Regio- and stereoselective hydroxylation of optically active α-ionone enantiomers by engineered cytochrome P450 BM3 mutants

Article abstract of DOI:10.1002/adsc.201200067

The selective hydroxylation of an unactivated C-H bond is a crucial step in the synthesis of fine chemicals such as hydroxylated terpenoids. In the present study, the ability of 40 cytochrome P450 BM3 mutants to perform the regio- and stereoselective hydroxylation of α-ionone has been investigated. Based on their activity and selectivity to produce 3-hydroxy-α-ionone from racemic α-ionone, 6 BM3 mutants were selected. Out of these, 3 mutants (M01 A82W, M11 A82W and M11 V87I) showed high selectivity for trans-3-hydroxy-α- ionone formation while 3 other mutants (M11 L437N, M11 L437S and M11 L437T) formed almost equal amounts of both cis-3-hydroxy- and trans-3-hydroxy-α- ionone. Incubation with individual enantiomers showed that M11 L437N, M11 L437S and M11 L437T exhibited opposite stereoselectivity producing (3S,6S)-hydroxy-α-ionone with the (6S)-enantiomer and (3S,6R)-hydroxy- α-ionone with the (6R)-enantiomer. Thus for the first time, BM3 mutants that can selectively produce diastereomers of 3-hydroxy-α-ionone (>90% de), with high turnover numbers and minimal secondary metabolism, have been identified. Docking studies have been performed to rationalize the basis of the experimentally observed selectivity. In conclusion, engineered P450 BM3s are promising biocatalysts for regio- and stereoselective production of hydroxylated α-ionones for industrial applications. Copyright

Full text of DOI:10.1002/adsc.201200067

FULL PAPERS
DOI: 10.1002/adsc.201200067
Regio- and Stereoselective Hydroxylation of Optically Active
a-Ionone Enantiomers by Engineered Cytochrome P450 BM3
Mutants
Harini Venkataraman,a Stephanie B. A. de Beer,a Daan P. Geerke,a
Nico P. E. Vermeulen,a and Jan N. M. Commandeura,*
a
Division of Molecular and Computational Toxicology, LACDR, Leiden/Amsterdam Center for Drug Research, Faculty of
Sciences, Vrije Universiteit, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands
Fax : (+31)-20-598-7610; e-mail: j.n.m.commandeur@vu.nl
Received: January 25, 2012; Revised: May 9, 2012; Published online: July 29, 2012
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201200067.
Abstract: The selective hydroxylation of an unacti- opposite stereoselectivity producing (3S,6S)-hydroxy-
À
vated C H bond is a crucial step in the synthesis of a-ionone with the (6S)-enantiomer and (3S,6R)-hy-
fine chemicals such as hydroxylated terpenoids. In droxy-a-ionone with the (6R)-enantiomer. Thus for
the present study, the ability of 40 cytochrome P450 the first time, BM3 mutants that can selectively pro-
BM3 mutants to perform the regio- and stereoselec- duce diastereomers of 3-hydroxy-a-ionone (>90%
tive hydroxylation of a-ionone has been investigated. de), with high turnover numbers and minimal secon-
Based on their activity and selectivity to produce 3- dary metabolism, have been identified. Docking
hydroxy-a-ionone from racemic a-ionone, 6 BM3 studies have been performed to rationalize the basis
mutants were selected. Out of these, 3 mutants of the experimentally observed selectivity. In conclu-
(M01A82W, M11A82W and M11V87I) showed high sion, engineered P450 BM3s are promising biocata-
selectivity for trans-3-hydroxy-a-ionone formation lysts for regio- and stereoselective production of hy-
while 3 other mutants (M11L437N, M11L437S and droxylated a-ionones for industrial applications.
M11L437T) formed almost equal amounts of both
cis-3-hydroxy- and trans-3-hydroxy-a-ionone. Incuba-
tion with individual enantiomers showed that Keywords: biocatalysis; cytochrome P450 BM3; dia-
M11L437N, M11L437S and M11L437T exhibited stereoselectivity; hydroxylation; a-ionone
Introduction
products and the lack of selectivity of conventional
synthetic methods.[2,6] In particular, the stereoselective
3
À
Ionones are a class of nor-isoprenoids which are hydroxylation of unactivated sp C H bonds is a chal-
widely used in the flavor and fragrance industry,[1] and lenging aspect in modern day chemistry.[7] Because
as building blocks for many chemicals.[2] a-Ionone is enzymes can catalyze reactions with remarkable
one of the nor-isoprenoids which possesses character- regio- and stereoselectivity under mild conditions,
istic organoleptic properties. Interestingly, the individ- their use as biocatalysts for the selective production
ual enantiomers of a-ionone possess different odors,[3] of fine chemicals has gained increased interest over
thereby making them important for the fragrance in- the years.[8] A variety of mono-oxygenases has been
dustry. The hydroxylated variants of a-ionone are also used for selective oxyfunctionalization of organic sub-
of commercial interest. For example, derivatives of 3- strates.[9] An important class of mono-oxygenases are
hydroxy-a-ionone are used as an intermediate for the the cytochrome P450s, which are heme-dependent en-
total synthesis of lutein and its stereoisomers.[4] Also, zymes with the ability to perform regio- and stereose-
3-hydroxy-a-ionone has been shown to be a potent at- lective hydroxylations of a wide range of chemicals.[10]
tractant for B. latifrons during pollination.[5]
Efforts over the last decade have been focussed on
The synthesis of hydroxylated variants of natural engineering P450s to obtain robust and useful biocat-
products such as terpenes and terpenoids still poses alysts for industrial applications.[11] Protein engineer-
a challenge due to the formation of a variety of side ing of P450s has proved to be an excellent means to
2172
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2172 – 2184

Regio- and Stereoselective Hydroxylation of Optically Active a-Ionone Enantiomers
broaden their substrate scope and specificity as well
The microbial biotransformation of racemic a-
as to improve activity.[12] For example, engineered var- ionone using different Streptomyces strains,[19] fungi
iants of cytochrome P450cam have been shown to ex- such as Aspergillus niger,[20] immobilized cells of Nico-
hibit increased regio- and stereoselectivity towards tium tabacum,[21] and cultured cells of Caragana cham-
terpenes like pinene and limonene.[13]
lagu[22] has been reported. Moreover, different bacte-
Cytochrome P450 CYP102A1 from Bacillus mega- rial P450s from Bacillus subtillus, Sorangium cellulo-
terium, commonly referred to as P450 BM3, has good sum, Novaspinghobium aromaticivorans and Strepto-
perspectives for the use as biocatalyst for hydroxyl- myces have been shown to hydroxylate a-ionone.[23–27]
ation reactions of industrial significance because it is In most of the biotransformation studies of racemic
the most active P450 so far identified.[14] The fact that a-ionone that have been reported so far, trans-3-hy-
the heme and reductase domains of P450 BM3 are droxy-a-ionone was found to be the main prod-
fused into a single polypeptide makes the electron uct.[19,23,25] Other studies reported the formation of 2
transfer very efficient.[15] The active site of P450 BM3 diastereomers of 3-hydroxy-a-ionone from racemic a-
has been re-engineered to accept a broader variety of ionone hydroxylation.[17,27] However, from these stud-
substrates including terpenoids such as ionones.[16] Al- ies it is not clear whether these enzymes are able to
though wild-type P450 BM3 exhibits a low activity to- catalyze both cis- and trans-hydroxylation on each in-
wards a-ionone hydroxylation, some cytochrome dividual enantiomer, or whether they hydroxylate
P450 BM3 mutants engineered by rational design each enantiomer with a different diastereoselectivity.
showed an increased activity. A triple mutant contain-
In the present study a library of P450 BM3 mutants
ing A74G F87V L188Q mutations, was shown to con- was evaluated for their ability to selectively produce
vert racemic a-ionone to 3-hydroxy-a-ionone (49%) the individual diastereomers of 3-hydroxy-a-ionone.
in addition to a number of other products.[17] Also, This library was generated by a combination of site-
BM3 mutant F87V produced a mixture of 4 oxidized directed mutagenesis of active site residues and
products with a-ionone without any selectivity.[18]
random mutagenesis by error-prone PCR as described
Hydroxylation of racemic a-ionone at the C-3 posi- previously.[28,31,32] Thirty one mutants of the library
tion can result in the formation of four different dia- were selected based on their proven catalytic diversity
stereomers of 3-hydroxy-a-ionone. Individual enantio- towards a variety of drugs and steroids.[28–31] Nine ad-
mers (6R)- or (6S)-a-ionone can be hydroxylated to ditional mutants were prepared by site-directed muta-
(3R,6R) and (3S,6R) or (3R,6S) and (3S,6S) forms, re- genesis of active site residues A82, V87 and L437
spectively, depending on the selectivity of the reac- using M01 or M11 as template, see the Supporting In-
tion. Structures of the 4 possible hydroxylated diaste- formation, Table S1. A detailed list of the combina-
reomers are shown in Figure 1.
tion of mutations present in the 40 mutants used in
the present study is provided in the Supporting Infor-
mation, Table S2. Because some of our P450 BM3
mutants showed both cis- and trans-hydroxy product
formation with racemic a-ionone as substrate, it was
decided to investigate the hydroxylation of each enan-
tiomer individually.
In this paper, it is shown for the first time that
BM3 mutants can hydroxylate individual enantiomers
of a-ionone in a highly regio- and stereoselective
manner and with high turnover numbers for 3 out of
the 4 possible diastereomers. Thus the introduction of
a hydroxy group in a single step by engineered BM3
in a highly regio- and stereoselective manner gives
them good perspectives for industrial applications.
Results and Discussion
Screening of the P450 BM3 Library with Racemic a-
Ionone
Figure 1. Structures of a-ionone enantiomers and their 3’ hy-
droxylated products. A: (6R)-a-ionone; B: (6S)-a-ionone;
C: (3S,6R)-hydroxy-a-ionone; D: (3R,6S)-hydroxy-a-
ionone; E: (3R,6R)-hydroxy-a-ionone; F: (3S,6S)-hydroxy-
a-ionone. C and D are cis isomers while E and F are trans
isomers (see the Supporting Information, Figure S1 for GC-
MS fragmentation).
The goal of the initial screening was to identify P450
BM3 mutants that can produce 3-hydroxy-a-ionone
with high selectivity. 40 cytochrome P450 BM3 mu-
tants that were tested converted racemic a-ionone to
Adv. Synth. Catal. 2012, 354, 2172 – 2184
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2173

Harini Venkataraman et al.
FULL PAPERS
Figure 3. Mutants M01A82W, M11A82W, M11V87I
converted racemic a-ionone mainly to trans-3-hy-
droxy-a-ionone (>85% de), whereas mutants
M11L437N, M11L437S and M11L437T produced
almost equal amounts of trans-3-hydroxy-a-ionone
and cis-3-hydroxy-a-ionone. Consistent with a previ-
ous study,[17] wild-type P450 BM3 showed only very
low activity, with a preference for trans-hydroxylation.
Conversion to 3-Oxo-a-ionone
Some of the mutants (M11A74E, M11L437T) con-
verted racemic a-ionone to 3-oxo-a-ionone and an
unidentified over-oxidation product (see the Support-
ing Information, Figure S4), in addition to trans-3-hy-
droxy- and cis-3-hydroxy-a-ionone. A representative
gas chromatogram illustrating this product profile is
shown in Figure 2, C. Incubation of purified P450
BM3 mutants with a-ionone also resulted in the for-
mation of 3-oxo-a-ionone indicating that the secon-
dary oxidation of alcohol to ketone is also mediated
by P450 BM3. 3-Oxo-a-ionone has been recently
shown to be a phagostimulant and also has commer-
cial value because of its fragrance.[34] The exact mech-
anism of formation of 3-oxo-a-ionone is still to be in-
vestigated, but the formation of ketones by engi-
neered P450 BM3 has been reported before for
octane hydroxylation.[35] A possible mechanism as
mentioned in these studies is the formation of a gemi-
nal diol that readily dehydrates to form the corre-
sponding ketone group. Also, in the case of 7-hy-
droxy-delta8-tetrahydrocannabinol (THC) hydroxyl-
ation by hepatic microsomes from Japanese monkeys,
the formation of 7-oxo-delta8-THC by a similar
mechanism has been reported.[36]
Figure 2. Gas chromatograms of biotransformation products
of racemic a-ionone by (A) M01A82W; (B) M11L437N;
(C) M11A74E. Identification of the peaks: (1) cis-3-hy-
droxy-a-ionone tR =9.84; (2) trans-3-hydroxy-a-ionone tR =
9.92; (3) 3-oxo-a-ionone tR =10.02; (4) unknown product
tR =10.84.
trans-3-hydroxy- and cis-3-hydroxy-a-ionone with dif- Incubation with Individual Enantiomers of a-Ionone
ferent selectivities and activities. As shown in
Figure 2, three distinct product profiles were ob- As seen in Figure 3, from the screening of the mu-
served: mutants that converted racemic a-ionone pre- tants with racemic a-ionone, mutants M11L437N,
dominantly to trans-3-hydroxy-a-ionone (Figure 2, M11L437S and M11L437T initially appeared to be
A), mutants which produced almost equal amounts of not stereoselective by producing almost equal
both cis- and trans-3-hydroxy-a-ionone (Figure 2, B) amounts of trans- and cis-3-hydroxy-a-ionone. In
and mutants with produced significant amounts of order to elucidate which diastereomers are formed
over-oxidation products (Figure 2, C). The retention from each a-ionone enantiomer, these three mutants
times (tR) of cis-3-hydroxy-a-ionone and trans-3-hy- were also incubated with the individual enantiomers
droxy-a-ionone were 9.84 and 9.92 min, respectively, (6R)- and (6S)-a-ionone. All three mutants turned
based on the tR of the available reference compounds. out to be highly stereoselective with individual enan-
The product profile of all the mutants tested with tiomers as substrate, producing (3S,6S)-hydroxy-a-
racemic a-ionone is shown in Figure 3. In general, ionone with the (6S)-enantiomer and (3S,6R)-hy-
most of the mutants were capable of highly regiose- droxy-a-ionone with the (6R)-enantiomer. Figure 4
lective C-3 hydroxylation of racemic a-ionone. C-3 is (A, B) shows the results obtained with mutant
energetically the most favorable site for hydroxylation M11L437N. Incubation of M11L437T with (6R)-a-
because it is an allylic methylene group. P450 BM3 ionone at a concentration of 500 mM resulted in the
mutants exhibiting high selectivity are highlighted in formation of about 80% of (3S,6R)-hydroxy-a-ionone,
2174
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2172 – 2184

Regio- and Stereoselective Hydroxylation of Optically Active a-Ionone Enantiomers
Figure 3. Product profile and conversion of racemic a-ionone by wild-type P450 BM3 and engineered variants. Conversions
in % were measured from substrate consumption after 1 hour incubation of 500 mM of racemic a-ione with 200 nM of P450
BM3. Values represent averages of 2 independent experiments with less than 10% error. Product selectivity in % was calcu-
lated from gas chromatograms by integrating the product peaks. Representative P450 BM3 variants showing high selectivity
towards formation of trans-3-a-ionone (M11V87I, M11A82W, M01A82W), variants that were not selective (M11L437N,
M11L437S, M11L437T, M11L437N V87F) and a variant showing high over-oxidation product formation (M11A74E) are
highlighted. Wild-type P450 BM3 and mutants M01 and M11 are also highlighted for comparison. Of the 40 mutants tested,
only one variant showed <0.5% conversion (M11V87C) and was therefore not included in Figure 3. An expanded form of
Figure 3 is included in the Supporting Information (Figure S3) with all the variants labelled.
12% of (3R,6R)-hydroxy-a-ionone and 8% of 3-oxo- L437N, and M11A82W and M11A82W L437N. As
a-ionone as measured after 1 hour. After 12 h of incu- shown in the Supporting Information, Figure S9A,
bation, a-ionone was completely converted to 3-oxo- substitution of L437N also increased the relative
a-ionone (see the Supporting Information, Fig- amount of (3S,6R)-hydroxy-a-ionone when applied to
ure S5). In contrast, 12-hour incubations of M11V87F and M11A82W although to a much lower
M11L437N with a-ionone enantiomer showed only extent when compared to that when applied to M11.
small amounts of 3-oxo-ionone. This shows that BM3 When incubated with (6S)-a-ionone, the L437N sub-
mutant M11L437N is capable of highly selective hy- stitution led to a slight increase in formation of
droxylation with minimal formation of secondary oxi- (3R,6S)-hydroxy-a-ionone when applied to M11V87F,
dation products whereas mutant M11L437T can whereas the substitution had no effect when applied
result in the formation of a significant amount of 3- to M11A82W, see the Supporting Information, Figur-
oxo-a-ionone.
e S9B. These results confirm that the effect of
The fact that incubations with mutant M11 showed a single amino acid substitution is strongly dependent
also significant amounts (30%) of (3R,6R)-hydroxy-a- on additional nearby mutations.[40, 60]
ionone with the (6R)-enantiomer as substrate is
As shown in Figure 3, M01A82W metabolized rac-
worthy of note, Table 1, shows that substitution emic a-ionone with high activity, producing mainly
L437N in mutant M11 has improved the stereoselec- trans-3-hydroxy-a-ionone, implicating that both enan-
tivity of the (6R)-a-ionone hydroxylation. Interesting- tiomers of a-ionone were stereoselectively hydroxy-
ly, L437 has been recently reported to be a ꢁhotspotꢂ lated. This was confirmed by incubation of
affecting the selectivity of terpene hydroxylation.[40] M01A82W with the individual enantiomers. As
To test whether the L437N substitution also influen- shown in Figure 4 (C, D), (6R)-a-ionone was specifi-
ces stereoselectivity in other M11-related mutants, the cally metabolized to (3R,6R)-hydroxy-a-ionone,
enantiomers of a-ionone were also incubated with the whereas (6S)-a-ionone was specifically metabolized
available combinations M11V87F and M11V87F to (3S,6S)-hydroxy-a-ionone. In both cases only mini-
Adv. Synth. Catal. 2012, 354, 2172 – 2184
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2175

Harini Venkataraman et al.
FULL PAPERS
Figure 4. Representative gas chromatograms showing opposite stereoselectivity for (A) hydroxylation of (6R)-a-ionone by
M11L437N and (B) hydroxylation of (6S)-a-ionone by M11L437N and similar stereoselectivity for (C) hydroxylation of
(6R)-a-ionone by M01A82W and (D) hydroxylation of (6S)-a-ionone by M01A82W. In all cases, the substrate (6R)- or
(6S)-a-ionone was completely oxidized to respective products.
Table 1. Stereoselective hydroxylation of (R)-a-ionone and (S)-a-ionone by wild-type P450 BM3 engineered P450 BM3 mu-
tants.
Mutant
Substrate
Product Distribution [%][a]
Diastereomeric Total turn-
excess [de %]
over number[b]
Wild-type P450
BM3
(R)-a-
ionone
3R,6R (53%), 3S,6R (17%), other-oxidation*
(30%)
51 (3R,6R)
–
(S)-a-ionone 3S,6S (46%), 3R,6S (36%), other-oxidation* (18%) 12 (3S,6S)
–
–
M01
(R)-a-
3R,6R (59%), 3S,6R (30%), 3-oxo-a-ionone (11%) 32 (3R,6R)
ionone
(S)-a-ionone 3S,6S (84%), 3R,6S (14%), 3-oxo-a-ionone (2%)
71 (3S,6S)
–
–
M11
(R)-a-
3R,6R (35%), 3S,6R (55%), 3-oxo-a-ionone (10%) 22 (3S,6R)
ionone
(S)-a-ionone 3S,6S (83%), 3R,6S (7%), 3-oxo-a-ionone (10%)
84 (3S,6S)
91 (3R,6R)
–
M01A82W
M11 L437N
(R)-a-
3R,6R (91%), 3S,6R (5%), 3-oxo-a-ionone (4%)
3500
ionone
(S)-a-ionone 3S,6S (95%), 3R,6S (3%), 3-oxo-a-ionone (2%)
(R)-a-
95 (3S,6S)
90 (3S,6R)
4000
3000
3R,6R (4%), 3S,6R (90%), 3-oxo-a-ionone (6%)
ionone
(S)-a-ionone 3S,6S (91%), 3R,6S (5%), 3-oxo-a-ionone (4%)
91 (3S,6S)
3800
[a]
Values (in %) were calculated from 2 independent experiments with not more than 10% error. For wild-type P450 BM3
the activity towards both enantiomers was too low to accurately determine the product distribution under the conditions
used for other P450 BM3 mutants. Therefore, the incubations were performed at a higher enzyme concentration of 1 mM.
Other-oxidation refers to the peak with m/z=208 as detected on GC-MS.
Total turnover numbers are reported as nmol product per nmol enzyme. The values are calculated from 2 independent
experiments with Æ10% error. Total turnover numbers were measured only for the enzymes which showed high diaste-
reoselectivity towards product formation.
[b]
2176
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2172 – 2184

Regio- and Stereoselective Hydroxylation of Optically Active a-Ionone Enantiomers
mal amounts of secondary oxidation products were sired product with minimal formation of side prod-
observed. When (6R)-a-ionone was incubated with ucts. Purified M01A82W was used as biocatalyst with
purified M01, significant amounts of both diastereo- (S)-a-ionone (48 mg) as substrate in a volume of
mers were found, Table 1, indicating that mutation 250 mL to illustrate the preparative use of this engi-
A82W has strongly increased stereoselectivity of hy- neered P450 BM3. The reaction resulted in >90%
droxylation of a-ionone.
substrate conversion to produce (3S,6S)-hydroxy-a-
Wild-type P450 BM3, which showed only very low ionone (95% de) under the experimental conditions
activity when incubated with racemic a-ionone, mentioned. Therefore, apparently the biotransforma-
showed very low diastereoselectivity, when incubated tion reaction can be scaled up without any loss of dia-
with the enantiomers of a-ionone, Table 1. When in- stereoselectivity.
cubated with (S)-a-ionone, almost equal amounts of
(3S,6S)- and (3R,6S)-diastereomers were found, next
to significant amounts of unidentified oxidation prod- Binding Spectra of a-Ionone Enantiomers
ucts. When incubated with (R)-a-ionone, the diaste-
reomeric excess for the (3R,6R)-diastereomer was The interactions of the a-ionone enantiomers with the
found to be 51%.
stereoselective mutants M01A82W and M11L437N
and their respective templates M01 and M11 were
studied by spectral binding experiments. Titration of
(R)- or (S)-a-ionone to solutions of purified P450
BM3 mutants in all cases resulted in typical type I dif-
ference spectra, with a peak at 390 nm and a trough
Total Turnover Numbers for Stereoselective
Hydroxylation
An optimal biocatalyst should ideally be able to cata- at 422 nm, indicative of the conversion of heme iron
lyze the formation of the desired product in a highly from low spin to high spin. The Supporting Informa-
selective manner (regio-, enantio- or stereo-) in com- tion, Figure S6 shows the binding spectra of both
bination with high catalytic turnovers and with mini- enantiomers to mutant M01A82W. The dissociation
mal amount of secondary products.[37] Among constants obtained from these titration studies are
M11L437N, M11L437S and M11L437T, M11L437N, shown in Table 2. The introduction of the A82W mu-
which showed the highest diastereoselectivity (> tation in M01 increased binding affinity towards both
90%), see Figure 4, was selected for determination of enantiomers: the dissociation constant (Kd) for
the total turnover number. With respect to hydroxyl- M01A82W was decreased approximately 3.5-fold in
ation of optically active a-ionone by P450s, mutant the case of (R)-enantiomer and approximately 22-fold
M11L437N is the first engineered P450 BM3 specifi- in the case of (S)-enantiomer when compared to M01.
cally producing one of the cis-3-hydroxy products, Improved affinity of hydrophobic substrates by A82W
that is, (3S,6R), with high stereoselectivity. Addition- mutation in P450 BM3 has been reported previously
ally, among mutants that produced predominantly with other substrates like fatty acids and testoster-
trans-3-hydroxy-a-ionone, M01A82W was selected one.[29,39]
for incubation with individual enantiomers because it
Mutation L437N in M11 also apparently increases
exhibited the highest diastereoselectivity as well as ac- affinity towards both enantiomers of a-ionone; disso-
tivity. M01A82W catalyzed the formation of (3S,6S)- ciation constants were lowered 2.4–3.6-fold in com-
hydroxy-a-ionone and (3R,6R)-a-ionone with (6S)- parison to its precursor M11, which exhibited Kd
and (6R)-enantiomers respectively, with remarkable values of 17 mM and 27 mM for (R)-enantiomer and
selectivity.
(S)-enantiomer respectively.
The fact that the P450 BM3 mutants M01A82W
Wild-type P450 BM3 also showed a typical type I
and M11L437N exhibit different stereoselectivities to- difference spectrum. Dissociation constants obtained
wards hydroxylation of (R)-and (S)-enantiomers, were comparable to those obtained with the much
makes them highly useful for industrial biotechnologi- more active P450 BM3 mutants, being 17.5 and
cal applications. M01A82W catalyzes the formation 10.3 mM, respectively, for (R)-a-ionone and (S)-a-
of (3S,6S)-hydroxy-a-ionone and (3R,6R)-hydroxy-a- ionone.
ionone with turnover numbers in the range of 3500–
4000, Table 1. Similarly M11L437N can also catalyze
the formation of (3S,6R)-hydroxy-a-ionone with high NADPH Consumption and Product Formation
turnover numbers of about 3000 nmol product/nmol
enzyme. These high turnover numbers indicate that To study the coupling efficiency of the mutants and
these enzymes might be useful for preparative scale wild-type P450 BM3, the rates of NADPH oxidation
biosynthesis.[38]
and product formation were determined at 200 mM of
For both M01A82W and M11L437N these reac- substrate, Table 2. In case of wild-type BM3, relative-
tions are carried out with high selectivity to the de- ly low (4–6%) coupling efficiencies were observed for
Adv. Synth. Catal. 2012, 354, 2172 – 2184
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2177

Harini Venkataraman et al.
FULL PAPERS
both enantiomers: the NADPH oxidation rates upon
addition of (R)- or (S)-a-ionone were found to be ap-
proximately 30 minÀ1 whereas the product formation
was only about 1.5 minÀ1. Wild-type P450 BM3 pro-
duced a mixture of 2 diastereomers of C-3 hydroxyl-
ation and an additional peak with an m/z of 208 indi-
cating hydroxylation at another carbon with both the
enantiomers. Interestingly, no over-oxidation peaks
were observed with both enantiomers. (Table 1)
In case of the P450 BM3 mutants significantly
higher coupling efficiencies, up to 45%, were ob-
served when compared to wild-type P450 BM3. Intro-
duction of mutation A82W in M01 did not significant-
ly affect the product formation with (R)-a-ionone as
substrate, but decreased NADPH oxidation 3-fold.
The increased affinity towards (R)-a-ionone due to
the A82W mutation in combination with a more pro-
ductive binding mode therefore may explain the in-
creased coupling efficiency.
For M11 the introduction of L437N did not increase
the coupling efficiency of C-3 hydroxylation of both
enantiomers, Table 2. However, the rate of both
NAPDH oxidation and product formation was in-
creased almost 3- to 4-fold. The higher affinity to
M11L437N variant and the higher stereoselectivity
when compared to M11 are indicative to a different,
more productive, binding mode due to the L437N mu-
tation.
For (S)-a-ionone, the A82W mutation in M01 led
to approximately 1.7-fold increase in rate of NADPH
oxidation and 2.7-fold increase in the rate of product
formation, resulting in an increased coupling efficien-
cy of 44%. For the (S)-enantiomer all four variants
formed predominantly trans-hydroxy-a-ionone with
M01A82W exhibiting the highest diastereoselectivity
(95%).
The results obtained show that the tested BM3 mu-
tants were able to catalyze the formation of 3 out of
the 4 possible hydroxy diastereomers with high dia-
steromeric excess, Table 1. The only diastereomer that
could not be produced in high enantiomeric excess
was (3S,6R)-3-hydroxy-a-ionone when using (S)-a-
ionone as substrate. Almost all BM3 mutants formed
mainly (3S,6S)-3-hydroxy-a-ionone with this enantio-
mer as substrate, with only small amounts of (3R,6S)-
3-hydroxy-a-ionone. The mutant producing the high-
est amounts of (3R,6S)-3-hydroxy-a-ionone was found
to be M11L437N V87F, which also showed an aselec-
tive product profile with racemic-a-ionone. This
mutant produced (3R,6S)-hydroxy-a-ionone with 40%
selectivity (see the Supporting Information, Fig-
ure S7). Interestingly, wild-type P450 BM3 also pro-
duced about 36% (3R,6S)-hydroxy-a-ionone but with
a much lower activity compared to the BM3 mutant
M11L437N V87F. Apparently, the presence of phe-
nylalanine at position 87 seems to increase the
amount of (3R,6S)-hydroxy-a-ionone produced. Al-
2178
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2172 – 2184

Regio- and Stereoselective Hydroxylation of Optically Active a-Ionone Enantiomers
though this mutant still shows only moderate selectiv- binding poses as observed in our docking studies on
ity for the formation of this product, it could be the binding of (R)- and (S)-a-ionone to M01A82W
a good starting point to improve the stereoselectivity and M11L437N. In the case of mutant M01A82W,
by directed evolution.
the majority of obtained binding poses for both (R)-
and (S)-a-ionone was found in a position suitable for
C-3 hydroxylation, that is, C-3 was found within 6 ꢃ
from the heme iron catalytic center (Figure 5, A and
B)[41] in line with the experimental observation that
Docking Studies
Molecular docking studies were performed in order C-3 is the dominant site-of-metabolism. The introduc-
to rationalize the basis for the experimentally ob- tion of 82W substantially reduces the size of the bind-
served regio- and stereoselectivity of hydroxylation of ing cavity, which results in directing the substrate
a-ionone enantiomers. Figure 5 shows representative closer to the heme group and in an increase in activi-
Figure 5. Binding poses of (R)-a-ionone (depicted in green in A and C) and of (S)-a-ionone (depicted in magenta in B and
D) in mutant M01A82W (A and B) and M11L437N (C and D), as obtained from molecular docking studies. Selected amino
acid residues within the binding cavity of the mutants, and the catalytic active heme moiety are depicted as well. The sub-
strateꢂs main site-of-metabolism (C-3) is marked by a sphere. Hydrogen atoms are omitted, except for the C-3 bound hydro-
gens, which are shown in small spheres; cis-hydrogens are depicted in grey, and trans-hydrogens are depicted in orange. In
addition, the hydrogen atoms at the stereo-center (C-6) are explicitly shown, in small grey spheres, to highlight which stereo-
isomer is depicted. The thin blue line represents the hydrogen bond formed between the side-chain hydroxy moiety of the
Ser72 residue, and the carbonyl oxygen of a-ionone; the dashed line represents distances (in ꢃ) of C-3 toward the heme
iron atom.
Adv. Synth. Catal. 2012, 354, 2172 – 2184
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2179

Harini Venkataraman et al.
FULL PAPERS
ty, in line with previous suggestions based on a study
The role of A82W mutation was experimentally fur-
on the hydroxylation of palmitate by an A82W ther confirmed by incubating P450 BM3 mutant
mutant of BM3.[39] Also among the docking poses ob- M11L437N A82W with (R)-a-ionone. In line with the
tained for M11L437N in which the substrate was posi- above discussion, the addition of A82W to
tioned within the catalytic cut-off, C-3 was found to M11L437N was found to substantially decrease the
be in close proximity of the heme iron atom stereoselectivity and cis-hydroxylation at C-3 position
(Figure 5, C and D). Based on the poses in which C-3 of (R)-a-ionone, see the Supporting Information, Fig-
was found to be the carbon atom closest to the heme ure S8. In conclusion, these docking studies support
iron, the experimentally observed regioselective hy- the observations that residue 82W plays an important
droxylation at C-3 can be explained by the formation role in the stereoselective hydroxylation of a-ionone
of a hydrogen bond between the carbonyl oxygen and enantiomers by M01A82W and M11L437N.
active site residue Ser72, which seems to anchor the
The presented docking studies used a wild-type
ligand such that C-3 points toward the heme catalytic BM3 crystal structure (PDB code 1BU7) as structural
center, see Figure 5. Interestingly, substitution of ala- template for the mutants and did not explicitly ac-
nine by tryptophan has been shown to affect the re- count for the role of possible differences in active-site
gioselectivity[29] and along with hydrophobic substitu- conformations and water molecule orientations, which
tions at position 72 to affect stereoselectivity of hy- may affect docking outcomes.[44,45] As a first explora-
droxylation of larger substrates like steroids.[42] The tion of the effect of active-site flexibility on our dock-
observed anchoring of a-ionone by Ser72 is compara- ing results, we additionally performed docking studies
ble to the anchoring of a-ionone by His94 upon bind- using another crystal structure (PDB code 1 JPZ)[46]
ing to CYP109D1 as observed in a recent docking of wild-type BM3 as structural template for the mu-
study of Khatri et al.,[25] and by Thr285 as observed in tants. Results of these additional docking studies
a recent docking study by Ly et al.[26] upon a-ionone (data not shown) did not change the structural ration-
binding to CYP264B1.
alization discussed above in terms of the orientation
When docking either (R)- or (S)-a-ionone into of the substratesꢂ C-3 carbon atoms (with respect to
M01A82W, binding poses were predominantly ob- the catalytical iron center) and the role played by
tained in which the C-3-bound trans-hydrogen is ori- Trp82 and Ser72 in substrate binding orientation. Cur-
ented toward the heme iron atom (Figure 5, A and B, rently, the structural analysis of the binding of (R)-
respectively). Such binding poses will be referred to and (S)-a-ionone to BM3 mutants such as M01A82W
as “trans-orientated” binding poses. Experimentally, and M11L437N is extended towards a more rigorous
more than 90% of the products were found to be hy- study using molecular dynamics simulations, thereby
droxylated at the trans position. These findings sug- allowing inclusion of the effects of protein plastici-
gest that the observed binding orientation (Figure 5, ty[45,47] and water-mediated protein-substrate interac-
A and B) directs the stereoselectivity of the hydroxyl- tions,[48] to make possible a more detailed analysis of
ation process, in which the substrate is oxidized by the a-ionone binding process.
the activated iron-oxo ferryl species.[43] When docking
(S)-a-ionone into M11L437N, predominantly trans-
orientated binding poses were obtained as well Conclusions
(Figure 5, D), but interestingly, only for (R)-a-ionone,
a number of docking poses were found in which the The present study shows that individual enantiomers
cis-hydrogen was found closest to the heme iron atom of a-ionone can be hydroxylated by P450 BM3 mu-
(Figure 5, C). This offers a possible explanation for tants with high regio- and stereoselectivity along with
the experimental finding that hydroxylation at cis-po- high turnover numbers, which makes them an attrac-
sition of C-3 was observed for this substrate by tive option for the green production of diastereomers
M11L437N.
of 3-hydroxy-a-ionone. This is the first time that such
For M01A82W, the (R)-enantiomer is oriented stereoselectivity of optically active a-ionone hydroxyl-
with the trans-hydrogen of C-3 closest to the heme ation by P450 BM3 mutants has been reported. More-
iron atom, due to a steric conflict of the two methyl over, the formation of (3S,6R)-hydroxy-a-ionone by
groups bound to C-1 and the bulky tryptophan side M11L437N (90% de) exhibits the ability of engi-
chain at the 82 position (Figure 5, A), whereas neered P450 BM3 to produce cis-3-hydroxy-a-ionone.
M11L437N (lacking this bulky side chain) leaves While initial docking results yield some insights into
more space in the active site, making it possible for the regio- and stereoselectivity of a-ionone hydroxyl-
the double methyl groups to point towards the resi- ation and the role played by the A82W mutation in
due-82 side of the cavity as seen in Figure 5, C. For M01A82W, extensive free energy calculations and
the (S)-enantiomer, no such direction of the binding binding affinities of the individual enantiomers to
orientation to prevent sterical hindrance is observed, M01A82W and M11L437N are needed to further un-
as seen in Figure 5, B and D.
derstand the basis of stereoselectivity.
2180
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2172 – 2184

Regio- and Stereoselective Hydroxylation of Optically Active a-Ionone Enantiomers
buffer in Vivaspin 20 filtration tube (10,000 MWCO PES,
Sartorius) at 4000 g until the histidine concentration was
Experimental Section
below 250 nM.
Materials
Racemic a-ionone, enantiomers (R)-and (S)-a-ionone and
the hydroxy diastereomers (3S,6R)-hydroxy-a-ionone and
(3S,6S)-hydroxy-a-ionone were obtained from DSM Innova-
tive synthesis B.V., Geleen, The Netherlands. All other
chemicals were of analytical grade and purchased from
Sigma unless otherwise mentioned. Restriction endonucleas-
es were obtained from Westburg, Pfu polymerase and iso-
propyl b-d-thiogalactopyranoside (IPTG) were obtained
from Fermentas.
Incubations of P450 BM3 Mutants with a-Ionone and
Analysis by GC-MS
For the biotransformation reaction, cytosolic fraction con-
taining 200 nM of P450 BM3 mutants was incubated with
500 mM of racemic a-ionone in a final volume of 250 mL.
The reaction was initiated by the addition of 25 mL NADPH
regeneration system containing 5 mM NADPH, 50 mM glu-
cose 6-phosphate and 20 units of glucose 6-phosphate dehy-
drogenase. The reaction was allowed to proceed for 1 hour
and then stopped by placing the tubes on ice. After addition
of 10 mL of 5 mm carbazole as internal standard, samples
were extracted by vortexing with 2 mL of ethyl acetate for
30 sec. Subsequently samples were centrifuged at 4000 g for
5 min to separate phases. About 1 mL of the organic layer
was then transferred to GC vials for analysis. The extracts
were analyzed using a Hewlett–Packard 6890 series with
ATAS Programmable Injector 2.1. Samples were separated
on a DB-5 GC-column (30 mꢄ0.25 mm). Ionization was
done by electron impact (EI) at 70 eV with the mass selec-
tive detector 5973. The injector temperature was maintained
at 2508C. The temperature program for the GC was as fol-
lows: 2 min. stationary at 608C, ramp at 208C minÀ1 to
2808C and maintained at 2808C for 4 min. For incubations
with individual enantiomers, same conditions were used as
described above.
Enzymes and Plasmids
The 31 P450 BM3 mutants used in this work were construct-
ed as described previously.[28,31,49] Detailed information about
the mutations present is tabulated in the Supporting Infor-
mation, Table S2. Nine other mutants were made by site di-
rected mutagenesis using appropriate oligonucleotides (see
the Supporting Information, Table S1). In general, the muta-
tions were introduced in the corresponding templates in
pBluescript II KS(+) vector by the QuikChange mutagene-
sis protocol using the forward primers mentioned in the
Supporting Information, Table S1. The reverse primers were
exactly complementary to the forward primers. After muta-
genesis, the presence of the right mutations was verified by
DNA sequencing (Service XS, Leiden, The Netherlands).
The whole gene was subcloned to pET28a+ vector using
BamHI and EcoRI sites and later transformed to BL21
(DE3) cells for expression.
Dissociation Constant Determination by Optical
Spectroscopy
Expression and Purification of P450 BM3 Mutants
The binding of (R)- or (S)-a-ionone to wild-type P450 BM3
and mutants M01, M11, M01A82W and M11L437N was an-
alyzed by optical titrations using a Shimadzu UV-2501PC
spectrophotometer (Shimadzu Duisburg, Germany). The
spectra for the substrate-free enzyme (1 mM) were recorded
at 248C in 1 mL of 100 mM potassium phosphate buffer
(pH 7.4) prior to addition of a-ionone enantiomers. Increas-
ing amounts of (R)- or (S)-a-ionone dissolved in methanol
were added from appropriate stocks up to a final volume of
not more than 2% of the total volume of the solution. Spec-
tra were recorded after each addition of substrate, and a dif-
ference spectrum was computed by subtraction of the start-
ing (substrate-free) spectrum from those generated at each
point in the titration. The maximum apparent absorption
change induced at each point in the titration (DA390–422) was
determined by subtraction of the minimum absorption value
at the trough in each difference spectrum from the maxi-
mum value at the peak and these values were plotted versus
corresponding concentrations of (R)- or (S)-a-ionone after
correction for dilution. Data were analyzed by non-linear re-
gression using GraphPad Prism 4 (GraphPad Sotware, San
Diego, CA, USA). The spectral dissociation constants (KD)
were determined by fitting the titration binding curves to
the following equation:
For screening of products of racemic a-ionone hydroxyl-
ation, cytosolic fractions of the enzymes were used. Wild-
type P450 BM3, mutants M01, M11, M01A82W and
M11L437N were grown on a large scale and purified as fol-
lows: 600 mL terrific broth medium with 30 mg/mL kanamy-
cin were inoculated with 15 mL of overnight culture. The
cells were grown at 378C and 175 rpm until the OD600
reached 0.6. The protein expression was then induced by the
addition of 0.6 mM IPTG. The temperature was lowered to
208C and 0.5 mM of heme precursor delta-aminolevulinic
acid was added. Expression was allowed to proceed for 18 h.
Cells were harvested by centrifugation (4600 g, 48C, 20 min)
and the cell pellet was resuspended in 20 mL KPi-glycerol
buffer (100 mM potassium phosphate [KPi] pH 7.4, 10%
glycerol, 0.5 mM EDTA, and 0.25 mM DTT). Cells were dis-
rupted using a French press (1000 psi, 3 repeats), and the cy-
tosolic fraction was separated from the membrane fraction
by ultracentrifugation of the lysate (120,000 g, 48C, 60 min).
Then the enzymes were purified using Ni-NTA agarose
(Sigma). To prevent aspecific binding, 1 mM histidine was
added to the cytosolic fraction. 3 mL of Ni-NTA slurry were
added to 20 mL of cytosol and the mixture was equilibrated
at 48C for 2 h. The Ni-NTA agarose was retained in a poly-
propylene tube with porous disc (Pierce,Rockford,USA)
and was washed 4 times with 4 mL Kpi-glycerol buffer con-
taining 2 mM histidine. The P450 was eluted in 10 mL Kpi-
glycerol containing 200 mM histidine. The histidine was sub-
sequently removed by repeated washing with Kpi-glycerol
Adv. Synth. Catal. 2012, 354, 2172 – 2184
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2181

Harini Venkataraman et al.
FULL PAPERS
where DA represents peak-to-trough absorbance difference,
DAmax is the maximum absorbance difference, [S] is the sub-
strate concentration and Kd is the dissociation constant of
the enzyme-substrate complex. For substrate showing tight
Preparative-Scale Synthesis
Large-scale incubation was performed in a 3-L baffled flask
with 250 nM of purified mutant M01A82W, 48 mg of (S)-a-
ionone (final concentration 1 mM) in a volume of 250 mL in
100 mM potassium phosphate buffer (pH 7.4) at 248C. The
reaction was initiated by addition of 25 mL NADPH regen-
erating system. The reaction was allowed to proceed for 6 h
and then the products were extracted by using ethyl acetate.
The organic layer was dried over anhydrous Na2SO4 and an-
alyzed by GC-MS to determine the substrate conversion
and product formation.
binding
tion[50]
ACHTUNGTRENNUNG(Kd<1 mM), data were fitted to tight quadratic equa-
where DA represents peak-to-trough absorbance difference,
DAmax is the maximum absorbance difference, [S] is the sub-
strate concentration and [E] is the enzyme concentration.
Computational Methods
Mutants M01A82W and M11L437N were selected for mo-
lecular docking studies as they showed distinct stereoselec-
tivity in C-3 hydroxylation. Molecular structures of the mu-
tants were modeled based on chain A of the substrate-free
crystal structure (PDB code 1BU7),[53] using the molecular
modeling software MOE (Molecular Operating Environ-
ment).[54] Side-chains of the mutated amino acids were man-
ually replaced using the builder tool, and polar and aromatic
hydrogen atoms were added. The mutant structures were
energy minimized in vacuo using the GROMOS05 biomo-
lecular simulation package[55] and the GROMOS force field
parameter set 45A4.[56] Subsequently, aliphatic hydrogen
atoms were added using MOE. Molecular structures of the
substrate for use in docking were manually built using the
MOE program and energetically optimized using the
MMFF94x force field, as implemented in MOE. Subse-
quently, (R)- and (S)-a-ionone were docked into the
M01A82W and M11L437N active sites using GOLD (Ge-
netic Optimization for Ligand Docking),[58] version 4.0, and
the Chemscore scoring function.[59] The center point of the
docking sphere was placed in the middle of the protein
binding cavity, in between the side chain of residue 437 and
the heme iron atom, from which the radius was set to
1.5 nm to define the active site of the protein. For each
enantiomer, four independent docking simulations were per-
formed. Per docking simulation, maximally 50 docked poses
were retained. A maximum of 50000 operations in the
GOLD genetic algorithm were performed, using a popula-
tion of 100 genes. Generated binding poses for which all of
the substrateꢂs carbon atoms were positioned more than 6 ꢃ
away from the heme iron center (which is the hypothetical
cut-off for catalytic activity,[41] and which is comparable to
the distance of 5.6 ꢃ between the heme iron and the C-3
atom of b-ionone in the crystal structure with PDB code
3OFU[27]) were discarded for further analysis. The remaining
binding poses were visually inspected using MOE.
Determination of Coupling Efficiency
NADPH oxidation rates were measured at 258C using
a Libra S12 Biochrom UV-VIS spectrophotometer and 1 cm
path length cuvettes. To measure the NADPH consumption
rate, P450 BM3 (final concentration 200 nM) was mixed
with 100 mM KPi buffer pH 7.4 containing 200 mM a-ionone
enantiomersACHTUNGTRENNUNG(R or S) in methanol (2% final) in a volume of
1 mL. The reaction was initiated by the addition of NADPH
(end concentration 200 mM) and the decrease in absorption
was monitored over 30 seconds. Rates were calculated based
on
the
extinction
coefficient
of
NAPDH
e=
6,200MÀ1 cmÀ1.[51] To measure the coupling efficiency, the
product formation was quantified under the same condi-
tions. The percentage of coupling was determined as the
ratio of the amount of product formed to the amount of
NADPH consumed.
Determination of Total Turnover Numbers of
M01A82W and M11L437N
200 nM of purified enzymes (M01A82W and M11L437N)
were incubated with 1 mM of (R)- or (S)-enantiomers (2%
methanol) in a total volume of 500 mL. The reaction was ini-
tiated by the addition of 50 mL NADPH regenerating
system as mentioned in above. The reaction was performed
in duplicate in 1.5-mL Eppendorf tubes at 248C for 12 h
with shaking. After this incubation time, benzoxyresorufin-
O-dealkylation[52] was used to check the enzyme inactivation
to ensure completion of the reaction. Briefly, an aliquot of
the incubation (62.5 mL) was mixed with Kpi buffer to
a final volume of 125 mL. Benzoxyresorufin (25 mL) was
added in the assay buffer at a final concentration of 10 mM.
100 mM of NADPH regenerating system were added to the
reaction mixture and the reaction was monitored on
a micro-plate reader. No increase in resorufin fluorescence
was observed indicating that the enzyme was inactive. The
samples were then extracted and analyzed by GC-MS as de-
scribed above. Total turnover numbers were calculated as
nmol of product formed per nmol enzyme by using the cali-
bration curve of (3S,6S)-hydroxy-a-ionone (R2 =0.9982)
with concentrations ranging from 0.1 mM to 1 mM.
Acknowledgements
This work was financially supported by the Netherlands Or-
ganization for Scientific Research (NWO) within the IBOS
(Integration of Biosynthesis and Organic Synthesis) program.
2182
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2172 – 2184

Regio- and Stereoselective Hydroxylation of Optically Active a-Ionone Enantiomers
References
[25] Y. Khatri, M. Girhard, A. Romankiewicz, M. Ringle, F.
Hannemann, V. B. Urlacher, M. C. Hutter, R. Bern-
hardt, Appl. Microbiol. Biotechnol. 2010, 88, 485–495.
[26] T. T. Ly, Y. Khatri, J. Zapp, M. C. Hutter, R. Bernhardt,
Appl Microbiol Biotechnol 2012, 95, 123–133.
[27] M. Ma, S. G. Bell, W. Yang, Y. Hao, N. H. Rees, M.
Bartlam, W. Zhou, L. L. Wong, Z. Rao, Chembiochem
2011, 12, 88–99.
[28] M. C. Damsten, B. M. van Vugt-Lussenburg, T. Zelden-
thuis, J. S. de Vlieger, J. N. M. Commandeur, N. P. E.
Vermeulen, Chem.-Biol. Interact. 2008, 171, 96–107.
[29] V. Rea, A. J. Kolkman, E. Vottero, E. J. Stronks, K. A.
Ampt, M. Honing, N. P. E. Vermeulen, S. S. Wijmenga,
J. N. M. Commandeur, Biochemistry 2012, 51, 750–760.
[30] J. Reinen, J. S. van Leeuwen, Y. Li, L. Sun, P. D. Groo-
tenhuis, C. J. Decker, J. Saunders, N. P. E. Vermeulen,
J. N. M. Commandeur, Drug Metab. Dispos. 2011, 39,
1568–1576.
[31] E. Vottero, V. Rea, J. Lastdrager, M. Honing, N. P. E.
Vermeulen, J. N. M. Commandeur, J. Biol. Inorg.
Chem. 2011, 16, 899–912.
[32] B. M. A. van Vugt-Lussenburg, E. Stjernschantz, J.
Lastdrager, C. Oostenbrink, N. P. E. Vermeulen,
J. N. M. Commandeur, J. Med. Chem. 2007, 50, 455–
461.
[1] P. Winterhalter, R. Rouseff, in: Carotenoid-Derived
Aroma Compounds, Vol. 802, American Chemical Soci-
ety, Washington DC, 2001, pp 1–17.
[2] S. T. Withers, J. D. Keasling, Appl. Microbiol. Biotech-
nol. 2007, 73, 980–990.
[3] E. Brenna, C. Fuganti, S. Serra, P. Kraft, Eur. J. Org.
Chem. 2002, 33, 967–978.
[4] F. Khachik, A.-N. Chang, J. Org. Chem. 2009, 74, 3875–
3885.
[5] T. Ishida, H. Enomoto, R. Nishida, J. Chem. Ecol.
2008, 34, 1532–1535.
[6] W. A. Duetz, H. Bouwmeester, J. B. van Beilen, B.
Witholt, Appl. Microbiol. Biotechnol. 2003, 61, 269–
277.
[7] a) R. Fasan, ACS Catalysis 2012, 647–666; b) A. T.
Larsen, E. M. May, K. Auclair, J. Am. Chem. Soc. 2011,
133, 7853–7858.
[8] a) C. M. Clouthier, J. N. Pelletier, Chem. Soc. Rev.
2012, 41, 1585–1605; b) F. Hollmann, I. W. C. E.
Arends, K. Buehler, A. Schallmey, B. Buhler, Green
Chem. 2011, 13, 226–265.
[9] D. E. Torres Pazmino, M. Winkler, A. Glieder, M. W.
Fraaije, J. Biotechnol. 2010, 146, 9–24.
[10] G. Grogan, Curr. Opin. Chem. Biol. 2011, 15, 241–248.
[11] a) E. OꢂReilly, V. Kohler, S. L. Flitsch, N. J. Turner,
Chem. Commun. 2011, 47, 2490–2501; b) V. B. Urlach-
er, M. Girhard, Trends Biotechnol. 2012, 30, 26–36.
[12] a) U. T. Bornscheuer, M. Pohl, Curr. Opin. Chem. Biol.
2001, 5, 137–143; b) D. Bottcher, U. T. Bornscheuer,
Curr. Opin. Microbiol. 2010, 13, 274–282; c) P. C.
Cirino, F. H. Arnold, Curr. Opin. Chem. Biol. 2002, 6,
130–135; d) S. T. Jung, R. Lauchli, F. H. Arnold, Curr.
Opin. Biotechnol. 2011, 22, 809–817.
[13] a) S. G. Bell, X. Chen, R. J. Sowden, F. Xu, J. N. Wil-
liams, L. L. Wong, Z. Rao, J. Am. Chem. Soc. 2003,
125, 705–714; b) S. G. Bell, R. J. Sowden, L.-L. Wong,
Chem. Commun. 2001, 635–636.
[14] A. Chefson, K. Auclair, Mol. Biosyst. 2006, 2, 462–469.
[15] R. Bernhardt, J. Biotechnol. 2006, 124, 128–145.
[16] C. J. Whitehouse, S. G. Bell, L. L. Wong, Chem. Soc.
Rev. 2012, 41, 1218–1260.
[17] D. Appel, S. Lutz-Wahl, P. Fischer, U. Schwaneberg,
R. D. Schmid, J. Biotechnol. 2001, 88, 167–171.
[18] V. B. Urlacher, A. Makhsumkhanov, R. D. Schmid,
Appl. Microbiol. Biotechnol. 2006, 70, 53–59.
[19] S. Lutz-Wahl, P. Fischer, C. Schmidt-Dannert, W. Wohl-
leben, B. Hauer, R. D. Schmid, Appl. Environ. Micro-
biol. 1998, 64, 3878–3881.
[20] Y. Yamazaki, Y. Hayashi, M. Arita, T. Hieda, Y.
Mikami, Appl. Environ. Microbiol. 1988, 54, 2354–
2360.
[21] Y. X. Tang, T. Suga, Phytochemistry 1994, 37, 737–740.
[22] H. Sakamaki, K.-i. Itoh, W. Chai, Y. Hayashida, S. Ki-
tanaka, C. A. Horiuchi, J. Mol. Catal. B: Enzym. 2004,
27, 177–181.
[23] A. Celik, S. L. Flitsch, N. J. Turner, Org. Biomol. Chem.
2005, 3, 2930–2934.
[24] M. Girhard, T. Klaus, Y. Khatri, R. Bernhardt, V. B.
Urlacher, Appl. Microbiol. Biotechnol. 2010, 87, 595–
607.
[33] V. Rea, S. Dragovic, J. S. Boerma, F. J. de Kanter,
N. P. E. Vermeulen, J. N. M. Commandeur, Drug Metab.
Dispos. 2011, 39, 2411–2420.
[34] R. Nishida, H. Enomoto, T. E. Shelly, T. Ishida, Ento-
mol. Exp. Appl. 2009, 131, 85–92.
[35] a) E. T. Farinas, U. Schwaneberg, A. Glieder, F. H.
Arnold, Adv. Synth. Catal. 2001, 343, 601–606; b) S. C.
Maurer, K. Kꢅhnel, L. A. Kaysser, S. Eiben, R. D.
Schmid, V. B. Urlacher, Adv. Synth. Catal. 2005, 347,
1090–1098.
[36] T. Matsunaga, Y. Iwawaki, A. Komura, K. Watanabe,
T. Kageyama, I. Yamamoto, Biol. Pharm. Bull. 2002,
25, 42–47.
[37] N. J. Turner, Nat. Chem. Biol. 2009, 5, 567–573.
[38] K. Zhang, S. El Damaty, R. Fasan, J. Am. Chem. Soc.
2011, 133, 3242–3245.
[39] W.-C. Huang, A. C. G. Westlake, J.-D. Marꢆchal, M.
Gordon Joyce, P. C. E. Moody, G. C. K. Roberts, J.
Mol. Biol. 2007, 373, 633–651.
[40] A. Seifert, M. Antonovici, B. Hauer, J. Pleiss, Chem-
BioChem 2011, 12, 1346–1351.
[41] E. Stjernschantz, B. M. A. van Vugt-Lussenburg, A. Bo-
nifacio, S. B. A. de Beer, G. van der Zwan, C. Gooijer,
J. N. M. Commandeur, N. P. E. Vermeulen, C. Oosten-
brink, Proteins 2008, 71, 336–352.
[42] H. Venkataraman, S. B. de Beer, L. A. van Bergen, N.
van Essen, D. P. Geerke, N. P. E. Vermeulen, J. N. M.
Commandeur, ChemBioChem 2012, 13, 520–523.
[43] V. B. Urlacher, S. Lutz-Wahl, R. D. Schmid, Appl. Mi-
crobiol. Biotechnol. 2004, 64, 317–325.
[44] a) S. B. A. de Beer, L. A. van Bergen, K. Keijzer, V.
Rea, H. Venkataraman, C. F. Guerra, F. M. Bickel-
haupt, N. P. E. Vermeulen, J. N. M. Commandeur, D. P.
Geerke, Curr. Drug Metab. 2012, 13, 155–166;
b) S. B. A. de Beer, N. P. E. Vermeulen, C. Oostenbrink,
Curr. Top. Med. Chem. 2010, 10, 55–66; c) J. Hritz, A.
Adv. Synth. Catal. 2012, 354, 2172 – 2184
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2183

Harini Venkataraman et al.
FULL PAPERS
de Ruiter, C. Oostenbrink, J. Med. Chem. 2008, 51,
7469–7477.
[45] F. P. Guengerich, Proc. Natl. Acad. Sci. USA 2006, 103,
13565–13566.
[46] D. C. Haines, D. R. Tomchick, M. Machius, J. A. Peter-
son, Biochemistry 2001, 40, 13456–13465.
[47] C. Oostenbrink, A. de Ruiter, J. Hritz, N. Vermeulen,
Curr. Drug Metab. 2012, 13, 190–196.
[55] a) W. F. van Gunsteren, S. R. Billeter, A. A. Eising,
P. H. Hꢅnenberger, P. Krꢅger, A. E. Mark, W. R. P.
Scott, I. G. Tironi, Biomolecular simulation: The
GROMOS96 manual and user guide, Vdf Hochschul-
verlag AG an der ETH Zꢅrich, Zꢅrich, 1996; b) M.
Christen, P. H. Hꢅnenberger, D. Bakowies, R. Baron,
R. Bꢅrgi, D. Geerke, T. N. Heinz, M. A. Kastenholz, V.
Krꢇutler, C. Oostenbrink, C. Peter, D. Trzesniak, W. F.
van Gunsteren, J. Comput. Chem. 2005, 26, 1719–1751.
[56] R. D. Lins, P. H. Hunenberger, J. Comput. Chem. 2005,
26, 1400–1412.
[57] W. D. Cornell, P. Cieplak, C. I. Bayly, I. R. Gould,
K. M. Merz, D. M. Ferguson, D. C. Spellmeyer, T. Fox,
J. W. Caldwell, P. A. Kollman, J. Am. Chem. Soc. 1995,
117, 5179–5197.
[58] G. Jones, P. Willett, R. C. Glen, A. R. Leach, R. Taylor,
J. Mol. Biol. 1997, 267, 727–748.
[48] S. B. A. De Beer, N. P. E. Vermeulen, C. Oostenbrink,
Curr. Top. Med. Chem. 2010, 10, 55–66.
[49] J. Reinen, S. Ferman, E. Vottero, N. P. E. Vermeulen,
J. N. M. Commandeur, J. Biomol. Screening 2011, 16,
239–250.
[50] J. W. Williams, J. F. Morrison, Methods Enzymol. 1979,
63, 437–467.
[51] G. Truan, J. A. Peterson, Arch. Biochem. Biophys.
1998, 349, 53–64.
[52] B. M. A. Lussenburg, L. C. Babel, N. P. E. Vermeulen,
J. N. M. Commandeur, Anal. Biochem. 2005, 341, 148–
155.
[59] M. D. Eldridge, C. W. Murray, T. R. Auton, G. V. Paoli-
ni, R. P. Mee, J. Comput. Aid. Mol. Des. 1997, 11, 425–
445.
[60] A. Seifert, S. Vomund, K. Grohman, S. Kriening, V. B.
Urlacher, S. Laschat, J. Pleiss, ChemBioChem 2009, 10,
853–861.
[53] I. F. Sevrioukova, H. Li, H. Zhang, J. A. Peterson, T. L.
Poulos, Proc. Natl. Acad. Sci. USA 1999, 96, 1863–1868.
[54] Molecular Operating Environment (MOE), Montreal,
Quebec, Canada: Chemical Computating Group Inc.
2184
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 2172 – 2184