Asymmetric whole cell biotransformations in biphasic ionic liquid/water-systems by use of recombinant Escherichia coli with intracellular cofactor regeneration

Article abstract of DOI:10.1016/j.tetasy.2007.08.003

Ionic liquids such as [BMIM][PF₆] and [BMIM][NTF] are already known as good alternatives to organic solvents in biphasic biotransformation. Herein, we report about a systematic procedure based on physical properties to identify more commercially available ionic liquids exhibiting the potential to improve the efficiency of whole cell biocatalyses. This approach resulted in the identification of seven other water immiscible ionic liquids. These ionic liquids were rated by their biocompatibility, their substrate- and product-specific distribution coefficients and by for example performed asymmetric reductions of several prochiral ketones. With the use of a recombinant Escherichia coli as biocatalyst, overproducing a Lactobacillus brevis alcohol dehydrogenase and a Mycobacterium vaccae N10 formate dehydrogenase for cofactor regeneration, the great potential of asymmetric whole cell biotransformations in biphasic ionic liquid/water-systems were demonstrated in simple batch processes.

Full text of DOI:10.1016/j.tetasy.2007.08.003

Tetrahedron: Asymmetry 18 (2007) 1883–1887
Asymmetric whole cell biotransformations in biphasic ionic
liquid/water-systems by use of recombinant Escherichia coli
with intracellular cofactor regeneration
Stefan Bra¨utigam,a,* Stephanie Bringer-Meyerb and Dirk Weuster-Botza,*
aLehrstuhl fu¨r Bioverfahrenstechnik, Technsiche Universita¨t Mu¨nchen, Boltzmannstrasse 15, 85748 Garching, Germany
bInstitut fu¨r Biotechnologie 1, Forschungszentrum Ju¨lich GmbH, 52425 Ju¨lich, Germany
Received 23 April 2007; accepted 1 August 2007
Abstract—Ionic liquids such as [BMIM][PF6] and [BMIM][NTF] are already known as good alternatives to organic solvents in biphasic
biotransformation. Herein, we report about a systematic procedure based on physical properties to identify more commercially available
ionic liquids exhibiting the potential to improve the efficiency of whole cell biocatalyses. This approach resulted in the identification of
seven other water immiscible ionic liquids. These ionic liquids were rated by their biocompatibility, their substrate- and product-specific
distribution coefficients and by for example performed asymmetric reductions of several prochiral ketones. With the use of a recombinant
Escherichia coli as biocatalyst, overproducing a Lactobacillus brevis alcohol dehydrogenase and a Mycobacterium vaccae N10 formate
dehydrogenase for cofactor regeneration, the great potential of asymmetric whole cell biotransformations in biphasic ionic liquid/
water-systems were demonstrated in simple batch processes.
Ó 2007 Elsevier Ltd. All rights reserved.
1. Introduction
approach for the identification of IL’s for whole cell bio-
transformation still does not exist.
Due to high product selectivities, the biocatalytic produc-
tion of non-racemic chiral fine chemicals with isolated
enzymes or whole cells has gained importance.1–4
Comparing these two alternatives, the use of whole micro-
bial cells avoids the cost-intensive enzyme-purification and
profits from intracellular cofactor regeneration, rendering
redundant the supplementary addition of cofactors
(NAD(P)H) and cofactor regeneration enzymes.5,6 In the
case of low water solubility or high toxicity of substrate
and product a biphasic process design is often applied, in
which an additional organic solvent functions as a sub-
strate reservoir and an in situ-extractant.7,8 However, or-
ganic solvents are often toxic to the whole cell
biocatalyst, as well.9 To avoid this disadvantage, water
immiscible ionic liquids (IL’s) are recommended as inter-
esting alternatives.10–15 Criteria for the rating of ionic liq-
uids have already been proposed,15 but a systematic
Regarding the specific biocatalyst performance, biotechno-
logically designed Escherichia coli strains overproducing a
desired enzyme are superior to the corresponding wild-
types. A further improvement of biocatalytic activity can
be achieved by implementing an additional suitable intra-
cellular cofactor regeneration system.16–19 To determine
the possible negative impacts on a biocatalyst with cofactor
regeneration, we applied a recombinant E. coli coexpress-
ing a Lactobacillus brevis alcohol dehydrogenase gene
(adhL. brevis) for asymmetric syntheses and a Mycobacterium
vaccae N10 formate dehydrogenase gene (fdhM. vaccae) for
NADH regeneration with formate.20
The performance of a recombinant E. coli whole cell
biocatalyst with cofactor regeneration in combination
with a biphasic process design using different ionic liquids
as the second liquid phase was evaluated for the asym-
metric reductions of 4-chloroacetophenone (4-Cl-AP) to
(R)-1-(4-chlorophenyl)ethanol [(R)-4-Cl-PE], of ethyl
4-chloroacetoacetate (4-Cl-ACE) to ethyl (S)-4-chloro-3-
hydroxybutyrate [(S)-4-Cl-HBE] and of phenacyl chloride
*
Corresponding authors. Tel.: +49 89 28915712; fax: +49 89 28915714
(S.B.); e-mail addresses: S.Braeutigam@lrz.tu-muenchen.de; D.Weuster-
Botz@lrz.tu-muenchen.de
0957-4166/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2007.08.003

1884
S. Bra¨utigam et al. / Tetrahedron: Asymmetry 18 (2007) 1883–1887
(a-Cl-AP) to (S)-a-chloro-1-phenylethanol [(S)-a-Cl-PE],
respectively.
reaction system in Figure 1b, all the tested ionic liquids
with the exception of [BMPL][E3FAP] and [HMIM][E3-
FAP] agree with this directive. [NTF]-ILs exhibited the
highest product-related distribution coefficients. Further-
more, regarding the first six ionic liquids in Figure 1b, it
was obvious, that ILs with hexyl-cation possessed generally
higher distribution coefficients than the corresponding cat-
ions with a butyl-chain. This effect could be explained by
the increasing surfactant character of ILs with increasing
alkyl chain length.21
2. Results and discussion
In order to identify IL’s suitable for biphasic biotransfor-
mation, the total number of commercially available ionic
liquids was narrowed down by their physical properties.
Obviously, important preconditions for the application as
a second liquid phase in a biphasic process are immiscibil-
ity with water and a melting point below 30 °C. An IL-den-
sity above 1.2 g/cm3 is necessary for a simple and efficient
phase separation after biotransformation. The viscosity
of ionic liquids saturated with water should be below
400 mm2/s, due to its major impact on dispersion quality
and mass transfer limitations. Finally, regarding these
exclusion criteria, the total number of currently commer-
cially available ionic liquids was reduced to the nine ILs
listed in Table 1.
The results for the biocompatibility of ILs and their distri-
bution coefficients implicated, that [PF6]-ILs possessed, in
comparison to the other ionic liquids, the best qualifica-
tions for utilization in biphasic whole cell biotransforma-
tion. However, in exemplarily performed conversions of
600 mM 4-Cl-AP on a 1.4 ml scale, [NTF]-ILs show like-
wise high yields as [PF6]-ILs (Fig. 1c). As this example
indicated, the interactions between IL and biocatalyst
(MI) and between IL and substrate/product (logD) were
not sufficient for an estimation of the best suited IL for
an entire reaction system.
Subsequently, these ionic liquids had to be rated with
respect to their applicability for whole cell biotransforma-
tion. An essential criterion for this purpose was their bio-
compatibility.14,15 Due to the fact, that toxic solvents’
main target is the cell membrane, experiments in continu-
ously stirred 4 ml reactors were performed to determine
the influence of respective ILs on biocatalyst’s membrane
integrity (MI). The results are illustrated in comparison
to a pure aqueous system (Fig. 1a). As the investigations
show, [PF6]-anions affected the cell membrane of E. coli
(fdh,adh) marginally: After a 5 h incubation in a biphasic
system consisting of buffer and 20% (v/v) IL, membrane
integrity decreased only to 70%, as compared to 95% for
the pure aqueous system. In contrast, [NTF]-anions
seemed to be more toxic to the cell membrane. The most
negative effect was measured for the ILs with [E3FAP]-an-
ions. The influence of the anions on membrane integrity of
E. coli (fdh,adh) were neither counteracted nor amplified by
the concomitant cations within the observation variance.
Due to the unsatisfactory conversion results, [E3FAP]-ILs
seemed to be unsuitable for the biphasic biotransformation
under study. The remaining six ionic liquids show good
performances in simple batch processes. Hexyl-ILs resulted
in slightly improved conversions, compared to the corre-
sponding butyl-ILs. The enantiomeric excess for the reac-
tions with ionic liquids was, in all cases, P99.5%. Thus
the ee was improved, compared to the aqueous system
(ee = 96%). The major advantage of a biphasic process
design with the selected ionic liquids was the much higher
chemical yield in the batch processes with E. coli (fdh,adh)
(ꢀ60% compared to 8%).
The toxicity of substrate and product can have a significant
influence on chemical yield. Therefore, the results obtained
with 4-Cl-AP were also verified for other reaction systems,
for example, 4-Cl-ACE and a-Cl-AP, as shown in Table 2.
However, [E3FAP]-ILs were neglected in this investiga-
tions due to the negative results for their biocompatibilities,
product-related distribution coefficients, and the so far per-
formed biocatalytic conversions of 4-Cl-AP. The identified
distribution coefficients and chemical yields for the other
six ionic liquids show that hexyl-ILs were better qualified
than the corresponding ILs with butyl-cations. Further-
more, for the 4-Cl-ACE reaction system, it could be stated,
that the target of a nearly complete conversion was not
Another criterion for the rating of ILs, regarding their abil-
ity for biotransformation, is the decadic logarithm of the
distribution coefficient between the IL and aqueous phase
for substrate and product. The objective of providing sub-
strate concentrations in the ionic liquid of several 100 mM,
without exceeding a toxic substrate- and product-concen-
tration in the aqueous phase, states the demand of a logD
above 2.0.10–13 As displayed for the 4-Cl-AP/(R)-4-Cl-PE
Table 1. Potential ionic liquids suitable for whole cell biotransformation in biphasic systems
IL-abbreviation
IL-name
[BMIM][PF6]
1-Butyl-3-methylimidazolium hexafluorophosphate
[HMIM][PF6]
1-Hexyl-3-methylimidazolium hexafluorophosphate
[BMIM][NTF]
[HMIM][NTF]
[BMPL][NTF]
[HMPL][NTF]
[BMPL][E3FAP]
[HMIM][E3FAP]
[EWTMG][E3FAP]
1-Butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide
1-Hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide
1-Butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)imide
1-Hexyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)imide
1-Butyl-1-methylpyrrolidinium tris(pentafluoroethyl)trifluorophosphate
1-Hexyl-3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate
N,N,N0,N0-Tetramethyl-N00-ethylguanidinium tris(pentafluoroethyl)trifluorophosphate

S. Bra¨utigam et al. / Tetrahedron: Asymmetry 18 (2007) 1883–1887
1885
bution coefficients, but, most probably as a result of a
kinetic limitation, the chemical yields were reduced. How-
ever, nearly complete conversion of a-Cl-AP was obtained
in the following 60 min (results not shown). Another indi-
cation for a kinetic limitation could be seen in the fact that
[PF6]-ILs offered in these cases the highest conversions,
although distribution coefficients recommended [BMIM]-
and [HMIM][NTF] for biphasic biotransformation.
3. Conclusion
In conclusion, we have demonstrated that a proceeding
analogous to the one displayed in Figure 2 is an appropri-
ate strategy for the identification and rating of ionic liquids
suitable for biphasic whole cell biotransformation. In com-
bination with a recombinant E. coli biocatalyst, and on the
basis of three reaction systems, we have shown for the first
time that [HMIM][PF6] and -[NTF], as well as [BMPL]-
and [HMPL][NTF], could improve the efficiency of whole
cell biocatalyses in a similar way as [BMIM][PF6] and
-[NTF]. These results also show that ionic liquids do not
negatively affect the intracellular NADH-regeneration
ensured by formate dehydrogenase. Accordingly, in com-
parison to the aqueous system, space–time-yield as well
as chemical yield of, for example, (S)-4-Cl-HBE were
increased 13-fold to 20 g lꢁ1 hꢁ1, and 99% conversion,
respectively. Moreover, using ionic liquids as a second
phase slightly improved the enantiomeric excess to 99.7%,
compared to 99.1% without IL.
4. Experimental
E. coli BL21 (DE3) (pBBR1MCS2-fdhM. vaccae; pBtac-
adhL. brevis), designated in this work E. coli (fdh,adh), was
constructed as described in the literature.20 Precultures of
E. coli (fdh,adh) were grown overnight in shaking flasks
without baffles filled with 20% complex medium containing
10 g lꢁ1 yeast extract, 5 g lꢁ1 peptone, 5 g lꢁ1 NaCl, 6 g lꢁ1
glucoseÆH2O and 50 mg lꢁ1 kanamycin and carbenicillin,
respectively, at 30 °C, 250 rpm and 5 cm excentricity. Cells
were concentrated by centrifugation at OD600 ꢀ 2 and were
used for inoculation of 4 l medium (same medium as
described above, but containing 15 g lꢁ1 instead of 6 g lꢁ1
glucoseÆH2O) in a 7.5 Labfors stirred tank reactor (Infors,
Switzerland). Batch cultivations were performed. After glu-
cose was completely depleted, 0.7 mM IPTG was added.
Cells were harvested by centrifugation (4.500g, 20 min,
4 °C) when measurements of the in vitro activities of
ADH and FDH at time intervals show constant values.
Afterwards, the biocatalyst was stored at 4 °C.
Figure 1. Criteria for the rating of ionic liquids: (a) membrane integrity
MI of E. coli (fdh,adh) after 5 h incubation in a biphasic system with 20%
(v/v) IL in comparison to buffer; (b) logarithm of distribution coefficients
for 4-Cl-AP and 4-Cl-PE between IL and buffer; (c) chemical yield Y of
(R)-4-Cl-PE after 1 h biotransformation in a biphasic system with 20% IL
and 600 mM substrate (related to IL volume) at a cell density of
ꢁ1
To determine the specific ADH- and FDH-activity cell sus-
pension was diluted to an OD600 of ꢀ1 and disrupted by
adding 10% (v/v) PopCulture (Novagen). The activity of
alcohol dehydrogenase was determined photometrically in
an assay mixture of 20 ll acetophenone solution
(50 mM), 20 ll NADH solution (6.4 mg mlꢁ1), 140 ll
potassium phosphate buffer (100 mM, pH 7.0) and 20 ll
disrupted cells. The assay for formate dehydrogenase con-
tained 20 ll sodium formate solution (800 mM), 20 ll
50 g
l
in comparison to a biotransformation without IL.
CDW
affected negatively by the relatively low distribution coeffi-
cients. [BMPL]- and [HMPL][NTF] seemed to be particu-
larly suitable due to comparatively high distribution
coefficients, which resulted in the highest chemical yield
for (S)-4-Cl-HBE. In comparison to that, the phenacyl
chloride reaction system offered significantly higher distri-

1886
S. Bra¨utigam et al. / Tetrahedron: Asymmetry 18 (2007) 1883–1887
Table 2. Logarithm of distribution coefficient (logD) for 4-Cl-ACE, 4-Cl-HBE, a-Cl-AP and a-Cl-PE between IL and buffer, chemical yield Y after 1 h in
a biphasic system with 20% (v/v) IL and 600 mM substrate for a cell density of 50 gCDW lꢁ1 in comparison to the biotransformations without IL and
enantiomeric excess
Without IL
[BMIM][PF6]
[HMIM][PF6]
[BMIM][NTF]
[HMIM][NTF]
[BMPL][NTF]
[HMPL][NTF]
logD (4-Cl-ACE)
logD (4-Cl-HBE)
Y [%] ((S)-4-Cl-HBE)
ee [%] ((S)-4-Cl-HBE)
—
—
1.92 ( 0.02)
1.03 ( 0.05)
95.8 ( 2.2)
99.7 ( 0.3)
1.97 ( 0.00)
1.12 ( 0.03)
98.2 ( 1.8)
99.6 ( 0.4)
1.88 ( 0.01)
1.14 ( 0.03)
97.1 ( 0.7)
99.7 ( 0.0)a
1.88 ( 0.02)
1.15 ( 0.00)
97.5 ( 1.7)
99.7 ( 0.0)a
1.80 ( 0.01)
1.53 ( 0.02)
99.3 ( 1.0)
99.7 ( 0.0)a
1.81 ( 0.00)
1.55 ( 0.00)
99.1 ( 1.4)
99.6 ( 0.4)
7.5 ( 0.4)
99.1 ( 0.2)
logD (a-Cl-AP)
logD (a-Cl-PE)
Y [%] ((S)-a-Cl-PE)
ee [%] ((S)-a-Cl-PE)
—
—
2.97 ( 0.01)
1.94 ( 0.01)
69.5 (n.b.)b
99.7 ( 0.0)a
3.05 ( 0.02)
2.02 ( 0.03)
89.0 (n.b.)b
99.7 ( 0.0)a
3.01 ( 0.00)
2.05 ( 0.00)
51.1 (n.b.)b
99.7 ( 0.0)a
3.01 ( 0.02)
2.07 ( 0.05)
69.8 (n.b.)b
99.7 ( 0.0)a
2.87 ( 0.03)
1.85 ( 0.01)
66.6 ( 3.9)
99.7 ( 0.0)a
2.93 ( 0.01)
1.88 ( 0.07)
74.2 (n.b.)b
99.7 ( 0.0)a
3.3 ( 0.2)
99.7 ( 0.0)a
a No (R)-enantiomer detectable.
b Not determined.
rated on a chiral BGB-174 column (BGB Analytik AG).
Typical retention times with a flow-rate of 9 ml minꢁ1 he-
lium and a temperature gradient from 75 °C to 140 °C
(2.5 °C minꢁ1) were 20.8 min with 4-Cl-AP, 24.4 min with
a-Cl-AP, 25.2 min with (R)-a-Cl-PE, 25.5 min with (S)-a-
Cl-PE, 26.1 min with (R)-4-Cl-PE and 26.5 min with (S)-
4-Cl-PE. 4-Cl-ACE and 4-Cl-HBE were separated on a chi-
ral Lipodex-E column (Macherey Nagel). Typical retention
times with a flow-rate of 4 ml minꢁ1 helium and a temper-
ature of 105 °C were 4.5 min with 4-Cl-ACE, 6.2 min with
(R)-4-Cl-HBE and 6.7 min with (S)-4-Cl-HBE.
With an initial substrate concentration of 600 mM (related
to IL volume), the biotransformation was performed in
4 ml glass vials equipped with magnetic stirrers by the
use of E. coli (fdh,adh)-cells, resuspended in buffer (0.5 M
KPi, pH 6.5, 1 M sodium formate) exhibitinꢁg1 an in vitro
NAD(H)-specific ADH-activity of 500 U gCDW and an
ꢁ1
in vitro FDH-activity of 180 U g
.
CDW
Figure 2. Criteria pyramid for the identification and rating of ionic liquids
Acknowledgments
concerning their applicability in biphasic whole cell biotransformation.
NAD solution (6.4 mg mlꢁ1), 140 ll potassium phosphate
buffer (100 mM, pH 8.0) and 20 ll disrupted cells. One unit
of enzyme activity (U) was defined as the amount of en-
zyme catalyzing the conversion of 1 lmol acetophenone
per min at 30 °C.
We are grateful to U. Welz-Biermann, Merck KGaA,
Darmstadt, Germany, for the supply of ionic liquids and
N. Esaki, Kyoto University, Japan, for providing the plas-
mid pMcFDH.
Membrane integrity was determined by the use of the via-
bility test kit LIFE/DEAD Baclight (Molecular Probes). In
contrast to the 1 h—biotransformation assays, the incuba-
tion of E. coli (fdh,adh) with 20% ionic liquid was per-
formed for 5 h, under consideration of a prospective
long-term application of the biocatalyst.
References
1. Daußmann, T.; Hennemann, H.-G.; Rosen, T. C.; Dunkel-
¨
mann, P. Chem. Ing. Tech. 2006, 78, 249–255.
2. Liese, A.; Seelbach, K.; Wandrey, C. Industrial Biotransfor-
mations; Wiley-VCH: Weinheim, 2000; pp S.25–S.27.
3. Straathof, A. J. J.; Panke, S.; Schmid, A. Curr. Opin.
Biotechnol. 2002, 13, 548–556.
4. Bertau, M. Curr. Org. Chem. 2002, 6, 987–1014.
5. Faber, K. Biotransformations in Organic Chemistry: A Text-
book; Springer: Berlin, 1997; pp 8–10.
6. Schmid, A.; Dordick, J. S.; Hauer, B.; Kiener, A.; Wubbolts,
M.; Witholt, B. Nature 2001, 409, 258–268.
7. Lye, G. J.; Woodley, J. M. Tibtech 1999, 17, 395–402.
8. Stark, D.; v Stockar, U. In Scheper, T., Ed.; Advances in
Biochemical Engineering/Biotechnology; Springer: Berlin,
Heidelberg, 2003; 80, pp 149–175.
To determine the concentrations of the substrates and
products in aqueous phase and ionic liquid, extractions
were performed with ethylacetate and hexane, respectively.
Extracts were analyzed afterwards with gas chromatogra-
phy (CP-3800, Varian) equipped with a flame ionization
detector (FID). Standards of all pure ketones and all pure
chiral (R)- or (S)-alcohols under study were delivered by
Sigma–Aldrich and Julich Chiral Solutions with purum
grade. 4-Cl-AP, 4-Cl-PE, a-Cl-AP, and a-Cl-PE were sepa-

S. Bra¨utigam et al. / Tetrahedron: Asymmetry 18 (2007) 1883–1887
1887
9. Vermue, M.; Sikkema, J.; Verheul, A.; Bakker, R.; Tramper,
J. Biotechnol. Bioeng. 1993, 42, 747–758.
10. Cull, S. G.; Holbrey, J. D.; Vargas-Mora, V.; Seddon, K. R.;
Lye, G. J. Biotechnol. Bioeng. 2000, 69, 227–233.
11. Howarth, J.; James, P.; Dai, J. Tetrahedron Lett. 2001, 42,
7517–7519.
12. Kragl, U.; Eckstein, M.; Kaftzik, N. In Ionic Liquids in
Synthesis; Wasserscheid, P., Welton, T., Eds.; Wiley-VCH:
Weinheim, 2003; pp 336–346.
16. Kataoka, M.; Kita, K.; Wada, M.; Yasohara, Y.; Hasegawa,
J.; Shimizu, S. Appl. Microbiol. Biotechnol. 2003, 62, 437–445.
17. Kroutil, W.; Mang, H.; Edegger, K.; Faber, K. Curr. Opin.
Chem. Biol. 2004, 8, 120–126.
18. Kizaki, N.; Yasohara, Y.; Hasegawa, J.; Wada, M.; Kataoka,
M.; Shimizu, S. Appl. Microbiol. Biotechnol. 2001, 55, 590–
595.
19. Matsuyama, A.; Yamamoto, H.; Kobayashi, Y. Org. Process
Res. Dev. 2002, 6, 558–561.
13. Matsuda, T.; Yamagishi, Y.; Koguchi, S.; Iwai, N.; Kitaz-
ume, T. Tetrahedron Lett. 2006, 47, 4619–4622.
14. Pfrunder, H.; Amidjojo, M.; Kragl, U.; Weuster-Botz, D.
¨
20. Ernst, M.; Kaup, B.; Muller, M.; Bringer-Meyer, S.; Sahm,
¨
H. Appl. Microbiol. Biotechnol. 2005, 66, 629–634.
21. Ranke, J.; Mo¨lter, K.; Stock, F.; Bottin-Weber, U.; Poczo-
butt, J.; Hoffmann, J.; Ondruschka, B.; Filser, J.; Jastorff, B.
Ecotoxicol. Environ. Saf. 2004, 58, 396–404.
Angew. Chem. 2004, 116, 4629–4631; Angew. Chem., Int. Ed.
2004, 43, 4529–4531.
15. Pfruender, H.; Jones, R.; Weuster-Botz, D. J. Biotechnol.
2006, 124, 182–190.