Development of a practical, biocatalytic reduction for the manufacture of (s)-licarbazepine using an evolved ketoreductase

Article abstract of DOI:10.1021/op4003483

This contribution describes the development of a ketoreductase enzyme over four rounds of directed evolution and the associated process development that enabled a practical, scalable process to (S)-licarbazepine (eslicarbazepine).

Full text of DOI:10.1021/op4003483

Article
pubs.acs.org/OPRD
Development of a Practical, Biocatalytic Reduction for the
Manufacture of (S)‑Licarbazepine Using an Evolved Ketoreductase
Naga K. Modukuru,† Joly Sukumaran,† Steven J. Collier,† Ann Shu Chan,† Anupam Gohel,†
Gjalt W. Huisman,‡ Raquel Keledjian,‡ Karthik Narayanaswamy,† Scott J. Novick,‡ S. M. Palanivel,†
Derek Smith,† Zhang Wei,† Brian Wong,† Wan Lin Yeo,† and David A. Entwistle*,†,‡
†Codexis Laboratories Singapore, 61 Science Park Drive, The Galen, Science Park II, Singapore
‡Codexis Inc., 200 Penobscot Drive, Redwood City, California 94063, United States
S
* Supporting Information
ABSTRACT: This contribution describes the development of a ketoreductase enzyme over four rounds of directed evolution
and the associated process development that enabled a practical, scalable process to (S)-licarbazepine (eslicarbazepine).
render them unsuitable for practical, large-scale use. However,
biocatalytic reduction does offer the possibility of high yield and
high selectivity similar to the hydrogenation approaches, but by
using sustainably produced enzymes under inherently safer and
greener conditions.
The continuing advances in modern directed-evolution
technologies, such as Codexis’ Code Evolver platform, have
provided rapid and efficient access to improved enzyme variants
such that substandard enzyme performance no longer needs to
be tolerated.9 Enzymes can now be engineered to operate
under a much wider range of nonphysiological, process relevant
conditions and can provide greener processes than traditional
chemistries when the catalytic activity is high and the reaction
workup is straightforward.10 A number of high-volume API
intermediates and APIs can be produced with such evolved
proteins as catalysts which often provide economic advantages
over traditional chemical methods.11
INTRODUCTION
■
Eslicarbazepine acetate (Aptiom in U.S.A.; Exalief in EU) (3) is
approved as a voltage-gated sodium channel inhibitor for the
treatment of epileptic seizures in adults.1 It is an acetate ester
prodrug of (S)-licarbazepine (2), which itself is an active
metabolite of oxcarbazepine (Trileptal) (1) (Figure 1).2
Figure 1. Structural relationships of carbazepine APIs.
Herein we report the biocatalytic reduction of oxcarbazepine
(1) using an isolated ketoreductase (KRED) specifically
evolved for the efficient manufacture of (S)-licarbazepine (2)
in >99% conversion with >95% isolated yield, >98% chemical
purity and >99% ee with a substrate concentration of 100 g/L
(0.39 M) and beyond and enzyme loading of 1% w/w with
respect to substrate (Scheme 1).12 The enzymatic step was
conducted in aqueous triethanolamine buffer using isopropanol
(IPA) as both cosolvent and terminal reductant, and was
successfully demonstrated at a pilot 500 mL scale.
Reported syntheses of eslicarbazepine are summarized in
Table 1 and include classical resolution of diastereomeric
intermediates,1b,3 asymmetric reduction of the ketone in
oxcarbazepine (1) to (S)-licarbazepine (2) via asymmetric
hydrogenation4 and Corey−Bakshi−Shibata catalyzed hydro-
boration followed by esterification.5 Enzymatic processes
involving hydrolytic resolution of racemic eslicarbazepine esters
have also been reported.6 The chemical and biocatalytic
resolution approaches are naturally limited to a maximum
yield of 50% and in general are two-step processes with modest
yields in the recovery of eslicarbazepine in the second step.
Asymmetric transfer hydrogenation is attractive as it can be
high yielding and highly selective, but this has to be weighed
against the dependence on complex ligands that are potentially
air sensitive and the use of expensive rare metals, the prices of
which are market driven and difficult to predict in the long
term.
RESULTS AND DISCUSSION
■
Enzyme Screening and Evolution. Initial screening of the
KRED Codex panel plates showed that the parent wild-type
KRED from Lactobacillus kefir was inactive, but gratifyingly
identified several variants capable of the conversion of
oxcarbazepine to (S)-licarbazepine with high enantioselectivity
(>98% ee) albeit with low substrate loading (∼2 g/L). The top
Biocatalytic asymmetric reductions of oxcarbazepine (1) have
been described using either Saccharomyces cerevisiae7 or Pichia
methanolica8 whole cell systems, and although these processes
provide the desired product in high enantiomeric purity, their
overall poor volumetric productivities and high catalyst loadings
Special Issue: Biocatalysis 14
Received: December 9, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/op4003483 | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Table 1. Synthetic approaches to eslicarbazepine
HPLC
selectivity
(% ee)
ref
3a
synthetic approach
yield (%)
purity (%)
diacetyl tartrate half ester formation and crystallization followed by base hydrolysis
(S)-ibuprofen half ester formation; purification, followed by base hydrolysis
(S)-ibuprofen half ester formation; purification, followed by base hydrolysis
49 (ester) 84 (eslicarbazepine)
23 (ester) 87 (eslicarbazepine)
not given
99.8
96
3b
3c
99.7 (ester)
93 (ester) no yield given
(eslicarbazepine)
92 (ester)
98
3d
acetyl (R)-mandelic half ester formation followed by base hydrolysis
no yield given (ester) 88
(eslicarbazepine)
not given 97.6 (ester)
4a
4b
4c
4d
4e
transfer hydrogenation of oxacarbazepine
transfer hydrogenation of oxacarbazepine
transfer hydrogenation of oxacarbazepine
transfer hydrogenation of oxacarbazepine
transfer hydrogenation of enol acetate
no yield given
not given
99.6
>99
97.8
98.8
99.4
95
74
86
99.4
not given
not given
no yield given
28−94 (of
acetate)
5a, b R-MeCBS catalyzed borane reduction
55
96.6
93.6
5a
naproxen ester followed by base hydrolysis
25 (ester) 50 (eslicarbazepine)
91.8
(ester)
96.7
93.9 (ester)
92.3
5c
6a
D-tartaric acid and phenylboronic acid moderated NaBH4 reduction of oxacarbazepine
70
99.5
99.4
96.4
enzymatic hydrolysis of methoxyacetate ester, removal of undesired alcohol by
hemisuccinate formation; sodium hydroxide hydrolysis of isolated ester
40 (methoxyacetate) no yield
given (eslicarbazepine)
99.99
Scheme 1. Enzymatic synthesis of (S)-licarbazepine
targeted mutations (and variants) that enabled low cofactor
loading (to minimize cost), high IPA concentrations (to
maximize solubility and drive the equilibrium), and high
temperature (to increase reaction rate and improve solubility).
The screening of the enzyme libraries was performed in 96-well
high throughput (HTP) format (HTP assay >600 variants/day)
under conditions mimicking the final process conditions, vide
infra. The DNA of variants that showed improved product
formation compared to control variants was sequenced to
identify the mutations beneficial for the desired process. The
mutations found in the improved variants were collected,
analyzed, and combined in subsequent rounds of evolution.
KRED-mediated reductions are equilibrium reactions that
need to be driven, in this case with IPA as the reductant for
cofactor (NADP+) recycling. A low level of NADP+, 0.05 g/L,
was used in screening from the outset to minimize the use of
expensive cofactor in the process. Initial process development
scoping studies performed using Variant 1 confirmed the
importance of high IPA concentration in shifting the reaction
equilibrium to the desired product and increasing the solubility
of substrate in order to achieve product titers of ∼100 g/L.
Therefore, round 1 of evolution targeted increased IPA
tolerance. In the subsequent rounds of evolution the reaction
parameters were gradually changed so that sufficient activity
performing variants were then tested at small scale under more
challenging conditions (100 g/L oxcarbazepine and 5 g/L
enzyme) and Variant 1 was found to give ∼5−10% conversion.
This activity was insufficient for a practical process for the large-
scale manufacture of (S)-licarbazepine, and Variant 1 was
improved via directed evolution towards a practical and
economic process.
Libraries of variant KREDs were generated and in total
approximately 15,000 KRED variants were screened under
process-like conditions over four rounds of evolution. It is
critical to reflect the desired end process conditions in the
screening assay used for enzyme evolution so that the correct
evolutionary pressure is maintained. The screening therefore
Table 2. Round-to-round HTP screening conditions and performance improvements
evolution variant selected improvement compared to predecessor total # of mutations vs wild type
round
wild type L. kefir
initial Variant 1
from round
(improvement from Variant 1)
(∼# of variants screened)
screening conditions
no activity for wild type
N/A
−
−
a
9 (400)
5 g/L oxcarbazepine, 5% (v/v) DMSO, 20% IPA,
70% (v/v) crude lysate, 25 °C
5 g/L oxcarbazepine, 5% (v/v) DMSO, 70% IPA,
20% (v/v) crude lysate, 40 °C
5 g/L oxcarbazepine, 5% (v/v) DMSO, 70% IPA,
10% (v/v) crude lysate, 40 °C
75 g/L oxcarbazepine, 70% IPA, 10% (v/v) crude
lysate, 50 and 57 °C
screen
a
round 1
round 2
round 3
round 4
Variant 2
Variant 3
Variant 4
1.5 (1.5)
10 (15)
1.3 (20)
1.3 (26)
10 (5800)
13 (1200)
19 (6900)
30(1300)
a
Variant 5 aka
CDX-021
75 g/L oxcarbazepine, 70% IPA, 1% (v/v) acetone,
10% (v/v) crude lysate, 57 and 62 °C
a
Small amount of DMSO was used to prepare a solution of oxcarbazepine to aid the dispensing of the oxcarbazepine into 96 well plates. Not
required in final process.
B
dx.doi.org/10.1021/op4003483 | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Table 3. Round-to-round process parameters and enzyme performance
a
a
b
c
c
parameter
substrate (g/L)
Variant 1 (screening hit)
Variant 2
Variant 3
Variant 4
Variant 5
100
5
100
5
100
3
100
1
100
1
enzyme (g/L)
NADP+ (g/L)
IPA (% v/v)
temp (°C)
0.5
20
25
7.0
99
24
7
0.1
70
45
9.0
>99
24
15
2
0.1
70
0.1
70
0.1
70
45
55
55
pH
9.0
>99
24
9.0
>99
28
9.0
>99
<24
99
ee (% S)
time (h)
conversion (%)
n-fold improvement over Variant 1
99
99
1
18
60
71
a
b
c
Lyophilized enzyme powders on 10 mL scale. Lyophilized enzyme powders on 10 mL scale with nitrogen sweep (acetone removal). Lyophilized
enzyme powders on 500 mL scale with nitrogen sweep (acetone removal).
Figure 2. (a) Effect of temperature on reaction kinetics. (b) Effect of % v/v IPA on reaction kinetics. (c) Effect of buffer pH on reaction kinetics. (d)
Effect of substrate loading on % conversion and g/L product produced. (e) Effect of NADP+ loading on reaction kinetics. (f) 500 mL scale reaction
under optimized conditions.
could be detected in the HTP screening as shown in Table 2. In
the final round of evolution, as acetone is produced as a
stoichiometric byproduct during the reaction, a small amount of
acetone was introduced to the assay to ensure that the enzyme’s
acetone tolerance was not lost. The temperature was also raised
further to instill greater thermostability to the enzyme.
Although only a small 1.3-fold improvement was seen in the
HTP performance in the final round from Variant 4 to Variant
5 (Table 2) and a similar 1.2-fold improvement in the 10-mL-
scale reactions (Table 3), the thermostability of Variant 5 was
significantly improved. Variant 5 retained 83% of its activity
after 16 h incubation challenge at 60 °C compared to 53%
retained activity for Variant 4. This improved thermostability
gives a more robust catalyst that can tolerate short temperature
excursions and localized hot spots during scale up to
manufacturing scale.
factorial design in a DoE study in order to analyze the effects on
reaction rate, process robustness and to define any process
limitations. Thirty-two experiments were conducted including
eight center points, on 10 mL scale. These reactions were
monitored over 30 h and quenched, and the results were used
for DoE analysis. The center points for the study were set at 55
°C, 60% v/v IPA, pH 9.5, and 120 g/L substrate loading, and
the parameters ranged across 45−65 °C, 50−70% v/v IPA and
pH 8.5−10.5. Enzyme and NADP+ loading were kept constant
at 0.25 g/L and 0.1 g/L, respectively.
Effect of Temperature. As expected the reaction initial
rate increased with increasing temperature (Figure 2a).
However, at 65 °C the reaction slowed down after 10 h,
indicating poor stability of either the enzyme or the cofactor
NADP+ (or both) at this temperature. At all pHs in the design
space (with fixed 120 g/L substrate loading and 60% v/v IPA,
Figure 3a) the maximum of the response curve shown in green
indicated that a temperature range of 53−57 °C was
operationally optimal. A temperature of 55 °C was chosen as
the optimum process temperature as the enzyme demonstrated
considerable stability and activity at that temperature.
After each round of screening selected top variants were
produced at gram scale and isolated as a lyophilized powder.
These enzyme powders were then used in a small scale
preparative reactions to confirm enzyme performance (Table
3).
Process Optimization and Robustness Analysis via
DoE. Critical process parameters such as temperature, % v/v
IPA, reaction pH, and substrate loading were stressed using
Effect of isopropanol/buffer ratio. Comparable initial
rates were observed for reactions conducted with 50−70% v/v
IPA (Figure 2b). A decrease in final conversion was observed
C
dx.doi.org/10.1021/op4003483 | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Figure 3. (a) Response surface at 120 g/L substrate and 60% IPA. (b) Response surface at 120 g/L substrate and 55 °C. (c) Response surface at 55
°C and pH 9.5. (d) Response surface at 60% v/v IPA and pH 9.5.
with greater than 60% v/v IPA and was most likely due to
decreased enzyme stability. Analysis across a range of 50−70%
v/v IPA suggested an operationally optimal range between 55−
65% v/v IPA (Figure 3b). The optimum process parameter for
the enzymatic reaction chosen was 60% v/v IPA.
enzyme manufacturing process, enabling the reaction to
progress without additional NADP+. However, for a more
robust process, it is desirable to add NADP+ (between 0.05 and
0.1 g/L), as the amount of intrinsic cofactor within the
manufactured enzyme powder may vary from batch to batch.
Reaction Scale Up. Upon identification of optimal and
robust process conditions (100 g/L substrate, 1 g/L CDX-021,
0.1 g/L NADP+, 60% v/v IPA, pH 10, 55 °C), a scaled up
reaction was performed at 500 mL scale. The reaction reached
equilibrium at approximately 90% conversion and was driven to
completion by the removal of acetone by employing a nitrogen
sweep and the simultaneous addition of a pre-mixed solution of
IPA and buffer to maintain the reaction volume. Upon
completion of the enzymatic reaction, IPA and any remaining
acetone were removed by distillation under reduced pressure.
Solvent replacement of IPA with water triggered crystallization
which upon subsequent filtration afforded the crude desired
product 2 in high yield, purity, and enantiomeric excess (96%
yield, 98.7 area %, >99% ee). A further recrystallization from
methanol improved the purity to 99.6 area % in 90% recovery
and removed low level traces of residual enzyme present in the
crude product.13 A typical reaction kinetic profile is shown in
Figure 2f.
The ACS Green Chemistry Round Table Process Mass
Intensity (PMI) Tool is an open source tool freely available for
the standardization of PMI calculations.14 By using the PMI
tool, the biotransformation and the purification were calculated
as 15.0 and 19.7, respectively. As PMIs for existing reductions
of oxcarbazepine have not been published, direct comparison
has not been possible. However, given the catalytic reaction
mechanism, low molecular mass of the terminal reductant, and
the direct isolation of product from the reaction by distillation,
the reaction PMI can only be significantly lowered further if
Effect of Buffer pH. No significant differences in reaction
rate were observed in the pH range of 8.5−10.5 (Figure 2c). 0.1
M Triethanolamine buffer at pH 10.0 was prepared at room
temperature and used for the enzymatic reaction without any
adjustments upon either addition of substrate or heating.
Effect of Substrate Loading. Although a higher percent
conversion was seen with lower substrate loading (Figure 2d,
solid lines), the absolute amount of product formed remained
the same, (Figure 2d, dotted lines) which indicated little or no
substrate inhibition of the enzyme. It was anticipated that with
efficient acetone removal from the reaction mixture combined
with replenishment of reaction volume, it would be possible to
fully convert at least 100 g/L substrate with 1 g/L enzyme.
Analyses of substrate loading against both the optimum
temperature and optimum IPA loading (55 °C and 60% v/v
IPA, Figure 3c,d) suggested an optimal substrate loading of
110−130 g/L in order to maximize product titer in g/L rather
than % conversion. Eventually a slightly lower substrate loading
of 100 g/L was chosen for the technical transfer in order to
minimize potential mass transfer effects on scale up and ensure
robust and reproducible performance.
Effect of NADP+ (Cofactor) Loading. NADP+ is a
cofactor essential for KRED enzyme activity which was not
included as a variable in the DoE. However, isolated robustness
tests showed no detrimental effect on reaction rate at a NADP+
loading as low as 0.05 g/L (Figure 2e, tested at 100 g/L
substrate loading, 60% IPA, pH 10.0, 55 °C). The
manufactured enzyme powder may contain NADP+ from the
D
dx.doi.org/10.1021/op4003483 | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Table 4. Isolated CDX-021 KRED vs whole cell performance
biotransformation system
substrate (g/L) substrate:catalyst ratio (w/w) conversion (isolated yield) (%) ee (%)
CDX-021 (60%v/v aq IPA, 0.1 g/L NADP+, 55 °C)9
S. cerevisiae CGMC 2267 (5%v/v EtOH, 70 g/L glucose, 30 °C)
P. methanolica 1036608 (1:1 aq:hexane, 30 °C)
100
73
100:1 (KRED)
14:1 (dry cells)
0.01:1 (wet cells)
>99 (96)
>99
10
99.2 (not reported)
>98 (>85)
1.5
>98
IPA recycling is implemented and the heptane wash of the
product cake is optimized or eliminated.15 The PMI of the
current downstream purification method is higher than that of
the biotransformation and isolation due to the relatively high
volume of solvents required for dissolution. Future develop-
ment would be aimed towards optimization of the downstream
processing, including solvent recycling where practical, to
further reduce the overall PMI.
In summary we have developed a practical, scalable isolated
enzyme-mediated process for the production of (S)-licarbaze-
pine, that provides important benefits over previously described
whole cell processes (Table 4) and chemocatalytic processes
(Table 1). The process is currently being assessed for
commercial manufacture of eslicarbazepine.
turbid reaction mixture was drained from the reactor into a 1 L
round-bottom flask. IPA was distilled by rotary evaporation (75
Torr, 50 °C bath). Upon partial concentration of the reaction
by distillation, water (100 mL) was added to the white slurry
and the distillation continued to completely remove IPA. The
crude product was collected by filtration on a Buchner funnel,
washed with water (100 mL) and heptane (200 mL),15 and
dried for 24 h in a vacuum oven (2 mbar) at 30 °C. Crude (S)-
licarbazepine (2) (48.0 g, 96% yield) was obtained as an off
white solid with a chemical purity of 98.7% (HPLC Method 5)
with >99.9% e.e. (HPLC Method 3) and a residual protein
content of 80 ppm.13
Purification of crude (S)-licarbazepine (2). A suspension
of crude 2 from above (10.0 g) in methanol (100 mL) was
heated to 40 °C (internal temperature) to allow maximum
dissolution of product. Celite (2.0 g) was added to the slightly
turbid solution and the mixture was stirred at 40 °C for 15−20
min. The Celite was removed by filtration through a sintered
funnel and the residue was washed with preheated methanol
(20 mL, ∼40 °C). The clear filtrate was then distilled under
reduced pressure to approximately 30 mL volume. The thick
mixture was cooled to 5 °C, cold water (50 mL, 5 °C) was
added over 30 min to the white precipitate, and the resulting
slurry was stirred at 5 °C for a further 30 min. The precipitated
product was filtered through a sintered funnel, rinsed, and
washed with 20 mL water before being dried in a vacuum oven
for 16 h (30 °C, 2 mbar). Purified product was isolated in a
single crop (9.0 g, 90% recovery) as a white solid with 99.6%
chemical purity (HPLC Method 5) and a residual protein
content of <10 ppm.13
CONCLUSIONS
■
A ketoreductase was evolved to enable a commercially
attractive process for a highly efficient biocatalytic reduction
of oxcarbazepine, a first for an isolated enzyme to the best of
our knowledge. The resulting enzymatic process outperforms
the whole cell processes previously reported in terms of
volumetric productivity and downstream processing. This
process affords a key chiral intermediate of eslicarbazepine
acetate in high purity and yield. The enzyme and the process
have been successfully transferred to an API manufacturer and
proven to achieve results similar to those observed in the
laboratory setting.
EXPERIMENTAL SECTION
For analytical Methods 1−5 please refer to the Supporting
Information.
■
ASSOCIATED CONTENT
* Supporting Information
Spectra and analytical methods. This material is available free of
charge via the Internet at http://pubs.acs.org.
■
Preparation of Crude (S)-Licarbazepine (2). Triethanol-
amine buffer (TEoA) solution (0.1 M, pH 10.0) was prepared
in a separate vessel by dissolving triethanolamine (13.3 mL) in
deionized water (1 L) at room temperature. MgSO4·7H2O
(0.25 g) was added to obtain a final magnesium concentration
of 1 mM. The pH of the resulting clear solution was 10.0
( 0.1) at room temperature and was used for reaction without
any pH adjustments. A 1 L jacketed reactor was charged
sequentially with IPA (300 mL), TEoA buffer (190 mL), and
solid oxcarbazepine (50 g). The mixture was stirred at 200 rpm
and heated to reach an internal temperature of 55 °C
whereupon a pale-yellow slurry was obtained with a pH of
8.7. A stock solution of enzyme (0.5 g) and NADP+ (50 mg)
was prepared separately in buffer (10 mL, final pH 8.6), and
was charged to the reaction mixture whereupon the reaction
mixture was stirred at 55 °C and 200 rpm under a nitrogen
atmosphere with a sweep flow rate of 0.8 L min−1. The reaction
volume was maintained by the intermittent addition of a
premixed solution of 60% IPA and 40% buffer (0.1 M TEoA
with 1 mM MgSO4, pH 10.0). The reaction course was
followed by periodically taking samples from the reaction
mixture, quenching, and analyzing as described in Method 1.
Samples were also frequently monitored for acetone content
using the procedure described in Method 4. After in-process
analyses indicated >99% conversion in 24 h, the pale-yellow
S
AUTHOR INFORMATION
Corresponding Author
*E-mail: david.entwistle@codexis.com
■
Notes
The authors declare no competing financial interest.
REFERENCES
■
(1) (a) Benes, J.; Soares Da Silva, P. EP 0751129, 1997. (b) Benes, J.;
Parada, A.; Figueiredo, A. A.; Alves, P. C.; Freitas, A. P.; Learmonth, D.
A.; Cunha, R. A.; Garrett, J.; Soares-da-Silva, P. J. Med. Chem. 1999, 42,
2582−2587. (c) Dulsat, C.; Mealy, N.; Castaner, R.; Bolos, J. Drugs
Future 2009, 34, 189−195.
(2) (a) Schindler, W. DE 2011087, 1970. (b) Mazza, M.; Della
Marca, G.; Di Nicola, M.; Martinotti, G.; Pozzi, G.; Janiri, L.; Bria, P.;
Mazza, S. Epilepsy Behav. 2007, 10, 397−401. (c) Schmidt, D.; Elger,
C. E. Epilepsy Behav. 2004, 5, 627−635. (d) Wellington, K.; Goa, K. L.
CNS Drugs 2001, 15, 137−163.
(3) (a) Learmonth, D. A. WO/2002/092572, 2002. (b) Gharpure,
M. M.; Rane, D.; Zope, S. S.; Narawade, K. B.; Thanedar, A. A. WO/
2012/156987, 2012. (c) Crasta, S. R. F.; Joshi, A. V.; Bhanu, M. N.
WO/2012/121701, 2012. (d) Desai, S. J.; Pandya, A. K.; Sawant, S. P.;
E
dx.doi.org/10.1021/op4003483 | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Mehariya K. R. WO/2011/117885, 2011. (e) Koilpillai, J. P.; Kulkarni,
P. B.; Sawant, S. B.; Limbekar, N. D. WO/2010/113179, 2010.
(4) (a) Mathes, C.; Sedelmeier, G.; Blatter, F.; Pfeffer, S.; Grimler, D.
WO/2004/031155, 2004. (b) Learmonth, D. A.; Grasa, G. A.; Zanotti-
Gerosa, A. WO/2007/012793, 2007. (c) Wisdom, R.; Jung J.; Meudt,
A. EP2383261, 2011. (d) Daqing, C.; Guoliang Z. CN102250005,
2011. (e) For asymmetric hydrogenation of the corresponding enol
acetate see , Yu, B.; Li, W.; Learmonth D. A. WO/2007/117166, 2007.
(5) (a) Katkam, S.; Sagyam, R. R.; Buchikonda, R.; Muttavarapu, M.
M.; Munagala, S.; Chikka, R. WO/2011/091131, 2011. (b) Satyanar-
ayana,, R. M.; Eswaraia, S.; Kondal, R. B.; Venkatesh, M. WO/2011/
138795, 2011. (c) Biswas, S.; Dubey, S. K.; Bansal, V.; Masand, M.;
Vir, D. WO/2012/120356, 2012.
(6) (a) Husain, M.; Datta, D. WO/2011/045648, 2010. (b) Karusala,
N. R.; Husain, M.; Thorpunuri, S.; Datta, D. IN2009CH00220, 2009.
(7) Ou, Z.; Shen, W.; Che, D. CN102465159, 2010.
(8) Singh, M.; Singh, S.; Deshaboina, S.; Krishnen, H.; Lloyd, R.;
Holt-Tiffin, K.; Bhattacharya, A.; Bandichhor, R. Catal. Sci. Technol.
2012, 2, 1602−1605.
(9) Huisman, G. W.; Collier, S. J. Curr. Opin. Chem. Biol. 2013, 17,
284−292 and the references therein..
(10) (a) Huisman, G. W.; Collier, S. J. Development of Green-by-
Design, Practical Biocatalytic Processes. In Scalable Green Chemistry;
Koenig, S. G., Ed.; Pan Stanford Publishing, Singapore, 2013
(b) Huisman, G. W.; Liang, J.; Krebber, A. Curr. Opin. Chem. Biol.
2010, 14, 122−129.
(11) (a) Savile, C. K.; Janey, J. M.; Mundorff, E. C.; Moore, J. C.;
Tam, S.; Jarvis, W. R.; Colbeck, J. C.; Krebber, A.; Fleitz, F. J.; Brands,
J.; Devine, P. N.; Huisman, G. W.; Hughes, G. J. Science 2010, 329,
305−309. (b) Busto, E.; Gotor-Fernandez, V.; Gotor, V. J. Org. Chem.
2012, 77, 4842−4848. (c) Ang, E. L.; Clay, M. D.; Behrouzian, B.;
Eberhard, E.; Collier, S. J.; Fu Fan, J.; Smith, D.; Song, S.; Alvizo, O.;
Widegren, M.; Zhu, J.; Xu, J.; Wilson, R. WO/2012/078800, 2012.
(d) Bong, Y. K.; Clay, M. D.; Collier, S. J.; Mijts, B.; Vogel, M.; Zhang,
X.; Zhu, J.; Nazor, J.; Smith, D.; Song, S. WO/2011/071982, 2011.
(e) Gilson, L.; Collier, S. J.; Sukumaran, J.; Yeo, W. L.; Alviso, O.; Teo,
E. L.; Wilson, R.; Xu, J. WO/2011/041231, 2011. (f) Collier, S. J.;
Teo, E. L.; Sukumaran, J.; Wilson, R. J.; Xu, J. WO/2011/041233,
2011. (g) Xie, X.; Watanabe, K.; Wojcicki, W. A.; Wang, C. C. C.;
Tang, Y. Chem. Biol. 2006, 13, 1161−1169. (h) Ma, S. K.; Gruber, J.;
Davis, C.; Newman, L.; Gray, D.; Wang, A.; Grate, J.; Huisman, G. W.;
Sheldon, R. A. Green Chem. 2010, 12, 81−86. (i) Martinez, C. A.; Hu,
S.; Dumond, Y.; Tao, J.; Kelleher, P.; Tully, L. Org. Process. Res. Dev.
2008, 12, 392−398.
(12) Gohel, A.; Smith, D.; Wong, B.; Sukumaran, J.; Yeo, W. L.;
Collier, S. J. WO/2012/142302, 2012.
(13) Protein levels of 80 ppm and <10 ppm were present in the crude
and purified products, respectively. The level of residual protein was
measured using a commercial UV absorbance-based protein
quantification assay (SPN - Assay commercially available from
GBiosciences). All materials used were those provided with the
commercial kit, and the experimental protocol provided was followed
without deviation. The amount of residual protein was calculated on
the basis of absorbance measurements from a UV spectrophotometer.
(14) (a) For the freely available PMI tool see www.acs.org/content/
dam/acsorg/greenchemistry/industriainnovation/roundtable/process-
mass-intensity-calculation-tool.xls. (b) For other green chemistry tools
and information see ACS Chemistry Institute Pharmaceutical
Roundtable website at www.acs.org/gcipharmaroundtable.
(15) The heptane cake wash was used on this scale to assist more
rapid drying and should be readily removed in future scale-up
development.
F
dx.doi.org/10.1021/op4003483 | Org. Process Res. Dev. XXXX, XXX, XXX−XXX