubations of (E)-NDMA-d, in Table I, the calculated V, value was 6.98 nmol/(mg of protein-min), while a mean of 7.81 nmol/ (mg of protein-min) was actually observed. The calculated K , was 45 FM, compared with an observed mean of 35 pM. For (Z)-NDMA-d,, the calculated and observed values were, respectively, 6.92 vs 8.10 nmol/(mg of protein-min) for V,,, and 44 vs 32 FM for K,. The kinetic data are thus consistent with the conclusion that the orientation of the nitroso oxygen atom has little effect on the interaction between NDMA and the microsomal enzymes metabolizing it. Because this conclusion is a t variance with the tenet that enzyme-catalyzed reactions should normally be highly stereospecific ( 2 4 ) , we sought an independent line of evidence to support it by examining the isotopic composition of the reaction's products. As shown in Table 111, both (2)-and ( E ) NDMA-d3 were metabolized primarily to CHzO rather than CDzO. Significantly more CHzO was generated from the (2)-NDMA-d3than from the E-isomer, suggesting that there might be some net preference for oxidation a t the anti methyl group, but the difference was small compared

544

Chem. Res. Toxicol., Vol. 3, No. 6, 1990

to the total amount of formaldehyde produced. The only substrate from which CD,O production above the mass spectral background (i.e., significantly above the apparent value for the nitrosamine-free control) could be observed was NDMA-d6. The results show that the microsomal preparations employed, whose catalytic activity toward NDMA is predominantly due to cytochrome P450IIE1 (5-7), have little ability to distinguish between the stereochemical]~nonequivalent methyl groups of NDMA. The ( V,,,/K,JH/ ( V,,,,/K,JD isotope effect of 4.8 on the metabolism of NDMA is due almost completely to the difference in K , for NDMA vs NDMA-d,, as reflected in the fact that the corresponding V,,, values are almost identical (Table I). This implies that the K,, as determined in these experiments, is a complex function of several parameters, including the rate constant for the rate-determining step. The data of Table 111 show that, while (Z)-NDMA-d, produced no CD,O above background, (E)-NDMA-d3gave rise to a slightly higher amount of CD20. Moreover, (E)-NDMA-d, showed a significantly lower amount of CH,O than the Z-isomer. These data suggest that the Z-isomer, whose CH, group is anti to the nitroso oxygen, is a slightly better substrate for the demethylase than the E-isomer, whose CH, group is on the same (syn) side as the oxygen. Faced with a choice of breaking a C-H vs a C-D bond with a rate preference for the former of 4.8-fold (Table I), the choice between the anti methyl group and the stereochemically less preferred syn position must be considerably less costly energetically than that between protium and deuterium. Assuming that the observed isotope effect of 4.8 is attributable to simple bond breakage in the rate-determining step, then the activation energy difference between C-H and C-D bond breakage must be a factor of approximately 1.6 (assuming that the preexponential factors for the respective isotopomers are the same). It can be concluded, therefore, that the energy difference between the syn and the apparently preferred anti oxidation of NDMA is much smaller than that, which implies that the substrate-enzyme interaction is very loose and stereochemically poorly defined. The results contrast rather sharply with those for the hepatocyte-induced metabolism of the related dialkylnitrosamine, N-nitrosomethyl(2-oxopropyl)amine, for which there is an all-or-none difference in reactivity between the inert Z-conformer and the active E-form (171, and the microsomal metabolism of N-nitrosomethyl(npentyllamine, for which the relative position of the nitroso group can have a profound effect on the metabolism of the two different alkyl groups (18). Since these are the only other nitrosamines for which the regioselectivity of metabolic attack has to our knowledge been described and the enzymes responsible for their metabolism are probably not the same as those for NDMA, considerable additional work will be required before the reason for the differences observed can be elucidated. Acknowledgment. We thank C. Riggs for performing many of the statistical analyses. Research was supported in part by National Institutes of Health Grants ES-03938 and CA-37037 and Contracts N01-CO-74101 and N01CO-23910.

References (1) Trager, W.F. (1977)Enantiomeric selectivity and perturbation of product ratios as methods for studying the multiplicity of microsomal enzymes. In Drug Metabolism Concepts (Jerina, D. M., Ed.) pp 81-98,ACS Symposium Series 44, American Chemical Society, Washington, DC. (2) Retey, J. (1986)Enzymic stereospecificity as a probe for occurrence of radical intermediates. In Mechanisms of Enzymatic

Keefer et al. Reactions: Stereochemistry (Frey, P. A., Ed.) pp 217-232, Elsevier, New York. (3) Yang, C. S., and Lu, A. Y. H. (1987)The diversity of substrates for cytochrome P-450.In Mammalian Cytochromes P-450, Volume I I (Guengerich, F. P., Ed.) pp 1-17, CRC Press, Boca Raton, FL. (4) Dipple, A., Michejda, C. J., and Weisburger, E. K. (1985)Metabolism of chemical carcinogens. Pharmacol. Ther. 27,265-296. (5) Yang, C. S.,Tu, Y. Y., Koop, D. R., and Coon, M. J. (1985) Metabolism of nitrosamines by purified rabbit liver cytochrome P-450isozymes. Cancer Res. 45, 1140-1145. (6) Patten, C. J., Ning, S. M., Lu, A. Y. H., and Yang, C. S. (1986) Acetone-inducible cytochrome P-450:Purification, catalytic activity, and interaction with cytochrome b5. Arch. Biochem. Biophys. 251, 629-638. (7) Levin, W., Thomas, P.E., Oldfield, N., and Ryan, D. E. (1986) N-Demethylation of N-nitrosodimethylamine catalyzed by purified rat hepatic microsomal cytochrome P-450: Isozyme specificity and role of cytochrome b6. Arch. Biochem. Biophys. 248, 158-165. (8) Karabatsos, G. J.,and Taller, R. A. (1964)Structural studies by nuclear magnetic resonance. IX. Configurations and conformations of N-nitrosamines. J . Am. Chem. SOC.86,4373-4378. (9) Muller, E., Hoppe, W., Hagenmaier, H., Haiss, H., Huber, R., Rundel, W., and Suhr, H. (1962)Struktur isomerer Diazotate: Das Hantzschsche Methyldiazotat. Chem. Ber. 96, 1712-1719. (10) Keefer, L. K., Lijinsky, W., and Garcia, H. (1973)Brief communication: Deuterium isotope effect on the carcinogenicity of dimethylnitrosamine in rat liver. J . Natl. Cancer Inst. 51, 299-302. (11) Keefer, L. K., Wang, S., Anjo, T., Fanning, J. C., and Day, C. S. (1988) Preparation of a thallium(1) diazotate. Structure, physicochemical characterization, and conversion to novel N nitroso compounds. J . Am. Chem. SOC.110, 2800-2806. (12) Tu,Y.Y., Peng, R., Chang, 2.-F., and Yang, C. S. (1983)Induction of a high affinity nitrosamine demethylase in rat liver microsomes by acetone and isopropanol. Chem.-Biol.Interact. 44, 247-260. 13) Lowry, 0.H.,Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951)Protein measurement with the folin phenol reagent. J . Biol. Chem. 193,265-275. 14) Wade, D.,Yang, C. S., Metral, C. J., Roman, J. M., Hrabie, J. A., Riggs, C. W., Anjo, T., Keefer, L. K., and Mico, B. A. (1987) Deuterium isotope effect on denitrosation and demethylation of N-nitrosodimethylamine by rat liver microsomes. Cancer Res. 47, 3373-3377. 15) Nash, T.(1953)The colorimetric estimation of formaldehyde by means of the Hantzsch reaction. Biochem. J . 55, 416-421. 16) Keefer, L.K., Hrabie, J. A., Hilton B. D., and Wilbur, D. (1988) Nitrogen protonation of N-nitrosodimethylamine. J . Am. Chem. SOC.110,7459-7462. 17) Farrelly, J. G., Stewart, M. L., Farnsworth, D. W., and Saavedra, J. E. (1988) Metabolism of the 2 and E isomers of N nitroso-N-methyl-(2-oxopropyl)amine by rat hepatocytes. Cancer Res. 48,3347-3349. 18) He, C., and Farrelly, J. G. (1989)Microsomal metabolism of the 2 and E isomers of N-nitroso-N-methyl-N-n-pentylamine. Cancer Lett. (Shannon,Irel.) 45, 189-194. 19) Dagani, D., and Archer, M. C. (1976)Brief communication: Deuterium isotope effect in the microsomal metabolism of dimethylnitrosamine. J . Natl. Cancer Inst. 57,955-957. See also correction in: Dagani, D., and Archer, M. C. (1978)J . Natl. Cancer Inst. 61,619. (20) Kroeger-Koepke, M. B., and Michejda, C. J. (1979)Evidence for several demethylase enzymes in the oxidation of dimethylnitrosamine and phenylmethylnitrosamine by rat liver fractions. Cancer Res. 39, 1587-1591. (21) Swann, P.F., Mace, R., Angeles, R. M., and Keefer, L. K. (1983) Deuterium isotope effect on metabolism of N-nitrosodimethylamine in uiuo in rat. Carcinogenesis (London) 4, 821-825. (22) Mico, B. A., Swagzdis, J. E., Hu, H. S.-W., Keefer, L. K., Oldfield, N. F., and Garland, W. A. (1985)Low-dose in uiuo pharmacokinetic and deuterium isotope effect studies of N-nitrosodimethylamine in rats. Cancer Res. 45,6280-6285. (23) Streeter, A. J., Nims, R. W., Sheffels, P. R., Heur, Y.-H., Yang, C. S., Mico, B. A,, Gombar, C. T., and Keefer, L. K. (1990)Metabolic denitrosation of N-nitrosodimethylamine in uiuo in the rat. Cancer Res. 50, 1144-1150. (24) Rose, I. A. (1972)Enzyme reaction stereospecificity: A critical review. Crit. Reu. Biochem. 1, 33-57.