Identification of the Heme Adduct and an Active Site Peptide Modified

Apr 16, 1996 - A [14C]SB-modified 2B1 peptide was thus isolated, purified, electrotransferred onto a poly(vinylidene) membrane, and identified by micr...
1 downloads 10 Views 179KB Size
Download PDF
614

Chem. Res. Toxicol. 1996, 9, 614-622

Identification of the Heme Adduct and an Active Site Peptide Modified during Mechanism-Based Inactivation of Rat Liver Cytochrome P450 2B1 by Secobarbital† Kan He,‡ Arnold M. Falick,§,| Baili Chen,‡ Fredrik Nilsson,§,⊥ and Maria Almira Correia*,‡,§,# Departments of Cellular and Molecular Pharmacology, Pharmacy, and Pharmaceutical Chemistry and the Liver Center, University of California, San Francisco, California 94143 Received October 17, 1995X

The olefinic barbiturate secobarbital (SB) is a sedative hypnotic known to be a relatively selective mechanism-based inactivator of rat liver cytochrome P450 2B1. Previous studies have demonstrated that such inactivation results in prosthetic heme destruction and irreversible drug-induced protein modification, events most likely triggered by P450 2B1-dependent oxidative activation of the olefinic π-bond. However, the precise structure of the SB-modified heme and/or the protein site targeted for attack remained to be elucidated. We have now isolated the SB-heme adduct from P450 2B1 inactivated by [14C]SB in a functionally reconstituted system and structurally characterized it by electronic absorption spectroscopy and tandem collision-induced dissociation (CID), matrix-assisted laser desorption ionization on time of flight (MALDI-TOF), and liquid secondary ion mass spectrometry in the positive mode (+LSIMS) as the N-(5-(2-hydroxypropyl)-5-(1-methylbutyl)barbituric acid)protoporphyrin IX adduct. The [14C]SB-modified 2B1 protein has also been isolated from similar inactivation systems and subjected to lysyl endopeptidase C (Lys-C) digestion and HPLC-peptide mapping. A [14C]SB-modified 2B1 peptide was thus isolated, purified, electrotransferred onto a poly(vinylidene) membrane, and identified by micro Edman degradation of its first N-terminal 17 residues (S277NH(H)TEFH(H)ENLMISLL293) as the Lys-C peptide domain comprised of amino acids 277-323. This peptide thus includes the peptide domain corresponding to the distal helix I of P450 101, a region highly conserved through evolution, and which is known not only to flank the heme moiety but also to intimately contact the substrates. This finding thus suggests that SB-induced protein modification of P450 2B1 also occurs at the active site and, together with heme N-alkylation, contributes to the SB-induced mechanism-based inactivation of P450 2B1.

Introduction The superfamily of hemoproteins termed cytochromes P450 (P450s)1 consists of a highly versatile class of biological catalysts that convert a vast array of chemically diverse endobiotic and xenobiotic substrates to their corresponding oxidized/hydroxylated products. Conventionally, P450s catalyze various oxidation reactions including the insertion of an activated atom of molecular oxygen into various bonds (C-H, N-H, etc.) to yield the corresponding hydroxylated product, or into a π-bond of an olefin, aromatic ring, or other unsaturated species to † Some of the data in this paper were presented in abstract form: Proceedings, 10th International Symposium on Microsomes & Drug Oxidations, Toronto, 1994, p 558, and FASEB J. 9, A1488, 1995. * Address correspondence to this author. ‡ Department of Cellular and Molecular Pharmacology and the Liver Center. § Department of Pharmaceutical Chemistry. | Present address: PerSeptive Biosystems, West Coast Technical Center, 871 Dubuque Ave., South San Francisco, CA 94080. ⊥ Present address: Center for Biostructure, Karolinska Institute & Novum, S-14157, Huddinge, Sweden. # Department of Pharmacy. X Abstract published in Advance ACS Abstracts, March 15, 1996. 1 Abbreviations: collision-induced dissociation (CID), cytochromes P450 (P450s), dilauroylphosphatidylcholine (DLPC), dithiothreitol (DTT), electrospray ionization mass spectrometry (ESI-MS), liquid secondary ion mass spectrometry in the positive mode (+LSIMS), matrix-assisted laser desorption ionization on time of flight (MALDITOF) mass spectrometry (MALDI-MS), secobarbital (SB), sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), poly(vinylidene difluoride) (PVDF).

0893-228x/96/2709-0614$12.00/0

give the corresponding epoxide or rearranged product (1). In most, if not all, of these instances the resulting product(s) is (are) more polar and more readily eliminated from the biological membranes and the participating P450 is left unscathed, as expected of a normal catalyst. In some instances, however, highly reactive intermediates of the chemicals formed in the course of their oxidative catalysis result in irreversibly inactivating the participating P450, in a process termed as mechanismbased or suicide inactivation of the enzyme (2-4). Three currently known mechanisms of such inactivations include the following: (i) N-alkylation of the P450 prosthetic heme moiety; (ii) direct alkylation of the apoprotein at the active site; and (iii) chemical-induced heme modification of the protein at the active site. Although, with most mechanism-based inactivators one of these inactivation mechanisms usually predominates, it is conceivable that two or all three mechanisms could operate concurrently to contribute to some extent to the overall inactivation process. This appears to be the case with the olefinic compound, secobarbital (SB). Our studies and those of others indicate that SB is a relatively selective mechanism-based inactivator of rat liver P450 2B1 (5, 6), and this inactivation process appears to partition largely between N-alkylation of the prosthetic heme and SB-mediated alkylation of the protein, with e10% of P450 destroyed being accounted by prosthetic © 1996 American Chemical Society

Modified Heme and Protein from Secobarbital-Inactivated P450

heme modification of the protein (5). Furthermore, the partitioning between P450 2B1 suicide inactivation and its catalytic oxidation of SB to the corresponding epoxide appears to be of the order of 1:16 (7). Although SB is known to convert P450 2B1 heme to the corresponding N-alkylated heme products (8, 9), to our knowledge, these adducts have never been structurally characterized. Similarly, although we have shown that SB alkylates P450 2B1 protein in a process unaffected by external nucleophiles such as bovine serum albumin (5), and thus this is very likely an active site event, the structural identity of either the SB-derived alkylating species or the SB-modified peptide remained to be elucidated. Identification and structural characterization of the SB-modified peptide would be essential to determine whether such inactivation also occurs at the active site of 2B1 and, in common with the N-alkylation of the heme, can be classified as a truly suicidal process. Furthermore, structural characterization of the SBderived alkylating species in the heme and peptide adducts would reveal whether the two inactivation mechanisms entail a common reactive intermediate. The findings detailed below provide evidence that the 2B1 peptide modified by SB is an active site peptide and that the SB-heme adduct is an N-hydroxy SB-porphyrin adduct.

Materials and Methods Purification of Cytochrome P450 2B1 and Cytochrome P450-NADPH Reductase. P450 2B1 and NADPH-P450 reductase were isolated and purified from liver microsomes of phenobarbital (PB)-induced male Sprague-Dawley rats by the method of Waxman and Walsh (6) and by the method of Shephard et al. (10), respectively. Synthesis of [2-14C]Secobarbital. [2-14C]SB was synthesized with minor modifications and its structure confirmed by 1H NMR and mass spectrometric analyses as described previously (5). The specific activity of 14C-SB was 0.34 µCi/µmol, and its radiochemical purity was >97%. [2-14C]SB-Mediated Inactivation and Covalent Binding of P450 2B1. SB-mediated inactivation of P450 2B1 and corresponding SB-induced covalent binding were examined in liver microsomes from PB-pretreated rats according to the procedures described previously (5). Purified P450 2B1 was inactivated by incubating the enzyme (2 nmol) with NADPHP450 reductase (2 nmol; 6000 nmol of cytochrome c reduced min-1 nmol-1), DLPC (80 nmol), EDTA (2 µmol), catalase (240 U), NADPH (2 µmol), [14C]SB (2 µmol, 0.68 µCi), GSH (2 mM), and 20% glycerol in 2 mL of potassium phosphate buffer (0.1 M, pH 7.4) at 37 °C for 30 min. Corresponding control incubations were carried out in the absence of NADPH and used as reference. P450 content was determined as described previously (11). Irreversible [14C]SB binding to the 2B1 protein was determined either after precipitation of the protein by MeOH/ 5% H2SO4, followed by extensive organic solvent washes, as previously (5), or after physical separation of P450 2B1 (after inactivation) by preparative SDS-PAGE from the other components of the reconstituted system, isolation, and dialysis (as described below). Assays of Pentoxyresorufin O-Deethylase (PROD) and Testosterone Hydroxylase Activities. PROD activity was assayed essentially by the method of Lubet et al. (12). For its assay, a 50 µL aliquot of the SB inactivation mixture was taken at the end of the incubation and added directly to cuvettes containing NADPH-P450 reductase (50 pmol), DLPC (20 nmol), MgCl2 (2.5 µmol), pentoxyresorufin (10 nmol), and Tris-HCl buffer (50 mM, pH 7.5) in a total volume of 0.5 mL. The rate of fluorescence increase with time was recorded over 5 min, at excitation and emission wavelengths of 526 and 586 nm, respectively, at 32 °C. Testosterone hydroxylase activity was

Chem. Res. Toxicol., Vol. 9, No. 3, 1996 615

similarly assayed using a 50 µL aliquot of the SB inactivation mixture taken at the end of the incubation with the unlabeled SB, added to a reaction mixture consisting of [14C]testosterone (0.2 µmol), P450 reductase (50 pmol), DLPC (20 nmol), isocitrate (2.5 µmol), MgCl2 (2.5 µmol), NADPH (1 µmol), and isocitrate dehydrogenase (0.25 unit) in potassium phosphate buffer containing 20% v/v glycerol and 0.5 mM EDTA in a total volume of 0.5 mL, and incubated at 37 °C for 10 min. Hydroxylated [14C]testosterone metabolites were extracted and isolated as previously described (13). Isolation of P450 2B1 from the Reconstituted System and SDS Removal. After functional reconstitution and inactivation, P450 2B1 was reisolated by preparative SDS-PAGE (Prep. Cell, Bio-Rad). For this purpose, the incubation mixture (2 mL) was mixed with 0.5 mL of a sample buffer containing 0.3 M Tris-HCl, pH 6.8, 5% SDS, 50% glycerol, 100 mM DTT, and a lane marker tracking dye (Pierce). The fractions containing P450 2B1 were identified by mini-SDS-PAGE, before SDS was removed by precipitation with ice-cold 0.1 M NaCl. After centrifugation at 2000 rpm at 4 °C for 10 min, the supernatant was first dialyzed against 0.05 M Tris-HCl, pH 7.4, containing 0.5% cholate for at least 4 h at room temperature, and then concentrated by ultrafiltration (YM 10, Amicon membrane). It was then dialyzed against 0.1 M Tris-HCl, pH 9, containing 30% glycerol. Residual SDS content was determined by the method of Sokoloff and Frigon (14). P450 2B1 Digestion by Lysyl Endopeptidase (Lys-C) and Peptide Mapping by HPLC. After denaturation with urea (8 M) in 0.1 M Tris-HCl buffer (pH 9.0) for 30 min at 60 °C, P450 2B1 was digested with Lys-C (Wako BioProducts) at a molar ratio of 40:1, overnight, as previously described (15). The resulting peptides were separated on a C4 reversed-phase column (Applied Biosystems, BU 300, 2.1 × 220 mm, 7 µm) with a solvent system of H2O/0.1% TFA (A) and 90% acetonitrile/ 0.1% TFA (B) by gradient elution from 0% to 80% over 120 min. The radioactivity was monitored by sequential scintillation counting of 2 min fractions. Electrophoretic and Amino Acid Sequencing Analysis of HPLC Fractions Containing the SB-Modified Peptide. The Lys-C peptide fraction containing the highest radioactivity was resolved by Tricine-SDS-PAGE according to the method of Schager and van Jagow (16). After electrophoresis, the peptide was electrotransferred onto a PVDF membrane (ProBlot). The PVDF membrane-bound peptide was sequenced by the micro Edman degradation method, as described previously (15). Mass Spectrometric Analysis of SB-Modified Peptide. Analysis of the HPLC fraction was carried out by reversed-phase HPLC/electrospray mass spectrometry. An Applied Biosystems dual syringe pump was used to provide the mobile phase, consisting of buffer A (0.1% aqueous formic acid) and buffer B [a mixture of ethanol and 1-propanol (ratio 5:2), made up to 0.05% formic acid] (17). The peptides present were separated on a 1.0 × 150 mm Vydac C-4 column (5 µm) using a gradient of 10-80% B in 40 min at a flow rate of 50 µL/min. The column effluent was directed into a variable wavelength UV detector (Applied Biosystems) set at 215 nm and equipped with a high sensitivity flow cell (“Z cell”; LC Packings). The HPLC system was directly interfaced to a VG Biotech/Fisons Platform electrospray mass spectrometer by means of a short length of fused silica capillary tubing. The total column effluent was injected into the mass spectrometer, which was scanned repetitively over the m/z range 350-2000 every 5 s. CNBr Cleavage of SB-Inactivated P450 2B1 and Corresponding HPLC Peptide Mapping. After functional reconstitution, P450 2B1 was incubated with SB and NADPH as described above, and the mixture was concentrated by centrifugation in a Centricon 10 (Amicon) at 5000g for 2 h at 4 °C, followed by two washes with 50 mM phosphate buffer (pH 7.4). The sample was then concentrated to near dryness under vacuum (Speed-Vac) and resuspended in 0.2 mL of 70% TFA. A cyanogen bromide solution (20 µL, 1.0 mg/µL in 70% TFA) was added and the cleavage carried out in the dark, overnight at

616 Chem. Res. Toxicol., Vol. 9, No. 3, 1996 room temperature, as reported (18). Before HPLC analysis, the sample was concentrated to near dryness. The HPLC system was the same as that described above for Lys-C peptide mapping, except that the linear gradient of buffer B varied from 0% to 30% over 30 min, increased to 60% over the next 70 min, and finally to 90% of buffer B over 20 min. This HPLC peptide mapping of the CNBr digest yielded a radioactive fraction eluting at 68-70 min which was repurified by subjecting it to HPLC on a C8 column (Microsorb-MVTM, 5 µm, 0.46 × 25 cm) with the same solvent system [H2O/0.1% TFA (A) and 90% acetonitrile/0.1% TFA (B)] but with a linear gradient increasing from 40% to 70% over 30 min, then to 90% of buffer B over 10 min. MALDI-MS of CNBr-Cleaved Peptides. Average masses of the peptides detected in the relevant HPLC fractions were determined with a VG/Fisons TofSpec MALDI time-of-flight mass spectrometer or a PerSeptive Biosystems Voyager linear MALDI-TOF mass spectrometer (PerSeptive Biosystems, Framingham, MA) equipped with a nitrogen laser (337 nm) and operated in the linear mode. Samples were crystallized with R-cyano-4-hydroxycinnamic acid (saturated solution in 50% acetonitrile, 0.1% TFA). The mean mass for each peptide from all spectra recorded is reported. All MALDI spectra were externally calibrated using a standard peptide mixture. Isolation and Purification of SB-Heme Adducts from Rat Livers. SB-heme adducts were isolated from intact rat livers as previously (5) and purified using the procedures outlined by Ortiz de Montellano (19). Rats (N ) 10) pretreated with sodium phenobarbital (80 mg/kg, daily for 5 days) were treated with a single dose of sodium secobarbital (100 mg/kg, ip) 3 h before sacrifice. The livers were perfused with ice-cold 1.15% KCl, homogenized in a minimum volume of 1.15% KCl, and subsequently extracted with 2 L of 5% sulfuric acid/ methanol with stirring in the cold room, overnight. The SBheme adducts were extracted with an equal volume of water and methylene chloride (CH2Cl2) after removal of the liver tissue by filtration. The CH2Cl2 phase was concentrated by rotary evaporation after addition of 0.5 mL of saturated methanolic zinc acetate solution. The resulting zinc complexed modified porphyrin methyl esters were then purified by thin layer chromatography on silica gel G (2000 µm) plates using 2:1 (v/v) chloroform/acetone as the solvent. The green (red fluorescing) pigment band was extracted with acetone and repurified by HPLC on a cyano-NH2 column (Partisil PXS 10/25 PAC, 0.46 × 250 mm) using a 0-100% linear gradient of methanol into 1:1 (v/v) hexane/tetrahydrofuran over 30 min. The zinc-complexed SB-porphyrin adduct, which migrated as a single peak, was collected and concentrated by rotary evaporation. The zinc free adduct was generated by treatment with 5% sulfuric acid/ methanol and then extracted with an equal volume of water and methylene chloride. The visible electronic spectra were determined with an UV-vis SLM Aminco 2000 spectrophotometer. High-energy positive ion CID mass spectra were acquired on a Kratos Analytical Instruments Concept IIHH four-sector tandem mass spectrometer equipped with a continuous flow, liquid inlet probe, and a scanning CCD array detector. Both MS1 and MS2 were operated at 1000 resolution (M/Dm) to determine monoisotopic masses. HPLC fractions were concentrated to ≈5 µL and diluted to ≈15 µL with an aqueous 5% thioglycerol/5% acetonitrile/0.1% TFA matrix solution, prior to 3 µL/min introduction into the mass spectrometer source.

Results SB-Mediated Inactivation and Alkylation of P450 2B1. Incubation of P450 2B1 with SB in the reconstituted system resulted in an apparent ≈50% loss of the chromophore, but in complete inhibition of characteristic P450 2B1 marker PROD and 16R- and 16β-testosterone hydroxylase activities, and androstenedione formation (Table 1). More recent findings indicate that, depending on the buffer system used in the inactivation system, varying amounts of residual SB-heme adduct (absor-

He et al. Table 1. SB-Mediated Inactivation of P450 2B1-Dependent PROD and Testosterone Hydroxylase Activities testosterone hydroxylase activity -SB + NADPH +SB - NADPH +SB + NADPH

PROD

16R

16β

androstenedione

0.35 0.31 ND

0.42 0.32 ND

0.32 0.24 ND

0.25 0.22 ND

a P450 2B1 was inactivated by SB in the complete functionally reconstituted system at 37 °C for 30 min as detailed (Materials and Methods). Controls with either SB or NADPH omitted were run in parallel. Aliquots (50 µL) of the inactivation system were then used to assay for PROD or testosterone hydroxylase activity as described (Materials and Methods). PROD is expressed in nmol of resorufin formed min-1 per incubation; testosterone hydroxylase in nmol of hydroxytestosterone formed min-1 per incubation. ND, not detected.

bance ≈445 nm) may contribute to these P450 spectral measurements and thus lead to overestimation of the remaining spectrally detectable P450 content in the P450 2B1 incubations containing SB and NADPH. The molar ratio of SB irreversibly bound to the P450 2B1 initially present in the purified reconstituted inactivation system was of the order of 0.23, when determined after MeOH/ H2SO4 precipitation and extensive washes of the inactivated P450 2B1 (5). A corresponding value of 0.15 was obtained when SB was incubated with NADPH-supplemented liver microsomes from PB-treated rats, consistent with our previous report (5). Isolation of P450 2B1 from the Reconstituted System. Peptide mapping and unambiguous structural analyses of the SB-modified peptide(s) of P450 2B1 required physical separation of P450 from P450 reductase, before proteolytic digestion. We found that neither HPLC (RH1 Poros column, PerSeptive Biosystems) nor ADP-Sepharose chromatography was satisfactory in completely separating P450 2B1 from the reductase, with acceptable recoveries of the SB-modified P450. In contrast, SDS preparative electrophoresis was capable of fully separating P450 2B1 from the other coincubated proteins (reductase, catalase), albeit with the inevitable and undesirable contamination with SDS (Figure 1). To avoid probable SDS-induced interference with mass spectrometric analysis, SDS was effectively removed (99.8%) from the P450 2B1 fraction. The resulting solution containing P450 2B1 was dialyzed, yielding a protein recovery of about 75%. HPLC Peptide Mapping and Electrophoresis of Lys-C Digest of SB-Modified P450 2B1. The HPLC peptide mapping profile of the Lys-C digest of [14C]SBmodified P450 2B1 is shown (Figure 2A). A single major peak containing [14C]SB-derived radioactivity eluted at ≈78-79 min, at the relatively higher organic solvent concentration (Figure 2B). Tricine-SDS-PAGE of this peak followed by silver staining revealed a single peptide band (Mr ≈5 kDa) (Figure 3). Identification of the SB-Modified Peptide by Amino Acid Sequencing and Electrospray Mass Spectrometry. After electrophoresis, the peptide electrotransferred onto ProBlot was subjected to microEdman degradation. This yielded a single N-terminal sequence of SNH(H)TEFH(H)ENLMISLL for the first 17 amino acids identified, which corresponded to residues 277-293 of the P450 2B1 Lys-C peptide (Table 2). Preliminary electrospray mass spectrometric analysis of the corresponding HPLC fraction revealed that this

Modified Heme and Protein from Secobarbital-Inactivated P450

Chem. Res. Toxicol., Vol. 9, No. 3, 1996 617

Figure 1. Isolation of SB-inactivated P450 2B1 from the incubation mixture by preparative electrophoresis. P450 2B1 was inactivated as detailed (Materials and Methods). The incubation mixture was dissolved in sample buffer and applied to a preparative SDSPAGE cell (Materials and Methods). Five milliliter fractions (1 mL/min) were collected, and the protein was monitored by miniSDS-PAGE. Fractions containing P450 2B1 and NADPH-P450 reductase were found to elute 80-100 and 165-195 min, respectively, after the dye front. Lanes 1-6 represent aliquots of fractions eluting at 195, 180, 165, 150, 120, and 110 min after the dye front, respectively. Lanes 7 and 8 contain molecular weight standards (top to bottom, 97.4, 66, 43, 21, and 14.4 kDa) and an aliquot of the incubation mixture, respectively. Lanes 9-13 represent aliquots of fractions eluting at 100, 90, 80, 70, and 60 min after the dye front, respectively.

Figure 3. Tricine-SDS-PAGE of [14C]SB containing peptide fraction isolated by HPLC from the Lys-C digest of SBinactivated P450 2B1. The peak eluting around 78-79 min and containing most of the [14C]SB radioactivity (Figure 2) was collected and subjected to Tricine-SDS-PAGE (lane 2), along with CNBr digests of myoglobin (Sigma) as molecular weight standards (lane 1). The gels were developed with silver stain to help identify the presence, if any, of minor contaminating peptides in the peptide fraction of interest. Table 2. Amino Acid Sequence Determination of SB-Modified Peptide Generated from Lys-C Digests of SB-Inactivated P450 2B1

Figure 2. HPLC profiles of Lys-C digest of SB-modified P450 2B1 (A, top panel) and corresponding [14C]SB-derived radioactivity of the peptide fractions (B, bottom panel). Aliquots of Lys-C digest of [14C]SB-inactivated P450 2B1 (≈100 pmol) were subjected to HPLC as detailed (Materials and Methods). Peptide fractions were monitored at 214 nm, and their radioactivity was determined by scintillation counting. The inset depicts the corresponding 400 nm profile of these fractions.

peptide had an average mass of 5420.7 ( 1 Da (Figure 4), which matched the theoretical average mass of 5421.24 Da for the peptide 277-323, expected from the Lys-C cleavage of P450 2B1 (20, 21). Although faint protonated signals corresponding to the mass of the modified peptide may be extracted from the spectrum (Figure 4), a clear signal with the expected mass of ≈5675.7 Da for the SB-modified peptide has yet to be detected using either the Fison/VG Platform Bio-Q electrospray quadrupole mass spectrometer or a Perkin

cycle

amino acid

pmol

cycle

amino acid

pmol

1 2 3 4 5 6 7 8 9

S N H H? T V F H H?

5.0 >2.6 nc nc 5.5 2.7 3.9 nc nc

10 11 12 13 14 15 16 17

E N L M I S? L L

1.2 0.7 2.1 0.8 1.3 nc 0.4 >0.4

a The Lys-C peptide fraction isolated by HPLC and subjected to Tricine-SDS-PAGE was electrotransferred to a ProBlott membrane and subjected directly to 17 cycles of micro Edman degradation. “nc” refers to amino acids that were not conclusively identified, although from homology four of these are predicted to be His. Indeed, a slight signal for His could be seen in cycles 3 and 8.

Elmer/Sciex API system equipped with a prototype turbo ion spray source (courtesy of Mr. Leo Raftogianis and Dr. Ling Chen, Applied Biosystems Division, Perkin Elmer Corp.). This was also true when the SB-modified peptide

618 Chem. Res. Toxicol., Vol. 9, No. 3, 1996

He et al.

Figure 4. Electrospray mass spectrum of HPLC fraction of interest containing the SB-modified and unmodified peptides. This spectrum was obtained on-line as the fraction of interest eluted from the HPLC column. Several scans were summed over the peak. The peaks marked “An” correspond to the mass-tocharge (m/z) ratios and are due to the unmodified peptide with n protons added. The peptide mass was determined by averaging the values calculated from each of the marked peaks.

sample was subjected to matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-MS), which was successfully used to characterize the corresponding 2B1 and 2B4 peptides modified by 2-ethynylnaphthalene (2-EN) (18, 22). In all these analyses, however, the signal (≈5420.7 Da) for the parent unmodified peptide (≈70% of the SB-inactivated 2B1 peptide population) was unambiguously detectable. Since the [14C]SB radioactivity comigrates with the peptide after electrotransfer as well as on a subsequent HPLC repurification of the initial HPLC fraction, the reasons for the failure to detect a corresponding mass signal for the modified peptide are unclear. Such a failure could be due to poor recovery and thus low abundance relative to the unmodified species, suppression by the SB modification and/or the traces of residual SDS, or the potential instability of the SB-peptide adduct under conditions used for its isolation and/or mass spectrometric analysis. HPLC Mapping and Mass Spectral Characterization of CNBr-Cleaved Peptides from SB-Inactivated 2B1. We found it noteworthy that the successful MALDIMS characterization of the 2-ethynylnaphthalene (2EN)modified 2B1 and 2B4 peptides was achieved after CNBr cleavage (18). Because CNBr cleavage would considerably shorten the size of the SB peptide from 46 to 24 amino acids and correspondingly reduce its mass, we examined whether such a peptide size reduction would improve the detection of the SB-modified peptide by mass spectrometry. Purified P450 functionally reconstituted with P450 reductase and inactivated by SB, and corresponding non-inactivated P450 control incubations without NADPH, as detailed above, were subjected to CNBr cleavage followed by HPLC mapping and [14C]SB radioactivity profiling, as previously with the Lys-C digests (Figure 5A,B). Although the HPLC peptide resolution of the CNBr digest was not as good as that of the Lys-C digest, nevertheless a distinct radioactive peptide fraction eluting at ≈70 min was detected (Figure 5C). This fraction was radiolabeled to a considerably higher extent than the Lys-C peptide and, when analyzed by MALDIMS, was found to contain several peptides. When this radioactive fraction was further purified by a subsequent HPLC on a reversed-phase C-8 column (Figure 6A), a single radioactive fraction eluting at ≈29 min was

Figure 5. HPLC reversed-phase separation of the peptides generated by CNBr cleavage of [14C]SB inactivated P450 2B1. For this purpose, the entire inactivation mixture without isolation of the [14C]SB-inactivated P450 2B1 was subjected to CNBr cleavage. The peptides were monitored at 214 nm (A, top panel, SB + NADPH; B, middle panel, SB - NADPH), and their radioactivity was determined by scintillation counting (C, bottom panel). The insets in the top and middle panels depict the corresponding 400 nm absorbance of the fractions.

obtained (Figure 6B), which, when analyzed by MALDIMS, yielded a mass (MH+) of 817.8 Da (Figure 6C). Since this mass could not be assigned either to any known CNBr-cleaved 2B1 peptide fragments or SB-derived metabolites, we suspected that this product might actually be the SB-heme adduct (562.6 Da for the porphyrin and 254 Da for the OH-SB species). This suspicion was confirmed when the 400 nm absorbance of the HPLC peptide profile was obtained (insets, Figures 5A and 6A). No such 400 nm absorbing radioactive peptide peak was found in 2B1+SB incubations devoid of NADPH (inset, Figure 5B), or in the Lys-C digest of SB-inactivated 2B1 (inset, Figure 2A). In the former case, no SB-heme adducts are expected in the absence of NADPH, and in the latter case, although SB-heme adducts are formed, they were removed from the SB-modified P450 2B1 by the preparative SDS-PAGE step that precedes Lys-C digestion. Their absence was confirmed by mass spectrometric analyses of the radiolabeled Lys-C peptide

Modified Heme and Protein from Secobarbital-Inactivated P450

Chem. Res. Toxicol., Vol. 9, No. 3, 1996 619

Figure 7. Electronic absorption spectra of the zinc complex of the SB-modified protoporphyrin IX dimethyl ester. The UVvis absorbance of N-alkylporphyrins, isolated and purified from livers of PB-pretreated rats given SB, is depicted. The characteristic Soret absorbance at 430.5 nm of the Zn-complexed N-alkylporphyrin is shown (peak 1). The absorbance scale for the 500-700 nm range was expanded 10×, in order to magnify the characteristic post-Soret triplet peak absorbances (peaks 2-4) detected at 544.2, 586.6, and 630.7 nm.

porphyrin adduct (Figure 7). Mass spectrometric analyses (+LSIMS) of the demetalated species gave a mass of 845.4 Da (MH+), which is consistent with a structure stoichiometrically assembled from protoporphyrin IX (dimethyl ester, Mr ) 591), SB (Mr ) 237), and a hydroxyl group (Mr ) 17). Furthermore, this proposed composition (dimethyl protoporphyrin plus hydroxylated SB) was confirmed by tandem CID mass spectrometry (Figure 8). Figure 6. Repurification and mass spectrometric characterization of the radioactive fraction isolated from C4-HPLC peptide mapping of CNBr digest of P450 2B1. The peptide fraction eluting at 68-70 min (Figure 5) was subjected to further reversed-phase C8-HPLC purification (A, top panel). The inset depicts the corresponding 400 nm absorbance of the eluates. The corresponding radioactivity of these eluates is shown (B, middle panel). The peak fraction eluting at 29 min, exhibiting a strong 400 nm absorbance as well as containing the bulk of the [14C]SB radioactivity, was collected and subjected to MALDI mass spectrometric analyses as detailed (Materials and Methods) (C, bottom panel).

fraction which failed to detect any mass corresponding to the SB-heme adduct. Furthermore, no peptides could be detected in this repurified radioactive fraction when it was subjected to Tricine-SDS-PAGE and silver staining (not shown), indicating that the radioactive fraction isolated from the CNBr digests exclusively contained the SB-heme adduct. The structural identity of this N-alkylated heme-SB adduct was confirmed, when authentic adducts obtained from SB-treated rat livers were isolated (after their conversion to porphyrin dimethyl esters), purified, and characterized by their characteristic absorption spectra of the Zn complex and demetalated species as well as LSIMS analyses (not shown). Thus, typical visible spectra of N-alkylporphyrins (green pigments) were observed with absorbances of 416, 512, 546, 597, and 652 nm for the metal free form (not shown), and 430, 546, 589, and 630 nm for the zinc-complexed form of the SB-

Discussion The above findings with the olefinic compound SB reveal that the drug completely inactivates P450 2B1 functionally with partial loss of the heme chromophore. Isolation of the N-alkylated SB-heme adduct and of the modified P450 2B1 protein reveals that the compound indeed partitions largely between heme N-alkylation, 2B1 protein modification, and epoxidation. A negligible fraction of the prosthetic heme modifies the protein and also contributes to SB-mediated 2B1 inactivation (5). The N-modified porphyrin has been isolated and characterized by electronic absorption spectroscopy and mass spectrometry as N-(5-(2-hydroxypropyl)-5-(1-methylbutyl)barbituric acid)protoporphyrin IX, the N-adduct of a porphyrin pyrrole and hydroxy-SB. The P450 2B1 Lys-C peptide that is modified by the drug has also been isolated and sequenced by micro Edman degradation. The first 17 residues (SNH(H)TEFH(H)ENLMISLL) identify this peptide domain as that spanning residues 277-293 of the 2B1 protein. This, together with the apparent mass of the SB-modified peptide on Tricine-SDS-PAGE as well as the mass of the corresponding unmodified peptide determined by both ESI-MS and MALDI-MS, is consistent with the masses of the expected Lys-C peptide comprising residues 277-323 (20, 21). This 2B1 region aligns with the peptides corresponding to the distal helices I of P450s 101 and 102 (23, 24). This observation

620 Chem. Res. Toxicol., Vol. 9, No. 3, 1996

He et al.

Figure 8. CID spectrum of the demetalated SB-modified protoporphyrin dimethyl ester. The precursor peak at m/z 845.8 is due to the intact modified dimethyl ester (MH+ ion), while the fragment peak at m/z 591.5 corresponds to protoporphyrin IX dimethyl ester from which the expected hydroxylated SB metabolite has been lost (difference ≈ 254.3 Da).

not only confirms that this 2B1 peptide indeed contacts the substrates, but that the inactivation occurs at the active site, and thus fulfils an essential criterion of mechanism-based inactivations. Although, the specific residue(s) alkylated has(ve) not been identified, several nucleophilic residues, including histidine, serine, threonine, and tyrosine, are present in the modified peptide and may be likely candidates. The structure of the SBderived alkylating species is also unclear since, as discussed earlier, for unknown reasons the SB-modified peptide has remained recalcitrant to mass spectrometric characterization, in spite of exhaustive attempts using a variety of different approaches. Interestingly, this SB-modified peptide is identical to that isolated from 2-EN-inactivated rat P450 2B1 and analogous to that of 2-EN-inactivated rabbit P450 2B4 (18, 20). Studies with radiolabeled 2EN and P450 2B1 have led, after CNBr cleavage, to the isolation of a radiolabeled peptide (ISLLSLFFAGTETSSTTLRYGFLLM) comprising residues 290-314 of the protein (18). This is also the peptide region modified after 2-EN inactivation of 2B4 (18), revealing that the peptides alkylated in both cases are similar, corresponding to the highly conserved substrate and heme contacting distal I helix of P450 101, and therefore truly representative of the P450 active site. On the other hand, peptide mapping and amino acid sequence analyses of the corresponding radiolabeled peptides from rat and rabbit P450s 1A2 revealed that 2EN modifies peptides comprised of amino acid residues 67-78 and 175-184, respectively (25). However, the instability of the adduct did not permit definitive identification of the specific residue(s) that was (were) modified. Sequence alignment of these peptides with those of P450 101, for which X-ray crystallographic data exist (23, 26), maps these 1A2 peptides to the corresponding helices A and D of P450 101. Accordingly, the rat 1A2 peptide modified by 2EN apparently contains putative substrate-accessible peptide domains that topologically differ from those of P450 2B1 (or 2B4), modified by either 2EN or SB. It is noteworthy that the conditions for CNBr cleavage that led to the successful isolation and mass spectrometric characterization of the corresponding 2-EN-modified peptide (presumably, a potentially unstable ketenederived adduct) (18) were unsuccessful for the isolation and detection of the SB-modified peptide. They led instead to the isolation and detection of the SB-heme adduct. To our knowledge, this is the first report of the isolation and characterization of such SB-heme adducts from in vitro purified reconstituted P450 inactivation systems. Furthermore, this heme adduct was not detected in Lys-C peptide digests of SB-inactivated 2B1, because it is effectively removed from the inactivated P450 by the preparative SDS-PAGE step that precedes

Scheme 1. Proposed Scheme for the Formation of the Reactive SB Intermediates in SB-Mediated P450 2B1 Inactivation

Lys-C digestion. Since, most likely, the formation of both the P450 2B1 heme and protein adducts of SB entails a common reactive SB-derived alkylating species, it is instructive that the two adducts should differ so markedly in their stability under the highly acidic conditions required for CNBr cleavage. Such differential stability of the two adducts may be related to the greater intrinsic bond strength of the hydroxy-SB-pyrrole nitrogen(s) of the heme, than that of the hydroxy-SB and the nucleophilic amino acid residues targeted, which as discussed above, include serine, threonine, and tyrosine residues. The possibility that the olefinic π-bond of SB is activated to a common reactive intermediate that alkylates both P450 2B1 heme and protein is compelling, in view of the fact that, unlike SB, its saturated analog, 5-ethyl-5-(1-methylbutyl)barbituric acid, fails to inactivate P450 2B1 (5, 8). This strongly implicates the olefinic π-bond as the enabling functionality in the compound (Scheme 1). The findings that SB forms the corresponding epoxide (7) and that the porphyrin N-alkylating species is hydroxy-SB strongly argue for the oxidative activation of this olefinic π-bond by P450 2B1. However, epoxides have been exonerated as formal participants in olefin-mediated P450 inactivation (4, 27). Thus, in common with other inactivating olefins, the key inactivating species most likely is the SB-derived cation radical intermediate formed on P450-dependent activation of the olefinic π-bond before it collapses to the epoxide. Although the structure of the SB-heme adduct remains to be fully characterized, the mass of 817.8 amu is consistent with an adduct of the porphyrin (562.6 Da) and hydroxy-SB (254 Da). In addition, the electronic absorption spectral characteristics of both the free and the Zn-complexed adducts in the visible range are consistent with its identification as an N-alkylporphyrin (19). It remains to be determined whether, as in other olefinic substrates, this cation radical alkylates the pyrrolic heme nitrogen on the terminal carbon while incurring O-attack on the internal carbon. The particular pyrrole rings that are N-alkylated by SB also remain to be identified.

Modified Heme and Protein from Secobarbital-Inactivated P450

Regioselective alkylation of pyrrole ring A and D nitrogens (NA and ND) of P450 2B1 heme by 1-alkynes and 1-alkenes (28), coupled with active-site-directed probes such as phenyldiazene that form in situ Nphenylheme adducts, reveals that pyrrole rings B and C of P450 2B1 heme are shielded from attack by the protein, whereas its pyrrole rings A and D are relatively open and thus accessible for N-arylation (29). In P450 101, the distal helix I completely extends over pyrrole ring B (26), masking it and thus apparently protecting its pyrrolic nitrogen (NB) from phenyl attack (30). With the exception of P450s 1A2 and 102, this also appears to be true of most P450s examined that fail to yield appreciable NB-arylated regioisomers (30-33). Thus, SB modification of the 2B1 distal helix peptide not only suggests that it is in sufficiently close proximity for intercepting the SB cation radical, at least part of the time, but also that such interception may shield pyrrole ring B from N-modification, a feature consistent with this general P450 active site motif. However, given the predicted topology of P450 2B1 with active-site-directed probes which favor ND over NA (29), it would be important to determine whether its pyrrole ring NA or ND is the principal one targeted for SB-mediated N-alkylation. Accordingly, the SB cation radical would have to partition between the distal helix region (above pyrrole ring B) and either or both of these sites, an outcome presumably also influenced by the SB orientation in the active site cavity and the shortness of the olefinic bond length. In summary, the findings described above extend our previous report (5) that SB suicidally inactivates P450 2B1 by heme N-alkylation as well as protein modification. The structural identification of the labeled peptide as that corresponding to the highly conserved distal helix I of P450s 101 provides the key evidence that in fact this peptide is at the active site of the enzyme, contacting both substrates and the heme-bound O2. Further substantiation of this architectural assignment for this peptide is provided by active site topological analyses of sitedirected mutants (31) as well as by our identification that this peptide domain is also that modified by heme during cumene hydroperoxide inactivation of P450 3A1 (34).

Acknowledgment. The authors wish to acknowledge the technical assistance of Mr. F. C. Walls in the tandem MS/MS analyses. We also wish to sincerely thank Prof. Douglas Gage and Dr. Pao-Chi Liao, MSU-NIH Mass Spectrometry Facility (Michigan State University, East Lansing), and Mr. Leo Raftogianis and Dr. Ling Chen, Applied Biosystems Division, Perkin Elmer Corp., for their generous assistance in the attempts to characterize the SB-modified adduct, using a variety of mass spectrometric techniques. We also acknowledge the UCSF Bioorganic Biomedical Mass Spectrometry Resource (A. Burlingame, Director; supported by NIH Division of Research Resources Grant RR-01614, and NIADKK 26743) and the UCSF Liver Center Core Facility in Spectrophotometry (supported by NIADKK 26743). This research project was supported by NIH-NIDDK Grant DK 26506 (M.A.C.).

References (1) Ortiz de Montellano, P. R. (1995) Biotransformations of Drugs and Chemicals. In Encyclopedia of Molecular Biology and Biotechnology (Meyers, R. A., Ed.) VCH, New York, in press. (2) Halpert, J. R., Guengerich, F. P., Bend, J. R., and Correia, M. A. (1994) Selective inhibitors of cytochromes P450. Toxicol. Appl. Pharmacol. 125, 163-175.

Chem. Res. Toxicol., Vol. 9, No. 3, 1996 621

(3) Halpert, J. R. (1995) Structural basis of selective cytochrome P450 inhibition. Annu. Rev. Pharmacol. Toxicol. 35, 29-54 and references therein. (4) Ortiz de Montellano, P. R., and Correia, M. A. (1995) Inhibition of cytochrome P450 enzymes. In Cytochrome P450: Structure, Function, and Mechanism (Ortiz de Montellano, P. R., Ed.) 2nd ed., Plenum Press, New York, pp 305-363, and references therein. (5) Lunetta, J. M., Sugiyama, K., and Correia, M. A. (1989) Secobarbital-mediated inactivation of rat liver cytochrome P-450b: A mechanistic reappraisal. Mol. Pharmacol. 35, 10-17. (6) Waxman, D. J., and Walsh, C. T. (1982) Phenobarbital-induced rat liver cytochrome P-450. J. Biol. Chem. 257, 10446-10457. (7) Walsh, C. T. (1982) Suicide substrates, mechanism-based enzyme inactivators: Recent developments. Annu. Rev. Biochem. 53, 493535. (8) Levin, W., Sernatinger, E., Jacobson, M., and Kuntzman, R. (1973) Breakdown of cytochrome P-450 by secobarbital and other allylcontaining barbiturates. Drug Metab. Dispos. 1, 275-285 and references therein. (9) De Matteis, F., and Cantoni, L. (1979) Alteration of the porphyrin nucleus of cytochrome P450 caused in the liver by treatment with allyl-containing drugs. Biochem. J. 183, 99-103. (10) Shephard, E. A., Pike, S. F., Rabin, B. R., and Phillips, I. R. (1983) A rapid one-step purification of NADPH-cytochrome c (P450) reductase from rat liver microsomes. Anal. Biochem. 129, 430433. (11) Correia, M. A., Decker, C., Sugiyama, K., Underwood, M., Bornheim, L., Wrighton, S. A., Rettie, A. E., and Trager, W. F. (1987) Degradation of rat hepatic cytochrome P-450 heme by 3,5dicarbethoxy-2,6-dimethyl-4-ethyl-1,4-dihydropyridine to irreversibly bound protein adducts. Arch. Biochem. Biophys. 258, 436451. (12) Lubet, R. A., Mayer, R. T., Cameron, J. W., Nims, R. W., Burke, M. D., Wolff, T., and Guengerich, F. P. (1985) Dealkylation of pentoxyresorufin: A rapid and sensitive assay for measuring induction of cytochrome(s) P-450 by phenobarbital and other xenobiotics in the rat. Arch. Biochem. Biophys. 238, 43-48. (13) Underwood, M. C., Cashman, J. R., and Correia, M. A. (1992) Specifically designed thiosteroids as active site-directed probes for functional dissection of cytochrome P-450 3A isozymes. Chem. Res. Toxicol. 5, 42-53. (14) Sokoloff, R. L., and Frigon, R. P. (1981) Rapid spectrophotometric assay of dodecyl sulfate using acridine orange. Anal. Biochem. 118, 138-141. (15) Yao, K., Falick, A. M., Patel, N., and Correia, M. A. (1993) Cumene hydroperoxide-mediated inactivation of cytochrome P450 2B1: Identification of an active site heme-modified peptide. J. Biol. Chem. 268, 59-65. (16) Schagger, H., and van Jagow, G. (1987) Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range of 1 to 100 kDa. Anal. Biochem. 166, 368379. (17) Medzihradszky, K. F., Maltby, D. A., Hall, S. C., Settineri, C. A., and Burlingame, A. L. (1994) Characterization of protein Nglycosylation by reversed-phase microbore liquid chromatography/ electrospray mass spectrometry, complementary mobile phases, and sequential exoglycosidase digestion. J. Am. Soc. Mass Spectrom. 5, 350-358. (18) Roberts, E. S., Hopkins, N. E., Zalulec, E. J., Gage, D. A., Alworth, W. L., and Hollenberg, P. F. (1994) Identification of active-site peptides from 3H-labeled 2-ethynylnaphthalene-inactivated P450 2B1 and 2B4 using amino acid sequencing and mass spectrometry. Biochemistry 33, 3766-3771. (19) Ortiz de Montellano, P. R. (1992) Mechanism-based inactivation of cytochrome P450: Isolation and characterization of N-alkyl heme adducts. Methods Enzymol. 206, 533-540. (20) Fujii-Kuryama, Y., Mizukami, Y., Kawajiri, K., Sogawa, K., and Muramatsu, M. (1982) Primary structure of a cytochrome P-450: coding nucleotide sequence of phenobarbital-inducible cytochrome P-450 cDNA from rat liver. Proc. Natl. Acad. Sci. U.S.A. 79, 2793-2797. (21) Frey, A. B., Waxman, D. J., and Kreibich, G. (1985) The structure of phenobarbital-inducible rat liver cytochrome P450 isoenzyme PB-4. J. Biol. Chem. 260, 15253-15265. (22) Roberts, E. S., Hopkins, N. E., Alworth, W. L., and Hollenberg, P. F. (1993) Mechanism-based inactivation of cytochrome P450 2B1 by 2-ethynylnaphthalene: Identification of an active-site peptide. Chem. Res. Toxicol. 6, 470-479. (23) Nelson, D. R., and Strobel, H. W. (1988) On the membrane topology of vertebrate cytochrome P-450 proteins. J. Biol. Chem. 263, 6038-6050. (24) Nelson, D. R. (1995) Cytochrome P450 nomenclatures and alignment of selected sequences. In Cytochrome P450: Structure, Function, and Mechanism (Ortiz de Montellano, P. R., Ed.) 2nd ed., Plenum Press, New York, pp 575-606.

622 Chem. Res. Toxicol., Vol. 9, No. 3, 1996 (25) Yun, C.-H., Martin, M. V., Hopkins, N. E., Alworth, W. L., Hammons, G. J., and Guengerich, F. P. (1992) Modification of cytochrome P4501A2 enzymes by the mechanism-based inactivator 2-ethynylnaphthalene. Biochemistry 31, 10556-10563. (26) Poulos, T. L., Finzel, B. C., and Howard, A. J. (1987) Highresolution crystal structure of cytochrome P450cam. J. Mol. Biol. 195, 687-700. (27) Ortiz de Montellano, P. R., Yost, G. S., Mico, B. A., Dinizo, S. E., Correia, M. A., and Kumbara, H. (1979) Destruction of cytochrome P-450 by 2-isopropyl-4-pentenamide and methyl-2-isopropyl-4pentenoate: mass spectrometric characterization of prosthetic heme adducts and non-participation of epoxide metabolites. Arch. Biochem. Biophys. 197, 524-533. (28) Kunze, K. L., Mangold, B. L. K., Wheeler, C., Beilan, H. S., and Ortiz de Montellano, P. R. (1983) The cytochrome P-450 active site. J. Biol. Chem. 258, 4202-4207. (29) Tuck, S. F., and Ortiz de Montellano, P. R. (1992) Topological mapping of the active sites of cytochromes P4502B1 and P4502B2 by in situ rearrangement of their aryl-iron complexes. Biochemistry 31, 6911-6916. (30) Tuck, S. F., Peterson, J. A., and Ortiz de Montellano, P. R. (1992) Active site topologies of bacterial cytochromes P450 101 (P450cam),

He et al.

(31)

(32)

(33)

(34)

P450 108 (P450terp), and P450 102 (P450BM-3): In situ rearrangement of their phenyl-iron complexes. J. Biol. Chem. 267, 56145620. Tuck, S. F., Hiroya, K., Shimizu, T., Hatano, M., and Ortiz de Montellano, P. R. (1993) The cytochrome P450 1A2 (CYP1A2) active site: Topology and perturbations caused by Glu-318 and Thr-319 mutations. Biochemistry 32, 2548-2553. Swanson, B. A., Dutton, D. R., Yang, C. S., and Ortiz de Montellano, P. R. (1991) The active sites of cytochromes P450 IA1, IIB1, IIB2, and IIE1. Topological analysis by in situ rearrangement of phenyl-iron complexes. J. Biol. Chem. 266, 1925819264. Swanson, B. A., Halpert, J. R., Bornheim, L. M., and Ortiz de Montellano, P. R. (1992) Topological analysis of the active sites of cytochromes P450IIB4 (rabbit), P450IIB10 (mouse) and P450IIB11 (dog) by in situ rearrangement of phenyl-iron complexes. Arch. Biochem. Biophys. 292, 42-46. Correia, M. A. (1993) Drug-mediated heme-modification of cytochrome P450 apoproteins: Structural characterization and physiological implications. Toxicologist 13, 15.

TX950177K