Computational Biotransformation Profile of Paracetamol Catalyzed by

Dec 30, 2014 - Shangwei Zhang , Dominik Wondrousch , Myriel Cooper , Stephen H. Zinder ... Myriel Cooper , Anke Wagner , Dominik Wondrousch , Frank ...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/crt

Computational Biotransformation Profile of Paracetamol Catalyzed by Cytochrome P450 Li Ji*,†,‡,§ and Gerrit Schüürmann*,‡,§ †

MOE Key Lab of Environmental Remediation and Ecosystem Health, College of Environmental and Resource Sciences, Zhejiang University, Hangzhou 310058, China ‡ UFZ Department of Ecological Chemistry, Helmholtz Centre for Environmental Research, Permoserstrasse 15, 04318 Leipzig, Germany § Institute for Organic Chemistry, Technical University Bergakademie Freiberg, Leipziger Strasse 29, 09596 Freiberg, Germany S Supporting Information *

ABSTRACT: The P450-catalyzed biotransformation of the analgesic drug paracetamol (PAR) is a long-debated topic, involving different mechanistic hypotheses as well as experimental evidence for the metabolites N-acetyl-pbenzoquinone imine (NAPQI), p-benzoquinone, acetamide, and 3-hydroxyPAR. During the catalytic cycle of P450, a high-valent iron(IV)-oxo species known as Compound I (Cpd I) is formed as the ultimate oxidant, featuring two energetically close-lying ground states in the doublet (low-spin) and quartet (high-spin) spin states, respectively. In order to clarify the catalytic mechanism, a computational chemistry analysis has been undertaken for both the high- and low-spin routes, employing density functional theory (DFT) including PCM (polarized continuum-solvation model) that yields an approximate simulation of the bulk polarization exerted through the protein. The results demonstrate that hydrogen abstraction transfer (HAT) by the P450 oxidant Cpd I (FeO) is kinetically strongly preferred over the alternative pathways of an oxygen addition reaction (OAR) or two consecutive single-electron transfers (SET). Moreover, only the respective high-spin route yields N-acetyl-p-semiquinone imine (NAPSQI) as an intermediate that is converted to the electrophile N-acetyl-p-benzoquinone imine (NAPQI). By contrast, 3-hydroxy-PAR, acetamide, and p-benzoquinone as electrophilic and redox-active agent are formed predominantly in the low-spin state through reactions that do not involve NAPSQI. Thus, all experimentally observed PAR metabolites are in accord with an initial HAT from the phenolic oxygen, and NAPSQI should indeed be the precursor of NAPQI, both of which are generated only via the high-spin pathway.



INTRODUCTION Paracetamol (N-acetyl-p-aminophenol, PAR) is a widely used over-the-counter analgesic and antipyretic drug and appears to be safe if used in normal therapeutic doses. However, overdosing PAR causes P450-dependent centrilobular hepatotoxicity in humans and experimental animals, as observed by the release of alanine aminotransferase into serum, which is often used as a monitoring parameter for hepatic damage.1−5 More specifically, metabolic PAR activation leads to electrophilic intermediates that attack proteins and DNA. This may result in oxidative stress, peroxynitrite formation, mitochondrial dysfunction through the opening of mitochondrial membrane permeability transition pores, and hepatic necrosis,2,3 the latter of which appears to involve different molecular mechanisms such as a cell death program called necropoptosis.6,7 Much evidence has been presented in support of N-acetyl-pbenzoquinone imine (NAPQI) being the electrophilic and oxidizing intermediate responsible for the observed toxicity of PAR, leading to 3-(cystein-S-yl) protein adducts as the major molecular initiating event after depletion of glutathione (GSH).1,4,5,8 In this regard, NAPQI resembles other α,β© XXXX American Chemical Society

unsaturated Michael acceptors, for which both chemoassays and in silico approaches are available to sense their toxicityrelated GSH reactivity.9−11 However, the direct in vitro detection of NAPQI is very difficult, and the exact mechanism of its formation still has not been unequivocally identified. So far, several oxidation mechanisms have been proposed5,12,13 for the P450-catalyzed generation of NAPQI from PAR, involving the active-site ferryl oxo porphyrin complex FeIVO, called Compound I (Cpd I),14 and its reduced form FeIVOH (Cpd II),15 as summarized in Schemes 1 and 2: (a) oxygen addition reaction (OAR), where the initial addition of the FeIVO oxygen at the amide electron lone pair of PAR (1 in Scheme 1) yields the N-oxide 2 and rearranges to N-hydroxy PAR (4) with subsequent decomposition to NAPQI (6) and H2O; (b) H atom transfer (HAT) to FeIVO and FeIVOH from the amide N and phenol O of PAR, yielding 3 and 6 (either directly or stepwise through 4) via HAT(N) followed by HAT(O) and, alternatively, 5 and 6 through an initial HAT(O) Received: September 7, 2014

A

DOI: 10.1021/tx5003645 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology Scheme 1. Alternative Reaction Mechanisms of the P450-Catalyzed Biotransformation of Paracetamol (PAR)a

a

An oxygen addition reaction (OAR) to the amide nitrogen and hydrogen atom transfer from the phenolic oxygen (HAT(O)) or amide nitrogen (HAT(N)) yielding NAPQI (6) without or with passing through its precursor NAPSQI (5), the OH-PAR recombination adducts 8 (ipso-Chydroxylated intermediate), 9 (peroxide), and 10 (meta-C-hydroxylated intermediate), and the experimentally observed follow-up products 3hydroxy-PAR (11), p-benzoquinone (7), and acetamide (CH3CONH2).

involves recombination of the OH radical (generated through abstraction from the initially formed FeOH) at the ipso carbon of PAR to yield the benzoquinone precursor 8. Alternative OH rebound reactions concern the meta carbon and phenoxy oxygen (attached at the para carbon), generating 3-hydroxyPAR 11 and p-hydroperoxo acetanilide 9, respectively. So far, the OH recombination to the meta and ipso phenyl ring carbons or to the phenolic oxygen of PAR has been understood to be a possible follow-up reaction of NAPSQI.1,12 As we will show below, however, an alternative route to 8, 9, and 10 (yielding 11), and thus to benzoquinone 7, may proceed directly from the HAT(O) pathway without intermediate formation of NAPSQI 5, which is indicated in Scheme 1.

followed by HAT(N); and (c) single electron transfer (SET) to FeO (Cpd I) that results in electromers (resonance forms) 12 and 14 (aminium and oxonium radical, Scheme 2) followed by proton transfer (PT) to give 3 (amine radical) or 5 (oxoradical) and subsequently by a SET to FeOH (Cpd II) to yield 13 or 15 (oxonium and aminium diradicals) that both finally form NAPQI upon elimination of H2O. According to Schemes 1 and 2, possible precursors of the electrophilic benzoquinone imine NAPQI 6 include N-acetyl-psemiquinone imine (NAPSQI) 5 and its tautomer 3 along the HAT(N) and HAT(O) pathways, the N-hydroxylated PAR 4 generated on the OAR pathway, and the radical cations 13 and 15 that result from electron-transfer reactions. While NAPQI may decompose upon hydrolysis to acetamide and benzoquinone 7 as a second electrophilic metabolite, a further route to 7 B

DOI: 10.1021/tx5003645 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology

Scheme 2. Electron Transfer Mechanism as an Alternative Option for the P450-Catalyzed Biotransformation of Paracetamol (PAR, 1) to NAPQI (6)a

a

Involving Cpd I (FeO) and Cpd II (FeOH) as consecutive oxidants and passing through the following intermediates: aminium radical 12 and oxonium radical 14 as electromers (resonance forms of the PAR radical cation), the tautomers amine radical 3 and oxoradical 5, and the tautomers oxonium diradical 13 and aminium diradical 15. SET = single-electron transfer from N (SET(N)) or O (SET(O)) of PAR and PAR metabolites to the P450 oxidants Cpd I and Cpd II, respectively; PT = proton transfer. (1), respectively. Regarding the latter, initial reaction coordinate calculations to identify the TS region were followed by full geometry optimizations without symmetry constraints to generate the exact TS geometries (within the model chemistry chosen). The B3LYP hybrid density functional has been reported to reproduce experimental kinetic isotope effects, electron paramagnetic resonance parameters, and vibrational spectra.22 In order to assess the potential impact of the DFT functional on the reaction mechanism of interest, B3PW9117,23 was used as an alternative functional (with the basis set BSI on the HS route) to calculate the H-abstraction step from the PAR phenolic oxygen and the anilide site by Cpd I of P450. More accurate energies were determined by single-point calculations at the UB3LYP/LACVP+** level (denoted BSII). Bulk polarity effects at the active site were evaluated with the polarized continuumsolvation model (PCM)24 using the solvent chlorobenzene (ε = 5.6) at the UB3LYP/BSII level. ZPE (zero-point energy) corrections are based on UB3LYP/BSI frequency calculations. In order to elucidate the impact of dispersion effects that are not covered properly by standard DFT, single-point B3LYP-D225,26 calculations at the BSI level were performed to evaluate the H-abstraction barrier from the PAR phenolic oxygen and the PAR anilide N as well as the oxygen addition barrier, employing the B3LYP-optimized structures. We also attempted B3LYP-D2 geometry optimizations; however, our attempts to locate a respective H-abstraction transition state failed. Only the B3LYP-D2-optimized reactant complexes featuring the substrate H atom associated with Cpd I of P450 could be found, being very similar to the corresponding B3LYP-optimized counterparts with differences in the interatomic distances below 0.003 Å. Regarding the potential impact of the basis set on the optimized geometries and energies, UB3LYP/BSII, including geometry optimization, was also employed for H abstraction from the PAR phenolic oxygen and the PAR anilide site by Cpd I of P450 via the HS route. The resultant BSII-level transition states are very similar to their BSIlevel counterparts with deviations below 0.01 Å for distances and less than 1° for angles and dihedrals, indicating that UB3LYP/BSII//BSI is appropriate for the present work. The nonenzymatic formation of 3-hydroxy-PAR (11), p-benzoquinone (7), and acetamide was computed at the B3LYP/6-31G* level and verified through intrinsic reaction coordinate (IRC) calculations. Here, bulk aqueous solvation effects were evaluated with PCM//

In early ab initio Hartree−Fock calculations with singlet oxygen replacing Cpd I, NAPSQI 5 was energetically more stable than 3 by ca. 30 kcal/mol, suggesting a thermodynamic preference of the phenoxy radical pathway (initiated by HAT(O)) over the nitrogen radical pathway (initiated by HAT(N)).12 More recent DFT (density functional theory) calculations employing Cpd I as a P450 model confirmed this trend, but with a reduced energy difference of ca. 12 kcal/ mol.16 So far, however, the respective reaction kinetics have not been addressed, which also holds for the above-mentioned follow-up reactions initiated by radical recombination. In the present investigation, DFT quantum chemistry with Cpd I representing the active site of P450 is applied to identify the kinetically preferred pathway of PAR oxidation. Moreover, the radical recombination of OH• from FeOH (Cpd II) is comparatively analyzed for the ipso, meta, and phenoxy O pathways in order to elucidate the degree of associated regioselectivity. The results provide new insight into the dependence of the PAR biotransformation profile on the electronic structure of the relevant reactants, and they unravel the energetically feasible pathways leading to the ultimate electrophilic metabolites NAPQI 6 and benzoquinone 7 as well as to 3-hydroxy-PAR 11 and acetamide as an additional experimentally observed metabolite.



COMPUTATIONAL METHODOLOGY

In this article, a six-coordinate oxo-ferryl species, Fe4+O2−Por−SH−, corresponding to the valence form [FeIVO(Por+•)(SH)]0 was used as a Cpd I model of P450, and its reduced form, Fe4+OH−Por2−SH−, corresponding to [FeIV-OH(Por)(SH)]0, was used as a Cpd II model of P450. Full geometry optimization was carried out by employing the unrestricted B3LYP hybrid density functional method17,18 in combination with the Los Alamos effective core potential plus double-ζ basis set (Lanl2dz)19−21 on iron and a 6-31G basis set on the remaining atoms (LACVP level, denoted BSI). Vibrational frequencies were computed to confirm the proper number of imaginary frequencies for optimized ground states (0) and transition states C

DOI: 10.1021/tx5003645 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology B3LYP/6-311++G** using ε = 78.4. All quantum chemical calculations were carried out with the Gaussian 0327 package except for single-point B3LYP-D2 calculations that were done with Gaussian 09.28

(4TSO), which is in line with other heteroatom oxidation reactions.32−34 The latter also holds for the HS and LS transition-state geometries as shown in Scheme 3, where the



Scheme 3. High-Spin (HS) and Low-Spin (LS) Transition States of the Oxygen Addition Reaction (OAR) of Cpd I (FeO) at the Amide Nitrogen of PARa

RESULT AND DISCUSSION Electron Transfer and Oxygen Addition Reaction (OAR) Mechanisms. On the basis of in vitro investigations of the P450-catalyzed bioactivation of PAR 1 to NAPQI 6, the latter has been concluded13 to proceed through an electron transfer mechanism. This mechanism is shown in some detail in Scheme 2, involving either the elementary steps SET(N), PT (from the aminium radical 12 to FeO−), and SET(O) or the alternative sequence SET(O), PT (from the phenoxy radical cation 14 to FeO−), and SET(N). The activation free energies (ΔG⧧) of these electron transfer reactions can be estimated through application of Marcus theory29 (for details, see the Supporting Information), which has also been used to investigate the electron transfer reductive pathway of P450-mediated halogenated alkanes,30 keeping in mind that DFT computational chemistry cannot treat the initial SET(N) and SET(O) pathways separately. Employing PCM// B3LYP/BSI, the initial electron transfer from PAR to Cpd I that yields the electromers 12 ↔ 14 is associated with a freeenergy reaction barrier ΔG⧧ of 40.3 kcal/mol in the high-spin quartet (HS) route and 40.4 kcal/mol in the low-spin doublet (LS) route. Interestingly, the calculated spin densities of the radical cation represented by the two resonance forms 12 and 14 are 0.21 at the amide nitrogen, 0.15 at the phenoxy oxygen, 0.15 at the phenyl moiety, and 0.25 at the acetyl unit, suggesting a slight preference of 12 over 14. For the second SET from the anilide radical 3 and its tautomer 5 (NAPSQI) to Cpd II (FeOH), the calculated ΔG⧧ values are still significantly larger, with 109.3/64.9 kcal/mol (HS/LS) for the anilide radical pathway and 175.1/100.0 kcal/ mol for the phenoxy radical pathway. Moreover, the calculated ionization potential (IP) of PAR is 7.5 eV (B3LYP/6-311+ +G**), as opposed to a much smaller computed electron affinity of Cpd I of 3.0 eV.31 Overall, these results indicate that the P450-catalyzed biotransformation of PAR 1 to the Michael acceptor and redox-active agent NAPQI 6 is unlikely to proceed through an electron transfer mechanism. The activation energies for the OAR pathway of Scheme 1 are shown in Table 1. Here, the initial O addition to the N electron lone pair is more favorable by 9 kcal/mol (PCM bulk polarization) in the LS state (2TSO) than in the HS state

a

Interatomic distances (angstroms) and angles (deg) are given in the order HS (LS).

Fe−O−N anilide angle is clearly larger for 4TSO (134°) than for 2TSO (123°). However, even in the LS pathway, the energy barrier of 29.6 kcal/mol (PCM) for the FeO oxygen addition to the PAR anilide nitrogen is too large to contribute significantly to the P450-catalyzed NAPQI formation. The latter may also explain the lack of reports of N-hydroxy-PAR 4 (with a half-life of 15 min)35 as a metabolic precursor of NAPQI. Accordingly, the OAR mechanism also does not offer a kinetically efficient way for the metabolic conversion of PAR to NAPQI. Inclusion of the dispersion correction through B3LYP-D2 (single-point, BSI level) lowers the oxygen addition barrier by 7−9 kcal/mol (B3LYP-D2/BSI: 31.93/19.52 kcal/mol for the HS/LS state; B3LYP/BSI: 38.95/28.88 kcal/mol for the HS/LS state), but it yields the same qualitative picture: the O addition barrier is (still) large, and the LS pathway is again more favorable than the HS pathway. HAT Phenoxy Radical Pathway vs Anilide Radical Pathway. As outlined in Scheme 1, homolytic hydrogen abstraction can take place both at the phenol oxygen of PAR, yielding NAPSQI (5), and at the anilide nitrogen, resulting in NAPSQI tautomer 3. For the former route, the calculated energy profile with Cpd I representing the catalytic center of P450 as well as some geometric details are shown in Scheme 4, calculated with UB3LYP/BSII//BSI including ZPE and PCM corrections (ε = 5.6; associated gas-phase values without ZPE are in parentheses). As is often seen for P450 reactions, both HS and LS routes are available that result from the neardegenerate states of Cpd I.36 Initially, the reactant complexes (4,2RCHO) are formed in which the H atom of the phenolic group of PAR interacts weakly with the oxygen atom of the iron-oxo moiety (note that the most stable conformation had been selected for 4,2RCHO as the starting point of the subsequent reaction profile). Here, the spin density of the Cpd I moiety is about two (HS: 2.02; LS: 2.11) on the FeO unit and one (HS: 0.96; LS: 1.12) on the Por + SH fragment, which is in line with the triradicaloid FeIV HS and LS states characterized earlier.37 These reactant complexes lead to a pair of closely lying Habstraction transition states 4TSHO (2TSHO) with ZPE- and PCM-corrected barriers of only 1.8 (1.4) kcal/mol, which are much lower than the corresponding H-abstraction barriers of C−H hydroxylations.36,38,39 The HS transition state 4TSHO is a

Table 1. Transition State and Product Energies (kcal/mol) of the Oxygen Addition Reaction (OAR) of Cpd I (FeO) at the Amide Nitrogen of PARa 4

TSO 2 TSO 4 PN 2 PN

UB3LYP/BSII//BSI

UB3LYP/BSII//BSI + Bulk Polarity + ZPE

39.25 28.09 19.09 10.31

38.62 29.62 17.86 12.91

a

Quantified relative to the reactant complex energy. 4,2TSO: doublet (low-spin, LS) and quartet (high-spin, HS) transition state of the Oaddition at the amide electron lone pair of PAR; PN: N-oxide product. BSI = basis set I: Lanl2dz (Fe), 6-31G (C,H,O,N,S); BSII = basis set II: Lanl2dz (Fe), 6-31+G** (all other atoms); PCM solvation with ε = 5.62 provides an approximate simulation of the bulk polarization induced through the protein. D

DOI: 10.1021/tx5003645 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology

Scheme 4. Potential Energy Profiles for Hydrogen Abstraction from the Phenolic OH Group of Paracetamol (PAR) by Cpd I of P450 and for Subsequent Rebound Reactions between the OH Radical and PARa

a Cpd I is represented computationally as the oxo-ferryl complex Fe4+O2−Por−SH−. Bond lengths in angstroms and angles in degrees of the key reaction species in the high-spin (HS, quartet) and low-spin (LS, doublet) states. The HS route includes an H-bonded complex between Cpd II (represented by Fe4+OH−Por2−SH−) and the oxoradical metabolite of PAR. Energies, in kcal/mol, are given relative to the quartet reactant complex 4 RC at the UB3LYP/BSII//BSI level, including ZPE and bulk polarization (PCM, ε = 5.6), with corresponding gas-phase values shown in parentheses. 4,2RCHO = quartet and doublet reactant complex; TSHO = transition state of the H-abstraction from the phenolic hydroxyl group of PAR; CIO = intermediate complex of NAPSQI H-bonded to Cpd II (FeOH) resulting from the HS HAT(O) step; TSrebO = transition state of the rebound reaction between the OH radical (released from FeOH) and the PAR phenoxy radical metabolite; TSreb‑metaC = rebound transition state of the OH radical attack at the meta carbon of the PAR-derived phenoxy radical; TSreb‑ipsoC = rebound transition state of the OH radical attack at the ipso carbon of the PAR-derived phenoxy radical; PO = complex of the peroxide rebound product and FeIII(PorSH); INTHN = intermediate complex of the PAR-derived phenoxy radical and Cpd II with an H bond between the FeOH oxygen and the anilide hydrogen; INTipsoC = intermediate complex of the ipso-C OH-PAR adduct and FeIII(PorSH); and INTmetaC = intermediate complex of the meta-C OH-PAR adduct and FeIII(PorSH).

For the HAT(N) pathway passing through NAPSQI tautomer 3, the calculated energy profile is shown in Scheme 5, again starting with the reactant complexes that are now termed 4,2 RC HN (and again refer to the most stable conformation). Here, the H-abstraction barrier (4,2TSHN) is 6.9/7.4 kcal/mol larger than that for HAT(O), with HS/LS values of 8.7/8.8 kcal/mol, including ZPE and PCM bulk polarization (larger by 5.9/6.5 kcal/mol in the gas phase), yielding an endothermic product that we could locate only on the quartet surface (4CIN). The latter corresponds to the situation found for the HAT(O) pathway (Scheme 4) except that 4CIN has an endothermic reaction energy of ca. 5−6 kcal/ mol, whereas 4CIO of the corresponding HAT(O) pathway is energetically more favorable by ca. 5 kcal/mol than the reactant complex. In accord with the Hammond postulate implying that increasingly endothermic reactions show an increasing similarity between the product geometry and the transition state structure, the HAT(N) transition state structure is later on the reaction coordinate than for the exothermic HAT(O)

bit later on the reaction coordinate than its LS counterpart 2 TSHO, as can be seen through comparison of their PAR−O···H and H···OFe distances (1.201 vs 1.192 Å and 1.211 vs 1.227 Å). Moreover, the O···H···O configuration is almost linear as that for other H-abstraction transition states,40,41 and the imaginary frequency is as large as expected for an H-abstraction process (HS: −1538.3 i cm−1; LS: −1506.7 i cm−1). The iron-hydroxo complex of FeOH (Cpd II) with the phenoxy radical, CIO, is found only on the HS surface, with an exothermic reaction energy of about 5 kcal/mol and with spin densities of 2.11 and 0.11 on the FeOH and Por + SH moieties, respectively. By contrast, the corresponding LS pathway offers, directly from 2TSHO, an essentially barrier-free rebound reaction of the OH radical with PAR, forming a covalent bond at the ipso or meta carbon to yield the intermediates 2 INTipsoC (8 in Scheme 1) and 2INTmetaC (10), as discussed in more detail below. For INTmetaC, only the meta carbon with the smaller distance to the attacking OH as highly reactive radical was considered. E

DOI: 10.1021/tx5003645 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology

Scheme 5. Potential Energy Profiles for Hydrogen Abstraction from the Anilide Nitrogen of Paracetamol (PAR) by Cpd I of P450 and for Subsequent Rebound Reactions between the OH Radical and PARa

Cpd I is represented computationally as the oxo-ferryl complex Fe4+O2−Por−SH−. Bond lengths in angstroms and angles in degrees of the key reaction species in the high-spin (HS, quartet) and low-spin (LS, doublet) states. The HS route includes an H-bonded complex between Cpd II (represented by Fe4+OH−Por2−SH−) and the anilide radical metabolite of PAR. Energies (UB3LYP/BSII//BSI), in kcal/mol, are given relative to the quartet reactant complex 4RC, including ZPE and bulk polarity effect (PCM, ε = 5.6), with corresponding gas-phase values shown in parentheses. 4,2 RCHN = quartet and doublet reactant complex; TSHN = transition state of H abstraction from the anilide nitrogen atom of PAR; CIO = intermediate complex of the PAR-derived anilide radical H-bonded to Cpd II (FeOH) resulting from the HS HAT(N) step; TSrebN = transition state of the rebound reaction between the OH radical (released from FeOH) and the anilide nitrogen of the PAR metabolite; TSreb‑metaC = rebound transition state of the OH radical attack at the meta carbon of the PAR-derived anilide radical; TSreb‑ipsoC = rebound transition state of the OH radical attack at the ipso carbon of the PAR-derived anilide radical; PN = complex of the N-hydroxylated rebound product of PAR and FeIII(PorSH); INTHO = intermediate complex of the PAR-derived anilide radical and Cpd II with an H bond between the FeOH oxygen and the PAR-metabolite OH hydrogen; INTipsoC = intermediate complex of the ipso-C OH-PAR adduct and FeIII(PorSH); and INTmetaC = intermediate complex of the meta-C OH-PAR adduct and FeIII(PorSH). a

reaction: whereas at 4TSNH the PAR−N···H distance has increased by 0.35 Å (from 1.021 to 1.372 Å, Scheme 5) to a value larger than the FeO···H distance by 0.24 Å (1.372 vs 1.134 Å), 4TSOH shows a PAR−O···H distance increased by 0.21 Å (from 0.996 to 1.201 Å, Scheme 4) that is close to the FeO···H distance (1.201 vs 1.211 Å). Similar to the abovedescribed phenoxy radical pathway, the LS route offers an essentially barrier-free recombination between the OH radical and PAR that proceeds directly from 2TSHN to the respective meta-C, ipso-C, and anilide-N adducts, contrasting with the significant reaction barriers for the corresponding HS recombination reactions. With single-point B3LYP-D2 at the BSI level, the HAT(N) barrier is lower by about 6 kcal/mol (B3LYP-D2/BSI: 8.6/9.0 kcal/mol for the HS/LS state; B3LYP/BSI: 14.7/14.8 kcal/mol for the HS/LS state), whereas the HAT(O) barrier is lowered by less than 1 kcal/mol (B3LYP-D2/BSI: 4.4/2.5 kcal/mol for the HS/LS state; B3LYP/BSI: 5.2/3.5 kcal/mol for the HS/LS state). Never-

theless, inclusion of the dispersion correction does not alter the conclusion that the initial H abstraction from the PAR phenolic oxygen is more favorable than H abstraction from the PAR anilide nitrogen. The same qualitative picture with a much lower barrier for HAT(O) than for HAT(N) is obtained from B3LYP geometry optimizations using the much larger basis set BSII as well as when employing the alternative B3PW91 functional (for details, see Tables S4 and S5 in the Supporting Information). In an early proton NMR investigation,42 the geometric orientation of the association of PAR with P450 was analyzed, using two hepatic P450 isoforms from rats, of which only one was catalyzing the generation of NAPQI. Interestingly, only for the PAR-biotransforming P450 isoform did the phenolic oxygen turn out to approach the heme iron closely, whereas the opposite orientation with a short distance between Fe3+ and the amide nitrogen of PAR was observed for the isoform that was not catalyzing NAPQI formation. Moreover, the rate of F

DOI: 10.1021/tx5003645 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology Scheme 6. Potential Energy Profile for the Second H Abstraction from PAR Catalyzed through Cpd IIa

Cpd II is represented by Fe4+OH−Por2−SH−. The reaction generates NAPQI, and takes place at the anilide nitrogen of the phenoxy radical NAPSQI. Bond lengths in angstroms and angles in degrees of the key reaction species in the HS quartet state. Energies (UB3LYP/BSII//BSI), in kcal/mol, are given relative to the quartet intermediate complex 4INTHN (NAPSQI H-bonded through its anilide hydrogen atom to the FeOH oxygen of Cpd II; see also Scheme 4), including ZPE and bulk polarization (PCM, ε = 5.6), with corresponding gas-phase values shown in parentheses. TSH = transition state of the H-abstraction from the anilide nitrogen atom of NAPSQI; 4PNAPQI = product complex of NAPQI and FeOH2(PorSH). a

NAPQI formation from PAR in rat liver microsomes was reported to increase by about 3-fold in the presence of caffeine,43 and subsequent NMR work pointed out that the addition of caffeine disrupted the close distance between the acetamido group of PAR and the heme iron of P450 3A4, resulting in the enhanced rate of oxidation to the toxic metabolite NAPQI.44 Our present results show further that even in the case of appropriate geometric orientations for both the phenoxy radical and anilide radical pathways the former is significantly preferred with regard to both kinetics and thermodynamics. In this context, the apparent inability to detect phenoxy radicals in rat liver P450-containing systems deserves attention.45 Although ESR spectra demonstrated the presence of NAPSQI and further phenoxy radicals generated from PAR and PAR derivatives as well as a lack of respective anilide radicals in horseradish peroxidase incubations augmented by H2O2, the absence of any radical signals in cytochrome P450 incubations with PAR and PAR derivatives was explained by the fact that both the cofactor NADPH and protein thiol groups are able to reduce phenoxy radicals efficiently.45 OH Radical Rebound vs Secondary Hydrogen Abstraction. As mentioned above, the complex between Cpd II (FeOH) and the phenoxy radical NAPSQI 5, formed after initial H abstraction from the phenolic hydroxyl group of PAR, is found only on the HS surface (4CIO in Scheme 4). Subsequently, there are three opportunities for an OH−PAR rebound reaction, all of which are accompanied by substantial activation barriers in the quartet state (4TSrebO, 4TSreb‑metaC, 4 TSreb‑ipsoC) of around 20−25 kcal/mol. Note that these rebound barriers are significantly larger than the very low Habstraction barriers of 1.8 kcal/mol (HS) and 1.4 kcal/mol (LS), respectively, and thus represent the rate-determining step to the HS rebound products along the P450-catalyzed HAT

pathway. By contrast, the corresponding LS route rebound reactions proceed directly from 2TSHO and essentially without activation barriers, thus providing an intermediate-free (onestep) route to rebound products 8−10 (the last of which rearranges to 3-hydroxy-PAR 11), as already indicated in Scheme 1. For the OH rebound reaction at the PAR phenoxy oxygen forming peroxide 9, both spin routes yield thermodynamically unfavorable and energetically similar products, 4PO (12.8 kcal/ mol PCM) and 2PO (14.7 kcal/mol PCM; Scheme 4). This result may explain why peroxides have not been found in PARbiotransformation studies.12 The alternative rebound reactions at the ipso and meta carbons of PAR that lead to 8 and 10 are exothermic, forming intermediates 4,2INTmetaC and 4,2INTipsoC with reaction energies of −19.0 to −25.1 kcal/mol (PCM), respectively. The latter are preceded by significant activation energies of ca. 21 kcal/mol only in case of the HS reactions (see above and Scheme 4). This demonstrates that OH recombination with the phenyl ring of PAR takes place in the LS state (but apparently not in the HS state), and, indeed, 3-hydroxy-PAR (11), p-benzoquinone (7), and acetamide (a further product of the reaction 8 → 7, Scheme 1) have been found as PAR metabolites.1 Coming back to the quartet state, the intermediate 4CIO, formed upon the initial H abstraction from the PAR OH group, has a finite lifetime in its local minimum of −5.4 kcal/mol (PCM). It allows the phenoxy radical to rotate and eventually form, in an essentially barrier-free manner, a more stable complex (4INTHN, −8.4 kcal/mol PCM, Scheme 4) that now features an H bond between the Cpd II (FeOH) oxygen and the H atom attached to the PAR amide nitrogen. Subsequently, a second H abstraction, this time by Cpd II, yields NAPQI (6) upon elimination of H2O (reaction 5 → 6 in Scheme 1). Note that, according to the present computational analysis, the P450G

DOI: 10.1021/tx5003645 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology

Scheme 7. Potential Energy Profile for the Final Nonenzymatic Formation of the Experimentally Observed PAR Metabolitesa

a

Formation of PAR metabolites 3-hydroxy-PAR, p-benzoquinone, and acetamide catalyzed through two water molecules is shown, along with bond lengths in angstroms and angles in degrees of the species. Energies (B3LYP/6-311++G**//6-31G*), in kcal/mol, are given relative to the reactant complex, including ZPE correction and single-point PCM simulation of a bulk aqueous (nonenzymatic) environment (ε = 78.4). (A) Formation of 3-hydroxy-PAR through water-assisted rearrangement of the meta-C-hydroxylated OH-PAR rebound adduct (reaction 10 → 11 in Scheme 1). (B) Formation of p-benzoquinone and acetamide through a water-assisted decomposition of the ipso-C-hydroxylated OH-PAR rebound adduct (reaction 8 → 7 in Scheme 1).

from PAR, taking place at the anilide nitrogen of the phenoxy radical NAPSQI 5. This reaction has a small activation barrier of only 5.8 kcal/mol (including ZPE and bulk polarization) and is exothermic by 11.7 kcal/mol. The latter indicates that the first H abstraction from the PAR phenolic hydroxyl group

catalyzed PAR biotransformation to NAPQI via NAPSQI (5) takes place only in the HS state, whereas the LS route yields the above-mentioned meta-C and ipso-C metabolites. The oxidizing capability of Cpd II in P450 reactions has been demonstrated earlier,46−50 and Scheme 6 shows the potential energy profile for the present case of a second H abstraction H

DOI: 10.1021/tx5003645 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology

calculated with the B3LYP/6-311++G**//6-31G* PCM model chemistry (including ZPE). Regarding the formation of 3-hydroxy-PAR from its preceding meta-C metabolite 10, the direct H rearrangement would involve a very large activation barrier of more than 50 kcal/mol (PCM bulk simulation of aqueous solution employing ε = 78.4). By contrast, microsolvation with a single water molecule reduces the activation barrier by more than a factor of 2 (20.6 kcal/mol in aqueous solution, see Scheme S2a of the Supporting Information), and two water molecules make the pathway still more feasible with a relatively small barrier of only 8.9 kcal/mol (Scheme 7a), indicating that ultimate 3-hydroxy-PAR formation is apparently catalyzed by a solvating water bridge involving at least two water molecules, a result similar to the one found for the keto− enol conversion mechanism of dopamine,51 and to the porphyrin degradation involving heme oxygenase and general acid catalysis.52 Concerning benzoquinone (7) as an additionally observed electrophilic metabolite in addition to NAPQI (6), Scheme 1 outlines two pathways for its generation. First, the ipso-Chydroxylated intermediate 8 decomposes into acetamide and benzoquinone. Second, hydrolysis of NAPQI via an addition− elimination reaction leads to these two final products (reaction 6 → 7 in Scheme 1). According to our present calculations, the water addition step of the latter pathway would be accompanied by a high activation barrier of ca. 45 kcal/mol (Scheme S3 of the Supporting Information), whereas the route via the ipso-C-rebound intermediate 8 can be assisted through catalyzing water molecules, reducing the activation barrier for the benzoquinone formation to ca. 19 kcal/mol in the presence of two water molecules (Scheme 7b; catalytic microsolvation through only one water molecule reduces the barrier to ca. 25 kcal/mol, as shown in Scheme S2b of the Supporting Information). It follows that benzoquinone, as an experimentally observed PAR metabolite, is likely to result mainly from the water-assisted decomposition of 8 that, itself, is generated predominantly along the LS route of the HAT(O) pathway through an ipso-C OH-PAR rebound reaction across 2TSHO, yielding 2INTipsoC without passing through NAPSQI (Scheme 4). The alternative HS route via 4TSHO, 4INTipsoC, and 4 TSreb‑ipsoC (Scheme 4) is much less efficient because of a substantial rebound barrier of ca. 21 kcal/mol, and the hydrolysis of NAPQI 6 is unlikely because of its very high nonenzymatic reaction barrier for the initial water addition step (see above). As discussed above, the HAT rebound step proceeds through an intermediate (4CIO in Scheme 4) followed by a substantial and, in fact, rate-determining reaction barrier only in the HS state, as opposed to an essentially barrier-free route from 2TSHO to the rebound products in the LS state. Interestingly, a similar pattern has been observed recently for the HAT pathway of the N-hydroxylation of primary aromatic amines, featuring a P450catalyzed one-step reaction to the rebounded N-hydroxylamine product only in the LS state.53 By contrast, the respective HS route proceeded through an intermediate that could convert into the rebound product only through a significant and ratedetermining reaction barrier. Note further that regarding the P450-catalyzed α-CH hydroxylation of nitrosamines, which is also subject to a HAT mechanism as well as most alkanes, for both the HS and LS state there have been found intermediates, followed by a reaction barrier only in the HS route (which, however, was lower than the preceding H-abstraction barrier).54

facilitates a subsequent H abstraction from the PAR amide nitrogen with regard to both kinetics and thermodynamics. Interestingly, at 4TSH, the N···H distance is larger by 0.24 Å than that in the preceding complex 4INTHN between NAPSQI and FeOH (1.278 vs 1.040 Å) (Scheme 6), whereas the O···H distance has already decreased by 0.53 Å (from 1.732 to 1.207 Å), also being closer to the product geometry (0.998 Å). The latter indicates that, in this case, the TS geometry does not follow the Hammond postulate (which would imply that the TS structure should be more similar to the reactant geometry); at the same time, the 4TSH N···H distance of the exothermic second H abstraction is smaller by 0.09 Å than the 4TSHN N···H distance of the endothermic first H abstraction from the same PAR amide nitrogen. We also studied the second H-abstraction process from the NAPSQI anilide nitrogen as catalyzed by Cpd I instead of Cpd II (see Scheme S1 of the Supporting Information for the respective potential energy profiles). In this case, the respective triplet state yields a quite similar reaction barrier (4.67 vs 5.77 kcal/mol PCM). It demonstrates that, at least for some reactions, the oxidizing power of Cpd II is very strong and is, indeed, comparable to that of Cpd I. As mentioned above, after initial NAPSQI formation, the associated intermediate complex 4CIO (Scheme 4) has a finite lifetime due to its local minimum and thus may stay in the same P450 heme pocket for the second H abstraction to yield NAPQI. Accordingly, both H abstractions may take place in the same P450 heme pocket and could be catalyzed consecutively by Cpd I and Cpd II. With regard to recombination reactions of the OH radical with PAR along the HAT(N) pathway, Scheme 5 suggests that the LS route from 2RCHN over 2TSHN followed by an essentially barrier-free adduct formation at meta-C (INTmetaC), ipso-C (INTipsoC), and anilide-N (PN) could provide a relevant contribution to the overall metabolic fate of PAR under P450 catalysis. However, the 2TSHN reaction barrier of ca. 8.8 kcal/ mol (PCM) is significantly larger than the corresponding 2 TSHO reaction barrier of only ca. 1.4 kcal/mol, indicating a strong kinetic preference for the phenoxy radical pathway as noted above. Moreover, 3-hydroxy-PAR (11) formed from the meta-C rebound adduct of the HAT(O) route (Scheme 1 and below) has been found experimentally,12 in contrast with the lack of experimental reports regarding 2-OH PAR as a metabolite that would have been generated through the HAT(N) pathway. These findings suggest that the strong kinetic advantage of the HAT(O) route over the HAT(N) route prevents the latter from playing a competitive role. Accordingly, the following discussion of the nonenzymatic formation of follow-up products will be confined to reactions with intermediates generated on the phenoxy radical pathway. Nonenzymatic Formation of Ultimate Products 3Hydroxy-PAR, p-Benzoquinone, and Acetamide. From Scheme 4, it was seen that the first HAT(O) step yields NAPSQI (5) only in the HS state and that, both from the NAPSQI complex with Cpd II, 4CIO, and the LS transition state 2 TSHO, the initially formed OH radical may add covalently to the meta or ipso carbon of the PAR phenyl ring. We will now investigate how these meta-C and ipso-C metabolites (10 and 8) are transformed further to the ultimate products, 3-hydroxyPAR (11), p-benzoquinone (7), and acetamide, that are observed experimentally.12 The relevant energy profiles for these nonenzymatic followup reactions are summarized in Scheme 7 and have been I

DOI: 10.1021/tx5003645 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology

from the anilide nitrogen catalyzed by Cpd II (FeOH). Both consecutive HAT steps (HAT(O) followed by HAT(N)) may take place in the same P450 heme pocket, whereas NAPSQI leaving the protein may be easily quenched by NADPH and protein thiols, both of which make it difficult to trap NAPSQI experimentally. A respective joint effort of Cpd I and Cpd II, resulting in two consecutive oxidations during one catalytic cycle of P450, may also apply to other bifunctional substrates such as catechols, hydroquinones, and amino phenols. In contrast to NAPQI, benzoquinone, as a second electrophilic metabolite, together with acetamide and 3-hydroxy-PAR is predicted to be generated mainly through the LS route. Here, the initial HAT(O) doublet transition state proceeds to essentially barrier-free rebound reactions of the OH radical at the ipso and meta carbons of PAR, with nonenzymatic followup reactions catalyzed by aqueous microsolvation that yield benzoquinone and acetamide (through decomposition of the ipso-C adduct) as well as 3-hydroxy-PAR (through rearrangement of the meta-C adduct). Overall, the present investigation unravels a distinct two-state biotransformation profile of PAR, with the HS and LS routes feeding different subsets of the PAR metabolites observed so far, and also explains why further previously discussed metabolites are unlikely to be formed.

The two major P450-catalyzed biotransformation pathways of paracetamol (PAR) that both result from an initial HAT attack of Cpd I at the phenol oxygen atom of PAR are summarized in Scheme 8, showing all major metabolites also Scheme 8. Major P450-Catalyzed Biotransformation Pathways of Paracetamol (PAR)a



ASSOCIATED CONTENT

S Supporting Information *

Absolute and relative energies at various levels for all molecular species; spin densities and charges; estimation of activation barriers for electron transfer reactions by the Marcus theory; potential energy profiles for the hydrogen abstraction from the acetylamino nitrogen of the phenoxy radical of PAR giving rise to NAPQI by Cpd I of P450; potential energy profiles for the nonenzymatic 3-hydroxy-PAR, p-benzoquinone, and acetamide formation catalyzed by one water molecule; potential energy profile for the nonenzymatic water addition to NAPQI as first step of its hydrolysis; Cartesian coordinates of all molecular structures discussed in this work. This material is available free of charge via the Internet at http://pubs.acs.org.

a

In the high-spin (quartet) state, the initial hydrogen atom abstraction from the phenol oxygen of PAR (HAT(O) reaction) yields NAPSQI (N-acetyl-p-semiquinone imine), which is metabolized further through a subsequent HAT(N) step to the electrophile NAPQI (N-acetyl-pbenzoquinone imine). By contrast, the alternative initial low-spin (doublet) HAT(O) attack leads to the metabolites 3-hydroxyparacetamol, benzoquinone (as an electrophile and redox-active agent), and acetamide (CH3CONH2).



AUTHOR INFORMATION

Corresponding Authors

found experimentally. While the intermediate N-acetyl-psemiquinone imine (NAPSQI) as a precursor of the electrophile N-acetyl-p-benzoquinone imine (NAPQI) is formed only on the high-spin (quartet) route, 3-hydroxy-paracetamol, benzoquinone (as an electrophilic and redox-active metabolite), and acetamide are predominantly generated through the lowspin (doublet) pathway.

*(L.J.) E-mail: [email protected]. *(G.S.) E-mail: [email protected]. Tel: +49-341-2351262. Fax: +49-341-235-1785. Funding

This work was supported by the National Natural Science Foundation of China (grant no. 21307107), the Fundamental Research Funds for the Central Universities of China, and the EU project OSIRIS (no. GOCE-CT-2007-037017).



CONCLUSIONS Our DFT computational analysis of the P450-catalyzed biotransformation of paracetamol (PAR) suggests that, among the different reaction mechanisms (hydrogen atom transfer (HAT), oxygen addition reaction (OAR), and electron transfer) discussed so far, only the HAT pathway contributes significantly to the formation of the observed metabolites Nacetyl-p-benzoquinone imine (NAPQI), p-benzoquinone, acetamide, and 3-hydroxy-PAR. Surprisingly, NAPQI formation appears to be confined to the high-spin (HS) route of the HAT reaction, passing through N-acetyl-p-semiquinone imine (NAPSQI) as an initial phenoxy radical intermediate catalyzed by Cpd I (FeO), which is followed by a second H abstraction

Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We gratefully acknowledge the high performance computing cluster of the UFZ for providing computational resources, and the China National Supercomputing Center in Shenzhen for providing Gaussian 09 software and CPU time. Moreover, G.S. dedicates this work to the 80th birthday (in autumn 2014) of Prof. Dr. Martin Klessinger, former Chair of Theoretical Organic Chemistry at the University of Münster in Germany. J

DOI: 10.1021/tx5003645 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology



(16) Alves, C. N., Borges, R. S., and Da Silva, A. B. F. (2006) Density functional theory study of metabolic derivatives of the oxidation of paracetamol. Int. J. Quantum Chem. 106, 2617−2623. (17) Becke, A. D. (1993) Density-functional thermochemistry. 3. The role of exact exchange. J. Chem. Phys. 98, 5648−5652. (18) Lee, C. T., Yang, W. T., and Parr, R. G. (1988) Development of the Colle−Salvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B 37, 785−789. (19) Hay, P. J., and Wadt, W. R. (1985) Ab initio effective core potentials for molecular calculations. Potentials for K to Au including the outermost core orbitals. J. Chem. Phys. 82, 299−310. (20) Hay, P. J., and Wadt, W. R. (1985) Ab initio effective core potentials for molecular calculations. Potentials for the transition metal atoms Sc to Hg. J. Chem. Phys. 82, 270−283. (21) Wadt, W. R., and Hay, P. J. (1985) Ab initio effective core potentials for molecular calculations. Potentials for main group elements Na to Bi. J. Chem. Phys. 82, 284−298. (22) Kumar, D., Karamzadeh, B., Sastry, G. N., and de Visser, S. P. (2010) What factors influence the rate constant of substrate epoxidation by Compound I of cytochrome P450 and analogous iron(IV)-oxo oxidants? J. Am. Chem. Soc. 132, 7656−7667. (23) Perdew, J. P., and Wang, Y. (1992) Accurate and simple analytic representation of the electron−gas correlation energy. Phys. Rev. B 45, 13244−13249. (24) Miertus, S., Scrocco, E., and Tomasi, J. (1981) Electrostatic interaction of a solute with a continuum. A direct utilization of ab initio molecular potentials for the prevision of solvent effects. Chem. Phys. 55, 117−129. (25) Grimme, S. (2006) Semiempirical GGA-type density functional constructed with a long-range dispersion correction. J. Comput. Chem. 27, 1787−1799. (26) Lonsdale, R., Harvey, J. N., and Mulholland, A. J. (2010) Inclusion of dispersion effects significantly improves accuracy of calculated reaction barriers for cytochrome P450 catalyzed reactions. J. Phys. Chem. Lett. 1, 3232−3237. (27) Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A., Cheeseman, J. R., Montgomery, J. A., Jr., Vreven, T., Kudin, K. N., Burant, J. C., Millam, J. M., Iyengar, S. S., Tomasi, J., Barone, V., Mennucci, B., Cossi, M., Scalmani, G., Rega, N., Petersson, G. A., Nakatsuji, H., Hada, M., Ehara, M., Toyota, K., Fukuda, R., Hasegawa, J., Ishida, M., Nakajima, T., Honda, Y., Kitao, O., Nakai, H., Klene, M., Li, X., Knox, J. E., Hratchian, H. P., Cross, J. B., Bakken, V., Adamo, C., Jaramillo, J., Gomperts, R., Stratmann, R. E., Yazyev, O., Austin, A. J., Cammi, R., Pomelli, C., Ochterski, J. W., Ayala, P. Y., Morokuma, K., Voth, G. A., Salvador, P., Dannenberg, J. J., Zakrzewski, V. G., Dapprich, S., Daniels, A. D., Strain, M. C., Farkas, O., Malick, D. K., Rabuck, A. D., Raghavachari, K., Foresman, J. B., Ortiz, J. V., Cui, Q., Baboul, A. G., Clifford, S., Cioslowski, J., Stefanov, B. B., Liu, G., Liashenko, A., Piskorz, P., Komaromi, I., Martin, R. L., Fox, D. J., Keith, T., Al-Laham, M. A., Peng, C. Y., Nanayakkara, A., Challacombe, M., Gill, P. M. W., Johnson, B., Chen, W., Wong, M. W., Gonzalez, C., and Pople, J. A. (2004) Gaussian 03, Revision C.02, Gaussian, Inc., Wallingford, CT. (28) Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A., Cheeseman, J. R., Scalmani, G., Barone, V., Mennucci, B., Petersson, G. A., Nakatsuji, H., Caricato, M., Li, X., Hratchian, H. P., Izmaylov, A. F., Bloino, J., Zheng, G., Sonnenberg, J. L., Hada, M., Ehara, M., Toyota, K., Fukuda, R., Hasegawa, J., Ishida, M., Nakajima, T., Honda, Y., Kitao, O., Nakai, H., Vreven, T., Montgomery, J. A., Jr., Peralta, J. E., Ogliaro, F., Bearpark, M., Heyd, J. J., Brothers, E., Kudin, K. N., Staroverov, V. N., Kobayashi, R., Normand, J., Raghavachari, K., Rendell, A., Burant, J. C., Iyengar, S. S., Tomasi, J., Cossi, M., Rega, N., Millam, N. J., Klene, M., Knox, J. E., Cross, J. B., Bakken, V., Adamo, C., Jaramillo, J., Gomperts, R., Stratmann, R. E., Yazyev, O., Austin, A. J., Cammi, R., Pomelli, C., Ochterski, J. W., Martin, R. L., Morokuma, K., Zakrzewski, V. G., Voth, G. A., Salvador, P., Dannenberg, J. J., Dapprich, S., Daniels, A. D., Farkas, Ö ., Foresman, J. B., Ortiz, J. V., Cioslowski, J., and Fox, D. J. (2013) Gaussian 09, Revision D.01, Gaussian, Inc.: Wallingford, CT.

ABBREVIATIONS PAR, paracetamol; NAPQI, N-acetyl-p-benzoquinone imine; Cpd I, Compound I; Cpd II, Compound II; LS, low-spin; HS, high-spin; DFT, density functional theory; PCM, polarized continuum-solvation model; ZPE, zero-point energy; HAT, hydrogen abstraction transfer; OAR, oxygen addition reaction; SET, single-electron transfer; NAPSQI, N-acetyl-p-semiquinone imine; PT, proton transfer; TS, transition state; IRC, intrinsic reaction coordinate



REFERENCES

(1) Bessems, J. G. M., and Vermeulen, N. P. E. (2001) Paracetamol (acetaminophen)-induced toxicity: molecular and biochemical mechanisms, analogues and protective approaches. Crit. Rev. Toxicol. 31, 55−138. (2) Jaeschke, H., McGill, M. R., and Ramachandran, A. (2012) Oxidant stress, mitochondria, and cell death mechanisms in druginduced liver injury: lessons learned from acetaminophen hepatotoxicity. Drug Metab. Rev. 44, 88−106. (3) Jaeschke, H., Williams, C. D., Ramachandran, A., and Bajt, M. L. (2012) Acetaminophen hepatotoxicity and repair: the role of sterile inflammation and innate immunity. Liver Int. 32, 8−20. (4) James, L. P., Mayeux, P. R., and Hinson, J. A. (2003) Acetaminophen-induced hepatotoxicity. Drug Metab. Dispos. 31, 1499−1506. (5) Vermeulen, N. P., Bessems, J. G., and van de Straat, R. (1992) Molecular aspects of paracetamol-induced hepatotoxicity and its mechanism-based prevention. Drug Metab. Rev. 24, 367−407. (6) Narayan, N., Lee, I. H., Borenstein, R., Sun, J., Wong, R., Tong, G., Fergusson, M. M., Liu, J., Rovira, I. I., Cheng, H. L., Wang, G., Gucek, M., Lombard, D., Alt, F. W., Sack, M. N., Murphy, E., Cao, L., and Finkel, T. (2012) The NAD-dependent deacetylase SIRT2 is required for programmed necrosis. Nature 492, 199−204. (7) Zhou, W., and Yuan, J. (2012) Cell biology: death by deacetylation. Nature 492, 194−195. (8) Dahlin, D. C., Miwa, G. T., Lu, A. Y. H., and Nelson, S. D. (1984) N-Acetyl-p-benzoquinone imine: a cytochrome P-450-mediated oxidation product of acetaminophen. Proc. Natl. Acad. Sci. U.S.A. 81, 1327−1331. (9) Böhme, A., Thaens, D., Paschke, A., and Schüürmann, G. (2009) Kinetic glutathione chemoassay to quantify thiol reactivity of organic electrophilesapplication to α,β-unsaturated ketones, acrylates, and propiolates. Chem. Res. Toxicol. 22, 742−750. (10) Mulliner, D., Wondrousch, D., and Schüürmann, G. (2011) Predicting Michael-acceptor reactivity and toxicity through quantum chemical transition-state calculations. Org. Biomol. Chem. 9, 8400− 8412. (11) Wondrousch, D., Böhme, A., Thaens, D., Ost, N., and Schüürmann, G. (2010) Local electrophilicity predicts toxicity-relevant reactivity of Michael acceptors. J. Phys. Chem. Lett. 1, 1605−1610. (12) Koymans, L., van Lenthe, J. H., van de Straat, R., Donné-Op den Kelder, G. M., and Vermeulen, N. P. E. (1989) A theoretical study on the metabolic activation of paracetamol by cytochrome P-450: indications for a uniform oxidation mechanism. Chem. Res. Toxicol. 2, 60−66. (13) Van de Straat, R., Vromans, R. M., Bosman, P., de Vries, J., and Vermeulen, N. P. E. (1988) Cytochrome P-450-mediated oxidation of substrates by electron-transfer; role of oxygen radicals and of 1electron and 2-electron oxidation of paracetamol. Chem.−Biol. Interact. 64, 267−280. (14) Lai, W., and Shaik, S. (2011) Can ferric-superoxide act as a potential oxidant in P450cam? QM/MM investigation of hydroxylation, epoxidation, and sulfoxidation. J. Am. Chem. Soc. 133, 5444−5452. (15) Borovik, A. S. (2011) Role of metal-oxo complexes in the cleavage of C−H bonds. Chem. Soc. Rev. 40, 1870−1874. K

DOI: 10.1021/tx5003645 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology (29) Marcus, R. A. (1993) Electron transfer reactions in chemistry theory and experiment (Nobel Lecture). Angew. Chem., Int. Ed. Engl. 32, 1111−1121; () Angew. Chem. 105, 1161−1172. (30) Ji, L., Zhang, J., Liu, W., and de Visser, S. P. (2014) Metabolism of halogenated alkanes by cytochrome P450 enzymes. Aerobic oxidation versus anaerobic reduction. Chem.Asian J. 9, 1175−1182. (31) Ogliaro, F., de Visser, S. P., and Shaik, S. (2002) The ‘push’ effect of the thiolate ligand in cytochrome P450: a theoretical gauging. J. Inorg. Biochem. 91, 554−567. (32) Li, C., Wu, W., Cho, K. B., and Shaik, S. (2009) Oxidation of tertiary amines by cytochrome P450-kinetic isotope effect as a spinstate reactivity probe. Chem.Eur. J. 15, 8492−8503. (33) Li, C., Zhang, L., Zhang, C., Hirao, H., Wu, W., and Shaik, S. (2007) Which oxidant is really responsible for sulfur oxidation by cytochrome P450? Angew. Chem., Int. Ed. 46, 8168−8170; () Angew. Chem. 119, 8316−8318. (34) Rydberg, P., Ryde, U., and Olsen, L. (2008) Sulfoxide, sulfur, and nitrogen oxidation and dealkylation by cytochrome P450. J. Chem. Theory Comput. 4, 1369−1377. (35) Calder, I. C., Hart, S. J., Healey, K., and Ham, K. N. (1981) NHydroxyacetaminophen: a postulated toxic metabolite of acetaminophen. J. Med. Chem. 24, 988−993. (36) Shaik, S., Kumar, D., de Visser, S. P., Altun, A., and Thiel, W. (2005) Theoretical perspective on the structure and mechanism of cytochrome P450 enzymes. Chem. Rev. 105, 2279−2328. (37) Chen, H., Song, J. S., Lai, W. Z., Wu, W., and Shaik, S. (2010) Multiple low-lying states for Compound I of P450 cam and chloroperoxidase revealed from multireference ab initio QM/MM calculations. J. Chem. Theory Comput. 6, 940−953. (38) De Visser, S. P., Kumar, D., Cohen, S., Shacham, R., and Shaik, S. (2004) A predictive pattern of computed barriers for C−H hydroxylation by compound I of cytochrome P450. J. Am. Chem. Soc. 126, 8362−8363. (39) Shaik, S., Kumar, D., and de Visser, S. P. (2008) Valence bond modeling of trends in hydrogen abstraction barriers and transition states of hydroxylation reactions catalyzed by cytochrome P450 enzymes. J. Am. Chem. Soc. 130, 10128−10140. (40) De Visser, S. P., Ogliaro, F., Sharma, P. K., and Shaik, S. (2002) What factors affect the regioselectivity of oxidation by cytochrome P450? A DFT study of allylic hydroxylation and double bond epoxidation in a model reaction. J. Am. Chem. Soc. 124, 11809−11826. (41) Ogliaro, F., Harris, N., Cohen, S., Filatov, M., de Visser, S. P., and Shaik, S. (2000) A model “rebound” mechanism of hydroxylation by cytochrome P450: stepwise and effectively concerted pathways, and their reactivity patterns. J. Am. Chem. Soc. 122, 8977−8989. (42) Van de Straat, R., de Vries, J., de Boer, H. J. R., Vromans, R. M., and Vermeulen, N. P. E. (1987) Relationship between paracetamol binding to and its oxidation by two cytochromes P-450 isozymesa proton nuclear magnetic resonance and spectrophotometric study. Xenobiotica 17, 1−9. (43) Lee, C. A., Manyike, P. T., Thummel, K. E., Nelson, S. D., and Slattery, J. T. (1997) Mechanism of cytochrome P450 activation by caffeine and 7,8-benzoflavone in rat liver microsomes. Drug Metab. Dispos. 25, 1150−1156. (44) Cameron, M. D., Wen, B., Roberts, A. G., Atkins, W. M., Campbell, A. P., and Nelson, S. D. (2007) Cooperative binding of acetaminophen and caffeine within the P450 3A4 active site. Chem. Res. Toxicol. 20, 1434−1441. (45) Bessems, J. G., de Groot, M. J., Baede, E. J., te Koppele, J. M., and Vermeulen, N. P. E. (1998) Hydrogen atom abstraction of 3,5disubstituted analogues of paracetamol by horseradish peroxidase and cytochrome P450. Xenobiotica 28, 855−875. (46) Hackett, J. C., Brueggemeier, R. W., and Hadad, C. M. (2005) The final catalytic step of cytochrome P450 aromatase: a density functional theory study. J. Am. Chem. Soc. 127, 5224−5237. (47) Sen, K., and Hackett, J. C. (2012) Coupled electron transfer and proton hopping in the final step of CYP19-catalyzed androgen aromatization. Biochemistry 51, 3039−3049.

(48) Sen, K., and Hackett, J. C. (2010) Peroxo-iron mediated deformylation in sterol 14alpha-demethylase catalysis. J. Am. Chem. Soc. 132, 10293−10305. (49) Wang, Y., Yang, C., Wang, H., Han, K., and Shaik, S. (2007) A new mechanism for ethanol oxidation mediated by cytochrome P450 2E1: bulk polarity of the active site makes a difference. ChemBioChem 8, 277−281. (50) De Visser, S. P., and Tan, L. S. (2008) Is the bound substrate in nitric oxide synthase protonated or neutral and what is the active oxidant that performs substrate hydroxylation? J. Am. Chem. Soc. 130, 12961−12974. (51) Schyman, P., Lai, W., Chen, H., Wang, Y., and Shaik, S. (2011) The directive of the protein: how does cytochrome P450 select the mechanism of dopamine formation? J. Am. Chem. Soc. 133, 7977− 7984. (52) Kumar, D., de Visser, S. P., and Shaik, S. (2005) Theory favors a stepwise mechanism of porphyrin degradation by a ferric hydroperoxide model of the active species of heme oxygenase. J. Am. Chem. Soc. 127, 8204−8213. (53) Ji, L., and Schüürmann, G. (2013) Model and mechanism: Nhydroxylation of primary aromatic amines by cytochrome P450. Angew. Chem., Int. Ed. 52, 744−748; () Angew. Chem. 125, 772−776. (54) Ji, L., and Schüürmann, G. (2012) Computational evidence for α-nitrosamino radical as initial metabolite for both the P450 dealkylation and denitrosation of carcinogenic nitrosamines. J. Phys. Chem. B 116, 903−912.

L

DOI: 10.1021/tx5003645 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX