Hypersensitive mechanistic probe studies of cytochrome P450-catalyzed hydroxylation reactions. Implications for the cationic pathway

Article abstract of DOI:10.1021/ja981157i

Details of the mechanism of cytochrome P450-catalyzed hydroxylation reactions were investigated by oxidation of trans-2-phenyl-1- alkylcyclopropanes (alkyl = methyl (1), ethyl (2), 1-propyl (3), 1- methylethyl (4)) and trans-2-(4-(trifluoromethyl)phenyl)-1-alkylcyclopropanes (alkyl = methyl (5), ethyl (6)). The syntheses of 3 and 6 and their possible oxidation products are reported. Oxidation of the probes with the cytochrome P450 isozyme CYP2B1 gave unrearranged cyclopropylcarbinols as major products and small amounts of ring-opened alcohol products in all cases except for 4. Phenolic products also were produced from substrates 1-4. The maximum lifetimes of putative radical intermediates were less than 1 ps, and the results with substrate 4 require that no intermediate was formed. The results were analyzed in the context of recent mechanistic proposals for cytochrome P450-catalyzed hydroxylations. Oxidation of a 'radical' component in the transition state of an insertion reaction to produce a cation is inconsistent with the results. The results also provide little support for a new alternative mechanism for hydroxylation, the agostic complex model (Collman, J.P.; Chien, A. S.; Eberspacher, T. A.; Brauman, J. I. J. Am. Chem. Soc. 1998, 120, 425). Formation of 'cationic' rearrangement products via solvolysis of first-formed protonated alcohol products produced by insertion of the 'OH⁺' moiety from iron-complexed hydrogen peroxide also is not supported by the results. The most consistent mechanistic description is the recently reported multistate reactivity paradigm (Shaik, S.; Filatov, M.; Schroder, D.; Schwarz, H. Chem. Eur. J. 1998, 4, 193).

Full text of DOI:10.1021/ja981157i

J. Am. Chem. Soc. 1998, 120, 7719-7729
7719
Hypersensitive Mechanistic Probe Studies of Cytochrome
P450-Catalyzed Hydroxylation Reactions. Implications for the
Cationic Pathway
Patrick H. Toy,† Martin Newcomb,*,† and Paul F. Hollenberg*,‡
Contribution from the Department of Chemistry, Wayne State UniVersity, Detroit, Michigan 48202, and
Department of Pharmacology, The UniVersity of Michigan, Ann Arbor, Michigan 48109
ReceiVed April 6, 1998
Abstract: Details of the mechanism of cytochrome P450-catalyzed hydroxylation reactions were investigated
by oxidation of trans-2-phenyl-1-alkylcyclopropanes (alkyl ) methyl (1), ethyl (2), 1-propyl (3), 1-methylethyl
(4)) and trans-2-(4-(trifluoromethyl)phenyl)-1-alkylcyclopropanes (alkyl ) methyl (5), ethyl (6)). The syntheses
of 3 and 6 and their possible oxidation products are reported. Oxidation of the probes with the cytochrome
P450 isozyme CYP2B1 gave unrearranged cyclopropylcarbinols as major products and small amounts of ring-
opened alcohol products in all cases except for 4. Phenolic products also were produced from substrates 1-4.
The maximum lifetimes of putative radical intermediates were less than 1 ps, and the results with substrate 4
require that no intermediate was formed. The results were analyzed in the context of recent mechanistic
proposals for cytochrome P450-catalyzed hydroxylations. Oxidation of a “radical” component in the transition
state of an insertion reaction to produce a cation is inconsistent with the results. The results also provide little
support for a new alternative mechanism for hydroxylation, the agostic complex model (Collman, J. P.; Chien,
A. S.; Eberspacher, T. A.; Brauman, J. I. J. Am. Chem. Soc. 1998, 120, 425). Formation of “cationic”
rearrangement products via solvolysis of first-formed protonated alcohol products produced by insertion of
the “OH+” moiety from iron-complexed hydrogen peroxide also is not supported by the results. The most
consistent mechanistic description is the recently reported multistate reactivity paradigm (Shaik, S.; Filatov,
M.; Schroder, D.; Schwarz, H. Chem. Eur. J. 1998, 4, 193).
Introduction
hydrogen abstraction by the iron-oxo to give an alkyl radical
that subsequently displaced hydroxy from iron in a homolytic
substitution.4-7 Recent studies with hypersensitive radical
probe/clock substrates, however, indicate that the “lifetime” of
a radical in P450 hydroxylation is too short (<100 fs) for a
true intermediate, implicate an insertion process, and further
suggest that cationic species can be produced in P450 hydrox-
ylations.8 Details of the studies that led to the various con-
clusions are in the Discussion. As evidence militating against
a true radical intermediate in P450 hydroxylations has appeared,
so also have mechanistic alternatives to the abstraction-rebound
pathway. For example, Shestakov and Shilov discussed the
possibility that a five-coordinate carbon species might explain
isomerization results in P450 hydroxylations.9 Collman, Brau-
man, and co-workers have recently proposed an “agostic
alternative” to the oxygen rebound mechanism and demonstrated
that hydroxylation reactions by iron-oxo species can be
inhibited by both hydrogen and methane.10 Shaik, Schwarz,
and co-workers have presented an insightful analysis of the P450
The cytochrome P450 (P450) enzymes are broadly distributed
in plants, animals, and bacteria.1-3 These enzymes are capable
of catalyzing a variety of reactions, the most energetically
difficult of which is the hydroxylation of unactivated C-H
positions at ambient temperatures. The active sites of P450
enzymes are characterized by a protoporphyrin IX-iron com-
plex (heme) with thiolate from a cysteine of the enzyme serving
as the fifth ligand to iron. Considering the invariant structure
of the active sites of the P450 enzymes, it is reasonable to
believe that a common mechanism exists for hydrocarbon
hydroxylation reactions catalyzed by these enzymes and that
substrate specificity and regio- and stereoselectivity results from
the tertiary protein structures. Considerable information con-
cerning the timing of events in P450 hydroxylations, such as
substrate and oxygen binding, is known, but the mechanism of
the hydroxylation step is not resolved.
For the past two decades, the consensus mechanism for P450-
catalyzed hydroxylation involved production of an iron-oxo
species that reacted with hydrocarbons in a two-step reaction,
(4) Sono, M.; Roach, M. P.; Coulter, E. D.; Dawson, J. H. Chem. ReV.
1996, 96, 2841-2887.
(5) (a) Groves, J. T.; Han, Y.-z. In ref 2; Chapter 1. (b) Ortiz de
Montellano, P. R. In ref 2; Chapter 8.
(6) Woggon, W. D.; Fretz, H. In AdVances in Detailed Reaction
Mechanisms; Coxon, J. M., Ed.; JAI: Greenwich, CT, 1992; Vol. 2, pp
111-147.
(7) Guengerich, F. P.; Macdonald, T. L. FASEB J. 1990, 4, 2453-2459.
(8) Newcomb, M.; Le Tadic-Biadatti, M.-H.; Chestney, D. L.; Roberts,
E. S.; Hollenberg, P. F. J. Am. Chem. Soc. 1995, 117, 12085-12091.
(9) Shestakov, A. F.; Shilov, A. E. J. Mol. Catal., A 1996, 105, 1-7.
(10) Collman, J. P.; Chien, A. S.; Eberspacher, T. A.; Brauman, J. I. J.
Am. Chem. Soc. 1998, 120, 425-426.
* To whom correspondence should be addressed.
† Wayne State University.
‡ University of Michigan.
(1) Nelson, D. R.; Koymans, L.; Kamataki, T.; Stegeman, J. J.;
Feyereisen, R.; Waxman, D. J.; Waterman, M. R.; Gotoh, O.; Coon, M. J.;
Estabrook, R. S.; Gunsalus, I. C.; Nebert, D. W. Pharmacogenetics 1996,
6, 1-42.
(2) Cytochrome P450 Structure, Mechanism, and Biochemistry, 2nd ed.;
Ortiz de Montellano, P. R., Ed.; Plenum: New York, 1995.
(3) Guengerich, F. P. Am. Sci. 1993, 81, 440-447.
S0002-7863(98)01157-3 CCC: $15.00 © 1998 American Chemical Society
Published on Web 07/24/1998

7720 J. Am. Chem. Soc., Vol. 120, No. 31, 1998
Toy et al.
Scheme 1a
Figure 1. Probes employed in P450-catalyzed hydroxylations.
a Key: (a) LAH, THF, Reflux; (b) (i) Et2Zn, CH2I2, CH2Cl2, -30
°C; (ii) m-CPBA, CH2Cl2, rt; (c) Oxalyl Chloride, DMF, Benzene, rt;
(d) CH2N2, CH2Cl2, Ether, rt; (e), Ag2O, Na2S2O3, Dioxane, Water, 75
°C; (f) LAH, THF, 0 °C; (g) (i) Et3N, MsCl, THF, -30 °C; (ii) LiEt3BH,
-78 °C to rt.
hydroxylation reaction in terms of a multistate reactivity
paradigm that permits stepwise processes in competition with
concerted processes and relates P450 hydroxylations to reactions
of metal oxenoid cations.11
The consensus abstraction-rebound mechanism and most of
the mechanistic alternatives begin from the premise that an
iron-oxo species is the active oxidant in P450, but it is also
possible that a hydroperoxo-iron or an iron-hydrogen peroxide
complex is the active oxidant in P450 (see the Discussion). An
understanding of the details of the pathway for formation of a
cationic species in P450 hydroxylation appears to be critical
for descriptions of both the mechanism of hydroxylation and
the active oxidant. In an attempt to study this aspect of P450
hydroxylations, we have employed two sets of “radical probes”
with varying alkyl substitution at the probe center in P450
hydroxylations. The study was designed specifically to test for
the possibility that cationic rearrangement products are formed
by oxidation of “radicals” in transition states for insertion
reactions; large changes in product distributions would be
expected to follow changes in the radical oxidation potentials.
Our results are not consistent with radical oxidation events as
the sources of cationic species and are most in line with
predictions of the multistate reaction model described by Shaik,
Schwarz, and co-workers.11
carbinyl position gives an unrearranged alcohol product which
can be oxidized to a carbonyl compound. Functionalization of
the cyclopropylcarbinyl position with a radical or cationic ring
opening ultimately gives a ring-opened alcohol product. The
phenyl-containing probes 1-4 also can be oxidized to phenolic
products, but phenols have not been observed with (trifluoro-
methyl)phenyl-containing probes 5 and 6. For simplicity, we
have labeled the oxidation products consistently with the number
corresponding to the probe from which they derive and a letter
indicating the type of oxidation product. Thus, the products
from each probe are unrearranged alcohol U, aldehyde or ketone
product K, rearranged alcohol R, and phenol P as shown below
for probe 1.
Results
Probes and Oxidation Products. Oxidations of two series
of probes (Figure 1) are compared in this work. One series
(1-4) contained the trans-2-phenylcyclopropyl moiety. The
parent for this series, methyl-containing probe 1, has been used
previously in studies of hydroxylations by P450,12,13 chloro-
peroxidase,14,15 soluble methane monooxygenase (sMMO),16 and
an alkane monooxygenase.17 The ethyl- and isopropyl-contain-
ing probes 2 and 4 were used with equivocal results in a study
of chloroperoxidase.15 Probes 5 and 6 contain the trans-2-(p-
(trifluoromethyl)phenyl)cyclopropyl moiety. Both enantiomers
of probe 5 were used in a study of P450 hydroxylations.18
A variety of oxidation products can be formed in the P450-
catalyzed hydroxylations. Hydroxylation at the cyclopropyl-
The syntheses of probes 1, 2, 4, and 5 and several of their
possible oxidation products were reported previously.15,16,18
Probes 3 and 6 are new, and the synthetic routes to these
compounds are shown in Scheme 1. Probe 3, previously
reported as a mixture of diastereomers from a carbenoid addition
to 1-pentene,19 was prepared as a pure diastereomer in two steps
from commercially available 1-phenyl-1-pentyne. Lithium
aluminum hydride (LAH) reduction by the method of Schwan20
afforded selectively the trans-alkene 7. Alkene 7 was cyclo-
propanated by diethylzinc and diiodomethane to afford 3.
Compound 3 was conveniently purified by treating the crude
product mixture from the cyclopropanation reaction with
3-chloroperoxybenzoic acid, to epoxidize unreacted alkene, and
chromatography. The synthesis of probe 6 started with car-
boxylic acid 8, which was an intermediate in the synthesis of
probe 5.18 Homologation of 8 by the Arndt-Eistert sequence21
followed by esterification with diazomethane afforded the
corresponding cyclopropylacetic acid derivative 9. The acid
resulting from the homologation sequence was esterified for
purification purposes. Reduction of ester 9 with LAH to give
(11) Shaik, S.; Filatov, M.; Schroder, D.; Schwarz, H. Chem. Eur. J.
1998, 4, 193-199.
(12) Atkinson, J. K.; Hollenberg, P. F.; Ingold, K. U.; Johnson, C. C.;
Le Tadic, M.-H.; Newcomb, M.; Putt, D. A. Biochemistry 1994, 33, 10630-
10637.
(13) Atkinson, J. K.; Ingold, K. U. Biochemistry 1993, 32, 9209-9214.
(14) Zaks, A.; Dodds, D. R. J. Am. Chem. Soc. 1995, 117, 10419-10424.
(15) Toy, P. H.; Newcomb, M.; Hager, L. P. Chem. Res. Toxicol., in
press.
(16) Liu, K. E.; Johnson, C. C.; Newcomb, M.; Lippard, S. J. J. Am.
Chem. Soc. 1993, 115, 939-947.
(19) Schlosser, M.; Heinz, G. Chem. Ber. 1970, 103, 3543-3552.
(20) Schwan, A. L.; Roche, M. R.; Gallagher, J. F.; Ferguson, G. Can.
J. Chem. 1993, 72, 312-324.
(21) Newcomb, M.; Johnson, C. C.; Manek, M. B.; Varick, T. R. J. Am.
Chem. Soc. 1992, 114, 10915-10921.
(17) Fu, H.; Newcomb, M.; Wong, C.-H. J. Am. Chem. Soc. 1991, 113,
5878-5880.
(18) Toy, P. H.; Dhanabalasingam, B.; Newcomb, M.; Hanna, I. H.;
Hollenberg, P. F. J. Org. Chem. 1997, 62, 9114-9122.

Cytochrome P450-Catalyzed Hydroxylation Reactions
J. Am. Chem. Soc., Vol. 120, No. 31, 1998 7721
Scheme 2a
The phenol 1P was previously prepared independently in
order to demonstrate its formation in enzyme-catalyzed hy-
droxylations of 1.16 This phenol displayed a strong molecular
ion in its electron-impact mass spectrum and a significantly
longer retention time than other oxidation products of 1 in GC
analyses on Carbowax. The production of phenols competes
with oxidations at the cyclopropylcarbinyl position of probes
1-4 but has little significance in regard to the mechanistic
analysis of this work. Thus, phenolic products from P450-
catalyzed oxidations of probes 2-4 were not synthesized but
were identified in the product mixtures from the enzyme
oxidations by their GC retention times and mass spectra.
Enzyme-Catalyzed Oxidations. In this work, probes 2-4
and 6 were oxidized by the phenobarbital-induced rat liver
isozyme CYP2B1.26 This enzyme is known to be a relatively
nonselective oxidation catalyst and has been employed widely
in mechanistic studies, including previous studies of probes 1
and 5.12,18 The hepatic P450 enzymes require a reductase
enzyme which delivers reducing equivalents from NADPH to
the substrate-enzyme-oxygen complex and to the subsequently
formed substrate-enzyme-superoxide complex. The reductase
used in this study was expressed in E. coli.27
Oxidations were conducted with reconstituted systems con-
sisting of the P450 and reductase enzymes in a 1:2 molar ratio,
NADPH, and dilauroylphosphatidylcholine. These oxidations
were performed under the same conditions as previously used
for the “low concentration” oxidations of probe 5.18 Specifi-
cally, a solution of approximately 1.2 µmol of substrate in 10
µL of methanol was added to the enzyme mixture, and the
reactions were initiated by addition of NADPH. Reactions were
terminated after 30 min, and the products were extracted into
methylene chloride and analyzed by GC and GC-MS.
a Key: (a) Oxalyl Chloride, DMF, Benzene, rt; (b) MeONHMe‚HCl,
Pyridine, CH2Cl2, rt; (c) PrMgBr, Ether, 0 °C; (d) LAH, THF, 0 °C;
(e) (i) Et3N, MsCl, THF, -30 °C; (ii) Water, Reflux; (f) MeMgBr,
Ether, 0 °C; (g) TBSCl, Imidazole, CH2Cl2, rt; (h) OsO4, NaIO4, Ether,
Water, rt; (i) Ph3PEtBr, BuLi, Benzene, rt; (j) Bu4NF, THF, rt.
the corresponding alcohol was followed by a two-step meth-
anesulfonation/lithium triethylborohydride reduction sequence22
that afforded probe 6. We note that all probes used in this study
were characterized by 1H and 13C NMR spectroscopy and
HRMS and were demonstrated to be >98% pure by NMR
spectroscopy and GC analysis.
Either four or eight P450-catalyzed oxidations of each probe
were conducted, and complete results are given in the Supporting
Information. Table 1 contains a summary of those hydroxyla-
tion reactions as well as the results previously reported for 1
and 5.12,18,28 Oxidations of probes 1-4 gave unrearranged
alcohols U and phenols P. Rearranged alcohol products R were
observed for 1-3, but product 4R from hydroxylation of 4 was
not present at the limit of our detection capabilities (ca. 1%
relative to 4U). Some oxidation of 2U and 3U occurred to give
2K and 3K, respectively. Because oxidation to the carboxylic
acid occurs when 1 reacts with chloroperoxidase,14,15 we have
previously searched for this product in the form of its methyl
ester (after diazomethane treatment of a product mixture) as
well as for product 1K but did not detect them.28 In a similar
manner, the carboxylic acid from probe 5 was shown not to be
formed in P450 hydroxylations.18 The (trifluoromethyl)phenyl
probes 5 and 6 gave results similar to those for 1 and 2 with
the exception that phenolic products were not formed. Although
hydroxylation reactions might occur at positions other than the
cyclopropylcarbinyl center and the phenyl ring, such oxidation
reactions were apparently quite minor as deduced from the
observation that no unidentified products with GC retention
times similar to those of products U, K, and R were present in
more than 2% relative to other products. Favored oxidation of
the cyclopropylcarbinyl position is consistent with the fact that
The products from oxidation of probe 3 were synthesized
from commercially available trans-2-phenyl-1-cyclopropane-
carboxylic acid (Scheme 2). This acid was converted to an
amide by the two-step procedure of Weinreb.23 Addition of
ethylmagnesium bromide to the Weinreb amide afforded ketone
3K that was reduced to a 2:1 mixture of diastereomeric alcohols
3U24 with LAH. Methanesulfonation of 3U followed by
hydrolysis in refluxing water afforded the diastereomeric acyclic
alcohols 3R25 in a 13:1 ratio.
The cyclic products from the oxidation of probe 6 were
prepared from Weinreb amide 10 (Scheme 2), also an interme-
diate in the synthesis of probe 5.18 Addition of methylmagne-
sium bromide to 10 afforded ketone 6K. Reduction of 6K with
LAH afforded an inseparable 1:1 mixture of diastereomeric
alcohols 6U. Attempts to convert alcohols 6U to acyclic
alcohols 6R by formation and solvolysis of a mesylate failed;
apparently the requisite sulfonate was not formed. Therefore,
a more circuitous synthesis of 6R was developed using 5R,18
itself a product from the P450 oxidation of probe 5, as the
starting material. The first step was to protect the hydroxyl
group of 5R as a silyl ether. Oxidation of the double bond to
the aldehyde was achieved by treatment of the silyl ether with
osmium tetroxide and sodium periodate. The resulting aldehyde
was olefinated with the ylide derived from ethyltriphenylphos-
phonium bromide, and removal of the silyl group with tetrabu-
tylammonium fluoride afforded acyclic alcohols 6R as a 4:1
mixture of diastereomers.
(26) Saito, T.; Strobel, H. W. J. Biol. Chem. 1981, 256, 984-988.
(27) Hanna, I. H.; Teiber, J. F.; Kokones, K. L.; Hollenberg, P. F. Arch.
Biochem. Biophys. 1998, 350, 324-332.
(28) The results for probe 1 are from reactions with microsomes from
livers of phenobarbital-treated rats. The major isozyme induced by this
treatment is CYP2B1. In unpublished work, virtually identical ratios of
unrearranged to rearranged product from probe 1 oxidations by the isozyme
CYP2B1 with expressed reductase were observed. Toy, P. H.; Choi, S.-Y.;
Newcomb, M. Unpublished results.
(22) Gajewski, J. J.; Squicciarini, M. P. J. Am. Chem. Soc. 1989, 111,
6717-6728.
(23) Nahm, S.; Weinreb, S. M. Tetrahedron Lett. 1981, 32, 3815-3818.
(24) Charette, A. B.; Lebel, H. J. Org. Chem. 1995, 60, 2966-2967.
(25) Barluenga, J.; Alvarez, F.; Concellon, J. M.; Bernad, P.; Yus, M.
Synthesis 1987, 3, 318-320.

7722 J. Am. Chem. Soc., Vol. 120, No. 31, 1998
Toy et al.
Table 1. Results from CYP2B1 Hydroxylation of Probes
yield of products (nmol)
substrate
acyclic alcohol
cyclic alcohol
ketone
phenol
C/Aa
4.2
25.0
51.7
% convb
% recc
turnoverd
1e
2
3
60.0
9.5
2.8
<0.5
34.1
3.4
224.0
228.4
137.8
126.7
128.9
78.6
149.0
15.8
19.3
15.2
NA
9.3
7.1
17.0
10.7
9.8
13.0
8.5
62.0
69.3
79.1
454
290
236
280
218
4
NAf
6.4h
24.3
>100.0
5g
6
4.0
NA
30.2
72.3
a Observed ratio of cyclic to acyclic products. b Percent conversion of substrate. c Percent recovery of substrate. d Number of enzyme turnovers.
e Data from ref 12. f NA ) not applicable. g Data from ref 18. h Aldehyde formed.
Table 2. Recovery and Stability Control Reactions for Oxidation
Products
as 100% for these reactions. The yields of the other two
reactions relative to the calibration reaction are given in Table
2.
relative percent yields
amount
(nmol)
% yield
The reactions conducted without NADPH were recovery
experiments. The high yields in these experiments demonstrate
that in all cases the enzyme mixture does not retain the products.
The reactions conducted with fully competent enzyme systems
are stability experiments. Cyclic oxidation products from probes
2, 4, and 6 were returned in high yields even though a portion
of the cyclic alcohols 2U and 6U was oxidized to the
corresponding ketone. The total recovery from cyclic product
3U was reduced (75% recovery of 3U + 3K), and ketone 3K
was clearly unstable, suggesting that the origin of reduced yield
for 3U involved oxidation to 3K and further conversion to
unknown products. Given the excellent stabilities of the other
secondary probe oxidation products (alcohols 2U and 6U and
ketones 2K and 6K), it is tempting to ascribe the instability of
the oxidation products from 3 to the presence of the propyl
group, but it is not apparent to us why this should be the case.
As one might expect on the basis of the presence of double
bonds, the rearranged alcohol products R were generally
unstable in the presence of the fully competent system. Ring-
opened alcohols 3R and 6R were completely destroyed, and
somewhat more than half of product 4R was lost, whereas only
about 20% of acyclic alcohol product 2R was lost in the enzyme
oxidation.
substrate
recoverya
stabilityb
admixturec
2U
2K
2R
2R
3U
3K
3R
3R
4U
4R
4R
6U
6K
6R
6R
740
649
518
11
85
12
97
95
109
89 (12)d
109
81
118
96
88
96
31 (44)e
27
0
79
20
65
86
477
431
17
608
351
52
91
95
88
43
105
99
98
71 (30)f
94
0
13
38
a NADPH was omitted from the system. b The enzyme oxidation
system was fully competent. c The parent probe and test substrate were
subjected to oxidation by the fully competent enzyme system; the
percent yields are based on the initial amount of test compound U
corrected by subtraction of the expected amount of this product from
oxidation of the probe. d Yield of ketone 2K. e Yield of ketone 3K.
f Yield of ketone 6K.
the cyclopropyl group slightly weakens a C-H bond on an
adjacent carbon.29
The stability test was quite severe for two reasons: (1) the
test substrates were added at the beginning of the reactions rather
than formed as products throughout the reactions, and (2) no
substrate other than the oxidation product being tested was
present. The latter point is important because the P450 and/or
reductase enzymes are destroyed during the course of the
reactions13 and the high concentrations of probe substrate relative
to oxidation products might provide some protection for the
oxidation products by competitive inhibition. Therefore, we
designed a potentially milder stability test for the rearranged
oxidation products. In the admixture stability test, the probe
substrate mixed with the corresponding rearranged alcohol
product R (in a small amount consistent with the amount of
oxidation observed in the probe studies) was subjected to the
enzymatic oxidation. The objective was to determine whether
rearranged oxidation product R would survive when a competent
oxidation reaction was in progress. The yields of rearranged
products were determined by GC and then corrected for the
amount of product R that was expected to be observed in the
enzyme oxidation of the probe (i.e., the amount observed in
Table 1 studies). In the admixture test, the yield of product
2R, which was formed in the enzyme oxidation of 2 (see Table
1), exceeded that originally added, suggesting good stability in
this more realistic test. Products 3R and 6R, which were
consumed completely in the severe stability test, were returned
in about half of the initial amounts added (65% and 38%,
respectively). Product 4R, which was not observed in the
Under the conditions used in this work, approximately 10%
of the probe substrate was converted to oxidation products. The
recoveries of unreacted probes averaged approximately 70%,
resulting in over 80% mass balance in most cases, a very high
recovery given that a small amount of relatively volatile probe
evaporates when the product mixture is concentrated for GC
analysis. An interesting and expected trend apparent from Table
1 is that there was a decrease in the efficiency of the enzyme
hydroxylations that correlates with the size of the probes.
Nevertheless, the oxidation products were formed in sufficient
quantities such that accurate quantitation was possible.
A series of control experiments was performed to determine
the recoverability and stability of the products to the oxidation
conditions (Table 2). In the first group of experiments, a stock
solution of substrate was prepared and equal portions of this
solution were added to three reaction mixtures, one lacking
enzymes, one lacking NADPH, and one completely competent
oxidizing mixture. The reactions were treated in the same
manner as the probe oxidation reactions. Following the reaction
period, equal amounts of a stock solution containing the
hydrocarbon standard were added to each mixture, and the
reactions were worked up and analyzed by GC. The GC results
from the buffer mixtures were used for calibration, and the
relative amount of oxidation product to standard was defined
(29) Walton, J. C. Magn. Reson. Chem. 1987, 25, 998-1000.

Cytochrome P450-Catalyzed Hydroxylation Reactions
J. Am. Chem. Soc., Vol. 120, No. 31, 1998 7723
conclusion from the cumulative observations with calibrated
radical clocks was that an unidentified process could be
contributing to the amount of rearranged products obtained from
the probe/clocks. Up to that point, all of the probes that gave
some rearranged products in P450-catalyzed hydroxylations had
the deficiency that the same skeletal reorganization would occur
for both radical and cationic rearrangements, and the possibility
existed that the unidentified process involved a cationic species.
We developed a probe/clock design (11) that maintained the
ultrafast radical rearrangement necessary for enzyme studies but
gave distinct products from radical and cationic rearrange-
ments.33,34 P450-catalyzed hydroxylation of probe 12 gave
Figure 2. Hydrogen abstractionsoxygen rebound mechanism for
P450-catalyzed hydroxylation.
oxidation of probe 4 (see Table 1), was returned in quite good
yield from the admixture test, consistent with the partial stability
found in the more severe test. In the Discussion, we will argue
that the absence of 4R from oxidation of probe 4 is one of the
most important observations of this work, and we consider the
stability results with 4R to be a clear indication that the absence
of detectable amounts of 4R from oxidations of 4 requires that
very little, if any, rearranged product was formed from probe
4.
In one sense, even the relatively mild admixture test should
be considered as a “worst case scenario” because the average
concentrations of the products in the probe oxidation reactions
are only half of the final concentrations. Further, the initial
concentrations of the oxidation products being tested in both
stability tests were greater than the final concentrations of these
products detected in the probe oxidation reactions, and the final
concentrations of products in all cases were 10% or less than
those of the probe substrates. The partial consumption of
oxidation products in the 3 and 6 series equivocates the absolute
quantitation of studies with these two probes, and we attempt
to address this in the Discussion. However, we conclude that
quantitation problems are relatively minor, especially for systems
2 and 4. In earlier studies, the same conclusion was reached
for P450-catalyzed hydroxylations of probes 1 and 5.12,18
products from both types of rearrangements, and the radical
lifetime from studies of hydroxylation of 12 was only 70 fs.8
Such a lifetime is too short for a true intermediate, the lifetime
for an intermediate calculated from conventional transition state
theory must exceed 150 fs, and requires that the radical can
only be a component of the transition structure for a concerted,
albeit nonsynchronous, insertion process.
The picture that emerges from studies with 12 is that the
exceptionally fast ring openings of the radicals from hypersensi-
tive probes (rate constants of about 1 × 1011 s-1 21
)
are minor
processes in competition with collapse of the transition structure
that gives unrearranged product (rate constant of ca. 1 × 1013
s-1) and cationic rearrangement products are produced by some
competing pathway. But this view leads to new questions. How
do “cationic” rearrangements occur during the hydroxylation
reactions, and what are the implications of these reactions in
regard to the nature of the oxidant?
Discussion
Mechanistic Descriptions of P450-Catalyzed Hydroxyla-
tions. The mechanism of P450-catalyzed hydroxylation that
held a consensus position for nearly two decades is shown in
Figure 2.4-7 An iron-oxo species similar to compound I of
peroxidase is the oxidant. This species is highly electrophilic,
described either as an FeVdO or an FeIVdO/porphyrin radical
cation. The iron-oxo abstracts hydrogen from an unactivated
CsH to give an alkyl radical that subsequently reacts with the
iron-hydroxy species in a homolytic substitution reaction
termed the oxygen rebound step. This mechanism was deduced
from the results of probe studies that gave rearranged products
consistent with a radical intermediate and was permissively
supported by the observation that kinetic isotope effects were
found in P450-catalyzed oxidations of deuterated substrates.
The series of studies that dispute the abstraction-rebound
mechanism had the initial objective of “timing” the oxygen
rebound step via the use of calibrated radical clocks. A wide
range of “radical” lifetimes were found,12,13,30-32 consistent with
free energies of activation for oxygen rebound ranging from 0
to 4 kcal/mol. Although enzymes can easily affect overall
activation barriers by more than 4 kcal/mol, the important points
here are the zero activation energy found in some cases and
the necessity that a relatively large barrier for collapse of a
radical pair would have to be imparted in other cases by steric
effects that have been demonstrated to be too small to hinder
free tumbling of probe 1 in the active site.12 An eventual
Recent mechanistic proposals contain competitive processes
that might explain the above results. We envisioned two
possible routes to cations shown in Figure 3.8 In (A), an
electron-transfer process would compete with collapse in the
transition state of the hydroxylation reaction. In (B), the first-
formed product in the hydroxylation step is a protonated alcohol
subject to a solvolytic-type rearrangement that competes with
deprotonation. The former process could be cast in the context
of the iron-oxo as the reactive intermediate, whereas the latter
seems to require that the actual oxidant is iron-complexed
hydrogen peroxide that inserts the OH+ moiety. The recent
mechanistic description by Collman, Brauman, and co-workers10
offers an alternative explanation in the context of the iron-
oxo as oxidant; in this model, an agostic complex involving an
iron interaction with the C-H bond could rearrange in a manner
expected for cations before reductive elimination of alcohol
product occurred ((C) in Figure 3).
The mechanistic description by Shaik, Schwarz, and co-
workers11 is a new paradigm for P450 hydroxylations that relates
them to reactions of simple metal oxenoid complexes. In this
model, a competition exists between two reaction manifolds, a
high-spin (HS) manifold that is effectively “spin-unpaired” and
a low-spin (LS) manifold that is effectively “spin-paired”. A
portion of this complex model is shown in (D) in Figure 3. The
(30) Ortiz de Montellano, P. R.; Stearns, R. A. J. Am. Chem. Soc. 1987,
109, 3415-3420.
(31) Bowry, V. W.; Ingold, K. U. J. Am. Chem. Soc. 1991, 113, 5699-
(33) Le Tadic-Biadatti, M.-H.; Newcomb, M. J. Chem. Soc., Perkin
Trans. 2 1996, 1467-1473.
(34) Newcomb, M.; Chestney, D. L. J. Am. Chem. Soc. 1994, 116, 9753-
9754.
5707.
(32) Newcomb, M.; Le Tadic, M. H.; Putt, D. A.; Hollenberg, P. F. J.
Am. Chem. Soc. 1995, 117, 3312-3313.

7724 J. Am. Chem. Soc., Vol. 120, No. 31, 1998
Toy et al.
products U largely survived the control reactions although some
oxidation to aldehyde/ketone products K occurred, and the
rearranged alcohol product 4R from the isopropyl-containing
probe 4 was relatively stable even to the severe stability test.
Thus, the exact ratios of rearranged to unrearranged oxidation
products might not be known for some of the probes, but good
approximate ratios can be calculated. In the following discus-
sion, we use the “worst case” calculations of product ratios in
that we estimate that the ratios of unrearranged to rearranged
oxidation products (C/A) are smaller than those observed in
the probe oxidation studies. Specifically, for cases where
instability of rearranged product was indicated in the control
reactions, we have calculated C/A ratios that are adjusted for
the amounts of materials lost in the control reactions, and
accordingly, these are lower limits of the actual C/A ratios.
The rate constants for rearrangements of putative intermedi-
ates are necessary for quantitative discussions. A 2-aryl-
substituted cyclopropylcarbinyl cation has no “rate constant”
for rearrangement because it is not a true intermediate. Wiberg
has demonstrated experimentally and computationally that the
actual species is a composite of a cyclopropylcarbinyl cation
and a homoallyl cation and, importantly, that production of these
aryl-substituted cations in solvolytic reactions of 3-substituted
cyclobutyl tosylates resulted in exclusive formation of the
homoallyl products.35 Therefore, although no rate constant is
available, one knows that a “cyclopropylcarbinyl cation”
produced from probe 1 or 5 must proceed to ring-opened
product. The Wiberg results are for primary cyclopropylcarbinyl
systems, and alkyl substitution will stabilize the cationic center.
Given that an unsubstituted primary “cyclopropylcarbinyl
cation” is not an intermediate, that aryl substitution strongly
distorts the species toward the homoallyl resonance structure,35
and that a tertiary carbocation is less stable than a secondary
aryl-substituted carbocation as judged by gas-phase hydride ion
affinities,36 we assume that the same situation holds for all other
cases of interest here; any cation formed from a probe will result
in a rearranged product.
Figure 3. Possible competitions in P450 hydroxylations. In (A), an
electron transfer competes with collapse of a transition structure. In
(B), solvolysis and deprotonation of a first-formed protonated alcohol
product compete. In (C), rearrangement of an iron-bound species
competes with reductive elimination. In (D), a high-spin ensemble can
react to give a radical by H atom abstraction or experience a state
crossing to a low-spin ensemble that can react by insertion. See the
text for details and references.
reaction starts in the HS surface and either crosses to the LS
surface via spin inversion or remains on the HS surface. If
crossing to the LS surface occurs, insertion is possible and
stereochemical information is conserved. If, however, the
reaction remains on the HS surface, insertion is not possible,
and stereochemical information would be lost (or probe rear-
rangement would occur) due to the formation of a relatively
long-lived radical intermediate. A key point in this model is
that the probability of spin inversion, or state crossing from the
HS to LS surface, is a function of the position of the state
crossing point on the reaction coordinate and should be probe
dependent.
Implications of Probe Results. The two series of probes
employed in the present work provide some insight to the details
of the P450-catalyzed hydroxylation mechanism. A single
isozyme of P450 was employed, so the results are not
complicated by minor processes ascribed to reactions of other
enzymes. The enzyme turnover was reasonably high, and the
slight decrease in turnover with increasing bulk of the probes
is consistent with expected steric effects in binding the
substrates.
Because the enzyme system is reconstituted in liposomes,
the experiments were designed such that small amounts of
substrate were present in order to limit disruption of the enzyme
complex. We have previously observed smaller enzyme
turnover numbers when the concentrations of substrate were
increased.18 The small substrate concentrations and good
turnover numbers result in oxidation of about 10% of the probe
substrates, and secondary oxidation reactions occur. The severe
control reactions showed that ring-opened products 3R and 6R
are unstable in the absence of substrate, but the more realistic
admixture control reactions indicated that about half of these
materials would survive the reactions. The unrearranged
The rate constant for ring opening of the radical derived from
probe 1 is known.21 Using the same method for measuring the
ring-opening rate constants, indirect competition kinetics em-
ploying benzeneselenol trapping, we have determined the rate
constants for ring opening of the cyclopropylcarbinyl radicals
derived from probes 2 and 4;37 these radicals open somewhat
less rapidly than that from 1. Ring opening of the cyclopro-
pylcarbinyl radical derived from probe 5 has also been
determined by the same method;38 this radical opens about 3
times as fast as that from 1 at ambient temperature due to the
stabilizing effect of the CF3 group. We assume that the radical
derived from 3 will ring open with a rate constant equal to that
for the radical from 2, and we estimate the rate constant for
ring opening of the radical derived from 6 from the measured
value for that from 5 and the observed relative reactivity of the
radicals from 1 and 2.
The minimum value for the ratio of unrearranged to rear-
ranged products and the rate constants at 37 °C, the temperature
of the P450 studies, for the radicals derived from each probe
are collected in Table 3. Using these values, we compute the
minimum value for the oxygen rebound rate constant (kox) that
(35) Wiberg, K. B.; Shobe, D.; Nelson, G. L. J. Am. Chem. Soc. 1993,
115, 10645-10652.
(36) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic
Chemistry, 3rd ed.; Harper & Row: New York, 1987; p 407.
(37) Choi, S.-Y.; Toy, P. H.; Newcomb, M. Submitted for publication.
(38) Choi, S.-Y.; Toy, P. H.; Newcomb, M. Unpublished results.

Cytochrome P450-Catalyzed Hydroxylation Reactions
J. Am. Chem. Soc., Vol. 120, No. 31, 1998 7725
Table 3. Limiting Values for P450 Hydroxylations
a 20 kcal/mol barrier. More to the point, the observed substrate
turnover velocities for 1-6 (ca. 0.2 substrate/enzyme per
second) are 1-2 orders of magnitude greater than the velocities
of the cyclization reactions of the acyclic radicals for these
systems (rate constants on the order of 10-2 s-1). Thus, the
ring-opened radicals are demonstrably kinetically incompetent
as intermediates for formation of cyclic products.
We believe the combined evidence is compelling that radical
intermediates are not formed and some process other than radical
ring opening in competition with rebound contributes to the
amounts of rearranged products. The objective of studying the
series of probes discussed here was to attempt to evaluate the
possibility that an electron-transfer process, an oxidation of the
radical in the transition state to a cation, competes with collapse
to product. Such a pathway seems possible in light of the
apparent alkyl radical oxidation reactions that occur with, for
example, the P450-related prostacyclin and thromboxane syn-
thetase enzymes,39 and it has been proposed both in the “simple”
view of P450 hydroxylation as an insertion with a competing
process8 and as one possible reaction in the more complex
mechanistic picture presented by Shaik et al.11
The results strongly suggest that cation formation by radical
oxidation is not important for the series of probes 1-4 and the
pair 5 and 6. An increase in methyl substitution at a radical
center results in a decrease in the oxidation potential. Given
that the radical does not exist as an intermediate, an electron-
transfer reaction would have to be an outer-sphere process, and
the distances between the radical and the iron species that
oxidizes it must be similar in all cases because there is too little
time for atomic translation of more than a few tenths of an
angstrom. Therefore, the amount of cation formation should
track with the oxidation potentials. As noted above, the 2-aryl-
substituted cyclopropylcarbinyl cations that would be formed
will rearrange. If we assume an essentially constant, small
amount of radical ring opening in the TS of the insertion process
as a background reaction, the results found here clearly are going
in the wrong direction for the oxidation model. One sees less
rearrangement with more highly substituted systems, and
ultimately, the tertiary radical precursor 4, progenitor to the
radical that should be oxidized most efficiently, apparently reacts
with no cation formation. In a qualitative sense, related results
were reported by Ingold and Bowry, who found no rearranged
products from P450-catalyzed hydroxylation of isopropylcy-
clopropane, the substrate among their probe set most likely to
give a cation.31 This inconsistency of the electron-transfer
hypothesis was previously noted in a comparison of the results
from probe 5 with those from probe 1,18 and we believe the
results of the present study effectively exclude this mechanism.
The agostic complex alternative presented by Collman et al.10
also appears not to be supported by the results of the present
study. In that model, cationic-type rearrangements would occur
for the iron complex in competition with reductive elimination
that gives alcohol. One might expect that both the energetic
preference for more highly substituted cations and the steric
compression resulting from increased alkyl substitution would
favor more rearrangement for the more highly substituted system
in contradistinction to the observations for 1-4. It is conceiv-
able, however, that the steric demand from addition of methyl
groups at the reactive center results in a large increase in the
rate of the reductive elimination from the complex, which would
reduce the lifetime of the complex and the amount of rear-
rangement. In that case, one might compare the results from 1
probe
C/Aa
kRb (s-1
)
koxc (s-1
)
τd (ps)
1
2
3
4
5
6
4.3
25
26
3.0 × 1011
1.5 × 1011
(1.5 × 1011)e
1.2 × 1011
9 × 1011
1.3 × 1012
0.8
0.3
0.3
<0.08
0.3
4 × 1012
4 × 1012
>100
>1.2 × 1013
4 × 1012
4
12
(4 × 1011)f
5 × 1012
0.2
a Minimum ratio of unrearranged to rearranged product; see the text.
b Rate constant for ring opening of radicals at 37 °C; see the text.
c Minimum oxygen rebound rate constant. d Maximum radical lifetime.
e Rate constant assumed to be equal to that for the radical from 2.
f Estimated rate constant.
would be necessary if radical ring opening was the only pathway
to rearranged products and the corresponding radical lifetime
(τ ) 1/kox).
The results in Table 3 in isolation from other results might
be rationalized in terms of an extremely fast rebound rate
constant if one assumes that the results from 1 and 4 are
anomalous. Two obvious problems with this conclusion are
(1) the minimum value for the oxygen rebound at which the
results appear to converge (4 × 1012 s-1) is nearly as fast as
collapse of a transition state as calculated from conventional
transition state theory (6 × 1012 s-1) and (2) as judged from
the product stability studies, the results with probes 1 and 4
should be among the most reliable. In addition, other probes
with the same P450 isozyme have given apparent rebound rate
constants ranging from 2 × 1010 to 1.5 × 1013 s-1 8,12,13,30-32
.
Thus, if radical rearrangements were the only pathway possible
for ring-opened products, then the rate constant for the oxygen
rebound step would be highly variable which is, in our opinion,
unreasonable for a process that must have virtually no activation
energy in many cases. We emphasize that the results with the
isopropyl-containing probe 4 are especially noteworthy because
the products were demonstrated to be relatively stable. The
apparent oxygen rebound rate constant for 4 is a minimum value
limit based on the failure to detect any rearranged product, but
this minimum value exceeds the theoretical maximum predicted
from conventional transition state theory and is quite close to
the value obtained from probe 12, which permitted separation
of the radical- and cation-derived rearrangement products.
The possible explanation that some steric effects of the
enzyme have slowed the rearrangement rates of all of the
radicals derived from probes 1-6 (and those from other probes
that gave subpicosecond radical lifetimes) by 2 or more orders
of magnitude and the oxygen rebound rate constant is actually
approximately 2 × 1010 s-1 (the smallest apparent rebound rate
constant yet found) has, in our opinion, reached the point of an
absurdity. Another especially damaging fact in regard to this
steric constraint argument is the observation that isotopic
substitution at the methyl position in probe 1 resulted in
metabolic switching between methyl and phenyl oxidation which
requires that the probe tumbled freely in the active site of the
enzyme;12 that is, no steric constraints on probe 1 were apparent
with the isozyme used in this work.
Another rationalization of the results in the context of radical
reactions, involving selective trapping of an equilibrating system
of ring-opened and cyclic radicals, is not possible for several
reasons. The equilibrium constants for ring opening of the
radicals derived from 1-6 and other aryl-substituted cyclopro-
pane systems21 are about 1013, which would require that trapping
of cyclic radicals is about 15 orders of magnitude faster than
trapping of ring-opened radicals; i.e., if the cyclic radical
collapsed with no energy barrier, the acyclic radical would have
(39) Ullrich, V.; Brugger, R. Angew. Chem., Int. Ed. Engl. 1994, 33,
1911-1919.

7726 J. Am. Chem. Soc., Vol. 120, No. 31, 1998
Toy et al.
to those from 5 and the results from 2 to those from 6. Both
pairs would have similar steric compression effects, and less
rearrangement should be observed for 5 and 6 due to the
trifluoromethyl group that strongly destabilizes a cation. This
was not the observed trend.
Figure 4. “Normal” reaction sequence in the protonated alcohol model.
Threonine in its neutral or alcoholate form serves as a base to activate
iron-complexed hydrogen peroxide, and insertion of OH+ is followed
by rapid deprotonation.
The agostic complex alternative might be in conflict with
the overall steric demands presented by various probes. The
complex involves an iron-carbon bond, and it is not clear that
the large probes can approach an iron atom held in a porphyrin
ring. This problem seems to be especially severe in the case
of oxidation of (2,2-diphenylcyclopropyl)methane, a known
substrate for P450,13 which apparently would have to form a
complex with one phenyl group protruding into the porphyrin.
in the probes or the presence of external magnetic fields, as
suggested by the authors,11 are definitely warranted.
The electron-transfer, agostic complex, and multistate models
are cast in terms of an iron-oxo species as the oxidant in P450.
It is possible, however, that the actual oxidant in P450 might
be iron-complexed hydrogen peroxide, a precursor to the iron-
oxo species. Oxidation by an iron-hydrogen peroxide complex
would be similar to oxidation by protonated hydrogen peroxide,
a reaction demonstrated by Olah decades ago,40 and supported
computationally much more recently.41 Some evidence for this
possibility exists in addition to previous probe results; Pratt
concluded that a hydrogen peroxide complex was the oxidant
in P450-catalyzed oxidations on the basis of the observation
that H2O2-shunted P450 reacted differently from P450 activated
by iodosobenzene.42
If an iron-complexed hydrogen peroxide was the oxidant in
P450 hydroxylations, then the first-formed product would be a
protonated alcohol from insertion of the elements of OH+ into
a C-H σ bond, and solvolysis of this protonated alcohol to
give a cation would compete with a deprotonation that gives
neutral alcohol product. The “normal” sequence is represented
in Figure 4, where, for the sake of speculation below, we show
deprotonation by threonine, which is known to be highly
conserved in the active sites of P450 enzymes43 and, in the case
of cytochrome P450cam, is located near the heme.44
The formation of rearrangement products in P450 hydrox-
ylations of the probes studied here and previously is consistent
with a “protonated alcohol” model in some ways but not in
others. Most importantly, some type of cationic species is
implicated from the results with probe 12 discussed earlier where
radical and cationic rearrangement products differ, and both were
observed in P450 oxidations.8 A protonated alcohol with aryl
substitution on the cyclopropane ring would be expected to
undergo an internally assisted solvolysis reaction leading to ring-
opened product as shown below for 13. Cyclopropane-based
probes lacking the aryl group would be expected to be less
reactive toward the solvolysis, and little or no ring opening
occurs with such probes with the exception of the sterically
strained probes 14.13,31 Finally, the observation that bicyclo-
[2.1.0]pentane (15) suffers partial rearrangement upon P450
hydroxylation30,31 is consistent with the protonated alcohol
model in that the bicyclo[2.1.0]pent-2-yl system is known to
be especially reactive in solvolytic reactions.45
The Shaik/Schwarz mixed electronic state paradigm presents
a number of pathways for rearrangements,11 but one important
point must be appreciated at the outset. The amount of
stereochemical or structural rearrangement observed with a given
probe can be controlled by the efficiency of intersystem crossing
(spin inversion) processes and, as such, will not necessarily track
with the kinetics of rearrangement of a radical or cationic
intermediate. If the reacting complex fails to cross from the
HS surface to the LS surface, then “insertion” is not possible,
and a radical intermediate would be produced. In addition, with
the very fast rates of rearrangement of the radicals corresponding
to the probes used in this study, some rearrangement could occur
even when the reacting ensemble crosses to the LS surface; this
portion of the model is effectively the same picture one has for
rearrangement occurring in competition with collapse in an
insertion process.
For a meaningful analysis of our results in the context of the
mixed state model, one can assume that the extent of rearrange-
ment that occurs on the LS surface will be nearly the same for
each of the probes 1-4, an assumption based on the observation
that the rate constants for ring opening of the cyclopropylcar-
binyl radicals corresponding to 1, 2, and 4 are in a ratio of 1.0:
0.5:0.4. Accordingly, because no rearranged product was
observed with probe 4, the “background” rearrangement on the
LS surface is effectively undetectable, and all rearranged
products would be ascribed to reactions occurring on the HS
surface.
The probability of state crossing is linked to the position of
the HS/LS crossing point on the reaction coordinate, and this
point will be a function of the donor property of the reaction
center, with a better electron donor center having an earlier
position and, thus, a higher probability of crossing. In this
context, the probe results of the present study are consistent
with the two-state reactivity model. The increasing donor
properties of the reactive center in moving from primary probe
1 to the secondary (2 and 3) and tertiary (4) probes result in
increasingly earlier and more efficient HS to LS crossings, with
the tertiary probe 4 ultimately crossing completely. The results
with the CF3-substituted probes 5 and 6 also fit this model, not
only because the amount of rearrangement for the secondary
probe 6 was less than that with primary probe 5 but also because
the effect of the remote CF3 group should be to reduce slightly
the donor abilities at the reactive center and disfavor HS to LS
crossing, giving slightly more rearrangement for 5 and 6 relative
to 1 and 2, respectively, as is observed.
Nevertheless, the results of the present study do not lend
support to this model. On one hand, the cation-stabilizing
effects of alkyl groups in the series of probes 1-4 should have
(40) Olah, G. A.; Parker, D. G.; Yoneda, N. Angew. Chem., Int. Ed.
Engl. 1978, 17, 909-931.
(41) Bach, R. D.; Mintcheva, I.; Estevez, C. M.; Schlegel, H. B. J. Am.
Chem. Soc. 1995, 117, 10121-10122.
(42) Pratt, J. M.; Ridd, T. I.; King, L. J. J. Chem. Soc., Chem. Commun.
1995, 2297-2298.
The Shaik/Schwarz model presents several competing pro-
cesses that could result in rearranged products which might be
considered a weakness, but it is the mechanistic description most
in line with our results. Other tests of this model that attempt
to alter the probability of spin inversion by heavy atom effects
(43) Nelson, D. R.; Strobel, H. W. Biochemistry 1989, 28, 656-660.
(44) Raag, R.; Martinis, S. A.; Sligar, S. G.; Poulos, T. L. Biochemistry
1991, 30, 11420-11429.
(45) Wiberg, K. B.; Williams, V. Z., Jr.; Friedrich, L. E. J. Am. Chem.
Soc. 1970, 92, 564-567.

Cytochrome P450-Catalyzed Hydroxylation Reactions
J. Am. Chem. Soc., Vol. 120, No. 31, 1998 7727
Conclusion
The probe results reported here, combined with previous
results, seem to leave little doubt that the hydrogen abstraction-
oxygen rebound mechanism for cytochrome P450 hydroxylation
is not complete. The mixtures of unrearranged and rearranged
alcohol products and the widely disparate values of the oxygen
rebound rate constants found with many hypersensitive probes
require that some competition exists in P450 hydroxylations,
either in the product-forming processes from a single oxidizing
species or from reactions of two oxidants. Accepting that the
two oxidants model cannot be evaluated from extant probe
results, the probe results of the present study seem to be best in
line with predictions of the multiple-state paradigm recently
presented by Shaik, Schwarz, and co-workers.11 New tests of
the various recent hypotheses are in order, but we believe that
one of the most important advances in understanding the
mechanisms of P450-catalyzed hydroxylations that can be made
at this time is an unequivocal demonstration of the active
oxidant.48,49
given more cationic rearrangement for more highly substituted
systems; i.e., probe 4 should have given the most rearrangement,
not the least. In a similar manner, the effect of the electron-
withdrawing CF3 group in probes 5 and 6 should be to disfavor
the aryl-assisted solvolysis reaction shown in 13, and less
rearrangement should have been observed for 5 relative to 1
and for 6 relative to 2 and 3.
A weak rationalization for the results in the context of the
protonated alcohol model would note that, unlike the radical
oxidation model with its outer-sphere electron-transfer event
competing with collapse to product, the deprotonation reaction
that would compete with solvolysis in the protonated alcohol
model is a bond-forming event and will be subject to steric and
(with the requisite short lifetimes) orientational effects. We offer
as speculation the possibilities that methyl substitution at the
cyclopropylcarbinyl positions in the probes could increase the
kinetic acidity by favorable positioning of the protonated alcohol
moiety with respect to the base that deprotonates it (Figure 4)
or disfavor the solvolytic process by a similar orientational
biasing. In either case, increased substitution at the reactive
center would lead to less cationic rearrangement. On balance,
we consider the present and past probe results to be inconclusive
at best regarding the formation of a protonated alcohol product
formed by insertion of the elements of OH+ from a hydrogen
peroxide complex.
The mechanistic probe results require some type of competi-
tion, and the above models present competing processes in the
product-forming reactions. An alternative possibility that must
now be considered is that a competition exists between oxidants.
From comparisons of the product mixtures in oxidations by a
truncated version of the P450 isozyme 2B4 and its mutant
lacking the highly conserved threonine in the active site, Vaz,
Coon, and co-workers deduced that multiple oxidant forms were
possible.46 In quite recent work, the same group showed that,
for two P450 isozymes (2B4 and 2E1), the ratios of hydrox-
ylation to epoxidation products observed in oxidations of alkenes
changed in proceeding from the wild-type active site isozymes
to mutants in which active site threonine was replaced with
alanine.47 This observation suggests not only that multiple
oxidant forms are possible but also that two distinct electrophilic
oxidants exist in the natural course of P450 oxidation reactions,
possibly a hydroperoxo-iron complex and the iron-oxo
species.47 The possibility that multiple oxidant forms can exist
definitely confuses the mechanistic issue for P450 hydroxylation.
One could readily rationalize all probe results in the context of
the “two oxidants” model because, if two oxidants actually exist,
one does not know how the probes should behave with either
of the oxidants in isolation. P450 2B4 is known to oxidize the
hypersensitive probe 1,13 and it seems apparent that comparisons
of probe oxidations by the wild-type and mutant enzymes would
be enlightening. Such studies might address not only the two
oxidants model of Vaz and Coon but also the “protonated
alcohol” model in which iron-complexed hydrogen peroxide is
the oxidant.
Experimental Section
General Methods. Commercially available reagents were purchased
from either the Sigma or Aldrich Chemical Co. and were used as
received. All moisture sensitive reactions were carried out in flame-
dried glassware under a nitrogen atmosphere. Tetrahydrofuran (THF)
and diethyl ether were distilled under a nitrogen atmosphere from
sodium and benzophenone ketyl. Methylene chloride was distilled
under a nitrogen atmosphere from phosphorus pentoxide. Benzene was
distilled under a nitrogen atmosphere from calcium hydride. Dimethyl
sulfoxide (DMSO) and dimethyl formamide (DMF) were distilled in
vacuo from calcium hydride.
NMR spectra were acquired on a Varian Gemini 300 spectrometer.
Gas chromatography analyses were performed using flame ionization
detection on a Varian 3400 chromatograph (15 m × 0.54 mm bonded
phase Carbowax column, Alltech). Gas chromatography-mass spectral
(GC-MS) analyses were performed using a Hewlett-Packard Model
5890 GC interfaced to a Hewlett-Packard Model 5971 mass selective
detector (30 m × 0.25 mm capillary bonded phase Carbowax column,
Alltech). High-resolution mass spectral analyses were performed by
the Central Instrumentation Facility at Wayne State University (Detroit,
MI). Melting points were determined using a Unimelt capillary melting
point apparatus (Thomas-Hoover) and are uncorrected. Radial chro-
matography was performed on a Chromatotron Model 7294T (Harrison
Research Corp.) using plates coated with TLC grade silica gel (Merck)
with gypsum binder and fluorescent indicator.
Enzymatic oxidation reactions were performed in reconstituted
systems using the purified isozyme CYP2B1 and expressed reductase
at 37 °C in 50 mM potassium phosphate buffer, pH 7.4, containing 1
mM desferrioxamine and 0.075 M KCl, with a total reaction volume
of 2 mL. A mixture of 0.6 nmol of CYP2B1 and 1.2 nmol of reductase
in buffer was allowed to stand on ice for 5 min. A suspension of DLPC
(0.96 nmol) in buffer was sonicated. The enzyme and micelle mixtures
were combined, and this mixture was allowed to stand on ice for 5
min. The mixture was diluted with buffer, and the substrate was added
as a solution in methanol (10 µL). The reaction mixture was incubated
for 2 min at 37 °C, and the reaction was initiated by the addition of
NADPH (final concentration of 1.2 mM). The reaction mixture was
shaken gently at 37 °C for 30 min and then extracted with methylene
chloride (3 × 2 mL). The combined organic phase was dried over
(48) We note that a spectroscopic detection of an iron-oxo in P450
probably will not provide such an unequivocal demonstration. Recent studies
of microperoxidase oxidations indicate that hydroxylic solvent can react
with an iron-oxo species, the reverse of formation of an iron-oxo from
an iron-hydrogen peroxide complex, on the time scale of oxidation
reactions. See ref 49.
(46) Vaz, A. D. N.; Pernecky, S. J.; Raner, G. M.; Coon, M. J. Proc.
Natl. Acad. Sci. U.S.A. 1996, 93, 4644-4648.
(47) Vaz, A. D. N.; McGinnity, D. F.; Coon, M. J. Proc. Natl. Acad.
Sci. U.S.A. 1998, 95, 3555-3560.
(49) Osman, A. M.; Boeren, S.; Boersma, M. G.; Veeger, C.; Rietjens,
I. M. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 4295-4299. Osman, A. M.;
Koerts, J.; Boersma, M. G.; Boeren, S.; Veeger, C.; Rietjens, I. M. Eur. J.
Biochem. 1996, 240, 232-238.

7728 J. Am. Chem. Soc., Vol. 120, No. 31, 1998
Toy et al.
MgSO4 and filtered through a plug of glass wool. An internal standard
was added, and the solution was concentrated under a slow stream of
nitrogen at room temperature to a final volume of approximately 0.2
mL. Quantitation of the product ratios was performed using gas
chromatography with flame ionization detection, and product identities
were determined by comparison with authentic samples using selected
ion monitoring (SIM) GC-MS analysis. Phenolic products from 2-4
were identified by their mass spectra and GC retention times.
(trans-2-Phenylcyclopropyl)propane (3). A solution of trans-1-
phenyl-1-pentene20 (7) (1.90 g, 13.00 mmol) in CH2Cl2 (150 mL) was
cooled to -30 °C. To this was added a solution of Et2Zn (1.0 M in
hexanes, 65.0 mL, 65.0 mmol). The reaction mixture was stirred for
5 min, and CH2I2 (10.50 mL, 130.3 mmol) was added dropwise over
30 min. The reaction mixture was allowed to warm slowly to room
temperature and was stirred for 16 h. The resulting suspension was
poured into a saturated, aqueous NH4Cl solution (300 mL) and extracted
with CH2Cl2 (3 × 200 mL). The combined organic phase was washed
with brine (200 mL), dried over MgSO4, filtered, and concentrated in
vacuo at 0 °C.
The crude product was dissolved in CH2Cl2 (100 mL), and
3-chloroperoxybenzoic acid (2.00 g, 11.59 mmol) was added. The
reaction mixture was stirred at room temperature for 6 h and then
washed sequentially with 15% aqueous NaOH solution (200 mL), water
(200 mL), and brine (200 mL). The organic layer was separated, dried
over MgSO4, filtered, and concentrated in vacuo at 0 °C. The crude
product was purified by radial chromatography (pentane) to afford 3
(0.98 g, 6.12 mmol, 47%) as a clear, colorless oil. 1H NMR (CDCl3):
δ 0.72-0.79 (1H, m), 0.88 (1H, dt, J1 ) 8.4 Hz, J2 ) 4.8 Hz), 0.94
(3H, t, J ) 7.2 Hz), 0.97-1.08 (1H, m), 1.32-1.53 (4H, m), 1.60 (1H,
dt, J1 ) 8.4 Hz, J2 ) 4.8 Hz), 7.02-7.27 (5H, m). 13C NMR
(CDCl3): δ 13.9, 16.0, 22.5, 23.2, 23.5, 36.5, 125.1, 125.6 (2C), 128.2
(2C), 144.1. HRMS: calcd for C12H16, 160.1252; found, 160.1248.
1-(trans-2-Phenylcyclopropyl)propanone (3K). A solution of
trans-2-phenyl-1-cyclopropanecarboxylic acid (8.00 g, 49.3 mmol) and
DMF (5 drops) in benzene (100 mL) under a nitrogen atmosphere was
cooled with a water bath as oxalyl chloride (8.60 mL, 98.6 mmol) was
added dropwise. The reaction mixture was stirred for 1 h, and volatiles
were removed in vacuo. The resulting crude carbonyl chloride was
dissolved in CH2Cl2 (200 mL) under a nitrogen atmosphere and cooled
to 0 °C. N,O-Dimethylhydroxylamine hydrochloride (5.30 g, 54.3
mmol) and pyridine (8.80 mL, 108.8 mmol) were added sequentially,
and the resulting solution was stirred at room temperature for 2 h. The
reaction mixture was washed sequentially with 10% aqueous HCl
solution (150 mL), saturated aqueous NaHCO3 solution (150 mL), and
brine (150 mL). The organic layer was dried over MgSO4, filtered,
and concentrated in vacuo. The crude product was chromatographed
on silica gel (40% ethyl acetate in hexanes) to afford N-methoxy-N-
methyl-trans-2-phenylcyclopropanecarboxamide (9.25 g, 45.1 mmol,
91%) as a clear, colorless oil.
NaOH solution (1.20 mL) and water (3.60 mL). The resulting
suspension was stirred at room temperature for 1 h and then filtered.
The filtrate was concentrated in vacuo, and the resulting crude product
was chromatographed on silica gel (20% ethyl acetate in hexanes) to
afford the two diastereomers of 3U as clear, colorless oils.
1
3Ua (2.20 g, 12.5 mmol, 50%). H NMR (CDCl3): δ 0.93-0.99
(2H, m), 1.04 (3H, t, J ) 7.5 Hz), 1.22-1.31 (1H, m), 1.61-1.76 (2H,
m), 1.91 (1H, dt, J1 ) 8.4 Hz, J2 ) 5.7 Hz), 2.11 (1H, br s), 3.11 (1H,
dt, J1 ) 7.2 Hz, J2 ) 6.0 Hz), 7.10-7.31 (5H, m). 13C NMR
(CDCl3): δ 10.1, 13.7, 20.6, 29.2, 30.0, 76.8, 125.5, 125.9 (2C), 128.3
(2C), 142.9. HRMS: calcd for C12H16O, 176.1201; found, 176.1202.
3Ub (1.20 g, 6.81 mmol, 27%). 1H NMR (CDCl3): δ 0.95-1.07
(2H, m), 1.04 (3H, t, J ) 7.5 Hz), 1.23-1.31 (1H, m), 1.72 (2H, dt, J1
) 13.8 Hz, J2 ) 6.9 Hz), 1.85 (1H, dt, J1 ) 9.0 Hz, J2 ) 5.1 Hz), 2.22
(1H, br s), 3.13 (1H, dt, J1 ) 7.8 Hz, J2 ) 6.3 Hz), 7.09-7.32 (5H,
m). 13C NMR (CDCl3): δ 10.2, 13.2, 21.2, 29.4, 30.3, 77.0, 125.6,
125.8 (2C), 128.4 (2C), 142.5. HRMS: calcd for C12H16O, 176.1201;
found, 176.1194.
(E)- and (Z)-1-Phenyl-3-hexen-1-ol (3R). A solution of 3Ua (0.80
g, 4.54 mmol) in THF (20 mL) was cooled to -30 °C under a nitrogen
atmosphere. To this were added sequentially triethylamine (1.20 mL,
8.61 mmol) and methanesulfonyl chloride (0.40 mL, 5.17 mmol). After
1 h, water (15 mL) was added, and the reaction mixture was heated at
reflux for 2 h. The reaction mixture was cooled to room temperature
and extracted with ether (3 × 25 mL). The combined organic phase
was washed sequentially with saturated, aqueous NaHCO3 solution (50
mL) and brine (50 mL), dried over MgSO4, and filtered. Concentration
of the filtrate in vacuo gave crude product that was purified by radial
chromatography (15% ethyl acetate in hexanes) to afford 3R25 (0.26
g, 1.48 mmol, 33%) as a clear, colorless oil. GC analysis showed this
to be a 13:1 mixture of isomers.
Mixture of isomers. 1H NMR (CDCl3): δ 0.92-1.03 (3H, m), 2.02-
2.11 (2H, m), 2.39-2.50 (2H, m), 2.41 (1H, s), 4.64-4.59 (1H, m),
5.36-5.46 (1H, m), 5.54-5.67 (1H, m), 7.27-7.36 (5H, m).
Major isomer. 13C NMR (CDCl3): δ 13.8, 25.7, 42.7, 73.6, 124.6,
125.9 (2C), 127.3, 128.3 (2C), 136.4, 144.2. MS: calcd for C12H16O,
176; (M)+ found at m/z ) 176 for each isomer.
Methyl trans-2-(4-(Trifluoromethyl)phenyl)cyclopropylacetate
(9). A solution of 818 (1.69 g, 7.34 mmol) and DMF (2 drops) in
benzene (50 mL) under a nitrogen atmosphere was cooled with a water
bath as oxalyl chloride (1.3 mL, 14.9 mmol) was added. The reaction
mixture was stirred for 2 h, and volatiles were removed in vacuo. The
crude carbonyl chloride was dissolved in CH2Cl2 (25 mL), and the
resulting mixture was added slowly to a solution of CH2N2 in ether,
prepared from Diazald (10 gm, 46.7 mmol). The reaction mixture was
stirred at room temperature for 16 h and then concentrated in vacuo.
The resulting diazoketone was chromatographed on silica gel (20%
ethyl acetate in hexanes).
The diazoketone was dissolved in dioxane (30 mL), and the solution
was added portionwise to a suspension of Ag2O (3.00 g, 12.9 mmol)
and Na2S2O3 (2.04 g, 12.9 mmol) in water (100 mL) at 75 °C. The
resulting suspension was stirred for 2 h, then cooled to room
temperature, and filtered through Celite. The filtrate was made basic
(NaOH) and washed with ether (2 × 100 mL). The aqueous layer
was acidified (concentrated HCl) and extracted with ether (2 × 100
mL). The combined organic layer was washed with brine (100 mL),
dried over MgSO4, filtered, and concentrated in vacuo.
The amide (9.25 g, 45.1 mmol) was dissolved in ether (200 mL)
and cooled to 0 °C under a nitrogen atmosphere. To this was added a
solution of ethylmagnesium bromide (3.0 M in ether, 20.0 mL, 60.0
mmol). After 2 h, the reaction mixture was quenched by carefully
pouring it into saturated, aqueous NH4Cl solution (200 mL). The
mixture was extracted with ether (3 × 100 mL). The combined organic
phase was washed sequentially with water (150 mL) and brine (150
mL), dried over MgSO4, filtered, and concentrated in vacuo. The crude
product was chromatographed on silica gel (15% ethyl acetate in
hexanes) to afford 3K24 (5.95 g, 34.1 mmol, 76%) as a clear, colorless
oil. 1H NMR (CDCl3): δ 1.12 (3H, t, J ) 7.5 Hz), 1.36 (1H, ddd, J1
) 8.4 Hz, J2 ) 6.6 Hz, J3 ) 4.2 Hz), 1.68 (1H, ddd, J1 ) 9.3 Hz, J2
) 5.4 Hz, J3 ) 4.5 Hz), 2.20 (1H, ddd, J1 ) 8.1 Hz, J2 ) 5.4 Hz, J3
) 3.9 Hz), 2.52 (1H, ddd, J1 ) 9.3 Hz, J2 ) 6.6 Hz, J3 ) 4.2 Hz),
2.64 (2H, q, J ) 7.2 Hz), 7.09-7.32 (5H, m). 13C NMR (CDCl3): δ
7.9, 18.8, 28.7, 31.9, 37.1, 126.0 (2C), 126.4, 128.5 (2C), 140.5, 209.4.
HRMS: calcd for C12H14O, 174.1045; found, 174.1047.
The resulting acid was dissolved in CH2Cl2 (25 mL) and added
slowly to a solution of CH2N2 in ether, prepared from Diazald (10 gm,
46.7 mmol). The mixture was stirred at room temperature for 16 h
and then concentrated in vacuo. The crude product was purified by
radial chromatography (10% ethyl acetate in hexanes) to afford 9 (1.04
gm, 4.04 mmol, 55% combined yield) as a clear, colorless oil. 1H
NMR (CDCl3): δ 0.94 (1H, dt, J1 ) 9.0 Hz, J2 ) 5.4 Hz), 1.05 (1H,
dt, J1 ) 8.4 Hz, J2 ) 5.4 Hz), 1.37-1.48 (1H, m), 1.81 (1H, dt, J1 )
9.0 Hz, J2 ) 4.8 Hz), 2.42 (2H, d, J ) 7.2 Hz), 3.70 (3H, s), 7.17 (2H,
d, J ) 8.1 Hz), 7.49 (2H, d, J ) 8.1 Hz). 13C NMR (CDCl3): δ 15.8,
19.1, 22.8, 38.5, 51.6, 124.3 (q, JC-F ) 270.3 Hz), 125.1 (2C, q, JC-F
) 3.4 Hz), 126.1 (2C), 127.8 (q, JC-F ) 33.1 Hz), 146.8, 172.8.
HRMS: calcd for C13H13F3O2, 258.0868; found, 258.0871.
1-(trans-2-Phenylcyclopropyl)propanol (3U). To a solution of 3K
(4.38 g, 25.1 mmol) in THF (150 mL) at 0 °C under a nitrogen
atmosphere was added LiAlH4 (1.20 g, 31.6 mmol). The mixture was
stirred under a nitrogen atmosphere for 2 h, and the reaction was
quenched by the sequential addition of water (1.20 mL), 15% aqueous

Cytochrome P450-Catalyzed Hydroxylation Reactions
J. Am. Chem. Soc., Vol. 120, No. 31, 1998 7729
(trans-2-(4-(Trifluoromethyl)phenyl)cyclopropyl)ethane (6). To
a solution of 9 (0.95 g, 3.68 mmol) in THF (25 mL) at 0 °C under a
nitrogen atmosphere was added LiAlH4 (0.15 g, 3.95 mmol). The
mixture was stirred under a nitrogen atmosphere for 2 h, and the reaction
was quenched by the sequential addition of water (0.15 mL), 15%
aqueous NaOH solution (0.15 mL), and water (0.45 mL). The resulting
suspension was stirred at room temperature for 1 h and then filtered.
The filtrate was concentrated in vacuo, and the resulting crude product
was purified by radial chromatography (30% ethyl acetate in hexanes)
to afford the corresponding alcohol (0.67 g, 2.91 mmol, 79%) as a
clear, colorless oil. 1H NMR (CDCl3): δ 0.89-1.04 (2H, m), 1.11-
1.22 (1H, m), 1.46 (1H, br s), 1.61-1.77 (3H, m), 3.77 (2H, t, J ) 6.3
Hz), 7.12 (2H, J ) 8.4 Hz), 7.49 (2H, J ) 8.4 Hz). 13C NMR
(CDCl3): δ 16.3, 21.1, 22.7, 27.1, 62.5, 124.4 (q, JC-F ) 270.2 Hz),
125.1 (2C, q, JC-F ) 33 Hz), 125.7 (2C), 127.7 (q, JC-F ) 33.1 Hz),
147.8. HRMS: calcd for C12H13F3O, 230.0918; found, 230.0916.
A solution of the above alcohol (0.47 g, 2.04 mmol) in THF (20
mL) was cooled to -30 °C under a nitrogen atmosphere. To this were
added sequentially triethylamine (0.60 mL, 4.30 mmol) and methane-
sulfonyl chloride (0.20 mL, 2.58 mmol). The mixture was stirred at
-30 °C for 45 min and then cooled to -78 °C. A solution of LiBEt3H
(1.0 M in THF, 10.0 mL, 10.0 mmol) was added dropwise. The mixture
was allowed to warm slowly to room temperature and was stirred for
16 h. The reaction mixture was quenched by the successive addition
of 15% aqueous NaOH solution (5 mL) and 30% H2O2 solution (5
mL). The resulting mixture was heated at reflux for 1 h and then cooled
to room temperature. The organic layer was separated, and the aqueous
layer was extracted with ether (3 × 25 mL). The combined organic
phase was washed with brine (50 mL), dried over MgSO4, filtered,
and concentrated in vacuo at 0 °C. The crude product was purified by
radial chromatography (pentane) to afford 6 (0.34 g, 1.59 mmol, 78%)
as a clear, colorless oil. 1H NMR (CDCl3): δ 0.82-9.95 (2H, m),
1.01 (3H, t, J ) 7.2 Hz), 1.02-1.11 (1H, m), 1.42 (2H, dt, J1 ) 14.4
Hz, J2 ) 6.9 Hz), 1.66 (1H, dt, J1 ) 9.0 Hz, J2 ) 4.8 Hz), 7.11 (2H,
d, J ) 8.1 Hz), 7.48 (2H, d, J ) 8.1 Hz). 13C NMR (CDCl3): δ 13.3,
16.7, 23.0, 26.5, 27.4, 124.5 (q, JC-F ) 269.1 Hz), 125.5 (2C, q, JC-F
) 3.3 Hz), 125.7 (2C), 127.4 (q, JC-F ) 32.0 Hz), 148.6. HRMS:
calcd for C12H13F3, 214.0969; found, 214.0971.
1-(trans-2-(4-(Trifluoromethyl)phenyl)cyclopropyl)ethanone (6K).
Amide 1018 (5.24 g, 19.2 mmol) was dissolved in ether (200 mL) and
cooled to 0 °C under a nitrogen atmosphere. To this was added a
solution of methylmagnesium bromide (3.0 M in ether, 8.0 mL, 24.0
mmol). After 2 h, the reaction mixture was quenched by carefully
pouring it into a saturated, aqueous NH4Cl solution (200 mL). This
was extracted with ether (3 × 150 mL). The combined organic phase
was washed sequentially with water (200 mL) and brine (200 mL),
dried over MgSO4, filtered and concentrated in vacuo. The crude
product was chromatographed on silica gel (15% ethyl acetate in
hexanes) to afford 6K (4.02 g, 17.6 mmol, 92%) as a clear, colorless
oil. 1H NMR (CDCl3): δ 1.34 (1H, ddd, J1 ) 8.1 Hz, J2 ) 6.6 Hz, J3
) 4.5 Hz), 1.65 (1H, ddd, J1 ) 9.3 Hz, J2 )5.4 Hz, J3 ) 4.5 Hz), 2.26
(1H, ddd, J1 ) 8.4 Hz, J2 ) 5.1 Hz, J3 ) 4.2 Hz), 2.26 (3H, s), 2.51
(1H, ddd, J1 ) 9.6 Hz, J2 ) 6.9 Hz, J3 ) 4.5 Hz), 7.14 (2H, d, J ) 8.1
Hz), 7.47 (2H, d, J ) 8.1 Hz). 13C NMR (CDCl3): δ 12.1, 28.1, 30.5,
32.7, 124.2 (q, JC-F ) 270.2 Hz), 125.3 (2C, q, JC-F ) 3.3 Hz), 126.2
(2C), 128.5 (q, JC-F ) 32.0 Hz), 144.6, 206.0. HRMS: calcd for
C12H11F3O, 228.0762; found, 228.0764.
(E)- and (Z)-1-(4-(Trifluoromethyl)phenyl)-3-penten-1-ol (6R). A
solution of 5R18 (4.76 g, 22.0 mmol), imidazole (3.00 g, 44.1 mmol),
and tert-butyldimethylsilyl chloride (5.00 g, 33.2 mmol) in CH2Cl2 (150
mL) was stirred under a nitrogen atmosphere at room temperature for
24 h. The reaction mixture was diluted with additional CH2Cl2 (200
mL) and then washed sequentially with 10% aqueous HCl solution
(200 mL), saturated, aqueous NaHCO3 solution (200 mL), and brine
(200 mL). The organic phase was dried over MgSO4, filtered, and
concentrated in vacuo. The crude product was chromatographed on
silica gel (10% ethyl acetate in hexanes) to afford the silyl ether (5.70
g, 21.7 mmol, 99%) as a clear, colorless oil. 1H NMR (CDCl3): δ
-0.12 (3H, s), 0.04 (3H, s), 0.89 (9H, s), 2.33-2.49 (2H, m), 4.74
(1H, t, J ) 5.7 Hz), 4.96-5.03 (2H, m), 5.67-5.81 (1H, m), 7.41 (2H,
d, J ) 8.1 Hz), 7.56 (2H, d, J ) 8.1 Hz). 13C NMR (CDCl3): δ -5.0,
-4.8, 18.1, 25.7 (3C), 45.3, 74.4, 117.5, 125.0 (2C, q, JC-F ) 3.3 Hz),
126.1 (2C), 124.2 (q, JC-F ) 270.2 Hz), 129.2 (q, JC-F ) 32.0 Hz),
134.3, 149.1.
To a solution of the above silyl ether (1.91 g, 7.28 mmol) in ether
(20 mL) was added an aqueous solution (20 mL) of NaIO4 (17.50 g,
81.8 mmol) and OsO4 (0.05 g, 0.19 mmol). The reaction mixture was
stirred vigorously at room temperature for 24 h. After cooling, the
suspension was filtered. The filtrate was extracted with ethyl acetate
(3 × 50 mL), and the combined organic phase was washed with brine
(75 mL), dried over MgSO4, filtered, and concentrated in vacuo. The
resulting aldehyde was used without purification. 1H NMR (CDCl3):
δ -0.13 (3H, s), 0.57 (3H, s), 0.87 (9H, s), 2.64 (1H, ddd, J1 ) 16.2
Hz, J2 ) 4.5 Hz, J3 ) 1.8 Hz), 2.86 (1H, ddd, J1 ) 16.2 Hz, J2 ) 7.8
Hz, J3 ) 2.4 Hz), 5.28 (1H, dd, J1 ) 7.5 Hz, J2 ) 4.2 Hz), 7.46 (2H,
d, J ) 8.1 Hz), 7.60 (2H, d, J ) 8.1 Hz), 9.78 (1H, t, J ) 2.1 Hz).
To a suspension of ethyltriphenylphosphonium bromide (5.57 g, 15.0
mmol) in benzene (50 mL) under a nitrogen atmosphere was added a
solution of BuLi (2.5 M in hexanes, 5 mL, 12.5 mmol). The mixture
was stirred at room temperature for 15 min, and a solution of the above
crude aldehyde in benzene (20 mL) was added via cannula. After 15
min, the reaction mixture was quenched by the addition of acetone (10
mL), and the resulting suspension was filtered through silica gel. The
filtrate was concentrated in vacuo, and the crude product was filtered
through silica gel (20% ethyl acetate in hexanes). The crude alkene
was used without further purification.
The above alkene was dissolved in THF (20 mL), and a solution of
Bu4NF (1.0 M in THF, 10.0 mL, 10.0 mmol) was added. The reaction
mixture was stirred at room temperature for 2 h and then poured into
a saturated, aqueous NH4Cl solution (50 mL). The resulting mixture
was extracted with ether (3 × 75 mL), and the combined organic phase
was washed sequentially with saturated, aqueous NaHCO3 solution (75
mL) and brine (75 mL), dried over MgSO4, filtered, and concentrated
in vacuo. The crude product was purified by radial chromatography
(10% ethyl acetate in hexanes) to afford 6R (0.86 g, 3.74 mmol, 51%
combined yield). GC analysis showed this to be a 4:1 mixture of
isomers. 1H NMR (CDCl3) of the mixture: δ 1.58-1.71 (3H, m), 2.00
(1H, br s), 2.33-2.61 (2H, m), 4.72-4.81 (1H, m), 5.36-5.46 (1H,
m), 5.59-5.75 (1H, m), 7.45-7.62 (4H, m). MS: calcd for C12H13F3O,
230; (M)+ found at m/z ) 230 for each isomer.
Acknowledgment. We thank the National Institutes of
Health for support of this work from Grants GM-48722 (M.N.)
and CA-16954 (P.F.H.). We are grateful to Professors S. Shaik
and H. Schwarz and Dr. D. Schroder for discussions regarding
their mechanistic model and comments on an earlier version of
the manuscript.
1-(trans-2-(4-(Trifluoromethyl)phenyl)cyclopropyl)ethanol (6U).
To a solution of 6K (2.92 g, 12.8 mmol) in THF (100 mL) at 0 °C
under a nitrogen atmosphere was added LiAlH4 (0.50 g, 13.2 mmol).
The mixture was stirred under a nitrogen atmosphere for 2 h, and the
reaction was quenched by the sequential addition of water (0.50 mL),
15% aqueous NaOH solution (0.50 mL), and water (1.50 mL). The
resulting suspension was stirred at room temperature for 1 h and then
filtered. The filtrate was concentrated in vacuo, and the resulting crude
product was chromatographed on silica gel (20% ethyl acetate in
hexanes) to afford 6U (2.93 g, 12.7 mmol, 99%) as a clear, colorless
oil. 1H NMR (CDCl3): δ 0.95-1.13 (2H, m), 1.25-1.30 (1H, m),
1.33 (3H, d, J ) 6.0 Hz), 1.60 (1H, br s), 1.82-1.99 (1H, m), 3.38-
3.47 (1H, m), 7.13-7.17 (2H, m), 7.49 (2H, d, J ) 7.8 Hz). MS:
calcd for C12H13F3O, 230; (M)+ found at m/z ) 230 for each isomer.
Supporting Information Available: Table of results from
CYP2B1 hydroxylation of probes 2-4 and 6, 1H and 13C NMR
1
spectra of 3, 6, 9, 3K, 3Ua, 3Ub, 3R, and 6K, and H NMR
spectra of diastereomeric mixtures of 6U and 6R (19 pages,
print/PDF). See any current masthead page for ordering
information and Web access instructions.
JA981157I