Chiral-Substrate-Assisted Stereoselective Epoxidation Catalyzed by H(2) O(2) -Dependent Cytochrome P450(SPα).

Abstract

The stereoselective epoxidation of styrene was catalyzed by H(2) O(2) -dependent cytochrome P450(SPα) in the presence of carboxylic acids as decoy molecules. The stereoselectivity of styrene oxide could be altered by the nature of the decoy molecules. In particular, the chirality at the α-positions of the decoy molecules induced a clear difference in the chirality of the product: (R)-ibuprofen enhanced the formation of (S)-styrene oxide, whereas (S)-ibuprofen preferentially afforded (R)-styrene oxide. The crystal structure of an (R)-ibuprofen-bound cytochrome P450(SPα) (resolution 1.9 Å) revealed that the carboxylate group of (R)-ibuprofen served as an acid-base catalyst to initiate the epoxidation. A docking simulation of the binding of styrene in the active site of the (R)-ibuprofen-bound form suggested that the orientation of the vinyl group of styrene in the active site agreed with the formation of (S)-styrene oxide.

Full text

Page 1

DOI: 10.1002/asia.201200250
Chiral-Substrate-Assisted Stereoselective Epoxidation Catalyzed by H2O2-
Dependent Cytochrome P450SPa
Takashi Fujishiro,[a]Osami Shoji,[a]Norifumi Kawakami,[a]Takahiro Watanabe,[a]
Hiroshi Sugimoto,[b]Yoshitsugu Shiro,[b]and Yoshihito Watanabe*[a, c]
Introduction
Oxidation reactions catalyzed by biocatalysts have been of
much interest in the field of organic synthesis because these
biocatalysts are able to produce fine chemicals, including
chiral building blocks and pharmaceuticals, under mild con-
ditions.[1]Cytochrome P450s (P450s) are ubiquitous enzymes
that are comprised of a superfamily of heme-containing
monooxygenases and are involved in oxidative metabolism,
detoxification, and in the synthesis of steroids.[2]P450s have
been regarded as attractive candidates as oxidation catalysts
because of their high catalytic activity for the direct inser-
tion of oxygen into unactivated C?H bonds.[3]However,
P450s consume a stoichiometric amount of an expensive co-
factor (NAD(P)H) in the reductive activation of molecular
oxygen.[3,4]Thus, the use of P450s in organic synthesis has
been limited to the formation of valuable products, such as
fine chemicals and drugs.[4]To overcome this limitation, Shi-
mizu and co-workers developed a hydrogen-peroxide-depen-
dent P450BM3by site-directed mutagenesis.[5]Whilst wild-
type P450BM3did not efficiently use hydrogen peroxide for
the generation of the active species (Compound I) followed
by the hydroxylation of fatty acids,[6]the F87A mutant of
P450BM3accepted hydrogen peroxide and catalyzed hydrox-
ylation reactions without the consumption of NADPH. Al-
though this mutant has been further developed to improve
its catalytic activity by directed evolution,[7]the final mutant
still requires a relatively high concentration of hydrogen
peroxide.
In contrast to most P450s, including P450BM3, hydrogen-
peroxide-dependent P450s, such as P450BSb,[8]P450SPa,[9]and
P450CLA,[10]exclusively use hydrogen peroxide as an oxidant
and efficiently catalyze the site-specific hydroxylation of
fatty acids. Although these P450s are expected to be practi-
cal biocatalysts, these enzymes usually show very high sub-
strate specificity. The crystal structure of a palmitic-acid-
bound form of P450BSbrevealed that the carboxylate of pal-
mitic acid interacted with the Arg-242 moiety that was locat-
ed near the heme.[11]The location of the carboxylate group
was almost the same as that of the distal glutamate in chlor-
operoxidase (CPO), which is one of the most efficient hy-
drogen-peroxide-dependent biocatalysts.[12]This observation
supported a proposal that the general acid–base function for
the facile generation of the active species was provided by
the carboxy group of the fatty acid that was bound to
P450BSb(Scheme 1a).[13]This unique reaction mechanism of
P450BSbalso contributed to its high substrate specificity and
P450BSb did not oxidize substrates other than long-alkyl-
chain fatty acids. However, we have demonstrated that the
oxidation reactions of non-natural substrates other than
long-alkyl-chain fatty acids could be catalyzed by P450BSbin
the presence of a series of short-alkyl-chain carboxylic acids
as “decoy molecules” that induced the substrate-misrecogni-
tion of P450BSb.[14]The addition of decoy molecules allowed
Abstract: The stereoselective epoxida-
tion of styrene was catalyzed by H2O2-
dependent cytochrome P450SPain the
presence of carboxylic acids as decoy
molecules. The stereoselectivity of sty-
rene oxide could be altered by the
nature of the decoy molecules. In par-
ticular, the chirality at the a-positions
of the decoy molecules induced a clear
difference in the chirality of the prod-
uct: (R)-ibuprofen enhanced the for-
mation of (S)-styrene oxide, whereas
(S)-ibuprofen preferentially
(R)-styrene oxide. The crystal structure
of an (R)-ibuprofen-bound cytochrome
afforded
P450SPa(resolution 1.9 ?) revealed that
the carboxylate group of (R)-ibuprofen
served as an acid–base catalyst to ini-
tiate the epoxidation. A docking simu-
lation of the binding of styrene in the
active site of the (R)-ibuprofen-bound
form suggested that the orientation of
the vinyl group of styrene in the active
site agreed with the formation of (S)-
styrene oxide.
Keywords: carboxylic acids · chir-
ality · decoy molecules · epoxida-
tion · hydrogen peroxide
[a] Dr. T. Fujishiro, Dr. O. Shoji, Dr. N. Kawakami, T. Watanabe,
Prof. Dr. Y. Watanabe
Department of Chemistry
Graduate School of Science
Nagoya University
Furo-cho, Chikusa-ku, Nagoya 464-8602 (Japan)
Fax: (+ +81)52-789-3557
E-mail: yoshi@nucc.cc.nagoya-u.ac.jp
[b] Dr. H. Sugimoto, Dr. Y. Shiro
RIKEN SPring-8 Center Harima Institute
1-1-1 Kouto, Mikazuki-cho, Sayo, Hyogo 679-5148 (Japan)
[c] Prof. Dr. Y. Watanabe
Research Center for Materials Science
Nagoya University
Furo-cho, Chikusa-ku, Nagoya 464-8602 (Japan)
2286
? 2012 Wiley-VCH Verlag GmbH&Co. KGaA, WeinheimChem. Asian J. 2012, 7, 2286–2293
FULL PAPER

Page 2

P450BSbto generate the active species and catalyze a variety
of reactions, such as the oxidations of guaiacol, styrene, eth-
ylbenzene, 1-methoxynaphthalene,
Thus, the use of decoy molecules is a useful method for the
generation of active species and for the alteration of sub-
strate specificity. In these “substrate-misrecognition sys-
tems,” the carboxylate group of the decoy molecule serves
as a general acid–base catalyst. X-ray crystal-structure anal-
ysis of a heptanoic-acid (decoy molecule)-bound form of
P450BSbshowed electron density that corresponded to the
carboxylate of heptanoic acid at the same position as that of
palmitic acid.[13]More recently, we solved the X-ray struc-
ture ofthepalmitic-acid-bound
(Figure 1),[16]which had 44% amino-acid-sequence identity
to P450BSb and also catalyzed the hydroxylation of fatty
andthioanisole.[14–15]
form ofP450SPa
Abstract in Japanese:
Scheme 1. a) Proposed reaction mechanism for the hydroxylation of long-alkyl-chain fatty acids catalyzed by P450SPa. The carboxylate group of the fatty
acid served as a general acid–base catalyst to generate Compound I species by using H2O2as an oxidant. b) Possible reaction mechanism for the oxida-
tion of non-natural substrates (styrene) other than long-alkyl-chain fatty acids catalyzed by P450SPain the presence of a carboxylic acid as a “decoy mole-
cule” ((R)-ibuprofen, green) that induced the substrate-misrecognition of P450SPa.
Chem. Asian J. 2012, 7, 2286–2293? 2012 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
www.chemasianj.org
2287

Page 3

acids. The regioselectivity of P450SPawas 100% a-selective,
whilst P450BSbhydroxylated both the a- and b positions with
a ratio of 43:57.[8–9]
The crystal structure of the palmitic-acid-bound form of
P450SParevealed that the key interactions between the car-
boxylate group of palmitic acid and the guanidine group of
the arginine moiety near the heme site were conserved in
P450SPa. Although two alternative conformations of palmitic
acid were observed (Figure 1b), the distance between the
carboxylate oxygen atom and the heme iron was 5.2 ? for
Conformation A and 5.5 ? for Conformation B (a produc-
tive conformation that gives the a-hydroxy fatty acid), thus
indicating that the location of the oxygen atom was similar
to that of the terminal carboxylate group of palmitic acid in
P450BSb(5.3 ?). These observations indicated that the reac-
tion mechanism for the formation of compound I of P450SPa
was the same as that of P450BSb(Scheme 1a). Encouraged
by the crystallographic studies of P450SPa, we decided to ex-
plore whether the monooxygenation system with decoy mol-
ecules could be expanded to P450SPa(Scheme 1b). We ex-
pected that P450SPawould show a different stereoselectivity
to that of P450BSbbecause P450SPacatalyzes the hydroxyl-
ation of fatty acids with high regioselectivity (a-selective)
and stereoselectivity (S-selective).[16]We were interested in
how the catalytic properties of the oxidation of unnatural
substrates were affected by the structure of the active site of
P450SPa, as well as the structure of decoy molecules. Herein,
we report the stereoselective oxidation of styrene catalyzed
by P450SPain the presence of a variety of decoy molecules,
including chiral decoy molecules such as (R)- and (S)-ibu-
profen. In addition, we performed the crystal-structure anal-
ysis of (R)-ibuprofen-bound P450SPain which electron densi-
ty that corresponded to (R)-ibuprofen was observed. The
crystal structure of the (R)-ibuprofen-bound form allowed
us to discuss the effect of the chirality of (R)-ibuprofen on
its binding and the stereoselectivity of the styrene-epoxida-
tion reaction.
Results and Discussion
Styrene Oxidation in the Presence of Carboxylic Acids
Initially, we investigated whether P450SPacatalyzed the ep-
oxidation of styrene in the presence of a series of short-
alkyl-chain carboxylic acids as decoy molecules. The sty-
rene-epoxidation reaction was catalyzed in the presence of
a decoy molecule to give the corresponding styrene oxide.
The turnover numbers were heavily dependent on the alkyl-
chain length of the carboxylic acids and were up to
165 min?1in the presence of heptanoic acid for wild-type
P450SPa(Table 1). This result clearly showed that the carbox-
ylate group of short-alkyl-chain carboxylic acids served as
an acid–base catalyst for the efficient generation of active
species, as was observed for P450BSb, thereby indicating that
Figure 1. a) Overall structure of H2O2-dependent cytochrome P450SPa
with palmitic acid:[16]The two channels were represented as light-brown
(Channel I) and light-blue surfaces (Channel II). b) Structure of the
active site around the heme group. Heme, Arg-241, Leu-78, and Phe-288
are represented as stick models. Palmitic acids of Conformation A (blue)
and Conformation B (yellow, the productive conformation that afforded
the a-hydroxy fatty acid) are shown as stick models.
Table 1. Epoxidation of styrene catalyzed by WT, L78F, and F288G
P450SPain the presence of carboxylic acids.
Decoy molecule (carbon-
chain length)
P450SPa Rate
[min?1][a]
ee [%] (R/
S)
SO/
PAA
octanoic acid (C8)WT
L78F
F288G
WT
L78F
F288G
WT
L78F
F288G
WT
L78F
F288G
WT
L78F
F288G
96(?8)22(?1) (R) 76:24
30(?1) (R) 73:27
38(?2) (S) 66:34
25(?2) (R) 76:24
34(?1) (R) 72:27
34(?4) (S) 66:34
28(?1) (R) 76:24
32(?1) (R) 74:26
27(?3) (S) 63:37
36(?6) (R) 76:24
31(?1) (R) 73:27
18(?3) (S)
33(?6) (R) 75:25
34(?1) (R) 73:27
16(?1) (S) 69:31
164(?18)
124(?17)
heptanoic acid (C7)165(?11)
277(?32)
180(?59)
hexanoic acid (C6)139(?16)
502(?37)
228(?10)
pentanoic acid (C5)145(?8)
319(?74)
322(?5)
106(?6)
67:33
butyric acid (C4)
310(?26)
354(?28)
[a] The unit for catalytic activity is (nmolproduct)min?1(nmolP450)?1;
uncertainty is given as the standard deviation for three measurements.
2288
www.chemasianj.org
? 2012 Wiley-VCH Verlag GmbH&Co. KGaA, WeinheimChem. Asian J. 2012, 7, 2286–2293
FULL PAPER
Yoshihito Watanabe et al.

Page 4

this monooxygenation system with decoy molecules was not
only applicable to P450BSbbut also to P450SPa. Interestingly,
in contrast to P450BSb, which gave (S)-styrene oxide, P450SPa
preferentially gave (R)-styrene oxide in the presence of hep-
tanoic acid under the same conditions. Structural compari-
son of the active site of P450SPawith that of P450BSbshowed
that Phe-288 and Leu-78 moieties in P450SPa
placed by Gly-290 and Phe-79 in P450BSb, respectively.[11]
These amino acids appeared to be key residues in determin-
ing the stereoselectivity; therefore, we examined the stereo-
selectivity of the styrene-epoxidation reactions catalyzed by
L78F and F288G mutants. Whereas the stereoselectivity of
the L78F mutant was unchanged, the F288G mutant gave
the opposite stereoselectivity to that catalyzed by WT
P450SPa. The mutation of Phe-288 into Gly-288 in P450SPa
might have provided a similar active site of P450BSband,
hence, gave (S)-styrene oxide. Interestingly, the L78F and
F288G mutants gave higher catalytic activity than WT. In
particular, the turnover of the L78F mutant with hexanoic
acid was more than 500 min?1, which is, to the best of our
knowledge, the highest catalytic activity for the epoxidation
of styrene among the heme enzymes with hydrogen perox-
ide reported so far.[7,17]To examine the effect of decoy mole-
cules on the catalytic properties, we carried out the epoxida-
tion of styrene with a variety of decoy molecules that con-
tained an aromatic ring because the aromatic ring of the
decoy molecule was expected to interact with both Phe-288
and with the aromatic group of styrene. The decoy mole-
cules used herein, as well as the results of the styrene-epoxi-
dation reactions, are summarized in Table 2. The stereose-
lectivity was drastically altered in the range 39% ee (R) to
88% ee (S) by changing the combination of the decoy mole-
cule and the mutant. Among the decoy molecules examined,
(R)-ibuprofen afforded the highest enantioselectivity in the
case of WT P450SPaand the combination of (R)-ibuprofen
and the F288G mutant gave 88% ee (S). In addition, the
chirality of ibuprofen induced a clear difference: (R)-ibupro-
fen gave (S)-styrene oxide, whilst (S)-ibuprofen gave (R)-
styrene oxide with WT P450SPa, thus showing that (R)-ibu-
profen was effective in enhancing S-selectivity. The R-chiral-
ity at the a-carbon atoms of the decoy molecules seemed to
be important for enhancing S-selectivity. Indeed, (R)-nap-
roxen also enhanced S-selectivity. These results showed that
the stereoselectivity could be controlled by simply selecting
the R- or S-enantiomer of ibuprofen.
[16]were re-
Crystal-Structure Analysis
To understand the asymmetric epoxidation reaction cata-
lyzed by P450SPawith ibuprofen, in particular the effect of
the chirality of the a-carbon atom of ibuprofen on the chir-
ality of the products, we determined the crystal structure of
an (R)-ibuprofen-bound form of wild-type P450SPaat a reso-
lution of 1.9 ? (Figure 2), which clearly showed the electron
density of (R)-ibuprofen in the active site. No remarkable
structural differences were observed between the overall
structures of the palmitic-acid-bound form and the (R)-ibu-
profen-bound form of wild-type P450SPa. Although the car-
boxylate was slightly separated from the heme moiety
(0.7 ?, Figure 2c), the location of the carboxylate group of
(R)-ibuprofen was essentially the same as that of palmitic
acid. The distance between one of the carboxylate oxygen
atoms of (R)-ibuprofen and the heme iron atom was 5.2 ?.
Thus, the carboxylate group of (R)-ibuprofen was expected
to serve as an acid–base catalyst for the generation of active
species with hydrogen peroxide. The phenyl ring of (R)-ibu-
profen interacted with the phenyl rings of Phe-287 and Phe-
288 in an edge-to-face manner. The para-isobutyl group of
(R)-ibuprofen was held close to the side-chain of Leu-77
through hydrophobic interactions. The methyl group at the
a-position of (R)-ibuprofen protruded toward the heme
moiety. Even after the binding of (R)-ibuprofen, there was
space to accommodate a styrene molecule. The channel that
was used for the access of fatty acids (Channel I) was occu-
pied by (R)-ibuprofen (Figure 1a), but Channel II was still
accessible for additional foreign substrates. Styrene may
have accessed the heme cavity through Channel II.
Although we wanted to examine the differences between
the active-site structures upon binding (R)- and (S)-ibupro-
fen, we could not obtain crystals of P450SPawith (S)-ibupro-
fen. The crystal structure of the (R)-ibuprofen-bound form
showed that the distance between the a-carbon atom and
the side-chain methyl group of Ala-245 was 3.5 ? (Fig-
ure 2b). This result implied that, if (S)-ibuprofen tended to
stay in the same location as (R)-ibuprofen, the methyl group
Table 2. Epoxidation of styrene catalyzed by WT and F288G P450SPα
in the presence of carboxylic acids that contained an aromatic ring.
Decoy moleculeP450SPa Rate
[min?1][a]
ee [%] (R/
S)
SO/
PAA
WT
F288G
WT
F288G
WT
F288G
189(?30)39(?2) (R)
22(?1) (S)
27(?1) (R)
29(?1) (S)
11(?3) (R)
72:28
74:26
77:23
72:28
77:23
74:26
89(?15)
217(?25)
180(?32)
302(?31)
281(?36)50(?1 (S)
WT
F288G
94(?9) 26(?1) (R)
22(?1) (S)
23(?1) (S)
37(?1) (S)
76:24
68:3290(?15)
WT
F288G
160(?5)83:17
69:31151(?20)
WT
F288G
43(?4)
12(?1)
4(?1) (R)
31(?2) (S)
82:18
73:27
WT
F288G
92(?11)63(?3) (S)
88(?1) (S)
95:05
90:10176(?27)
[a] The unit for catalytic activity is (nmolproduct)min?1(nmolP450)?1;
uncertainty is given as the standard deviation for three measurements.
Chem. Asian J. 2012, 7, 2286–2293? 2012 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
www.chemasianj.org
2289
Chiral-Substrate-Assisted Stereoselective Epoxidation

Page 5

of (S)-ibuprofen at the a-carbon atom would induce a steric
repulsion with the side-chain methyl group of Ala-245. This
steric repulsion may have led to a higher dissociation con-
stant of (S)-ibuprofen, whereas we could not estimate the
dissociation constants of (R)- and (S)-ibuprofen owing to no
appreciable change in the UV/Vis spectra of P450SPaupon
the addition of ibuprofen, as was observed for the binding
of fatty acids.[16]Whilst the possible location of (S)-ibupro-
fen is still unclear, it would be different from that of (R)-
ibuprofen.
Docking Simulations
To understand the formation of the (S)-epoxide in the pres-
ence of (R)-ibuprofen, the binding mode of styrene in the
active site of the (R)-ibuprofen-bound form was simulated
by using Autodock4 (Figure 3).[18]In the simulated structure,
the vinyl group of styrene was close to the heme iron
(4.7 ?). The phenyl ring of styrene interacted with the
phenyl rings of Phe-288. The orientation of the vinyl group
of styrene in the active site agreed with the formation of
(S)-styrene oxide. Because (R)-
styrene oxide was obtained in
the presence of pentanoic acid
(36% ee) as a decoy molecule,
the binding of styrene toward
the pentanoic acid-bound form
of P450SPa was also simulated
for comparison. We assumed
that pentanoic acid would be
accommodated at the active
site of P450SPa with the same
fashion of palmitic acid. The
structure of the
acid-bound form was generated
by simply shortening the alkyl
chain of palmitic acid in the
palmitic-acid-bound
P450SPa. Because
structure of P450SPa with pal-
mitic acid showed two alterna-
tive conformations of palmitic
acid (Conformation A and Con-
formation B), the docking simu-
lations were carried out toward
both conformations. In the si-
mulated structures, the vinyl
groups of styrene-A and sty-
rene-Bwereslightly
from the heme iron atom (Fig-
ure 3b) compared with that in
the active site of the (R)-ibu-
profen-bound form, but the ori-
entation agreed with formation
of (R)-styrene oxide. Because
styreneinteracted
phenyl ring of Phe-288 and the
pentatonic-
form
the
of
crystal
further
with the
side-chain of Gln-84 rather than (R)-ibuprofen and pentano-
ic acid (Figure 4), the orientation of styrene appeared to be
governed by the conformation of Phe-288. This assumption
was supported by the fact that the stereoselectivity was
heavily affected by the mutation at Phe-288: the replace-
ment of Phe-288 by glycine largely enhanced the formation
of (S)-styrene oxide (Table 1 and 2).
Conclusions
We have demonstrated the stereoselective epoxidation of
styrene catalyzed by P450SPawith carboxylic acids as decoy
molecules andconfirmed
system[14–15]was applicable to P450SPa. The crystal structure
of P450SPawith (R)-ibuprofen revealed that the carboxylate
group of (R)-ibuprofen served as an acid–base catalyst to in-
itiate the epoxidation. The stereoselectivity of the styrene-
epoxidation reaction was largely affected by the chirality of
ibuprofen and (R)-ibuprofen enhanced the formation of (S)-
styrene oxide. We believe that our strategy with decoy mole-
thatthis monooxygenation
Figure 2. X-ray crystal structure of the (R)-ibuprofen-bound form of P450SPa(PDB code: 3VM4). Views from
the propionate side of the heme (a) and from the opposite side (b). The distance between the heme iron atom
and the oxygen atoms of the carboxy group, as well as between the a-carbon and the methyl group in the side-
chain of Ala-245, are shown with dotted lines. c) 2Fo-Fcelectron-density map of (R)-ibuprofen, contoured at
the 1.0s level (blue mesh). The hydrophobic amino-acid residues, heme, Arg-241, and (R)-ibuprofen are repre-
sented as stick models. Stereo view of the co-crystal structure with (R)-ibuprofen (pink) superimposed on the
palmitic-acid-bound form of P450SPa(PDB code: 3AWM, light blue).
2290
www.chemasianj.org
? 2012 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim Chem. Asian J. 2012, 7, 2286–2293
FULL PAPER
Yoshihito Watanabe et al.

Page 6

cules can be applied to most of the hydrogen-peroxide-de-
pendent P450s having the substrate-assisted reaction mecha-
nism.[19]Furthermore, the combination of the decoy mole-
cule and a mutant (or other P450s) enabled us to improve
the catalytic activity, as well as the stereoselectivity of the
products.
Experimental Section
Materials
All reagents and solvents were purchased from commercial sources and
used without further purification, except for styrene. n-Butyric acid, hex-
anoic acid, octanoic acid, CH2Cl2, phosphoric acid, glycerol, KCl, HCl,
and di-potassium hydrogen phosphate were purchased from Nacalai
Figure 3. Docking simulations of styrene toward P450SPawith (R)-ibuprofen (a) and pentatonic acid (b); the structure of P450SPawith pentatonic acid
that was used for docking simulations was generated from the crystal structure of P450SPawith palmitic acid (Conformations A and B). Styrene-A
(purple) and styrene-B (green) correspond to the results of docking simulations toward P450SPawith pentanoic acid of Conformation A and Conforma-
tion B, respectively. The location of styrene is shown on the right.
Figure 4. Docking simulations of styrene toward P450SPawith (R)-ibuprofen (a) and pentatonic acid (b, c). Styrene-A (purple) and styrene-B (light
green) correspond to the results of docking simulations toward P450SPawith pentanoic acid of Conformation A and Conformation B, respectively.
Chem. Asian J. 2012, 7, 2286–2293 ? 2012 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
www.chemasianj.org
2291
Chiral-Substrate-Assisted Stereoselective Epoxidation

Page 7

Tesque Inc. (Kyoto, Japan). Heptanoic acid, 1,2-epoxyethylbenzene, (R)-
(+ +)-1,2-epoxyethylbenzene, (p-isopropylphenyl)acetic acid, (S)-(+ +)-nap-
roxen, and hydrogen peroxide were obtained from WAKO Pure Chemi-
cal Industries, Ltd (Osaka, Japan). Styrene, methyl phenylacetate, n-vale-
ric acid, and (S)-(+ +)-ibuprofen were obtained from Tokyo Chemical In-
dustry Co. (Tokyo, Japan). p-Methylphenylacetic acid, (R)-(?)-naproxen,
4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), and (?)-2-
methyl-2,4-pentanediol (MPD) were purchased from Sigma–Aldrich Co.
(USA). (R)-(?)-Ibuprofen was purchased from Funakoshi Co., Ltd.
(Tokyo, Japan). 2-(N-morpholino)ethanesulfonic acid (MES) was pur-
chased from Dojindo Laboratories (Kumamoto, Japan). Phenylacetic
acid was prepared by the hydrolysis of methyl phenylacetate and purified
by re-crystallization.
Measurements
UV/Vis spectra were recorded on a Shimadzu UV-2400 PC spectropho-
tometer. GC analysis was performed on a Shimadzu GC-2014 that was
equipped with a cyclosil-b column (Agilent Technologies, Inc., 30 m?
0.25 mm).
Preparation of WT, L78F, and F288G P450SPa
WT, L78F, and F288G P450SPawere prepared according to literature pro-
cedures.[16]The concentrations of WT P450SPaand the mutants were de-
termined from their CO difference spectra.[20]
Epoxidation of Styrene
Styrene was purified by column chromatography on alumina to remove
the polymerization inhibitor before measurements; styrene was used im-
mediately on purification.[21]The epoxidation reactions were performed
as follows: 4 mm styrene, 4 mm H2O2, and 1 mm WT P450SPawere mixed
in 0.1m potassium phosphate buffer (pH 7.0) at 258 8C. A solution of the
carboxylic acid in EtOH was added as a decoy molecule to a final con-
centration of 20 mm, except for (R)- and (S)-naproxen; because of their
poor solubility, the concentrations of (R)- and (S)-naproxen were 5 mm in
the reaction mixture. The reactions were performed at least three times
with each decoy molecule. CH2Cl2was added immediately into the reac-
tion mixture for quenching and methyl phenylacetate was added as an in-
ternal standard. The extract was evaporated and the resulting solution
was analyzed by GC that was equipped with a cyclosil-b column (Agilent
Technologies, Inc.). The absolute configuration of styrene oxide was de-
termined by using an authentic sample of (R)-styrene oxide. GC analyti-
cal conditions were as follows: injector temperature: 2008 8C, detector
temperature: 2508 8C, initial oven temperature: 908 8C (30 min), ramp rate:
108 8Cmin?1, final oven temperature 1908 8C (20 min), carrier gas: He.
Co-Crystallization of (R)-Ibuprofen-Bound P450SPaWT
P450SPaWT that is expressed in E. coli contains palmitic acid, even after
purification.[16]After the removal of glycerol and KCl from P450SPaby
solvent exchange to 0.1m KPi buffer (pH 7.0), residual fatty acids, includ-
ing palmitic acid, were removed from P450SPaby passing thorough a Hy-
droxyalkoxypropyl-Dextran, Type VI column (Sigma–Aldrich, Co., USA)
that was equilibrated with 0.1m potassium phosphate buffer (pH 7.0) at
RT. The fractions were collected and the solvent was exchanged to
50 mm MES buffer that contained 20% glycerol (pH 7.0). The resulting
P450SPawas used as its substrate-free form. Co-crystallization of sub-
strate-free P450SPaWT and (R)-ibuprofen were performed by a sitting-
drop vapor-diffusion method. The reservoir solution was prepared by
mixing 0.1m HEPES-NaOH (99 mL, pH 7.0), 35% MPD, and 1m (R)-ibu-
profen (1 mL) in EtOH to prepare a total solution volume of 100 mL. A
solution of substrate-free P450SPaWT (2 mL) was mixed with the reser-
voir solution (2 mL). Crystals were grown at 208 8C over 1 week.
Data Collection and Refinement
Crystals were flash-cooled in liquid nitrogen. X-ray-diffraction data sets
were collected on beam-line BL26B1 that was equipped with an ADSC
Quantum 315 CCD detector at the RIKEN SPring-8 (Hyogo, Japan) with
a 1.0 ? wavelength at 100 K. The program HKL2000[22]was used to inte-
grate the diffraction intensities and scaling. The structure of (R)-ibupro-
fen-bound P450SPawas solved by molecular replacement with MolRep.[23]
Model building and refinement were performed by using COOT[24]and
REFMAC5.[25]
The (R)-ibuprofen model was generated by using
a Dundee PRODRG server[26]and used in the refinement with COOT
and REFMAC5. Alternative conformations were introduced to Glu-89
with occupancies of 0.6 and 0.4, Arg-187, Val-313, and Ser-344 with occu-
pancies of 0.7 and 0.3, and to Glu-64, Val-292, and His-412 with occupan-
cies of 0.5 and 0.5. TLS refinement[27]was performed at the final stage of
the refinement by defining each chain in the asymmetric unit as a sepa-
rate TLS group. The produced model showed a final Rfact=15.8% and
Rfree=18.6% (Table 3). The final model consisted of one polypeptide
chain with residues 9–415 of P450SPa, 1 heme, 1 (R)-ibuprofen, 2 (R)-2-
methyl-2,4-pentanediol (MRD), and 274 water molecules. Structure vali-
dation was performed by using PROCHECK.[28]All protein figures were
depicted by using PyMOL.[29]
Docking Simulations
The docking experiments were performed by using AutoDock4[18]and
AutoGrid4 in combination with AutoDock Tools according to literature
procedures.[30]The styrene model was generated by using Dundee Pro-
drug Server and used as a ligand.[26]The X-ray crystal structure of (R)-
ibuprofen-bound P450SPaWT was used as a rigid receptor for the docking
of styrene. Met-69, Leu-77, Leu-78, and Phe-288, which were located
around the hydrophobic and hydrophilic channels, were set as flexible
residues for the access of styrene into the active site from the protein sur-
face. For the docking of styrene to the pentanoic-acid-bound-form of
P450SPa, the model was generated by shortening the alkyl-chain length of
palmitic acid in the crystal structure of palmitic-acid-bound WT P450SPa.
The model was used as a rigid receptor and Met-69, Leu-77, Leu-78, and
Phe-288 were set as flexible residues. Docked conformations were
ranked automatically by Autodock4 as a free-energy-scoring function.
These results were visualized by using PyMOL.[29]
Table 3. Data collection and refinement of (R)-ibuprofen-bound WT
P450SPa.
Data collection
l [?]
space group
cell dimensions
a, b, c [?]
a, b, g [8 8]
resolution [?]
no. of total observed reflections
no. of unique reflections
Rmerge
completeness[a][%]
I/s(I)[a]
redundancy[a]
1000
P3121
94.390, 94.390, 112.863
90.000, 90.000, 120.000
20.0–1.94 (2.01-1.94)
452029
43560
4.8 (29.7)
100 (100)
53.1 (8.2)
10.4 (10.7)
[a,b][%]
Refinement statistics
resolution range [?]
no. of monomer/asymmetric unit
Rfact/Rfree
r.m.s.d. bond length[e][?]
r.m.s.d. bond angles[e][8 8]
no. of atoms
average B-factor [?2]
19.76–1.94
1
15.8/18.6
0.012
1.193
3614
19.5
[c,d][%]
[a] The values in parentheses are for the highest-resolution shell.
[b] Rmerge=SjI?<I> j/SI. [c] Rfact=Sj jFoj?kjFcj j/SjFoj where Foand
Fcare the observed and calculated structure-factor amplitudes, respec-
tively. [d] Rfreewas calculated as the Rfactfor 5% of the reflection that
were not included in the refinement. [e] r.m.s.d.=root mean square devi-
ation.
2292
www.chemasianj.org
? 2012 Wiley-VCH Verlag GmbH&Co. KGaA, WeinheimChem. Asian J. 2012, 7, 2286–2293
FULL PAPER
Yoshihito Watanabe et al.

Page 8

Acknowledgements
This work was supported by a Grant-in-Aid for Scientific Research (S) to
Y.W. (19105044), a Grant-in-Aid for Scientific Research on Innovative
Areas to Y.S. (22105012), and a Grant-in-Aid for Young Scientists (A) to
O.S. (21685018) from the Ministry of Education, Culture, Sports, Science
and Technology (Japan). T.F. was supported by the JSPS Research Fel-
lowships for Young Scientists. We thank Dr. Go Ueno, Dr. Hironori Mur-
akami, Dr. Masatomo Makino, and Dr. Nobuyuki Shimizu for their assis-
tance with the data collection at SPring-8. We thank Dr. Isamu Matsuna-
ga for his kind gift of the expression system of P450SPa.
[1] a) Z. Li, J. B. van Beilen, W. A. Duetz, A. Schmid, A. de Raadt, H.
Griengl, B. Witholt, Curr. Opin. Chem. Biol. 2002, 6, 136–144; b) F.
van Rantwijk, R. A. Sheldon, Curr. Opin. Biotechnol. 2000, 11, 554–
564.
[2] M. Sono, M. P. Roach, E. D. Coulter, J. H. Dawson, Chem. Rev.
1996, 96, 2841–2888.
[3] a) F. P. Guengerich, Nat. Rev. Drug Discovery 2002, 1, 359–366;
b) B. Rita, J. Biotechnol. 2006, 124, 128–145; c) V. B. Urlacher, S.
Eiben, Trends Biotechnol. 2006, 24, 324–330; d) P. C. Cirino, F. H.
Arnold, Curr. Opin. Chem. Biol. 2002, 6, 130–135; e) V. B. Urlacher,
S. Lutz-Wahl, R. D. Schmid, Appl. Microbiol. Biotechnol. 2004, 64,
317–325; f) N. Kawakami, O. Shoji, Y. Watanabe, Angew. Chem.
2011, 123, 5427–5430; Angew. Chem. Int. Ed. 2011, 50, 5315–5318.
[4] J. B. van Beilen, W. A. Duetz, A. Schmid, B. Witholt, Trends Bio-
technol. 2003, 21, 170–177.
[5] Q.-S. Li, J. Ogawa, S. Shimizu, Biochem. Biophys. Res. Commun.
2001, 280, 1258–1261.
[6] a) L. O. Narhi, A. J. Fulco, J. Biol. Chem. 1986, 261, 7160–7169;
b) S. S. Boddupalli, R. W. Estabrook, J. A. Peterson, J. Biol. Chem.
1990, 265, 4233–4239; c) H. Li, T. L. Poulos, Nat. Struct. Biol. 1997,
4, 140–146.
[7] P. C. Cirino, F. H. Arnold, Angew. Chem. 2003, 115, 3421–3423;
Angew. Chem. Int. Ed. 2003, 42, 3299–3301.
[8] I. Matsunaga, A. Ueda, N. Fujiwara, T. Sumimoto, K. Ichihara,
Lipids 1999, 34, 841–846.
[9] I. Matsunaga, T. Sumimoto, A. Ueda, E. Kusunose, K. Ichihara,
Lipids 2000, 35, 365–371.
[10] M. Girhard, S. Schuster, M. Dietrich, P. D?rre, V. B. Urlacher, Bio-
chem. Biophys. Res. Commun. 2007, 362, 114–119.
[11] D.-S. Lee, A. Yamada, H. Sugimoto, I. Matsunaga, H. Ogura, K.
Ichihara, S.-i. Adachi, S.-Y. Park, Y. Shiro, J. Biol. Chem. 2003, 278,
9761–9767.
[12] M. Sundaramoorthy, J. Terner, T. L. Poulos, Structure 1995, 3, 1367–
1378.
[13] O. Shoji, T. Fujishiro, S. Nagano, S. Tanaka, T. Hirose, Y. Shiro, Y.
Watanabe, J. Biol. Inorg. Chem. 2010, 15, 1331–1339.
[14] O. Shoji, T. Fujishiro, H. Nakajima, M. Kim, S. Nagano, Y. Shiro, Y.
Watanabe, Angew. Chem. 2007, 119, 3730–3733; Angew. Chem. Int.
Ed. 2007, 46, 3656–3659.
[15] a) T. Fujishiro, O. Shoji, Y. Watanabe, Tetrahedron Lett. 2011, 52,
395–397; b) O. Shoji, Y. Watanabe, Metallomics 2011, 3, 379–388.
[16] T. Fujishiro, O. Shoji, S. Nagano, H. Sugimoto, Y. Shiro, Y. Wata-
nabe, J. Biol. Chem. 2011, 286, 29941–29950.
[17] A. Zaks, D. R. Dodds, J. Am. Chem. Soc. 1995, 117, 10419–10424.
[18] G. M. Morris, D. S. Goodsell, R. S. Halliday, R. Huey, W. E. Hart,
R. K. Belew, A. J. Olson, J. Comput. Chem. 1998, 19, 1639–1662.
[19] T. Furuya, K. Kino, Appl. Microbiol. Biotechnol. 2010, 86, 991–
1002.
[20] T. Omura, R. Sato, J. Biol. Chem. 1964, 239, 2370–2378.
[21] W. L. F. Armarego, D. D. Perrin, Purification of Laboratory Chemi-
cals, 4th ed., Butterworth-Heinemann, Oxford, 1996.
[22] Z. Otwinowski, D. Borek, W. Majewski, W. Minor, Acta Crystallogr.
Sect. A 2003, 59, 228–234.
[23] A. Vagin, A. Teplyakov, J. Appl. Crystallogr. 1997, 30, 1022–1025.
[24] P. Emsley, K. Cowtan, Acta Crystallogr. Sect. D 2004, 60, 2126–2132.
[25] M. D. Winn, G. N. Murshudov, M. Z. Papiz, Methods Enzymol. 2003,
374, 300–321.
[26] A. W. Sch?ttelkopf, D. M. van Aalten, Acta Crystallogr. Sect. D
2004, 60, 1355–1363.
[27] M. D. Winn, M. N. Isupov, G. N. Murshudov, Acta Crystallogr. Sect.
D 2001, 57, 122–133.
[28] R. A. Laskowski, M. W. Macarthur, D. S. Moss, J. M. Thornton, J.
Appl. Crystallogr. 1993, 26, 283–291.
[29] W. L. DeLano, The PyMOL Molecular Graphics System (2002) on
World Wide Web http://www.pymol.org. 2002.
[30] M. F. Sanner, J. Mol. Graphics Modell. 1999, 17, 57–61.
Received: March 22, 2012
Published online: June 14, 2012
Chem. Asian J. 2012, 7, 2286–2293? 2012 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
www.chemasianj.org
2293
Chiral-Substrate-Assisted Stereoselective Epoxidation