Effective synthesis of (S)-3,5-bistrifluoromethylphenyl ethanol by asymmetric enzymatic reduction

Article abstract of DOI:10.1016/j.tetasy.2006.01.039

The synthesis of (S)-3,5-bistrifluoromethylphenyl ethanol, a pharmaceutically important alcohol intermediate for the synthesis of NK-1 receptor antagonists, was demonstrated from a ketone via asymmetric enzymatic reduction. The isolated enzyme alcohol dehydrogenase from Rhodococcus erythropolis reduced the poorly water soluble substrate with excellent ee (>99.9%) and good conversion (>98%). The optimized process was demonstrated up to pilot scale using high substrate concentration (390 mM) using a straightforward isolation process achieving excellent isolation yields (>90%) and effective space time yield (100-110 g/L d). Process improvements, demonstrated at preparative scale, increased the substrate concentration to 580 mM achieving a space time yield of 260 g/L d.

Full text of DOI:10.1016/j.tetasy.2006.01.039

Tetrahedron: Asymmetry 17 (2006) 554–559
Effective synthesis of (S)-3,5-bistrifluoromethylphenyl ethanol
by asymmetric enzymatic reduction
David Pollard,a,* Matthew Truppo,a Jennifer Pollard,a Cheng-yi Chenb and Jeffrey Moorea
aBioprocess R&D, Merck Research Laboratories, Merck & Co. Inc. Rahway, NJ 07065, USA
bProcess R&D, Merck Research Laboratories, Merck & Co. Inc. Rahway, NJ 07065, USA
Received 3 November 2005; accepted 7 January 2006
Available online 10 March 2006
Abstract—The synthesis of (S)-3,5-bistrifluoromethylphenyl ethanol, a pharmaceutically important alcohol intermediate for the syn-
thesis of NK-1 receptor antagonists, was demonstrated from a ketone via asymmetric enzymatic reduction. The isolated enzyme
alcohol dehydrogenase from Rhodococcus erythropolis reduced the poorly water soluble substrate with excellent ee (>99.9%) and
good conversion (>98%). The optimized process was demonstrated up to pilot scale using high substrate concentration
(390 mM) using a straightforward isolation process achieving excellent isolation yields (>90%) and effective space time yield
(100–110 g/L d). Process improvements, demonstrated at preparative scale, increased the substrate concentration to 580 mM achiev-
ing a space time yield of 260 g/L d.
Ó 2006 Elsevier Ltd. All rights reserved.
1. Introduction
conditions without the need for ee upgrade, which is
often required for routes using metal catalyzed transfer
hydrogenation. Biocatalytic reductions can be com-
pleted using whole cell systems but are frequently ham-
pered by low productivity, and complicated by the
presence of multiple ketoreductases, which can lower
the selectivity and produce additional impurities.2 The
isolated enzymes provide simple reaction systems with
few reaction components, making them amenable to
the rapid process development required to meet the
aggressive development timelines for clinical phase drug
supply. The use of recombinant technology with high
expression systems along with protein engineering has
made isolated ketoreductases readily available with high
activity.3 Cofactor regeneration is no longer an issue due
to the availability of both glucose dehydrogenase and
formate dehydrogenase to recycle NADPH and NADH,
respectively. The cofactor can be provided by in situ
regeneration using formate dehydrogenase (FDH),
which irreversibly oxidizes the formate to carbon diox-
ide that is easily removed as a gaseous product.4 This
has successfully been demonstrated on a tonne scale
for the production of tert-leucine.5
Efficient routes to enantiomerically pure compounds are
required to meet the increasing complexity of pharma-
ceutical drugs with multiple chiral centers. Optically
active phenylethanol and its derivatives are useful build-
ing blocks for the synthesis of complex molecules as the
alcohol functionality can be easily transformed, without
racemization, into other useful functional groups. The
trifluoromethyl substituents are important to a number
of active research compounds as a useful tool to prevent
degradation of the active compound by human metabo-
lism. For example, both enantiomers of 3,5-bistri-
fluoromethylphenyl ethanol are intermediates for the
synthesis of NK-1 receptor antagonists. (R)-3,5-Bistri-
fluoromethylphenyl ethanol is incorporated into the
orally active NK1 receptor antagonist (EMENDÒ) for
chemotherapy induced emesis,1 while (S)-3,5-bistrifluoro-
methylphenyl ethanol is incorporated into a number of
antagonists currently under clinical evaluation.
The use of commercially available isolated enzymes for
the reduction of ketones can provide efficient and eco-
nomic processes with good yields and excellent enantio-
selectivity (>99%). They generally use mild reaction
This contribution demonstrates the first example of the
effective synthesis of (S)-3,5-bistrifluoromethylphenyl
ethanol by enzymatic reduction. Process development
is presented from the initial ketoreductase screening
success, which identified the alcohol dehydrogenase
*
Corresponding author. E-mail: david_pollard@merck.com
0957-4166/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2006.01.039

D. Pollard et al. / Tetrahedron: Asymmetry 17 (2006) 554–559
555
Rhodococcus erythropolis, through to the effective pilot
plant demonstration. Particular challenges were over-
come including the poor solubility of ketone substrate
in water, which often limits enzymatic reactions to low
substrate concentrations (5–10 mM) and product crys-
tallization, which was hampered by the high solubility
of the alcohol product in most organic solvents.
of ketone 1 was amenable for development into an effec-
tive process. Process development for the pilot plant
demonstration was pursued using formate dehydroge-
nase (FDH) to regenerate NADH.
2.2. Process development using formate dehydrogenase
for cofactor regeneration
2.2.1. Substrate delivery and mass transfer. The poor
solubility of ketone substrates in aqueous buffer systems
can lead to mass transfer issues and limit the reaction to
slow rates and low substrate concentrations. Despite the
poor solubility of ketone 1 in aqueous buffer (<0.05 g/L
(<2 mM)), reactions proceeded to completion provided
that adequate mixing was applied to disperse the hydro-
phobic liquid ketone as a second phase. For this bipha-
sic system, the mixing intensity affected the rate of
reaction by impacting the ketone droplet size. This
dependency of the initial reaction rate to mixing is
shown in Figure 1. A minimum speed was required, in
the case of the lab reactor >600 rpm, before the initial
rate of reaction was independent of agitation. Reac-
tions, which were run below the minimum agitation,
were slow and often led to incomplete conversion. Ade-
quate mixing was especially important as the alcohol
product was a solid at 30 °C and so precipitated out of
solution during the reaction. Once conversions reached
>60% completion, the residual liquid ketone 1, which
formed a second phase in water, became encapsulated
by the precipitated alcohol. This limited the availability
of the ketone to the enzyme causing reduced rates and
incomplete conversion. The rate of reaction was main-
tained by increasing agitation, in the case of the lab reac-
tor this meant P700 rpm (Fig. 1), which minimized the
alcohol solid particle size and the amount of encapsu-
lated ketone.
2. Results and discussion
2.1. Ketoreductase library screen
A library of commercially available ketoreductases were
screened for their ability to reduce ketone 1 (Scheme 1).
An example of some of the screening results is shown in
Table 1 where the alcohol dehydrogenase (ADH) from
R. erythropolis (ADH RE) gave an ee >99% for the
(S)-alcohol 2 followed by the ADH’s of Candida boidinii
and Candida parapsilosis with ee of 94%. The screen also
showed that the (R)-enantiomer can be synthesized by a
number of enzymes including KRED 101 with ee of
99%.
Alcohol dehydrogenase
R.erythropolis
O
OH
pH 7 phosphate buffer
30oC
F3C
F3C
NADH
NAD
Formate
CO2
CF3
1
CF3
2
Formate dehydrogenase
Scheme 1.
Table 1. A sample of the screening data from the ketoreductase library
for the reduction of 3,5-bistrifluoromethyl acetophenone
80
60
40
20
0
Enzyme catalyst
ee (%)
Cofactor
dependency
ADH Rhodococcus erythropolis
ADH Candida parapsilosis
ADH Candida boidinii
KRED 101
KRED 110
KRED 112
KRED 113
ADH Lactobacillus brevis
ADH Lactobacillus kefir
>99 (S)
96 (S)
94 (S)
>99 (R)
98 (R)
97 (R)
96 (R)
87 (R)
83 (R)
NADH
NADH
NADH
NADPH
NADPH
NADPH
NADPH
NADPH
NADPH
0
200
400
600
800
1000
1200
1400
1600
The NADH cofactor dependent alcohol dehydrogenase
isolated from R. erythropolis (ADH RE) is known to
catalyze the conversion of ketones to chiral alcohols,
particularly medium chain alkanones and aromatic
ketones, including acetophenone with halogenated
substituents.6 Preparative scale examples of the aceto-
phenone reductions with 4-chloro and 4-bromo aceto-
phenone have been shown with good yields and
enantiomeric excesses (>95%).7 The relative activity of
ADH RE for ketone 1 was found to be similar to that
of 4-chloroacetophenone indicating that the reduction
Agitation speed (rpm)
Figure 1. Effect of agitation on the initial rate of reaction for 30,50-
bistrifluoromethyl acetophenone (ADH RE 5 KU/L, FDH 4 KU/L,
NAD 1 g/L, 30 °C, pH 7.0, substrate 4 g, 40 mL Multimax reactor).
2.2.2. Optimization of the activity and stability of two
enzyme process. A good understanding of the factors
affecting the activity and stability of both enzymes is
essential for process optimization of isolated enzyme
systems. The pH for the optimum activity of ADH RE

556
D. Pollard et al. / Tetrahedron: Asymmetry 17 (2006) 554–559
was 6.5, whereas the optimum for FDH was closer to
pH 8.0 (Fig. 2). This meant that a compromise is re-
quired for the two enzyme systems to work efficiently.
An operating region (Fig. 2: shaded area) around pH 7
was chosen as the optimum region for both enzymes.
Both ADH RE and FDH showed improved activity at
elevated temperatures (Fig. 2) but the stability of both
enzymes must also be considered. Studies of the enzyme
stability in the buffer showed the deactivation profiles, at
a range of pH and temperatures, followed the first order
kinetic model (r2 > 0.98) enabling half-life times to be
compared. Figure 3 shows the inferior thermostability
of FDH compared to ADH RE. This limits the reaction
operating temperature to around 30 °C when using
FDH for cofactor regeneration. The buffer pH was also
shown to affect enzyme stability as ADH RE was most
stable at pH <7.0 while FDH was most stable at pH
values >6.5. Combining the activity and stability data
showed that the optimum conditions for activity and
stability for the two enzyme system (ADH RE with
FDH) was to be at the temperature of 30–32 °C and
pH between 6.8 and 7.2.
2.2.3. Effect of substrate and enzyme concentration. The
relationship of substrate and enzyme concentrations
must be understood for any enzymatic system. Ketone
reductions with >95% completion were achieved for
substrate concentrations up to 390 mM (100 g/L)
(Fig. 4) with a satisfactory space time yield (STY) of
107 g/L d. Comparison of initial rates showed a linear
increase with substrate concentration, so no substrate
inhibition was observed up to 200 g/L. The 200 g/L
reaction could reach completion if the initial concentra-
tions of ADH RE and FDH were increased by 50%. The
effect of enzyme concentration was also studied and as
expected, the initial rates of reaction were proportional
to the enzyme concentration.
100
80
60
40
20
0
140
200 g/L (65% conv.)
120
100
100 g/L (>95% conv.)
80
60
50g/L (>97% conv.)
40
25g/L (>99% conv.)
10g/L (>99% conv.)
20
0
6.00
6.25
6.50
6.75
7.00
pH
7.25
7.50
7.75
8.00
0
5
10
15
Time (h)
20
25
Figure 2. Effect of temperature and pH on the activity of ADH
RE&FDH. (25 °C: closed circle, 30 °C: open circle, 34 °C: closed
triangle, 38 °C open triangle). One hundred percent relative activity
refers to maximum activity for each enzyme across all temperature and
pH ranges.
Figure 4. Effect of substrate concentration on the reduction of 30,50-
bistrifluoromethyl acetophenone (ADH RE 5 KU/L, FDH 4 KU/L,
NAD 1 g/L, 30 °C, pH 7.2).
2.2.4. Isolation process development. Crystallization of
alcohol 2 proved challenging as an antisolvent could
not be found due to the high solubility of the alcohol
in most organic solvents. Heptane was chosen as the
crystallization solvent due to a steep solubility curve
with temperature. Above 35 °C, the solubility of alcohol
in heptane was >200 g/L, which fell to <5 g/L at 0 °C.
Therefore a simple isolation was developed, which
enabled heptane to be used as a single organic solvent
for the complete isolation process (extraction and crys-
tallization). Hot heptane was used for the extraction of
the alcohol from the aqueous buffer reaction, followed
by a water wash to remove residual enzyme and then
concentration by vacuum distillation. The heptane
extract was cooled to ꢀ10 °C with crystal seeding at
35 °C. This crystallization was found to reject up to
20% residual ketone in the unlikely event that a reaction
did not reach completion. This ketone could then be
recycled back for enzymatic reduction. For the prepara-
tive scale demonstration this isolation procedure gave an
overall yield of 95% with high purity (>99.9% purity).
160
140
120
100
GDH 103
80
60
ADH RE
40
20
FDH 101
0
20
25
30
35
40
45
50
55
Temperature (deg oC)
Figure 3. Stability comparison of alcohol dehydrogenase (R. erythro-
polis), formate dehydrogenase (101), and glucose dehydrogenase (103):
the effect of temperature on the half life t1/2 of the enzyme in 100 mM
phosphate buffer (pH 7).

D. Pollard et al. / Tetrahedron: Asymmetry 17 (2006) 554–559
557
2.2.5. Scale up of the 3,5-bistrifluoromethyl acetophenone
reduction. The optimized process of pH 7 and 30 °C,
was demonstrated with ketone 1 at 100 g/L (390 mM)
and scaled to 1.1 kg with conversion >99%, ee >99.5%
and overall yield after crystallization of 95% with purity
>99%. Similar results were achieved at 25 kg scale with
space time yield (STY) of 100–110 g/L/d (Fig. 5). The
process was robust and able to withstand operating
ranges for the pilot plant of 28–32 °C and pH 6.8–7.2.
The process ran in a standard glass lined pilot scale reac-
tor with the addition of a recycle loop for pH monitor-
ing and subsurface acid addition for pH control. The
initial rate of reaction was similar at all scales indicating
that mixing was sufficiently maintained (Fig. 5).
recombinant fermentation from the bioreduction. The
biocatalytic routes certainly provide superior enantio-
meric excess (>99%) over the conventional chemo cata-
lysis routes, which often require a difficult upgrade
procedure from ee values of 90–92%.1,9 The isolated
enzyme route provides an ease of operation with the
aqueous process compared to the handling issues often
faced for chemo catalysis, which include air and mois-
ture sensitive catalysts and hazardous reagents such as
in the case of borane reduction.
2.3. Further process improvements using glucose
dehydrogenase for cofactor regeneration
After implementation of the ADH RE/FDH process
into the pilot plant, a glucose dehydrogenase became
available (GDH 103, Biocatalytics Inc.) that had
enhanced temperature tolerance and was capable of
regenerating either NADH or NADPH with similar
activity. The thermostability of GDH 103 was signifi-
cantly improved to that of FDH, as shown in Figure
3. This allowed the reaction temperature to be raised
to 45–50 °C above the 30 °C limit set by the process
using FDH. The pH activity optimum for GDH is 6.5,
which matched the optimum for ADH RE. This allowed
both enzymes to work at their optimum pH for activity
compared to the previous process using FDH where
both enzymes had to work at the suboptimal pH of
7.0 (between the two activity optimums of pH 6.5 for
ADH RE and pH 8.0 for FDH). The GDH was also
10-fold more active than FDH, which meant the protein
levels in the reaction were reduced by 10-fold. This made
a cleaner interface during the heptane extraction and
simplified the phase separation between aqueous and
organic solvent for pilot plant operation. These com-
bined effects enabled reactions to be operated at 45 °C
and pH 6.5 (Scheme 2). Preparative scale demonstra-
tions at 100 g/L, 390 mM showed the increased rate of
reaction at 45 °C allowed space time yield improvements
from 100 to 170 g/L d. with similar 99% ee, 96% conver-
sion and overall yield of 93%. A twofold increase of the
concentration of ADH RE allowed the substrate con-
centration to be increased to 580 mM with a space time
yield of 260 g/L d.
ADH RE
O
OH
pH 7 phosphate buffer
30o
C
F3C
F3C
NADH
NAD
CO2
Formate
FDH
CF3
CF3
100
80
60
40
20
0
0
5
10
15
Time (h)
20
25
30
Figure 5. Scale up of the enzymatic reduction of 30,50-bistrifluoro-
methyl acetophenone to 25 kg (3 g: closed circles, 1.1 kg: triangles,
25 kg: open circles) (ADH RE 3KU/L, FDH 4 KU/L, NAD 1 g/L,
30 °C, pH 7.2).
This enzymatic process with the bioconversion stage fol-
lowed by simple isolation using organic solvent extrac-
tion and crystallization was certainly more convenient
and practical to operate than existing technologies. Sim-
ilar space time yields (STY) (60–104 g/L h) to this work
have been shown for continuous acetophenone reduc-
tion.8 The system required a specialized enzyme mem-
brane reactor and solvent extractor connected by
external loop that was far more complicated than the
procedure with isolated enzymes described in this paper.
The reduction of 4-chloroacetophenone by ADH RE,
using cofactor regeneration from FDH, co-expressed
in the same E. coli cell has also been demonstrated.7
However, this intensive process required several steps:
fermentation of the cells, cell separation/concentration,
bioconversion then product isolation. The ability to
use commercially available isolated enzymes conve-
niently separates the issues of enzyme manufacture by
ADH RE
O
OH
pH 6.5 phosphate buffer
45 o
C
F3C
F3C
NADH
NAD
Gluconic acid
Glucose
CF3
Glucose dehydrogenase 103
CF3
Scheme 2.
3. Conclusion
This work successfully demonstrated the first example of
the synthesis of (S)-3,5-bistrifluoromethylphenyl ethanol
up to pilot scale by enzymatic reduction using the iso-
lated alcohol dehydrogenase from R. erythropolis. The

558
D. Pollard et al. / Tetrahedron: Asymmetry 17 (2006) 554–559
enzymatic reduction using in situ cofactor regeneration
was shown to be an efficient and practical alternative
to existing methods achieving superior selectivity of
>99.9%. The reduction using high substrate concentra-
tion, up to 580 mM, was demonstrated without the need
for additional organic co-solvent. The substrate to cata-
lyst molar ratio for substrate 1 was high at 1,000,000:1
and total turnover numbers for NAD estimated to be
>1000. The ability to use the improved glucose dehydro-
genase 103 for cofactor recycling of either NADH or
NADPH allows the same reaction method to be used
to synthesize either enantiomers of alcohol 2 by simply
switching between the ketoreductase enzyme such as
ADH RE or KRED 101. This biocatalytic method pro-
vides general applicability of chiral alcohol production
for robust manufacturing with effective space time yields
using standard chemical processing equipment.
relative activity % compared to a standard ADH RE
solution 2 U/mL (for 40-chloroacetophenone). The activ-
ity of the NAD dependent FDH and GDH 103 was
determined spectrophotometrically by measuring the
reduction of NAD at 340 nm (E340 = 6.22 mMꢀ1 cmꢀ1
)
in the presence of 150 mM sodium formate for FDH or
150 mM glucose for GDH. Activity of the reaction sam-
ples was expressed as relative activity % compared to a
standard FDH solution or GDH solution 2 U/mol (for
sodium formate and glucose, respectively). Studies of
ADH RE, FDH, and GDH stability were carried out at
a range of pH and temperature conditions at 1 mL scale.
Samples (10 lL) were removed during a 48 h period and
activity measured as described above. The main activity
loss was described by the first order deactivation where
the half life (t1/2) for each pH and temperature condition
with both enzymes was measured from the Ln plot of rel-
ative activity against time.
4. Experimental
4.3. Enzymatic reduction of ketone 1 using formate
dehydrogenase for cofactor regeneration
All solvents and reagents including enzymes were pur-
chased commercially. The ADH RE, formate dehydro-
genase 101 and glucose dehydrogenase 103 were from
Biocatalytics Inc. Enzyme activity was measured spec-
trophotometrically using a Biotek powerwave HT plate
reader. Quantification of ketone to alcohol conversion
was completed using reverse phase HPLC with a Zorbax
extend C18 column. The mobile phase was 40% aceto-
nitrile/60% water with 0.1% TFA. The isocratic flow
rate was 0.75 mL/min at 25 °C with absorbance moni-
toring at 265 nm. Chiral analysis for ee determination
was by normal phase HPLC with Chiralcel OD-H
column using 98% hexanes/2% 2-propanol at 1 mL/
min, 25 °C and monitoring at 265 nm.
Development work was completed at the 4 g scale:
phosphate monobasic buffer (50 mM, pH 7.2) was
added to a Multimax reactor (Mettler Toledo) at a vol-
ume of 40 mL with overhead agitation. NAD (1 g/L) at
160 mg was added followed with sodium formate 6 g
and the two enzymes ADH RE (3 KU/L) and FDH
(2.88 KU/L). The specific activity for ADH RE was
35 U/mg and FDH activity was 4.7 U/mg, calculated
using the activity assays outlined in Section 4.2. Reac-
tions were run at 30 °C with automatic pH control at
7.0 using 2 N sulfuric acid. Ketone 1 was added as a
liquid. The same procedure was used for the 1.1 kg
process using a 20 L glass stirred tank reactor. The
reactions were extracted with two half volumes of hot
heptane when the substrate was shown to be >95% con-
verted. The combined heptane extracts were washed
with a 1/4 volume of water to remove remaining en-
zyme protein residues, then concentrated by distillation
to a 200 g/L concentration. For crystallization the solu-
tion was slowly cooled from 45 to 35 °C. Seeding with
alcohol product with 1 g (0.1% gram/gram of product)
was completed at 35 °C, followed by 1 h of aging and
cooled down to ꢀ10 °C. The crystallization procedure
rejected impurities such as residual ketone and final
purity of >99% was obtained. 1H NMR: d 7.85 (s,
2H), 7.80 (s, 1H), 5.05 (qd, J = 6.5, 3.3, 1H), 2.04
(d, J = 3.3, 1H), 1.56 (d, J = 6.5, 3H). 13C NMR: d
148.44, 131.99 (q, J = 33.2), 125.87 (br q, J = 2.8),
123.58 (q, J = 272.6), 121.53 (septet, J = 3.9), 69.31,
25.79.
4.1. Screening protocol
A total of 40 commercially available ketoreductases
(KRED 101–131) from Biocatalytics Inc. and alcohol
dehydrogenases from Julich Fine Chemicals were used
at 2 g/L in 100 mM phosphate buffer (pH 7) containing
1.2 M equiv of NADPH or NADH for the dependency
of the enzyme. Substrate (20 mM) was delivered to the
reaction in 5% v/v toluene. After overnight incubation
at 30 °C, reactions were extracted with three volumes
of acetonitrile for reverse phase HPLC, then dried and
resuspended in methanol for chiral HPLC analysis.
4.2. Enzyme activity and stability
The activity of the NADH dependent ADH RE was
determined spectrophotometrically by measuring the
oxidation of NADH at 340 nm (E340 = 6.22 mMꢀ1 cmꢀ1
)
The same procedure was followed for the 25 kg pilot
plant batch at the 250 L scale with the inclusion of a
recycle loop in order to monitor pH.
in the presence of 2.5 mM 4-chloroacetophenone. Activ-
ity was measured at 25 °C in 96 well plate format (200 lL)
containing 200 mM potassium phosphate buffer at
pH 7.0 with 5 mM NADH. The reaction was started by
adding the assay mixture to the enzyme sample. The
reaction kinetics were monitored over a 2 min period
and one unit of ADH activity was defined as the amount
of enzyme that converted 1 lmol of NADH per minute.
The activity of the reaction samples was expressed as
4.4. Enzymatic reduction of ketone 1 using glucose
dehydrogenase 103 for cofactor regeneration
Development work was completed at the 8 g scale fol-
lowing the same protocol as Section 4.3, except that
GDH 103 replaced FDH at 5 KU/L. The specific activ-

D. Pollard et al. / Tetrahedron: Asymmetry 17 (2006) 554–559
559
3. (a) Zhu, D.; Rios, B. E.; Rozzell, D.; Hua, L. Tetrahedron:
Asymmetry 2005, 16, 1541–1546; (b) Kaluzna, I.; Rozzell,
D.; Kambourakis, S. Tetrahedron: Asymmetry 2005, 16,
3682–3689.
4. Hummel, W. TIBTECH 1999, 17, 487–492.
5. Bommarius, A. S.; Drauz, K.; Hummel, W.; Kula, M.-R.;
Wandrey, C. Biocatalysis 1994, 10, 37–47.
ity for GDH 103 was 65 U/mg calculated using the
activity assay outlined in Section 4.2. The pH was con-
trolled at 6.5 using 2 M sodium hydroxide with temper-
ature at 45 °C.
References
6. Hummel, W.; Abokitse, K.; Drauz, K.; Rollman, C.;
Groger, H. Adv. Synth. Catal. 2003, 345, 153–159.
7. Groger, H.; Hummel, W.; Rollman, C.; Chamouleau, F.;
Husken, H.; Werner, H.; Wunderlich, C.; Abokitse, K.;
Drauz, K.; Buchholz, S. Tetrahedron 2004, 60, 633–640.
8. Kruse, W.; Hummel, W.; Kragl, U. Recl. Trav. Chim. Pays-
Bas 1996, 115, 239–243.
9. Hansen, K. B.; Chilenski, J. R.; Desmond, R.; Devine, P.
N.; Grabowski, E. J. J.; Heid, R.; Kubryk, M.; Mathre, D.
J.; Varsolona, R. Tetrahedron: Asymmetry 2003, 14, 3581–
3587.
1. Brands, K. M. J.; Payack, J. F.; Rosen, J. D.; Nelson, T.
D.; Candelario, A.; Huffman, M. A.; Zhao, M. M.; Li, J.;
Craig, B.; Song, Z. J.; Tschaen, D.; Hansen, K.; Devine, P.
N.; Pye, P. J.; Rossen, K.; Dormer, P. G.; Reamer, R. A.;
Welch, C. J.; Mathre, D. J.; Tsou, N. N.; McNamera, J. M.;
Reider, P. J. J. Am. Chem. Soc. 2003, 125, 2129–2135.
2. Pollard, D. J.; Telari, K.; Lane, J.; Humphrey, G.;
McWilliams, C.; Nidositko, S.; Salmon, P.; Moore, J.
Biotechnol. Bioengng. 2006, 93, 674–686.