Review
Two-state reactivity mechanisms of hydroxylation and epoxidation by cytochrome P-450 revealed by theory

https://doi.org/10.1016/S1367-5931(02)00363-0Get rights and content

Abstract

Recent computational studies of alkane hydroxylation and alkene epoxidation by a model active species of the enzyme cytochrome P-450 reveal a two-state reactivity (TSR) scenario in which the information content of the product distribution is determined jointly by two states. TSR is used to reconcile the dilemma of the consensus ‘rebound mechanism’ of alkane hydroxylation, which emerged from experimental studies of ultra-fast radical clocks. The dilemma, stated succinctly as ‘radicals are both present and absent and the rebound mechanism is both right and wrong’, is simply understood once one is cognizant that the mechanism operates by two states, one low-spin (LS) the other high-spin (HS). In both states, bond activation proceeds in a manner akin to the rebound mechanism, but the LS mechanism is effectively concerted, whereas the HS is stepwise with incursion of radical intermediates.

Introduction

Cytochromes P-450s constitute a super-family of enzymes that appear in all aerobic bioorganisms and perform vital bioregulatory functions such as detoxification and biosynthesis of sex hormones, as well as of compounds having anti-inflammatory and anti-hypertensive roles [1]. It is no wonder, therefore, that these enzymes present an exciting agenda, directed to identify the key species that appear in the respective catalytic cycles and the mechanisms by which the active species oxidize organic compounds. Two of the most important reactions are hydroxylation of CH and epoxidation of CC bonds. The mechanisms of both reactions have been subjects of intense study but, nevertheless, many of the crucial mechanistic features remain disputed. Let us exemplify these controversies by reference to the hydroxylation reaction [1].

The consensus [2] mechanism formulated in the pioneering studies of Groves [3] is the rebound mechanism shown in Fig. 1. This involves initial hydrogen abstraction from the alkane (RH) by the active iron-oxo species 1, known also as Compound I (Cpd I) 1., 2., 3., 4., 5., 6., followed by radical rebound on the iron-hydroxo intermediate 2, to generate the ferric-alcohol complex 3, which then releases the alcohol and restores the resting state (i.e. the water complex 4). The rebound mechanism has gained support from large kinetic isotope effects (KIEs), which indicated CH bond breaking in the rate-determining step, and isolation of rearranged alcohol products, which points to the presence of radicals with finite lifetimes 2., 3., 4., 5., 6., 7., 8., 9.. Evidence for radicals emerges, for example, from the hydroxylation products 6 and 7 obtained from deuterated norbornane 5 (Fig. 2) 10., 11. In this case, both cytochrome P-450LM2 [10] and artificial models [11] of Cpd I converge to the same conclusion that a free norbornyl radical exists as an intermediate, which is subsequently hydroxylated from its two faces, thereby leading to both exo- and endo-alcohols with corresponding scrambling of the deuterium.

Additional support for the rebound mechanism has recently been provided by the KIE measurements of Dinnocenzo, Jones and co-workers [12] who found that for a series of aryl-methanes, the KIE values of hydroxylation are virtually equal to those of a corresponding hydrogen abstraction reaction by a t-C4H9O radical, thus implying isostructural transition states for the two processes with a colinear OHC arrangement of the reacting atoms in the transition state.

The picture started to cloud, however, when radical clocks were used to gauge the rate of the rebound step [2]. Thus, several studies by Newcomb and co-workers 13••., 14••. using ultra-fast radical clocks, have cast doubts on the presence of free radicals. Apparent lifetimes are too short to correspond to free radical intermediates (e.g. τ = 80–200 fs), and have no correlation with independently clocked rearrangement lifetimes of the free radicals. To elucidate the mechanism, Newcomb and co-workers have used clocks such as 8 (Fig. 3a) which can undergo rearrangement in two distinct modes: one is typical for a cyclopropyl methyl radical and the other for a cyclopropyl methyl cation. On the basis of the rearranged products, the ‘two-oxidant hypotheses’ 15., 16. has been suggested as an alternative mechanism 13••., 14••., whereby cytochrome P-450 operates via two oxidants (Fig. 3b). The primary oxidant is Cpd I (1), and the second oxidant, a ferric-peroxide species (9), is suggested to insert OH+ and lead to the protonated alcohol that rearranges via anchimeric assistance in a manner typical of cationic species. The extent of rearranged products that do not originate in carbocations is associated with apparent lifetimes that are too short to be assigned to free radicals 13••., 14••.. This has led Newcomb to conclude that radicals are not present and that the rebound mechanism cannot account for hydroxylation.

The Groves–Newcomb alternative mechanisms are both supported by seemingly straightforward evidence. Thus, one set of experiments, based on rearrangement data and KIE profiles, supports the rebound mechanism and the presence of radicals. The other set of experiments, based on radical lifetimes extracted from the rearrangement patterns of radical clocks, questions the validly of both the rebound mechanism and the presence of radicals. This situation poses an intellectually fascinating dilemma that can be stated as follows: radicals are both present and absent and the rebound mechanism is both right and wrong.

Although the dilemma has not been as focused for the mechanism of olefin epoxidation by cytochrome P-450, similar and even more complex problems appear in this mechanism too because of the involvement of intermediates with different oxidation states 3., 17., 18. on the iron centre (FeIII and FeIV) and on the alkene unit (radical, cation, radical–cation) as well as incursion of aldehydes and suicidal complexes that lead to destruction of the porphyrin 3., 19..

The mechanistic dilemma posed by the reactivity patterns is perhaps too subtle for current experimental means. Here, theory must come into play and offer complementary insight, which can possibly guide experiment in a new direction. This paper reviews some recent theoretical elucidations of the mechanism by which Cpd I hydroxylates and epoxidizes organic compounds 20••., 21••., 22., 23., 24•.. The theoretical calculations reveal that the reactions of Cpd I proceed by two-state reactivity (TSR), which involves high-spin (HS) and low-spin (LS) states that jointly determine the mechanism and product distribution. Consideration of the reactivity in terms of TSR offers a resolution of the essential elements of the dilemma.

Section snippets

Background: TSR of FeO+ — a model iron-oxo species and its analogy to the electronic structure of Cpd I

The recognition that two electronic states contribute to the overall reactivity in the gas-phase bond activation by bare transition-metal cations can be traced back to the pioneering studies of Armentrout using guided ion beams [25]. Subsequently, the Berlin group reported that the ‘hydroxylation’ of H2 by the smallest iron-oxo species (i.e. the diatomic FeO+ cation in the gas phase) is sluggish and proceeds with efficiency of 10−3–10−2 despite the facts that the reaction is highly exothermic,

The ground state of Cpd I of P-450

Fig. 6 shows the key orbitals and their occupancy in the low-lying electronic states of Cpd I. These are the five d-block orbitals, and two additional orbitals, labeled as ‘a2u’ 3., 38., 39., 40••., 41., 42. and pπ(S). The ‘a2u’ orbital is a mixed a2u(Por)–pσ(S) orbital with a′ symmetry in Cs and no-symmetry in C1, but for brevity we call it in the following discussion a2u after its parent porphyrin orbital, using the common nomenclature in the literature 38., 39.. The second orbital is termed p

TSR for alkane hydroxylation by Cpd I

Fig. 7 shows energetic features of the critical species for the CH bond hydroxylation of methane as a model alkane [20••]. The reaction has two phases differing in their functionality. The first corresponds to CH bond activation and the second, a rebound phase, to the CO bond making. The first phase is characterized by a pair of closely lying transition states, 2,4TSH, which are nascent from the 2,4A2u states of Cpd I, discussed above by appeal to Fig. 6.

The LS and HS profiles remain

Why do lifetimes of radical clocks appear too short?

In radical-clock experiments, the apparent lifetime of the radical (τapp) is customarily determined from the measured ratio of unrearranged (U) to rearranged (R) alcohol products, and the knowledge of the rearrangement rate constant for the putative free radical clock (kr), as shown in Eq. (1) and Eq. (2):kreb(app)=kr([U/R])τapp=1/kreb(app)=[1/kr]/[U/R]

In a classical situation of SSR, the expression in Eq. (2) is perfectly correct. However, Eq. (2) may become misleading under TSR conditions,

Reconciliation of the Newcomb–Groves/Dinnocenzo–Jones dilemma

The TSR scenario offers a satisfactory solution of the mechanistic controversy discussed in the introduction. Thus, the argument raised by Dinnocenzo and Jones [12] that the KIE for hydroxylation is virtually identical to genuine hydrogen abstraction, is in perfect harmony with the computations, which reveal that the transition states for hydroxylation are hydrogen abstraction-like. Indeed, KIE appears as a sensitive probe for the transition states of hydroxylation. In turn, the apparent

TSR in ethene epoxidation by Cpd I

Fig. 9 shows the reaction profiles for the epoxidation of ethene by Cpd I [21••]. An identical picture emerged recently for propene epoxidation [43]. Much as in alkane hydroxylation, here too, there are both LS and HS processes. The HS process is stepwise with a barrier for ring closure to yield the ferric-epoxide complex, whereas the LS process is effectively concerted. The HS and LS transition states for the CC bond activation are very similar and involve a one centre OC attack of Cpd I on

The multistate nature of the rebound phase

The difference between the HS and LS rebounds is fundamental and merits a few comments. The TSR scenario was reproduced in recent results of Yoshizawa et al. [24•] who found that the LS and HS states for the hydroxylation of ethane by another Cpd I model (with L = CH3S) are energetically close. However, in this later study, both HS and LS states have significant barriers for rebound (see, however, Update). These results, obtained with a small basis set, are in contrast with previous higher

Predictions of rearrangement patterns in alkane hydroxylation

The ratio of rearranged to unrearranged alcohol, [U/R], is an experimentally accessible quantity. As such, it is important to outline factors that control this quantity. One factor is the relative concentration of HS/LS complexes that reach the bifurcation point of the rebound phase (Fig. 7). The second factor depends on the partition of the HS radical into its various available pathways.

The relative concentration of HS/LS complexes near the bifurcation point of the two phases of Fig. 7 depends

TSR vis-à-vis other hypotheses

There could exist alternative mechanistic scenarios, which can in principle resolve the apparent controversies. These hypotheses have in common that they rule out operation of a single mechanism, and instead involve at least two pathways. One such conceivable hypothesis is that one of these pathways involves a synchronous concerted oxygen insertion into the CH bond. Another hypothesis, which was discussed originally by Coon and co-workers [16], involves two oxidant species, Cpd I and

Conclusion

The active species of cytochrome P-450, Cpd I, is a two-state reagent, having closely lying quartet (HS) and doublet (LS) states. DFT investigations of alkane hydroxylation and alkene epoxidation show that these two states are competitive. As such, oxidation by Cpd I occurs by a TSR scenario, whereby reactivity patterns and product distribution are determined by the interplay of the two states 20••., 21••..

The reaction mechanism consists of two phases (Fig. 7 and 9): a bond activation phase,

Update

Recent work by Yoshizawa (Yoshizawa K, unpublished data) reveals that the LS rebound has indeed a tiny barrier of < 1 kcal mol−1, in better accord now with the picture in 20••., 43., 49.. Jones et al. [66] have shown that sulfoxidation and N-dealkylation by P-450 arise from two different enzyme species. This result was concluded to be consistent either with TSR or with the two-oxidant hypothesis. Recent investigation of camphor hydroxylation by P450cam by Davydov et al. [67] showed that product

Acknowledgements

This research at the HU was supported by the Israeli Science Foundation (ISF), and in part by the Ministry of Sciences, Culture and Sports. The Berlin group acknowledges continuous financial support by the Deutsche Forschungsgemeinschaft, the Fonds der Chemischen Industrie, and the Gesellschaft von Freunden der TU Berlin.

References and recommended reading

Papers of particular interest, published within the annual period of review, have been highlighted as:

  • • of special interest

  • •• of outstanding interest

References (68)

  • J.H Dawson

    Probing structure-function relations in heme-containing oxygenases and peroxidases

    Science

    (1988)
  • B Meunier

    Metalloporphyrins as versatile catalysts for oxidation reactions and oxidative DNA cleavage

    Chem Rev

    (1992)
  • E.J Mueller et al.

    Twenty-five years of P-450cam research

  • P.R Ortiz de Montellano

    Control of the catalytic activity of prosthetic heme by the structure hemoproteins

    Acc Chem Res

    (1987)
  • T.G Traylor et al.

    Aliphatic hydroxylation catalyzed by iron(III) porphyrins

    J Am Chem Soc

    (1992)
  • J.I Manchester et al.

    A new mechanistic probe for cytochrome P-450: An application of isotope effect profiles

    J Am Chem Soc

    (1997)
  • M Newcomb et al.

    Hypersensitive radical probes and the mechanisms of cytochrome P-450 catalyzed hydroxylation reactions

    Acc Chem Res

    (2000)
  • M Newcomb et al.

    Cytochrome P-450-catalyzed hydroxylation of mechanistic probes that distinguish between radicals and cations. Evidence for cationic but not radical intermediates

    J Am Chem Soc

    (2000)
  • A.D.N Vaz et al.

    Peroxo-iron and xenoid-iron species as alternative oxygenating agents in cytochrome P-450-catalyzed reactions: switching by threonine-302 to alanine mutagenesis of cytochrome P-450 2B4

    Proc Natl Acad Sci USA

    (1996)
  • A.D.N Vaz et al.

    Epoxidation of olefins by cytochrome P-450: evidence from site-specific mutagenesis for hydroperoxo-iron as an electrophilic oxidant

    Proc Natl Acad Sci USA

    (1998)
  • Z Gross et al.

    Reaction profile of the last step in cytochrome P-450 catalysis revealed by studies of model complexes

    J Biol Inorg Chem

    (1997)
  • J.T Groves et al.

    Preparation and reactivity of oxoiron(IV) porphyrins

    Inorg Chem

    (1994)
  • S.P de Visser et al.

    How does ethene inactivate cytochrome P-450 en route to its epoxidation? A density functional study

    Angew Chem Int Ed Engl

    (2001)
  • F Ogliaro et al.

    A model “rebound” mechanism of hydroxylation by cytochrome P-450: a stepwise and effectively concerted pathways, and their reactivity patterns

    J Am Chem Soc

    (2000)
  • S.P de Visser et al.

    Multi-state epoxidation of ethene by cytochrome P-450: a quantum chemical study

    J Am Chem Soc

    (2001)
  • de Visser SP, Ogliaro F, Shaik S: Stereospecific oxidation by Compound I of cytochrome P-450 does not proceed in a...
  • K Yoshizawa et al.

    Kinetic isotope effects in a CH bond dissociation by the iron-oxo species of cytochrome P-450

    J Phys Chem B

    (2000)
  • K Yoshizawa et al.

    A theoretical study of the dynamic behavior of alkane hydroxylation by a Compound I model of cytochrome P-450

    J Am Chem Soc

    (2001)
  • P.B Armentrout

    Chemistry of excited electronic states

    Science

    (1991)
  • D Schröder et al.

    Surprisingly low reactivity of base FeO+ in its spin-allowed, highly exothermic reaction with molecular hydrogen to generate Fe+ and water

    J Phys Chem

    (1994)
  • A Fiedler et al.

    Electronic structures and gas-phase reactivities of cationic late transition-metal oxides

    J Am Chem Soc

    (1994)
  • D.E Clemmer et al.

    State-specific reactions of Fe+ (a 6D, a 4F) with D2O and reactions of FeO+ with D2

    J Phys Chem

    (1994)
  • S Shaik et al.

    Two-state reactivity in organometallic gas-phase ion chemistry

    Helv Chim Acta

    (1995)
  • D Danovich et al.

    Spin-orbit coupling in the oxidative activation of HH by FeO+. Selection rules and reactivity effects

    J Am Chem Soc

    (1997)
  • Cited by (313)

    • Computational insight into biotransformation of halophenols by cytochrome P450: Mechanism and reactivity for epoxidation

      2022, Chemosphere
      Citation Excerpt :

      It is also notable that there are spin-crossing effects between LS and HS states during the epoxidation of 2,4,6-trichlorophenol and 2,4,6-trifluorophenol, as shown in Fig. 1, which may potentially affect the reaction mechanism. This phenomenon probably involves the participation of spin inversion in the rate-determining O-addition step and thus results in a two-state reactivity (TSR) process (Schröder et al., 2000; de Visser et al., 2001a; Shaik et al., 2002). By contrast, for some other XPs, such as 2-chlorophenol (Table S9), 2,3,5-trifluorophenol (Table S43), and 2-bromophenol (Table S45), they have no spin crossover phenomenon and show relatively significant energy difference between the LS and HS O-addition barriers, which may only lead to a spin-selective single-state reactivity (SSR) (Shaik et al., 2007; Wang et al., 2007).

    • Iron and manganese biomimetic compounds

      2021, Comprehensive Coordination Chemistry III
    • Ion-molecule reactions of mass-selected ions

      2024, Mass Spectrometry Reviews
    View all citing articles on Scopus
    View full text