Oxidative Deboronation of the Peptide Boronic Acid Proteasome

Mar 25, 2006 - Bortezomib (1) is a potent first-in-class dipeptidyl boronic acid proteasome ... aShadowed boxes depict the potential intermediates inv...
0 downloads 0 Views 390KB Size
Chem. Res. Toxicol. 2006, 19, 539-546

539

Oxidative Deboronation of the Peptide Boronic Acid Proteasome Inhibitor Bortezomib: Contributions from Reactive Oxygen Species in This Novel Cytochrome P450 Reaction Jason Labutti,† Ian Parsons,‡ Ron Huang, Gerald Miwa, Liang-Shang Gan, and J. Scott Daniels* Department of Drug Metabolism and Pharmacokinetics, Drug Safety and Disposition, and Department of Analytical DeVelopment, Millennium Pharmaceuticals, Inc., Cambridge, Massachusetts 02139 ReceiVed NoVember 10, 2005

Bortezomib (1) is a potent first-in-class dipeptidyl boronic acid proteasome inhibitor employed in the treatment of patients with relapsed multiple myeloma where the disease is refractory to conventional lines of therapy. The potency of 1 is owed primarily to the presence of the boronic acid moiety, one which is suited to establish a tetrahedral intermediate with the active site N-terminal threonine residue of the proteasome. Hence, deboronation of 1 represents a deactivation pathway for this chemotherapeutic agent. Deboronation of 1 affords a near equal mixture of diastereomeric carbinolamide metabolites (M1/ M2) and represents the principal metabolic pathway observed in humans. In vitro results from human liver microsomes and human cDNA-expressed cytochrome P450 enzymes (P450) indicate a role for P450 in the deboronation of 1. Use of 18O-labeled oxygen under controlled atmospheres confirmed an oxidative mechanism in the P450-mediated deboronation of 1, as 18O was found incorporated in both M1 and M2. Chemically generated reactive oxygen species (ROS), such as those generated as byproducts during P450 catalysis, were also found to deboronate 1 resulting in the formation of M1 and M2. Known to undergo efficient redox cycling, P450 2E1 was found to catalyze the deboronation of 1 predominantly to the carbinolamide metabolites M1 and M2, as well as to a pair of peroxycarbinolamides, 2 and 3. The presence of superoxide dismutase (SOD) and catalase prevented the deboronation of 1, thus, supporting the involvement of ROS in the P450 2E1-catalyzed deboronation reaction. The presence of SOD and catalase also protected 1 against P450 3A4-catalyzed deboronation, albeit to a lesser extent. The remaining deboronation activity observed in the P450 3A4 reaction may suggest the involvement of the more conventional activated enzyme-oxidants previously described for P450. Our present findings indicate that the oxidase activity of P450 (i.e., formation of ROS) represents a mechanism of deboronation. Introduction Bortezomib (1, VELCADE, formerly known as PS-341) is a novel dipeptidyl boronic acid inhibitor of the 26S proteasome approved for the treatment of patients with relapsed multiple myeloma where the disease is refractory to conventional lines of therapies. Inhibition of the 26S proteasome involves the formation of a dative bond between the N-terminal threonine residue of the chymotryptic site and the boron atom of 1 (1). Hence, deboronation represents a deactivation pathway for this chemotherapeutic agent. Our laboratory recently described the in vitro and in vivo metabolism of 1 in humans (2). The identification and characterization of metabolites of 1 in humans indicated deboronation to be the principal route of metabolism, resulting in the formation of a pair of diastereomeric carbinolamide metabolites, M1 and M2 (Scheme 1). Secondary metabolism of M1 and M2 resulted in a number of metabolites, most of which carried an additional oxidation on the leucine moiety. It was also demonstrated that multiple cytochrome P4501 enzymes contributed to the in vitro deboronation of 1 in humans, including P450s 1A2, 2C9, 2C19, 2D6, and 3A4 (2, 3). * To whom correspondence should be addressed. Tel: 617-444-1559. E-mail: [email protected]. † Current address: Department of Chemistry, University of MissouriColumbia, Columbia, MO. ‡ Department of Analytical Development, Millennium Pharmaceuticals, Inc.

Figure 1. Select activated oxygen species generated by P450: the nucleophilic peroxo-iron intermediate and the electrophilic peroxo-iron intermediate oxidants.

However, the mechanism of P450-catalyzed oxidative deboronation of 1 resulting in the formation of diastereomeric carbinolamide metabolites M1 and M2 remained unknown. We recognized that multiple activated enzyme oxidants, such as the generally accepted iron-peroxo and iron-oxo species produced in the P450 catalytic cycle (Figure 1) (4), may be contributing to the deboronation of 1. Specifically, the ironperoxo intermediate appeared a likely candidate in the deboronation of 1, given the literature surrounding the stereocontrolled preparation of alcohols via peroxidation of boranes and boronates (Scheme 2A) (5-7). Consistent with the peroxidation of boranes and boronates, the iron-peroxo-mediated reaction 1 Abbreviations: LC/MS, liquid chromatography/mass spectrometry; LC/ MS/MS, liquid chromatography/tandem mass spectrometry; HPLC, highperformance liquid chromatography; m/z, mass-to-charge ratio; NMR, nuclear magnetic resonance spectroscopy; UV, ultraviolet/visible spectrophotometry; P450, cytochrome P450; cDNA, complementary DNA; NADPH, β-nicotinamide adenine dinucleotide phosphate, reduced form; SOD, superoxide dismutase; MeCN, acetonitrile.

10.1021/tx050313d CCC: $33.50 © 2006 American Chemical Society Published on Web 03/25/2006

540 Chem. Res. Toxicol., Vol. 19, No. 4, 2006

Labutti et al.

Scheme 1. Metabolism of 1 in Humans

Scheme 2. General Pathways of Boronate Oxidationa

a

Shadowed boxes depict the potential intermediates involved in the oxidation of 1.

would result in the stereocontrolled oxidation of 1 (i.e., retention of stereochemistry at the migrating carbon center; highlighted in Scheme 2A) (6). Interconversion of M1 via reversible dehydration (8) may then result in the observed mixture of diastereomers (M1/M2). However, given that multiple P450 enzymes catalyze the deboronation reaction (2), we also entertained the possibility that a general oxidative mechanism involving byproduct reactive oxygen species (ROS) such as superoxide anion (O2-•) and hydroxyl radical (•OH), generated as a consequence of P450 futile cycling (9-13), may also contribute to the oxidative deboronation of 1. Autoxidation of chiral boranes and boronates by molecular oxygen (O2) is generally accepted to occur via a

radical mechanism and, thus, exhibits a significant loss of stereocontrol upon the recombination with O2 and subsequent formation of alcohols (Scheme 2B) (7, 14, 15). Similarly, peroxyl radicals have also been shown to initiate the deboronation of alkyl boranes via a radical chain mechanism (16, 17). Oxidative deboronation of 1 via a transient carbon-centered radical (7) followed by recombination with O2 (highlighted in Scheme 2B) would be consistent with the formation of a diastereomeric mixture of carbinolamides (M1/M2) observed in vitro. In an effort to understand this novel P450-catalyzed biotransformation, we examined the in vitro metabolism of 1 in human liver microsomes, cDNA-expressed P450 enzymes, and chemi-

Mechanism of P450-Catalyzed OxidatiVe Deboronation

Chem. Res. Toxicol., Vol. 19, No. 4, 2006 541

cal model reactions. To differentiate the mechanisms of deboronation, we also employed the use of controlled atmosphere techniques and isotopically labeled reagents (e.g., 18O2). Liquid chromatography/tandem mass spectrometry (LC/MS/MS) was utilized throughout our investigation to identify the metabolites of bortezomib generated from the various in vitro experiments. Confirmation of several metabolite structures was achieved by direct comparison of their LC/MS/MS data to that obtained from synthesized and isolated metabolites that were characterized by high-field NMR (2). Results from the present investigation indicate that multiple oxidants may play a role in the P450catalyzed deboronation of the peptidyl boronic acid 1 in human liver microsomes.

Experimental Procedures Materials. [(1R)-3-Methyl-1-[[(2S)-1-oxo-3-phenyl-2-[(pyrazinylcarbonyl)amino]propyl]amino]butyl]boronic acid (1), pyrazine2-carboxylic acid [1-(1-hydroxy-3-methyl-butylcarbamoyl)-2-phenylethyl]-amide (M1/M2), pyrazine-2-carboxylic acid (1-carbamoyl2-phenyl-ethyl)amide (M3), 3-phenyl-2-[(pyrazine-2-carbonyl)amino]-propionic acid (M34), and pyrazine-2-carboxylbic acid [1-(1-hydroperoxy-3-methyl-butylcarbamoyl)-2-phenyl-ethyl]amide (2 and 3; see Supporting Information) were synthesized and characterized by Albany Molecular Research, Inc. (Albany, NY). Potassium phosphate, ammonium formate, β-nicotinamide adenine dinucleotide phosphate, reduced form (NADPH), trifluoroacetic acid (TFA), formic acid, magnesium chloride (MgCl2), 18O-labeled H2O (H218O, 95% 18O), superoxide dismutase (SOD), and catalase were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO). Human liver microsomes were obtained from Xenotech LLC (Lenexa, KS), and human cDNA-expressed Supersomes P450 3A4 and P450 2E1 were purchased from BD Gentest (Waltham, MA). HPLC-grade solvents were purchased from Burdick and Jackson (Muskegon, MI). Argon was purchased from BOC Gases (Murray Hill, NJ), while 18O-labeled oxygen (18O2, 95% 18O) was purchased from Cambridge Isotope Laboratories (Andover, MA). All reagents and solvents were obtained at the highest purity available. Human Liver Microsomal and cDNA-Expressed P450 Metabolism of 1, Metabolites M1 and M2, and Peroxycarbinolamides 2 and 3. A potassium phosphate-buffered solution (0.1 M, pH 7.4) containing 1 (50 µM), human liver microsomes (1 mg/ mL), MgCl2 (3 mM), and NADPH (2 mM) was incubated at 37 °C under ambient oxygenation for 1 h. Protein from human liver microsomes was precipitated by the addition of MeCN (2 mL), and the resulting mixture was chilled at 4 °C for 20 min followed by centrifugation. The supernatant was dried under a stream of nitrogen and reconstituted in 90:10 H2O/MeCN in preparation for LC/MS analysis. When indicated, test article 1 was replaced with 2, 3, M1, or M2 at 50 µM, respectively. Incubations with cDNAexpressed P450 enzymes were performed in an analogous manner, substituting cDNA-expressed P450s 3A4 or 2E1 (150 pmol/mL) for human liver microsomes. The cDNA-expressed P450 experiments were centrifuged to remove protein, and subsequent supernatants were directly analyzed by LC/MS. When required, superoxide dismutase (SOD; 100 U) and catalase (4000 U) were added to the P450 3A4 or 2E1 incubations of 1, prior to the addition of the cofactor NADPH. Microsomal Metabolism of 1 in the Presence of H218O or 18O2. 1. Incubations Containing 18O-Labeled Water. Conditions were identical to those previously described for microsomal incubations with the following exceptions: a potassium phosphate-buffered solution (0.1 M, pH 7.4) containing MgCl2 (3 mM) and NADPH (2 mM) was dried in vacuo and reconstituted in H218O. The substrate 1 (50 µM) was dissolved in potassium phosphate buffer (reconstituted in H218O) and added to initiate the reaction. The reaction was quenched by the addition of MeCN (2 mL) and prepared for LC/ MS analysis as previously described. 2. Incubations Containing 18O2. Conditions were identical to those previously described for microsomal incubations with the

Figure 2. 1H NMR (600 MHz) analysis of the solution stability of 1 at pH 7.4. The relative abundance of each metabolite was determined by integrating the following 1H signals: M1 (solid diamond), methyls at 0.68 and 0.71 ppm; M2 (solid box), methyls at 0.59 and 0.60 ppm; M3 (solid triangle), 3-methylbutanal methyls at 0.72 and 0.76 ppm.

following exceptions: a potassium phosphate-buffered solution (0.1 M, pH 7.4) of human liver microsomes (1 mg/mL) containing MgCl2 (3 mM) and NADPH (2 mM) was subjected to three vacuum-18O2 purge cycles (4 °C) before the addition of 1 (50 µM; 18O -purged solution). The reaction was quenched by the addition 2 of MeCN (2 mL) and prepared for LC/MS analysis as previously described. Oxidation of 1 with FeSO4. FeSO4 (0.3 mM) was added to a potassium phosphate-buffered solution (0.1 M, pH 7.4) of 1 (50 µM) and allowed to incubate at 37 °C for 1 h under ambient oxygenation. The FeSO4 reaction was directly analyzed by LC/MS with no additional workup. High-Field 1H NMR Analysis of the Solution Stability of M1 and M2. The solution stability of M1 and M2 (1.4 mM; 10% DMSO-d6/D2O) was monitored via high-field 1H NMR (600 MHz) over a period of 12 h at 25 °C. The pH of the solution was adjusted to 7.4 with NaOD (1 N). The stability of M1 and M2 over time was determined by monitoring a pair of distinct methyl proton signals for each diastereomer at 0.68/0.71 ppm and 0.59/0.60 ppm, respectively. The formation of M3 over time was determined indirectly by monitoring the appearance of its degradation product, 3-methylbutanal, bearing a pair of methyl signals at 0.72/0.76 ppm. Liquid Chromatography/Mass Spectrometry. Aliquots (40 µL) of the previously described in vitro reaction extracts were injected onto an Agilent 1100 HPLC system (Agilent Technologies, Palo Alto, CA) that was coupled to a Symmetry C18 column (5 µm, 3.8 mm × 150 mm; Waters Corporation, Milford, MA). Solvent A was 10 mM ammonium formate (pH 4.1) and solvent B was MeCN. The initial mobile phase was 85:15 A/B (v/v) and by linear gradient transitioned to 20:80 A/B over 20 min. The flow rate was 0.400 mL/min. The HPLC eluent was introduced via electrospray ionization directly into a Finnigan LCQ Deca XPPLUS ion trap mass spectrometer (Thermoelectron Corporation, San Jose, CA) operated in the positive ion mode. Ionization was assisted with sheath and auxiliary gas (nitrogen) set at 60 and 40 psi, respectively. The electrospray voltage was set at 5 kV with the heated ion transfer capillary set at 300 °C and 30 V. Relative collision energies of 25-40% were used when performing MS/MS operations with the ion trap.

Results Human Liver Microsomal Metabolism of 1, M1, and M2. Human liver microsomes metabolize 1 predominantly to the diastereomeric carbinolamides M1 and M2 and a range of secondary metabolites (Scheme 1) (2). We recently reported that multiple P450 enzymes (3A4, 2C19, 1A2, 2D6, and 2C9) contribute to the in vitro oxidative deboronation of 1, as evident from the NADPH dependency to the metabolism of 1 (2) and

542 Chem. Res. Toxicol., Vol. 19, No. 4, 2006

Labutti et al.

Figure 3. Incorporation of 18O2 into metabolites M1 and M2 upon incubation of 1 with human liver microsomes.

the inhibition thereof with selective chemical and antibody inhibitors of P450 (3). While we initially accepted that the P450catalyzed formation of the well-described peroxo-iron species (Figure 1) (4) could be a likely mediator of the peroxidation of 1, leading to the formation of M1, it was unclear as to the mechanism(s) leading to the formation of M2. The fact that M1 and M2 were formed at near equal mixtures in vitro (2) initially prompted us to consider the possibility that 1 was oxidatively deboronated to M1 followed by reversible dehydration to M2. Originally proposed by Stella and co-workers while investigating the stability of 1 in various formulations (8), we investigated reversible dehydration as a mechanism to account for the presence of both diastereomeric carbinolamides in vitro. To address whether reversible dehydration could account for the observed mixture of carbinolamides, M1 was incubated in human liver microsomes fortified with NADPH and subsequently analyzed by LC/MS to identify secondary metabolites, specifically the presence of the diastereomer M2. We find that M1 undergoes NADPH-dependent metabolism in human liver microsomes to its expected secondary metabolites (e.g., M5) (2), with no formation of M2 observed. Similarly, M2 is converted to its secondary metabolites (e.g., M6), yet fails to produce the diastereomer M1 when incubated in human liver microsomes (see Supporting Information). We also find that metabolites M1 and M2 are stable toward reversible dehydration in human liver microsome solutions lacking the cofactor NADPH (data not shown). Consistent with the present data generated in human liver microsomes, NMR analysis of aqueous solutions (pH 7.4) of M1 also demonstrates an inability of the carbinolamide metabolite to convert to its diastereomer M2 via reversible dehydration (Figure 2). In fact, the only reaction observed at physiological pH was the gradual hydrolysis of M1 to the amide metabolite M3 (Figure 2), a fate previously reported for these metabolites in vitro (2). Similarly, M2 was stable toward reversible dehydration under identical conditions (data not shown). Delocalization of the lone electron pair of nitrogen into the carbonyl of the amide system is most likely contributing to the stability of the carbinolamide metabolites. Evidence indicating the involvement of P450-catalyzed formation of activated oxygen species in the deboronation of 1

and formation of both carbinolamide metabolites (M1/M2) was provided from experiments in which 18O-labeled oxygen (18O2) was introduced into degassed, buffered (pH 7.4) solutions of 1 containing human liver microsomes. We find that human liver microsome metabolism of 1 under a controlled atmosphere of 18O results in the NADPH-dependent incorporation of 18O into 2 the carbinolamide metabolites M1 and M2 (Figure 3). LC/MS analysis indicates an [M + Na]+ for M1 and M2 at m/z 381 (see Supporting Information), clearly demonstrating the incorporation of 18O into the carbinolamide metabolite, whereas the [M + Na]+ for M1 under ambient oxygenation (16O2) produces an [M + Na]+ at m/z 379. Importantly, the mass spectra for M1 and M2 do not produce an [M + Na]+ at m/z 381 when 1 is incubated in human liver microsomes (+NADPH) which are reconstituted in H218O (see Supporting Information). These data not only confirm the lack of reversible dehydration of M1 or M2 but also invoke a mechanism involving the P450-catalyzed activation of O2 in the deboronation of 1. Deboronation of 1 by P450 2E1, P450 3A4, and Chemically Generated ROS. Uncoupling of the P450 catalytic cycle is a common event and results in the reduction of O2 and the formation of byproduct ROS, such as superoxide anion (O2-•) and other reactive oxygen species (ROS) (10-12). Given the lability of alkyl boronates toward autoxidation (7), we anticipated that byproduct ROS may contribute to the oxidative deboronation of 1. Furthermore, given the lack of stereocontrol associated with the autoxidation of boronates, we considered ROS-mediated oxidation of 1 would also account for the observed mixture of diastereomeric carbinolamides M1 and M2 produced in vitro. Therefore, we investigated the deboronation of 1 with cDNA-expressed human P450 2E1, an enzyme recognized to uncouple from its catalytic cycle and undergo futile redox cycling with subsequent ROS production (18-21). 1. Deboronation of 1 by P450 2E1. Buffered solutions of 1 were incubated with cDNA-expressed P450 2E1 (+ NADPH) and the metabolites profiled by LC/MS. We find that P450 2E1 metabolism of 1 results in the formation of M1 and M2 as the predominant metabolites (Figure 4A) in this reaction. We also observe that P450 2E1-catalyzed deboronation of 1 results in the formation of a mixture of peroxycarbinolamides 2 and 3

Scheme 3. Proposed Mechanisms of P450-Mediated Oxidative Deboronation of 1

Mechanism of P450-Catalyzed OxidatiVe Deboronation

Chem. Res. Toxicol., Vol. 19, No. 4, 2006 543

Figure 4. ESI-LC/MS/UV analysis of the metabolism of 1 by P450 2E1 (A, [TIC]) and the effect of SOD and catalase (B [TIC] and C [UV]) on oxidative deboronation.

Figure 5. Structures of peroxycarbinolamides 2 and 3.

(Figure 5), the LC/MS characteristics of which match data obtained from authentic standards of each. Peroxy intermediates are likely the consequence of radical recombination of O2 with a putative carbon-centered radical, formed as a result of the fragmentation of the carbon-boron bond (7, 14). The present data appear to indicate that the deboronation of 1 by P450 2E1 results in the formation of a putative carbon-centered radical (Scheme 2B), which in turn combines with O2 to yield a pair of peroxycarbinolamides (2/3). It is likely then that subsequent metabolism of the peroxycarbinolamides 2 and 3 results in the formation of M1 and M2, respectively (Scheme 3) (7). This is indeed the case, for we find that 2 is efficiently converted to its respective carbinolamide M1 when incubated with human liver microsomes fortified with NADPH (Figure 6A). The presence of the amide metabolite M3 is consistent with the subsequent degradation of M1 at physiological pH (2). The degradation of 2 to M1 and the imide M34 was also observed in buffered solutions containing no human liver microsomes (Figure 6B), albeit to a lesser extent, and is most likely mediated by transition metal contaminants present in solution at trace levels (22). Additional evidence supporting the involvement of ROS (e.g.,O2-•) in the CYP2E1-catalyzed deboronation reaction was provided from experiments utilizing the ROS-mitigating enzymes superoxide dismutase (SOD) and catalase (23). As a result, when incubations of 1 with P450 2E1 were fortified with SOD and catalase, no oxidative deboronation of 1 was observed, as determined by the substantial reduction in the levels of M1

and M2 (Figure 4B,C). Not surprisingly, when SOD alone was incubated in the P450 2E1 reaction, a near quantitative conversion of 1 to M1 was observed (data not shown). This observation indicates the efficient conversion of O2-• to hydrogen peroxide, a species capable of mediating the stereoselective oxidation of 1 and subsequent formation of M1 (Scheme 3). 2. Deboronation of 1 by Chemically Generated ROS. In view of the P450 2E1 data indicating a role of ROS in the oxidative deboronation of 1, specifically the contribution of O2-•, we utilized a system designed to model the ROS-mediated deboronation of 1 in the redox environment of P450. The reduction of molecular oxygen by ferrous iron (Fe2+) has been investigated and shown to produce ROS, such as the diffusible oxidant O2-• (24). We find that incubation of 1 with buffered (pH 7.4) solutions of FeSO4 (0.3 mM) results in a similar metabolic profile to that obtained with P450 2E1, specifically the production of M1 and M2 as the principal metabolites (Figure 7A,B). As expected, the addition of SOD and catalase to the FeSO4 reaction inhibited the metabolism of 1 (data not shown), providing additional evidence for the involvement of the diffusible oxidant O2-• in the oxidative deboronation of 1 in vitro. Together, these data indicate that byproduct ROS, formed as a result of the P450 catalytic cycle, are capable of deboronating 1. 3. Deboronation of 1 by P450 3A4. To determine the extent of the involvement of byproduct ROS across P450 enzymes, we investigated the metabolism of 1 by cDNA-expressed P450 3A4, the principal enzyme found to metabolize 1 in vitro (3). LC/MS analysis indicates that P450 3A4 metabolizes 1 to the deboronated metabolites M1 and M2, as well as to a range of secondary metabolites (Figure 8A) which were described previously (e.g., M5/M6) (2). The presence of SOD and catalase in the P450 3A4 experiment resulted in a more modest decrease in the levels of the primary deboronated metabolites M1 and M2 formed in the reaction (Figure 8B,C). These data indicate

544 Chem. Res. Toxicol., Vol. 19, No. 4, 2006

Labutti et al.

Figure 6. ESI-LC/MS analysis of the incubation of 2 in human liver microsomes (A) and K+ phosphate buffer (0.1 M, pH 7.4) (B).

Figure 7. ESI-LC/MS/UV analysis of the stability of 1 in FeSO4 solution at pH 7.4 (A [TIC] and B [UV]) and K+ phosphate buffer (0.1 M, pH 7.4) (C [TIC]).

Mechanism of P450-Catalyzed OxidatiVe Deboronation

Chem. Res. Toxicol., Vol. 19, No. 4, 2006 545

Figure 8. ESI-LC/MS/UV analysis of the metabolism of 1 by P450 3A4 (A [TIC]) and the effect of SOD and catalase (B [TIC] and C [UV]) on oxidative deboronation.

a lesser role of byproduct ROS in the P450 3A4-catalyzed deboronation of 1 and, furthermore, may point to the involvement of multiple P450-catalyzed oxidants, such as the generally accepted oxo-iron species (Figure 1) generated during the P450 catalytic cycle (4).

Discussion The deboronation of 1 in human liver microsomes may involve multiple mechanisms of P450 oxidation. The present enzyme and chemical model experiments clearly demonstrate a role for ROS in the deboronation of 1 in vitro. The most compelling example of ROS-mediated deboronation was that of the reaction of 1 with the redox-cycling P450 2E1 enzyme, where 1 was efficiently deboronated to the diastereomeric carbinolamides (M1/M2), a finding consistent with metabolism of 1 in human liver microsomes. The identification of a near equal mixture of diastereomeric carbinolamides (M1/M2) as the major deboronated metabolites in vitro is also consistent with historical accounts of boronate reactivity, where the lack of stereocontrolled autoxidation is indicative of a putative epimerizing carbon-centered radical generated upon alkoxyl radicalor O2-mediated scission of the carbon-boron bond (Scheme 2B) (14, 15). The formation of the peroxycarbinolamide metabolites (2/3) in the P450 2E1 reaction serve to bolster the present claim of a key contribution from ROS in the P450catalyzed metabolism of 1, given the well-described mechanisms of boronate autoxidation involving the radical recombination of O2 with the aforementioned putative carbon-centered radical to form peroxy intermediates (7). The transition metal-catalyzed formation of ROS (e.g., O2-•) also resulted in the deboronation of 1 and generation of the mixture of carbinolamide diastereomers (M1/M2) that were observed in the human liver microsome and P450 2E1 reactions.

Decreasing the levels of ROS in solution with SOD and catalase represented a means to protect 1 against oxidative deboronation in the P450 (e.g., 2E1) and FeSO4 reactions and, thus, further substantiate the proposed mechanism. The fact that these ROS-mitigating enzymes were unable to completely protect 1 from 3A4-catalyzed deboronation, suggested additional mechanisms may be contributing to the overall in vitro metabolism of 1 by P450. A Baeyer-Villiger type oxidation of 1 by the nucleophilic peroxo-iron oxidant (Figure 1) would be directly analogous to the oxidation of aldehydes to carboxylic acids (25) as well as the oxidative deformylation of the same (26). In fact, a similar mechanism has been proposed by Walsh and co-workers to account for the flavoenzyme (FAD-4a-OOH) oxidation of chiral boronic acids to secondary alcohols with retention of configuration (6). However, given the present data that near equal mixtures of the diastereomer carbinolamide metabolites M1 and M2 were formed in vitro, it is unlikely that the peroxo-iron species is contributing to the P450-catalyzed deboronation. The hypervalent oxo-iron species (Figure 1) on the other hand has been implicated as the predominant oxidant in most P450-catalyzed reactions (4) and may account for the residual deboronation activity observed for P450 3A4. While a mechanism involving oxo-iron participation in the deboronation reaction is currently unclear, preliminary data from experiments employing a suitable oxometalloporphyrin biomimetic (27-29) appear to indicate a potential role for the oxo-iron species in the P450-catalyzed deboronation of 1.2 The present investigation has served to illuminate the mechanisms contributing to this novel P450-catalyzed reaction. Considering the number of P450 enzymes reported to metabolically deactivate 1 (3A4, 2C19, 1A2, 2D6, 2C9, and presently 2E1) (3), it is likely that multiple activated enzyme-oxidants, 2

Unpublished data.

546 Chem. Res. Toxicol., Vol. 19, No. 4, 2006

generated throughout their respective catalytic cycles, are capable of deboronating this peptidyl boronic acid. Our present findings indicate that the oxidase activity of P450 (i.e., formation of ROS) represents a mechanism of deboronation. Determining the range of reactive oxygen species contributing to the deboronation of 1 (e.g., O2-•, •OH), as well as identifying additional P450-intermediates participating in the reaction, remains an ongoing investigation. Acknowledgment. The authors wish to thank Professor Kent Gates (University of Missouri-Columbia) and Professor Paul Ortiz de Montellano (University of California, San Francisco) for helpful discussions regarding the present investigation. Supporting Information Available: HPLC-UV chromatograms and MS spectral data from select microsomal incubations of 1, M1, and M2; the 1H NMR and LC/MS data for the peroxycarbinolamides 2 and 3. This information is available free of charge via the Internet at http://pubs.acs.org.

References (1) McCormack, T., Baumeister, W., Grenier, L., Moomaw, C., Plamondon, L., Pramanik, B., Slaughter, C., Soucy, F., Stein, R., Zuhl, F., and Dick, L. (1997) Active site-directed inhibitors of Rhodococcus 20 S proteasome. Kinetics and mechanism. J. Biol. Chem. 272, 2610326109. (2) Pekol, T., Daniels, J. S., Labutti, J., Parsons, I., Nix, D., Baronas, E., Hsieh, F., Gan, L.-S., and Miwa, G. (2005) Human metabolism of the proteasome inhibitor bortezomib: identification of circulating metabolites. Drug Metab. Dispos. 33, 771-777. (3) Uttamsingh, V., Lu, C., Miwa, G., and Gan, L.-S. (2005) Relative contributions of the five major human cytochromes P450, 1A2, 2C9, 2C19, 2D6 and 3A4 to the hepatic clearance of the proteasome inhibitor bortezomib. Drug Metab. Dispos. 33, 1-6. (4) Coon, M. J., Vaz, A. D. N., McGinnity, D. F., and Peng, H.-M. (1998) Multiple activated oxygen species in P450 catalysis: contributions to specificity in drug metabolism. Drug Metab. Dispos. 26, 1190-1193. (5) House, H. O. (1972) Modern Synthetic Reactions, W. A. Benjamin, Inc., Menlo Park, CA. (6) Latham, J. A., Jr., and Walsh, C. (1986) Retention of configuration in oxidation of a chiral boronic acid by the flavoenzyme cyclohexanone oxygenase. J. Chem. Soc., Chem. Commun., 527-528. (7) Cadot, C., Dalko, P. I., Cossy, J., Ollivier, C., Chuard, R., and Renaud, P. (2002) Free-radical hydroxylation reactions of alkylboronates. J. Org. Chem. 67, 7193-7202. (8) Wu, S., Waugh, W., and Stella, V. J. (2000) Degradation pathways of a peptide boronic acid derivative, 2-Pyz-(CO)-Phe-Leu-B(OH)2. J. Pharm. Sci. 89, 758-765. (9) Ioannides, C. (1996) Cytochromes P450 Metabolic and Toxicological Aspects, CRC Press, Boca Raton, FL. (10) Goeptar, A. R., Scheerens, H., and Vermeulen, N. P. E. (1995) Oxygen and xenobiotic reductase activities of cytochrome P 450. Crit. ReV. Toxicol. 25, 25-65. (11) Ekstrom, G., and Ingelman-Sundberg, M. (1986) Mechanisms of lipid peroxidation dependent upon cytochrome P-450 LM2. Eur. J. Biochem. 158, 195-201. (12) Kuthan, H., and Ullrich, V. (1982) Oxidase and oxygenase function of the microsomal cytochrome P450 monooxygenase system. Eur. J. Biochem. 126, 583-588.

Labutti et al. (13) Fisher, M. B., Thompson, S. J., Ribeiro, V., Lechner, M. C., and Rettie, A. E. (1998) P450-catalyzed in-chain desaturation of valproic acid: isoform selectivity and mechanism of formation of D3-valproic acid generated by baculovirus-expressed CYP3A1. Arch. Biochem. Biophys. 356, 63-70. (14) Brown, H. C., Midland, M. M., and Kabalka, G. W. (1971) Stoichiometrically controlled reaction of organoboranes with oxygen under very mild conditions to achieve essentially quantitative conversion into alcohols. J. Am. Chem. Soc. 93, 1024-1025. (15) Mirviss, S. B. (1967) Mechanism of the oxidation of trialkylboranes. J. Org. Chem. 32, 1713-1717. (16) Davies, A. G., and Roberts, B. P. (1971) Homolytic organometallic reactions. IV. Homolytic alkylthiyl dealkylation of organoboranes. J. Chem. Soc. B, 1830-1837. (17) Pokidova, T. S., and Denisov, E. T. (2001) The reactivity of organoboranes in radical substitution reactions. Russ. Chem. Bull. 50, 390-395. (18) Persson, J. O., Terelius, Y., and Ingelman-Sundberg, M. (1990) Cytochrome P-450-dependent formation of reactive oxygen radicals: isozyme-specific inhibition of P-450-mediated reduction of oxygen and carbon tetrachloride. Xenobiotica 20, 887-900. (19) Ingelman-Sundberg, M., and Johansson, I. (1984) Mechanisms of hydroxyl radical formation and ethanol oxidation by ethanol-inducible and other forms of rabbit liver microsomal cytochromes P-450. J. Biol. Chem. 259, 6447-6458. (20) Gorsky, L. D., Koop, D. R., and Coon, M. J. (1984) On the stoichiometry of the oxidase and monooxygenase reactions catalyzed by liver microsomal cytochrome P-450. Products of oxygen reduction. J. Biol. Chem. 259, 6812-6817. (21) Koop, D. R. (1992) Oxidative and reductive metabolism by cytochrome P450 2E1. FASEB J. 6, 724-730. (22) Headlam, H. A., and Davies, M. J. (2002) Cell-mediated reduction of protein and peptide hydroperoxides to reactive free radicals. Free Radical Biol. Med. 34, 44-55. (23) Fridovich, I. (1972) Superoxide radical and superoxide dismutase. Acc. Chem. Res. 5, 321-326. (24) Kosaka, H., Katsuki, Y., and Shiga, T. (1992) Spin trapping study on the kinetics of iron(2+) autoxidation: formation of spin adducts and their destruction by superoxide. Arch. Biochem. Biophys. 293, 401408. (25) Guengerich, F. P. (2001) Common and uncommon cytochrome P450 reactions related to metabolism and chemical toxicity. Chem. Res. Toxicol. 14, 611-650. (26) Vaz, A. D. N., Pernecky, S. J., Raner, G. M., and Coon, M. J. (1996) Peroxo-iron and oxenoid-iron species as alternative oxygenating agents in cytochrome P450-catalyzed reactions: switching by threonine-302 to alanine mutagenesis of cytochrome P450 2B4. Proc. Natl. Acad. Sci. U.S.A. 93, 4644-4648. (27) Anzenbacher, P., Niwa, T., Tolbert, L. M., Sirimanne, S. R., and Guengerich, F. P. (1996) Oxidation of 9-alkylanthracenes by cytochrome P450 2B1, horseradish peroxidase, and iron tetraphenylporphine/iodosylbenzene systems: anaerobic and aerobic mechanisms. Biochemistry 35, 2512-2520. (28) Nam, W., Jin, S. W., Lim, M. H., Ryu, J. Y., and Kim, C. (2002) Anionic ligand effect on the nature of epoxidizing intermediates in iron porphyrin complex-catalyzed epoxidation reactions. Inorg. Chem. 41, 3647-3652. (29) Nam, W., Choi, S. K., Lim, M. H., Rohde, J.-U., Kim, I., Kim, J., Kim, C., and Que, L., Jr. (2003) Reversible formation of iodosylbenzene-iron porphyrin intermediates in the reaction of oxoiron(IV) porphyrin p-cation radicals and iodobenzene. Angew. Chem., Int. Ed. 42, 109-111.

TX050313D