THE JOURNAL 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

OF BIOLOGICAL CHEMISTRY Vol. 268, No. 1, Issue of January 5, pp. 59-65.1993 Printed in U. S. A.

Cumene Hydroperoxide-mediated Inactivation of Cytochrome P450 2B 1 IDENTIFICATION OF AN ACTIVE SITE HEME-MODIFIED PEPTIDE*

(Received for publication, April 21, 1992)

Kunquan Yao, Arnold M. Falick, Naina Patel, and Maria Almira CorreiaS From the Departments of Pharmacology and Pharmaceutical Chemistry and the Liver Center, University of California, San Francisco, California 94143

Cumene hydroperoxide (CuO0H)-mediated inacti- vation of cytochromes P450 (P450) results in the deg- radation of their prosthetic heme to products that al- kylate the apoprotein. Indirect approaches suggest that this alkylation occurs at the active site. in order to identify the specific apoprotein site(s) alkylated, puri- fied ‘H- or 14C-heme-labeled P450 2B1 was incubated with CuOOH and subjected to lysyl endopeptidase-C digestion. Two major peaks (L1 and L2) containing 3H- or “C-labeled peptides were detected by reverse-phase high pressure liquid chromatography of the digest. L1 contained the highest specific radioactivity and after Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis yielded 3 peptide bands (Mr = 3,500 (Pl) , 5,000 (P2), and 7,000 (P3)). Although all 3 bands were found radiolabeled, the yield of P1 was higher than that of P2 or P3. Amino acid sequence analysis of the first 13 N-terminal residues of P1 revealed the sequence RICLGEGIARNEL, corresponding to resi- dues 434-446 of the reported 2B1 sequence. A species with the molecular mass of 3771 f 1 Da was detected in preliminary electrospray mass spectrometric analy- sis of L1. Since the theoretical average mass of the predicted peptide (residues 434-466) is 3721.99 Da, the additional 49 f 1 Da are considered to be contrib- uted by the alkylating heme fragment. This alkylated 2B1 sequence contains not only CYS~~’, the conserved residue that provides the SH ligand for heme, but also other highly conserved residues, and therefore corre- sponds to the heme-sandwiching helix L of P450cam. To our knowledge, this is the first report to localize CuOOH-induced heme alkylation of 2B1 to its active site.

The hepatic microsomal hemoproteins collectively termed cytochromes P450 (P450),’ include multiple constitutive and

* This work was supported in part by National Institutes of Health (NIH) Grant GM 44037 (to M. A. C.). The Liver Center Core Facility on spectrophotometry, supported by NIH Grant DK 26743, and the University of California, San Francisco Mass Spectrometry Facility (A. L. Burlingame, Director), supported by NIH Grants RR01614 and DK26743 and National Science Foundation Grant DIR8700766, are also acknowledged. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom correspondence should be addressed Dept. of Phar- macology, Box 0450, University of California, San Francisco, CA 94143. Tel.: 415-476-3992; Fax: 415-476-5292.

’ The abbreviations used are: P450, cytochrome P450; ALA, ami- nolevulinic acid; CuOOH, cumene hydroperoxide; DDEP, 3,5-dicar- bethoxy-2,6-dimethyl-4-ethyl-1,4-dihydropyridine; HPLC, high pres- sure liquid chromatography; Lys-C, lysyl endopeptidase-C; Tricine, N-[2-hydroxy-l,l-bis(hydroxymethyl)ethyl]glycine; PAGE, poly- acrylamide gel electrophoresis.

inducible enzymes, which recruit heme (iron-protoporphyrin IX), as their essential prosthetic moiety (1-3). These hemo- proteins are monomeric (M, = 50 kDa) containing one heme per mole of enzyme. In spite of their identical heme moieties, P450s differ functionally, a property conferred by individual heme-apocytochrome microenvironments. P450s are instru- mental in the oxidative/reductive metabolism of various phys- iologically relevant endobiotics and xenobiotics. However, although all these reactions result in the formation of readily excretable products, not all are beneficial (2). It is now rec- ognized that in the course of certain redox reactions, the participating P450 is sacrificed in a process classified as a mechanism-based or “suicide” inactivation (Ref. 4, and ref- erences therein). Three modes of such substrate-mediated P450 inactivation are currently known: (i) prosthetic heme destruction via N-alkylation/arylation (i .e. allylisopropylacet- amide, secobarbital); (ii) apocytochrome alkylation by a re- active intermediate (chloramphenicol, secobarbital, Il-unde- cynoic acid); and (iii) destruction of the prosthetic heme to products that irreversibly bind to the apocytochrome (CCl,, spironolactone, 3,5-dicarbethoxy-2,6-dimethyl-4-ethyl-l,4- dihydropyridine (DDEP)) (4-16). By definition, suicide in- activations occur at the active site (Ref. 4, and references therein). However, this criterion has not been rigorously applied in all the cases examined. Isolation and structural characterization of the N-alkylated heme has unequivocally established this criterion for N-alkylation of P450 heme (mode i) (Ref. 4, and references therein). Such definitive mechanistic classification of modes ii or iii of drug-induced P450 destruction as “suicide inactivation” is largely precluded by the inaccessibility to structural analyses of the highly hydrophobic apoP450 active site regions and their resistance to proteolytic digestion with an array of proteases. Further- more, in addition to methodological approaches to the isola- tion and structural characterization of the P450 active site cores, such mechanistic classification of modes ii or iii inac- tivation processes requires the delineation of each P450 active site microenvironment.

However, to our knowledge, no crystallographic structure is yet available for the eukaryotic P450s. Because the x-ray crystal structure of the bacterial P450cam (P450 101) is known (17-20), its heme-apoprotein domain has served as an excellent motif for probing the corresponding domains of the mammalian P450s. Accordingly, in analogy to P450cam, the heme of the membrane-bound P450s is well sequestered within the protein interior, sandwiched by two helices (prox- imal L and distal I in P450cam) that are highly conserved through evolution (17, 20). A cysteine residue located in the highly conserved pentadecapeptide region (helix L/P450cam) in the COOH terminus, provides the thiolate ligand for coor-

59

60 Heme-modified P450 2B1 Active Site Peptide

dinating the prosthetic heme iron. This relatively "loose" assembly of the heme with the apocytochrome permits dy- namic heme movement in and out of the active site, a feature generally retained in modes i and ii of suicide inactivation of the P450.

In contrast, the inactivation of P450s (3A, 2E1) by certain substrates (DDEP, spironolactone, and CClJ (5, 10, 11, 21, 22) results in prosthetic heme destruction to products that irreversibly bind to the apocytochrome. It is unknown whether oxidative uncoupling and H20, accumulation are responsible for this, or even whether such heme-alkylation occurs at the active site and thus qualifies as truly suicidal. The notorious

TABLE I CuOOH-induced "C-heme-derived alkylation of apoP450 2Bl

"C-heme-labeled P450 2B1 was purified as described and incubated with CuOOH (0.5 mM) at 37 "C for 15 min. The protein was precipi- tated as described and the cpm in the supernatant and pellet deter- mined bv scintillation counting.

Preparation

"C-P450 2B1 (2 nmol) Pellet Supernatant

Recovery (total cpm) Recovery (% of initial) Irreversible "C-heme binding (%) Loss of P450 2B1 contentC (%)

Radioactivity

-CuOOH +CuOOH

cPm 1547 1547

1276 594

1306 1372 84 89

20 >99

30 778"

2.3 50.3b

a Corrected for 71 f 11% protein recovery observed after extensive solvent washes in 3 separate experiments.

This value compares favorably with the corresponding value of 50 f 2 of irreversible "C-heme-binding to P450 loss observed in CuOOH-incubated liver microsomes from phenobarbital-pretreated rats. (Because this estimate is based on the assumption that the alkylating species is the intact heme skeleton, the values for irrevers- ible heme binding may be considerably underestimated.)

e Spectrally determined.

FIG. 1. Visible electronic absorp- tion spectra of native and CuOOH- inactivatedP450 2B1. The visible ab- sorption spectra of aliquots of native (-) or CuOOH-incubated (- - - - -) P450 2B1 were determined with a UV- Vis SLM-Aminco 2000TM spectropho- tometer.

0.1

I

I

/ .

L

recalcitrance of the purified P450 3A1 to functional reconsti- tution has largely deterred the mechanistic characterization of its mode iii inactivation. However, because such inactiva- tion apparently is also a hallmark of prosthetic heme degra- dation of microsomal or purified P450 2B1 (23,24) and P450s 3A (6, 10)' by H202 or cumene hydroperoxide (CuOOH), as an initial approach to characterizing this inactivation process, we examined the site of heme-alkylation of CuOOH-inacti- vated P450 2B1. The results of this investigation are reported below.

MATERIALS AND METHODS

Purification of 14C-/3H-Heme-labeled P450 2Bl"Male Sprague- Dawley rats were pretreated with sodium phenobarbital (80 mg/kg/ day, intraperitoneally) for 5 days. To label the prosthetic heme moiety of P450, the heme precursor 6-[4-"C]- or [3,5-3H]aminolevulinic acid (ALA) hydrochloride ( ~ 2 5 pCi/rat) was injected intraperitoneally 2- 3 h before killing. Liver microsomes were prepared and P450 2B1 was isolated and purified by the method of Waxman and Walsh (26) and its content assayed as described previously (7).

CuOOH-mediated Inactivation of P450 2Bl"The incubation mix- ture contained purified P450 2B1 (2 nmol), CuOOH (0.5 mM), EDTA (1.5 mM), and 0.1 M phosphate buffer, pH 7.4, in a final volume of 1 ml. When the CuOOH-mediated heme alkylation of apoP450 2B1 was examined, purified "C- or 3H-heme-labeled P450 2B1 was used instead of the unlabeled 2B1 and incubated under identical condi- tions. The reaction was terminated by the addition of CHaOH con- taining 5% H2S04, rat liver microsomes (10 mg of protein) were added as protein carrier and the heme-derived radioactivity determined after extensive solvent washing of the protein pellet to remove the nonco- valently bound radiolabeled material, as previously reported (5, 6). The percent pellet protein recovery after washing was of the order of 71 f 11 in three separate determinations.

Lysyl Endopeptidase-C Digestion of CuOOH-inactivated and Native P450 2Bl Preparations"P450 2B1 (10 nmol) was incubated with or without CuOOH as described above. After incubation, P450 2B1 was dialyzed against 0.1 M Tris HCI buffer, pH 9.0, at 4 "c overnight. Solid urea was added to a final concentration of 8 M. The mixture was incubated at 60 "C for 30 min. Lysyl endopeptidase-C (LYS-C)

K. Yao and M. A. Correia, preliminary observations.

Heme-modified P450 2Bl Active Site Peptide 61

FIG. 2. Reverse-phase HPLC of Lys-C digests of native ( A ) and CuOOH-inactivated ( B ) P450 2B1. Aliquots (400 pl) of control or CuOOH- incubated enzyme were subjected to HPLC as detailed under "Materials and Methods." Peptide fractions were moni- tored at 220 nm and at 400 nm (insets).

r I A

600

3

E 300

2 00

100

0

A E20 I

20 40

~B J I

600:

500:

200

0 -

fl 220

20 4 0 60 80 100 120 T n m r ( m l n 1 J

was added to the 2B1 incubation at a molar ratio of 1/40. The mixture was incubated at 30 "C overnight.

Reversed-phase HPLC of the Lys-C Digest-An aliquot of the Lys- C digest of 2B1 was chromatographed on a 2.1 x 220-mm Applied Biosystems BU 300 C4 (7 pm) reversed phase column. The solvent system used was Hz0 containing 0.1% heptafluorobutyric acid (sol- vent A), and 1:l (v/v) mixture of isopropanol and acetonitrile con- taining 0.08% heptafluorobutyric acid (solvent B). The peptides were separated by gradient elution from 15% to 85% B over 120 min, at a flow rate of 0.4 ml/min. The distribution of radioactivity was moni- tored by sequential scintillation counting of 2-min fractions.

Acid Butanone Extraction ofP450 2BI Heme-Heme was extracted from native or CuOOH-inactivated 3H-heme-labeled P450 2B1 (10 nmol/incubation) as described elsewhere (27), and then subjected to Lys-C digestion and HPLC peptide mapping as described above, except that larger aliquots of the digests were injected.

Amino Acid Sequencing and Composition Analysis of the HPLC- separated Peptide Fractions-The relevant HPLC peptide peaks were collected, subjected to Tricine-SDS-PAGE on a 20% polyacrylamide gel and then blotted on to a ProBlottTM membrane, a polyvinylidine difluoride membrane for protein sequencing (Applied Biosystems Inc.). The peptide bands were sliced and submitted for sequencing by the Edman degradation method at the Biomolecular Resource Center, University of California, San Francisco. The major peptide band was also subjected to direct HCl hydrolysis on ProBlott, the HC1 was removed, and the amino acid hydrolysate was dissolved in citrate buffer (0.2 M, pH 2.2); aliquots were directly subjected to HPLC at the University of California, Davis Protein Structure Laboratory.

Electrospray Mass Spectrometric Analyses of the HPLC-separated Peptide Fraction-The HPLC fraction was dried under NZ, and then dissolved in a mixed solvent (CHC13:MeOH:H20, 2:5:2 (v/v/v)) con- taining 1% acetic acid. The sample was run on a Fisons/VG (Man- chester, UK) Bio-Q mass spectrometer equipped with an electrospray ion source. The same mixed solvent was used in the electrospray source at a flow rate of 4 pl/min. Data were acquired at a scan rate of 10 s/scan, and several scans were averaged.

RESULTS AND DISCUSSION

CuOOH-mediated Inactivation of P450 2Bl "CuOOH in- cubation with purified 14C-heme-labeled P450 2B1 at 37 "C for 15 min resulted in a nearly complete loss (>99%) of the spectrally detectable hemoprotein, consistent with previous reports (6, 8, 13, 23, 24). In parallel, nearly 50% of the 14C- heme-derived radioactivity was found covalently bound to the inactivated P450, when the inactivated enzyme was denatured and extensively washed free of adventitiously bound 14C-heme and/or heme derivatives (Table I). In contrast, 20% of the spectrally detectable content of the non-CuOOH-incubated 2B1 (control) was lost during this time,3 with only 2.3% of the 14C-heme radioactivity covalently attached to the protein (Table I). The corresponding recoveries of the initial 14C radioactivity were of the order of 84 and 89% for the CuOOH- incubated and control P450 2B1. The relative visible absolute absorption spectrum of the CuOOH-incubated P450 2B1 ver- sus that of the control P450 2B1 preparation attests to the virtual abolition of the conjugated electronic .rr-system of the intact iron-porphyrin or tetrapyrrole skeleton (Fig. 1). Fur- thermore, this spectrum was virtually identical when the CuOOH incubation was carried out under N P atmosphere (not shown), indicating that it was not an artifact created by secondary radical reactions in the aerobic environment. To- gether, these findings confirm previous reports of the effects of CuOOH-mediated inactivation of various P450s (6, 8, 13, 23).

In preliminary studies, dialysis against 0.1 M Tris buffer,

This loss is somewhat higher than that (-5%) usually observed after incubation of substrate-free P450 2B1.

62 Heme-modified P450 2B1 Active Site Peptide

FIG. 3. HPLC profiles of Lys-C digests of native ( A ) , CuOOH-inac- tivated ( B ) , and CuOOH-inacti- vatedbutanone-extracted (C) 3H- heme-labeled P450 2B1 and the corresponding radioactivity of the peptide fractions. Aliquots (400 pl) of control ( A ) or CuOOH-incubated ( B ) “H-heme-labeled enzyme were subjected to HPLC as detailed under “Materials and Methods,” except that the incuba- tion (I ml) contained 10 nmol of P450 2B1. Peptide fractions were monitored at 220 nm and at 400 nm (insets), col- lected and their radioactivity determined by scintillation counting (bottornpanels). Aliquots (500 pl) of Lys-C digests of CuOOH-inactivated 3H-heme-labeled P450 2B1 (a different preparation of slightly higher specific radioactivity) ex- tracted with 2-butanone to remove resid- ual heme were also subjected to HPLC under similar conditions ( C ) . The radio- activity profiles of the corresponding na- tive 3H-heme-labeled P450 2B1 Lys-C digests (500 pl) are shown in the inset.

PI

500: 3 1 E 4 0 0 j

300: 4

200;

100:

R 220

2 0 4 0 6 0 80 100 120

2 0 0

E 1 0 0

/3 11 ,1140

3 9.0. was found to remove -20% of the 3I 3- or I4C-heme- derived radioactivity from the CuOOH-inactivated but not control P450 2B1. Such dialyzable polar products, we believe, may largely reflect heme-derived mono- (hematinic acid and methylvinylmaleimide) and/or dipyrrolic (propentdyopents) fragments, known coproducts of CuOOH-mediated P450 heme degradation (6, 13).

Lys-C Peptide Mapping of P450 2BI“When CuOOH-in- activated and native P450 2Bl preparations were subjected to Lys-C digestion, followed by reversed-phase HPLC of the digest, the prototypic peptide profiles depicted in Fig. 2, were obtained. Comparison of the two peptide profiles reveals that CuOOH inactivation of P450 2B1 produces minor changes such as the attenuation of the peaks eluting at 67,81, and 89 min. The most salient and consistent feature was the poor resolution of the peptides eluting between 110 and 130 min (Fig. 2, A and B). The corresponding peptide profiles detected at the iron-protoporphyrin characteristic 400-nrn absorbance are shown in the insets. While a single peak eluting at 56 min was detected in the native preparation, no such peak was detected after CuOOH inactivation of P450 2B1 (Fig. 2, insets). Because synthetic hemin also elutes at -56 rnin (with appreciable tailing), these findings merely indicate that the native P450 2B1 preparation retained substantial amounts of its prosthetic moiety, whereas the CuOOH-inactivated P450 2B1 was devoid of it.

Retention time (min)

HPLC Separation of CuOOH-inactivated 3H- or 14C-Heme- labeled 2Bl Peptides-To identify the specific apoprotein site(s) modified, purified 3H- or 14C-heme-labeled P450 2B1 incubated with CuOOH was subjected to Lys-C digestion as well as reversed-phase HPLC separation essentially under the conditions described above (except for a slightly shorter linear gradient period), and the radioactivity of individual HPLC fractions was quantitated by scintillation counting (Fig. 3B). Nan-CuOOH-incubated 3H- or 14C-heme-labeled native P450 2B1 served as its parallel control and was subjected to iden- tical procedures (Fig. 3A). The radioactive peak (~53-54 min; Fig. 3, A and B ) corresponding to heme in control P450 2B1 preparations was considerably diminished after CuOOH-me- diated inactivation of P450 2B1. Extraction of the native P450 2B1 with acidic butanone (which is known to extract heme from hemoproteins (27)) considerably attenuated this radioactive peak (Fig. 3C, inset), while completely abolishing its corresponding 400 nm-absorbance (not shown). The radio- activity profile of the reversed-phase HPLC of Lys-C digests from CuOOH-inactivated 2B1 (Fig. 3B) revealed, in addition to appreciable radioactivity in the void peak, several peptide fractions with considerably higher radioactivity than that in corresponding native 2B1 fractions (Fig. 3A). These peaks were clearly detected after butanone extraction of the CuOOH-inactivated 2B1 (Fig. 3C).

Of all these radioactive peaks, two major fractions (L1 and

Heme-modified P450 2Bl Active Site Peptide 63

R 220

2 0 4 0 60 80 100 120 T i m e ( m i n . 1

FIG. 3"continued

2 0 0

E, 1 0 0

1

L2) containing 3H- or '*C-labeled peptides could be singled out in the CuOOH-inactivated P450 2B1 chromatogram: L1, eluting between 72 and 74 min (with the shorter gradient), and L2, eluting between 110 and 130 min (Fig. 3, B and C). The peptide fractions corresponding to L1 in the native 2B1 chromatogram had relatively little detectable radioactivity after butanone extraction: whereas the fractions correspond- ing to the L2 peak exhibited a low level of radioactivity (Fig. 3A) that was only slightly attenuated by butanone extraction (Fig. 3C, inset):

The L1 peak (fractions eluting at 72-74 min; Fig. 3B) contained the highest specific radioactivity and after Tricine- SDS-PAGE, yielded three peptide bands ( M , = 3,500 (P l ) , 5,000 (P2), and 7,000 (P3)) (Fig. 4). Although all the three bands were radioactive, the yield and the corresponding ra- dioactivity of P1 were higher than that of P2 or P 3 (Fig. 4). Further confirmation that the radioactivity was intrinsically

The radioactivity detected in the corresponding native P450 2B1 fraction is thus largely due to the tailing of the radioactive heme eluting a t 56 min, since it is largely attenuated by 2-butanone extrac- tion. Cold hemin also yields a similar chromatographic tailing effect. However, the persistence of small, albeit detectable, radioactive peak fractions 2 7 4 and 120-126 min in the native P450 2B1 chromatogram after the butanone extraction (largely detected when rats are injected with large doses of highly radiolabeled ALA, Fig. 3C, inset), leads us t o suspect that they may indeed reflect the previously noted low extent of endogenous heme modification (21).

2 0 4 0 i

6 0 a o 1 0 0 1 2 0 1 4 0

Rctcnllon time (min)

associated with peak L1 was obtained when the Lys-C digests were subjected to further pronase or thermolysin digestion, which attenuated its spectroscopically monitored peak height and corresponding radioactivity in parallel (not shown).

The L2 radioactive peak was associated with the broad spectrally detectable peak (eluting between 110 and 130 min) observed after P450 2B1 inactivation. Tricine-SDS-PAGE of the L2 peak material yielded a smear rather than any distinct peptide bands, and once collected off the HPLC column and dried, the L2 peak could not be easily redissolved. I t was also largely unaffected by further digestion with either thermolysin or Staphylococcus aureus V8 protease and only minimally attenuated by pronase. This combined behavior together with its poor chromatographic and electrophoretic resolution, leads us to suspect that peak L2 might largely consist of apoP450 peptides cross-linked by heme fragments: a possibility under current examination.

Structural Characterization of Peak LI Derived from CuOOH-inactivated P450 2B1 by Amino Acid Sequencing and Electrospray Mass Spectrometry-The L1 peptide bands were transferred electrophoretically onto a ProBlott membrane, and each peptide band was sliced and individually subjected

Alternatively, such poor resolution of these peptides could be due to heme-mediated alkylation of a lysine residue or residues in its immediate vicinity and consequent impediment to Lys-C digestion.

64

C

6011

4 O(1

a u E

20a

Heme-modified P450 2B1 Active Site Peptide

$0 4'0 1 60 i I

8.0 m . 00 ll

1 1 i u I L

140

RETENTION TIME (MINI FIG. 3"continuecl

to amino acid sequence analysis. Such sequencing6 of the first 13 N-terminal residues of P1 revealed the sequence RICLGE- GIARNEL, corresponding to residues 434-446 of the reported 2B1 sequence (25). N-terminal amino acid sequencing of P2 revealed the sequence SEAFM, which corresponds to the Lys- C-induced cleavage between residues Lys4** and Ser423 of P450 2B1. The apparent molecular mass on SDS-PAGE of P2 is fully consistent with the generation of a peptide containing residues 423-466 (Fig. 4). Similar analysis of P3 revealed multiple N termini including RICLG and SEAF, thereby revealing incompletely proteolyzed peptide fragments, a find- ing consistent both with its radioactivity as well as the relative molecular mass on SDS-PAGE. It is noteworthy that in spite of the differential molecular masses and amino acid sequences, their elution as a single peak reveals that the hydrophobicity of the three peptide peaks is similar and largely determines their chromatographic behavior in this system.

Electrospray mass spectrometric analysis of L1 peptides yielded the molecular mass of 3771 f 1 Da. Since the theo- retical average mass of the predicted P1 peptide (434-466)

The amino acid yields (pmol) of P1 peptide sequenced off the

E (87); G (60); I (85); A (97); R (86); N (57); E (49); L (45). No other ProBlott membrane, were as follows: R (402); I (146); L (117); G (90);

amino acid peak constituting 210% of the first value was detected at each cycle.

derived from the Lys-C-induced cleavage a t L Y S ' ~ ~ and Lys4= is 3721.99, the additional 49 f 1 mass units apparently are contributed by the alkylating heme fragment.7

Conclusions and Implications-The heme-alkylated P450 2B1 sequence RICLGEGIARNEL not only contains Cys"', the conserved residue that is believed to provide the SH- ligand for heme, but also other highly conserved Leu Gly ( G ~ Y ~ ~ ' ) , and Arg (Arg443) residues and therefore corre- sponds to the heme-sandwiching helix L of P450cam. Al- though the precise amino acid residue that is modified remains to be identified: to our knowledge, this is the first report that localizes the site of heme alkylation of CuOOH-inactivated P450 2B1 to its active site. This finding thus rationalizes the failure of either GSH or albumin to attenuate heme-apoP450

' The possibility that these additional 49 & 1 atomic mass units are due to either an amino acid difference in this particular P450 2B1 sequence from that reported (25) and/or posttranslational modifica- tions (glycosylation, phosphorylation), is minimized by the finding that the 3H radioactivity of the P1 peptide from CuOOH-inactivated P450 2B1 is 2-3-fold higher than that of an equivalent amount of the corresponding P1 peptide from the native P450 2B1.

Although the amino acid composition of the P1 peptide was consistent with its sequence analyses, the typically low recoveries of His, Arg, and Lys extracted from membrane blotted peptides pre- cluded definitive identification of the amino acid(s) modified by the alkylating heme-derived species.

Heme-modified P450 2Rl Active Site Peptide 65

PEAK* SlDS KDa

17

- 14.4

- 10.6 CPM PHI - 8.1

112 - 6.2

221 2.5 152

FIG. 4. Tricine-SDS PAGE of L1 peak obtained from CuOOH-inactivated 3H-heme-labeled P450 2B1. The peak elut- ing at -72-74 min (Fig. 3) was collected and subjected to Tricine- SDS-PAGE, with CNBr-digests of myoglobin (Sigma) as molecular weight markers. Each of the peptide bands was excised, solubilized, and its radioactivity determined by scintillation counting. The radio- activity of each peptide band is indicated after injection of an aliquot of L1 (-700 cpm). Additionally, the separated peptides were trans- ferred electrophoretically to ProBlott membrane, the membranes stained to localize the proteins, and then each band was excised and subjected to amino acid sequencing or amino acid composition. P1, P2, and P3 refer to peptides corresponding to molecular masses of -3500, 5000, and 7000, respectively.

covalent binding on inclusion in the inactivation system (6, 13, 23), thereby legitimizing the use of such indirect ap- proaches for probing active site events of this nature.

Peroxidative inactivation of myoglobin and myoglobin-me- diated reductive metabolism of CCl, and CBrC13 have been reported to respectively entail heme-alkylation to TyrIo3 and Hisg3 residues at its active site (27-29). The CC12-heme-Hisg3 ligation is to the a-carbon of the heme-pyrrole ring I vinyl (28, 29), whereas that of herne-Tyrlo3 apparently occurs a t either the a- or P-meso-bridge (27). However, in both these instances, heme-mediated alkylation could be monitored a t 400 nm, thereby revealing that not only is the alkylating species the intact heme moiety, but that inactivation of these hemoproteins apparently does not disrupt the heme electronic rr conjugation system. In contrast, heme alkylation of apo- P450s by CuOOH, DDEP, or spironolactone destroys the heme moiety beyond spectroscopic recognition. Indeed, the presence of additional radiolabeled peptide fractions in the Lys-C digests of CuOOH-inactivated P450 2B1 reveals that several other P450 2B1 peptides may also be modified. Ap- parently, CuOOH inactivation converts the P450 2B1 heme moiety into fragments that attack several accessible amino acid residues. Whether such heme dismemberment occurs during catalysis or during post catalytic processing is unclear. Nevertheless, in addition to the several possible alkylating heme species, such heme destruction is associated with sig- nificant generation of soluble heme-derived dipyrrolic and monopyrrolic products along with HCOOH (6, 23, 24). With the possible exception of HCOOH (or a relevant precursor): none of the other known heme derivatives could account for the 49 f 1 mass units of the alkylating heme species detected in the mass spectrometric analyses of the alkylated P450 2B1 peptide. The observation that the modified P1 peptide was radioactive, irrespective of whether P450 2B1-heme was la-

' Although HCOOH is also a major product of heme destruction by HZ02 (30), it arises from one of the four methene bridges that are not labeled by [4-I4C]ALA, and thus unlikely to be the source of the alkylating heme fragments.

beled by either [4-I4C]- or [3,5-'H]ALA, limits the origin of its alkylating heme species to those tetrapyrrolic sites con- taining both [-l4C]ALA-derived carbons as well as [3,5-3H] ALA hydrogen atoms i.e. the two vinyl substituted C2 and C4 atoms of pyrrolic rings A and B, and the two propionic acid side chain bearing C6 and C7 atoms of rings C and D. Unfortunately, the size of the P1 peptide (33 amino acids) has precluded direct structural characterization of its heme- derived modifier by tandem mass spectrometric analysis. Thus the precise chemical structure of this fragment is cur- rently unknown and attempts to examine it after further size reduction of the P1 peptide are currently in progress. If drug- mediated heme alkylation of apoP450s 3A and 2E1 proceeds through a similar catalytic trajectory as that of CuOOH- inactivated P450 2B1, then by inference, it may also occur at their active site and qualify as truly suicidal in character. Studies are currently in progress to explore this particular possibility.

Acknowledgments-We sincerely thank Jane Presas for the assist- ance in the purification of P450 2B1. We also gratefully acknowledge Ralph Reid, University of California, San Francisco Biomolecular Resource Facility for the protein sequencing and valuable discussions, as well as Jack Presley, Protein Structure Laboratory (University of California, Davis) for the amino acid composition and discussions.

REFERENCES 1.

2.

3. 4.

5.

6.

7.

8.

9.

10.

11. 12.

13. 14. 15.

16.

17.

18.

19.

20. 21.

22.

23. 24.

25.

26. 27.

28.

29.

30.

Ortiz de Montellano, P. R. (1986) in Cytochrome P-450, Structure, Mech- anism & Biochemistry (Ortiz de Montellano, P. R., ed) pp. 217-271,

Guengerich, F. P. (1987) in Mammalian Cytochromes P-450, Vols. I and 11, Plenum Press, New York

Gonzalez, F. J. (1989) Pharmacol. Rev. 40,243-288 CRC Press, Boca Raton, FL

Ortiz de Montellano, P. R., and Reich, N. 0. (1986) in Cytochrome P-450, Structure, Mechanism & Biochemlstry (Ortiz de Montellano, P. R., ed)

Sugiyama, K., Yao, K., Rettie, A. E., and Correia, M. A. (1989) Chem. Res. pp. 273-314, Plenum Press, New York

Correia, M. A, Decker, C., Sugiyama, K., Caldera, P., Bornheim, L. M., Toxicol. 2,400-410

Wrighton, S. A,, Rettie, A. E., and Trager, W. F. (1987) Arch. Biochem. Biophys. 258,436-451

Lunetta, J. M., Sugiyama, K., and Correia, M. A. (1989) Mol. Pharmacol. 35, 10-17

Decker. C. J.. Rashed. M.. Baillie, T. A,. Maltbv. D., and Correia, M. A. (1989) Biochemistry'28,'5128-5i36

Bornheim, L. M., Underwood, M. C., Caldera, P., Rettie, A. E., Trager, W. F., Wrighton, S. A,, and Correia, M. A. (1987) Mol. Pharmacol. 32, 299- m

Decker, C., Sugiyama, K., Underwood, M., and Correia, M. A. (1986)

Osawa, Y., and Pohl, L. R. (1989) Chem. Res. Toxicol. 2, 131-141 Davles, H. S., Britt, S. G., and Pohl, L. R. (1986) Arch. Biochem. Biophys.

Guengerich, F. P. (1986) Biochem. Biophys. Res. Commun. 138,193-198 Halpert, J. (1981) Biochem. Pharmacol. 30,875-881 Halpert, J., Naslund, B., and Betner, I. (1983) Mol. Pharmucol. 23 , 445-

Biochem. Biophys. Res. Commun. 136,1162-1169

244,387- 352

CaJacob, C. A:. Chan, W. K., Shephard, E., and Ortiz de Montellano, P. R.

Poulos. T. L.. Finzel, B. C.. and Howard, A. J. (1987) J. Mol. Biol. 195,

4.52

(1988) J. Blof. Chem. 263, 18640-18649

Poulos, ,T. L. (1986) in Cytochrome P-450, Structure, Mechanism & Bio- 687-700

chemlstry (Ort~z de Montellano, P. R., ed) pp. 505-523, Plenum Press, New York

Poulos, T. L. (1988) in Heme Proteins (Eichorn, G . L., and Marzielli, L.

Nelson, D. R., and Strobel, H. W. (1988) J. Biof. Chem. 263,6038-6050 Correia, M. A., Sugiyama, K., and Yao, K. (1989) Dry: Metab. Rev. 20,

G., eds) pp. 1-36, Elsevier, New York

Tierney, D. J., Haas, A. L., and Koop, D. R. (1992) Arch. Biochem. Biophys.

Guengerich, F. P. (1978) Biochemist 17,3633-3639 Schaefer, W. H., Harris, T. M., andxuengerich, F. P. (1985) Biochemistry

Fujii-Kuryama, Y., Mizukami, Y., Kawajiri, K., Sogawa, K., and Mura-

Waxman, D. J., and Walsh, C. (1982) J. Biof. Chem. 257,10446-10457 Catalano, C. E., Choe, Y. S., and Ortiz de Montellano, P. R. (1989) J . Biof.

Osawa, Y., Martin, B. M., Griffin, P. R., Yates, J. R., 111, Shabanowitz, J., Chem. 264, 10534-10541

Hunt, D. F:. Murphy, A. C., Chen, L., Cotter, R. J., and Pohl, L. R. (1990) J . B~ol. Chem. 2 6 5 , 10340-10346

Osawa, Y., Highet, R. J. , Bax, A., and Pohl, L. R. (1991) J. Biof. Chem. 266,3208-3214

Cantoni, L., Gibbs, A. H., and De Matteis, F. (1981) Int. J. Biochem. 13, 823-830

615-628

293.9-16

2 4 , 3254-3263

matsu, M. (1982) Proc. Natl. Acad. Sci. U. S. A. 79, 2793-2797