Revisiting the mechanism of P450 enzymes with the radical clocks norcarane and spiro[2,5]octane

Article abstract of DOI:10.1021/ja025608h

Norcarane (1) and spiro[2.5]octane (2) yield different product distributions depending on whether they are oxidized via concerted, radical, or cationic mechanisms. For this reason, these two probes were used to investigate the mechanisms of hydrocarbon hydroxylation by two mammalian and two bacterial cytochrome P450 enzymes. Products indicative of a radical intermediate with a lifetime ranging from 16 to 52 ps were detected during the oxidation of norcarane by P450_cam (CYP101), P450_BM3 (CYP102), CYP2B1, and CYP2E1. Trace amounts of the cation rearrangement product were observed with norcarane for all but CYP2E1, while no cation or radical rearrangement products were observed for spiro[2.5]octane. The results for the oxidation of norcarane with a radical rearrangement rate of 2 × 10⁸ s^-1 are consistent with the involvement of a two-state radical rebound mechanism, while for the slower (5 × 10⁷ s^-1) spiro[2,5]-oct-4-yl radical rearrangement products were beyond detection. Taken together with earlier data for the hydroxylation of bicyclo[2.1.0]pentane, which also suggested a 50 ps radical lifetime, these three structurally similar and functionally simple substrates show a consistent pattern of rearrangement that supports a radical rebound mechanism for this set of cytochrome P450 enzymes.

Full text of DOI:10.1021/ja025608h

Published on Web 05/02/2002
Revisiting the Mechanism of P450 Enzymes with the Radical
Clocks Norcarane and Spiro[2,5]octane
Karine Auclair,‡ Zhengbo Hu,† Dorothy M. Little,† Paul R. Ortiz de Montellano,*,‡ and
John T. Groves*,†
Contribution from the Department of Chemistry, Princeton UniVersity,
Princeton, New Jersey 08544, and Department of Pharmaceutical Chemistry,
UniVersity of California, San Francisco, California 94143-0446
Received January 15, 2002
Abstract: Norcarane (1) and spiro[2.5]octane (2) yield different product distributions depending on whether
they are oxidized via concerted, radical, or cationic mechanisms. For this reason, these two probes were
used to investigate the mechanisms of hydrocarbon hydroxylation by two mammalian and two bacterial
cytochrome P450 enzymes. Products indicative of a radical intermediate with a lifetime ranging from 16 to
52 ps were detected during the oxidation of norcarane by P450cam (CYP101), P450BM3 (CYP102), CYP2B1,
and CYP2E1. Trace amounts of the cation rearrangement product were observed with norcarane for all
but CYP2E1, while no cation or radical rearrangement products were observed for spiro[2.5]octane. The
results for the oxidation of norcarane with a radical rearrangement rate of 2 × 108 s-1 are consistent with
the involvement of a two-state radical rebound mechanism, while for the slower (5 × 107 s-1) spiro[2,5]-
oct-4-yl radical rearrangement products were beyond detection. Taken together with earlier data for the
hydroxylation of bicyclo[2.1.0]pentane, which also suggested a 50 ps radical lifetime, these three structurally
similar and functionally simple substrates show a consistent pattern of rearrangement that supports a radical
rebound mechanism for this set of cytochrome P450 enzymes.
Introduction
The cytochrome P450 (P450) enzyme family encompasses
heme proteins that catalyze a large diversity of transformations,
the most common of which are hydroxylation, epoxidation, and
heteroatom oxidation.1 The most generally accepted mechanism
for substrate hydroxylation by P450 enzymes consists of 7 steps
(Figure 1): (a) the 6-coordinate low-spin ferric enzyme binds
a substrate with a concomitant shift to a 5-coordinate high-spin
state; (b) transfer of one electron from a redox partner reduces
the ferric enzyme to the ferrous state; (c) the ferrous enzyme
binds molecular oxygen to produce the FeII-O2 complex; (d) a
second electron followed by a proton are transferred to the
ferrous-dioxy species to afford an iron-hydroperoxo (FeIII-OOH)
intermediate; (e) the O-O bond is cleaved to release a molecule
of water and an iron-oxo ferryl (formally FeVdO) oxidizing
species; (f) a hydrogen atom is abstracted from the substrate
by the iron-oxo intermediate to yield a one-electron reduced
ferryl species (FeIV-OH) and a substrate radical (R•), followed
by radical recombination to give the enzyme-product complex;
and (g) release of the product (ROH) to regenerate the initial
low-spin ferric enzyme. Despite decades of intensive research,
many details of this mechanism remain unclear. Crystal
Figure 1. Hydroxylation mechanism for P450 enzymes. The square
represents heme bound to the enzyme (R).
structures have recently been determined in the case of P450cam
(CYP101) for several of the intermediates along this reaction
pathway.2,3 However, to date, the iron-oxo porphyrin π-cation
radical of P450 enzymes has not been unambiguously observed,
although its presence has been inferred by low-temperature EPR,
spectroscopic, and possibly crystallographic studies.3-5 The
* Authors to whom correspondence should be addressed. J.T.G.: FAX
609-258-0348, e-mail jtgroves@princeton.edu. P.R.O.M.: FAX 415-502-
4728, e-mail ortiz@cgl.ucsf.edu.
‡ University of California.
(2) Poulos, T. L.; Raag, R. FASEB 1992, 6, 674-679.
(3) Schlichting, I.; Berendzen, J.; Chu, K.; Stock, A. M.; Maves, S. A.; Benson,
D. E.; Sweet, R. M.; Ringe, D.; Petsko, G. A.; Sligar, S. G. Science 2000,
287, 1615-1622.
† Princeton University.
(1) Ortiz de Montellano, P. R. Cytochrome P450: Structure, Mechanism, and
Biochemistry, 2nd ed.; Plenum: New York 1995.
9
6020
J. AM. CHEM. SOC. 2002, 124, 6020-6027
10.1021/ja025608h CCC: $22.00 © 2002 American Chemical Society

Radical Clocks and the Cytochrome P450 Mechanism
A R T I C L E S
Figure 2. Concerted versus rebound mechanisms by an oxoiron(IV)porphyrin cation radical species. The porphyrin is designated as P and the protein is
omitted.
Figure 3. Structure of bicyclo[2.1.0]pentane and its radical and cationic
rearrangement products.
quest for the exact mechanism of oxidation by the iron-oxo
species has given rise to contradictory conclusions over the past
two decades. The initial view of a concerted mechanism (Figure
2) was dismissed in the late 1970s based on the observation of
Figure 4. One example of an ultrafast radical clock capable of differentiat-
ing radical from cationic intermediates. See ref 16.
large intramolecular kinetic isotope effects (kH/kD > 10),6 loss
of stereochemistry,6-8 and substrate rearrangements.6,9,10a
A
nonconcerted hydroxylation pathway, referred to as the oxygen
rebound mechanism, was first suggested by Groves et al.10b to
account for these rearrangements. This process involves a
hydrogen atom abstraction from the substrate by the iron-oxo
species to yield a carbon-centered radical intermediate. Subse-
quent transfer of the iron-bound hydroxyl group to the substrate
radical would give the product alcohols in this scenario.
However, the high level of stereospecificity (>90% retention)
observed in most experiments suggested either that any radical
intermediate would have to be short-lived or that substrate
motion was restricted within the enzyme active site.11 The first
use of radical clock substrates in the P450 area by Ortiz de
Montellano and Stearns,12 followed by timing of the clock,13,14
suggested a lifetime of 50 ps for the radical intermediate formed
during the P450 oxidation of bicyclo[2.1.0]pentane (Figure 3).
Subsequent extensive work by Newcomb and co-workers with
ultrafast radical clock substrates able to differentiate between
radical and cationic pathways (Figure 4) indicated radical life-
times of 80-200 fs, which are in the order of the lifetimes of
transition states rather than intermediates.15,16 These experiments,
which also yielded products indicative of cationic intermediates,
led them to propose the involvement of two oxidizing species,
the iron-oxo complex that inserts an oxygen atom, and an iron-
hydroperoxo intermediate that inserts an OH+ species.
In an effort to reconcile the divergent results obtained with
radical clocks, Shaik and co-workers have proposed a “two-
state reactivity” process.17-20 Using density functional theory,
they showed that the iron-oxo species adopts two reactive states
that explain the formation of products via both the rebound
(radical) and insertion (concerted) mechanisms. The low spin
doublet state leads to a substrate radical transition state with
no energy barrier to rebound collapse, whereas the high spin
quartet generates a radical with a significant barrier. This clearly
suggests a finite lifetime for the radical intermediate, especially
for the high-spin reaction pathway. Although the proposed two-
state rebound mechanism explains the formation of radical
transition states or intermediates, the model is not complete and
provides no explanation for some of the available observations
(e.g. variations in isotope effects, or cationic products). The
(4) Davydov, R.; Makris, T. M.; Kofman, V.; Werst, D. E.; Sligar, S. G.;
Hoffman, B. M. J. Am. Chem. Soc. 2001, 123, 1403-1415.
(5) Denisov, I. G.; Makris, T. M.; Sligar, S. G. J. Biol. Chem. 2001, 276,
11648-11652.
(6) Groves, J. T.; McGlusky, G. A.; White, R. E.; Coon, M. J., Biochem.
Biophys. Res. Commun. 1978, 81, 154-160.
(15) (a) Newcomb, M.; Shen, R.; Choi, S.-Y.; Toy, P. T.; Hollenberg, P. F.;
Vaz, A. D. N.; Coon, M. J. J. Am. Chem. Soc. 2000, 122, 2677-2686. (b)
Toy, P. H.; Newcomb, M.; Hollenberg, P. F. J. Am. Chem. Soc. 1998,
120, 7719-7729. (c) We thank Prof. Martin Newcomb for a sample of
norcarane oxidation products produced by CYP2B4.
(7) Gelb, M. H.; Heimbrook, D. C.; Malkonen, P.; Sligar, S. G. Biochemistry
1982, 21, 370-377.
(8) White, R. E.; Miller, J. P.; Favreau, L. V.; Bhattacharyaa, A. J. Am. Chem.
Soc. 1986, 108, 6024-.
(9) Gelb, M. H.; Heimbrook, D. C.; Malkonen, P.; Sligar, S. G. Biochemistry
1982, 21, 370-377.
(16) Newcomb, M.; Toy, P. H. Acc. Chem. Res. 2000, 33, 449-455.
(17) (a) Ogliaro, F.; Harris, N.; Cohen, S.; Filatov, M.; de Visser, S. P.; Shaik,
S. J. Am. Chem. Soc. 2000, 122, 8977-8989. (b) Ogliaro, F.; de Visser, S.
P.; Groves, J. T.; Shaik, S. Angew. Chem., Int. Ed. 2001, 40, 2874-2878.
(18) (a) de Visser, S. P.; Ogliaro, F.; Shaik, S. J. Chem. Soc., Chem. Commun.
2001, 2322-2323. (b) Ogliaro, F.; de Visser, S. P.; Cohen, S.; Sharma, P.
K.; Shaik, S. J. Am. Chem. Soc. 2002, 124, 2806-2817.
(10) (a) Groves, J. T.; Subramanian, D. V. J. Am. Chem. Soc. 1984, 106, 2177-
2181. (b) Groves, J. T.; McClusky, G. A. J. Am. Chem. Soc. 1976, 98,
859.
(11) Fretz, H.; Woggon, W.-D.; Voges, R. HelV. Chim. Acta 1989, 72, 391-
400.
(12) Ortiz de Montellano, P. R.; Stearns, R. A. J. Am. Chem. Soc. 1987, 109,
3415-3420.
(19) Filatov, M.; Harris, N.; Shaik, S. Angew. Chem., Int. Ed. 1999, 38, 3510-
3512.
(13) Bowry, V. W.; Ingold, K. U. J. Am. Chem. Soc. 1991, 113, 5699-5707.
(14) Atkinson, J. K.; Ingold, K. U. Biochemistry 1993, 32, 9209-9214.
(20) Shaik, S.; Filatov, M.; Schroder, D.; Schwarz, H. Chem. Eur. J. 1998, 4,
193-199.
9
J. AM. CHEM. SOC. VOL. 124, NO. 21, 2002 6021

A R T I C L E S
Auclair et al.
Figure 5. Rearrangement pathways and products resulting from formation of a radical or a cation intermediate upon oxidation of norcarane (1) by P450
enzymes.
alternative reaction intermediates to the iron-oxo species have
recently been reviewed.21
to a substrate radical intermediate that is sometimes followed
by a second oxidation to give a cationic intermediate.24b The
second radical clock used in our experiments, spiro[2.5]octane
(2), was selected for three reasons: (a) it differentiates radical
and cationic intermediates (Figure 6); (b) it is oxidized at a
secondary carbon instead of a methyl group; and (c) its radical
has been shown to rearrange ca. 4 times slower (k ) 5 × 107
s-1) than that of norcarane (k ) 2 × 108 s-1). We have
investigated the oxidation of norcarane and spiro[2.5]octane by
four different enzymes: the two bacterial enzymes P450cam and
P450BM3 and the two mammalian enzymes CYP2B1 and
CYP2E1. Our results clearly show the involvement of a radical
intermediate in the oxidation of norcarane by the four P450
enzymes. The calculated radical lifetime in this case was found
to range from 16 ps for CYP2B1 to 52 ps for P450cam. The
latter corresponds to a rebound rate of 2 × 1010 s-1. Trace
amounts of the cationic rearrangement product were also
detected in several of the enzymatic reactions. No rearrangement
products derived from either a radical or a cation intermediate
were observed for the hydroxylations of spiro[2.5]octane.
More recently, Hata and co-workers examined the mechanism
of methane oxidation by the iron-oxo species using density
functional theory.22 The template model used was based on the
structure of P450cam and consisted of a porphine ligated to CH3-
S-. Their calculations suggested a four-step process: (a)
formation of FeIII-OH and a substrate radical; (b) rotation of
the OH group of FeIII-OH; (c) interaction of the substrate radical
with FeIII-OH; and (d) elimination of the hydroxylated product.
Consistent with the isotope effect reported earlier and the
proposed radical intermediate, they calculated the rate-limiting
step to be the hydrogen abstraction (step 1, 15 kcal/mol).
It is interesting to note that bicyclo[2.1.0]pentane, the only
diagnostic substrate that allowed the measurement of a lifetime
for the radical intermediate, is also the only clock that involves
oxidation at a methylene group in a small and otherwise
featureless molecule.12 This led us to consider whether steric
hindrance at the hydroxylation site, the strength of the scissile
C-H bond of the substrate, and unencumbered molecular
rotation may play decisive roles in the rebound mechanism. We
therefore examined the P450-catalyzed oxidation of hydrocarbon
radical clocks of a similar nature containing only secondary and
tertiary carbons and no other complicating functionality. Nor-
carane (1) has been a valuable substrate in mechanistic studies
of other oxidative enzymes and model systems because it can
unambiguously differentiate between cation and radical inter-
mediates (Figure 5).23a-c For example, norcarane has been used
to establish the formation of a radical intermediate with a
lifetime of ∼1 ns during catalysis by the non-heme diiron
monooxygenases AlkB.24a The oxidation of norcarane by
methane monooxygenase has been used to demonstrate that the
Fe(IV)-Fe(IV) intermediate Q performs a first oxidation leading
Results
To validate the use of norcarane and spiro[2.5]octane as
diagnostic substrates, we first spectroscopically measured their
(23) (a) Groves, J. T.; Kruper, W. J., Jr.; Haushalter, R. C. J. Am. Chem. Soc.
1980, 102, 6375-77. The solvolysis of the dinitrobenzoate ester of
2-norcaranol in aqueous acetone has been reported to afford mostly
2-norcaranols and 10% 3-cyclohepten-1-ol as the only rearranged product.
Under more forcing conditions (HClO4/HOAc) that allow for repeated
reionizations, 2-norcaranol afforded 3-cyclohepten-1-yl acetate and 3-
acetoxymethylcyclohexene in a ratio of 96:4. Cf.: (b) Friedrich, E. C.;
Jassawalla, J. D. C. Tetrahedron Lett. 1978, 953-956. (c) Friedrich, E.
C.; Jassawalla, J. D. C. J. Org. Chem. 1979, 44, 4224-4227. (d) Likewise,
we have shown that solvolysis of the dinitrobenzoate ester of 3-hydroxy-
methylcyclohexanol in aqueous acetone gave the same mixture of products
as for the 2-norcaranol ester with 90% 2-norcaranols (12 and 13), 9%
cycloheptenol (19), and <1% hydroxymethylcyclohexene (18).
(21) Watanabe, Y. J. Biol. Inorg. Chem. 2001, 6, 846-856.
(22) Hata, M.; Hirano, Y.; Hishino, T.; Tsuda, M. J. Am. Chem. Soc. 2001,
123, 6410-6416.
(24) (a) Austin, R. N.; Chang, H.-K.; Zylstra, G. J.; Groves, J. T. J. Am. Chem.
Soc. 2000, 122, 11747-11748. (b) Brazeau, B. J.; Austin, R. N.; Tarr, C.;
Groves, J. T.; Lipscomb, J. D. J. Am. Chem. Soc. 2001, 123, 11831-11837.
9
6022 J. AM. CHEM. SOC. VOL. 124, NO. 21, 2002

Radical Clocks and the Cytochrome P450 Mechanism
A R T I C L E S
Figure 6. Rearrangement pathways and products resulting from formation of a radical or a cation intermediate upon oxidation of spiro[2.5]octane (2) by
P450 enzymes.
Table 1. Ligand Dissociation Constants (Kd Values)
enzyme
[norcarane], µM
[spiro[2.5]octane], µM
P450cam
P450BM3
CYP2B1
CYP2E1
14
70
11
3
200
1
4
150
binding to P450cam, P450BM3, rat CYP2B1, and human CYP2E1.
Both substrates produce a Type I shift (i.e., a shift to a high-
spin species absorbing at 390 nm) in the Soret absorption of
the four enzymes. The calculated Kd values are listed in Table
1.
To discard the possibility of oxidation by hydrogen peroxide
or superoxide (uncoupling products of P450 enzymes), catalase
and superoxide dismutase were added to the reaction mixtures.
Control reactions did not show any oxidation products in the
absence of either the P450 enzyme or NADH/NADPH. Spiro-
[2.5]octane was oxidized by P450cam, CYP2B1, and CYP2E1,
whereas norcarane was a good substrate for all four enzymes.
The reaction of P450cam with spiro[2.5]octane (Figure 7A)
yielded 63% of 5-hydroxyspiro[2.5]octane (3), the major
product, 19% of 4-hydroxyspiro[2.5]-octane (4), and 18% of
6-hydroxyspiro[2.5]octane (5). The reaction of CYP2B1 (Figure
7B) afforded the same products but in different proportions:
74% of 5, 12% of 4, 11% of 3, and 3% of the 6-ketone (6).
The conversion was 40% based on the starting spiro[2.5]octane
detected in the reaction mixture. The products of the reaction
of CYP2E1 with spiro[2.5]octane did not match any of the
standards available and could not be identified. None of the
enzymes tested with this substrate yielded detectable amounts
of the products expected from a radical intermediate, 1-(2-
hydroxyethyl)cyclohexene (9), or a cationic intermediate, spiro-
[2.5]oct-4-ene (10) and 1-hydroxybicyclo[4.2.0]octane (11).
Figure 7. Total ion current chromatograms for the oxidation of spiro[2.5]-
octane (2): Trace A shows the data resulting from the reaction of P450cam
and trace B is the data obtained from the reaction of CY2B1. The rearranged
products 9 and 11 would have appeared at 8.03 and 6.81 min, respectively.
Ion chromatograms showed negligible intensity at m/e 79 (base peak of 9)
at 8.03 min in both samples and intensity at m/e 98 (base peak of 11)
corresponding to <0.5% 11 in trace A and <0.2% 11 in trace B. Compound
3 has a negligible peak at m/e 98.
2% of the primary alcohol 1-(2-hydroxyethyl)cyclohexene (9).25
These results were confirmed to obtain authentic samples of
these materials for GC-MS comparison. The free radical
The solvolysis of 1-(2-tosyloxyethyl)cyclohexene (9-Ts) has
been shown to produce 4-hydroxyspiro[2.5]octane (4) and
1-hydroxybicyclo[4.2.0]octane (11) as the major products and
(25) (a) Hanack, M.; Schneider, H.-J. Angew. Chem., Int. Ed. Engl. 1965, 4,
976; 1967, 6, 666. (b) Wiberg, K. B.; Hiatt, J. E.; Hseih, K. J. Am. Chem.
Soc. 1970, 92, 544-553.
9
J. AM. CHEM. SOC. VOL. 124, NO. 21, 2002 6023

A R T I C L E S
Auclair et al.
Table 2: Product Distributions for the Reactions of
Spiro[2.5]octane
products
P450
CYP2B1
cam
3a
4
5
63
19
18
ndb
11
12
74
3
8
a Product yields were determined by digital integration of the total ion
current signal of the GC-MS using the resident HP ChemStation software.
Product identity was confirmed by retention time and fragmentation pattern
as compared to authentic standards. The rearrangement products 3-(2-
hydroxyethyl)cyclohexene (9) and 1-hydroxybicyclo[4.2.0]octane (11) were
shown to be absent (<0.5%) by comparing the TIC trace to authentic
standards. b Not detected.
Table 3. Product Distributions for the Reactions of Norcarane
products
P450
P450BM3
CYP2B1
CYP2E1
cam
12a
13
14
16
17
18
19
100
35
16
100
49
13
100
59
11
0.5
5.0
0.5
0.5
16
100
188
18
20
nd
nd
3.2
1.6
0.6
52
9.8
1.3
0.6
44
7.0
2.0
nd
radical lifetime (ps)
35
Figure 8. GC-MS total ion current (TIC) chromatogram in the region of
the product alcohols from the hydroxylation of norcarane by P450BM3
.
a Product yields are reported relative to endo-2-norcaranol and were
determined by digital integration of the total ion current signal of the GC-
MS using the resident HP ChemStation software (see Figure 8). Product
identity was confirmed by retention time and fragmentation pattern. Products
reported in less than 1% relative yield were determined by comparing ion
chromatograms of the four most intense ions to those of dilute standard
mixtures. endo-3-Norcaranol (15) did not resolve from the exo-2-norcaranol
peak and is reported as the sum. The mass spectrum of this peak indicated
that it is mostly the latter compound. Products designated as “nd” were
below detectable limits with a good baseline.
sec-1, and the lifetime of the radical formed falls in the range
of 16 ps for CYP2B1 and 52 ps for P450cam
.
Discussion
Our data clearly demonstrate the formation of products
resulting from a radical intermediate during the hydroxylation
of norcarane by the four different P450 enzymes: P450cam
,
chlorinations of norcarane and spiro[2.5]octane were carried out
in accord with a previous study.26 A competitive chlorination
of these two substrates afforded 4-chloromethylcyclohexene and
1-(2-chloroethyl)cyclohexene in relative ratios of 4:1 with
respect to the unrearranged chlorides, indicating a 4-fold slower
rearrangement rate of the intermediate radical of spiro[2.5]-
octane.27,28
P450BM3, CYP2B1, and CYP2E1. For CYP2B1 the amount of
radical rearrangement product was significantly smaller than
the others. In agreement with the reported radical rearrangement
observed for bicyclo[2.1.0]pentane,12 the maximum lifetime
calculated for the 2-norcarane radical is 16-52 ps, which is in
the order of magnitude that corresponds to an intermediate rather
than a transition state.
The oxidation of norcarane by P450cam, P450BM3, and
CYP2E1 yielded endo-2-norcaranol (12) as the major product,
smaller amounts of exo-2-norcaranol (13), exo-3-norcaranol (14),
endo-3-norcaranol (15), 2-norcaranone (16), and 3-norcaranone
(17), and 1.3-2.0% of the radical rearrangement product
4-hydroxymethylcyclohexene (18) (Table 3). The total ion
current chromatogram in the product region for the oxidation
of norcarane by P450BM3 is presented in Figure 8. The oxidation
of norcarane by CYP2B1 afforded the same products as the
other enzymes but yielded only 0.5% 4-(hydroxymethyl)-
cyclohexene (18). Small amounts (ca. 0.5%) of the product
expected from a cationic intermediate, 3-cyclohepten-1-ol (19),
No substrate radical intermediate was detected during the
oxidation of spiro[2.5]octane by P450cam, P450BM3, CYP2B1,
or CYP2E1. This is consistent with the 4-fold slower radical
rearrangement rate for the spiro[2.5]oct-4-yl radical measured
using the chlorination of spiro[2.5]octane by tert-butyl hy-
pochlorite. Thus, each of these three closely related hydrocarbon
substrates, bicyclo[2.1.0]pentane, norcarane, and spiro[2.5]-
octane, gives an amount of radical rearrangement consistent with
a 16-50 ps radical lifetime in a typical oxygen rebound
scenario. While the rearranged cyclopentenol product derived
from bicyclo[2.1.0]pentane could have derived from either a
radical or cation pathway, the quantitative correlation of
substrate rearrangement with the radical rearrangement rate
described here supports a radical process for this substrate.
Our results are in good agreement with Shaik’s two-state
rebound mechanism in which the iron-oxo species may oxidize
the substrate via a low-spin concerted pathway or a high-spin
nonconcerted mechanism leading to a radical intermediate.
Assuming that other factors change less, this model also allows
the prediction that, as the radical becomes a better electron
donor, the high-spin rebound barrier will decrease, and the high-
spin pathway will become more concerted. This corroborates
the trend observed in the oxidation of not only arylmethylpro-
were detected in the hydroxylation mixtures from P450cam
,
P450BM3, and CYP2B1, while none of this product was detected
for CYP2E1, which showed the most radical rearrangement
product 18. Based on these data and the rearrangement rate of
2 × 108 sec-1 for norcarane, the rate of rebound is ca. 1010
(26) Boikess, R. S.; Mackay, M.; Blithe, D. Tetrahedron Lett. 1971, 401-404.
(27) Tanko, J. M.; Blackert, J. F. J. Chem. Soc., Perkin Trans. 2 1996, 1775-
1779.
(28) Given the direct comparison reported here for the rearrangement rates of
the 2-norcaranyl and spiro[2, 5]oct-4-yl radicals, an earlier estimate (ca. 4
× 108 s-1) determined at 120 K by EPR spectroscopy for the rearrangement
rate of the latter must be ca. 10 times too fast. Cf.: Roberts, C.; Walton,
J. C. J. Chem. Soc., Perkin. Trans. 2 1985, 841-846.
9
6024 J. AM. CHEM. SOC. VOL. 124, NO. 21, 2002

Radical Clocks and the Cytochrome P450 Mechanism
A R T I C L E S
carbon radicals by Fe(III)phen was timed at ca. 108 M-1 s-1
Thus, the electron-transfer oxidation of an incipient carbon
radical at the active site in the P450 rebound process could well
be oxidized by an intermediate oxoiron(IV) species to explain
the products that appear to arise from cationic rearrangement
processes.
pane derivatives but also the three radical clocks, norcarane,
bicyclo[2.1.0]pentane, and spiro[2.5]octane, in which oxidation
occurs at a secondary carbon. Thus, hydrogen abstraction from
norcarane or bicyclo[2.1.0]pentane leads to radicals that are
relatively poor electron donors because of the strain, and are
more likely to rearrange (longer lifetime observed). Conversely,
because the spiro[2.5]oct-4-yl radical is a better electron donor
with a 4-fold slower rearrangement rate, it is expected to yield
less rearrangement products.
.
Another anomaly in the body of data on P450 hydroxylations
is the observation that some “fast” radical clocks have shown
less rearrangement than slower clocks.30 For a two-state rebound
process, the transition state for hydrogen abstraction will position
the active oxygen a few tenths of an a˚ngstrom farther from the
hydroxylated carbon atom than the transition state for the
subsequent C-O bond formation. Thus, the extent of rearrange-
ment detected by a particular probe may simply reflect a facile
molecular trajectory from the hydrogen abstraction transition
state to the hydroxylation transition state. Indeed, in cases for
which the radical is produced within the attractive potential well
of the new C-O bond, the timing of radical rearrangements,
and thus the timing of the radical clocks, is likely to be affected.
Such an effect would be most pronounced for hydroxylation at
the C-H bond of methyl groups, with a very strong C-H bond
and a small steric size. Both effects would push the reaction
coordinate toward a tighter radical cage and shorten the radical
lifetime.31b Furthermore, the effective C-H bond strength for
cyclopropylcarbinyl hydrogens will depend on the stereoelec-
tronics of the interaction of the C-H bond with the neighboring
cyclopropyl group. We note that these interactions are rigidly
enforced in the bicycloalkanes under investigation here. Indeed,
it is the more sterically hindered 2-endo-hydrogen of biyclo-
[2.1.0]pentane that is the more reactive.12
A clear case of differential reactivity of nominally similar
secondary radical intermediates is observed in the hydroxylation
of a chirally deuterated ethylbenzene by a chiral iron porphyrin.31b
Thus, for this P450 model system, the hydrogen-deuterium
inventory of the products showed that while the benzylic radical
produced by abstraction of the pro-R deuterium rebounded to
produce the alcohol with 95% retention of configuration at
carbon, significant racemization was observed following ab-
straction of the pro-S hydrogen. Clearly, the same carbon radical
is produced in both cases, but the effective lifetime of the pro-R
radical, as revealed by the very small amount of rearrangement,
must be shorter than that of the pro-S radical due to the
asymmetry of the steric environment. This result is consistent
with the observed epimerization at carbon observed during the
hydroxylation of norbornane by cytochrome P450 in the first
experiments using a mechanistically diagnostic substrate probe
that was indicative of a radical rebound process.6 Thus, radical
reactions at C2 of norbornane are known to occur with
epimerization while cationic processes are prevented from doing
so by neighboring σ-bond participation.
The detection of trace amounts of the cation product (19)
observed for the norcarane hydroxylation for three of the P450s
studied here leaves open the possibility that a cation rearrange-
ment pathway is accessible to some small extent for this
substrate. However, the absence of the cyclobutanol product
(11) during the hydroxylation of spiro[2.5]octane argues against
cationic intermediates during the hydroxylation of this substrate.
How then does one explain the appearance of cationic rear-
rangements with some substrates, as has been reported by
Newcomb et al.,16 while for other substrates no such rearrange-
ments are observed? While a hydroperoxyiron(III) heme inter-
mediate has been suggested as an alternate oxidant in the P450
reaction cycle, based on changes in product ratios upon
modification of the enzyme active site, there is no evidence
that such a hydroperoxide would be sufficiently electrophilic
at the distal oxygen to react with a C-H bond. In fact, high-
level DFT calculations17,18 indicate that such a species would
be nucleophilic but not electrophilic. Furthermore, Hofmann and
Sligar et al. have shown by cryo-ENDOR spectroscopy that in
a single turnover hydroxylation with P450cam, the proton of the
incipient alcohol product derived from the substrate C-H bond
with the alcohol oxygen coordinated to the heme iron.4 This is
as expected for an iron-oxo precursor but is not possible for an
iron-hydroperoxide.
We suggest that the cationic rearranged products that
sometimes appear during P450 turnover derive from an electron
transfer oxidation of the incipient carbon radical that competes
in some cases with oxygen rebound (Figure 5). This is an
expression of the inner-sphere vs outer-sphere paradigm for
carbon radical oxidation that has been extensively studied by
Kochi.29 In this very informative work, the rate of electron
transfer from an alkyl radical to an iron(III)phenanthroline
oxidant to form a carbocation and iron(II) was timed by
competition with a homolytic bromine atom transfer from
bromotrichloromethane. Significantly, carbocation formation via
electron-transfer oxidation of even primary alkyl radicals was
found to compete with inner-sphere atom transfer processes.
Moreover, the extent of carbocation formation was a very
sensitive function of the oxidation potentials of both the iron(III)
oxidant and the alkyl radical while the inner-sphere processes
were most sensitive to steric effects. It is very reasonable to
assume that similar competitions would be at work during the
stepwise hydroxylation process mediated by cytochrome P450.
Here, oxygen rebound to form the product alcohol is the inner-
sphere process while competing electron-transfer processes
would occur to the extent that the rebound rate, taken to be
about 1010 s-1 based on earlier data of Ingold13,14 and data
presented here, competes with the electron-transfer rate. Sig-
nificantly, the rate of electron-transfer oxidation for secondary
Summary and Conclusions
We have reexamined the mechanism of aliphatic hydroxyl-
ation by P450 enzymes using two new diagnostic substrates:
spiro[2.5]octane and norcarane. The formation of a radical
(30) Newcomb, M.; Le Tadic-Biadatti, M.-H.; Chestney, D. L.; Roberts, E. S.;
Hollenberg, P. F. J. Am. Chem. Soc. 1995, 111, 1927-1928 and references
therein.
(31) (a) Groves, J. T.; Viski, P. J. Am. Chem. Soc. 1989, 111, 8537-8538. (b)
Groves, J. T.; Shalyaev, K.; Lee, J. In The Porphyrin Handbook; Kadish,
K. M., Smith, K. M., Guilard, R., Eds.;, Academic Press: San Diego, CA
and Burlington, MA, 1999; Vol. 4, Chapter 27, pp 17-40.
(29) (a) Kochi, J. K.; Mains, H. E. J. Org. Chem. 1965, 30, 1862-1872. (b)
Rollick, K. L.; Kochi, J. K. J. Am. Chem. Soc. 1982, 104, 1319-1330.
9
J. AM. CHEM. SOC. VOL. 124, NO. 21, 2002 6025

A R T I C L E S
Auclair et al.
20 min. Analysis of the reaction mixtures by GC-MS revealed the
presence of various monochlorides I-VII.
intermediate is consistent with the oxidation products of
norcarane obtained with P450cam, P450BM3, CYP2B1, and
CYP2E1 with intermediate radical lifetimes of 52, 44, 16, and
35 ps, respectively. Although we could not detect such an
intermediate during the transformation of spiro[2.5]octane, the
4-fold slower radical rearrangement rate for the spiro[2.5]oct-
4-yl radical as compared to the 2-norcaranyl radical would be
expected to lower the yield of radical products and be below
our detection limit. The results presented here for norcarane
and spiro[2.5]octane along with earlier results with bicyclo-
[2.1.0]pentane are in good agreement with the two-state oxygen
rebound mechanism with a consistent timing for the lifetime of
the radical intermediate.
The absolute rate constant for abstraction of chlorine from tert-butyl
hypochlorite had been determined to be 2.6 × 109 M-1 s-1 by using
homoallyl free radical clock, cyclopropylcarbinyl (free radical ring-
opening rate at 20 °C: ko ) 7.9 × 107 s-1).27 Thus, the ratio of V:VII
allows a confirmation of the rearrangement rate for the 2-norcaranyl
radical to be 2 × 108 s-1. The competition reaction between norcarane
and spiro[2.5]octane with tert-butyl hypochlorite shows that the rate
for ring opening of the 2-norcaranyl radical is about 4 times as fast as
the rate for spiro[2.5]oct-4-yl radical (5 × 107 s-1).
The mechanisms of aliphatic hydroxylation by metal-oxo
species remain among the most interesting and enigmatic in
catalytic science. The challenge for the field has been to fold
the disparate and sometimes contradictory data into a compre-
hensible picture. We have shown here that the hydroxylation
event mediated by cytochrome P450 at the methylene group of
a closely related set of substrates gives a consistent set of data
supporting a radical rebound process. The traces of cation
products observed have been attributed to a competing electron-
transfer process that has strong precedent in chemical systems.
The lack of rearrangement observed for some fast rearranging
probes can be understood in terms of an interplay of sensitive
steric and stereoelectronic effects, especially for the hydroxyl-
ation of the C-H bond of a methyl group.
Preparation of the Enzymes and Binding Assays. P450cam from
Pseudomonas putida, Pd, and PdR were expressed heterologously in
Escherichia coli and were purified as reported previously.33 CYP2B134
and rat NADPH-cytochrome P450 reductase35 were expressed and
purified as described elsewhere. Human CYP2E1 was kindly provided
by F. P. Guengerich, Vanderbilt University. Cytochrome b5 was
generously provided by Almira Correia, University of California, San
Francisco. Kd values for the association of norcarane or spiro[2.5]octane
with the different P450 enzymes were determined by UV spectroscopy.
The camphor present in the P450cam sample was removed using a PD10
column eluted with 10 mM potassium phosphate at pH 7.5. The binding
assays were performed on the same day (P450cam is unstable to freezing/
thawing in the absence of camphor). The decrease in absorbance of
the low-spin state at 416 nm and the increase in absorbance of the
high-spin species at 390 nm were monitored upon addition of norcarane
or spiro[2.5]octane. Kd values for CYP2E1 were measured by displace-
ment of the pyrazole ligand and the corrected Kd was calculated using
the following equation: Kd ) Ks apparent/[1 + ligand concentration/
Ksligand]. When pyrazole was used the decrease in absorbance occurred
at 423 nm and the increase was again at 390 nm.
Cytochrome P450BM3 Expression. P450BM3 DNA in the pUC19
vector was kindly provided by Professor Armand J. Fulco (University
of California, Los Angeles). Quickchange was used to add a His6-tag
on the N-terminal end of the protein with the 5′-sense oligonucleotide
5′-GA AAG AGG GAT AAC ATG CAC CAC CAC CAC CAC CAC
ACA ATT AAA GAA ATG CCT CAG CC. Mutations were confirmed
by DNA sequencing. The DNA was then cleaved out of pUC119 and
ligated into the pCWori vector. Escherichia coli strain DH5R [F′ ara
D(lac-proAB)rpsLf80d lacZDM15 hsdR17] were transformed with the
plasmid and spread onto Luria-Bernati medium containing agar. One-
liter of 2YT liquid medium supplemented with 100 mg/L of ampicillin
Experimental Procedures
Materials. Unless otherwise mentioned, all the reagents were from
Sigma-Aldrich. The dichloromethane and toluene were Omnisolve grade
and purchased from EM Science (Darmstadt, Germany). The Quick-
Change site-directed mutagenesis kit and the nitrilotriacetic acid resin
(NTA-Agarose) were obtained from Stratagene (La Jolla, CA). Yeast
extract, tryptone, 2YT, sodium chloride, EDTA, and organic solvents
were obtained from Fisher Scientific (Fair Lawn, NJ). Isopropyl-â-D-
thiogalactopyranoside was from FisherBiotech (Fair Lawn, NJ), and
ampicillin from Promega (Madison, WI). High-purity argon (99.998%),
CO (99.9%), and the oxygen adsorbent, Oxisorb-W, used to further
purify the argon were purchased from Puritan-Bennett Medical Gases
(Overland Park, KS). All the chemicals were used without further
purification. The UV-visible spectra were recorded on a Hewlett-
Packard 8452A diode array spectrophotometer or on a Cary 1E Varian
UV-visible spectrophotometer. GC-MS analyses were performed on
a HP GC-MS (5890/ 5989B) with a HP-5MS cross-linked 5% PH ME
Siloxane capillary column.
Synthesis. Norcarane and spiro[2.5]octane were prepared by using
the Simmons-Smith reaction as previously described.32 Authentic
samples of the product norcaranols, spiro[2.5]octanols, the correspond-
ing ketones, and the rearrangement products 4-cyclohepten-1-ol and
4-hydroxymethylcyclohexene were prepared according to published
procedures.24 The purity of all products was assessed by 1H NMR and
GC-MS.
Chlorination of Norcarane and Spiro[2.5]bicyclooctane with tert-
Butyl Hypochlorite. Norcarane (196 mg, 2 mmol), spiro[2.5]octane
(220 mg, 2 mmol), and tert-butyl hypochlorite (215 mg, 2 mmol) in 5
mL of benzene were charged into a 25 mL flask under Ar at 20 °C
(water bath). The flask was irradiated with a 75 W tungsten lamp for
(32) (a) LeGoff, E. J. Org. Chem. 1964, 29, 2048-2050. (b) Simmons, H.;
Cairns, T.; Vladuchick, S.; Hoiness, C. Org. React. 1973, 20, 1. (c) Dauben,
W. G.; Berezin, G. H. J. Am. Chem. Soc. 1963, 85, 469. (d) Snider, B.;
Rodini, D. Tetrahedron Lett. 1980, 21, 1815-1818. (e) Akermark, B.;
Hanson, S.; Rein, T.; Vagberg, J. J. Organomet. Chem. 1989, 369, 433-
444.
(33) De Voss, J. J.; Sibbesen, O.; Zhang, Z.; Ortiz de Montellano, P. R. J. Am.
Chem. Soc. 1997, 119, 5489-5498.
(34) Tuck, S. F.; Ortiz de Montellano, P. R. Biochemistry 1992, 31, 6911-
6916.
(35) Dierks, E. A.; Davis, S. C.; Ortiz de Montellano, P. R. Biochemistry 1998,
37, 1839-1847.
9
6026 J. AM. CHEM. SOC. VOL. 124, NO. 21, 2002

Radical Clocks and the Cytochrome P450 Mechanism
A R T I C L E S
was used for the expression. P450BM3 expression was induced with 0.4
mM isopropyl-â-D-thiogalactopyranoside at OD600 ∼ 0.8 (after about
4 h at 37 °C and 250 rpm), then further grown at 30 °C for 11 h after
another 75 mg/L of ampicillin was added. The harvested cells (75 g
from 4 L cultures) were lysed for 1 h at 4 °C in 200 mL of 50 mM
potassium phosphate buffer (pH 7.4) containing 2 mM DTT, 0.1 mM
EDTA, 0.1 mM PMSF, 1 µg/mL leupeptin, 1 µg/mL pepstatin, and
0.3 g of lysosyme. The cells were further lysed by sonication. Following
centrifugation, the supernatant was directly loaded onto a nickel column
(15 mL) equilibrated with buffer A (20 mM Tris buffer at pH 8.0, 0.5
M NaCl, 10% glycerol, and 1 mM PMSF). The column was washed
with buffer A (100 mL) and the protein was eluted with buffer A
containing 100 mM imidazole. The colored fractions with Rz g 0.4
were pooled, diluted in one volume of 100 mM potassium phosphate
at pH 7.4 containing 0.5 mM of DTT, and loaded onto a 2′,5′-ADP
Sepharose column (20 mL) equilibrated with the same buffer. The
column was washed with 5 column volumes each of the equilibration
buffer and 500 mM potassium phosphate containing 0.5 mM DTT.
The protein was eluted with the latter buffer containing 10 mM 3′-
AMP. The red fractions were concentrated in an Amicon cell using a
YM30 membrane and dialyzed against 10 mM potassium phosphate,
pH 7.4.
Reactions of the Different P450 Enzymes with Norcarane or
Spiro[2.5]octane. Unless otherwise indicated, the buffer was 100 mM
potassium phosphate at pH 7.4. P450cam reactions [P450cam (2 µM), Pd
(4 µM), PdR (2 µM), substrate (500 µM of 1 or 2), superoxide dismutase
(2 µM), and catalase (100 µg/mL)] were dissolved in buffer. The
reaction was started by the addition of NADH (200 µM). P450BM3
reactions: To a solution of P450BM3 (2 µM) in buffer were added
superoxide dismutase (2 µM), substrate (500 µM), and NADPH (100
µM). CYP2B1 reactions: The following components were mixed in
strict order, DLPC (L-R-phosphatidylcholine dilauroyl, 200 µg/mL),
CYP2B1 (0.5 µM), P450 reductase (0.5 µM), cytochrome b5 (1 µM),
superoxide dismutase (0.5 µM), and catalase (100 µg/mL). The mixture
was incubated at room temperature for 10 min followed by the addition,
in order, of the buffer and the substrate (500 µM). After 5 min of
incubation at 37 °C, the reaction was started by the addition of NADPH
(1 mM). The enzymatic reactions with CYP2E1 were prepared by
mixing in order and incubating at room temperature for 10 min, DLPC
(30 µg/mL), P450 reductase (1.5 µM), CYP2E1 (0.5 µM), superoxide
dismutase (0.5 µM), and catalase (100 µg/mL). After addition, in order,
of the buffer and the substrate (500 µM) followed by incubation at 37
°C for 5 min, the reaction was initiated by mixing NADPH (1 mM).
Finally, control reactions were carried out without enzyme and,
separately, without NADPH/NADH. All the reaction mixtures (1 mL)
were incubated at 37 °C for 1 h before being extracted with
dichloromethane (3 × 300 µL). The combined organic layers were
carefully washed with brine (2 × 400 µL), before injection in the GC/
MS system. The rebound rates were calculated from the following
equation: krebound ) krearr. × [nonradical products]/[radical product]. The
radical lifetime is the reciprocal of krebound
.
Acknowledgment. We are grateful for support of this work
by the National Institutes of Health (GM25515, P.R.O.M.;
GM36928, J.T.G.) and by the National Science Foundation (NSF
9810248, J.T.G.) for support of the Center for Environmental
Bioinorganic Chemistry.
Supporting Information Available: Ion chromatograms show-
ing the product region for norcarane oxidation by P450BM3 and
CYP2E1 (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org.
JA025608H
9
J. AM. CHEM. SOC. VOL. 124, NO. 21, 2002 6027