Chem. Res. Toxicol. 1990, 3, 21-26
21
Low Kinetic Hydrogen Isotope Effects in the Dehydrogenation of 1,4-Dihydro-2,6-dimethyl-4-( 2-nitropheny1)3,5-pyridinedicarboxylic Acid Dimethyl Ester (Nifedipine) by Cytochrome P-450 Enzymes Are Consistent with an Electron/Proton/Electron Transfer Mechanism F. Peter Guengerich Department of Biochemistry and Center in Molecular Toxicology, Vanderbilt University, Nashville, Tennessee 37232 Received August 1 , 1989
Cytochrome P-450 enzymes have been postulated to oxidize amines through a variety of mechanisms. One of the means of distinguishing among potential pathways involves the use of kinetic hydrogen isotope effects: low isotope effects are characteristic of aminium radical mechanisms while high values are consistent with hydrogen atom abstraction, a process documented in alkane hydroxylation. Nifedipine [ 1,4-dihydro-2,6-dimethyl-4-(2-nitrophenyl)-3,5pyridinedicarboxylic acid dimethyl ester] was prepared with isotopic substitution a t the 4-position and utilized in determinations of the deuterium and tritium kinetic isotope effects. The noncompetitive intermolecular deuterium [DV, D(V/K>] and competitive intermolecular tritium [T(V / K ) ]isotope effects observed in the oxidation of nifedipine with liver microsomes prepared from untreated male rats were 99% chromatographically pure (A254) as judged by HPLC in the solvent system used for analysis (vide infra) and was identical with the undeuterated material in all of its physical properties except for the presence of the single deuterium atom. 'H NMR analysis a t high sensitivity and comparison to other signals indicated an atomic excess of >99.8% a t the 4-H position:
by using A,,,) was collected in an amber bottle, and the CH30H was removed under reduced pressure-the nifedipine was extracted three times with equal volumes of CH2C12,and the combined layers were dried with Na2S04and reduced in vacuo to yield 30 pmol of [4-3H]nifedipine. I t should be emphasized that all of these manipulations were done in amber glass or in the absence of room light. The final product was >99.8% pure as judged by analytical HPLC in a system similar to that used for the preparative chromatography (Azu), and >97.8% of the radioactivity injected on the column was recovered under the nifedipine peak. The material exhibited a UV spectrum that was identical with that of authentic (unlabeled) nifedipine; the specific radioactivity was determined to be 3.64 (i0.39) mCi/mmol. 2,6-Dimethyl-4-(2-nitrophenyl)-3,5-pyridinedicarboxylic acid dimethyl ester was prepared by H N 0 3 oxidation of nifedipine as described elsewhere (12, 38): yield 74%; mp 100-101 "C; UV 264, tm 10500 M-' cm-'; 'H NMR (C2HC13)6 2.65 (CH,OH) A, (s, 6 H, 2- and 6-CH3 groups), 3.52 (s, 6 H, O-CH3), 7.17-8.23 (m, 4 H, phenyl protons); MS (electron impact) m / z 345 (M, 9), 313 (M - 32,41), 299 (M - 46,100). Anal. Calcd for cl7Hl6N2o6:c , 59.30; H, 4.65; N, 8.14. Found: C, 59.84; H, 4.67; N, 8.05. Enzymes. Sprague-Dawley rats (150-200 g) were purchased from Harlan Industries, Indianapolis, IN, and treated as described elsewhere (39,40). Human liver samples were obtained from organ transplant donors through the Nashville Regional Organ Procurement Agency, Nashville, TN. Microsomal preparations were prepared and stored in 10 mM Tris-acetate buffer (pH 7.4) containing 1mM EDTA and 20% glycerol (v/v) at -80 "C in small aliquots. Rat liver P-450 P-450pB-~and rabbit liver NADPHP-450 reductase were prepared and used as described elsewhere (41). Horseradish peroxidase (type XII) was purchased from Sigma Chemical Co., St. Louis, MO, and used without further treatment (assays were done in 100 mM sodium phosphate buffer (pH 7.4) in the presence of 5 mM H202). Assays. The oxidation of nifedipine to 2,6-dimethyl-4-(2nitrophenyl)-3,5-pyridinedicarboxylicacid dimethyl ester was assayed essentially as previously described ( I I ) , with separation of the product by the use of HPLC (same column as used previously-Zorbax columns are now supplied by Mac-Mod, Chadds Ford, PA). When the release of tritium was measured as 3H20,separation was done by partitioning samples four times between the aqueous solution and CH2C1,, with the organic phase
Nifedipine Oxidation Mechanism
Chem. Res. Toxicol., Vol. 3, No. 1, 1990 23
Table I. Oxidation of Nifedipine by Various Microsomal Preparations, P-450pe.B, and Horseradish Peroxidase nmol of product formed/ (min-nmol of enzyme)O 2,6-dimethyl-4-(2-nitrophenyl)-3,5-pyridinedicarboxylic acid dimethyl ester
enzyme preparation rat microsomes (male, untreated) rat microsomes (male, isosafrole treated) rat microsomes (male, ciprofibrate treated) rat microsomes (male, phenobarbital treated) rat microsomes (male, isoniazide treated) rat microsomes (female, phenobarbital treated) rat microsomes (female, pregnenolone 16a-carbonitrile treated) human microsomes (HL 105) human microsomes (HL 110) human microsomes (HL 124) p - 4 5 0 (IIB1) ~~~ horseradish peroxidase
from 4-2H-labeled substrate 0.84 f 0.13 2.04 f 0.14 1.10 f 0.20 1.17 f 0.36 0.79 f 0.09 0.43 f 0.02 0.83 & 0.04 2.79 f 0.23 5.82 f 0.54 0.88 f 0.10 0.16 f 0.05 1.56 f 0.23
from unlabeled substrate 1.41 f 0.04 2.85 f 0.10 1.56 f 0.18 1.40 f 0.03 1.43 f 0.04 0.64 f 0.13 1.45 f 0.07 3.71 f 0.22 8.08 f 0.39 1.37 f 0.08 0.32 f 0.07 2.48 f 0.55
SH20 from 4-SH-labeled substrate 0.99 f 0.05 2.32 f 0.16 1.12 f 0.04 1.33 f 0.15 1.07 f 0.01 0.36 f 0.11 0.80 f 0.06 3.62 f 0.04 8.80 f 0.38 1.63 f 0.04 0.23 f 0.04 1.39 f 0.38
D ( V / K ) T(V/K) 1.7 1.4 1.4 1.2 1.8 1.5
1.4 1.2 1.4 1.1 1.3 1.8 1.8 1.0b 0.9b 0.8b 1.4 1.8
1.8
1.3 1.4 1.6 2.1 1.6
The reaction time was 4 min, and incubations contained 1 nmol of enzyme. All results are expressed as means of four replicates f SD except the isotope effects, which are simply expressed as ratios. bThese values are not as reliable as the others because of the large extent of sample oxidation. being used for the analysis of the oxidized nifedipine product (31). All incubations (as well as synthetic procedures-vide supra) were done in amber glass vials, because of the high sensitivity of nifedipine t o light (34). It should be emphasized that, in all work with light-sensitive materials, aluminum foil is not effective in blocking light, for it only acts as a mirror.2
Results The nifedipine compounds used in this study were carefully prepared and fully characterized, as described under Experimental Procedures. All work with nifedipine was done in amber glass because of the inherent light sensitivity of solutions of the material (34). Rates of oxidation of unlabeled ([4-'H]-) and [4J!H]nifedipine by rat liver microsomes are shown in Figure 1 as a function of the nifedipine concentration. The data were analyzed in several ways with the usual linear transformations (42),and both DV and D(V/K) were consistently in the range of 1.5-2 regardless of the method used. A Hanes-Woolf plot is shown in Figure 2. The possibility was considered that the full intrinsic kinetic deuterium isotope effect (Dk) might be attenuated in the noncompetitive experiments described in Figures 1 and 2. Since nifedipine oxidation is for all purposes irreversible, Northrop's method (3) for estimating DK should be applicable, if T( V / K ) can be utilized. [4-3H]Nifedipine was synthesized as described and used in what is necessarily a competitive experiment (Figure 3), with both 3Hz0 and the organic nifedipine product being measured. Estimation of the tangents to the early portions of the slopes yields an estimate of *( V / K ) of 1.4. Under the conditions of the entire incubation, the extent of oxidation was extensive (about half), but the measured specific radioactivity of the recovered substrate remained ~~
In response to a suggestion of one of the reviewers, nifedipine oxidation was carried out in the presence of room light (in clear glass) using the nifedipine samples that were prepared and characterized here. The measured rate of oxidation of nifedipine to the primary pyridine product was 67% of that observed under the usual conditions, but several major additional (uncharacterized) products were obtained in the unprotected reactions, including a prominent one migrating with a retention time between those of the pyridine product and nifedipine. However, the apparent DV (only one set of assays was done at a substrate concentration of 200 rM) was 1.3, which is still consistent with the results presented in this paper.
c $
100 [Nifedipine], uM
00
200
Figure 1. Rates of oxidation of nifedipine and [4-2H]nifedipine by rat liver microsomes as a function of nifedipine concentration. Microsomes were prepared from untreated male rats. The general methods are described under Experimental Procedures. ( 0 ) Unlabeled nifedipine; (0)[4-H2]nifedipine.
0'
0
'
'
10
'
'
20
'
"
30
'
'
40
'
50
'
'
60
[Nifedipinel, uM
Figure 2. Rates of oxidation of nifedipine and [4-2H]nifedipine by rat liver microsomes as a function of nifedipine concentration. The data points of Figure 1were fitted to the Hanes-Woolf linear transformation of S / u versus S (lines fitted by linear regression analysis). ( 0 )Unlabeled nifedipine; (0) [4-2H]nifedipine. This transform yielded the following values for unlabeled nifedipine: V-, 1.37 nmol of product formed/(min.nmol of P-450); K,, 5.9 p M V,,/K,, 0.234 nmol of product formed/(min.nmol of P450-bM). For [4-2H]nifedipine,the values were as follows: V-, 0.86 nmol of product formed/(minmmol of P-450); K,, 7.3 bM; V-/K ,0.118 nmol of product formed/(min.nmol of P-450-bM). Thus, $V = 1.59 and D ( V / K )= 1.98.
24
Chem. Res. Toxicol., Val. 3, No. 1, 1990
Guengerich
.
.-C
15
ti
8-
2
6-
%E 10
i
5 c
0
Time, min
Figure 3. Oxidation of [4-3H]nifedipineto the product 2,6-dimethyLC(2-nitrophenyl)-3,5-pyridinedicarboxylicacid dimethyl ester and to 3Hz0in rat liver microsomes as a function of time. The methods are described under Experimental Procedures, and the microsomes were prepared from untreated male rats. The specific radioactivity (A)was estimated at each indicated point from measurements of the integral of the nifedipine AzM HPLC peak and the radioactivity recovered in the recovered material. 3HzO;(m) The initial substrate concentration was 50 rM. (0) 2,6-dimethyl-4-(2-nitrophenyl)-3,5-pyridinedicarboxylic acid dimethyl ester.
$E .-
t
. I
5 2Q
!
0
2
,
4
,
6
,
,
,
6
,
1
1 0
nmol Nifedipine Oxidizedlminl nmol P-450
Figure 5. Correlation of rates of formation of 3Hz0and the product 2,6-dimethyl-4-(2-nitrophenyl)-3,5-pyridinedicarboxylic acid dimethyl ester from [4-3H]nifedipinein different microsomal preparations.
i
4t
2
;
210,i,
2 CT)
0 0
"
0 2 4 6 8 nmol Unlabeled Nifedipine Oxidizedl minlnmol P-450
Figure 4. Correlation of rates of formation of the product 2,6dimethyl-4-(2-nitrophenyl)-3,5-pyridinedicarboxylic acid dimethyl ester from unlabeled nifedipine and [4-2H]nifedipinein different microsomal preparations (noncompetitiveexperiments). Points are taken from Table I (r2 = 0.97). relatively constant, consistent with the view that only a low kinetic tritium isotope effect is seen. The low kinetic deuterium and tritium isotope effects seen with liver microsomes prepared from untreated rats might be considered peculiar to only the P-450s primarily involved in nifedipine oxidation there. Less detailed studies were carried out with a number of other preparations (fixed incubation time of 4 min, 80 p M nifedipine). This concentration of substrate should yield estimates of D V (Figure 1). As indicated in Table I, in all cases relatively low values for D Vand T( V / K )were estimated, even with a purified rat P-450P~.B(P-450IIB1) system and in the case of nifedipine oxidation supported by horseradish peroxidase in the presence of HzOz. When the rates of product formation are plotted against each other (Figures 4 and 5), it can be seen that, although the rates cover a wide range and probably reflect the contribution of at least three different enzymes in rats and humans (II), the correlation coefficients are excellent and there is little evidence that certain P-450s deviate from the basic finding of low isotope effects.
Discussion When alicyclic amines are oxidized by electrochemical
H3c02c)tJ
co2cn3
(fast)
H3C
H
CH3
H3C&CH3
H
means, the observed kinetic deuterium isotope effects are low ( 2) and similar kinetic isotope effects are observed in P-450-catalyzed amine oxidations (25, 26). Dihydropyridines, like flavins, can undergo either one- or twoelectron transfers. In reactions with two-electron acceptors, such as pyridine nucleotide linked dehydrogenases, shielded hydride transfer is observed (43). With obligate one-electron acceptors, however, electron/proton/electron transfer pathways are seen in model systems (44-47). With model acceptors where electron transfer is fast (e.g., high oxidation potential and no kinetic barriers), proton loss can be rate-limiting and the kinetic deuterium isotope effect for loss of the C-4 hydrogen can be high unless a general or specific base is present (45). If one-electron transfer is slow, then the rate of transfer of the C-4 proton is not rate-limiting and the observed kinetic deuterium isotope effect is low (Scheme 11). The available evidence argues that an aminium radical pathway is used in the oxidation of 4-substituted 1,4-dihydropyridines by P-450s. Previous evidence argues against C-4 hydrogen abstraction in that no large differences were observed in the rates of oxidation of the enantiomers of 1,4-dihydro-2,6-dimethyl-3-nitro-4[2-(trifluoromethyl)phenyl]-5-pyridinecarboxylicacid methyl N
Chem. Res. Toxicol., Vol. 3, No. 1, 1990 25
Nifedipine Oxidation Mechanism
mation about the tertiary structures of these enzymes. The ester (Bayer K8644) (13). Further, the rates of oxidation possibility can also be considered that the putative (FeOY+ of variously substituted 4-phenyl-l,4-dihydropyridine degenerated by the first electron abstraction might itself act rivatives are relatively similar (12, 48), while one might as a specific base. An explanation may lie in consideration expect that if a benzylic radical were an intermediate, then of the rates of loss of the 4-proton in model compounds. electron donation and withdrawal by ring constituents Manring and Peters (49)have measured rapid rearrangewould have an influence on the rates of oxidation (although ment rates in N-methylacridane radicals, and Sinha and it should be pointed out that a delocalized radical is also Bruice (47) have estimated the rate of general-base-catafound in the deprotonated intermediate shown in Scheme lyzed proton loss from the minium radical of a pyridine 11). Therefore, it was surprising to see a report of a large nucleotide analogue to be on the order of lo7 M-’ s-l for DV value for nifedipine oxidation. A large DV coupled with imidazole, acetate, and formate and lo4 M-’ s-l for HzO. a low D( V / K ) (33)might be interpreted as evidence that If these rate constants are operative in the enzyme interior, a step late in the overall reaction scheme (e.g., product then proton transfer is much more rapid than the initial release) is rate-limiting (3). However, the possibility seems unlikely since studies with 2,6-dimethyl-4-pheny1-3,5- electron abstraction in all of the cases examined here (Scheme 11). pyridinedicarboxylic acid dimethyl ester are consonant with rapid exchange in the P-450 site (Le., high kinetic Acknowledgment. This work was supported in part deuterium isotope effect for material labeled in the ester by Grants CA 44353, ES 00267, ES 01590, and ES 02205 methyl group in competitive intermolecular experiments) from the National Institutes of Health. I thank Dr. Ian (31). Further, there does not seem to be a reason why Blair and Brian Nobes for recording the mass spectra and nifedipine oxidation should proceed via C-4 hydrogen Dr. W. Griffith Humphreys for the NMR spectra. abstraction, as proposed by Born and Hadley (33),while the model 1,4-dihydro-2,6-dimethyl-4-phenyl-3,5- Registry No. P-450, 9035-51-2; nifedipine, 21829-25-4; [42H]nifedipine, 118950-27-9; [4-3H]nifedipine, 124460-99-7; mopyridinedicarboxylic acid does not (13), for an o-nitro nooxygenase, 9038-14-6; 2-nitrobenzoyl chloride, 610-14-0; deugroup would not be expected to stabilize a benzylic radical. terated 2-nitrobenzaldehyde, 50344-83-7; methyl acetoacetate, The question of whether a large or small kinetic deu105-45-3; 2-nitrobenzaldehyde, 552-89-6; 2-nitro[aldeh~de-~H]benzaldehyde, 124442-25-7; deuterium, 7782-39-0; tritium, terium isotope effect occurs in the removal of the C-4 10028-17-8. hydrogen was reexamined with carefully prepared substrates. The initial results (Figures 1and 2) indicated that References both low D V and D( V / K )values were observed in micro(1) Guengerich, F. P. (1987) Enzymology of rat liver cytochromes somal oxidation. The possibility still existed that a high P-450. In Mammalian Cytochromes P-450 (Guengerich, F. P., intrinsic isotope effect might be inherent in the system but Ed.) Vol. 1, pp 1-54, CRC Press, Boca Raton, FL. be attenuated. The approach of Northrop (3)is applicable (2) Nebert, D. W., Nelson, D. R., Adesnik, M., Coon, M. J., Estahere since in the effectively irreversible reaction the brook, R. W., Gonzalez, F. J., Guengerich, F. P., Gunsalus, I. C., magnitude of the commitment to reverse catalysis factor Johnson, E. F., Kemper, B., Levin, W., Phillips, I. R., Sato, R., and (C,) approaches zero. However, rather low T(V/K) values Waterman, M. R. (1989) The P450 superfamily: update on listing of all genes and recommended nomenclature of the chromosomal were measured (Figure 3). Therefore, a high intrinsic loci. DNA 8, 1-13. kinetic deuterium isotope effect cannot be associated with (3) Northrop, D. B. (1982) Deuterium and tritium kinetic isotope this system. The value of Dk cannot be measured accueffects on initial rates. Methods Enzymol. 87, 607-625. rately because of the large error associated with the use (4) Wislocki, P. G., Miwa, G. T., and Lu, A. Y. H. (1980) Reactions of this method with low values of D( V / K ) and T( V / K ) (3), catalyzed by the cytochrome P-450 system. In Enzymatic Basis but the only qualitative conclusion that can be reached is of Detoxication (Jakoby, W. B., Ed.) Vol. 1, pp 135-182, Academic Press, New York. that the isotope effects are probably
