The substrate spectrum of mandelate racemase: Minimum structural requirements for substrates and substrate model

Article abstract of DOI:10.1002/adsc.200505012

Mandelate racemase (EC 5.1.2.2) is one of the few biochemically well-characterized racemases. The remarkable stability of this cofactor-independent enzyme and its broad substrate tolerance make it an ideal candidate for the racemization of non-natural α-hydroxycarboxylic acids under physiological reaction conditions to be applied in deracemization protocols in connection with a kinetic resolution step. This review summarizes all aspects of mandelate racemase relevant for the application of this enzyme in preparative-scale biotransformations with special emphasis on its substrate tolerance. Collection and evaluation of substrate structure-activity data led to a set of general guidelines, which were used as basis for the construction of a general substrate model, which allows a quick estimation of the expected activity for a given substrate.

Full text of DOI:10.1002/adsc.200505012

REVIEWS
The Substrate Spectrum of Mandelate Racemase: Minimum
Structural Requirements for Substrates and Substrate Model
Ulfried Felfer, Marian Goriup, Marion F. Koegl, Ulrike Wagner,
Barbara Larissegger-Schnell, Kurt Faber,* Wolfgang Kroutil*
Department of Chemistry, Organic & Bioorganic Chemistry, University of Graz, Heinrichstrasse 28, A-8010 Graz, Austria
Phone: (þ43)-316-380-5332, fax: þ43-316-380-9840, e-mail: Kurt.Faber@uni-graz.at.
Received: January 11, 2005; Accepted: March 16, 2005
Abstract: Mandelate racemase (EC 5.1.2.2) is one of quick estimation of the expected activity for a given
the few biochemically well-characterized racemases. substrate.
The remarkable stability of this cofactor-independent
enzyme and its broad substrate tolerance make it an
ideal candidate for the racemization of non-natural
1
2
3
4
Introduction
Substrate Spectrum
Substrate Model for Mandelate Racemase
Application of Mandelate Racemase in Deracem-
ization Processes
a-hydroxycarboxylic acids under physiological reac-
tion conditions to be applied in deracemization proto-
cols in connection with a kinetic resolution step. This
review summarizes all aspects of mandelate racemase
relevant for the application of this enzyme in prepara-
tive-scale biotransformations with special emphasis
on its substrate tolerance. Collection and evaluation
5
Experimental Section
of substrate structure-activity data led to a set of gen- Keywords: deracemization; enzymatic racemization;
eral guidelines, which were used as basis for the con- a-hydroxycarboxylic acid; mandelate racemase; race-
struction of a general substrate model, which allows a mase-lipase two-enzyme process; substrate spectrum
1 Introduction
chemically classified as subgroup EC 5.1.X.X among the
diverse and heterogeneous group of isomerases. Despite
The interconversion of enantiomers – racemization – their rare occurrence in Nature, their importance in syn-
goes in hand with a loss of “chiral value”[1] and thus thetic organic chemistry lies in the fact that they can cat-
has been generally regarded as an undesired side reac- alyze “chemically impossible” isomerization reactions.
tion rather than a synthetically useful transformation. Besides numerous racemases acting on a-amino acids,
As a consequence, the controlled racemization of organ- the best-studied racemase so far is mandelate racemase
ic compounds has been scarcely studied and the major (EC 5.1.2.2).
body of data available to date originates from industrial
Mandelate racemaseis produced by the soil bacterium
research, which was mainly driven by the demand to im- Pseudomonas putida ATCC 12633 presumably to allow
prove the economic balance of kinetic resolution proc- the strain to funnel d-mandelate (via racemization) into
esses.[2] It was only recently that the importance of syn- the l-specific mandelate degradation pathway[6,7] and to
thetic protocols for the controlled racemization of or- use it as a carbon- and energy-source.[8] The biochemical
ganic compounds under mild reaction conditions for characteristics of this enzyme,[9] its structure,[10] (PDB
the development of so-called “deracemization process- entry 1MDR) and mechanism of action were studied
es” has been recognized.[3] In this context, enzymatic in great detail by Kenyon and Hegeman.[11,12] The indu-
racemization[4,5] holds great potential, not only in view cible octameric enzyme (subunit 39 kDa) has been made
of the mild (physiological) reaction conditions, but available in large amounts by fermentation of Pseudo-
also due to the high compatibility of biocatalysts with monas putida ATCC 12633 using rac-mandelate as in-
each other to be applied in dynamic kinetic resolu- ducer[13] and it has been cloned.[14] Enzyme immobiliza-
tion.
tion onto the cationic carrier DEAE-cellulose leads to
Due to the fact that the vast majority of biochemical enhanced activity and facilitates recovery.[15] Simple as-
processes are stereospecific, Nature has faced little says for mandelate-racemase activity based on circular
need for racemization and, as a consequence, “racemas- dichroism[16] or on the decline of optical rotation over
es” are a small group of enzymes, which have been bio- time[17] have been developed.
Adv. Synth. Catal. 2005, 347, 951–961
DOI: 10.1002/adsc.200505012
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
951

Ulfried Felfer et al.
REVIEWS
Graz. After her Ph. D. in Graz in crystallography
she went for a post-doc to the Department of
Structural Chemistry at Weizmann Institute of Sci-
ence (Rehovot, Israel). She collected one year of
industrial experience at the Sandoz Research In-
stitute in Vienna (Austria) before she became an
assistant professor at the University of Graz in
1991, where she was promoted to associate profes-
sor in 1997. In 1999 she did a sabbatical at the
ESRF (European synchrotron radiation facilities)
in Grenoble, France.
Barbara Larissegger-Schnell, born 1966 in Alten-
markt (Salzburg/Austria), studied chemistry at
the University of Graz, where she received her
Ph. D. in 1994. Since 1993 she is part-time assistant
(with maternity breaks 1995–1996 and 2002–
2003) at the Department of Chemistry. Until
1998, she was a co-worker in the research group
of Prof. Thomas Kappe and in 1998 she joined
the research group of Prof. Kurt Faber.
From left to right: Kurt Faber, Ulrike Wagner, Ulfried
Felfer, Barbara Larissegger-Schnell, Wolfgang Kroutil.
Ulfried Felfer, born 1970 in Oberzeiring (Styria/
Austria), did his M.Sc. on “man-made” enzymes
with Prof. H. Griengl (Graz/Austria) and Prof.
S. M. Roberts at the University of Exeter (UK).
After graduating from Graz University of Tech-
nology in 1996 he continued his research with
Prof. K. Faber on dynamic kinetic resolution,
where he received his Ph. D. in 1998. In 1999 he
moved to DSM Fine Chemicals Austria. Since
2002 he is responsible for scale-up into (pilot) pro-
duction scale.
Kurt Faber, born 1953 in Klagenfurt (Carinthia/
Austria), studied chemistry at the University of
Graz, where he received his Ph. D. in 1982. From
1982–1983 he moved to St. Johnꢁs (Canada) for
a post-doc and continued his career at the Univer-
sity of Technology (Graz), where he became asso-
ciate professor in 1997. The following year he was
appointed full professor at the University of Graz,
where he is heading his research group devoted to
the use of biocatalysts for the synthetic transfor-
mation of non-natural compounds. He was a visit-
ing scientist at University of Tokyo (1987/1988),
Exeter University (1990), University of Trond-
heim (1994) and Stockholm University (2001).
Marian Goriup, born 1975 in Leoben/Austria,
studied technical chemistry at the University of
Technology in Graz and wrote his Diploma Thesis
with Prof. Faber on mandelate racemase. Under
the supervision of Prof. Saf he got his Ph. D. at
the Institute for Chemistry and Technology of Or-
ganic Materials/Graz in 2001. Afterwards he went
to industry and spent one year in Germany at Mul-
tek before he changed in 2003 to Borealis in Vien-
na.
Wolfgang Kroutil (born 1972 in Graz, Austria) re-
ceived his undergraduate training in chemistry at
the University of Technology in Graz and com-
pleted his graduate studies in Exeter (UK) and
Graz. After his Ph. D. he collected two years of in-
dustrial experience in the biocatalysis research
group at Syngenta (formerly Novartis CP) in Basel
(Switzerland) and in the R&D department of
Krems Chemie Chemical Services (Austria). In
2000 he became assistant professor and was pro-
moted to associate professor in 2004 at the Univer-
sity of Graz.
Marion Koegl, born 1978 in Graz/Austria, finished
her undergraduate training with a Diploma Thesis
at the University of Graz under the supervision of
Prof. Faber in 2003. During her undergraduate
studies she participated at a student exchange pro-
gram with Montreal/Canada (September 1994 –
February 1995). She started her Ph. D. under the
supervision of Prof. Mulzer at the University of
Vienna in 2003.
Ulrike G. Wagner (born 1960) received her under-
graduate training in chemistry at the University of
The catalytic mechanism of mandelate racemase is a tion, made feasible through the combined synergistic ef-
masterpiece of chemical catalysis, where evolution has fects of a number of chemical operators on the substrate
managed to catalyze a “chemically impossible” reac- within the active site of the enzyme.[18–22] The major cat-
952
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 951–961

The Substrate Spectrum of Mandelate Racemase
REVIEWS
Scheme 1. Schematic representation of mandelate racemase
catalysis.
Scheme 2. Schematic docking of (S)-l-mandelate onto the
chemical operators within the active site of mandelate race-
mase. The pair of catalytically active bases and the a-H are
emphasized in bold.
The problem is overcome by the binding of both enan-
tiomers within the active site through a tight network of
salt bridges and hydrogen bonds involving two Brønsted
acids (Glu 317 and Lys 164) and a Lewis acid (Mg2þ in
the native enzyme[25]) (Scheme 2). This intimate ar-
rangement of the a-hydroxycarboxylic acid moiety of
the substrate within the electron-withdrawing ligands
within the active site diminishes the electron density at
the a-position to such an extent that the a-H abstraction
by an adjacent base becomes feasible through a “two-
base-mechanism”.[26] Racemization takes place via de-
protonation by two enantiomer-specific bases juxta-
posed on either side of the chiral a-carbon atom, i.e.,
His-297 and Lys-166 for (R)- and (S)-mandelate, respec-
tively; while one base acts as base, the other functions as
the corresponding acid for proton delivery from the op-
posite side. This “molecular ping-pong game” makes
mandelate racemase an extremely efficient catalyst
with a turnover frequency of 1000 sꢀ1 [27]
meaning that
1.0 g of the enzyme racemizes approximately 1.7 kg of
mandelic acid per hour.
,
Figure 1. Active site of mandelate racemase containing (S)-
mandelate. The main structural element (so-called TIM-bar-
rel) consisting of a circular arrangement of eight parallel b-
sheets is shown in dark red. On top of it, the catalytically ac-
tive Mg2þ ion (green) is positioned through a network of H-
bonds and salt bridges (dashed) within an array of Asp, Glu
and Lys residues. The polar a-hydroxycarboxylic acid moiety
of the substrate (orange) is docked onto this same array,
while the non-polar aryl moiety is accommodated in the lip-
ophilic binding pocket (yellow). Catalytically active bases for
proton abstraction juxtaposed on each side of the substrate
(Lysꢀ166, Hisꢀ297, turquoise) are annotated.
2 Substrate Spectrum
The value of an enzyme for the selective biotransforma-
tion of non-natural organic compounds on a preparative
scale is predominantly characterized by its substrate tol-
erance, i.e., how many substrates of more or less related
structure can be transformed at acceptable rates. Al-
though first hints on the relaxed substrate specificity of
mandelate racemase were published quite early,[14] it
alytic hurdle during the interconversion of mandelate was the availability of this enzyme by fermentation[13]
enantiomers is the abstraction of the a-proton of man- that facilitated the exploration of its substrate tolerance
delate possessing a (formal) pKa value of 29[23] to gener- considerably.[15,28] Moreover, molecular modeling
ate the corresponding achiral enolate intermediate showed that the hydrophobic binding pocket within
(Scheme 1), which is facilitated through resonance sta- the active site (which accommodates the phenyl moiety
bilization by the phenyl-substituent R.[24]
of the natural substrate, mandelate, see Fig. 1) shows a
Adv. Synth. Catal. 2005, 347, 951–961
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
953

Ulfried Felfer et al.
REVIEWS
Table 1. Substrate tolerance of mandelate racemase.
Substrate
Relative Activity
[%][a]
Ref.
Substrate
Relative Activity
Ref.
[%][a]
[32]
100
n. a.
96
[31]
[32]
[31]
[31]
[32]
[32]
[32]
[31]
[31]
[31]
[32]
[32]
326
61
X1[b]
376
45
17
0.03[b]
X0.02[b]
59[c]
38[c]
X1
26
24[c]
53
this study
this study
[31]
X1[b]
X0.01[b]
50
this study
X0.01[b]
[31]
[31]
5.4
this study
36
this study
954
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 951–961

The Substrate Spectrum of Mandelate Racemase
REVIEWS
Table 1 (cont.)
Substrate
Relative Activity
[%][a]
Ref.
Substrate
Relative Activity
Ref.
[%][a]
[33]
[31]
35
X0.01[b]
[33]
X0.01[b]
1
this study
[28]
[28]
15
22[b]
[28]
[28]
X1
X1[b]
[28]
[34]
X1[b]
n. a.e
[35]
8.5[d]
n. a. not applicable.
[a]
Relative activities are expressed as % relative to the natural substrate mandelate (R)-1 (100%).
Limit of sensitivity of the assay procedure used.
Previously published values were corrected mathematically by the initial ratio of the specific optical rotations of substrate to
reference (mandelate).
Substrate 32 acts as strong inhibitor.
[b]
[c]
[d]
[e]
Not applicable, substrate 33 undergoes elimination of HBr to form p-methylbenzoyl formate.
remarkable plasticity;[29] particularly the latter feature tive. Total replacement of the phenyl moiety by a bulky
makes this enzyme a highly desired “broad substrate naphthyl group (11) or by a heterocycle, such as furan
spectrum” catalyst.
(13, 14) or thiophene (12) gave moderate to good race-
Careful analysis of the substrate structure-activity mization rates.
data reported so far together with six novel data sets al-
(2) Non-aromatic substrates: A remarkable feature
lowed us to deduce a set of rules which facilitate the pre- of the flexibility of mandelate racemase lies in the fact
diction of the relative activity of a given substrate for that the aromatic system can be drastically reduced to
¼
mandelate racemase (Table 1).
a minimal p-system consisting of a single b,g-C C moi-
(1) Variation of the aryl moiety: Substituents on the ety while retaining a good part of its catalytic activity
aryl moiety of mandelate are well-accepted, in particu- (35–50%), which is valid not only for cyclic derivatives
lar in the (sterically less stringent) p-position, such as (21), but also for straight-chain analogues (22, 23).
halogen (2, 3, 6), hydroxy or alkoxy (7, 8). While some
(3) Aliphatic and aryl-aliphatic substrates: Aliphatic
reduction in activity is observed with m-substituents (24–26) and aryl-aliphatic a-hydroxycarboxylic acids
(4), o-substituted (5) and m, p-disubstituted mandelate (17–20), which have no (or an “electronically discon-
derivatives (9, 10) are not accepted, presumably due to nected”) aryl moiety cannot contribute to the resonance
steric hindrance, which is most striking for the o-deriva- stabilization of the a-carbanion intermediate[24] and are
Adv. Synth. Catal. 2005, 347, 951–961
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
955

Ulfried Felfer et al.
REVIEWS
Scheme 3. Substrate model for mandelate racemase.
therefore non-substrates. However, when the aryl moi- their aromaticity (expressed as the resonance energy):
ety is linked through a conjugated p-system (E-15), car- Whereas furan analogues (13) and (14) (resonance ener-
banion stabilization is feasible and, as a consequence, gy of furan ca. 56 kJ/mol) showed relative rates of 24–
good activity is observed. The inactivity of the corre- 38%, the corresponding thiophene analogue (12) gave
sponding Z-isomer (16)is most likely caused by steric re- an enhanced rate of 59% (resonance energy of thio-
strictions.
phene ca. 84 kJ/mol, cf. resonance energy of benzene
(4) Variation of the a-hydroxycarboxylic acid moi- 124 kJ/mol).
ety: Bearing in mind that the a-hydroxycarboxylic acid
(6) Steric effects: Negative effects of steric hindrance
moiety has a critical function as docking group though clearly appear to dominate in close vicinity of the a-cen-
multipoint attachment within the active site of the en- ter, i.e., o-substituted mandelates (5) are not racemized
zyme, any alterations in this region have to be regarded at all, and some hindrance seems to be true also for the
with caution. All attempts to alter the a-hydroxy group m-position (4, 9, 10). This is in good agreement with the
failed so far: a-phenylglycine (30) and the correspond- inactivity of the “bent” Z-configurated substrate (16). In
ing N-acetyl derivative (31) were totally inactive, most contrast, there is apparently quite some space available
likely due to the inability of the amino/ammonium or within the active site beyond the aryl moiety to accom-
aminoacetyl group to form a tight salt bridge with the modate a bulky naphthyl group (11) (cf. Fig. 1).
catalytically active Mg2þ ion. Most remarkably, varia-
(7) Substrate inhibitors: In rare cases, unexpected re-
tion of the carboxylate moiety to the corresponding car- sults were obtained with a-hydroxycarboxylic acids. The
boxamide was tolerated to some extent: mandeloamide acetylenic analogue propargyl glycolate (33) was slowly
(27) and its p-bromo analogue (28) were racemized with racemized, but acted as a strong inhibitor.[35] p-Bromo-
acceptable rates (15 and 22%, respectively).
methyl mandelate (32) underwent unexpected elimina-
(5) Electronic effects: Overall, the requirement for tion of hydrogen bromide to form p-methylbenzoyl for-
resonance stabilization of the a-carbanion intermediate mate. The latter reaction could be sufficiently explained
during catalysis appears to be a striking prerequisite for by expulsion of bromide anion from the a-carbanion in-
catalysis: Whereas mandelate derivatives bearing elec- termediate, followed by rapid tautomerization of the
tron-donating substituents (7, 8) show markedly re- formed exo-methylene enol to furnish the correspond-
duced racemization rates, electron-withdrawing groups, ing a-carbonyl compound.[34]
such as Cl (3) or Br (6) lead to significantly (more than
three-fold) enhanced racemization rates. A similar
rate-enhancing effect by a p-Br substituent was ob- 3 Substrate Model for Mandelate
served for the corresponding mandeloamides (27) and
(28). For the fluoro analogue (2), the electron-withdraw-
Racemase
ing effect is cancelled out by the well-known “back-don- The data on the substrate-activity pattern of mandelate
ation” effect of F.[30] The ability of heterocycles to stabi- racemase available so far allow us to construct a simple
lize the a-carbanion intermediate is nicely paralleled predictive “substrate model”[36] (Scheme 3). The “ideal”
with their ability for resonance stabilization through substrate consists of an a-hydroxycarboxylic acid. While
956
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 951–961

The Substrate Spectrum of Mandelate Racemase
REVIEWS
Scheme 4. Deracemization of (ꢁ)-mandelate using a stepwise two-enzyme process.
carboxamides are tolerated to some extent, replacement 5 Experimental Section
of the a-hydroxy group is forbidden. The requirement
for a p-system attached to the a-carbon is stringent,
thus no “electronic disconnecting” spacers, such as
Modeling
CH2 are allowed. Ideally, the p-system should be fully ar- Figure 1 was created using Pdb coordinates 1MRA by replac-
ing (S)-atrolactic acid (used as irreversible inhibitor) with
(S)-mandelate. Rendering was accomplished by render (Ras-
ter3d);[40] the picture was created using Molscript[41] and sur-
faces were generated as reported.[42]
omatic, although the minimum requirements consist of a
single C C bond. Electron-withdrawing groups on this
system dramatically enhance the activity up to several-
fold. Steric restrictions apply for the o-aryl position,
while ample space is available in the p-position, at the
far side of the substrate (cf. Fig. 1).
¼
Determination of Activity
A semi-purified enzyme preparation of mandelate racemase
was prepared as previously described, the specific activity
4 Application of Mandelate Racemase in
Deracemization Processes
was 107 mmol mgꢀ1 minꢀ1 [13]
Relative rates of racemization
.
were determined using an assay based on the online measure-
ment of the decline of optical rotation versus time with the cor-
responding corrections.[43] In experiments, where the substrate
was not optically pure, the ee of mandelate (used as standard)
was adjusted to the ee of the substrate in order to obtain com-
parable data for relative activity. Optical rotations were meas-
ured on a Perkin-Elmer Polarimeter 341 in a 1-mL cuvette of
10 cm length. The absence of spontaneous racemization and
undesired side-reactions under the reaction conditions was
The first application of mandelate racemase for the de-
racemization of a rac-a-hydroxycarboxylic acid on a
preparative scale was demonstrated recently.[37] Thus,
when rac-mandelate was subjected to kinetic resolution
via Pseudomonas sp. Lipase-catalyzed acyl transfer in
diisopropyl ether using vinyl acetate as acyl donor, (S)-
O-acetyl mandelate was formed while the (R)-mande-
late remained untouched. Since the formed product is verified for all substrates (i) in the absence of biocatalyst and
(ii) in the presence of heat-denatured enzyme; it was proven
to be <0.5% within 48 h for all substrates. For non-substrates
the negative result was verified for a period of 24 hours and
then analyzing the mixtures by HPLC.
not a substrate for mandelate racemase, the non-con-
verted substrate enantiomer could be racemized by
mandelate racemase in situ, i.e., without substrate-prod-
uct separation. Although mandelate racemase is re-
markable stable towards deactivation by organic sol-
vents, it requires high water activity for catalytic activity
and, as a consequence, the enzyme is inactive, but not de-
activated in organic solvents, such as diisopropyl ether.
Thus, the reaction medium had to be switched from or-
Crotonaldehyde, ethyl glyoxylate, ethynylbenzene, 1-cyclo-
hexene-1-carboxaldehyde,
(S)-2-hydroxy-4-phenylbutyric
acid ethyl ester and (R)-2-hydroxy-4-phenylbutyric acid were
commercially available. Column chromatography was per-
1
formed using Merck 60 silica gel (0.040–0.063 mm). H and
13C NMR were recorded on a Bruker 200 MHz spectrometer
ganic to aqueous.[38,39] After four cycles, (S)-O-acetyl at 200 and 50.3 MHz, respectively, using TMS as internal stand-
ard. Chemical shifts are reported in ppm (d) and coupling con-
stants (J) are given in Hz.
mandelate was obtained in x80% yield and>98% ee
as the sole product.
In summary, mandelate racemase constitutes a bio-
chemically well-characterizedand remarkably stable bio-
catalyst for the racemization of a-hydroxycarboxylic
acids under mild (physiological) conditions for applica-
tion in the deracemization of (ꢁ)-a-hydroxycarboxylic
acids. The enzyme shows a broad substrate tolerance
within certain well-defined structural limitations, which
can be deduced by a general substrate model proposed
in this paper.
Synthesis of (R)-(E)-2-Hydroxy-3-pentenoic Acid
[(R)-22] from Crotonaldehyde Employing (R)-
Hydroxynitrile Lyase from Almond Meal
Caution: Due to the handling of solutions containing hydrogen
cyanide, the following procedure was performed in a well-ven-
tilated hood. A crude enzyme preparation of (R)-hydroxyni-
trile lyase [(R)-HNL] from almonds was prepared according
Adv. Synth. Catal. 2005, 347, 951–961
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
957

Ulfried Felfer et al.
REVIEWS
to the literature.[44] Sodium cyanide (2.94 g, 60 mmol) was dis-
solved in water (75 mL) and the pH was adjusted to 5.5 with
glacial acetic acid. The solution was extracted once with methyl
tert-butyl ether (MTBE, 75 mL). The organic layer was sepa-
rated and immediately used for enzymic cyanohydrin syn-
thesis. The (R)-hydroxynitrile lyase preparation was rehydrat-
ed with citrate buffer (9 mL, 20 mM, pH 5.5) for 30 min. Fresh-
ly prepared MTBE-HCN solution (~40 mmol HCN) was add-
ed at once at þ38C before freshly distilled crotonaldehyde
(1.4 g, 20 mmol) was added via a syringe. The reaction was stir-
red at þ38C for 24 h after which the biocatalyst was separated
by filtration. The filtrate was extracted with MTBE (2ꢂ
50 mL), the combined organic layers were dried (Na2SO4)
and the solvent was evaporated under reduced pressure to
yield the crude, non-racemic cyanohydrin which was immedi-
ately transformed to the corresponding (R)-E-2-hydroxypen-
tenoic acid methyl ester via a modified Pinner reaction.[45]
Without purification, the cyanohydrin was dissolved in anhy-
drous diethyl ether (20 mL) and anhydrous methanol (2 mL)
was added. Dry HCl gas was bubbled through the solution
for about 15 min to precipitate the hydrochloride salt of the
corresponding imidate, which was then left at 48C for two
days to complete crystallization. The salt was filteredand wash-
ed with cold anhydrous diethyl ether (10 mL). The ice-cold hy-
drochloride salt was suspended in diethyl ether (5 mL/g salt)
and distilled water (same amount as ether) was added drop-
wise. The solution was stirred for 15 min at 08C before it was
allowed to warm to room temperature. The organic phase
was separated and the aqueous phase was extracted with dieth-
yl ether. The combined organic layers were washed with water
until the pH was neutral, dried (Na2SO4) and the solvent was
evaporated under reduced pressure to yield (R)-E-2-hydroxy-
pentenoic acid methyl ester, which was directly hydrolyzed
with sodium hydroxide solution [20 mL, 10% (w/v)]. After
completion (2 h), the reaction mixture was extracted with cy-
clohexane (4ꢂ30 mL), acidified to pH 2.0 (1 M aqueous
HCl) and extracted with ethyl acetate (5ꢂ70 mL). The organic
layers were combined, dried (Na2SO4) and concentrated to af-
ford (R)-E-2-hydroxy-3-pentenoic acid (R)-E-22 with 94% ee;
(3ꢂ20 mL). The combined organic phase was dried (Na2SO4)
and the organic solvent evaporated under reduced pressure
to yield the crude cyanohydrin, which was further transformed
into rac-(E)-2-hydroxy-3-pentenoic acid via the modified Pin-
ner reaction followed by ester hydrolysis as described above
for the (R)-(E)-2-hydroxy-3-pentenoic acid (R)-22; overall
yield of pure rac-(E)-2-hydroxy-3-pentenoic acid (rac-22):
1.4 g (54%). The NMR spectra were identical to those of the
non-racemic material.
Synthesis of (R)-(E)-2-Hydroxy-4-phenyl-3-butenoic
Acid [(R)-15] from Cinnamaldehyde Employing (R)-
Hydroxynitrile Lyase from Almond Meal
The reaction was performed as described for (R)-(E)-2-hy-
droxy-3-pentenoic acid [(R)-22] starting from cinnamalde-
hyde. The product (R)-(E)-2-hydroxy-4-phenyl-3-butenoic
acid was obtained as yellowish crystals; overall yield: 0.43 g
(30%); mp 114–1178C (petroleum ether/acetone) {ref.[46]
1398C}; [a]2D0: ꢀ81 (c 0.29, MeOH) {ref.[46] ꢀ98 (R) (c 1,
MeOH)}; ee 63% (HPLC). 13C and 1H NMR matched with re-
ported data.[47]
Synthesis of rac-(E)-2-Hydroxy-4-phenyl-3-butenoic
Acid (rac-15) from Cinnamaldehyde
The reaction was performed as described for rac-(E)-2-hy-
droxy-3-pentenoic acid (rac-22) starting from cinnamaldehyde
(3.32 g, 25 mmol) and TMSCN (2.97 g, 30 mmol) to afford rac-
(E)-2-hydroxy-4-phenyl-3-butenoic acid (rac-15); overall
yield: 2.6 g (62%). The structure was verified as described for
the non-racemic product.
Synthesis of (R)-2-Hydroxypentanoic Acid [(R)-26]
from (R)-(E)-2-Hydroxy-3-pentenoic Acid [(R)-22]
1
overall yield: 0.71 g (42%), [a]2D0: ꢀ60.2 (c 1.32, acetone). H
NMR (acetone-d6): d¼1.7 (3H, d, J¼7 Hz, CH3), 4.6 (1H, d,
(R)-(E)-2-Hydroxy-3-pentenoic
acid
[(R)-22;
0.25 g,
¼
J¼6 Hz, CH-2), 5.6 (1H, dd, J¼6, 15 Hz, CH-3), 5.8 (1H,
2.15 mmol) was dissolved in absolute methanol (20 mL) before
Pd/C (10%, 5 mg) was added. The flask was evacuated twice
and flushed with hydrogen. The reaction mixture was stirred
at room temperature for 2 h, the catalyst was removed by filtra-
tion and the organic solvent was removed under reduced pres-
sure to give (R)-2-hydroxypentanoic acid [(R)-26]; yield:
0.227 g (89%); 94% ee; [a]2D0: þ1 (c 1.05, MeOH). The structure
was confirmed by comparison of the 1H NMR spectrum with
literature data.[48]
m, CH-4); 13C-NMR: (acetone-d6): d¼17.6 (CH3), 71.8
¼
(CHOH), 128.3 (C-3), 129.5 (C-4), 174.4 (COOH).
Synthesis of rac-(E)-2-Hydroxy-3-pentenoic Acid (rac-
22) from Crotonaldehyde
Trimethylsilyl cyanide (TMSCN; 2.38 g, 3.0 ml, 24 mmol) was
added to a solution of freshly distilled crotonaldehyde (1.4 g,
1.66 ml, 20 mmol) in absolute dichloromethane (5 ml). After-
wards ZnI2 (5 mg) was added, the flask was sealed and the mix-
ture was stirred at 08C for 15 min before it was allowed to warm
to room temperature and stirred for another 45 min. The or-
ganic solvent was evaporated under reduced pressure to give
the crude TMS cyanohydrin. Without further purification the
crude intermediate was suspended in aqueous HCl (3 N,
20 mL) on an ice bath. After stirring for about 1 h at 08C the
mixture was allowed to warm to room temperature. Methyl
tert-butyl ether (MTBE, 20 mL) was added and the mixture
was vigorously stirred for 15 min, before the organic phase
was separated. The aqueous layer was extracted with MTBE
Racemic 2-hydroxy pentanoic acid (rac-26) was synthesized
by analogy starting from rac-2-hydroxy-3-pentenoic acid (rac-
22).
Synthesis of (S)-(Z)-2-Hydroxy-4-phenyl-3-butenoic
Acid [(S)-16]
Substrate (S)-16 was obtained via the following reaction se-
quence starting from ethynylbenzene and ethyl glyoxylate
with rac-2-hydroxy-4-phenyl-3-butynoic acid ethyl ester as
the first intermediate.
958
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 951–961

The Substrate Spectrum of Mandelate Racemase
REVIEWS
rac-2-Hydroxy-4-phenyl-3-butynoic acid ethyl ester:[49,50] n-
Butyllithium (18.5 mL, 2.5 M in hexane; 46 mmol) was added
dropwise to ethynylbenzene (5 g, 49 mmol) in THF (50 mL)
at ꢀ788C under argon atmosphere. The reaction mixture
was stirred for 30 min at ꢀ788C before ethyl glyoxylate
(9.7 mL, 43% in toluene)[51] was added within 15 min. After
6 h of stirring at ꢀ788C the reaction was stopped by addition
of glacial acetic acid (5 mL) and the mixture was warmed to
08C. After addition of brine (40 mL) the mixture was extracted
with ethyl acetate (3ꢂ50 mL), the combined organic phases
were dried (Na2SO4), the solvent was evaporated and the prod-
uct was purified using silica gel chromatography (petroleum
ether/ethyl acetate¼10:1) to give pure rac-2-hydroxy-4-phe-
nyl-3-butynoic acid ethyl ester as a yellow oil; yield: 200 mg
was evaporated and the product mixture was purified by silica
gel chromatography [petroleum ether/ethyl acetate/glacial
acetic acid 2:1:(50 mL/L)] to give (S)-(Z)-2-acetoxy-4-phe-
nyl-3-butenoic acid; yield: 5 mg (0.022 mmol, 39%); ee
(product):>97% (HPLC), enantioselectivity E>100.
1
ꢀ
ꢀ
H NMR (CDCl3): d¼2.12 (3H, s, CH3 COO ), 5.78 (1H, t,
ꢀ
¼
ꢀ
J¼11 Hz, CH CH CHOCOCH3), 5.86 (1H, d, J¼10.5 Hz,
ꢀ
¼
ꢀ
ꢀ
CH CH CHOH ),
6.94
(1H,
d,
J¼11.2 Hz,
ꢀ
¼
ꢀ
ꢀ
CH CH CHOH ); 7.27–7.45 (5H, m, Ar).
(S)-(Z)-2-Hydroxy-4-phenyl-3-butenoic acid [(S)-16]: (S)-
(Z)-2-Acetoxy-4-phenyl-3-butenoic acid (5 mg, 0.022 mmol)
was dispersed in Mg-HEPES buffer (3 mL, 50 mM HEPES,
3.3 mM MgCl2 ·6 H2O, pH 7.6) in a ultrasonic bath. After addi-
tion of lipase from Candida rugosa (40 mg), the mixture was
shaken at 308C at 130 rpm for 20 days. The mixture was centri-
fuged and the pH was adjusted to pH 1 using 1 N aqueous HCl.
Extraction with ethyl acetate (4ꢂ5 mL), drying (Na2SO4) and
evaporation of the organic solvent under reduced pressure af-
forded (S)-(Z)-2-hydroxy-4-phenyl-3-butenoic acid; yield:
3.5 mg (0.02 mmol, 91%), [a]2D0: þ59.8 (c 0.225, MeOH), ee
91% (HPLC). NMR data were identical to those of the racemic
compound. The absolute configuration was deduced using
three different indicators: (i) Kazlauskas rule,[52] ii) comparison
of the sense of the optical rotation with other homochiral b,g-
unsaturated-a-hydroxy acids, and (iii) comparison of elution
order on chiral HPLC with other b,g-unsaturated-a-hydroxy
acids.
1
(0.98 mmol, 2%). H NMR (CDCl3): d¼1.4 (3H, t, J¼7 Hz,
CH3), 3.4–3.6 (1H, br, OH), 4.3 (2H, q, J¼7 Hz, CH2), 5.1
ꢀ
ꢀ
(1H, s, CHOH ), 7.3–7.4 (3H, m, Ar), 7.4–7.5 (2H, m, Ar);
13C NMR: (CDCl3): d¼14.0 (CH3), 61.9, 62.8 (CH2-O,
ꢀ
CHOH), 84.3, 85.4 (Ar-CꢃC), 121.9–131.9 (6C, Ar), 170.4
¼
(C O).
rac-(Z)-2-Hydroxy-4-phenyl-3-butenoic acid ethyl ester:
rac-2-Hydroxy-4-phenyl-3-butynoic acid ethyl ester (200 mg,
0.98 mmol) was reduced to the corresponding Z-alkene using
quinoline (208 mg, 1.36 mmol) and Lindlar catalyst (80 mg,
5% wt Pd on CaCO3 with Pb, Sigma) in ethanol (2 mL). The
mixture was stirred under a hydrogen atmosphere (1 bar) for
6 h. The catalyst was filtered off and Amberlite IR-120
(240 mg, Hþ form) was added to remove the remaining quino-
line. The mixture was stirred at room temperature for 2 h, fil-
tered and the organic solvent was evaporated under reduced
pressure to give rac-(Z)-2-hydroxy-4-phenyl-3-butenoic acid
1
ethyl ester as a yellow oil; yield: 158 mg (79%). H NMR
rac-(Z)-2-Acetoxy-4-phenyl-3-butenoic Acid
(CDCl3): d¼1.4 (3H, t, J¼7 Hz, -CH3), 4.3 (2H, q, J¼7 Hz,
ꢀ
¼
ꢀ
ꢀ
A mixture of p-dimethylaminopyridine (10 mg) and acetic an-
hydride (38.8 mg, 0.38 mmol) was stirred at 48C for 15 min be-
fore rac-(Z)-2-hydroxy-4-phenyl-3-butenoic acid (50 mg,
0.28 mmol) was added. Stirring was continued at room temper-
ature for 16 h. Dichloromethane (11 mL) and ice water (6 mL)
were added and the mixture was stirred for 30 min, the organic
phase was separated, the aqueous phase was extracted with di-
chloromethane (5 mL), the combined organic phases were
dried (Na2SO4) and the organic solvent was evaporated under
reduced pressure. Purification using silica gel column chroma-
CH2), 5.1 (1H, d, J¼10 Hz, CH CH CHOH ), 5.7 (1H, t,
ꢀ
¼
ꢀ
ꢀ
J¼10 Hz, CH CH CHOH ), 6.8 (1H, d, J¼11 Hz,
CH CH CHOH ), 7.2–7.5 (5H, m, Ar); 13C NMR
ꢀ
¼
ꢀ
ꢀ
ꢀ
ꢀ
(CDCl3): d¼14.0 (CH3), 62.3, 67.3 (CH2-O, CHOH ),
¼
127.3, 127.8, 128.3, 128.9, 135.8, 134.7 (Ar-CH CH), 173.9
¼
(C O).
rac-(Z)-2-Hydroxy-4-phenyl-3-butenoic acid: rac-(Z)-2-
Hydroxy-4-phenyl-3-butenoic acid ethyl ester (158 mg,
0.767 mmol) was hydrolyzed in 1 M aqueous LiOH (1.8 mL)
and THF (0.2 mL) at room temperature for 24 h. The reaction
mixture was acidified with 1 M HCl to pH 2 and extracted with
ethyl acetate (4ꢂ4 mL), the combined organic phase was dried
(Na2SO4) and the solvent evaporated under reduced pressure
to give rac-(Z)-2-hydroxy-4-phenyl-3-butenoic acid as yellow
tography [petroleum ether (ethyl acetate/glacial acid
¼
2:1:(50 mL/L)] afforded the pure product as a colorless oil;
yield: 30 mg (0.18 mmol, 64%). The NMR data were identical
to those of (S)-(Z)-2-acetoxy-4-phenyl-3-butenoic acid.
crystals; yield: 120 mg (0.67 mmol, 87%); mp 102.2–1048C.
1
ꢀ
ꢀ
H NMR (CDCl3): d¼5.1 (1H, d, J¼10 Hz, CHOH ), 5.7
ꢀ
¼
ꢀ
ꢀ
(1H, t, J¼10 Hz, CH CH CHOH ), 6.87 (1H, d, J¼
10 Hz, CH CH CHOH ), 7.2–7.5 (5H, m, Ar); 13C-NMR
ꢀ
¼
ꢀ
ꢀ
(CDCl3): d¼67.1 (-CHOH), 126.6, 127.9, 128.4, 128.9, 135.5;
Synthesis of rac-2-Cyclohexenyl-2-hydroxyethanoic
Acid (rac-21)
¼
ꢀ
¼
134.7 (Ar-CH CH ), 176.7 (C O).
(S)-(Z)-2-Acetoxy-4-phenyl-3-butenoic acid: rac-(Z)-2-Hy-
droxy-4-phenyl-3-butenoic acid (10 mg, 0.056 mmol) was dis-
solved in diisopropyl ether (1.3 mL) before 260 mL vinyl ace-
tate and lipase from Candida rugosa (11 mg, Amano AY-30)
were added. The suspension was shaken at 130 rpm at 288C
in Eppendorf tubes (1.5 mL). Samples were taken regularly
to measure ee and conversion. The (Z)-isomer proved to be a
very bad substrate for the lipase-mediated kinetic resolution.
After 28 days the lipase was filtered off, the organic solvent
Starting from 1-cyclohexene-1-carboxaldehyde (0.59 g,
5.4 mmol) the synthesis was performed as described for rac-
(E)-2-hydroxy-3-pentenoic acid (rac-22) to afford rac-2-cyclo-
hexenyl-2-hydroxyethanoic acid (rac-21); overall yield: 0.46 g
(75.4%). 1H NMR (CDCl3): d¼1.53 (4H, m), 2.0 (4H, m), 4.3
(1H, s, CHOH-), 5.7 (1H); 13C-NMR (CDCl3): d¼21.8, 21.9,
23.6, 24.4 (aliphatic C, cyclohexenyl), 74.5 (CHOH), 124.3,
¼
¼
135.9 (C C, cyclohexenyl), 173.9 (C O).
Adv. Synth. Catal. 2005, 347, 951–961
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
959

Ulfried Felfer et al.
REVIEWS
Synthesis of (R)-2-Cyclohexenyl-2-hydroxyethanoic
Acid [(R)-21] Starting from rac-21 using Lipase
“Amano P”
References and Notes
[1] E. L. Eliel, S. H. Wilen, M. P. Doyle, Basic Organic Ster-
eochemistry, Wiley, New York, 2001, pp. 277–278.
[2] E. J. Ebbers, G. J. A. Ariaans, J. P. M. Houbiers, A. Brug-
gink, B. Zwanenburg, Tetrahedron 1997, 53, 9417–9476.
[3] K. Faber, Chem. Eur. J. 2001, 7, 5004–5010.
[4] B. Schnell, K. Faber, W. Kroutil, Adv. Synth. Catal. 2003,
345, 563–6.
[5] E. Adams, Adv. Enzymol. Relat. Areas Mol. Biol. 1976,
44, 69–138.
[6] The first “bottleneck” enzyme of the mandelate pathway
is the (S)-l-mandelate dehydrogenase, see: A. Y. Tsou,
S. C. Ransom, J. A. Gerlt, Biochemistry 1990, 29, 9856–
9862.
rac-2-Cyclohexenyl-2-hydroxyethanoic acid (rac-21; 0.398 g,
2.55 mmol), was dissolved in diisopropyl ether (50 mL), vinyl
acetate (10 mL, 108 mmol) and Lipase “Amano P” (0.5 g)
were added and the mixture was shaken at room temperature
at 180 rpm for 2 days and afterwards again for 2 days at 408C.
The lipase was removed by filtration and the organic solvent
was removed by evaporation under reduced pressure. (R)-2-
Cyclohexenyl-2-hydroxyethanoic acid [(R)-21] was purified
by silica gel chromatography (petroleum ether/ethyl
acetate¼1:1 with 50 mL AcOH per liter solvent) giving the
enantiopure product; yield: 0.179 g (corresponding to 90% of
50% theoretical yield); ee>99%; [a]2D0: ꢀ68 (c 0.2, H2O).
The absolute configuration was deduced as described for
(S)-(Z)-2-hydroxy-4-phenyl-3-butenoic acid but as third indi-
cator the elution order on chiral GC.
[7] C. A. Fewson, FEMS Microbiol. Rev. 1988, 54, 85–110.
[8] G. D. Hegeman, J. Bacteriol. 1966, 91, 1140–1154.
[9] G. L. Kenyon, G. D. Hegeman, Adv. Enzymol. Relat.
Areas Mol. Biol. 1979, 50, 325–360.
[10] D. J. Neidhart, P. L. Howell, G. A. Petsko, V. M. Powers,
R. Li, G. L. Kenyon, J. A. Gerlt, Biochemistry 1991, 30,
9264–9273.
[11] G. L. Kenyon, J. A. Gerlt, G. A. Petsko, J. W. Kozarich,
Acc. Chem. Res. 1995, 28, 178–186.
[12] J. A. Gerlt, G. L. Kenyon, J. W. Kozarich, D. J. Neidhart,
G. A. Petsko, V. M. Powers, Curr. Opinion Struct. Biol.
1992, 2, 736–742.
[13] H. Stecher, U. Felfer, K. Faber, J. Biotechnol. 1997, 56,
33–40.
GC Analysis
rac-(E)-2-Hydroxy-3-pentenoic acid (rac-22) was analyzed as
the corresponding methyl ester; column: Chrompack Chirasil
Dex (25 mꢂ0.32 mmꢂ0.25 mm, 0.75 bar N2), isotherm
1008C, 3.3 min (R), 3.8 min (S).
rac-2-Hydroxy-pentanoic acid (rac-26) was analyzed as the
corresponding methyl ester; column: Chrompack Chirasil
Dex (25 mꢂ0.32 mmꢂ0.25 mm, 1.0 bar N2), isotherm 1008C,
2.3 min (R), 2.6 min (S).
rac-2-Cyclohexenyl-2-hydroxyethanoic acid (rac-21) was
analyzed as the corresponding methyl ester; column: Chiraldex
B-TA, Astec (30 mꢂ0.25 mm, 0.55 bar H2), isotherm 1508C,
34.4 min (R), 37.0 min (S).
[14] A. Y. Tsou, S. C. Ransom, J. A. Gerlt, V. M. Powers,
G. L. Kenyon, Biochemistry 1989, 28, 969–975.
[15] U. T. Strauss, A. Kandelbauer, K. Faber, Biotechnol. Lett.
2000, 22, 515–520.
[16] T. R. Sharp, G. D. Hegeman, G. L. Kenyon, Anal. Bio-
chem. 1979, 94, 329–334.
[17] H. Stecher, A. Hermetter, K. Faber, Biotechnol. Tech.
1998, 12, 257–261.
[18] W. N. Lipscomb, Acc. Chem. Res. 1982, 15, 232.
[19] J. Kraut, Science 1988, 242, 533–540.
[20] F. M. Menger, Acc. Chem. Res. 1993, 26, 206–212.
[21] A. T. Kallarakal, B. Mitra, J. W. Kozarich, J. A. Gerlt,
J. G. Clifton, G. A. Patsko, G. L. Kenyon, Biochemistry
1995, 34, 2788–2797.
[22] B. Mitra, A. T. Kallarakal, J. W. Kozarich, J. A. Gerlt,
J. G. Clifton, G. A. Petsko, G. L. Kenyon, Biochemistry
1995, 34, 2777–2787.
[23] The pKa value of the a-H of mandelic acid was reported
to be ca. 22, of mandelate anion ca. 29; see: J. A. Gerlt,
J. W. Kozarich, G. L. Kenyon, P. G. Gassman, J. Am.
Chem. Soc. 1991, 113, 9667–9669.
HPLC analysis
Jasco HPLC system (pumps PU-980, multi-wavelength-detec-
tor MD-910, autosampler AS-950, degasser CMA/260), sol-
vent: n-heptane/2-propanol/TFA¼90:10:1, flow 0.65 mL/
min.
rac-(Z)-2-Hydroxy-4-phenyl-3-butenoic acid (rac-16), Chir-
acel OD-H (Daicel, Ø 0.46ꢂ25 cm), 11.4 min (R), 14.8 min
(S).
rac-(E)-2-Hydroxy-4-phenyl-3-butenoic acid (rac-15), Chir-
alpak AD (Daicel, Ø 0.46 cmꢂ25 cm), 16.8 min (R), 20.8 min
(S).
rac-2-Hydroxy-4-phenylbutyric acid (rac-17), Chiralpak
AD (Daicel, Ø 0.46 cmꢂ25 cm), 14.9 min (R), 17.6 min (S).
[24] G. L. Kenyon, G. D. Hegeman, Biochemistry 1970, 9,
4036–4043.
Acknowledgements
[25] Mg2þ may be replaced by Mn2þ, Fe2þ, Co2þ or Ni2þ; see:
J. A. Fee, G. D. Hegeman, G. L. Kenyon, Biochemistry
1974, 13, 2528–2532.
This study was performed within the Spezialforschungsbereich
“Biokatalyse” and the Competence Center “Applied Biocataly-
sis” in cooperation with BASF AG. Financial support by the
Fonds zur Fçrderung der wissenschaftlichen Forschung, the
FFG (Vienna), the City of Graz and the Province of Styria
and BASF AG (Ludwigshafen) is gratefully acknowledged.
[26] V. M. Powers, C. W. Koo, G. L. Kenyon, J. A. Gerlt, J. W.
Kozarich, Biochemistry 1991, 30, 9255–9263.
[27] E. T. Maggio, G. L. Kenyon, A. S. Mildvan, G. D. He-
geman, Biochemistry 1975, 14, 1131–1139.
960
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 951–961

The Substrate Spectrum of Mandelate Racemase
REVIEWS
[28] M. Goriup, U. T. Strauss, U. Felfer, W. Kroutil, K. Faber,
J. Mol. Catal. B: Enzym. 2001, 15, 207–212.
[29] M. St. Maurice, S. L. Bearne, Biochemistry 2004, 43,
2524–32.
[30] J. Clayden, N. Greeves, S. Warren, P. Wothers, in: Organ-
ic Chemistry, Oxford University Press, Oxford, 2001, pp.
566–568.
[31] G. D. Hegeman, E. Y. Rosenberg, G. L. Kenyon, Bio-
chemistry 1970, 9, 4029–4035.
[32] U. Felfer, U. T. Strauss, W. Kroutil, W. M. F. Fabian, K.
Faber, J. Mol. Catal. B: Enzym. 2001, 15, 213–222.
[33] R. Li, V. M. Powers, J. W. Kozarich, G. L. Kenyon, J. Org.
Chem. 1995, 60, 3347–4451.
[34] D. T. Lin, V. M. Powers, L. J. Reynolds, C. P. Whitman,
J. W. Kozarich, G. L. Kenyon, J. Am. Chem. Soc. 1988,
110, 323–324.
[35] J. A. Landro, G. L. Kenyon, J. W. Kozarich, Bioorg. Med.
Chem. Lett. 1992, 2, 1411–1418.
[36] For an extensive review on substrate- and active-site
models see: K. Hult, K. Faber, in: Encyclopedia of Catal-
ysis, Vol. 1 (Eds.: I. T. Horvath, E. Iglesia, M. T. Klein,
J. A. Lercher, A. J. Russell, E. I. Stiefel), John Wiley &
Sons, New York, 2004, pp. 86–119.
[39] C. Bauer, M. Boy, K. Faber, U. Felfer, H. Voss, J. Mol.
Catal. B: Enzym. 2001, 16, 91–100.
[40] E. A. Merritt, M. E. P. Murphy, Acta Cryst. D 1994, 50,
869–873.
[41] P. J. Kraulis, P. J. Domaille, S. L. Campbell-Burk, T. Van
Aken, E. D. Laue, Biochemistry 1994, 33, 3515–3531.
[42] G. A. Nicholls, K. Sharp, B. Honig, Prot. Struct. Funct.
Genet. 1991, 11, 281–296.
[43] H. Stecher, A. Hermetter, K. Faber, Biotechnol. Tech.
1998, 12, 257–261.
[44] A. M. C. H. van den Nieuwendijk, E. G. J. C. Warmer-
dam, J. Brussee, A. van der Gen, Tetrahedron: Asymme-
try 1995, 6, 801–806.
[45] J. Brussee, W. T. Loos, C. G. Kruse, A. van der Gen, Tet-
rahedron 1990, 46, 979–986.
[46] F. Nerdel, H. Rachel, Chem. Ber. 1956, 89, 671–677.
[47] A. Schummer, H. Yu, H. Simon, Tetrahedron 1991, 47,
9019–9034.
[48] M.-J. Kim, G. M. Whitesides, J. Am. Chem. Soc. 1988,
110, 2959–2964.
[49] M.-J. Chen, K. Narkunan, R.-S. Liu, J. Org. Chem. 1999,
64, 8311–8318.
[50] D. Obrecht, Helv. Chim. Acta 1991, 74, 27–46.
[51] Q. Jia, W. Xie, W. Zhang, A. Janczuk, S. Luo, B. Zhang,
J. P. Cheng, M. B. Ksebati, P. G. Wang, Tetrahedron Lett.
2002, 43, 2339–2342.
[37] U. T. Strauss, K. Faber, Tetrahedron: Asymmetry 1999,
10, 4079–4081.
[38] M. Pogorevc, H. Stecher, K. Faber, Biotechnol. Lett.
2002, 24, 857–60.
[52] R. J. Kazlauskas, A. N. E. Weissfloch, A. T. Rappaport,
L. A. Cuccia, J. Org. Chem. 1991, 56, 2656–2665.
Adv. Synth. Catal. 2005, 347, 951–961
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
961