The Catalytic Mechanism of Cytochrome P-450 - Semantic Scholar

THEJOURNALOF

BIOLOGICAL CHEMISTRY

Vol 257. No. 1Y. Issue of October 10. pp. 11288-11295. 1982 Prrnted En 1i.SA.

The Catalytic Mechanismof Cytochrome P-450 SPIN-TRAPPING EVIDENCE FOR ONE-ELECTRON SUBSTRATE OXIDATION* (Received for publication, March 8, 1982)

O h a r a Augusta$, Hal S. Beilan, and Paul R. Ortiz de Montellano$j From the Departmentof Pharmaceutical Chemistry, School of Pharmacy, a n d Liver Center, University of California, Sun Francisco, California 94143

Cytochrome P-450 is destroyed during catalytic oxidation of several 4-substituted 3,5-bis(ethoxycarbonyl)%,tdimethyl-1,4-dihydropyridine substrates. A qualitative correlation has been found between the ability to destroy cytochrome P-450 and the stability of the 4substituent as a radical. Destruction of the enzyme by the 4-ethyl (DDEP), 4-propyl, and 4-isobutyl analogues is due to transfer of the 4-alkyl group f r o m the substrate to a nitrogen of the prosthetic heme, a process which gives rise to isolable N-alkylprotoporphyrin IX derivatives. Little enzyme destruction is observed when the 4-alkyl group is of low radical stability (methyl, phenyl) and good destruction, but no isolable heme adducts when the 4-substituent is of very high radical stability (isopropyl, benzyl). Spin-trapping studies have established that the 4-ethyl group in DDEP is lost as a radical as a result of oxidation by cytochrome P-450. Of three commonly used spin traps, only a-(4-pyridyl-l-oxide) N-tert-butylnitrone was found suitable f o r such studies. The other spin traps, 5 4 dimethyl-1-pyrroline-N-oxide and a-phenyl N-tert-butylnitrone, were found to be ineffective, the latter because it strongly inhibits cytochrome P-450. Hydrogen peroxide formed in situ can support a part of the cytochromeP-450-catalyzed ethyl radical formation and DDEP-dependent self-inactivation.The results provide persuasive evidence that oxidation of the nitrogen in D D E P b y cytochrome P-450 proceeds in one-electron steps. Cytochrome P-450 may thus function, at least with certain substrates, as a one-electron oxidant.

m d y at the oxene oxidation state, from the enzyme to the acceptor function on the substrate. Themolecular details of the oxygen transfer, however, have only recently begun to yield to investigation. The principal questions to be resolved are whetheroxygen transfer occurs in aconcerted orstepwise fashion, and, if stepwise, whether the reaction proceeds by free radical or paired electron steps. Seminal experimentson carbon hydroxylation using deuterium scrambling as a probe have providedconvincingevidence that, at least in some instances, carbon hydroxylation proceeds via a nonconcerted (probably free radical) oxidative mechanism (1,Z). Our exploration of the autocatalytic destruction of cytochrome P-450 associated with terminal olefin metabolism points to the intervention of a transient intermediate which partitions between formation of the epoxide metabolite and prosthetic heme’ alkylation (3-6). The regiospecificity of the reactionof substituted olefins with prosthetic heme argues for a radical rather than cationic intermediate in olefin oxidation (7). Evidence also exists for radical intermediates in the enzymatic transfer of oxygen to nitrogen or sulfur electron pairs. The incubation of aminopyrine with liver microsomes or purified cytochrome P-450 in the presence of cumene hydroperoxide, for example, results in the formation of aminopyrine radicals (8-10). Further supportfor one-electron heteroatomoxidation mechanisms is provided by the finding that the rate of Soxidation of sulfur compounds by microsomes correlates well with the one-electronoxidation potentials of the sulfur atoms (11). A critical need nevertheless exists for new approaches and for new evidence relevant to the question of whether cytochrome P-450 functionsundernormalNADPH-supported conditions as a one- ortwo-electron oxidant. Cytochrome P-450 has recently been shown to be inactiCytochrome P-450 enzymes perform three general typesof of DDEP (12). Enzyme loss is oxidative reactions: ( a ) insertion of an oxygen atom into the vated during catalytic turnover bond between a hydrogen and a heavier atom to yield the coupled to the formation of an abnormal hepatic pigment (12). TheN-ethyl corresponding hydroxyl derivative, ( b ) addition of an oxygen identified asN-ethylprotoporphyrinIX atom acrom the two carbons of a 8-bond toyield an epoxide, group derives from the 4-ethyl moiety in the substrate. Enand ( c ) addition of an oxygen to the electron pair of a hetero- zymatic oxidation of DDEP thus generates an activated speatom to give a dipolar oxide. The most common examples of cies, the 4-ethylgroup of which is transferred to the prosthetic these reactions are carbon hydroxylation, olefin epoxidation, heme of the enzyme. Characterization of the activated interof the mechand N-oxide formation. The central featureof all these reac- mediate is therefore crucial to our understanding tions is the transfer of a catalytically activated oxygen, for- anism by which the enzyme is inactivated. The activated intermediate furthermore bears the imprint of the catalytic mechanismwhich engendersitand is thus,potentially, a * This work was supported by National Institutes of Health Grants unique and intimate probe of that mechanism. In particular, GM-25515, HL-15476,and P50 AM-19520. The costs of publication of this article were defrayed in part by the payment of page charges. three types of mechanisms can be envisagedforoxidative This article must therefore be hereby marked “aduertisement” in conversion of DDEP to a species which transalkylates the

accordance with 18 U.S.C. Section 1734 solely to indicate this fact. On leave of absence from the Universidade de Si0 Paulo, SBo Paulo, Brazil. Recipient of fellowships from the Fundacio deAmparo i Pesquisa do Estado de Si0 Paulo and Conselho Nacional de Desenvolvimento Cientifico e Tecnologico. 8 Research Fellow of the Alfred P. Sloan Foundation (1978-1982). Address correspondence to this author.

+

’

The abbreviations and restrictively defined term are: heme, iron protoporphyrin IX regardless of the iron oxidation state; DMPO, 5,5dunethyl-1-pyrroline-N-oxide; PBN,a-phenyl N-tert-butylnitrone; POBN, a-(4-pyridyl-l-oxide) N-tert-butylnitrone; DDC, 3.5-bis(ethoxycarbonyl)-2,4,6-trimethyl-l,4-dihydropyridine; DDEP, 3,5bis(ethoxycarbonyl)-4-ethyl-2,6-dimethyl-l,4-dihydropyridine.

11288

One-electron Oxidation

11289

treated Sprague-Dawley male rats (1 mg/ml), KC1 (150 mM), EDTA (1.5 mM), and the test substratein 0.1 M NaK phosphate buffer (pH 7.4) were initiated by addition of NADPH (1.0 mM final concentration). Incubations were carried out at 37 "C and lasted 30 min. The cytochrome P-450 content of the microsomes was determined by difference spectroscopy, as described by Estabrook et al. (18), before NADPH was added and after the30-min incubation period. The loss of cytochrome P-450 in the absence of substrate but in the presence of NADPH (lipid peroxidation control), and in the presence of substrate but in the absence of NADPH, was also routinely measured. The loss of enzyme in the control incubations, although negligible (12%), has been subtracted from the reported values. An Aminco DW2A instrument was used for the cytochrome P-450 assays. Alterations in the above standard incubation system and procedure are indicated for the appropriate experiments in the text. Cytochrome P-450 losses, generally measured in at least two separate experiments, are given in the tables as theper cent decrease relative to the amount of enzyme originally present in the microsomes. MATERIALS AND METHODS The binding of DDEP and the spin traps tounreduced cytochrome The spin traps PBN, DMPO, and POBN were purchased from P-450 was assayed by difference spectroscopy (18, 19). The peak-toAldrich Chemical Co. DMPO was purified before use as described by trough absorbance difference was measured as a function of the Buettner and Oberley (13). Nifedipine (3,5-bis(ethoxycarbonyl)-2,6- substrate concentration and wasused to determine the apparent dimethyl-4-(2-nitrophenyl)-1,4-dihydropyridine) was a kind gift from binding constant. Nominal concentrations were used in the studies Dr. F. C. Greenslade, Pfizer, Inc. DDC and 4-hydroxy-2,2,6,6-tetra- with spin traps. Preliminary experiments using a gas chromatographic methylpiperidinooxy radical were obtained from Aldrich. The syntheanalytical method established that approximately 0.06 mM DDEP sis of the 4-ethyl DDC analogue DDEP (3), has been reported (12). saturated a microsomal suspension. The 4-propyl (4), 4-isopropyl (5), 4-isobutyl (6), 4-benzyl (7), and 4Benzphetamine demethylation was assayed under the conditions phenyl (8) analogues were synthesized by method A of Loev et al. of Correia and Mannering (20). The formaldehyde generated in the (14). Briefly, a solution of ethyl acetoacetate (0.20 mol), concentrated reaction was quantitated by the procedure of Nash (21). N S O H (0.15 mol), and the appropriate aldehyde (0.10 mol) in 50 ml Isolation of Hepatic Pigments-The detailed procedure for isolaof ethanol was refluxed for approximately 12 h. The solid which tion of abnormal ("green") porphyrins from the livers of phenobarprecipitated out was collected by filtration of the cooled mixture and bital-pretreated Sprague-Dawley rats injected with dihydropyridine was recrystallized from ethanol/water. If the product separated as an derivatives (400 mg/kg dose) has been reported (12). The abnormal oil, it was added to ice water (250 m l ) and the solidified material was porphyrins in this studyhave been obtained in sufficient quantity for collected and recrystallized from ethanol/water. The 4-unsubstituted characterization by electronic absorption and field desorption mass analogue 1 was prepared similarly except that the mixture of ethyl spectrometry,butnot for detailed NMR structural analysis. The acetoacetate (26.0 g, 0.20 mol), paraformaldehyde (3.0 g,0.10 mol), electronic absorption spectra of the metal-free and zinc-complexed ammonium carbonate (48.0 g, 0.50 mol), andwater (500 ml) was porphyrins were recorded in methylene chloride on a Cary-Varian stirred at 25 "C for 24 h before product isolation and recrystallization. Model 118 spectrophotometer. Field desorption mass spectra were The dihydropyridine products, characterized by electronic absorption, obtained on a modified Kratos MS-9 instrument at the Biomedical NMR, andmass spectroscopy, had physical properties consistentwith and EnvironmentalMassSpectrometry Resource (Berkeley, CA) those previously reported for the assigned structures: 1, 5, 7, 8 (14), under conditions similar to those reported previously ( 6 ) . 4 (15),and 3 (12, 16). The elemental analysis for 6, a previously EPR Studies-A hepatic microsomal suspension containing 2.5 undescribed analogue, was consistent with the assigned structure. nmol/ml of cytochrome P-450 was used for the EPR experiments. The aromatic derivatives 10 and 11 were prepared by oxidation of The microsomal preparation was obtained in thesame manner as the dihydropyridine precursors 5 and 3, respectively, with nitrous acid preparation used for cytochrome P-450 assays. The standard incu(16). N-Ethyl DDEP (12)was synthesized by condensation of probation mixture contained, in addition to the microsomal protein, pionaldehyde, ethyl acetoacetate, and ethylamine hydrochloride by NADPH (1 mM), EDTA (1.5 mM), KCL (150 mM), DDEP (10 mM), the method used previously for the synthesis of N-ethyl DDC (12). and the desired spin trap (usually POBN, 20 mM), all in 0.1 M NaK Benzphetamine was kindly provided by Dr. G. Miwa, Merck Sharp phosphate buffer (pH 7.4). Variations in the incubation system are and Dohme. indicated in the text. EPR analyses were carried out on aliquots of Bovine thymol-free catalase (nominally 10,000-25,000 units/mg), the mixture taken after 30 min of incubation at 37 "C. The aliquots bovine superoxide dismutase (type I), NADPH,and NADH were were put into a capillary pipette or into gas-permeable tubing (inner purchased from Sigma Chemical Co. The specific activity of catalase diameter, 0.81 mm, wall thickness 0.05 mm; Zeus Industrial Products, (13,000units/mg) was measured as described in the Sigma catalogue. Inc., Raritan, NJ). EPR spectra were recorded a t room temperature The specific activity of superoxide dismutase (2000 units/mg) was on either a Varian Model E 109-E or E-3 spectrometer. To estimate assayed as described in the literature(17). the approximate concentration of spin-trapped ra&cals, the peak I n Vitro Cytochrome P-450Studies-The protocol used to measure height of the low field signal of the POBN adduct was compared with the in vitro loss of cytochrome P-450 due to incubation of hepatic that of known concentrations of the 4-hydroxy-2,2,6,6-tetramethylmicrosomes with destructiveagents has been described (3-5, 12). piperidinooxy radical. Incubations containing microsomal protein from phenobarbital-preIdentification of the POBN-trapped Ethyl Radical-In order to obtain an authentic standard with which to develop a purification sequence, ethyl radicals were first generated in the presence of POBN by CuC12-catalyzeddecomposition of ethylhydrazine (22). Ethylhydrazine (30 mg) and POBN (100 mg), dissolved in 5.0 ml of an 0.2 mM CuC12 solution in 0.05 M carbonate buffer (pH lo), were stirred 2 h at approximately 20-25 OC. The resulting product mixture was lyophilized and the residue was extracted with 5 ml of acetone. The extract was concentrated on a rotary evaporator and was then chromatographed on 500pm thick Silica Gel G (Analtech, Inc.) thin layer plates. The plates were developed in 5% (v/v) methanol/chloroform. H Et The band containing the adduct ( R Fapproximately 0.4) wasremoved from the plate, extracted with the same solvent, and rechromatographed under the same conditions. The spin-trapped radical was located on the thin layer plates by EPR analysis of extracts of I fractions taken from the plate. Quantitation of the radical prior to 0" purification indicated the formation of approximately 0.5 mg of adSCHEME 1 duct. Similar assay of the final purified material indicated a final yield

prosthetic heme group, although subtle aspects of the mechanisms may vary (Scheme 1).The dihydropyridine ring nitrogen can undergo a two-electron oxidation to the unstable hydroxylamine. The hydroxylamine can then aromatize by loss of the 4-ethyl group as a cation (path b) or, after further oxidation to thenitroxide radical, by loss of the 4-ethyl group ) , radical cation as a radical (pathc). Alternatively ( p a t h athe produced by one-electron oxidation of DDEP could aromatize by ejecting an ethyl radical in a reaction competitive with collapse to the hydroxylamine product of paths a and b. We report here a spin-trappingstudy which differentiates among these three alternatives and strongly suggests that the oxidation of nitrogen by cytochrome P-450,at least in this instance, is mediated by one-electron steps.

Oxidation

11290

One-electron

of about 0.2 mg of adduct. The mass spectrum of the isolated material was recorded on a Kratos MS-25 under electron impact conditions (70 eV ionization voltage). A largescale (100 ml)incubation of the standard microsomal preparation, containing DDEP(IO mM), NADPH(1mM), and POBN (20 mM), was placed in a reciprocating bath at 37 “C for 30 min. The mixture was then centrifuged for 30 min (4 “ C ) at 100,OOO X g. The supernatant was lyophilizedand the residue was extracted with acetone. The extract was concentrated and the adduct purified as

described for the ethylhydrazine-derived radical, except that the thin layer chromatographic purification was performed three times. The EPR yield of the adduct in the microsomal preparation was estimated to be approximately 0.1 mg. The purified yield was similarly estimated to be 0.02 mg. The mass spectrum of the biologically obtainedproduct was obtained underthe same conditions as those of the ethylhydrazine-derived adduct.

DDC (2), the 4-methyl dihydropyridineanalogue, is essentially undetectable as a cytochrome P-450 destructive agent in vitro and gives a verylow yield of N-methylprotoporphyrin IX in vivo (12). In contrast, DDEP (3), the 4-ethyl analogue, gives a good yield of N-ethylprotoporphyrin IX in vivo and causes a readily measured in vitro loss of cytochrome P-450 (12). A limited examination of the relationship between the nature of the 4-substituent, the in vitro destruction of cytochrome P-450, and the in vivo formation of prosthetic heme adducts was therefore undertaken to identify the most appropriate substrate for mechanistic studies.Analogues of 3,5bis(ethoxycarbonyl)-2,6-dimethyl-l,4-dihydropyridinewith the following substituents at the 4-position were synthesized (l), ethyl (3), propyl (4), by conventionalprocedures:H isopropyl (5), isobutyl (6),benzyl (7),and phenyl (8).Nifedipine (9), the 4-nitrophenyl derivative in current use as an antianginal agent, wasincludedin the study, as were the aromatized pyridine derivatives 10 and 11 (see Structures). The dihydropyridine analogues can be separated, on the basis of their biological activities (TableI), into three groups: (a)compounds which do not cause detectable destructionof cytochrome andwhich do not resultin vivo in the accumulation of abnormal hepatic porphyrins; ( b ) compounds which cause in vitro enzyme loss but do not give isolable abnormal porphyrins in vivo; and ( c ) compounds which are active in both assays. The 4-unsubstitutedanalog (l),compounds with 10 a 4-aryl group(8,9),and the aromatic pyridine derivatives and 11 fall into the first class of agents. However, since a specific search for abnormalhepaticporphyrins wasonly carried out in thecase of nifedipine, the possibility exists that low yield abnormal others of the first class of agents may cause porphyrin formation. The observation that DDCgives a low yield of such porphyrins even though it appears inactive in the in vitro enzyme destruction assay (12) confirms the greater

Me

Me

Me

EtOzC&CO;Et Me

H R -

I 2 3 4 5

10 II

H Me Et Pr

6 7

I-Bu Me benzyl

8

phenyl 2-Nitrophenyl

9

H Et

1 (H) 2 (Me) 3 (Et)

4 (pr) 5 (isoPr)

6 (isoBu) 7 (benzyl)

(phenyl) (2-nitrophenyl)

10 11 12

Me

Et

12

STRUCTURES

Porphyrin In he- field desorpLoss of p-450c, cytopatic porchrome phyrin fortion mass mationh spectra molecular ims‘ %

2c1

-I

2+1

Yes Yes Yes No Yes No No -

41 f 1 44 f 2 43 c 1 33 1 36 1

*

ND ND 2 f l l f 2 35 c 2

-

-

604

618‘ 632

646

-

-

-

Loss after 30 min of incubation as described under “Materials and Methods.” No loss is observed in the absence of NADPH in any of

these incubations: ND, not detectable. Hepaticpigmentisolatedasdescribedunder “Materialsand Methods.” ‘The unprotonated molecular ion value for the isolated hepatic porphyrin is given. A monoprotonated molecular ion is presentin all of the spectra. -, the experiment was not done. e Data taken from Refs. 12 and 23. dynamicrange of the invivo assay. The secondclass of substrates includes the 4-isopropyl (5) and 4-benzyl (7)dihydropyridine derivatives. Both of these compounds effectively destroy cytochrome P-450 but do not give detectable abnormal porphyrins. The absence of such porphyrins indicates that they are not formed or,if formed, that they are not stable to, or detected by, our experimental procedures. The third class of analogues, those which cause bothenzyme destruction and abnormal porphyrin accumulation, includes the 4-ethyl (3), 4-propyl (4),and 4-isobutyl (6) derivatives. DDC could be considered as a member of this last class, although, as already noted, itsin vitro cytochrome P 450 destructive activity is essentially not detectable. The N-ethyl derivative of DDEP (12) is, if anything, only slightly less effective than DDEP in the destructionof cytochrome P-450 (Table I). The difference in activity between DDC and DDEP is thus retained when thenitrogen is substituted,since N-ethyl DDEP is highly active whereas N-ethyl DDCwas previously shown to have littlein vitro activity (12). The structures of the abnormal porphyrins obtained with DDC (23) and DDEP (12) have been unambiguously elucidated in earlier work. A detailed structural analysis of the porphyrins produced by the 4-propyl and 4-isobutyl dihydropyridine analogues is not essential in the present context and has not been carried out. Nevertheless, enough data have been obtained for the assignment of tentative structures to these porphyrins. The virtual identity of the electronic absorption spectra of the porphyrins engendered by 4 and 6, both in the metal-free and zinc-complexed forms, with the corresponding spectra of the porphyrins obtainedwith DDC and DDEP(12,23) establishes that they are N-alkylprotoporphyrin IX derivatives. The field desorption mass spectra of theporphyrinsobtained with the 4-propyl and 4-isobutyl substrates (Table I) strengthens this conclusion. The monoprotonated molecular ions of the abnormal porphyrins ( m / e 633 and 647, respectively) are those expected for structures constructed from the dimethylesterof protoporphyrin IX (the alkyl ester is expected from the workup) and the appropriate moiety. The porphyrin obtainedwith 4 is therefore probably

Etozcq C02Et

1-Pr

Substrate (4-substituent)

8 9

RESULTS

E t 0 2 C 8CO, Et

TABLE I Inactivation of cytochrome P-450 by substituted dihydropyridine derivatives

Oxidation

One-electron

11291

TABLE I1 the dimethyl ester of N-propylprotoporphyrin IX, and that Spin-trapping of a radical in the microsomal metabolism of DDEP obtainedwith 6 is thecorrespondingN-isobutylprotoporEPR phyrin IX derivative, although these structures can only be Spin trap (mM) Changes in standard incubation system" signal tentatively assigned in the absenceof confiimation by further height" structural data. The 4-ethyl (3), 4-propyl (4), and 4-isobutyl (6) analogues ND ' None are shown by the structural survey to be most suitable for ND DMPO (50) 1.7 mechanistic studies. The destruction of cytochrome P-450 by PBN (50) 4.6 POBN ( 5 ) these agents is linked to prosthetic heme alkylation by the 6.4 POBN (10) formation of isolable alkylated-heme derivatives (Table I). 9.3 POBN (20) The three analogues,however, havecomparableactivities 8.1 POBN (50) both as cytochrome P-450 destructive agents and as precursors POBN (20) +PBN (50 mM) 3.1 for in vivo heme-adduct formation. The fact that the structure ND Heat-denatured microsomes POBN (20) ND -DDEP (3) has been of the abnormal porphyrin obtained with DDEP POBN (20) ND -NADPH POBN (20) conclusively established, in the absence of a compelling dif5.0 -NADPH, +NADH (1 mM) POBN (20) ference in the activitiesof analogues 3,4, and 6, has led us to POBN (20) -NADPH, +NADH (1 mM), +CO 3.0 choose DDEP for the ensuing mechanistic studies. 11.5" +NADH (1 mM) POBN (20) Incubation of DDEP with hepatic microsomes from phe-DDEP, +N-ethyl DDEP (IO mM)' 7.1 POBN (20) nobarbital-pretreated rats in the presence of DMPO did not -DDEP, -NADPH, +N-ethyl DDEP ND POBN (20) give rise to a detectable EPR signal, but in the presence of (10 mrdP a weak six-linesignal PBN rather than DMPO produced Standard incubationsystem, described under"Materials and pattern (Table 11).PBN was consequently used in a number Methods," contains 1 mM NADPH, 10 mM DDEP. of early experiments.' The subsequent observation that POBN Intensity oflow field EPR signal after 30 min of incubation at functioned much more effectively as a spin trap in our micro- 37 "C. ' ND, not detectable. somal system, however, has resulted in primary use of this This value, obtained in a separate experiment, has been normalagent. A direct comparison of the EPRsignal heights obtained ized. inparallel incubations of DDEP with microsomes in the e EDTA was replaced by diethylenetriaminepentaacetic acid in presence of POBN (20 mM) orPBN (50 mM) (Table 11) these incubations. indicates that POBN is approximately five times more effecg = 2 0060 tive than PBN in trapping the DDEP-derived radical. The IO G f EPR signal height in such incubations increases as the POBN a concentration is raised from 5 to 20 mM, although the signal is, if anything, attenuated by a further rise in concentration (to 50 mM) (Table 11). A 20 mM concentration of POBN was therefore routinely employed. b No EPR signal is observed in the presence of this concentration of POBN if DDEP is omitted from the microsomal incubation mixture, if heat-denatured microsomes are used, or if the incubation is carried out without added NADPH (Table 11).NADPH can be replaced by NADH, although this 11). dampenstherate of radicaladductformation(Table Cytochrome P-450 destruction is also observed to occur at a reduced but definite rate when NADH is used instead of NADPH (Table IV). Carbonmonoxide inhibits the reactions supported by both NADPH (Fig. 1) and NADH (Table 11),a fact which suggests thatbothreactionsaremediated by cytochrome P-450. NADH is known to support the catalytic C turnover of cytochrome P-450 in intact microsomal mema substantially slower rate than that obtained branes albeit at with NADPH (24, 25). The validity of spin-trapping studies depends critically on unambiguous identification of the trapped radical. This has FIG. 1. EPR spectra of the POBN-radical adduct obtained in been achieved by the usual (if not necessarily unambiguous) microsomal incubations of DDEP. a , spectrum after incubation correlation of the EPR parameters of the adduct in two with boiled microsomes; b, spectrum after incubation with the comdifferent solvents with those of an authentic standard and, plete standard incubation system; c, spectrum obtained after incubamore definitively,by the isolation and mass spectrometric tion with the standard system in the presence of carbon monoxide. characterization of the radical adduct. An authentic sampleof The incubations, described under"Materials and Methods," conthe POBN-ethyl radical adductwas generated by CuC12-cat- tained 10 m~ DDEP, 20 mM POBN. EPR spectrometer conditions: alyzed decomposition of ethylhydrazhe in the presence of microwave power, IO milliwatts; modulation amplitude, 1.0 G; time constant, 0.5 s; scan range, 200 G; scan time, 16 min; gain, 1.25 x lo". POBN under the conditions used previously to trap the radical with PBN (22). The POBN adducts obtained chemically and ysis. Approximately 0.2 mg of pure POBN-ethyl radical adduct biologically were purified by multiple thin layer chromatogwas thus isolated from the 0.5 mg of the adduct producedby raphy. The thin layer fractions containing the radical species CuClz-catalyzed decomposition of ethylhydrazine. Because of were identified during thepurification sequence by EPR anal- the inherently more complex reaction mixture, and conse-

' Preliminary results using PBN were reported at the Fifth International Symposium on Microsomes and Drug Oxidations, July 2629, 1981 in Tokyo, Japan.

quently more difficult purification, only 0.02 mg of POSNethyl radical adduct was isolated from an estimated 0.1 mg accumulated in the microsomal incubation with DDEP. The

Oxidation

11292

One-electron

ethylhydrazine- and DDEP-derived adducts exhibit identical six-line EPR signals both in benzene (Fig. 2, a N = 14.43, U H = 2.50) and in aqueous buffer (see Fig. 1, U N = 15.78, a H = 2.73). The electron impact mass spectra of the chemically and biochemically obtained spin adducts arefurthermore identical if allowance is made for the presence of a trace of unreacted POBN in the biological sample (Fig. 3). The molecular ion peak at m / e 223 and the more intense peak at m / e 222 due to loss of a hydrogen are those expected for a POBN-ethyl radical adduct ( M y = 223). The peak at m / e 166 can be attributed toloss of the t-butyl moiety from the parent ion. A strong peak at m / e 138 due to loss of the t-butyl group from the parent ion at m / e 194 is observed with POBN itself (not shown). In sum, the EPR and mass spectrometric data unambiguously establish that the metabolically formed species is an ethyl radical. To more precisely define the role of cytochrome P-450 in ethyl radical formation and to explore the reasons for the differential effectiveness of DMPO, PBN, and POBN as microsomal spin traps, the effects of these spin traps on in vitro cytochrome P-450 destruction by DDEP, and on benzphetamine N-demethylation by the enzyme, wereinvestigated (Fig. 4). DMPO, which gives no EPR-detectable products in our system, only weaklyinhibits benzphetamine N-demethylation or,except at high concentrations, DDEP-mediated cytochrome P-450 destruction. In stark contrast, PBN strongly inhibits both cytochrome P-450 destruction and benzphetamine metabolism. POBN, the most effective spin trap in the microsomal system, inhibits benzphetamine metabolism, albeit less effectively than PBN, but does not inhibit the destruction of cytochrome P-450 by DDEP (Fig. 4). The inhibition of benzphetamine N-demethylation by PBN and POBN, but notby DMPO, can be explainedif the former but not the lattercompete effectivelywith benzphetamine for occupancy of the enzyme active site. The binding of the three spin traps to cytochrome P-450 in microsomes from phenobarbital-pretreated rats has therefore been estimated by difference spectroscopy (18, 19). In effect, DMPO produces a type I1 difference spectrum when it binds to unreduced cytochrome P-450 but the interaction is characterized by a large apparent dissociation constant (Ks= 6100 p ~ (Fig. ) 5). PBN gives a type I difference spectrum with an apparent dissocia, POBN gives a type I1 spectrum tion constant K, = 150 p ~and with Ks= 700 p~ (Fig. 5). PBN thus appears to be bound five times more tightly than POBN and 35 times more tightly than DMPO, although it is to be noted that this experiment does not establish that the three spin traps are bound by the 4: 2006 IO G $ ”

b

FIG. 2. Comparison of the EPR spectra (in benzene under Na) of the POBN-trapped radicals. Spectra were obtained from ( a ) CuClz-catalyzed decompositionof ethylhydrazine and( b )microsomal incubations of DDEP. EPR spectrometerconditions:microwave power, 10 milliwatts;modulation amplitude, 0.5 G; time constant, 0.3 s; scan range,100 G scan time, 8 min; gain,8 X lo4for a and 1.5 X lo5 for b.

B

A.

I

al

z

c

0

I66

222 I

130

150

170

190 210

230

m /e rn /e FIG. 3. Upper mass region of the 70-Evelectronimpact specSpectra were obtained ( a ) tra of the POBN-ethyl radical adduct. from microsomal incubations with DDEP and ( b ) from CuClz-catalyzed decomposition of ethylhydrazine.The spectrum ofthe biological product has been corrected by subtraction of peaks dueto a trace of POBN in the sample.

0

10

20 40 Concentratlon of Spin-Trap (mM)

50

FIG. 4. Inhibition of DDEP-mediated cytochrome P-450 destruction and benzphetamine N-demethylation by three radicalspintraps. The left axis refers to the percentdecreasein cytochrome P-450 destruction inthe presence of the given concentrations of DMPO (m), PBN (O),and POBN (A).The r z h t axis refers to the per cent decrease in benzphetamine N-demethylation in the and presenceof the given concentration of DMPO (O),PBN (O), POBN (A).The incubation conditions are given under“Materialsand Methods.” Maximum P-450 destruction in the absence of inhibitors was 39%of the enzyme presentin the microsomes.

same cytochrome P-450fraction in the microsomal membrane. For comparison, the apparent spectral dissociation constants (K,) for DDEP, atype I substrate(Fig. 5), and benzphetamine , The inhibition of (not shown) are 2 and 10 p ~ respectively. benzphetamine N-demethylation by the three spin traps thus correlates well with their relative affinities for microsomal cytochrome P-450: DMPO is weakly bound and weakly inhibits the reaction, PBN is tightly bound and strongly inhibits,

One-electron Oxidation A

040r b

4 dl--." -10

20

IO

I 00

30

-03

0 3

06

09

v

II / /

11293

stituted cytochrome P-450 system ( 2 6 ) ,although we have not specifically examined the effect of catalase on benzphetamine metabolism in our system. Hydrogen peroxide is a known product of uncoupled cytochrome P-450 turnover (27). The ability of hydrogen peroxide to support DDEP metabolism to an ethyl radical has been examined directly in microsomal incubations without NADPH but with added hydrogen peroxide ( 2 mM) and sodium azide (1 mM), thelatterasan inhibitor of endogenous catalase. Ethyl radical formationwas observed albeit at a slower rate, the height of the EPR spintrapped radical signal measured after 30 min being approxiTABLE I11 The contribution of reactive oxygen intermediates to carbon radical formationfrom DDEP

05

I

EPR sig-6 0

60

120

180

-30

30

60

90

nal height I, cm

Additions to incubation system"

mM"

r n t 4 - I

I/concentrotion

FIG. 5. Binding of DDEP and the spin traps to cytochrome P-450fromphenobarbital-pretreated rats. Double reciprocal plots of the peak-to-trough absorbance difference as a function of the reciprocal of the nominal substrate concentration are given for (a) DDEP, (b) DMPO, (e) POBN, and (d) PBN.

and POBN is both less tightly bound and inhibits N-dealkylation less effectively than PBN. A rationale for the relative inefficiency of PBN asa spin trap for DDEP-derived radicals is suggested by these results. If the cytochrome P-450 enzyme which catalyzes DDEP-metabolism is strongly inhibited by the concentration of PBN required for efficient interception of carbon radicals, only a low yield of spin-trapped radicals would be expected. To test this hypothesis, the EPR signal provided by microsomal metabolism of DDEP in the presence individually of either POBN or PBN has been compared with the signal obtained in the presence of both spin traps. If the spin traps function independently, the net signal intensity in the presence of both traps should, if anything, increase. If PBN actually inhibits radical formation, however, the EPR signal intensity should decrease despite the presence of signal POBN. The results (Table 11) show that the EPR intensity observed in the presence of POBN alone is decreased by approximately 70% when PBN is also present. The role of reactive oxygen species in DDEP oxidation and ethyl radical formation has been assessed (Table 111). The addition of superoxide dismutase to the microsomal incubation system does not significantly impair ethyl radical formation. The reaction is also not inhibitedby a high concentration of the hydroxy radical scavenger mannitol. The intensity of the POBN-ethyl radical EPR signal, however, is reduced if the incubation is carried out in the presence of catalase (Table 111).Control incubations with heat-denatured catalase confirm that thecatalytically active enzyme is required for inhibition. Addition of catalase to the incubation mixture at the end of the incubation period established that the EPR signal was not quenched by catalase. The inhibition of ethyl radical formation as a function of the catalase concentration (Fig. 6) shows that a maximum of approximately 50% of the reaction can be suppressed by catalase. Ethyl radical formation under conditions where the catalase effect is maximal (0.50 mg/ml of catalase) remains susceptible to inhibition by carbon monoxide (Fig. 6), as expected if cytochrome P-450 catalyzes the reaction. The involvement of cytochrome P-450 is also implied by the observation that the DDEP-mediated loss of cytochrome P-450 in a typical incubation is reduced from 41 k 1 to 29 f 2% in a parallel incubation containing 0.5 mg/ml of catalase (Table IV). Hydrogen peroxide has been reported to support benzphetamine N-demethylation by a purified, recon-

B of control signal

7.1

None (control) +Superoxide dismutase (0.1mg/ml) +Catalase (0.1 mg/ml) +Catalase (0.1 mg/ml), superoxide dismutase (0.1 mg/ml) +Catalase after the incubation period finished +Boiled catalase +Mannitol (60 mM) -NADPH, +H202 (2 mM), +NaN3 (1 mM) -NADPH. +H202 (10 mM). +NaN? (1 mM)

3.5 3.9

100 93 49 55

7.1 7.4 7.5 2.3 ND'

100 104 105 33 ND

6.6

a The standardincubation system, described under "Materials and Methods," contained NADPH (1.0 mM) and DDEP (10 mM). Low field EPR signal after 30 min incubation at 37 "C. ' ND, none detectable.

't 0

I

005

I 010

I

.r

I

I

015 0 25 Catalase (rng/ml )

050

FIG. 6. Inhibition of POBN-radical adduct formation by catalase. The EPR signal height after 30 min of incubation of DDEP with hepatic microsomes in the presence of POBN (20 mM) is shown for incubations carried out in the presence of increasing amounts of catalase. The corresponding signal height in an incubation with the highest catalase concentration but in the presence also of carbon monoxide is shown by the arrow. TABLE IV Destruction of cytochrome P-450by DDEP Incubation conditions"

Loss of cytochrome P-450 %

Normal 41 f 1 -NADPH, +NADH (134 mM) f2 +Catalase (0.5 mg/ml) 29 f 2 +Ascorbic acid (10 mM) 39 f 2 a Normal incubation conditions are given under "Materials and Methods."

I1294

One-electron Oxidation

mately one-third of that observed in the normal NADPHsupported reaction (Table 111).No EPR signal was observed, however, if the hydrogen peroxide concentration was raised to 10 mM, or if 2 mM hydrogen peroxide was used but the microsomes were Fist thermally denatured or omitted altogether (not shown). The generation of ethyl radicals from DDEP by hydrogen peroxide is thus mediated by a microsomal enzyme, presumably cytochromeP-450, rather thanby a direct reaction of the two substances. In this context, it is interesting that, althoughhydrogen peroxide can replace oxygen and NADPH in cytochromeP-450 catalysis (26,28),the enzyme is irreversibly inactivated by high hydrogen peroxide concentrations (26, 29). The observation of an EPR signal with 2 but not 10 mM hydrogen peroxide, clearly inconsistent with a direct reaction of hydrogen peroxide with DDEP, can be rationalized by the latter inactivationprocess. We have previously shown that the N-ethyl derivative of DDC gives rise in vivo to a low yield of N-methylprotoporphyrin IX. The N-ethyl derivative of DDEP (12) has now been synthesized and shown, on incubationwith hepatic microsomes and POBN, to give rise to the EPR signal of the spin-trapped ethyl radical (Table 11).The N-ethyl derivative appears to be approximately 75% as effective as DDEP as a precursor of the ethyl radical.

little basis for differentiating between radical and cation mechanisms. Information of direct relevance to the mechanism of prostheticheme alkylation and to the catalytic mechanism of cytochrome P-450 has been obtained with radical spin traps. The absenceof data bothon the interactionof spin traps with cytochrome P-450 and on their relative utility for the interception of radicals in microsomal systems, however, required an initial study of the spin traps themselves. This study has established that POBNis a much more effective spin trapfor carbon radicals in microsomes than DMPO or PBN. DMPO, previously reported not to inhibit microsomal drug metabolism (31),is only weakly bound by microsomal cytochrome P450 ( K s = 6100 IJM, Fig. 5). It furthermore is onlyapoor inhibitor of benzphetamine N-demethylation or DDEP-mediated cytochromeP-450 destruction (Fig. 4) and completely is ineffective as a trap for DDEP-derived radicals (Table 11). The probable explanation for these results is that DMPO, a hydrophilic compound, is only present in low concentrations in the lipophilic membrane compartment occupied by cytochrome P-450. In contrast, the highly lipophilic (32) PBN is strongly bound by cytochrome P-450 (Ks = 150 IJM, Fig. 5) and strongly inhibits both benzphetamine N-demethylation and DDEP-mediated enzyme destruction (Fig. 4). It is not particularly effective as a spin trap, however, because it inDISCUSSION hibits theenzyme responsible forradical formation (Table11). A rough correlation exists betweenthe biological activity of POBN, on the other hand, is less tightly boundby cytochrome the 4-substituted dihydropyridine analogues and the relative P-450 ( K s = 700 IJM, Fig. 5) and is much less effective as a stability of the radicalwhich would be obtained by homolytic direct inhibitor of cytochrome P-450 (Fig. 4) than PBN. Itis, loss of the 4-substituent. Thehomolytic bond energies for the in consequence, the most effective spin trap forlipophilic formation of the following radicals from the hydrocarbons,a radicals generated by the catalytic action of cytochrome Pmeasure of the stability of the alkyl or aryl radicals, are (in 450 (Table 11). kilocalories/mol) (30): phenyl (103), methyl (104), ethyl (98), Incubation of DDEP with hepatic microsomal cytochrome propyl (98), isopropyl (94), and benzyl (85). Thus, dihydroP-450 in the presence of POBN results in the spin-trapping of pyridine analogues with 4-substituents of relatively low radical ethyl radicals(Fig. 1). The identity of the POBN-trapped stability (phenyl, methyl) cause little or no cytochrome P-450 radical has been unambiguously established by isolation and destruction, those with substituents of intermediate stability characterization of the resulting adduct (Figs. 2 and 3 ) . The (ethyl, propyl, butyl) areeffective destructive agents andgive involvement of cytochrome P-450 in ethyl radical formation good yields of prosthetic heme adducts, and analogues with is indicated by several lines of evidence. Radical formation substituents of high radical stability (isopropyl, benzyl) are requires NADPH, is inhibited by carbon monoxide, does not effective destructive agents but do not give detectable pros- occur after thermal denaturation of the microsomes, and is thetic heme adducts (Table I). The results with analogues inhibited in proportion to the ability of the three spin traps to bearing substituents of low or intermediate radical stability inhibit benzphetamine N-demethylation (Table 11, Figs. 1 and are consistent with the observation that cytochrome P-450 4). Particularly telling in this regard is the observation that destruction by the methyl (12), ethyl (12), and propyl (Table PBN inhibits the trappingof ethyl radicals by POBN (Table I) analogues involves transfer of the substituent to thepros- II), a result which requires interference with the catalytic thetic hemeof the enzyme, a reaction which involves cleavage mechanism responsible for ethyl radical formation. Furtherof the bond between the substituent and the dihydropyridine more, substitution of NADH for NADPH in the incubation ring. A simple correlation between substituent bond strength mixtures reduces, but does not terminate, both ethyl radical (radical stability), enzyme destruction, and prosthetic heme formation and DDEP-mediated cytochrome P-450 destrucalkylation cannot be drawn, however, because the isopropyl tion (Tables I1 and IV). Catalasealso inhibits cytochrome Pand benzyl derivatives destroy the enzyme but do not appar- 450 destruction and ethyl radical formation toa comparable isolable) by prosthetic degree (Tables 111 and IV). The finding that low concentraently do so (unless the adducts are not heme alkylation. This mechanistic discontinuityis less easily tions of hydrogen peroxide can support cytochrome P-450 rationalized if the 4-substituent is directly transferred from destruction and ethyl radical formation is consistent with the the substrate to the hemenitrogen than it is by mechanisms ability of hydrogen peroxide to support benzphetamine N in which loss of the substituent and heme alkylation occur as demethylation by purified, reconstituted cytochrome P-450 (26). These results, in conjunction with the demonstration discrete steps. The direct transfer of the methyl and benzyl groups, for example,should, if anything, be easier than trans- that the loss of cytochrome P-450 reflects alkylation of the fer of the ethyl and isobutyl groups because concerted dis- prosthetic hemegroup (12), provide astrong web of correlative placement reactions are highly sensitive to steric effects but evidence for the direct catalytic participationof cytochrome relatively insensitive to electronic effects. In a stepwise trans- P-450 in radical formation. In this context, it is important to of cytofer mechanism, however, the higher stability and lower reac- noie that POBN does not inhibit the destruction tivity of a secondary (isopropyl) or conjugated (benzyl) radical chrome P-450 by DDEP (Fig. 4). This is readily understood if the radicalinvolved in heme alkylation is formed within the or cation could'resultin protein rather than hemealkylation. activesite, where addition to POBN is not a competitive The study of substrate analogues thus suggests that heme process, but isdifficult to reconcile with amechanism in which provides alkylation is mediatedby a discrete intermediate but

Oxidation

One-electron

an externally generated radical must diffuse into the active site. The generation of ethyl radicals during cytochromeP-450catalyzed turnover of DDEP implies that a one-electron oxidative step intervenesin the metabolic sequence. The nature of this oxidative step is clarified by the finding that N-ethyl DDEP destroys cytochrome P-450 and gives rise to ethyl radicals as effectively as DDEP itself (Tables I and 11).We have previously shown that N-ethyl DDC causes prosthetic heme alkylation (12). Of the three mechanisms outlined in Scheme 1, only that of path a is consistent with the data obtained in this investigation. Path c is excluded because the presence of the N-ethyl group in the substrate prevents formation of the N-hydroxylderivative. If the nitrogen were oxidized despite the presence of the N-ethyl group, the product would be an N-oxide, a structure which could not give rise to an appropriate precursor for the ethyl radical. Path b, of course, is excluded by the activityof N-ethyl DDEP but even more so by the observation of ethyl radical^.^ The present results thus provide convincing evidence that, at least in the case of DDEP and related substrates, nitrogen oxidation by cytochrome P-450 proceeds in one-electron steps. This may reflect the particular situation in DDEP, where the radical cation obtained by one-electron abstraction canbe stabilized by extensive delocalization, but may in fact also point to a more general (if not necessarilyuniversal) mechanism for nitrogen oxidation bythis class of enzymes. In its more limited context,thepresent evidencefor one-electronabstraction from nitrogen in systems where theresulting radical is stabilized by extensive delocalization has a direct bearing on the mechanism of cytochrome P-450 catalyzed oxidation of substances like acetaminophen (33,34).A direct electron abstraction from the conjugated nitrogen in acetaminophen is suggested by these results rather than thehydrogen abstraction or other mechanisms currently in favor. In a more general context, the observation of one-electron nitrogen oxidation has important implications for N-dealkylation and other nitrogen-based metabolic transformation^.^ The use of a substrate constructed so that elimination of a reporter fragment (the ethyl radical) iscompetitive with the normal fate of the intermediatein the catalytic sequence has provided a novel probe for the mechanism of cytochrome P450. The present successfulutilization of such a substrate suggests that similar probes may be of value not only in further studiesof cytochrome P-450 but in the study of other redox enzymes. Acknowledgments-We gratefully acknowledge acquisition of the mass spectrometric data by Dr. Kent L. Kunze and generous access to the ESR spectrometers provided by Drs. Alex Quintanilha and Lester Packer.

Evidence has been obtained for the formation of 3,sbis(ethoxycarbonyi)-2,6-dimethylpyridine (the 4-dealkylated aromatic metabolite) by gas and high pressure liquid chromatographic analysis of incubation extracts. The ready autoxidationof the dihydropyridine precursor in the gas chromatograph inlet, however, has so f a r prevented product quantitation. Evidence for one-electron oxidation of a hydroxylamine has recently been reported (35).

11295 REFERENCES

1. Groves, J. T., McClusky, G. A., White, R. E., and Coon, M. J. (1978) Biochem. Biophys. Res. Commun. 81, 154-160 2. Groves, J . T., Akinbote, 0. F., and Avaria, G. E. (1980) in Microsomes, Drug Oxidations, and Chemical Carcinogenesis (Coon, M. J., Conney, A. H., Estabrook, R.W., Gelboin, H. V., Gillette, J. R., and OBrien,P. J.,eds) pp. 253-261, Academic Press, New York 3. Ortiz de Montellano, P. R., Yost, G. S., Mico, B. A., Dinizo, S. E., Correia, M. A,, and Kambara, H. (1979) Arch. Biochem. Biophys. 197,524-533 4. Ortiz de Montellano, P. R., and Mico, B. A. (1980) Mol. Pharmacol. 18, 128-135 5. Ortiz de Montellano, P.R., and Mico, B. A. (1981) Arch. Biochem. Biophys. 206.43-50 6. Ortiz de Montellano, P. R., Beilan, H. S., Kunze, K. L., and Mico, B. A. (1981) J.Biol. Chem. 256,4395-4399 7. Ortiz de Montellano, P. R., Kunze, K. L., Beilan, H. S., and Wheeler, C. (1982) Biochemistry 21, 1331-1339 8. Ashley, P. L., and Griffin, B. W.(1981) Mol. Pharmacol. 19, 146152 9. Griffin, B. W., Marth, C., Yasukochi, Y., and Masters, B. S. S. (1980) Arch. Biochem. Biophys. 205,543-553 10. Griffin, B. W., and Ting, P.L. (1978) Biochemistry 17,2206-2211 11. Watanabe, Y., Iyanagi, T., and Oae, S. (1980) Tetrahedron Lett. 3685-3688 12. Ortiz de Montellano, P. R., Beilan, H. S., and Kunze, K. L. (1981) J.Biol. Chem. 256,6708-6713 13. Buettner, G. R., and Oberley, L. W. (1978) Biochem. Biophys. Res. Commun. 83,69-74 14. Loev, B., Goodman, M. M., Snader,K.M.,Tedeschi, R., and Macko, E. (1974) J.Med. Chem. 17, 956-965 15. Kamal, A., and Qureshi, A. A. (1963) Pak. J.Sci. Res. 15, 35-42 16. Low, B., and Snader, K. M. (1965) J. Org. Chem. 30, 1914-1916 17. McCord, J. M., and Fridovich, I. (1969) J.Biol. Chem. 244,60496055 18. Estabrook, R. W., Peterson, J., Baron, J., and Hildebrandt, A. (1972) Methods Pharrnacol. 2, 303-350 19. Jefcoate, C. R. (1978) Methods Enzymol. 52, 258-279 20. Correia, M. A,, and Mannering, G . J. (1973) Mol. Pharrnacol. 9, 470-485 21. Nash, T. (1953) Biochem. J.55,416-421 22. Augusto, O., Ortiz de Montellano,P. R., andQuintanilha, A. (1981) Biochem. Biophys. Res. Commun. 101, 1324-1330 23. Ortiz de Montellano, P. R., Beilan, H. S., and Kunze, K. L. (1981) Proc. Natl. Acad. Sci. U. S. A . 78, 1490-1494 25. West, S. B., Levin, W., Ryan, D., Vore, M., and Lu, A. Y. H. (1974) Biochem. Biophys. Res. Commun. 58, 516-522 25. Correia, M. A,, and Mannering, G. J. (1972) Mol. Pharrnacol. 9, 455-469 26. Nordblom, G. D., White, R. E., and Coon, M. J. (1976) Arch. Biochem. Biophys. 175, 524-533 27. Hildebrandt, A. G., and Roots, I. (1975) Arch. Biochem. Biophys. 171,385-397 28. Hrycay, E. G., Gustafsson, J. A., Ingelman-Sundberg, M., and Ernster, L. (1975) Biochem. Biophys. Res. Commun. 66, 2092 16 29. O’Brien, P. J. (1978) Pharmacol. Ther. Part A . Chemother. Toxicol. Metab. Inhibitors 2, 517-536 30. Benson, S. W. (1965) J. Chem. Ed. 42, 502-516 31. Cederbaum, A. I., and Cohen, G. (1980) Arch. Biochem. Biophys. 204,397-403 32. Saprin, A. N., and Piette, L. H. (1977) Arch. Biochem. Biophys. 180,480-492 33. Hinson, J. A., Pohl, L. R., Monks, T. J., and Gillette, J. R. (1981) Life Sci. 29, 107-116 3 4 . Nelson, S. R., Dahlin, D. C., Rauckman, E. J., and Rosen, G. M. (1981) Mol. Pharmacol. 20, 195-199 35. Rauckman, E. J., Rosen, G. M., and Cavagnaro, J. (1982) Mol. Pharrnacol. 21,458-463