Identification of Candida tenuis xylose reductase as highly selective biocatalyst for the synthesis of aromatic α-hydroxy esters and improvement of its efficiency by protein engineering

Article abstract of DOI:10.1039/b616475g

Wild-type Candida tenuis xylose reductase and two Trp-23 mutants thereof catalyze NADH-dependent reduction of a homologous series of aromatic α-keto esters with absolute pseudo re-face stereoselectivity and broad tolerance for the substituent on the aromatic ring, producing the corresponding R-alcohols in high yield. The Royal Society of Chemistry.

Full text of DOI:10.1039/b616475g

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Identification of Candida tenuis xylose reductase as highly selective
biocatalyst for the synthesis of aromatic a-hydroxy esters and
improvement of its efficiency by protein engineering{
Regina Kratzer and Bernd Nidetzky*
Received (in Cambridge, UK) 10th November 2006, Accepted 15th January 2007
First published as an Advance Article on the web 26th January 2007
DOI: 10.1039/b616475g
Wild-type Candida tenuis xylose reductase and two Trp-23
mutants thereof catalyze NADH-dependent reduction of a
homologous series of aromatic a-keto esters with absolute
pseudo re-face stereoselectivity and broad tolerance for the
substituent on the aromatic ring, producing the corresponding
R-alcohols in high yield.
relationship well characterized. We introduce here xylose reductase
6
from the yeast Candida tenuis (CtXR) as a powerful hydrogen
transfer biocatalyst for the synthesis of aromatic a-hydroxy esters.
7
Considering its physiological function, CtXR was a very unlikely
a-keto ester reductase. However, we show that the enzyme displays
excellent stereochemical selectivity and broad substrate acceptance
during NADH-dependent reduction of 2a, 2b, 2d, 2g, thus
providing a useful combination of properties for organic synthesis.
CtXR belongs to a large superfamily of NAD(P)-dependent
Chiral a-hydroxy esters are valuable target structures for organic
synthesis. Mandelic acid ester and derivatives thereof (1a–1g;
Fig. 1) have attracted special attention because they are bioactive
compounds or key intermediates in the preparation of pharma-
8
oxidoreductases, the aldo-keto reductases, and salient features of
how it recognizes and converts carbonyl substrates have been
6a,9
elucidated at the molecular level (see Scheme 1). The enzyme is
1
ceuticals. Asymmetric reduction of a-keto precursor compounds
(
2
2a–2g; Fig. 1) is a promising route to synthesize them. Enzyme-
therefore tractable by structure guided protein engineering, and we
report on two site-directed mutants of the reductase that showed
up to eightfold enhanced specificity for the target keto ester
substrates. Energy-minimised docking of the open-chain xylose
catalyzed hydrogen transfer often outperforms counterpart
chemo-catalytic reductions of prochiral ketones in terms of
enantioselectivity. It usually proceeds at a reasonable rate and
with low environmental burden. However, the range of substrates
converted and unpredictable dependences of enzyme activity and
selectivity on the substrate structure sometimes limit the scope of
+
into the active site of CtXR bound to NADP revealed a probable
mode of substrate binding. The modeling predicts steric conflicts
between the side chain of Trp-23 and carbonyl group substituents
6a,9
3
reductases for biocatalytic applications.
replacing the aldehydic hydrogen (Scheme 1). The replacement
Although candidate catalysts for a particular transformation
of Trp-23 by phenylalanine and tyrosine caused a relative better
utilization of ketones than xylose, reflected by both a strong
4
5
may be identified from enzyme or genomic libraries by extensive
screening, it is often preferable to use a small set of practical
reductase systems that exhibit complementary patterns of activity
and stereoselectivity and have had their structure–function
Fig. 1 Aromatic a-hydroxy esters and their a-keto precursor compounds
which are substrates of NADH-dependent reduction by Candida tenuis
xylose reductase.
Research Centre Applied Biocatalysis, Petersgasse 14, A-8010 Graz,
Austria c/o Institute of Biotechnology and Biochemical Engineering,
Graz University of Technology, Petersgasse 12/I, A-8010 Graz, Austria.
E-mail: bernd.nidetzky@tugraz.at; Fax: +43 316 873 8434;
Tel: +43 316 873 8400
Scheme 1 Proposed orientation of 2a in the active site of CtXR and
stereochemical course of carbonyl group reduction by enzyme-bound
NADH. Dashed lines show hydrogen bonds or electrostatic interactions.
The indole ring of Trp-23 is close to the reactive carbonyl group of
{
Electronic supplementary information (ESI) available: Source of
chemicals and production of isolated enzymes. Description of steady state
kinetic assays, enzymatic reduction of a-keto esters and stereochemical
analysis of reduction products. See DOI: 10.1039/b616475g
6a
xylose, and its replacement by smaller groups in Tyr or Phe promotes
9
reaction of 2a compared with the natural substrate xylose.
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 1047–1049 | 1047

View Article Online
a
Table 1 Catalytic efficiencies (kcat/K
21 21
to 95%). Note that isolation of products was beyond the scope of
2,4a
m
M
tivities (e.e. %) of wild-type CtXR and Trp-23 mutants thereof for
s
) and stereoselec-
b
this study and can be done according to literature.
W23F and
NADH-dependent reduction of a series of aromatic a-ketoesters. The
corresponding kcat/K
W23Y retain the very high stereoselectivity of the wild-type
enzyme. In contrast to conversions of aromatic a-keto esters which
yielded R-configured alcohols, the opposite enantiomer, S-ethyl
lactate (e.e. . 99.9%), was obtained upon reduction of ethyl
pyruvate (90% conversion). These results suggest a stereochemical
9
m
values for xylose are shown for comparison
Wild-type
k
W23F W23Y
(e.e.)
Substrate
D-xylose
cat/K
m
(e.e.)
kcat/K
m
m
kcat/K (e.e.)
136
2.4
0.8
2
2
2
2
2
2
2
a
a
b
c
d
e
f
269 (.99.9 R)
576 (99.9 R)
53
2211 (.99.9 R)
911 (.99.9 R)
4029 (99.9 R)
369
13835 (.99.9 R)
636 (.99.9 R)
3158 (99.8 R)
186
11056 (.99.9 R)
course of the reaction catalyzed by CtXR (Scheme 1) where the
13
4-pro-R hydride of NADH attacks the carbonyl carbon from its
pseudo re-face, which according to Prelog’s generalization on the
27
21
208
131
128
74
enantioselectivity of reductases is determined by the relative steric
14
requirements of the substituents of the prochiral ketone.
A
c
c
c
g
197 (99.7 n.d. )
955 (99.7 n.d. )
717 (99.7 n.d. )
binding mode of aromatic a-keto esters is thus supported in which
the carbonyl group is positioned above the nicotinamide ring of
NADH and the phenyl ring points towards the substrate binding
pocket of the enzyme (Scheme 1). With the reactive carbonyl
group locked in place by active-site residues, the relatively open
From initial rate data recorded at varied concentration of a-keto
b
ester substrate. The a-keto esters (1.5 or 5 mM) were incubated in
the presence of excess NADH (2.5 or 6 mM) and enzyme (1–50 mM)
in 50 mM potassium phosphate buffer, pH 7.0. HPLC analysis of
the products was performed using authentic standards of the
c
respective alcohol antipodes. A single enantiomeric product was
6a
and flexible structure of the substrate binding site of CtXR
obtained. However, only racemic ethyl 4-chloromandelate was
available as standard. Details of the experimental procedures are
described in the Supporting Information.
underpins acceptance of various substituents at different positions
on the aromatic ring without, however, compromising the
stereoselectivity of reduction. Among the many isolated reductases
1
5
tested for the synthesis of aromatic a-hydroxy esters, this
decrease in efficiency for xylose reduction and an increase in
9
efficiency for ketone reduction.
property of CtXR seems to be shared only by ketoamide reductase
1
6
from Saccharomyces cerevisiae, which like CtXR belongs to the
aldo-keto reductase superfamily, and 7a-hydroxysteroid dehydro-
Table 1 summarizes results of a steady-state kinetic analysis of
NADH-dependent reduction of 2a–2g by wild-type CtXR and
1
7
10
genases from Bacteroides fragilis, a member of the short-chain
dehydrogenase/reductase superfamily. However, because a struc-
tural basis for these two enzymes is lacking, rational engineering of
their substrate specificities, as done here for CtXR, would seem to
be currently out of reach. Studies of lactate dehydrogenase are
relevant in showing the improvement of a biocatalyst for the
W23F and W23Y mutants. Specificity constants (kcat/K
m
) for
reduction of xylose are shown for comparison. (Values of kcat/K
m
for xylose and 2a were determined previously and are taken from
ref. 9). The series of a-keto esters were converted by the three
enzymes with efficiencies that were up to 14000-fold higher than
the corresponding observable kcat/K
9,11
xylose.
m
values for reaction with
1
8
reduction of a-keto acids by protein engineering. However, the
Replacement of Trp-23 by Phe or Tyr significantly
relatively narrow substrate specificity of the lactate dehydrogenase
2e,19
makes mandelate dehydrogenase
enhanced a-keto ester reductase activity of the wild-type. As
shown in Table 1, W23F was the best improved of the two
mutants, exhibiting three- to eightfold increased catalytic efficien-
cies for reduction of the chosen substrates in comparison with
wild-type. In a homologous series of chloro- or cyano-substituted
derivatives of 2a, reduction of 2d and 2g harbouring the
substituent in the para-position on the aromatic ring was strongly
preferred. Structure–activity correlations of kinetic substituent
a more suitable choice for the
conversion of a-keto acids with bulky side chains such as the
aromatic ring.
In conclusion, we report development of an unanticipated
activity of CtXR towards aromatic a-keto esters into a synthetic
application for which well established reductase systems are
lacking. Because substrate binding site plasticity is a characteristic
property of aldo-keto reductases, results for CtXR may promote
the more widespread structure-based systematic examination of
these reductases for synthetic purposes.
m
effects on kcat/K for carbonyl group reduction by CtXR
suggested that the intrinsic chemical reactivity of the substrate,
reflected by the ability to stabilise a partial positive charge on the
9,12
reactive carbon, determined the speed of the enzymatic reaction.
We thank Dr Wolfgang Skranc (DSM, Austria) for providing
some of the substrates used in this study, and Prof. Erich Leitner
and Dr Thomas Purkarthofer for help with the stereochemical
analysis of reduction products. Financial support from the
Austrian Science Funds (P18275-B09 and P15208-B09 to B. N.)
is gratefully acknowledged.
Considering the evidence from our previous studies, the results in
Table 1 indicate that direct resonance effects of the substituent on
the reaction center probably provide most of this electronic
stabilization in the conversion of 2b–2g (Scheme 2).
Table 1 also reports enantiomeric excess (e.e.) values for alcohol
products obtained by enzymatic reduction and measured in
reaction samples at a substrate conversion of 70% or greater (up
Notes and references
1
E. N. Jacobsen, A. Pfaltz and H. Yamamoto, in Comprehensive
Asymmetric Catalysis, Springer, Berlin, Germany, 1999; Vols. I–III.
(a) Y. Sun, X. Wan, J. Wang, Q. Meng, H. Zhang, L. Jiang and
Z. Zhang, Org. Lett., 2005, 7, 5425; (b) K. Sz o¨ ri, M. Sutyinszki,
K. Felf o¨ ldi and M. Bart o´ k, Appl. Catal., A, 2002, 237, 275; (c)
P. V. Ramachandran, S. Pitre and H. C. Brown, J. Org. Chem., 2002,
2
Scheme 2 Resonance structures of 2d. The cationic center is adjacent to
the aromatic ring, so substituents can undergo direct stabilizing resonance
interaction with it, as shown in the resonance hybrid on the right hand side.
67, 5315; (d) W. Zhang and M. Shi, Chem. Commun., 2006, 11, 1218; (e)
H. Gr o¨ ger, Adv. Synth. Catal., 2001, 343, 547.
3 W. Hummel, Adv. Biochem. Eng. Biotechnol., 1997, 58, 145.
This journal is ß The Royal Society of Chemistry 2007
1
048 | Chem. Commun., 2007, 1047–1049

View Article Online
4
5
(a) D. Kalaitzakis, J. D. Rozzell, S. Kambourakis and I. Smonou, Org.
Lett., 2005, 7, 4799; (b) M. Kataoka, K. Kita, M. Wada, Y. Yasohara,
J. Hasegawa and S. Shimizu, Appl. Microbiol. Biotechnol., 2003, 62, 437;
(c) D. Zhu, B. E. Rios, J. D. Rozzell and L. Hua, Tetrahedron:
Asymmetry, 2005, 16, 1541.
to apparent homogeneity by published protocols. See the Supporting
Information for details.
11 The open-chain free aldehyde of xylose is the real substrate of CtXR,
and its proportion in an aqueous solution of the pentose is only 0.02%.
m
The true kcat/K is therefore 5000 times the one measurable.
(a) I. A. Kaluzna, T. Matsuda, A. K. Sewell and J. D. Stewart, J. Am.
Chem. Soc., 2004, 126, 12827; (b) M. Katz, T. Frejd, B. Hahn-H a¨ gerdal
and M. F. Gorwa-Grauslund, Biotechnol. Bioeng., 2003, 84, 573; (c)
M. Katz, B. Hahn-H a¨ gerdal and M. F. Gorwa-Grauslund, Enzyme
Microb. Technol., 2003, 33, 163.
(a) K. L. Kavanagh, M. Klimacek, B. Nidetzky and D. K. Wilson,
Biochemistry, 2002, 41, 8785; (b) W. Neuhauser, D. Haltrich,
K. D. Kulbe and B. Nidetzky, Biochemistry, 1998, 37, 1116.
CtXR catalyzes reduction of the C1 aldehyde of xylose using NADPH
or NADH as the coenzyme, giving xylitol as the product. W. Neuhauser,
D. Haltrich, K. D. Kulbe and B. Nidetzky, Biochem. J., 1997, 326, 683.
(a) J. M. Jez, M. J. Bennett, B. P. Schlegel, M. Lewis and T. M. Penning,
Biochem. J., 1997, 326, 625; (b) D. Hyndman, D. R. Bauman,
V. V. Heredia and T. M. Penning, Chem.-Biol. Interact., 2003, 143–144,
12 P. Mayr and B. Nidetzky, Biochem. J., 2002, 366, 889.
13 B. Nidetzky, P. Mayr, W. Neuhauser and M. Puchberger, Chem.-Biol.
Interact., 2001, 130–132, 583.
14 V. Prelog, Pure Appl. Chem., 1946, 9, 119.
15 Note that our comparison does not include whole-cell bioreductions
using natural microorganisms because the enzyme(s) involved in
promoting the reactions is (are) usually not known.
16 K. Ishihara, H. Yamamoto, K. Mitsuhashi, K. Nishikawa, S. Tsuboi,
H. Tsuji and N. Nakajima, Biosci. Biotechnol. Biochem., 2004, 68, 2306.
17 D. Zhu, J. E. Stearns, M. Ramirez and L. Hua, Tetrahedron, 2006, 62,
4535.
18 (a) M.-J. Kim and G. M. Whitesides, J. Am. Chem. Soc., 1988, 110,
2959; (b) H. M. Wilks, K. W. Hart, R. Freeney, C. R. Dunn,
H. Muirhead, W. N. Chia, D. A. Barstow, T. Atkinson, A. R. Clarkee
and J. J. Holbrook, Science, 1988, 242, 1541; (c) M. A. Luyten, D. Bur,
H. Wynn, W. Parris, M. Gold, J. D. Friesen and J. B. Jones, J. Am.
Chem. Soc., 1989, 111, 6800.
6
7
8
6
21.
R. Kratzer, S. Leitgeb, D. K. Wilson and B. Nidetzky, Biochem. J.,
006, 393, 51.
9
2
10 The mutations W23F and W23Y were introduced using reported
9
methods. Wild-type and mutant enzymes were produced and purified
19 W. Hummel, H. Sch u¨ tte and M.-R. Kula, Appl. Microbiol. Biotechnol.,
1988, 28, 433.
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 1047–1049 | 1049