;Mechaiiisms
of cytochrome
xenobiotic PAUL
metabolism
-
F. HOLLENBERO,
Department
of Pharmacology,
....
Wayne State Universitr
The cytochrome P450 enzyme systems catalyze the metabolism of a wide variety of naturally occurring and foreign compounds by reactions requiring NADPH and 02. Cytochrome P450 also catalyzes peroxide-dependent hydroxylation of substrates in the absence of NADPH and 02. Peroxidases such as chloroperoxidase and horseradish peroxidase catalyze peroxidedependent reactions similar to those catalyzed by cytochrome P450. The kinetic and chemical mechanisms of the NADPH and 0a-supported dealkylation reactions catalyzed by P450 have been investigated and compared with those catalyzed by P450 and peroxidases when the reactions are supported by peroxides. Detailed kinetic studies demonstrated that chioroperoxidaseand horseradish peroxidase-catalyzed N-demethylations proceed by a Ping Pong Bi Bi mechanism whereas P450-catalyzed 0-dealkylations proceed by sequential mechanisms. Intramolecular isotope effect studies demonstrated that Ndemethylations catalyzed by P450s and peroxidases proceed by different mechanisms. Most hemeproteins investigated catalyzed these reactions via abstraction of an acarbon hydrogen whereas reactions catalyzed by P-450 and chloroperoxidase proceeded via an initial oneelectron oxidation followed by a-carbon deprotonation. ‘80-Labeling studies of the metabolism of NMC also demonstrated differences between the peroxidases and P450s. Because the hemeprotein prosthetic groups of P450, chloroperoxidase, and horseradish peroxidase are identical, the differences in the catalytic mechanisms result from differences in the environments provided by the proteins for the heme active site. It is suggested that the axial heme-iron thiolate moiety in P450 and chloroperoxidase may play a critical role in determining the mechanism of N-demethylation reactions catalyzed by these proteins. Hollenberg, P. F. Mechanisms of cytochrome P450 and peroxidase-catalyzed xenobiotic mechanism. FASEBJ. 6: 686-694; 1992. ABSTRACT
Ky Wordt: cytochrome P450 function oxidase . hydroxylation
.
and per xidaseCatIjzed
P450
peroxidse#{149}monooxygenase
mired-
ENZYME SYSTEMS IN THE ENDOPLASMIC reticulum of the liver and most other mammalian tissues catalyze the metabolism of a wide variety of endogenous and exogenous compounds including drugs, steroids, prostaglandins, chemical carcinogens, and other xenobiotics (1, 2). These enzymes, known as mixed-function oxidases, catalyze incorporation of one atom of molecular oxygen into the substrate to give product while the other oxygen atom is reduced by two electrons to give water. Although these enzymes play an important role in the detoxication of many drugs, chemical carcinogens, and other toxic agents, they are also responsible for catalyzing the metabolic activation of some substrates to highly reactive free-radical, alkylating, or arylating intermediates, which
School
-
-.
....
of Medicine, Detroit, Michigan 48J,
USA
then react with critical cellular macromolecules to initiate toxic and carcinogenic events (3). The critical role these enzymes play in the metabolic activation and detoxication of a wide variety of carcinogens and other toxic agents makes them of particular importance in light of human exposure to these compounds in the environment. The microsomal P450-dependent mixed-function oxidase enzyme
systems
are
composed
of NADPH-P450
reductase,
a phospholipid, and P4502 and exhibit an absolute requirement for NADPH and 02 for catalytic activity (1-3). Cytochrome P450 plays a critical role as it binds substrate, activates oxygen, and then catalyzes the insertion of the activated oxygen into the substrate (1-3). Therefore, the P450 determines the substrate specificity of the overall system and the structure (or structures) of the product (or products) formed during substrate metabolism.
CYTOCHROME
P450-CATALYZED
REACTIONS
The P450-dependent mixed-function oxidases catalyze a wide variety of reactions including epoxidations, N-dealkylations, 0-dealkylations, S-oxidations, and hydroxylations of aliphatic and aromatic substrates. The diversity of the reactions catalyzed can be understood by considering that the initial reaction in all cases involves insertion of a hydroxyl group into the substrate to form an hydroxylated intermediate which can then, depending on the nature of the substrate and the stability of the intermediate, undergo dealkylation, deamination, etc. (4). This hypothesis has led to the suggestion that a critical step in P450-catalyzed reactions is the formation of an hydroxyl donor, which can then transfer the hydroxyl group to the appropriate position on the substrate. The P450 family includes many different isozymes and more than 150 isoforms have been characterized (5). The purified isozymes exhibit different molecular weights, amino acid compositions, peptide maps, amino acid sequences, spectral, immunochemical, and catalytic properties (5, 6). They also exhibit different, although sometimes overlapping, substrate specificities, and the overall reactions catalyzed by P450s proceed with the stoichiometry shown: RH
+
02
+
NADPH + W H2O + NADP
-
ROH
+
Eq.
(1)
‘To whom correspondence should be addressed, at: Department of Pharmacology, Wayne State University School of Medicine, 540 E. Canfield, Detroit, MI 48201, USA. 2Abbreviations: P450, cytochrome P450, DMA, N,N-dimethylaniline;
DMA-N-oxide,
N,N-dimethylaniline-N-oxide;
roper.oxidase; HRP, horseradish peroxidase; bazole; NHMC, N-hydroxymethylcarbazole; bazole; and EtOOH, ethyl hydroperoxide.
CPO,
NMC, NFC,
chlo-
N-methylcarN-formylcar-
686 0892-6638/92/0006-686/$01 .50. © FASEB ww.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 15, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum
where RH represents the substrate. The catalytic mechanisms by which the P450s activate molecular oxygen appear to be very similar. The catalytic cycle for P450-catalyzed reactions is thought to consist of at least six discrete reactions (4, 5, 7). The first step involves binding of the substrate to the ferric form of the enzyme. The second step involves transfer to one electron from the NADPH-P450 reductase to the iron of the fernic P450 enzyme to give a ferrous enzymesubstrate complex. The reduced P450-substrate complex then binds 02 to form a ferrous enzyme-Os-substrate ternary complex with the 02 bound to the iron. The addition of a second electron to this ternary complex by the reductase results in the formation of an iron peroxo species best represented as Fe3O22. The next step involves cleavage of the oxygen-oxygen bond. One of the oxygens is released with the uptake of two protons at some stage, resulting in the formation of water. The retained oxygen remains associated with the heme iron as an activated oxygen, perhaps as an iron-oxene species. Considering the redox properties of the iron, oxygen, and sulfur atoms, numerous resonance forms could be suggested for the activated oxygen intermediate. The activated oxygen has been postulated to be a variety of forms of oxygen including superoxide, oxene, and peroxides such as penoxyacid or peroxyimidic acid. The activated oxygen atom associated with the iron is then inserted into the substrate, resulting in a two-electron oxidation of the substrate to the alcohol. The insertion of the oxygen into the substrate is thought to involve hydrogen abstraction from the substrate followed by radical recombination of the resulting transient carbon and hydroxyl radicals to form product (8). The product is then released, regenerating the native ferric P450 that is available to begin another catalytic cycle. The demonstration that P450 can use peroxides, hydroperoxides, and peracids to support N- and O-dealkylations, aliphatic hydroxylations, and olefin epoxidations suggested relationships with peroxidases and provided new approaches for studying its mechanism of action (4, 9, 10). The addition of one electron, a molecule of oxygen, a second electron and two protons during the catalytic cycle is formally equivalent to the addition of hydrogen peroxide (H2O2). The pathway involving substitution of a peroxy compound for NADPH and 02 is referred to as the peroxide shunt. P450-catalyzed oxygenations by 02 and by peroxy compounds are believed to have common mechanistic features. Therefore, the investigation of the mechanisms by which these oxidants support P450-catalyzed reactions is of interest because: 1) this system may be studied in the absence of NADPH, reductase, and 02, thereby simplifying it for study; 2) there are many similarities to reactions catalyzed by peroxidases, which have been studied in great detail; and 3) comparison of the reactions supported by NADPH and molecular oxygen with those supported by the organic oxidants may help us to understand both systems. The isolation of hydroperoxides of tetralin, fluorene, steroids, 9,10-dimethyl-1,2 -benzanthracene, and 9-methylfluorene from microsomal reaction mixtures has been cited as evidence for the role of hydroperoxides in biological hydroxylations (11).
REACTIONS
CATALYZED
Peroxidases
The function of the peroxidase may be to oxidize a particular substrate or to reduce a hydroperoxide. Most peroxidases have heme as their prosthetic group. Many peroxidasecatalyzed reactions proceed as shown in Eqs. 3-5, Peroxidase
Compound
Compound
+
I
II
ROOH
AH2
+
+
AH2
AH
Compound
-
Compound
-
Peroxidase
-
-
Nonradical
I
II
+
+
+
ROH Eq.
(3)
Eq.
(4)
Eq.
(5)
AH.
AH
products Eq. (6)
where ROOH is the hydroperoxide, AH2 is the reducing substrate, and AH is a free radical (12). This process predicts that free radicals would be the products of many organic molecules. The fate of these radicals is characteristic of each substrate. However, some substrates appear to reduce compound I directly to the native enzyme without the formation of compound II (13, 14). For some peroxidases such as HRP, compound I is a relatively stable, spectrally distinct enzyme species two oxidizing equivalents above the native ferric enzyme. The formation of compound I proceeds with the concomitant release of the alcohol, ROH, derived from the hydroperoxide. Based on studies of the CPO-catalyzed evolution of oxygen from 18O-labeled H2O2 and peracids, it has been suggested that compound I contained one, and only one, atom of oxygen from the peroxy substrate (15), and that it might exist as any of the configurations shown in Fig. 1. These structures are electronically equivalent and have the following features: 1) they are two oxidizing equivalents above the native ferric (+3) enzyme, and 2) they contain only one oxygen liganded to the heme iron. Several of these structures could serve as good hydroxylating agents.
OH
OH
OH
x
I
BY PEROXIDASES
are
widely distributed in nature. As shown in the oxidation of inorganic and organic substrates with hydrogen peroxide, alkyl hydropenoxides, or acyl hydroperoxides as the ultimate electron acceptor.
Eq. 2, they catalyze
ROOH
+
AH2
-
ROH
+
A
+
H20
Eq.
(2)
x Figure pound
1. Scheme I.
showing
possible
structures
for peroxidase
corn-
MECHANISMS OF HEMEPROTEIN CATALYSIS ww.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 15, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum
CPO is a heme protein that has been isolated from the mold Caldariomyces furnago and has been characterized in detail by Hewson and Hager (15). CPO is unique among peroxidases in the diversity of reactions it catalyzes (16). In the presence of halide anions (C1, Br, or 1, but not F; represented by X in Eq. 7), CPO catalyzes the formation of a carbon-halogen bond with a variety of nucleophiles (AH in Eq. 7). Like other protoheme peroxidases, CPO catalyzes the peroxide-supported oxidation of a wide variety of substrates (BH2 in Eq. 8). Like catalase, CPO also catalyzes the decomposition of H2O2 to give 0 (Eq. 9). However, unlike catalase, CPO catalyzes the decomposition of alkyl hydroperoxides and peracids to give 02. AH
+
BH2
X
+
2 H2O2
+
H2O2
H2O2 -
-
2H20
+
B +
H
+
AX
-
+
2H20
2H2O
02
Eq.
(7)
Eq.
(8)
Eq.
(9)
Another property of CPO is the striking similarity of its heme active site environment to that of bacterial and mammalian P450s (17-19). Similar to P450s, the reduced CPO forms a CO complex having the Soret maximum at an unusually long wavelength, Xm = 443 nm (17). In addition, Mossbauer spectroscopy, EPR spectroscopy, resonance Raman spectroscopy, magnetic circular dichroism spectroscopy, and extended X ray absorption fine structure studies suggest a strong similarity between the heme active site environments of P450 and CPO (18-20). Extensive spectral studies of mammalian and bacterial P450s and X ray crystallographic studies of bacterial P450 provide conclusive evidence that the fifth axial ligand to the heme iron is a thiolate provided by a cysteine residue of the protein (18, 19). Blanke and Hager (21) have demonstrated that CPO also has a thiolate axial ligand. The striking similarities in the physiochemical properties of these two hemeproteins and the potential for compound Ito serve as an hydroxylating agent prompted us to investigate the ability of CPO to catalyze reactions characteristic of P450 (20).
PEROXIDASE-CATALYZED REACTIONS
N-DEMETHYLATION
CPO catalyzes the peroxide-dependent dealkylation of DMA as well as other N-methyl arylamines in a manner similar to that described for the peroxide-supported demethylation of DMA and benzphetamine by purified rabbit liver microsomal P450 (20, 22). The reaction for both the CPOand P450-catalyzed dealkylations can be written as shown: ROOH
+ R R’-N-CH3 R-OH R R’NH + HCHO
+
Eq.
(10)
where ROOH is the peroxide and R’R”N-CH3 is the N-methylarylamine substrate (either secondary or tertiary). The stoichiometry for formation of N-methylaniline and formaldehyde from DMA is 1:1. The turnover number for the C P0-catalyzed dealkylation of DMA supported by EtOOH (1476) was much greater than that observed for P450-catalyzed reactions in the reconstituted system, which range up to 20-30 (20, 22). CPO exhibited normal Michaelis-Menten saturation kinetics with respect to both EtOOH and DMA. A variety of
different peroxides, hydroperoxides and peracids were able to support the demethylation reaction. The initial velocity of the demethylation reaction was dependent on the identity of the oxidant. EtOOH was the most effective. CPO was able to use various arylamine substrates and the turnover numbers were also dependent on the identity of the substrate. Although P450 inhibitors such as azide and n-propyl gallate gave marked inhibition, other P450 inhibitors such as SKF-525A, metyrapone, and piperonyl butoxide did not inhibit the reaction. Tiron and DL-epinephrine, trapping agents for the superoxide anion, markedly inhibited the demethylation reactions whereas superoxide dismutase had no significant effect. Spin trapping agents such as a-phenyl-tbutylnitrone or 5,5-dimethylpyrroline-N-oxide, had no significant inhibitory effect on the demethylations, suggesting that free radical formation does not play a major role in demethylation. Diphenylfuran and DL-histidine, which react with singlet oxygen, did not inhibit the reaction. These studies of the demethylation reactions catalyzed by CPO had been initiated to investigate the hypothesis that the structural similarities between CPO and P450 might be reflected in the ability of CPO to function catalytically like a P450. Because CPO was able to function like a P450 for N-demethylation reactions, the ability of hemeproteins that did not provide P450-like active site environments for the heme to catalyze N-dealkylation reactions was investigated. Two preparations of HRP (isozyme A and isozymes B-C) exhibited catalytic activities for DMA demethylation significantly greater than those for the liver microsomal P450-dependent demethylations supported by NADPH and molecular oxygen or cumene hydroperoxide (4, 22-24). Lactoperoxidase also exhibited catalytic activity comparable to that of the most active P450 in the demethylation reaction (24). HRP catalyzes the peroxide-supported oxidation of a variety of different organic substrates (Eq. 2). As with CPO, the first step in HRP-catalyzed reactions involves reaction of the ferric (Fe3) form of the enzyme with the peroxide to form compound I, The HRP-catalyzed N-demethylation of DMA supported by EtOOH exhibited a stoichiomety of hydroperoxide consumption, formaldehyde formation, and N-methylaniline formation of 1:1:1 (24). The HRP-catalyzed N-demethylations were supported by a variety of oxidants including hydroperoxides, peroxides, and peracids, and turnover numbers were highly dependent on the identity of the oxidant. Unlike CPO-catalyzed demethylations (20), sodium chlorite was very efficient at supporting the HRP-catalyzed reaction. Chlorite is utilized by HRP to catalyze both chlorination and peroxidation reactions (25). HPLC analysis of the chlorite-supported reaction mixture suggested that HRP also catalyzed chlorite-dependent chlorination of DMA (24). When conditions were optimized, the turnover number for DMA demethylation by the B-C fraction of HRP was 7061 for the hydrogen-peroxide-supported reaction. This rate is significantly greater than for CPO under optimal conditions. HRP catalyzed the demethylation of a variety of secondary and tertiary N-methylarylamines. The HRP-catalyzed demethylation reactions exhibited normal Michaelis-Menten saturation kinetics with respect to DMA (Km = 0.34 mM), and H202 (Km = 0.016 mM), as well as EtOOH (Km = 0.020 mM). If an enzyme-generated activated oxygen species such as the superoxide anion, singlet oxygen, or the hydroxyl radical were a free intermediate in the HRP-catalyzed demethylations, reagents that react with these activated oxygen species should markedly inhibit the demethylations. Although Tiron, a reagent that reacts with the superoxide anion, markedly in-
688 Vol. 6 January 1992 The FASEB Journal HOLLEN BERG ww.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 15, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum
hibited demethylation, superoxide dismutase did not inhibit the HRP-catalyzed demethylation. Methionine, diphenylfuran, azide, and histidine are all reported to be efficient singlet oxygen-quenching agents. The inability of these compounds to inhibit demethylation suggests that singlet oxygen is not an intermediate in this reaction (24). Reagents used as scavengers for the hydroxyl radical include thiourea, mannitol, benzoate, formate, ethanol, /3-methylthiopropionaldehyde, dimethyl sulfoxide, and ascorbate. Ascorbate was the only hydroxyl radical scavenger that significantly inhibited the HRP-catalyzed demethylations. Because ascorbate is a substrate for HRP, it is probably inhibiting demethylations by competing with DMA for the enzyme active site rather than by trapping a hydroxyl radical intermediate. If the HRP-catalyzed demethylations proceeded via the formation of a free radical intermediate, then spin trapping reagents such as a-phenyl-t-butylnitrone, nitrosobenzene, nitromethane, tri-t-butylphenol, 2-methyl-2-nitrosopropane, or 5,5 dimethylpyrroline-N-oxide should markedly inhibit the demethylation. However, even high concentrations (10-75 mM) of these agents caused relatively small inhibitions of the demethylation reactions, suggesting that formation of free radicals, which could then dissociate from the enzyme, does not play an important role in demethylation. Thus, these results suggest that substrate free radicals are not formed as free intermediates in the demethylation reaction; however, the results do not preclude the involvement of enzyme-bound radical species as intermediates in the reaction. The kinetic mechanism for the C P0-catalyzed dealkylation of DMA supported by EtOOH has been determined by using initial rate and inhibition studies (26). When the concentration of DMA was varied systematically in the presence of several different fixed concentrations of EtOOH, double reciprocal plots of the initial rate data yielded a series of parallel lines. When the data were plotted as a function of the concentration of the EtOOH, parallel lines were also observed. Although these results are consistent with a PingPong mechanism, they do not unequivocally rule out a sequential mechanism. To distinguish between sequential and Ping-Pong mechanisms, the concentrations of both substrates were varied simultaneously while keeping them in a constant ratio. Double reciprocal plots of the initial rate data were linear, consistent with the initial assignment of a Pin gPong mechanism for the N-demethylation reaction. For reactions that proceed via a Ping-Pong mechanism, competitive substrate inhibition is normally observed at higher concentrations of the two substrates due to binding of the substrate to the wrong form of the enzyme to yield an abortive binary complex. Competitive substrate inhibition was observed when the concentration of EtOOH was varied in the presence of several inhibitory concentrations of DMA. Competitive substrate inhibition was also observed when the concentration of DMA was varied in the presence of several inhibitory concentrations of EtOOH. 2,5-Dimethylfuran, a potent inhibitor of the CPO-catalyzed demethylation, was a competitive inhibitor with respect to DMA and uncompetitive with respect to EtOOH. These results are consistent with dead-end inhibition in a Ping-Pong system. Thus, the combined initial velocity and inhibition results are consistent with a Ping-Pong Bi Bi mechanism as the minimal kinetic model for the CPO-catalyzed N-demethylations. In the model shown in Fig. 2 the CPO (E) reacts with the EtOOH (A) to form compound 1(F), the oxidized enzyme intermediate, with the concomitant release of ethanol (P). The DMA (B) then binds to compound I and is oxidized, presumably
to give a carbinolamine (Q). This results in the regeneration of the native CPO and the presumed carbinolamine intermediate, which is unstable and spontaneously decomposes to give N-methylaniline and formaldehyde. Detailed initial rate and inhibition studies on the HRP-catalyzed N-demethylation of DMA indicate that this reaction also proceeds by a Ping-Pong Bi Bi mechanism (24). Although these results establish the kinetic mechanism for peroxidase-catalyzed demethylation reactions, they provide no information regarding the nature of the chemical events involved in DMA demethylation.
ROLE OF AN N-OXIDE DEMETHYLATION
INTERMEDIATE
IN
One mechanism suggested for the oxidative demethylation of DMA involves initial oxygenation of the amine nitrogen to form DMA-N-oxide as shown in Fig. 3. This could then rearrange via an enzyme-catalyzed reaction to give the carbinolamine which, for a substrate such as DMA, is relatively unstable in aqueous solution and readily decomposes to give formaldehyde and the amine product. An alternative route shown in Fig. 3 involves initial oxygenation of the a-carbon to give the carbinolamine directly, which readily decomposes to give the products indicated. The N-oxidation of DMA occurs concomitantly with N-demethylation in liver microsomal preparations. Because the N-oxide is metabolized by liver microsomes to form N-methylaniline and formaldehyde, the formation of a tertiary amine N-oxide as an intermediate in N-demethylation has been suggested (27). Although studies from several laboratories support the N-oxide as an intermediate in N-demethylations, studies from other laboratories suggest that the N-oxide is not an intermediate in demethylations. All these studies were performed using microsomal preparations. Because the metabolism of DMA to give formaldehyde and N-methylaniline can proceed via two different pathways (Fig. 3) involving different microsomal monooxygenase enzyme systems, the interpretation of these experiments was complicated by the use of microsomal preparations containing different enzyme systems. In addition, although microsomal metabolism of DMA results in N-oxide formation, it has not been conclusively demonstrated that the N-oxide is an intermediate in the reaction rather than a by-product of metabolism. Therefore, the metabolism of DMA to the N-oxide by four different purified isozymes of P450 (23) as well as by CPO (20) and HRP (24) was investigated to determine the possible role of the N-oxide as an intermediate in N-demethylation reactions. All four isozymes of P450 (rabbit liver forms 1A23 and 2B4 and rat liver 1A1 and 2B1) investigated in the reconstituted system did not catalyze the metabolism of DMA to give measurable amounts of the N-oxide. In addition, all four isozymes of P450 exhibited significantly greater turnover numbers (two to sevenfold) for the demethylation of DMA than for the N-oxide. If the N-oxide were an intermediate in the demethylation reaction catalyzed by a P450, then the N-oxide should be metabolized to give formalde-
3The nomenclature used here is that based on structural homology (5). P450 1A1 is the 3-methylcholanthrene-inducible form from rat liver microsomes, P450 1A2 is the 3-methylcholanthreneinducible form from rabbits, P450 2B1 is the major phenobarbitalinducible form from rats, and P450 2B4 is the major phenobarbitalinducible form from rabbits.
MECHANISMS OF HEMEPROTEIN CATALYSIS 689 ww.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 15, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum
2. Scheme for the Ping-Pong Bi Bi model for the demethylation reaction. E is the native ferric enzyme and F is compound I. A is EtOOH, P is ethanol, B is DMA, and Q is the oxidation product of DMA. (Reprinted from ref 26 with the permission of the authors.)
used. Support for this conclusion was provided by the observation that the isotope effects for the P450-catalyzed demethylations are comparable whether the reactions are performed using the reconstituted system containing NADPH-P450 reductase, NADPH, and 02 or cumene hydroperoxide is used as the oxidant (31, 32). The large isotope effects for the demethylations by the isozymes of HRP, hemoglobin, myoglobin, and lactoperoxidase are consistent with a radical mechanism of carbon-hydrogen bond cleavage and suggest a symmetrical transition state involving initial hydrogen atom abstraction, followed by hydroxyl radical recombination. This conclusion is consis-
hyde at a rate greater than, or at least equal to, that for DMA metabolism. Therefore, these results, in conjunction with the inability to detect N-oxide formation during P450-catalyzed metabolism of DMA, suggest that the N-oxide is not an intermediate in demethylation by these isozymes of P450 and that the microsomal N-oxidase activity is not associated with these P450s. When DMA N-oxide was incubated with CPO (20) or HRP (24) using the standard conditions for demethylation by these enzymes, there was no formaldehyde formation, indicating that the DMA demethylation by these enzymes also does not proceed via the formation of an N-oxide intermediate. Guengerich (28) has also provided evidence against N-oxides as intermediates in the demethylation of sparteine and its derivatives by rat liver microsomes. Therefore, these studies indicate that N-oxides are not intermediates in demethylations catalyzed by P450, HRP, and CPU, and suggest that a critical step in these demethylations involves mitial oxygenation of the a-carbon atom to form a carbinolamine intermediate rather than the N-oxide.
tent with the large intrinsic isotope effects (7.7-9.2) observed for hydrogen atom abstraction during P450-catalyzed whydroxylation of octane (33). These results provide strong evidence that these heme proteins catalyze the Ndemethylation of DMA via hydrogen atom abstraction from the a-carbon. Dinnocenzo and Banack (34) have recently reported an isotope effect of 7.68 associated with the deprotonation of the di-p-anisylaminium radical cation, a synthetic aminium cation radical not dissimilar from the radical cation expected to be formed from DMA. However, questions about the unexpectedly high pKa (about 7) and the rates of hydrogen loss need to be answered in order for the relevance of this model to be established. In contrast to the large isotope effects determined for most of the hemeprotein-catalyzed demethylations, relatively small isotope effects were observed from the reactions catalyzed by the two isozymes of P450 and CPU. The small isotope effects for the P450-catalyzed N-demethylations supported by NADPH and 02 or cumene hydroperoxide (kH/kD < 3.1) are comparable to those observed for demethylations of aromatic (35) and aliphatic (30, 36) tertiary amines by microsomal P450 and by reconstituted P450 sys-
B
A
Q
1L -
E
E A
-
F P
F
F B a E
Q
E
Figure
tems. MECHANISMS CATALYZED
OF N-DEMETHYLATIONS BY P450 AND PEROXIDASES
Because DMA demethylation involves cleavage of carbonhydrogen bonds, studies of the effect of substituting deuteriums for hydrogens of the N-methyl groups on the kinetics of demethylation can provide valuable information regarding both the chemical mechanism and the kinetic details for the overall reaction. Northrup (29) has shown that the magnitudes of isotope effects determined by intermolecular competition (separate incubations of deuterated and nondeuterated substrates) are subject to unknown degrees of suppression from numerous precatalytic Steps. Therefore, the use of this method to determine isotope effects may result in a substantial underestimation of the intrinsic isotope effect. Alternatively, isotope effects determined by intramolecular competition between two positions on the same substrate molecules which differ only in deuterium substitution can provide a more accurate estimate of the intrinsic isotope effect since they are subject to fewer precatalytic steps (30). The intramolecular isotope effects for N-methyl-N-trideuteriomethylaniline demethylation by rat liver microsomal P450 1A1 and 2B1, CPO, HRP, and several other proteins have been determined (31, 32). The isotope effects for the demethylations catalyzed by HRP (A and B-C), hemoglobin, myoglobin, and lactoperoxidase were relatively large (kH/kD > 8.5) whereas those for the two P450s and CPU were low (kH/k0 < 3.1). The observation of large and comparable isotope effects for demethylation by HRP (B-C isozyme) supported by either H2O2 or EtOOH suggests that the isotope effects are independent of the identity of the oxidant
The
small
isotope
effect
for the demethylations
cata-
lyzed by the two P450 isozymes and CPU suggests initial formation of an anilinium radical followed by deprotonation of the a-carbon. This conclusion is consistent with the small intrinsic isotope effects associatd with deprotonation of the acarbon of aminium cation radicals reported for photochemical oxidation of N-methyl-N-trideuteriomethyl-t-butylamine (kH/kD = 2.2) (37) and oxidation of trimethylamine by permanganate (kHIkD = 1.8) (38). Based on similarities of the intramolecular isotope effects for the electrochemical (kH/kD = 1.9) and microsomal (kH/kD = 1.6) demethylation of Nmethyl-N-trideuteriomethylimipramine, Shono and coworkers (39) suggested that N-demethylations catalyzed by P450 proceed by an initial one-electron transfer from the nitrogen. Guengerich (40) has demonstrated low deuterium and tritium isotope effects (-2) for P450-catalyzed metabolism of nifedipine and interpreted these in terms of
R
OH
“‘N
-
R
cH3
R”
R’2
R
+
HCHO
-a-
/0 N
-
CH3
R’
3. Scheme showing two possible mechanisms for the oxidative demethylation of tertiary arnines such as DMA by P450 and peroxidases. (Reprinted from ref 23 with the permission of the authors and the copyright holder, Pergamon Press, Oxford.) Figure
H OL LEN BERG The FASEBJournal 690 Vol. 6 January 1992 ww.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 15, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum
an electron/proton/electron transfer pathway rather than hydrogen atom abstraction. Several possible chemical mechanisms are shown in Fig. 4 for hemeprotein-catalyzed N,N-dimethylamine demethylations. One mechanism involves transfer of one electron (- e) from the amine to the iron-oxo species of the hemoprotein to form an aminium cation radical. The aminium radical could then undergo deprotonation (- H) to form a neutral carboncentered radical. This radical could then form the carbinolamine by recombination with the nascent heme iron-bound hydroxyl radical (+ OH.). The generally unstable carbinolamine would then decompose to yield formaldehyde and the secondary amine. Alternatively, the neutral carbon-centered radical could be formed directly by hydrogen atom abstraction (- H) from the amine. The enzyme-bound oxygen moiety could then recombine with the carbon-centered radical to form the carbinolamine, which would then decompose. The pathway involving nitrogen oxidation followed by acarbon deprotonation would exhibit relatively small isotope effects (
of cytochrome
xenobiotic PAUL
metabolism
-
F. HOLLENBERO,
Department
of Pharmacology,
....
Wayne State Universitr
The cytochrome P450 enzyme systems catalyze the metabolism of a wide variety of naturally occurring and foreign compounds by reactions requiring NADPH and 02. Cytochrome P450 also catalyzes peroxide-dependent hydroxylation of substrates in the absence of NADPH and 02. Peroxidases such as chloroperoxidase and horseradish peroxidase catalyze peroxidedependent reactions similar to those catalyzed by cytochrome P450. The kinetic and chemical mechanisms of the NADPH and 0a-supported dealkylation reactions catalyzed by P450 have been investigated and compared with those catalyzed by P450 and peroxidases when the reactions are supported by peroxides. Detailed kinetic studies demonstrated that chioroperoxidaseand horseradish peroxidase-catalyzed N-demethylations proceed by a Ping Pong Bi Bi mechanism whereas P450-catalyzed 0-dealkylations proceed by sequential mechanisms. Intramolecular isotope effect studies demonstrated that Ndemethylations catalyzed by P450s and peroxidases proceed by different mechanisms. Most hemeproteins investigated catalyzed these reactions via abstraction of an acarbon hydrogen whereas reactions catalyzed by P-450 and chloroperoxidase proceeded via an initial oneelectron oxidation followed by a-carbon deprotonation. ‘80-Labeling studies of the metabolism of NMC also demonstrated differences between the peroxidases and P450s. Because the hemeprotein prosthetic groups of P450, chloroperoxidase, and horseradish peroxidase are identical, the differences in the catalytic mechanisms result from differences in the environments provided by the proteins for the heme active site. It is suggested that the axial heme-iron thiolate moiety in P450 and chloroperoxidase may play a critical role in determining the mechanism of N-demethylation reactions catalyzed by these proteins. Hollenberg, P. F. Mechanisms of cytochrome P450 and peroxidase-catalyzed xenobiotic mechanism. FASEBJ. 6: 686-694; 1992. ABSTRACT
Ky Wordt: cytochrome P450 function oxidase . hydroxylation
.
and per xidaseCatIjzed
P450
peroxidse#{149}monooxygenase
mired-
ENZYME SYSTEMS IN THE ENDOPLASMIC reticulum of the liver and most other mammalian tissues catalyze the metabolism of a wide variety of endogenous and exogenous compounds including drugs, steroids, prostaglandins, chemical carcinogens, and other xenobiotics (1, 2). These enzymes, known as mixed-function oxidases, catalyze incorporation of one atom of molecular oxygen into the substrate to give product while the other oxygen atom is reduced by two electrons to give water. Although these enzymes play an important role in the detoxication of many drugs, chemical carcinogens, and other toxic agents, they are also responsible for catalyzing the metabolic activation of some substrates to highly reactive free-radical, alkylating, or arylating intermediates, which
School
-
-.
....
of Medicine, Detroit, Michigan 48J,
USA
then react with critical cellular macromolecules to initiate toxic and carcinogenic events (3). The critical role these enzymes play in the metabolic activation and detoxication of a wide variety of carcinogens and other toxic agents makes them of particular importance in light of human exposure to these compounds in the environment. The microsomal P450-dependent mixed-function oxidase enzyme
systems
are
composed
of NADPH-P450
reductase,
a phospholipid, and P4502 and exhibit an absolute requirement for NADPH and 02 for catalytic activity (1-3). Cytochrome P450 plays a critical role as it binds substrate, activates oxygen, and then catalyzes the insertion of the activated oxygen into the substrate (1-3). Therefore, the P450 determines the substrate specificity of the overall system and the structure (or structures) of the product (or products) formed during substrate metabolism.
CYTOCHROME
P450-CATALYZED
REACTIONS
The P450-dependent mixed-function oxidases catalyze a wide variety of reactions including epoxidations, N-dealkylations, 0-dealkylations, S-oxidations, and hydroxylations of aliphatic and aromatic substrates. The diversity of the reactions catalyzed can be understood by considering that the initial reaction in all cases involves insertion of a hydroxyl group into the substrate to form an hydroxylated intermediate which can then, depending on the nature of the substrate and the stability of the intermediate, undergo dealkylation, deamination, etc. (4). This hypothesis has led to the suggestion that a critical step in P450-catalyzed reactions is the formation of an hydroxyl donor, which can then transfer the hydroxyl group to the appropriate position on the substrate. The P450 family includes many different isozymes and more than 150 isoforms have been characterized (5). The purified isozymes exhibit different molecular weights, amino acid compositions, peptide maps, amino acid sequences, spectral, immunochemical, and catalytic properties (5, 6). They also exhibit different, although sometimes overlapping, substrate specificities, and the overall reactions catalyzed by P450s proceed with the stoichiometry shown: RH
+
02
+
NADPH + W H2O + NADP
-
ROH
+
Eq.
(1)
‘To whom correspondence should be addressed, at: Department of Pharmacology, Wayne State University School of Medicine, 540 E. Canfield, Detroit, MI 48201, USA. 2Abbreviations: P450, cytochrome P450, DMA, N,N-dimethylaniline;
DMA-N-oxide,
N,N-dimethylaniline-N-oxide;
roper.oxidase; HRP, horseradish peroxidase; bazole; NHMC, N-hydroxymethylcarbazole; bazole; and EtOOH, ethyl hydroperoxide.
CPO,
NMC, NFC,
chlo-
N-methylcarN-formylcar-
686 0892-6638/92/0006-686/$01 .50. © FASEB ww.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 15, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum
where RH represents the substrate. The catalytic mechanisms by which the P450s activate molecular oxygen appear to be very similar. The catalytic cycle for P450-catalyzed reactions is thought to consist of at least six discrete reactions (4, 5, 7). The first step involves binding of the substrate to the ferric form of the enzyme. The second step involves transfer to one electron from the NADPH-P450 reductase to the iron of the fernic P450 enzyme to give a ferrous enzymesubstrate complex. The reduced P450-substrate complex then binds 02 to form a ferrous enzyme-Os-substrate ternary complex with the 02 bound to the iron. The addition of a second electron to this ternary complex by the reductase results in the formation of an iron peroxo species best represented as Fe3O22. The next step involves cleavage of the oxygen-oxygen bond. One of the oxygens is released with the uptake of two protons at some stage, resulting in the formation of water. The retained oxygen remains associated with the heme iron as an activated oxygen, perhaps as an iron-oxene species. Considering the redox properties of the iron, oxygen, and sulfur atoms, numerous resonance forms could be suggested for the activated oxygen intermediate. The activated oxygen has been postulated to be a variety of forms of oxygen including superoxide, oxene, and peroxides such as penoxyacid or peroxyimidic acid. The activated oxygen atom associated with the iron is then inserted into the substrate, resulting in a two-electron oxidation of the substrate to the alcohol. The insertion of the oxygen into the substrate is thought to involve hydrogen abstraction from the substrate followed by radical recombination of the resulting transient carbon and hydroxyl radicals to form product (8). The product is then released, regenerating the native ferric P450 that is available to begin another catalytic cycle. The demonstration that P450 can use peroxides, hydroperoxides, and peracids to support N- and O-dealkylations, aliphatic hydroxylations, and olefin epoxidations suggested relationships with peroxidases and provided new approaches for studying its mechanism of action (4, 9, 10). The addition of one electron, a molecule of oxygen, a second electron and two protons during the catalytic cycle is formally equivalent to the addition of hydrogen peroxide (H2O2). The pathway involving substitution of a peroxy compound for NADPH and 02 is referred to as the peroxide shunt. P450-catalyzed oxygenations by 02 and by peroxy compounds are believed to have common mechanistic features. Therefore, the investigation of the mechanisms by which these oxidants support P450-catalyzed reactions is of interest because: 1) this system may be studied in the absence of NADPH, reductase, and 02, thereby simplifying it for study; 2) there are many similarities to reactions catalyzed by peroxidases, which have been studied in great detail; and 3) comparison of the reactions supported by NADPH and molecular oxygen with those supported by the organic oxidants may help us to understand both systems. The isolation of hydroperoxides of tetralin, fluorene, steroids, 9,10-dimethyl-1,2 -benzanthracene, and 9-methylfluorene from microsomal reaction mixtures has been cited as evidence for the role of hydroperoxides in biological hydroxylations (11).
REACTIONS
CATALYZED
Peroxidases
The function of the peroxidase may be to oxidize a particular substrate or to reduce a hydroperoxide. Most peroxidases have heme as their prosthetic group. Many peroxidasecatalyzed reactions proceed as shown in Eqs. 3-5, Peroxidase
Compound
Compound
+
I
II
ROOH
AH2
+
+
AH2
AH
Compound
-
Compound
-
Peroxidase
-
-
Nonradical
I
II
+
+
+
ROH Eq.
(3)
Eq.
(4)
Eq.
(5)
AH.
AH
products Eq. (6)
where ROOH is the hydroperoxide, AH2 is the reducing substrate, and AH is a free radical (12). This process predicts that free radicals would be the products of many organic molecules. The fate of these radicals is characteristic of each substrate. However, some substrates appear to reduce compound I directly to the native enzyme without the formation of compound II (13, 14). For some peroxidases such as HRP, compound I is a relatively stable, spectrally distinct enzyme species two oxidizing equivalents above the native ferric enzyme. The formation of compound I proceeds with the concomitant release of the alcohol, ROH, derived from the hydroperoxide. Based on studies of the CPO-catalyzed evolution of oxygen from 18O-labeled H2O2 and peracids, it has been suggested that compound I contained one, and only one, atom of oxygen from the peroxy substrate (15), and that it might exist as any of the configurations shown in Fig. 1. These structures are electronically equivalent and have the following features: 1) they are two oxidizing equivalents above the native ferric (+3) enzyme, and 2) they contain only one oxygen liganded to the heme iron. Several of these structures could serve as good hydroxylating agents.
OH
OH
OH
x
I
BY PEROXIDASES
are
widely distributed in nature. As shown in the oxidation of inorganic and organic substrates with hydrogen peroxide, alkyl hydropenoxides, or acyl hydroperoxides as the ultimate electron acceptor.
Eq. 2, they catalyze
ROOH
+
AH2
-
ROH
+
A
+
H20
Eq.
(2)
x Figure pound
1. Scheme I.
showing
possible
structures
for peroxidase
corn-
MECHANISMS OF HEMEPROTEIN CATALYSIS ww.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 15, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum
CPO is a heme protein that has been isolated from the mold Caldariomyces furnago and has been characterized in detail by Hewson and Hager (15). CPO is unique among peroxidases in the diversity of reactions it catalyzes (16). In the presence of halide anions (C1, Br, or 1, but not F; represented by X in Eq. 7), CPO catalyzes the formation of a carbon-halogen bond with a variety of nucleophiles (AH in Eq. 7). Like other protoheme peroxidases, CPO catalyzes the peroxide-supported oxidation of a wide variety of substrates (BH2 in Eq. 8). Like catalase, CPO also catalyzes the decomposition of H2O2 to give 0 (Eq. 9). However, unlike catalase, CPO catalyzes the decomposition of alkyl hydroperoxides and peracids to give 02. AH
+
BH2
X
+
2 H2O2
+
H2O2
H2O2 -
-
2H20
+
B +
H
+
AX
-
+
2H20
2H2O
02
Eq.
(7)
Eq.
(8)
Eq.
(9)
Another property of CPO is the striking similarity of its heme active site environment to that of bacterial and mammalian P450s (17-19). Similar to P450s, the reduced CPO forms a CO complex having the Soret maximum at an unusually long wavelength, Xm = 443 nm (17). In addition, Mossbauer spectroscopy, EPR spectroscopy, resonance Raman spectroscopy, magnetic circular dichroism spectroscopy, and extended X ray absorption fine structure studies suggest a strong similarity between the heme active site environments of P450 and CPO (18-20). Extensive spectral studies of mammalian and bacterial P450s and X ray crystallographic studies of bacterial P450 provide conclusive evidence that the fifth axial ligand to the heme iron is a thiolate provided by a cysteine residue of the protein (18, 19). Blanke and Hager (21) have demonstrated that CPO also has a thiolate axial ligand. The striking similarities in the physiochemical properties of these two hemeproteins and the potential for compound Ito serve as an hydroxylating agent prompted us to investigate the ability of CPO to catalyze reactions characteristic of P450 (20).
PEROXIDASE-CATALYZED REACTIONS
N-DEMETHYLATION
CPO catalyzes the peroxide-dependent dealkylation of DMA as well as other N-methyl arylamines in a manner similar to that described for the peroxide-supported demethylation of DMA and benzphetamine by purified rabbit liver microsomal P450 (20, 22). The reaction for both the CPOand P450-catalyzed dealkylations can be written as shown: ROOH
+ R R’-N-CH3 R-OH R R’NH + HCHO
+
Eq.
(10)
where ROOH is the peroxide and R’R”N-CH3 is the N-methylarylamine substrate (either secondary or tertiary). The stoichiometry for formation of N-methylaniline and formaldehyde from DMA is 1:1. The turnover number for the C P0-catalyzed dealkylation of DMA supported by EtOOH (1476) was much greater than that observed for P450-catalyzed reactions in the reconstituted system, which range up to 20-30 (20, 22). CPO exhibited normal Michaelis-Menten saturation kinetics with respect to both EtOOH and DMA. A variety of
different peroxides, hydroperoxides and peracids were able to support the demethylation reaction. The initial velocity of the demethylation reaction was dependent on the identity of the oxidant. EtOOH was the most effective. CPO was able to use various arylamine substrates and the turnover numbers were also dependent on the identity of the substrate. Although P450 inhibitors such as azide and n-propyl gallate gave marked inhibition, other P450 inhibitors such as SKF-525A, metyrapone, and piperonyl butoxide did not inhibit the reaction. Tiron and DL-epinephrine, trapping agents for the superoxide anion, markedly inhibited the demethylation reactions whereas superoxide dismutase had no significant effect. Spin trapping agents such as a-phenyl-tbutylnitrone or 5,5-dimethylpyrroline-N-oxide, had no significant inhibitory effect on the demethylations, suggesting that free radical formation does not play a major role in demethylation. Diphenylfuran and DL-histidine, which react with singlet oxygen, did not inhibit the reaction. These studies of the demethylation reactions catalyzed by CPO had been initiated to investigate the hypothesis that the structural similarities between CPO and P450 might be reflected in the ability of CPO to function catalytically like a P450. Because CPO was able to function like a P450 for N-demethylation reactions, the ability of hemeproteins that did not provide P450-like active site environments for the heme to catalyze N-dealkylation reactions was investigated. Two preparations of HRP (isozyme A and isozymes B-C) exhibited catalytic activities for DMA demethylation significantly greater than those for the liver microsomal P450-dependent demethylations supported by NADPH and molecular oxygen or cumene hydroperoxide (4, 22-24). Lactoperoxidase also exhibited catalytic activity comparable to that of the most active P450 in the demethylation reaction (24). HRP catalyzes the peroxide-supported oxidation of a variety of different organic substrates (Eq. 2). As with CPO, the first step in HRP-catalyzed reactions involves reaction of the ferric (Fe3) form of the enzyme with the peroxide to form compound I, The HRP-catalyzed N-demethylation of DMA supported by EtOOH exhibited a stoichiomety of hydroperoxide consumption, formaldehyde formation, and N-methylaniline formation of 1:1:1 (24). The HRP-catalyzed N-demethylations were supported by a variety of oxidants including hydroperoxides, peroxides, and peracids, and turnover numbers were highly dependent on the identity of the oxidant. Unlike CPO-catalyzed demethylations (20), sodium chlorite was very efficient at supporting the HRP-catalyzed reaction. Chlorite is utilized by HRP to catalyze both chlorination and peroxidation reactions (25). HPLC analysis of the chlorite-supported reaction mixture suggested that HRP also catalyzed chlorite-dependent chlorination of DMA (24). When conditions were optimized, the turnover number for DMA demethylation by the B-C fraction of HRP was 7061 for the hydrogen-peroxide-supported reaction. This rate is significantly greater than for CPO under optimal conditions. HRP catalyzed the demethylation of a variety of secondary and tertiary N-methylarylamines. The HRP-catalyzed demethylation reactions exhibited normal Michaelis-Menten saturation kinetics with respect to DMA (Km = 0.34 mM), and H202 (Km = 0.016 mM), as well as EtOOH (Km = 0.020 mM). If an enzyme-generated activated oxygen species such as the superoxide anion, singlet oxygen, or the hydroxyl radical were a free intermediate in the HRP-catalyzed demethylations, reagents that react with these activated oxygen species should markedly inhibit the demethylations. Although Tiron, a reagent that reacts with the superoxide anion, markedly in-
688 Vol. 6 January 1992 The FASEB Journal HOLLEN BERG ww.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 15, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum
hibited demethylation, superoxide dismutase did not inhibit the HRP-catalyzed demethylation. Methionine, diphenylfuran, azide, and histidine are all reported to be efficient singlet oxygen-quenching agents. The inability of these compounds to inhibit demethylation suggests that singlet oxygen is not an intermediate in this reaction (24). Reagents used as scavengers for the hydroxyl radical include thiourea, mannitol, benzoate, formate, ethanol, /3-methylthiopropionaldehyde, dimethyl sulfoxide, and ascorbate. Ascorbate was the only hydroxyl radical scavenger that significantly inhibited the HRP-catalyzed demethylations. Because ascorbate is a substrate for HRP, it is probably inhibiting demethylations by competing with DMA for the enzyme active site rather than by trapping a hydroxyl radical intermediate. If the HRP-catalyzed demethylations proceeded via the formation of a free radical intermediate, then spin trapping reagents such as a-phenyl-t-butylnitrone, nitrosobenzene, nitromethane, tri-t-butylphenol, 2-methyl-2-nitrosopropane, or 5,5 dimethylpyrroline-N-oxide should markedly inhibit the demethylation. However, even high concentrations (10-75 mM) of these agents caused relatively small inhibitions of the demethylation reactions, suggesting that formation of free radicals, which could then dissociate from the enzyme, does not play an important role in demethylation. Thus, these results suggest that substrate free radicals are not formed as free intermediates in the demethylation reaction; however, the results do not preclude the involvement of enzyme-bound radical species as intermediates in the reaction. The kinetic mechanism for the C P0-catalyzed dealkylation of DMA supported by EtOOH has been determined by using initial rate and inhibition studies (26). When the concentration of DMA was varied systematically in the presence of several different fixed concentrations of EtOOH, double reciprocal plots of the initial rate data yielded a series of parallel lines. When the data were plotted as a function of the concentration of the EtOOH, parallel lines were also observed. Although these results are consistent with a PingPong mechanism, they do not unequivocally rule out a sequential mechanism. To distinguish between sequential and Ping-Pong mechanisms, the concentrations of both substrates were varied simultaneously while keeping them in a constant ratio. Double reciprocal plots of the initial rate data were linear, consistent with the initial assignment of a Pin gPong mechanism for the N-demethylation reaction. For reactions that proceed via a Ping-Pong mechanism, competitive substrate inhibition is normally observed at higher concentrations of the two substrates due to binding of the substrate to the wrong form of the enzyme to yield an abortive binary complex. Competitive substrate inhibition was observed when the concentration of EtOOH was varied in the presence of several inhibitory concentrations of DMA. Competitive substrate inhibition was also observed when the concentration of DMA was varied in the presence of several inhibitory concentrations of EtOOH. 2,5-Dimethylfuran, a potent inhibitor of the CPO-catalyzed demethylation, was a competitive inhibitor with respect to DMA and uncompetitive with respect to EtOOH. These results are consistent with dead-end inhibition in a Ping-Pong system. Thus, the combined initial velocity and inhibition results are consistent with a Ping-Pong Bi Bi mechanism as the minimal kinetic model for the CPO-catalyzed N-demethylations. In the model shown in Fig. 2 the CPO (E) reacts with the EtOOH (A) to form compound 1(F), the oxidized enzyme intermediate, with the concomitant release of ethanol (P). The DMA (B) then binds to compound I and is oxidized, presumably
to give a carbinolamine (Q). This results in the regeneration of the native CPO and the presumed carbinolamine intermediate, which is unstable and spontaneously decomposes to give N-methylaniline and formaldehyde. Detailed initial rate and inhibition studies on the HRP-catalyzed N-demethylation of DMA indicate that this reaction also proceeds by a Ping-Pong Bi Bi mechanism (24). Although these results establish the kinetic mechanism for peroxidase-catalyzed demethylation reactions, they provide no information regarding the nature of the chemical events involved in DMA demethylation.
ROLE OF AN N-OXIDE DEMETHYLATION
INTERMEDIATE
IN
One mechanism suggested for the oxidative demethylation of DMA involves initial oxygenation of the amine nitrogen to form DMA-N-oxide as shown in Fig. 3. This could then rearrange via an enzyme-catalyzed reaction to give the carbinolamine which, for a substrate such as DMA, is relatively unstable in aqueous solution and readily decomposes to give formaldehyde and the amine product. An alternative route shown in Fig. 3 involves initial oxygenation of the a-carbon to give the carbinolamine directly, which readily decomposes to give the products indicated. The N-oxidation of DMA occurs concomitantly with N-demethylation in liver microsomal preparations. Because the N-oxide is metabolized by liver microsomes to form N-methylaniline and formaldehyde, the formation of a tertiary amine N-oxide as an intermediate in N-demethylation has been suggested (27). Although studies from several laboratories support the N-oxide as an intermediate in N-demethylations, studies from other laboratories suggest that the N-oxide is not an intermediate in demethylations. All these studies were performed using microsomal preparations. Because the metabolism of DMA to give formaldehyde and N-methylaniline can proceed via two different pathways (Fig. 3) involving different microsomal monooxygenase enzyme systems, the interpretation of these experiments was complicated by the use of microsomal preparations containing different enzyme systems. In addition, although microsomal metabolism of DMA results in N-oxide formation, it has not been conclusively demonstrated that the N-oxide is an intermediate in the reaction rather than a by-product of metabolism. Therefore, the metabolism of DMA to the N-oxide by four different purified isozymes of P450 (23) as well as by CPO (20) and HRP (24) was investigated to determine the possible role of the N-oxide as an intermediate in N-demethylation reactions. All four isozymes of P450 (rabbit liver forms 1A23 and 2B4 and rat liver 1A1 and 2B1) investigated in the reconstituted system did not catalyze the metabolism of DMA to give measurable amounts of the N-oxide. In addition, all four isozymes of P450 exhibited significantly greater turnover numbers (two to sevenfold) for the demethylation of DMA than for the N-oxide. If the N-oxide were an intermediate in the demethylation reaction catalyzed by a P450, then the N-oxide should be metabolized to give formalde-
3The nomenclature used here is that based on structural homology (5). P450 1A1 is the 3-methylcholanthrene-inducible form from rat liver microsomes, P450 1A2 is the 3-methylcholanthreneinducible form from rabbits, P450 2B1 is the major phenobarbitalinducible form from rats, and P450 2B4 is the major phenobarbitalinducible form from rabbits.
MECHANISMS OF HEMEPROTEIN CATALYSIS 689 ww.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 15, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum
2. Scheme for the Ping-Pong Bi Bi model for the demethylation reaction. E is the native ferric enzyme and F is compound I. A is EtOOH, P is ethanol, B is DMA, and Q is the oxidation product of DMA. (Reprinted from ref 26 with the permission of the authors.)
used. Support for this conclusion was provided by the observation that the isotope effects for the P450-catalyzed demethylations are comparable whether the reactions are performed using the reconstituted system containing NADPH-P450 reductase, NADPH, and 02 or cumene hydroperoxide is used as the oxidant (31, 32). The large isotope effects for the demethylations by the isozymes of HRP, hemoglobin, myoglobin, and lactoperoxidase are consistent with a radical mechanism of carbon-hydrogen bond cleavage and suggest a symmetrical transition state involving initial hydrogen atom abstraction, followed by hydroxyl radical recombination. This conclusion is consis-
hyde at a rate greater than, or at least equal to, that for DMA metabolism. Therefore, these results, in conjunction with the inability to detect N-oxide formation during P450-catalyzed metabolism of DMA, suggest that the N-oxide is not an intermediate in demethylation by these isozymes of P450 and that the microsomal N-oxidase activity is not associated with these P450s. When DMA N-oxide was incubated with CPO (20) or HRP (24) using the standard conditions for demethylation by these enzymes, there was no formaldehyde formation, indicating that the DMA demethylation by these enzymes also does not proceed via the formation of an N-oxide intermediate. Guengerich (28) has also provided evidence against N-oxides as intermediates in the demethylation of sparteine and its derivatives by rat liver microsomes. Therefore, these studies indicate that N-oxides are not intermediates in demethylations catalyzed by P450, HRP, and CPU, and suggest that a critical step in these demethylations involves mitial oxygenation of the a-carbon atom to form a carbinolamine intermediate rather than the N-oxide.
tent with the large intrinsic isotope effects (7.7-9.2) observed for hydrogen atom abstraction during P450-catalyzed whydroxylation of octane (33). These results provide strong evidence that these heme proteins catalyze the Ndemethylation of DMA via hydrogen atom abstraction from the a-carbon. Dinnocenzo and Banack (34) have recently reported an isotope effect of 7.68 associated with the deprotonation of the di-p-anisylaminium radical cation, a synthetic aminium cation radical not dissimilar from the radical cation expected to be formed from DMA. However, questions about the unexpectedly high pKa (about 7) and the rates of hydrogen loss need to be answered in order for the relevance of this model to be established. In contrast to the large isotope effects determined for most of the hemeprotein-catalyzed demethylations, relatively small isotope effects were observed from the reactions catalyzed by the two isozymes of P450 and CPU. The small isotope effects for the P450-catalyzed N-demethylations supported by NADPH and 02 or cumene hydroperoxide (kH/kD < 3.1) are comparable to those observed for demethylations of aromatic (35) and aliphatic (30, 36) tertiary amines by microsomal P450 and by reconstituted P450 sys-
B
A
Q
1L -
E
E A
-
F P
F
F B a E
Q
E
Figure
tems. MECHANISMS CATALYZED
OF N-DEMETHYLATIONS BY P450 AND PEROXIDASES
Because DMA demethylation involves cleavage of carbonhydrogen bonds, studies of the effect of substituting deuteriums for hydrogens of the N-methyl groups on the kinetics of demethylation can provide valuable information regarding both the chemical mechanism and the kinetic details for the overall reaction. Northrup (29) has shown that the magnitudes of isotope effects determined by intermolecular competition (separate incubations of deuterated and nondeuterated substrates) are subject to unknown degrees of suppression from numerous precatalytic Steps. Therefore, the use of this method to determine isotope effects may result in a substantial underestimation of the intrinsic isotope effect. Alternatively, isotope effects determined by intramolecular competition between two positions on the same substrate molecules which differ only in deuterium substitution can provide a more accurate estimate of the intrinsic isotope effect since they are subject to fewer precatalytic steps (30). The intramolecular isotope effects for N-methyl-N-trideuteriomethylaniline demethylation by rat liver microsomal P450 1A1 and 2B1, CPO, HRP, and several other proteins have been determined (31, 32). The isotope effects for the demethylations catalyzed by HRP (A and B-C), hemoglobin, myoglobin, and lactoperoxidase were relatively large (kH/kD > 8.5) whereas those for the two P450s and CPU were low (kH/k0 < 3.1). The observation of large and comparable isotope effects for demethylation by HRP (B-C isozyme) supported by either H2O2 or EtOOH suggests that the isotope effects are independent of the identity of the oxidant
The
small
isotope
effect
for the demethylations
cata-
lyzed by the two P450 isozymes and CPU suggests initial formation of an anilinium radical followed by deprotonation of the a-carbon. This conclusion is consistent with the small intrinsic isotope effects associatd with deprotonation of the acarbon of aminium cation radicals reported for photochemical oxidation of N-methyl-N-trideuteriomethyl-t-butylamine (kH/kD = 2.2) (37) and oxidation of trimethylamine by permanganate (kHIkD = 1.8) (38). Based on similarities of the intramolecular isotope effects for the electrochemical (kH/kD = 1.9) and microsomal (kH/kD = 1.6) demethylation of Nmethyl-N-trideuteriomethylimipramine, Shono and coworkers (39) suggested that N-demethylations catalyzed by P450 proceed by an initial one-electron transfer from the nitrogen. Guengerich (40) has demonstrated low deuterium and tritium isotope effects (-2) for P450-catalyzed metabolism of nifedipine and interpreted these in terms of
R
OH
“‘N
-
R
cH3
R”
R’2
R
+
HCHO
-a-
/0 N
-
CH3
R’
3. Scheme showing two possible mechanisms for the oxidative demethylation of tertiary arnines such as DMA by P450 and peroxidases. (Reprinted from ref 23 with the permission of the authors and the copyright holder, Pergamon Press, Oxford.) Figure
H OL LEN BERG The FASEBJournal 690 Vol. 6 January 1992 ww.fasebj.org by Univ of So Dakota Lommen Hlth Sci Library (192.236.36.29) on September 15, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNum
an electron/proton/electron transfer pathway rather than hydrogen atom abstraction. Several possible chemical mechanisms are shown in Fig. 4 for hemeprotein-catalyzed N,N-dimethylamine demethylations. One mechanism involves transfer of one electron (- e) from the amine to the iron-oxo species of the hemoprotein to form an aminium cation radical. The aminium radical could then undergo deprotonation (- H) to form a neutral carboncentered radical. This radical could then form the carbinolamine by recombination with the nascent heme iron-bound hydroxyl radical (+ OH.). The generally unstable carbinolamine would then decompose to yield formaldehyde and the secondary amine. Alternatively, the neutral carbon-centered radical could be formed directly by hydrogen atom abstraction (- H) from the amine. The enzyme-bound oxygen moiety could then recombine with the carbon-centered radical to form the carbinolamine, which would then decompose. The pathway involving nitrogen oxidation followed by acarbon deprotonation would exhibit relatively small isotope effects (
