A study towards efficient L-threonine aldolase-catalyzed enantio- and diastereoselective aldol reactions of glycine with substituted benzaldehydes: Biocatalyst production and process development

Article abstract of DOI:10.1016/j.tetasy.2011.04.016

The development of aldol reactions of glycine with substituted benzaldehydes in the presence of recombinant L-threonine aldolases from Escherichia coli or Saccharomyces cerevisiae, which were obtained with excellent overexpression data, has been carried o

Full text of DOI:10.1016/j.tetasy.2011.04.016

Tetrahedron: Asymmetry 22 (2011) 925–928
Contents lists available at ScienceDirect
Tetrahedron: Asymmetry
journal homepage: www.elsevier.com/locate/tetasy
A study towards efficient L-threonine aldolase-catalyzed enantio-
and diastereoselective aldol reactions of glycine with substituted
benzaldehydes: biocatalyst production and process development
Katrin Baer a, Nina Dückers b, Thorsten Rosenbaum c, Christian Leggewie b, Sabine Simon a,
a,
Marina Kraußer a, Steffen Oßwald d, Werner Hummel c, , Harald Gröger
⇑
⇑
a Department of Chemistry and Pharmacy, University of Erlangen-Nürnberg, Henkestr. 42, 91054 Erlangen, Germany
b evocatal GmbH, Merowinger Platz 1a, 40225 Düsseldorf, Germany
c Institute of Molecular Enzyme Technologie at the Heinrich-Heine-University of Düsseldorf, Research Centre Jülich, Stetternicher Forst, 52426 Jülich, Germany
d Evonik Degussa GmbH, Health & Nutrition, Rodenbacher Chaussee 4, 63457 Hanau-Wolfgang, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
The development of aldol reactions of glycine with substituted benzaldehydes in the presence of recom-
binant L-threonine aldolases from Escherichia coli or Saccharomyces cerevisiae, which were obtained with
excellent overexpression data, has been carried out. When using glycine and ortho-chlorobenzaldehyde, a
high conversion of >95%, an enantioselectivity of >99% ee, and a diastereoselectivity with d.r.(syn/
Received 5 January 2011
Revised 18 April 2011
Accepted 19 April 2011
anti) = 80:20 was obtained for the resulting b-hydroxy
a-amino acid in such a biotransformation. It
should be noted that this enzymatic process can be conducted at an elevated substrate concentration
of 250 mM.
Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction
however, an intensive search is currently ongoing to gain more in-
sight into the capabilities of known and novel threonine aldolases
L
-Threonine aldolases have the unique capability to catalyze
as catalysts for the direct enantio- and diastereoselective addition
of glycine to aldehydes. A major driving force for these studies is
not only the aforementioned straightforward synthetic approach
asymmetric aldol reactions via the formation of b-hydroxy L-a-
amino acids 3 without the need to protect glycine 2, which serves
as a donor molecule in this reaction (Scheme 1).1,2 The direct use of
(non-modified) glycine 2 is an advantage over typical asymmetric
chemo-catalytic aldol reactions which usually require glycine in
the protected form with respect to both the amino and carboxylic
acid functionalities.3 A comparison of the ‘classic’ multi-step ap-
proach with a chemo-catalyst on one hand and a one-step bio-
transformation as described above on the other hand is given in
of using
products 3,1,2 but also the pharmaceutical importance of the prod-
uct class of b-hydroxy
-amino acids.4
L-threonine aldolase-catalyzed reactions to access these
L-a
Herein, we report our preliminary studies on the process devel-
opment studies of aldol reactions of glycine 2 and several substi-
tuted benzaldehydes 1 in the presence of L-threonine aldolases
from Escherichia coli and Saccharomyces cerevisiae. An efficient and
Scheme 1. In addition,
formation of the -carbon center with excellent stereoselectivity,
thus leading to the desired products with a high enantioselectivity,
typically >99% ee.1,2 A drawback of
-threonine aldolase-catalyzed
aldol reactions, however, is the low diastereoselectivity at higher
conversions as a general process feature, which is due to thermo-
dynamic reasons. Often diastereoselectivities do not exceed
d.r. = 65:35, and for aromatic substrates, the highest known d.r.
values are in the range of d.r. = 78:22.1,2 In spite of this limitation,
L
- and
D
-threonine aldolases catalyze the
technically feasible fermentation process for the production of the
L-threonine aldolases from E. coli will be presented as well as bio-
a
catalytic processes with both enzymes, which show excellent
enantioselectivity and reasonable diastereoselectivity when using
ortho-substituted benzaldehydes. Furthermore, by using a recombi-
nant biocatalyst synthetic processes running at attractive substrate
concentrations of aldehydes 1 of up to 250 mM have been realized.
L
2. Results and discussion
At first, we chose as a biocatalyst a recombinant L-threonine
⇑
Corresponding authors. Current address: Faculty of Chemistry, University of
Bielefeld, Universitätsstraße 25, 33615 Bielefeld, Germany (H.G.). Fax: +49 2461 61
2490 (W.H.); +49 9131 85 21165 (H.G.).
aldolase from E. coli, which was developed by the Yamada et al. and
Wong et al.,5 due to its known high specific activity and favorable
expression data. However, there is limited data with regard to its
application range for aldol reactions with aromatic benzaldehydes.
E-mail addresses: w.hummel@fz-juelich.de (W. Hummel), harald.groeger@
uni-bielefeld.de, harald.groeger@chemie.uni-erlangen.de (H. Gröger).
0957-4166/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2011.04.016

926
K. Baer et al. / Tetrahedron: Asymmetry 22 (2011) 925–928
(A) "Classic" multi-step approach using a chemocatalyst:
O
R
H
1. esterification
2. imine
+
cleavage of both
protecting
O
O
OH
OH
R
R
formation
groups
CO2R
CO2H
OH
OR
Ph
O
chiral
NH2
N
N
NH2
Ph
Ph
synthetic
catalyst
-
Ph
Ph
mixture of
Ph
- ROH
L-syn-3 / L-anti-3
(B) One-step biocatalytic approach:
O
O
R
H
+
OH
R
CO2H
OH
L-enantioselective
threonine aldolase,
PLP-dependent
NH2
NH2
mixture of
L-syn-3 / L-anti-3
Scheme 1. Comparison of chemocatalytic versus biocatalytic approaches toward b-hydroxy L-a-amino acids L-3.
The isolation and cloning of the gene (ltaE) of this enzyme in E. coli
BL21(DE3) led to an excellent overexpression in shaking flask
experiments. This is underlined by the corresponding SDS–PAGE
gel shown in Figure 1. A volumetric activity of the crude protein
types of aromatic aldehydes. Initially, we were in particular inter-
ested in studying the influence of the position of the substituents
on the aromatic moiety on the diastereoselective course of the
aldol reaction. This study was carried out with the different
regioisomers of chloro-substituted benzaldehyde (Scheme 2), and
the experiments were conducted at a synthetically useful substrate
concentration of 100 mM of aldehydes 1a–c. When using both
meta- and para-chlorobenzaldehydes 1b and 1c, the diastereose-
lectivity was in the non-satisfactory range of only d.r.(syn/
anti) = 67:33 and d.r.(syn/anti) = 66:34 at conversions of 64% and
26%, respectively. However, we observed a significant increase in
the diastereoselectivity toward d.r.(syn/anti) = 80:20 when using
the ortho-substituted 2-chlorobenzaldehyde 1a as a substrate. In
addition, an excellent conversion of >95% was achieved in this
experiment. The same tendency and high diastereoselectivity in
the case of ortho-chlorobenzaldehyde 1a (2-chlorobenzaldehyde)
extract with recombinant L-TA from E. coli of 35 U/mL was ob-
tained, corresponding to a specific activity of 1.4 U/mg of crude
protein extract. Notably, the corresponding recombinant E. coli
strain also has been successfully produced in high cell density fer-
mentation on a 5 L scale, reaching a volumetric activity of 15 U/mL
and a specific activity of 0.8 U/mg of crude protein extract.
Although the volumetric activity was somewhat lower compared
with shaking flask conditions, this recombinant L-TA from E. coli
still represents an attractive economical and technically feasible
biocatalyst.
As a second biocatalyst, we prepared the L-threonine aldolase
from S. cerevisiae on a shaking flask scale according to a literature
protocol.6 The resulting crude protein extract showed an activity of
was also found when using the recombinant
L-threonine aldolase
25 U/mL for this
mg of crude protein extract (Fig. 1).
Next we used the recombinant -threonine aldolase (
E. coli as a biocatalyst for the aldol reaction of glycine with different
L
-TA, corresponding to a specific activity of 1.25 U/
(L-TA) from S. cerevisiae as a biocatalyst, leading to a high conver-
sion of 92% and are a diastereomeric ratio of d.r.(syn/anti) = 82:18
(Scheme 2).
L
L-TA) from
We next focused on the process development, in particular on a
further increase of the substrate concentration of aldehydes 1. To
the best of our knowledge, threonine aldolase-catalyzed aldol reac-
tions were carried out at substrate concentrations of the aldehyde
1 of 100 mM as the highest.3 Thus, processes running at elevated
substrate concentrations of >100 mM still represent a challenge
and are a prerequisite for the process to be attractive for technical
purposes. During process development of the aldol reaction of gly-
1
2
M
3
4
55 kDa
cine 2 with ortho-chlorobenzaldehyde 1a catalyzed by L-TA from
E. coli, we found that the reaction proceeds efficiently at such a de-
sired elevated substrate concentration of 250 mM, leading to the
S. cerevisiae
E. coli
40 kDa
35 kDa
formation of the desired b-hydroxy L-a-amino acid L-3a with still
excellent conversion of >95% and an unchanged diastereoselectiv-
25 kDa
ity of d.r.(syn/anti) = 80:20 (Scheme 3). In addition, the desired
major syn-diastereomer in the product L-3a was formed with an
excellent enantioselectivity of >99% ee. Thus, compared to the
experiment at a substrate concentration of 100 mM, the conversion
as well as diastereoselectivity remained unchanged. To the best of
our knowledge, this is the highest substrate concentration, which
has been successfully applied so far in threonine aldolase-
catalyzed diastereo- and enantioselective aldol reactions. At a
Figure 1. SDS–PAGE gel of crude protein extracts of the recombinant
L-TA from
E. coli (runs 1 and 2), recombinant -TA from S. cerevisiae (runs 3 and 4), and a
L
protein marker (run M). Runs 1 and 3 refer to the crude protein extracts obtained
prior to induction, runs 2 and 4 refer to the crude protein extracts obtained after
induction of the recombinant strains with IPTG.

K. Baer et al. / Tetrahedron: Asymmetry 22 (2011) 925–928
927
L-threonine aldolase
O
O
OH
*
O
(70 U/mmol),
PLP, pH 8, rt, 18h
OH
OH
+
Cl
Cl
NH2
NH2
mixture of
1a-c
100 mM
2
L-syn-3a-c / L-anti-3a-c
10 equiv.
Synthetic results
OH
O
Cl
OH
*
O
OH
*
O
Cl
*
OH
OH
OH
NH2
L-3b
NH2
NH2
Cl
L-3a
L-3c
L-TA from E. coli:
64% conversion,
d.r.(syn/anti)=67:33
L-TA from E. coli:
>95% conversion,
d.r.(syn/anti)=80:20
L-TA from E. coli:
26% conversion,
d.r.(syn/anti)=66:34
L-TA from S. cerevisiae:
85% conversion,
L-TA from S. cerevisiae:
92% conversion,
L-TA from S. cerevisiae:
30% conversion,
d.r.(syn/anti)=76:24
d.r.(syn/anti)=82:18
d.r.(syn/anti)=75:25
Scheme 2. Influence of the position of the chloro-substituent in 1 on the conversion and diastereoselectivity.
L-threonine aldolase
from E. coli
(44 U/mmol),
Cl
O
O
Cl
OH
*
O
PLP, pH 8, 25 °C, 6h
OH
OH
+
NH2
NH2
1a
250 mM
2
L-3a
>95% conversion,
d.r.(syn/anti)=80:20
>99% ee (syn)
8 equiv.
Scheme 3. Optimized process for the aldol reaction of ortho-chlorobenzaldehyde with glycine catalyzed by L-TA from E. coli.
Table 1
Overview about biotransformations with different types of ortho-substituted benzaldehydes 1a,d–g
L-threonine aldolase
(50 U/mmol),
R
O
O
R
OH
*
O
PLP, pH 8, rt, 18h
OH
OH
+
NH2
NH2
mixture of
1a,d-g
250 mM
2
L-syn-3a,d-g / L-anti-3a,d-g
10 equiv.
Entrya
Product
R
Type of
L
-TAd
Conversione (%)
d.r.f (syn/anti)
1b
2c
3
Cl
Cl
Br
81
93
51
24
69
9
82:18
80:20
73:27
78:22
75:25
n.d.g
L
L
L
L
L
L
-3a
-3a
-3d
-3e
-3f
L
L
L
L
L
L
-TA from E. coli
-TA from S. cerevisiae
-TA from E. coli
-TA from E. coli
-TA from E. coli
-TA from E. coli
4
CH3
OCH3
OH
5
6
-3g
a
The reactions were carried out in analogy to the representative biotransformation protocol given in the Section 4, however, using the reaction conditions indicated in the
reaction scheme shown above; for the derivatization of the products 3 for analytical purpose 1.5 equiv of benzoylchloride were used in these experiments.
b
In this experiment 8 equiv of glycine and a catalytic amount of the
L
-TA of 30 U/mmol were used; the reaction time was 6 h and the reaction was carried out at 25 °C.
c
d
e
f
In this experiment, a catalytic amount of the
L-TA of 70 U/mmol was used.
The enzymes were used in non-purified, recombinant form as a crude extract.
The conversion was determined from the crude product; for the determination of the diastereomeric ratio, see Section 4.
d.r. = diastereomeric ratio; the d.r. values were determined from the crude product; for the determination of the diastereomeric ratio, see Section 4.
n.d. = not determined.
g

928
K. Baer et al. / Tetrahedron: Asymmetry 22 (2011) 925–928
lower catalytic amount of the
saw a decrease in conversion to 81%, but still a high diastereoselec-
tivity of d.r.(syn/anti) = 82:18 was obtained for the desired product
L
-TA from E. coli of 30 U/mmol, we
enzymes. The enzyme activities (in U) were determined according
to the rate of the retro-aldol reaction with -threonine as a sub-
strate (Unit [U] definition: 1U corresponds to the formation of
mol of acetaldehyde from -threonine per minute at 30 °C
according to the previously reported assay.5,6
L
L-3a (Table 1, entry 1). This type of aldol reaction also runs
1
l
L
successfully when using an -TA from S. cerevisiae, leading to -3a
L
L
with a high conversion of 93% and a diastereoselectivity of
d.r.(syn/anti) = 80:20 (entry 2).
4.2. Procedure for the L-TA-catalyzed aldol reaction of glycine
Having demonstrated the suitability of the
L
-TA from E. coli to
with aromatic aldehydes (exemplified for the biotransformation
shown in Scheme 3).
catalyze aldol reactions with ortho-chlorobenzaldehyde 1a with
good diastereoselectivities and excellent conversions at elevated
substrate concentrations, we became interested in studying other
types of ortho-substituted benzaldehydes 1d–g. Again an elevated
substrate concentration of 250 mM was used in all of these exper-
iments. The results of these preliminary studies, which were done
A crude protein extract with a recombinant
[500 l, 44 U/(mmol of 1a)] was added to glycine 2 (1 mmol),
ortho-chlorobenzaldehyde 1a (2-chlorobenzaldehyde, 125 mol)
and pyridoxal-5-phosphate (50 M). The reaction mixture was
L-TA from E. coli
l
l
l
in the presence of
Table 1. Using ortho-bromobenzaldehyde 1d as a substrate led to
the formation of the desired product -3d with 51% conversion
and a diastereomeric ratio of d.r.(syn/anti) = 73:27 (entry 3).
Diastereoselectivities of d.r.(syn/anti) = 78:22 and d.r.(syn/anti) =
75:25 were found when using ortho-methylbenzaldehyde 1e and
ortho-methoxybenzaldehyde 1f, respectively, as substrates (entries
4 and 5). When using ortho-hydroxybenzaldehyde 1g, however, the
reaction proceeds with a low conversion of 9% only (entry 6).
L
-TA from E. coli as a catalyst, are summarized in
stirred at 25 °C for 6 h. For the determination of the conversion
as well as the diastereo- and enantioselectivity, a derivatization
of the reaction mixture was carried out on the basis of a protocol
reported in the literature7 as follows: to the reaction mixture we
L
added an aqueous solution of NaOH (5 M, 700 ll) and benzoylchlo-
ride (1.7 equiv, 1.69 mmol) were added and the resulting mixture
was stirred for 2 h. Next, the pH was adjusted to pH 1 by the addi-
tion of HCl (5 M), and after extraction with ethyl acetate
(3 Â 5 mL), the combined organic layers were dried over MgSO4.
Removal of the solvent gave the resulting crude product, which
was directly analyzed without further purification by 1H NMR
spectroscopy as well as chiral HPLC analysis [Daicel ChiralcelÒ col-
umn OJ-H, eluent: hexane/isopropanol/formic acid (95:5:0.1),
3. Conclusion
In conclusion, a process development of the aldol reactions of
glycine with substituted benzaldehydes in the presence of a re-
combinant L-threonine aldolase from E. coli or S. cerevisiae as a bio-
flow: 0.8 mL minÀ1, 230 nm, retention times: 63.05 min (
3a), 68.07 min ( -anti-3a)].
L-syn-
L
catalyst has been reported in addition to an efficient fermentation
process, which also turned out to be suitable on a large scale, for
production of the L-threonine aldolase from E. coli. This recombi-
Acknowledgments
nant biocatalyst was obtained with excellent overexpression data.
When using glycine and ortho-chlorobenzaldehyde 1a as a model
substrate, a high yield of >95%, an excellent enantioselectivity of
The authors thank the German Federal Environmental Founda-
tion (Deutsche Bundesstiftung Umwelt, DBU) for generous support
within the DBU-network ‘ChemBioTec’ of the funding priority ‘Bio-
technology’ (Project AZ 13217).
>99% ee, and
a sufficient diastereoselectivity with d.r.(syn/
anti) = 80:20 were obtained in the biotransformation. It should be
noted that this process can be conducted at an elevated substrate
concentration of 250 mM, which to the best of our knowledge rep-
resents the highest substrate concentration reported so far for
threonine aldolase-catalyzed aldol reactions. A remaining chal-
lenge to make this route synthetically more attractive for the
diastereo- and enantioselective preparation of ortho-substituted
phenylserines is in particular the further improvement of the
diastereoselectivity.
References
1. Review: Dückers, N.; Baer, K.; Simon, S.; Gröger, H.; Hummel, W. Appl. Microbiol.
Biotechnol. 2010, 88, 409–424.
2. For selected examples, see: (a) Steinreiber, J.; Fesko, K.; Reisinger, C.;
Schürmann, M.; van Assema, F.; Wolberg, M.; Mink, D.; Griengl, H. Tetrahedron
2007, 63, 918–926; (b) Steinreiber, J.; Fesko, K.; Mayer, C.; Reisinger, C.;
Schürmann, M.; Griengl, H. Tetrahedron 2007, 63, 8088–8093; (c) Kimura, T.;
Vassilev, V. P.; Shen, G.-J.; Wong, C.-H. J. Am. Chem. Soc. 1997, 119, 11734–
11742; (d) Gutierrez, M. L.; Garrabou, X.; Agosta, E.; Servi, S.; Parella, T.; Joglar,
J.; Clapés, P. Chem. Eur. J. 2008, 14, 4647–4656.
4. Experimental
3. (a) Gasparski, C. M.; Miller, M. J. Tetrahedron 1991, 47, 5367–5378; (b) Horikawa,
M.; Busch-Petersen, J.; Corey, E. J. Tetrahedron Lett. 1999, 40, 3843–3846; (c) Ooi,
T.; Taniguchi, M.; Kameda, M.; Maruoka, K. Angew. Chem. 2002, 114, 4724–4726;
Angew. Chem., Int. Ed. 2002, 41, 4542–4544.; Overview: (d) Berkessel, A.; Gröger,
H. Asymmetric Organocatalysis; Wiley-VCH: Weinheim, 2005. p 159.
4. Overview: Kleemann, A.; Engel, J.; Kutscher, B.; Reichert, D. Pharmaceutical
Substances, 5th ed.; Thieme: Stuttgart, 2009.
5. (a) Liu, J. Q.; Dairi, T.; Itoh, N.; Kataoka, M.; Shimizu, S.; Yamada, H. Eur. J.
Biochem. 1998, 255, 220–226; (b) Ref.. 2c.
6. Liu, J. Q.; Nagata, S.; Dairi, T.; Misono, H.; Shimizu, S.; Yamada, H. Eur. J. Biochem.
1997, 245, 289–293.
4.1. Procedure for the preparation of crude protein extracts
containing a recombinant
respectively.
L-TA from E. coli and S. cerevisiae,
The preparation of cell-free crude protein extracts, containing
recombinant threonine aldolases from E. coli (ltaE) and from Sac-
charomyces cerevisiae (GLY1) in overexpressed form, was carried
out according to the procedure described in the literature,5a,6
respectively, followed by a subsequent sterile filtration step. A host
organism E. coli (BL21(DE3)) was used for overexpression of both
7. Shiraiwa, T.; Saijoh, R.; Suzuki, M.; Yoshida, K.; Nishimura, S.; Nagasawa, H.
Chem. Pharm. Bull. 2003, 51, 1363–1367.