Desaturation of alkylbenzenes by cytochrome P450BM3 (CYP102A1)

Article abstract of DOI:10.1002/chem.200801927

A study was conducted to investigate the desaturation of alkylbenezenes by cytochrome P450_BM3 (CYP102A1). It was observed during the study that oxidation of alkylbenzenes with CYP102A1 involves a gamut of P450 activity types that are terminal, sub-terminal benzylic and aromatic hydroxylation; terminal and sub-terminal desaturation; and epoxidation of the olefins. It was also found that the desaturation of cumene by CYP102A1 is sensitive to α-deuteration and insensitive to β-deuteration. Intramolecular deuterium isotope study revealed that the first abstraction in the desaturation of valproic acid and ezlopitant by microsomal P450s take place from activated carbon atoms. The increased β-hydroxylation percentages show that the β-carbon lies closer to the ferryl oxygen in the KT5.

Full text of DOI:10.1002/chem.200801927

COMMUNICATION
DOI: 10.1002/chem.200801927
Desaturation of Alkylbenzenes by Cytochrome P450BM3 (CYP102A1)
[
a]
Christopher J. C. Whitehouse, Stephen G. Bell, and Luet-Lok Wong*
Cytochrome P450 monooxygenases catalyse the oxidation
of a broad array of endogenous and exogenous organic sub-
strates, typically inserting an oxygen atom from atmospheric
dioxygen into a carbonÀhydrogen bond to give the corre-
elution time (Figure S2), to arise from monooxygenase ac-
tivity. This was in due course identified as the CÀC desatura-
tion product, p,a-dimethylstyrene, using GC co-elution with
an authentic sample. Small amounts of p,a-dimethylstyrene
oxide, the further oxidation product, were also produced. In
the light of this finding, the oxidation of a series of alkylben-
zenes structurally related to p-cymene was investigated in a
search for complementary examples of desaturase activity.
In addition to WT CYP102A1, a selection of variants was
[1]
sponding alcohol. Other known types of activity include
olefin epoxidation, carbonÀcarbon bond cleavage, heteroa-
tom oxidation, dealkylation and dehydrogenation across CÀ
[2]
O, CÀN and CÀC bonds. Few examples of dehydrogena-
tion by bacterial P450 systems have been reported.
CYP102A1 (P450BM3), a medium- to long-chain fatty acid
hydroxylase from Bacillus megaterium, has been shown to
[15]
studied, with F87A/A330P/E377A/D425N (KT5)
giving
the most striking product distributions (Figure 1, Table 1).
[3]
aromatise Nifedipine by CÀN dehydrogenation, and a
triple mutant was able to oxidise acetaminophen by CÀN
[4]
and CÀO dehydrogenation. Wild-type CYP102A1 gave
0
.9% CÀC dehydrogenation (desaturation) with a-thujone
(
Figure S1), while CYP101A1 (P450 ) from Pseudomonas
cam
[5]
putida gave 13.7%. Substrates found to undergo desatura-
tion in microsomal systems include valproic and lauric
[6,7]
[8]
[9]
acids,
ezlopitant and capsaicinoids (Figure S1). In
some cases, aromatisation or the formation of conjugated
products, such as styrenes, appeared to provide a driving
[9–11]
force for the reaction.
Wild-type CYP102A1 (WT) monooxygenates most non-
natural substrates at nugatory rates, although directed evolu-
tion and site-specific mutagenesis may be employed to en-
[12–14]
hance activity.
Introducing mutations can cause sub-
strates to bind in altered orientations, allowing different CÀ
H bonds to be oxidised and transforming product profiles.
During an investigation into the oxidation of p-cymene by a
family of recently reported variants, we were interested to
encounter significant quantities of a product that seemed
unlikely, on the basis of its short gas chromatography (GC)
Figure 1. Product distributions in the oxidation of alkylbenzenes by wild-
type CYP102A1 (WT) and variant KT5 (F87A7A330P/E377A/D425N).
[15]
Desaturation took place between the a- and b-positions
of the alkyl substituents of all the substrates studied, though
to varying extents. WT gave small quantities, while higher
percentages were obtained from F87A-containing variants
such as KT5. Isopropylbenzene (cumene) gave 29% a,b-de-
saturation (including a-methylstyrene oxide formation) with
variant KT5, while sec-butylbenzene gave 11%, ethylben-
zene 3% and propylbenzene just 0.7%. p-Cymene and p-
isopropylanisole gave 22 and 23%, respectively (Tables S5
and S6). Desaturation was therefore most pronounced when
the a-carbon of the alkyl substituent was branched and the
[
a] C. J. C. Whitehouse, Dr. S. G. Bell, Dr. L.-L. Wong
Department of Chemistry, University of Oxford
Inorganic Chemistry Laboratory, South Parks Road
Oxford OX1 3QR (UK)
Fax : (+44)1865272690
E-mail: luet.wong@chem.ox.ac.uk
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200801927.
Chem. Eur. J. 2008, 14, 10905 – 10908
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
10905

Table 1. Product distributions in the oxidation of alkylbenzenes by wild-
type CYP102A1 (WT) and KT5 (F87A/A330P/E377A/D425N).
tion of capsaicinoids is thought to be initiated by abstraction
from the less activated w-position. Carbocation formation
[9]
%
a-
Hydroxylation
% Desaturation
a,b- b,g- oxides
via electron transfer from the incipient w-radical to the oxi-
dising species then occurs, followed by rearrangement to the
more stable w-1 carbocation. In principle, an analogous
mechanism could operate in CYP102A1, with successive ab-
stractions taking place from the less activated b-carbon. b-
Abstraction followed by electron transfer would yield a b-
cation with an intrinsic tendency to rearrange to the more
highly substituted benzylic a-cation (Scheme 1). Rearrange-
ment should be facile in cumene, where the a-carbon is ter-
tiary and the b-carbon terminal, potentially allowing a,b-de-
saturation to compete effectively against b-hydroxylation as
well as a-hydroxylation. In propylbenzene, however, there
would be a limited probability of rearrangement, which
could explain the low desaturation levels observed (0.7%
with KT5 vs 29% for cumene).
b-
g-
o-
propylbenzene
ethylbenzene
WT
KT5
98.7 0.3
20.3 78.3 0.3
–
0.6
–
0.2
0.6
0.2
0.4
–
0.1
WT
KT5
90.2
93.2 3.9
–
n/a 9.8
–
2.9
n/a
n/a
–
–
n/a
–
sec-butylbenzene
[
a]
WT
KT5
63.2 32.8
30.5 56.3 1.1
–
1.8
–
1.5
7.9
–
–
0.7
3.2
[
b]
cumene
WT
KT5
81.8
67.9 3.2
–
n/a 9.3
n/a
8.9
n/a
–
–
27.5 n/a
1.4
a-[D]cumene
WT
KT5
44.0
–
n/a 47.5 8.5
n/a
–
54.1 16.3 n/a
–
28.8 n/a
0.8
6
b-[D ]cumene
WT
KT5
84.3
72.3
–
–
n/a 8.8
n/a
6.9
25.5 n/a
n/a
–
2.2
–
p-cymene
WT
KT5
83.7
77.9
–
–
n/a 10.4 5.9
n/a
–
3.2
n/a 18.9 n/a
–
[
a] Includes 0.7% p-sec-butylphenol. [b] percentages do not sum to 100
due to a minority product. n/a=not applicable. See Tables S1–S6 for de-
tailed product profiles.
b-carbon(s) terminal. NADPH utilisation efficiencies (cou-
pling) and product formation rates (PFRs) did not appear to
be adversely affected by high desaturation levels. p-Cymene,
À1
À1
for example, gave a PFR of 702 nmolmin (nmol P450)
with KT5 at a coupling efficiency of 50% (Table S7). The
principal oxidation pathway was generally a-hydroxylation,
but significant levels of b-hydroxylation occurred in sub-
strates possessing a sub-terminal b-carbon (e.g., 33% from
sec-butylbenzene with WT). Substrates with terminal b-
carbon atoms gave smaller quantities of b-alcohols, consis-
tent with CÀH bond reactivity trends, while terminal g-alco-
Scheme 1. Possible reaction pathways in the oxidation of alkylbenzenes
by CYP102A1 (R , R , R =H or Me).
1
2
3
hols were produced by substrates possessing a g-carbon
atom. With certain substrates, o-hydroxylation or O-deme-
thylation (Table S6) were encountered. Interestingly, trace
amounts of b,g-desaturation were observed with propylben-
zene (though not with sec-butylbenzene). In addition to
giving higher desaturation percentages than WT across all
substrates, KT5 also gave more b-hydroxylation (e.g., 78 vs
In a search for evidence of a carbocationic rearrangement,
we investigated the oxidation of a-[D]cumene to establish
whether desaturation entailed the migration of the single a-
deuterium to the b-carbon. GC-MS analysis gave the mass
of the a-methylstyrene formed as 118.176 versus a theoreti-
cal mass for the non-deuterated product of 118.176, while
NMR spectroscopy showed the non-aromatic protons to be
in a 1:1:3 ratio, with no loss of fine structure at any of the
three resonances when compared to an authentic sample.
The absence of deuterium in the product demonstrated that
desaturation did not proceed via carbocationic rearrange-
ment with a-to-b proton transfer. The NADPH consumption
rates measured in these experiments were similar to those
observed with non-deuterated cumene, but coupling and
PFRs were substantially lower (Table 2). Peroxide uncou-
pling levels were unchanged, indicating that the oxidase
pathway was primarily responsible for the decrease in cou-
0.3% with propylbenzene), while WT gave more o-hydrox-
ylation than KT5 (e.g., 11 vs 0% with cumene).
The formation of b- and g-alcohols indicates that the first
hydrogen abstraction by the P450 oxyferryl compound I spe-
cies can take place from the b- or g-carbon of an alkyl sub-
stituent as well as from the more activated benzylic a-
carbon. Desaturation between C and C could therefore be
a
b
initiated either by a- or b-abstraction (Scheme 1). Intramo-
lecular deuterium isotope effects show that the first abstrac-
tion in the desaturation of valproic acid and ezlopitant (see
Figure S1) by microsomal P450s takes place from one of the
[6,8]
more activated carbon atoms.
By contrast, the desatura-
10906
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 10905 – 10908

Desaturation of Alkylbenzenes
COMMUNICATION
pling. a-Deuteration resulted in significant levels of meta-
bolic switching. KT5 gave higher b-hydroxylation levels at a
higher absolute PFR while WT gave higher o-hydroxylation
percentages, again at a higher PFR. The isotope effects ob-
served for desaturation were lower than those for a-hydrox-
ylation. These metabolic switching levels and isotopic sensi-
tivities were higher than those observed in the microsomal
Whereas desaturation competes effectively against hydroxyl-
ation in the oxidation of cumene by KT5, hydroxylation
dominates when propylbenzene is the substrate even though
C and C are both accessible to the ferryl oxygen (1% de-
a
b
saturation vs 20% a-hydroxylation and 78% b-hydroxyl-
ation). Evidently, the substitution pattern at Ca and C is
b
critical; in this context, it is worth noting that the isopropyl
group has featured prominently among the examples of de-
[8]
P450 oxidation of monodeuterated ezlopitant (Figure S1),
[5,8,9]
which has the same desaturation site structure as a-
saturation reported in other P450 enzymes.
It remains to
[16]
[D]cumene. b-Deuteration of cumene
had contrastingly
be established whether substitution levels influence desatu-
ration yields by retarding radical rebound, accelerating the
second hydrogen abstraction, causing substrates to orientate
differently in the active site, or by some other means. It is
also unclear at this point whether electronic effects are in-
volved. While we have demonstrated that deuterium does
not migrate to Cb during the oxidation of a-[D]cumene,
which argues against the involvement of b-cation rearrange-
ment, this mechanism cannot yet be ruled out as the
little impact on oxidation profiles, the only significant iso-
tope effect being that associated with the abolition of b-hy-
droxylation in KT5.
Table 2. Product formation rates (PFRs) and isotopic sensitivities for the
oxidation of cumene (H), a-[D]cumene (a-[D]) and b-[D ]cumene (b-
6
[
D]) by wild-type CYP102A1 (WT) and variant KT5.
PFR
PFR(H)/PFR(D)
III
H
a-[D]
b-[D]
a-[D]
b-[D]
Fe (OH) intermediate could capture the a-proton while in
transit to the b-carbon in preference to abstracting a second
b-proton.
WT
a-hydroxylation
b-hydroxylation
o-hydroxylation
a,b-desaturation
a,b-oxide formation
TOTAL
coupling (%)
peroxide (%)
KT5
a-hydroxylation
b-hydroxylation
o-hydroxylation
a,b-desaturation
a,b-oxide formation
a,b-desaturation
106
–
12
12
–
130
31
33
26
–
29
5.1
–
60
13
33
116
–
12
10
–
138
34
34
4.0
–
0.4
2.3
–
0.9
–
1.0
1.2
–
In conclusion, the oxidation of alkylbenzenes, a relatively
simple class of compounds, by CYP102A1 involves a gamut
of competing P450 activity types: terminal, sub-terminal,
benzylic and aromatic hydroxylation, terminal and sub-ter-
minal desaturation, epoxidation of the resulting olefins and,
in one case, O-dealkylation (p-isopropylanisole, Table S6).
Research into alkylbenzenes with different branching struc-
tures and other compound types will be required to support
and develop the mechanistic framework outlined above. It
will be interesting, for example, to see whether natural sub-
strates, such as unsaturated fatty acids, particularly branched
2.2
0.9
189
8.9
–
77
3.9
81
45
13
–
24
0.7
25
217
–
–
76
6.6
83
4.2
0.7
–
3.2
5.9
3.3
0.9
Inf.
–
1.0
0.6
1.0
[
17]
fatty acids,
substrates.
are as prone to desaturation as non-natural
CYP102A1 is a convenient vehicle to explore
+
a,b-oxide formation
[
7,18]
total
coupling [%]
peroxide [%]
279
37
14
83
10
14
300
41
14
3.4
0.9
desaturation as it is self-sufficient and has variants that can
rapidly oxidise a wide array of non-natural substrates in
Coupling=total product formation as % of NADPH consumption. Inf.=
infinite. PFRs expressed in nmolmin (nmol P450) . All data are means
of at least three experiments with standard deviations <5% of the mean.
vitro. KT5 is also well suited for the study of the partition
À1
À1
[19,20]
between desaturation and w-hydroxylation,
as it is an
effective desaturase and also possesses some w-hydroxylase
activity. However, even the wild-type enzyme shows slight
desaturase activity with most of the substrates investigated,
suggesting that this rarely reported P450 activity type may
occur more widely than hitherto supposed.
The fact that the desaturation of cumene by CYP102A1 is
sensitive to a-deuteration but relatively insensitive to b-
deuteration implies that the first of the two hydrogen ab-
[6]
stractions takes place predominantly from the a-carbon.
Given that a-hydroxylation is more sensitive to a-deutera-
tion than a,b-desaturation, there could, however, be a com-
peting desaturation pathway initiated by b-abstraction. The
increased b-hydroxylation percentages given by KT5 with
most substrates suggest that the b-carbon lies closer to the
ferryl oxygen in KT5 than in WT, but the concomitantly
high desaturation yields given by KT5 do not in fact sub-
stantiate the existence of a desaturation pathway initiated
by b-abstraction, since a partition shift in favour of desatura-
tion would also be expected if the first abstraction took
place from the a-carbon. Nor is the relatively short oxyferr-
Acknowledgements
This work was supported by the EPSRC and BBSRC, UK (grant ref. EP-
D048559-1). We thank Dr. Mark Moloney for assistance with organic
synthesis and Dr. Nick Rees for NMR spectra.
yl–C distance in KT5 sufficient in itself to tilt the hydroxyl-
ation/desaturation partition in favour of desaturation.
Keywords: CÀH
activation
·
cytochromes
·
b
dehydrogenation · isotope effects · oxidase
Chem. Eur. J. 2008, 14, 10905 – 10908
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10907

L.-L. Wong et al.
[
[
[
11] K. Kassahun, K. Skordos, I. McIntosh, D. Slaughter, G. Doss, T.
Baillie, G. Yost, Chem. Res. Toxicol. 2005, 18, 1427–1437.
12] S. G. Bell, N. Hoskins, C. J. C. Whitehouse, L. L. Wong, Met. Ions
Life Sci.2007, 3, 437–476.
13] V. B. Urlacher, S. G. Bell, L. L. Wong in Modern Biooxidation (Eds.:
R. D. Schmid, V. Urlacher), Wiley-VCH, Weinheim, 2007, pp. 99–
[
[
1] J. T. Groves in Cytochrome P450: Structure, Mechanism and Bio-
chemistry, 3rd ed. (Ed.: P. R. Ortiz de Montellano), Kluwer Academic/
Plenum, New York, 2005, pp. 1–85.
2] P. R. Ortiz de Montellano, J. J. De Voss in Cytochrome P450: Struc-
ture, Mechanism and Biochemistry, 3rd ed. (Ed.: P. R. Ortiz de Mon-
tellano), Kluwer Academic/Plenum, New York, 2005, pp. 183–245.
3] G. Di Nardo, A. Fantuzzi, A. Sideri, P. Panicco, C. Sassone, C.
Giunta, G. Gilardi, J. Biol. Inorg. Chem. 2007, 12, 313–323.
1
22.
[
[
[
14] J. D. Bloom, M. M. Meyer, P. Meinhold, C. R. Otey, D. MacMillan,
F. H. Arnold, Curr. Opin Struct. Biol. 2005, 15, 447–452.
15] C. J. Whitehouse, S. G. Bell, H. G. Tufton, R. J. Kenny, L. C. Ogilvie,
L. L. Wong, Chem. Commun. 2008, 966–968.
4] B. M. van Vugt-Lussenburg, M. C. Damsten, D. M. Maasdijk, N. P.
Vermeulen, J. N. Commandeur, Biochem. Biophys. Res. Commun.
[
2
006, 346, 810–818.
5] Y. Jiang, X. He, P. R. Ortiz de Montellano, Biochemistry 2006, 45,
33–542.
6] A. E. Rettie, M. Boberg, A. W. Rettenmeier, T. A. Baillie, J. Biol.
Chem. 1988, 263, 13733–13738.
[
[
16] K. Sugiyama, W. F. Trager, Biochemistry 1986, 25, 7336–7343.
17] M. J. Cryle, R. D. Espinoza, S. J. Smith, N. J. Matovic, J. J. De Voss,
Chem. Commun. 2006, 2353–2355.
[
[
[
5
[
[
18] P. H. Buist, Nat. Prod. Rep. 2004, 21, 249–262.
19] A. E. Rettie, P. R. Sheffels, K. R. Korzekwa, F. J. Gonzalez, R. M.
Philpot, T. A. Baillie, Biochemistry 1995, 34, 7889–7895.
20] D. Kumar, S. P. De Visser, S. Shaik, J. Am. Chem. Soc. 2004, 126,
7] X. Guan, M. B. Fisher, D. H. Lang, Y. M. Zheng, D. R. Koop, A. E.
Rettie, Chem.-Biol. Interact. 1998, 110, 103–121.
[
[
[
8] R. S. Obach, Drug Metab. Dispos. 2001, 29, 1599–1607.
9] C. A. Reilly, G. S. Yost, Drug Metab. Dispos. 2005, 33, 530–536.
5
072–5073.
[
10] H. Sun, W. J. Ehlhardt, P. Kulanthaivel, D. L. Lanza, C. A. Reilly,
Received: September 18, 2008
G. S. Yost, J. Pharmacol. Exp. Ther. 2007, 322, 843–851.
Published online: November 10, 2008
10908
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 10905 – 10908