Development of an Enzymatic Process for the Production of (R)-2-Butyl-2-ethyloxirane

Article abstract of DOI:10.1021/acs.oprd.7b00179

An epoxide resolution process was rapidly developed that allowed access to multigram scale quantities of (R)-2-butyl-2-ethyloxirane 2 at greater than 300 g/L reaction concentration using an easy-to-handle and store lyophilized powder of epoxide hydrolase from Agromyces mediolanus. The enzyme was successfully fermented on a 35 L scale and stability increased by downstream processing. Halohydrin dehalogenases also gave highly enantioselective resolution but were shown to favor hydrolysis of the (R)-2 epoxide, whereas epoxide hydrolase from Aspergillus Niger instead provided (R)-7 via an unoptimized, enantioconvergent process.

Full text of DOI:10.1021/acs.oprd.7b00179

Subscriber access provided by Olson Library | Northern Michigan University
Full Paper
Development of an Enzymatic Process for
the Production of (R)-2-butyl-2-ethyloxirane
Gheorghe-Doru Roiban, Peter W Sutton, Rebecca A Splain, Christopher W.
Morgan, Andrew Fosberry, Katherine Honicker, Paul Homes, cyril boudet, Alison
Dann, Jiasheng Guo, Kristin Brown, Leigh Anne Ihnken, and Douglas Fuerst
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.7b00179 • Publication Date (Web): 30 Jun 2017
Downloaded from http://pubs.acs.org on June 30, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Organic Process Research & Development is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 31
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Development of an enzymatic process for the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
production of (R)ꢀ2ꢀbutylꢀ2ꢀethyloxirane
GheorgheꢀDoru Roiban,*a Peter Sutton,*b† Rebecca Splain,c Christopher Morgan,c Andrew
Fosberry,d Katherine Honicker,e Paul Homes,d Cyril Boudet,f Alison Dann,f Jiasheng Guo,c
Kristin K. Brown,e Leigh Anne F. Ihnken,e Douglas Fuerste
a Synthetic Biochemistry, Advanced Manufacturing Technologies, GlaxoSmithKline, Medicines
Research Centre, Gunnels Wood Road, Stevenage SG1 2NY, United Kingdom
b API Chemistry, GlaxoSmithKline, Medicines Research Centre, Gunnels Wood Road,
Stevenage SG1 2NY, United Kingdom
c API Chemistry, GlaxoSmithKline, 709 Swedeland Road, King of Prussia, 19406, PA, US
d Protein and Cellular Sciences, GlaxoSmithKline, Medicines Research Centre, Gunnels Wood
Road, Stevenage SG1 2NY, United Kingdom
e Synthetic Biochemistry, Advanced Manufacturing Technologies, GlaxoSmithKline, 709
Swedeland Road, King of Prussia, PA, 19406, US
f Biotechnology and Environmental Shared Service, Global Manufacturing and Supply,
GlaxoSmithKline, Dominion Way, Worthing BN14 8PB, United Kingdom
ACS Paragon Plus Environment
1

Organic Process Research & Development
Page 2 of 31
1
2
g Molecular Design, Computational and Modeling Sciences, GlaxoSmithKline, 1250 S.
Collegeville Road, Collegeville, PA, 19426, US
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Abstract
An epoxide resolution process was rapidly developed that allowed access to multiꢀgram scale
quantities of (R)ꢀ2ꢀbutylꢀ2ꢀethyloxirane 2 at greater than 300 g/L reaction concentration using an
easy to handle and store lyophilised powder of epoxide hydrolase from Agromyces mediolanus.
The enzyme was successfully fermented on a 35L scale and stability increased by downstream
processing. Halohydrin dehalogenases also gave highly enantioselective resolution, but were
shown to favour hydrolysis of the (R)ꢀ2 epoxide, whereas epoxide hydrolase from Aspergillus
niger instead provided (R)ꢀ7 via an unoptimized, enantioconvergent process.
Keywords: Biocatalysis, Biotransformations, Epoxide Hydrolase, Halohydrin dehalogenase,
Enzyme
1. INTRODUCTION
GSK2330672 (5), an ileal bile acid transport (iBAT) inhibitor indicated for diabetes type II and
cholestatic pruritus, is currently under Phase IIb evaluation in the clinic.1 The API is a highly
complex molecule containing two stereogenic centres, one of which is quaternary. A previous
medicinal chemistry route to a related molecule had introduced the quaternary stereocentre using
an enantiopure activated aminoalcohol or aziridine.2 However, these building blocks themselves
required a lengthy synthesis starting from a quaternary amino acid, which was resolved using pig
ACS Paragon Plus Environment
2

Page 3 of 31
Organic Process Research & Development
1
2
liver esterase.3 In addition to the fact that the use of mammalian sourced enzymes should be
avoided unless absolutely necessary due to transmissible spongiform encephalopathies (TSE)
risks,4 a more concise route, via ring opening of the racemic epoxide 2 with thiophenol 3 and
conversion of the resulting alcohol to amine via a Ritter reaction was therefore developed
(Scheme 1).1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Current route to chiral intermediate 4 in the synthesis of GSK2330672
To avoid the need for a wasteful classical resolution to produce the tartrate salt 4, we wanted to
access the enantioenriched epoxide (R)ꢀ2 directly. Numerous chemocatalytic5 and biocatalytic
methods6 have been reported to access chiral epoxides by asymmetric or kinetic resolution
approaches from either unactivated alkenes or aldehydes, but few are capable of efficiently
producing chiral 2,2ꢀdisubstituted epoxides. Perhaps the most elegant chemical approach is via
asymmetric CoreyꢀChaykovsky reaction using a heterobimetallic LaꢀLi3ꢀBINOL complex (LLB)
as reported by the Shibasaki group.7 However, whereas excellent yields and enantioselectivities
ACS Paragon Plus Environment
3

Organic Process Research & Development
Page 4 of 31
1
2
3
4
5
6
are achieved for the synthesis of various aryl or alkylꢀ2ꢀmethyloxiranes, 2ꢀarylꢀ2ꢀethyloxiranes
are produced with only moderate enantioselectivity which was deemed too low for our needs.
7
8
9
We first turned our attention to halohydrin dehalogenases (HHDHs) to access (R)ꢀ2. HHDHs are
lyases that catalyze the ringꢀclosure of halohydrins to produce epoxide rings.8 They are also
capable of catalyzing epoxide ringꢀopening with numerous nucleophiles allowing the
enantioselective kinetic resolution of epoxides. We were interested in assessing the possibility of
accessing (R)ꢀ2, prepared readily in racemic form by achiral CoreyꢀChaykovsky reaction from
cheap starting materials,9 using a resolution approach based on the ring opening of racemic
epoxide with cyanide using halohydrin dehalogenases (HHDHs; EC 4.5.1) (Scheme 2).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Halohydrin dehalogenase catalysed enantioselective ring opening.
Although this approach is relatively new, a broad range of halohydrin dehalogenases have
recently become available that we were keen to apply and investigate for this methodology.10
We also set out to investigate a similar resolution approach based on the ring opening of
racemic epoxide with epoxide hydrolase enzymes (EHs), which can only ring open epoxides
using water as the nucleophile (EHs; EC 3.3.2.3). It is well known that a variety of epoxides,
including geminally disubstituted derivatives, can be hydrolytically cleaved with high
enantiospecificity using EHs under extremely mild and benign aqueous conditions.6b
Depending on the orientation of the epoxide within the enzyme active site, nucleophilic attack
may occur at different ends of the epoxide resulting in retention or inversion to afford diol and
ACS Paragon Plus Environment
4

Page 5 of 31
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
residual epoxide of opposite or the same absolute configuration (Scheme 3). Either through acid
hydrolysis,11 or primary alcohol activation and ring closure,12 it is then possible to access >50 %
yields of diol or epoxide respectively.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 3. Hydrolytic kinetic resolution using (i) retaining and (ii) inverting
enantioselective epoxide hydrolases (R, R’ = alkyl).
In a few limited examples, some epoxide hydrolases have also been reported to selectively
hydrolyse both epoxide enantiomers simultaneously at different carbon centres respectively to
afford >50 % yields of enantiopure diol in an “enantioconvergent” fashion (Scheme 4).13
ACS Paragon Plus Environment
5

Organic Process Research & Development
Page 6 of 31
1
2
3
4
Scheme 4. Hydrolytic deracemisation using an enantioconvergent epoxide hydrolase.
5
6
7
8
9
Although a close literature example reported by Mitchitz et al gave only moderate
enantioselectivity (E=7.2) from a hydrolytic resolution of a structurally related epoxide to
compound 2 using EH from Rhodococcus sp. NCIMB 11216,14 we hoped that by testing a small
range of EH homologues, we could find a good starting hit for the preparation of (R)ꢀ2 that could
be further improved through directed evolution, as successfully applied towards this enzyme
class for the synthesis of other epoxides.15
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2. RESULTS AND DISCUSSSION
Screening of two Codexis HHDH panels identified variant HHDHꢀP1H08 as the top hit in
terms of conversion and selectivity. Initial efforts toward process development uncovered that at
higher substrate loading (20 mg mLꢀ1), the racꢀ2 is immiscible in the TrisꢀSO4 buffer system used
as solvent. From this observation, we chose to focus on the use of biphasic reaction media using
either immobilised enzyme or whole cells. Enzyme immobilization and the use of enzymes in
whole cells are both known to increase enzyme stability, including in organic solvents. Several
coꢀsolvents such as toluene, tertꢀbutyl methyl ether (TBME), 2ꢀmethyl tetrahydrofuran,
isopropyl acetate, cyclohexane and ethyl acetate have been tested. We found that, on a small
scale (<1 mL), HHDHꢀP1H08 immobilized on Sepabeads ECꢀEP/M (90 wt%) could resolve 40
mg mLꢀ1 of racꢀ2 to 69% ee with 43% conversion, using 0.6 equiv NaCN in 20% v/v of 200 mM
TrisꢀSO4 in ethyl acetate. However, when increasing the scale of this reaction to 30 mL, the
conversion decreased significantly with a concomitant drop in ee. This could have resulted from
the reduced surface area between phases to volume, resulting in decreased rate of
substrate/product transfer between phases, but was not further investigated. This unfavourable
result led us to focus on the second strategy of whole cells in a biphasic system. Indeed, we
ACS Paragon Plus Environment
6

Page 7 of 31
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
found we could resolve 8 g of racꢀ2 to 95% ee with 52% conversion as determined by GC.
Unoptimized purification by extraction and distillation allowed us to produce 1.85 g of (S)ꢀ2 in
95% ee. This enantioenriched epoxide was subsequently found to afford the unwanted
enantiomer of a known intermediate in the current route to GSK2330672, indicating that the
epoxide (S)ꢀ2 had been obtained from the HHDH resolution.16 Unfortunately, no good HHDH
hits were found that produce the opposite enantiomer in sufficient enantioselectivity.
Turning our attention to EHs, we initially intended to select and order genes for a range of
homologues that are known to accept geminally disubstituted epoxides or are known to perform
enantioconvergent hydrolyses. We were surprised to find that the sequences of epoxide
hydrolases commonly reported for the resolution of geminally disubstituted epoxides or the
enantioconvergent hydrolysis of epoxides, such as the epoxide hydrolases from Rhodococcus or
Nocardia sp., have not been reported in the literature.17
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In spite of the diversity in sequence space within the largest, α/βꢀhydrolase fold family of
epoxide hydrolases, it is well known that their core structure is actually extremely similar.18 We
therefore decided to select a range of genes from this family to primarily include epoxide
hydrolases that are reported to express well (Table 1).
Table 1. Selected epoxide hydrolase constructs.
EH
Construct
Description
Reference
No.
1
2
S sp K8 EH
Sphingomonas sp. K8 ID: ZP_09141165.1/ 19
WP_010127425
S sp MC1 EH
Sphingopyxis sp MC1 (53% identity to 20
Sphingomonas sp. K8, but higher similarity to
Sphingomonas sp HXNꢀ200)
3
NaDSM 12444 Novosphinbium aromaticivorans DSM 12444 ID: 21
ACS Paragon Plus Environment
7

Organic Process Research & Development
Page 8 of 31
1
2
3
4
EH
ABD26703
5
6
7
8
4
5
An LCP 631 EH Aspergillus niger LCP 631 I: AJ238460
22
23
Am
Agromyces mediolanus ZJB1202030ID: JX467176
ZJB120203EH
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6
7
8
Ar AD1 echA
Agrobacterium radiobacter AD1 (echA) ID: 20
O31243
MaIEꢀAr AD1 MalE fused Agrobacterium radiobacter AD1
24
25
echA
Cm Ch34 EH
C. metallidurans Ch34 ID:YP_583993
The EH proteins were expressed from BL21(DE3) cells transformed with pET28a vectors
containing Nꢀterminal His6ꢀtagged genes. All EHs expressed well as expected, but most
displayed poor enantioselectivity or activity towards racꢀ2 in small scale scouting experiments
using whole cells (Table 2).
Table 2. Initial screening data for EH 1-8.a)
Selectivity
Epoxide Time
hydrolase (min)
Conv.
(%)
10
2
7
(E)
2
e.e. [%]
Config. e.e. [%] Config.
EH1 b
EH2 c
EH3 c
EH4 b
EH5 b
EH6 c
EH7 c
EH8 b
60
10
5
3
(R)
(S)
(R)
(S)
(R)
(S)
(S)
(R)
30
5
(S)
(R)
ꢀ
58
8
1
92
4
0
n.d.d
25
20
1
15
60
15
60
60
55
98
87
13
6
70
76
8
(R)
(S)
(R)
(R)
(S)
65
63
53
25
82
2
40
35
14
ACS Paragon Plus Environment
8

Page 9 of 31
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
a) Reaction conditions: racꢀ2 (1–20 ꢀL) was added to whole cell paste (10ꢀ15 mg) containing
expressed epoxide hydrolase in potassium phosphate buffer 100 mM, pH 7.4 (1.5 mL) and the
mixtures shaken at 1300 rpm at 30°C using a thermomixer. Aliquots (100 ꢀL) were extracted
into ethyl acetate (500 ꢀL) and analysed by chiral GC (for GC conditions check supplementary
information). Negative controls containing empty pET28a vector only or buffer, suggested low
substrate hydrolysis in the buffer (2ꢀ3% in 20 h). b) substrate concentration of 4.3 mM. c)
Substrate concentration of 86.5 mM. Absolute configuration assignment determined after
derivatizing the chiral epoxide to intermediate 4 and chiral HPLC analysis. d) n.d.= not
determined.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EH1 exhibited low activity when reactions were run at 4.3 mM, whereas EHs2ꢀ3 were
approximately 20 fold more active, although the enantioselectivities for the epoxide 2 were very
low in all cases. The EH4 from Aspergillus niger produced the diol (R)ꢀ7 with moderate
enantioselectivity (70% ee) at 55% conversion. A followꢀup experiment at the same scale, but
allowing the reaction to progress further (92 % conversion to the (S)ꢀdiol over 23 h) afforded the
(S)ꢀdiol in 88% ee, demonstrating that the enzyme operates towards racꢀ2 in an
enantioconvergent fashion. Although the resulting epoxide had the undesired stereochemistry for
our purposes, to the best of our knowledge, enantioconvergency had not previously been
reported for this enzyme. In contrast, EH5 favoured hydrolysis of (S)ꢀ2 with retention of
stereochemistry to afford (R)ꢀ2 and (S)ꢀ7 (Scheme 5).26 EHs6 and 7 provided poor
enantioselectivities, whereas EH8 performed similarly to EH5, but with slightly reduced activity
and enantioselectivity.
Scheme 5. Epoxide hydrolase from Agromyces mediolanus catalysed enantioselective ring opening.
ACS Paragon Plus Environment
9

Organic Process Research & Development
Page 10 of 31
1
2
3
4
5
6
7
8
9
Cost analysis of the potential resolution and enantioconvergent approaches to (R)ꢀ2
demonstrated little overall difference, whereas the latter would require an additional chemistry
step to generate epoxide from diol. We therefore only progressed the resolution approach using
the epoxide hydrolase from Agromyces mediolanus (EH5).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To optimize the reaction, several experiments were performed to improve parameters which
are typically known to influence enzyme enantioselectivity and activity: catalyst loading, coꢀ
solvents, pH and temperature. Addition of higher enzyme amounts provided the desired
enantioselectivity but the resulting faster reaction rate could impede easy monitoring and
quenching if inappropriately controlled. As a consequence, we found that the addition of
sufficient enzyme to allow the reaction to reach the selected end point (epoxide 60% conversion
or 95% epoxide ee – see below) in approximately 15h allowed sufficient time to monitor reaction
progress.
A reaction temperature of 25°C resulted in a slower rate and reduced selectivity, whereas
reaction at 35°C increased the rate of background hydrolysis over the enzyme catalysed reaction.
As a consequence, further reactions were performed at 30°C. Water immiscible coꢀsolvents were
next evaluated, to assess EH5 enantioselectivity/activity using a substrate concentration of 125
mg/mL (8 volume process) using two buffer/organic solvent ratios: 1:4 and 4:1 (v/v)
respectively. However, none of the conditions tested led to an improvement over buffer alone.
Table 3. Hydrolysis of rac-2 using EH5 in the presence of various organic co-solvents.a)
Entry
Coꢀsolvent
Convb)
(%)
(R)ꢀ2
(S)ꢀ7
e.e. (%)
n.d.c)
82
Selectivity
(%)
ꢀb
e.e. (%)
(E)
16
15
1
2
None
Ethyl acetate
63
28
98
39
20
ACS Paragon Plus Environment
10

Page 11 of 31
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
n.d.
84
3
4
5
6
7
8
9
Ethyl acetate
TBME
80
20
80
20
80
20
80
2
52
3
n.d.
73
n.d.
17
n.d.
90
TBME
n.d.
68
n.d.
16
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Heptane
Heptane
Toluene
Toluene
59
7
n.d.
73
n.d.
77
n.d.
17
50
4
n.d.
n.d.
n.d.
a) Aqueous/organic solvent / 100 mM potassium phosphate buffer pH 7.4, 4/1 and 1/4 (v/v),
EH5, 125 mg/mL racꢀ2, final volume 1 mL, 30°C. Two conical centrifuge tubes of EH5
containing 1.05 g cell paste dissolved in 8 mL 100 mM potassium phosphate buffer pH 7.4 in
total and then split equally to all reactions. b) Conversion time point taken after 20h, initial
experiments showed activity did not change when reaction times >20h. c) n.d.= not determined.
Experiments run at elevated pH (8.0 and 8.5) with EH5 displayed an increased rate of
background hydrolysis and therefore lower pH conditions (7.0 and 7.5) were selected in order to
obtain a higher enantioselectivity
To isolate the desired enantioenriched epoxide, several solvents such as heptane, tertꢀbutyl
methyl ether, hexane, diethyl ether, toluene and ethyl acetate were tested in order to effect
selective extraction. Heptane was found to preferentially extract epoxide, resulting in a 64/36
mixture of epoxide to diol in the combined organic phases after 3 x 1 vol extractions and only
<0.1% of residual epoxide in the aqueous phase. Further washing of the combined heptane phase
with water (2 x 4 vols) led to further enrichment, resulting in a 92:8 epoxide/diol mixture. In
contrast, extraction of both epoxide and diol into ethyl acetate (2 x 5 vols) led to a more practical
approach. The resultant combined ethyl acetate portions could then be readily concentrated under
reduced pressure and the epoxide isolated by vacuum distillation at 20±5 mbar with the jacket
temperature set to 90°C.
ACS Paragon Plus Environment
11

Organic Process Research & Development
Page 12 of 31
1
2
3
4
5
6
7
8
9
With an isolation method in hand and before further scaleꢀup, we needed to ensure that the
stereochemistry remained intact in the subsequent Ritter reaction, determine the required
enantiopurity of (R)ꢀ2 to meet the API specification and determine whether vacuum distillation
of the epoxide is necessary or whether the crude epoxide/diol mixture could be used directly.
Samples of (R)ꢀ2 (90% ee and 95% ee respectively) were thus transformed into the free amine of
compound 4 under the standard Ritter conditions without any observable deterioration in
enantiopurity. Isolation of compound 4 as the tartrate salt by crystallisation resulted in an
enhancement in enantiopurity to 96% ee and 98% ee respectively, the latter of which was
successfully converted through to API of acceptable enantiopurity. The minimum target for the
enantiopurity of (R)ꢀ2 was therefore set at 95% ee for subsequent process optimisation.
To assess whether vacuum distillation of the epoxide was necessary, the crude epoxide/diol
mixture was used as input for the Ritter transformation.. The desired product 4 was obtained in
typical purity with no deterioration in enantiomeric excess while the low molecular weight diol is
dicarded as waste. The telescoped conditions provided a more robust and streamlined sequence
moving forward.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For further scaleꢀup work, larger quantities of EH5 enzyme were required. A large scale
fermentation process using recombinant EH5 in pET28a vector and expressed in competent E
coli BL21(DE3) cells, was developed, producing 1.43 kg of cell paste from a 35L growth. This
cell paste gave significantly higher activity per gram than had been observed previously from
cells produced in shake flasks when the same conditions and scale were tested, due to a better
protein expression.
Having larger quantities of EH5 in hand, further process development work was performed
using 22 g (97% purity) quantities of racemic 2 in 250 mL controlled laboratory reactors (CLRs).
ACS Paragon Plus Environment
12

Page 13 of 31
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Initially, we intended to test several enzyme forms in parallel reactions, using a final reaction
volume of 200 mL at, 30°C and stirring at 300 rpm in a dishꢀbottomed CLR using a PTFE retreat
curve impeller. However, from initial trial reactions using fresh cell paste, we quickly found that
in contrast to our findings on a smaller scale, they could be performed in just 3 vols of buffer
media (338g/L substrate concentration) (Table 4, entry 1) without any impact on
enantioselectivity as long as the appropriate loading of cell paste was charged to keep the
reaction rate sufficiently high to allow completion times of 15ꢀ22 h. The reaction was also
successfully run over shorter reaction times such as 10 h, slow enough to allow monitoring
whilst also avoiding deterioration in enantioselectivity to maximise yield. Reaction sampling was
initially performed using achiral GC analysis. However, significant variation in the results were
obtained due to the biphasic nature of the reaction mixture and the condensation of some of the
volatile epoxide on the lid of the CLR. To avoid these two issues, reaction progression was
instead followed by monitoring changes in ratio of the two epoxide enantiomers by chiral GC,
and workꢀup commenced once the target enantiomeric excess of the epoxide had been achieved.
We also investigated the practicality of longer term storage and shipping of the enzyme.
Therefore, a batch of frozen cell paste was tested for stability after 8 weeks, with no significant
change in performance, which suggested that the cell paste could be stored at ꢀ80°C prior to
processing if required (Table 4, entry 2). The (R)ꢀ2 could be isolated by extraction into ethyl
acetate (1 x 2.4 vols). However, phase separation was not ideal, presumably due to cell debris,
with a persistent emulsion layer being observed. Slow stirring (50 rpm) did not break the
emulsion, although quick filtration through wet Celite grade 545 and subsequent washing with
ethyl acetate (1 x 1.4 vols) allowed the removal of the cell debris which was noticed to favour
emulsion formation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
13

Organic Process Research & Development
Page 14 of 31
1
2
3
4
5
6
7
8
9
However, the longerꢀterm need for low temperature storage of large volumes of frozen cell
paste, the potential need to ship large volumes of biocatalyst and the emulsion formation on
extraction made for a less than ideal process. We therefore looked towards the use of lyophilised
whole cell and clarified lysate powders as more convenient, lowꢀcost alternative. Enzyme
immobilisation was also considered at this stage and discounted based on the negligible impact
of biocatalyst on epoxide cost of goods and the high cost of supports. Hence, we decided that it is
best to discard the enzyme after each use and not immobilise.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Enzyme activity was unaffected upon lyophilisation suggesting that the enzyme is very robust
(Table 4, entry 3). However, emulsion formation on product extraction still persisted (Figure 1).
A batch stored at ꢀ20°C tested 3 weeks after lyophilisation again remained unchanged (Table 4,
entry 4).
Table 4. Enantioselectivities from the hydrolysis of rac-2 using different forms of EH5.a)
Entry Enzyme type
Enzyme loading
Conv.c) Time (R)ꢀ2
(%)
(h)
15
15
15
15
22
e.e. (%)
96.4
(g)
8
(%w/w)
36
1
2
3
4
5
Fresh cell paste
66.4
67.2
68.7
65.7
62.3
Frozen cell pasteb)
8
36
96.7
Lyophilised whole cells
Lyophilised whole cellsd)
2e)
2e)
9.1
96.6
9.1
95.0
Lyophilised
lysate
unclarified
2.2f)
10
92.0
6
Lyophilised clarified lysate
2.1f)
9.5
58.9
20
91.1
a) Reactions were run in a 250 mL CLR, final reaction volume 65 mL, 22g substrate (loading
338 g/L) and 100 mM potassium phosphate buffer pH 7.4 at 30°C. All enzyme forms originate
from the same enzyme batch. b) Batch retested after 8 weeks of storage at ꢀ80°C. c) Conversion
after 15 h, Product was not isolated but in these cases GC yield was reported. d) Batch retested
after 3 weeks of storage at ꢀ20°C. e) Amount of lyophilised enzyme was produced from 8 g of
cell paste (using the 35 L fermentation procedure 41 g of cell paste were obtained / 1 L of
fermentation broth). No buffer was added prior to lyophilisation. f) Amount of lyophilised
ACS Paragon Plus Environment
14

Page 15 of 31
Organic Process Research & Development
1
2
3
4
5
enzyme was produced from 8 g of cell paste. Prior to lysis, cell paste was resuspended in 32 mL
25 mM potassium phosphate buffer, pH 7.4.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. CLR screening (22 g)– lyophilised whole cells (left hand side) vs lyophilised
clarified lysate (right hand side). From top to bottom (time 0 = lyo powder addition, t=10
min and bottom after ethyl acetate and brine addition).
Initial attempts to produce lyophilised unclarified lysate (from 8 g of cell paste) by using
sonication or cell disruption resulted in a small loss of activity during lysis (92% ee after 22 h)
(Table 4, entry 5). However, the high activity of lyophilised unclarified lysate suggested that the
enzyme is stable upon cell disruption, although emulsion formation was still observed during
reaction workup. Attempts to break the emulsion by increasing jacket temperature to 40ꢀ45°C
combined with stirring failed to improve separation. Higher amounts of ethyl acetate (5ꢀ8 vols)
provided a better layer separation, whereas water addition proved to be detrimental.
ACS Paragon Plus Environment
15

Organic Process Research & Development
Page 16 of 31
1
2
3
4
5
6
7
8
9
Ultimately, lyophilised clarified lysate powder (9.5% w/w) was tested and gave comparable
activity to unclarified lyophilised lysate. The reaction occurred with very good
enantioselectivities, and the drop in activity was no higher than 5% (Table 4, entry 6).
Furthermore, we noticed a visible reduction in emulsion formation during workup, with layer
separation occurring after approximately 2 min, compared to >5ꢀ10 min in the previous cases.
After ethyl acetate extraction, brine addition (1.2 vols), and stirring at 50 rpm for 3 min,
complete layer separation was achieved within 5 min (Figure 1, right hand side). Biocatalyst
robustness experiments, showed only a slight drop in activity (<15%) on storage at ꢀ20°C over a
5 month period.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The wildꢀtype EH5 provided very good activity at high substrate loading. As such, there was
no significant value in engineering an improved enzyme. However, several mutants of EH5,
previously developed for improved enantioselectivity towards epichlorohydrin by Xue et al,26
were selected for further evaluation to gauge whether improvements could be gained through
further enzyme modification if desired.
Selected
mutants
(S207V,
N240D,
W128F,
S207V/N240D,
W128F/S207V,
S207V/N240D/W182F) were also screened (data not shown). We found that the N240D (EH15)
mutant gave a slight improvement in enantioselectivity (E=23) towards racemic epoxide 2
compared to wildꢀtype (E=20), and at reduced enzyme loading compared to wild type (5ꢀ8%
w/w). Small scale experiments and further reaction optimisation performed in the same manner
as EH5, and initial process development showed that using EH15 the enzyme loading could be
reduced to 5ꢀ8% w/w.
Given that EH15 appeared to offer significant advantages over EH5, fermentation and
downstream processing to produce a lyophilised, clarified lysate powder form of the biocatalyst
ACS Paragon Plus Environment
16

Page 17 of 31
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
was evaluated. A 10 litre fed batch fermentation was performed to generate high cell density
material and assess the efficacy of an industrial process. In comparison to the batch fermentation
the fed system achieved a higher biomass with maximum OD600 >150. The intracellularly overꢀ
expressed protein was recovered by mechanical cell lysis of the fermentation broth or
resuspended cell paste, followed by clarification by centrifugation in the presence or absence of a
flocculating agent. The broth lysate or resuspended cell lysate treated with polyethylenimine
acetate solution (PEI) as flocculating agent prior to lyophilisation (clarified lysate), was observed
to have a lower turbidity upon resuspension of resultant lyophilised protein powder within the
reaction mixture than lysates without flocculent (nonꢀclarified lysate). Furthermore, this had a
positive impact on the extraction step. In summary, a high activity clarified powder was
produced that could be readily stored refrigerated at 4°C for months with minimal activity loss.
The PEI treated lysates were observed to improve phase separation during product extraction as
less suspended cell debris were present at the recovery step.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3. CONCLUSIONS
Multiꢀgram scale quantities of (R)ꢀ2ꢀbutylꢀ2ꢀethyloxirane 2 were successfully prepared by the
kinetic resolution of racemic 2 at 338 g/L reaction concentration using an easy to handle and
store lyophilised powder of epoxide hydrolase from Agromyces mediolanus. The enzymes EH5
and EH15 were successfully fermented on a 35L scale. EH15 was also produced using high cell
density fermentation with the enzyme stabilized by using several downstream processing
conditions. EH15, a variant of EH5 gave improved activity and enantioselectivity and so it is
likely that enzyme engineering would result in further improvements. As we had tested only a
small set of EHs selected almost at random one would not be surprised if screening of further
ACS Paragon Plus Environment
17

Organic Process Research & Development
Page 18 of 31
1
2
3
4
5
6
7
8
9
natural enzymes would also lead to even better hits towards either resolution or
enantioconvergent approaches to the asymmetric synthesis of geminally disubstituted epoxides.
Epoxide hydrolase from Aspergillus niger provided instead (R)ꢀ2ꢀbutylꢀ2ꢀethyloxirane 2 and
shown to be enantioconvergent. On the same note HHDHs were also shown to favour the (S)ꢀ2
formation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4. EXPERIMENTAL SECTION
General Remarks: General Remarks: The genes encoding EHs1ꢀ8 and EH15 containing Nꢀ
terminal His6ꢀtags were codon optimised for expression in E. coli BL21(DE3), synthesised, and
cloned into pET28a at GenScript. Double strand sequence confirmation was performed at
GenScript. Racꢀ2 was prepared using the CoreyꢀTchaikovsky reaction and used without further
purification. Unless otherwise stated, the reactions were performed in test tubes (1.5 mL),
conical centrifuge tubes (50 mL), conical flasks (120 mL) and controlled laboratory reactors (250
mL). 1 Wt is defined as the weight of 2ꢀbutylꢀ2ꢀethyloxirane 2 charged to the reactor in grams.
All other weights, volumes and equivalents given are calculated relative to this figure. The
reaction progress was monitored by chiral GC and products were assigned by comparison with
authentic samples. Conversion values determined by GC were calculated based on racꢀ2
consumption and formation of the alcohol 7. NMR spectra were recorded on a Bruker Advance
13
400 (1H: 400 MHz, C: 101 MHz) spectrometer using TMS as internal standard (d=0). Small
amounts of cell paste were lysed using a Fisher Scientific sonicator model FB705 equipped with
a specific probe depending on the lysis volume (up to 100 mL). Higher amounts of cell paste
were disrupted using a microfluidiser M110Y from analytikLtd. Lyophilisation was performed
using a BPS VirTris SP Scientific Advantage Pro lyophiliser.
ACS Paragon Plus Environment
18

Page 19 of 31
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
For the enzyme fermentation a 50 L Fermenter SOP Stuart Pope (BPES) 07957 125 829 was
used. Run parameters: 35 L working volume. 30 L/min air, 200 rpm, 0.2 bar. For the second
fermentation procedure production stage was carried out in 12 L Braun bioreactors. The batch
fed enzyme fermentation was carried out in 12 L Braun bioreactors with Siemens control system,
10 L working volume employing rushton turbine impellors. Dissolved oxygen was monitored
using Hamilton Visiferm DO sensors and ph monitored using Hamilton Easyferm 325 pH
sensors. Feed solutions were introduced via Watson Marlow 502S peristaltic pumps with
feedback control via the Siemens control system. OD600 measurements were recorded using a
Cecil Spectrophotometer.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Small scale reactions. Two conical centrifuge tubes (2 x 50 mL) containing Luria Broth
medium (2 x 10 mL) and kanamycin (2 x 10 ꢁL, 1000x, final conc: 50 ꢁg/mL) were preꢀ
inoculated with a colony of Agromyces mediolanus ZJB120203 (EH5) in E. coli BL21(DE3)
cells (grown previously on an agar plate containing kan 50 ꢁg/mL and 1% glucose) and
incubated for 5 h at 37°C with shaking in an orbital shaker at 220 rpm (O.D. after 5 hours = 0.4).
This preculture was used to inoculate Terrific broth medium TB (2 x 1 L) containing kan (final
conc: 50 ꢁg/mL) and allowed to grow at 37°C (in a 2.8 L baffled shake flask) until O.D. of 0.6–
0.8 at 600 nm was reached (after approximately 3 hours), then IPTG was added (final conc:
0.5 mM) and the culture grown at 30 °C for 19 h keeping the same shaking (Kuhner shaker 220
rpm). Cells were pelleted by centrifugation (2 x 50 mL and 2 x 950 mL) using 50 mL conical
centrifuge tubes (20 min, 4000 rpm, 4°C) and 1L centrifuging bottles (20 min, 4000 rpm at 4°C).
The pellet was stored overnight in the freezer at ꢀ20°C. Pellet obtained from 50 mL fermentation
broth (cell wet mass =1.05 and 1.08) and 950 mL fermentation broth (cell wet mass =16.61 and
16.87) was resuspended in potassium phosphate buffer (pH 7.4, 100 mM), mixture being
ACS Paragon Plus Environment
19

Organic Process Research & Development
Page 20 of 31
1
2
3
4
5
6
7
8
9
transferred to an Erlenmeyer flask (see above for sizes). Reactions were started by addition of 2ꢀ
butylꢀ2ꢀethyloxirane (2) and were carried out in an Kuhner shaker at 220 rpm at 30°C. Reaction
evolution was monitored via chiral GC, (50 ꢁL samples were taken each hour and extracted with
ethyl acetate 600 ꢁL).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Preparation of (R)-2-butyl-2-ethyloxirane (2)
Lyophilized EH5 from clarified lysate (12 wt%) was charged to the reaction vessel. Potassium
phosphate buffer pH 7.4 (100 mM, 1.9 vol) was then charged to the same reaction vessel and the
agitation adjusted to 300 rpm. When suspended the reaction was started by the addition of
racemic 2ꢀbutylꢀ2ꢀethyloxirane (22.6 g, 176.3 mmol, 1 wt). The reaction mixture was stirred at
30°C and reaction monitored by chiral GC until the enantiomeric excess (ee) of (R)ꢀ2ꢀbutylꢀ2ꢀ
ethyloxirane 2 reached a value ≥ 95% (R) (typically conversion is around ≥ 62±2% over a
maximum 15 hr time period). The reaction was quenched by adding ethyl acetate (2.4 vol). The
resultant biphasic solution was then filtered over Celite. Additional ethyl acetate (1.2 vol) was
used to wash the Celite cake. The layers were then separated. The aqueous layer was discarded.
The organic layer was washed with brine (1.2 vol). The organic layer was then concentrated by
distillation under reduced pressure to afford a neat mixture of the desired epoxide (R)ꢀ2ꢀbutylꢀ2ꢀ
ethyloxirane (2) and diol byꢀproduct (S)ꢀ2ꢀethylhexaneꢀ1,2ꢀdiol 7. The mixture was distilled at
90°C and 20±5 mbar to give the desired epoxide (R)ꢀ2ꢀbutylꢀ2ꢀethyloxirane (4.58 g, 20% yield,
99.2% purity, 95% ee). 1H NMR (400 MHz, CDCl3) δ 2.61 (d, J = 4.9 Hz, 1H), 2.59 (d, J = 4.9
Hz, 1H), 1.72−1.46 (m, 4H), 1.42−1.26 (m, 4H), 0.99−0.87 (m, 6H). 13C NMR (125 MHz,
CDCl3) δ 60.2, 52.2, 33.7, 27.0, 26.9, 22.9, 14.0, 8.9.
Preparation of (R)-5-((2-amino-2-ethylhexyl)thio)-2-methoxyphenol (4)
ACS Paragon Plus Environment
20

Page 21 of 31
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
A reaction vessel was charged with 3ꢀhydroxyꢀ4ꢀmethoxythiophenol (564 mg, 3.61 mmol),
(R)ꢀ2ꢀbutylꢀ2ꢀethyloxirane (2) (509 mg, 3.97 mmol) and EtOH (3.4 mL). The mixture was
treated with a solution of NaOH (318 mg, 7.94 mmol) in water (2.3 mL). The mixture was stirred
at ambient temperature for 20 h under dry nitrogen gas. The mixture was treated with toluene (4
mL) and stirred for 2 min. The layers were separated and the organic layer was discarded. The
aqueous layer was neutralized with 2N HCl and extracted with toluene. The extract was washed
successively with saturated aqueous Na2CO3 solution and water and concentrated in vacuum to
give intermediate 3 as an oil. The oil intermediate 3 was dissolved in chloroacetonitrile (5.5 mL)
and HOAc (2 mL). The mixture was cooled to 0oC. H2SO4 (0.96 mL, 18.05 mmol), preꢀdiluted
with 0.33 mL of water) was added at a rate maintaining the temperature below 5oC. After stirring
at below 10oC for 0.5 h, the reaction mixture was treated with water, extracted with MTBE. The
extract was washed with saturated aqueous NaHCO3 and concentrated in vacuo to give
intermediate chloroacetamide protected amine as an oil. The oil intermediate chloroacetamide
protected amine was then dissolved in EOH (9.1 mL) and treated with HOAc (1.8 mL) and
thiourea (0.412 g, 5.42 mmol). The mixture was heated at reflux until completion, and then
cooled to ambient temperature. The solids were removed by filtration The filtrate was
concentrated in vacuo to give an oil. The oil was treated with ethyl acetate, washed successively
with saturated aqueous Na2CO3 solution and water, and then concentrated in vacuo to give
intermediate 4 (851mg, 83% yield over 3 steps, 79% purity) as an oil. 425 mg of the intermediate
4 was further purified by silica gel chromatography to give intermediate 4 (223 mg, 100% purity,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
94.8% ee). H NMR (400 MHz, CDCl3) δ 6.94 (d, J = 2.2 Hz, 1H), 6.85 (dd, J = 8.4, 2.2 Hz,
1H), 6.67 (d, J = 8.4 Hz, 1H), 5.23 (s, 1H), 3.78 (s, 3H), 2.87 (s, 2H), 1.46−1.28 (m, 4H),
13
1.25−1.05 (m, 4H), 0.81 (t, J = 6.9 Hz, 3H), 0.76 (t, J = 7.45 Hz, 3H). C NMR (125 MHz,
ACS Paragon Plus Environment
21

Organic Process Research & Development
Page 22 of 31
1
2
3
4
5
6
CDCl3) δ 146.8, 146.0, 129.1, 122.9, 117.7, 111.2, 56.0, 54.8, 47.5, 38.6, 31.8, 25.8, 23.3, 14.1,
8.0.
7
8
9
Enzyme expression in fermenter (50 L)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Seed stage: A sterile Corning Erlenmeyer flask (500 mL) containing Luria Broth medium (100
mL, LB) and kanamycin final conc: 50 ꢁg/mL) was inoculated with 10 µL of glycerol stock cells
and incubated overnight at 37°C with shaking in a Kuhner shaker at 200 rpm. A secondary seed
stage was inoculated using a 1% inoculum from the above primary seed into 8 sterile glass
Erlenmyer flasks (500 mL) containing Luria Broth medium (100 mL, LB), kanamycin final
conc: 50 ꢁg/mL) and glucose (final conc 1%) and incubated at 30°C at 200 rpm in a Kuhner
shaker overnight
Production stage: Production fermentations were carried out in a 50 L Pierre Guerin Biolaftte
bioreactor using 35 L of TB medium, 1% glycerol, 2 mL/L antifoam (DC1520), 50 µg/mL
kanamycin and 2% inoculum from the secondary seed stage. The fermenter was incubated at
37°C, agitated at 200 rpm with air supplied at 30 L/min with a pressure of 0.2 bar. At inoculation
the culture was at OD600 = 0.2. The culture was incubated at 37ºC for ~90 min, with continuous
OD monitoring until OD600 = 2.0. Temperature was reduced to 25ºC and then induced by the
addition of IPTG to a final concentration 0.5 mM. Induction period was 23 h.
Harvesting: Whole broth was harvested by centrifugation (6,427 rcf in Sorvall RC 12BP) for
20 min, preꢀchilled to 4°C, and pellets frozen and stored at ꢀ80°C. At harvest the cell wet mass
was 41 g/L. Total wet cell weight harvest was 1.43 Kg, OD600 = 35.0.
Cell lysis (small scale QC): Samples (1 mL) were centrifuged at 16,000 g for 1.5 minutes.
Supernatant was discarded and pellets frozen. Samples were prepared by evaluating soluble
ACS Paragon Plus Environment
22

Page 23 of 31
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
protein OD600 of Bugbuster Mastermix (Novagen) and centrifuged at 16,000g for 25 minutes at
4ºC. The lysate was analysed with an equal amount of SDSꢀReduced (soluble).
Fed Batch Fermentation (10L)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Seed Stage: 1 L baffled conical flasks containing Luria Broth medium (300 mL), 2% glucose
and 30 ꢁg/mL kanamycin were inoculated with 0.6 mL of EH15 glycerol stock. Flasks were
incubated for 12 h at 28°C and 220 rpm to an OD600 of between 1.0–2.0.
Production stage: 1% inoculum of the seed stage culture (described above) was added into a
12 L fermenter containing 5.5 L of growth medium, GF01B. This medium was sterilised at
natural pH and adjusted to pH 7.0 with ammonium hydroxide solution (9.0 M) prior to
inoculation. Oneꢀhour post inoculation kanamycin (30 µg/mL) was added. The fermentor was
incubated at 27°C and agitated at 300–950 rpm with air supplied at 5.5–20 L/min to maintain
dissolved oxygen levels at 20% saturation or greater. The pH was controlled at 6.90 by the
addition of ammonium hydroxide solution (9.0 M) and growth maintained by the addition of a
50% w/w glucose monohydrate solution. This feed (50% w/w glucose monohydrate solution)
was introduced exponentially over 12 hours to a maximum of 14.5 g/L/h initial volume from the
point of batched glucose exhaustion, as indicated by a rise in pH and dissolved oxygen content
(DO2). After the culture reached an OD600 of 80 (+/ꢀ 10) expression of the epoxide hydrolase was
induced by the addition of isopropylꢀβꢀDꢀthiogalactosidase (IPTG) (final conc 1.0 mM). Feed
rate was reduced to 11 g/L/Hr initial volume and the fermentation was harvested 40 hours post
induction. At harvest the fermenter was chilled to 12°C and broth collected for downstream
processing.
Harvesting: Two methods were used to isolate a clarified protein solution from 10 L
fermentation broth.
ACS Paragon Plus Environment
23

Organic Process Research & Development
Page 24 of 31
1
2
3
4
5
6
7
8
9
Method 1: Biomass from the fermentation broth was isolated by centrifugation (40 min, 4500
rpm at 4°C) to recover cell paste. The cell paste was resuspended to a solid content of 25% (w/w)
with a 100 mM potassium phosphate buffer (79 mM K2HPO4 and 21mM K2HPO4, pH 7–7.5).
The resuspended cells were mechanically disrupted at 28ꢀ30 KPSI at a constant temperature of
10°C. Half of the lysate was clarified by addition of polyethylenimine (PEI, MW 750000) to a
final concentration of 0.4% (w/v), mixing for 30 min, then centrifugation (30 min, 4500 rpm,
4°C) to recover a clarified supernatant. The other half was directly centrifuged (30 min, 4500
rpm, 4°C) to recover a turbid supernatant.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Method 2: The fermentation broth (pH 6.3) was pH adjusted (pH 6.8ꢀ7.0) by addition of 100
mM potassium phosphate buffer solution (79 mM K2HPO4 and 21 mM K2HPO4, pH 7ꢀ7.5). The
cells were mechanically disrupted at 28ꢀ30 KPSI at a constant temperature of 10°C. Half of the
lysate was clarified by addition of polyethylenimine (PEI, MW 750000) to a final concentration
of 0.4% (w/v), mixing for 30 min, then centrifugation (30 min, 4500 rpm, 4°C) to recover a
clarified supernatant. The other half was directly centrifuged (30 min, 4500 rpm, 4°C) to recover
a turbid supernatant. All supernatants volumes were aliquoted in large square Petri dishes and
frozen at –20°C. The frozen clarified protein solutions were lyophilised and milled. Protein
concentrations were measured by Bradford assays.
Lyophilization
All supernatant volumes were aliquoted in large square petri dishes and frozen at ꢀ20°C. The
frozen clarified protein solutions were lyophilised and milled and stored at ꢀ20°C.
ASSOCIATED CONTENT
Supporting Information.
ACS Paragon Plus Environment
24

Page 25 of 31
Organic Process Research & Development
1
2
3
4
The Supporting Information is available free of charge on the ACS Publications website at
5
6
DOI:
7
8
9
Media and buffer preparation and fermenter data (file type, PDF)
The following files are available free of charge (file type, PDF)
Epoxide Hydrolase SDSꢀPage Chiral GC methods and GC chromatograms (file type, PDF)
Enzyme Lyophilisation recipe (file type, PDF)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
AUTHOR INFORMATION
Corresponding Author
*Eꢀmail: gheorghe.d.roiban@gsk.com.
*Eꢀmail: peter.sutton@uab.cat.
ORCID
GheorgheꢀDoru Roiban: 0000ꢀ0002ꢀ5006ꢀ3240
Present Addresses
† Current address: Department of Chemical Engineering, School of Engineering, Universitat
Autònoma de Barcelona, 08193 Bellaterra (Cerdanyola del Vallès), Catalunya, Spain.
Author Contributions
The manuscript was written through contributions of all authors. All authors have given approval
to the final version of the manuscript.
ACKNOWLEDGMENT
ACS Paragon Plus Environment
25

Organic Process Research & Development
Page 26 of 31
1
2
3
4
5
6
We thank Angela Bridges and Anne Gronlund for ordering the epoxide hydrolase plasmids and
Matt John for reviewing the manuscript and providing constructive comments.
7
8
9
REFERENCES
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
. Guo, J.; Liu, B.; Martin, M. T.; Mitchell, M. B.; Zhou, X. Synthesis of benzothiazepines,
WO2016020785A1.
2
. Cowan, D. J.; Collins, J. L.; Mitchell, M. B.; Ray, J. A.; Sutton, P. W.; Sarjeant, A. A.;
Boros, E. E., Enzymaticꢀ and IridiumꢀCatalyzed Asymmetric Synthesis of
a
Benzothiazepinylphosphonate Bile Acid Transporter Inhibitor. J. Org. Chem. 2013, 78 (24),
12726−12734.
3
. Liu, W.; Ray, P.; Benezra, S. A., Chemoꢀenzymic synthesis of optically active α,αꢀ
disubstituted αꢀamino acids. J. Chem. Soc., Perkin Trans. 1 1995, 553−559.
4
. Wells, A. S.; Finch, G. L.; Michels, P. C. and Wong, J. W. Use of Enzymes in the
Manufacture of Active Pharmaceutical Ingredients ꢀ A Science and SafetyꢀBased Approach To
Ensure Patient Safety and Drug Quality. Org. Process Res. Dev. 2012, 16, 1986−1993.
5
. a) JinꢀAo Duan, MinꢀZhe Shen, WeiꢀZhang, Hao Tang, Wei Li, YuꢀPing Tang, NianꢀGuang
Li, ZhiꢀHao Shi, QianꢀPing Shi. Asymmetric Epoxidation of Olefins with Homogeneous Chiral
(Salen) Manganese (III) Complex, Curr. Org. Chem., 2013, 17, 2936−2970; b) Zhu, Y.; Wang,
Q.; Cornwall, R. G.; Shi, Y., Organocatalytic Asymmetric Epoxidation and Aziridination of
Olefins and Their Synthetic Applications. Chem. Rev. 2014, 114 (16), 8199−8256; c) Gelalcha,
F. G., Biomimetic IronꢀCatalyzed Asymmetric Epoxidations: Fundamental Concepts, Challenges
and Opportunities. Adv. Synth. Catal. 2014, 356 (2ꢀ3), 261−299.
ACS Paragon Plus Environment
26

Page 27 of 31
Organic Process Research & Development
1
2
3
4
5
6
6
. a) Hollmann, F.; Arends, I. W. C. E.; Buehler, K.; Schallmey, A.; Buhler, B., Enzymeꢀ
7
8
9
mediated oxidations for the chemist. Green Chemistry 2011, 13 (2), 226−265; b) Hasnaouiꢀ
Dijoux, G.; Majerić Elenkov, M.; Lutje Spelberg, J. H.; Hauer, B.; Janssen, D. B., Catalytic
Promiscuity of Halohydrin Dehalogenase and its Application in Enantioselective Epoxide Ring
Opening. ChemBioChem 2008, 9 (7), 1048−1051.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
7
. Sone, T.; Yamaguchi, A.; Matsunaga, S.; Shibasaki, M., Enantioselective Synthesis of 2,2ꢀ
Disubstituted Terminal Epoxides via Catalytic Asymmetric CoreyꢀChaykovsky Epoxidation of
Ketones. Molecules 2012, 17 (2), 1617−1634.
8
. Schallmey, A. and Schallmey, M. Recent advances on halohydrin dehalogenases—from
enzyme identification to novel biocatalytic applications. Appl. Microbiol. Biotechnol. 2016, 100,
7827−7839.
9
. Aggarwal, V. K.; Richardson, J., The complexity of catalysis: origins of enantioꢀ and
diastereocontrol in sulfur ylide mediated epoxidation reactions. Chem. Commun. 2003, (21),
2644−2651; b) Aggarwal, V. K.; Winn, C. L., Catalytic, Asymmetric Sulfur YlideꢀMediated
Epoxidation of Carbonyl Compounds:ꢂ Scope, Selectivity, and Applications in Synthesis.
Accounts Chem. Res. 2004, 37 (8), 611−620.
10. Fox, R. J.; Davis, S. C.; Mundorff, E. C.; Newman, L. M.; Gavrilovic, V.; Ma, S. K.;
Chung, L. M.; Ching, C.; Tam, S.; Muley, S.; Grate, J.; Gruber, J.; Whitman, J. C.; Sheldon, R.
A. and Huisman,G. W. Nature Biotechnol. 2007, 25, 338−344.
ACS Paragon Plus Environment
27

Organic Process Research & Development
Page 28 of 31
1
2
3
4
5
6
7
8
11. Orru, R. V. A.; Kroutil, W.; Faber, K., Deracemization of (±)ꢀ2,2ꢀdisubstituted epoxides
via enantioconvergent chemoenzymatic hydrolysis using Nocardia EH1 epoxide hydrolase and
sulfuric acid. Tetrahedron Lett. 1997, 38 (10), 1753−1754.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
12. Moussou, P.; Archelas, A.; Furstoss, R., Microbiological transformations 40. Use of fungal
epoxide hydrolases for the synthesis of enantiopure alkyl epoxides. Tetrahedron 1998, 54 (8),
1563−1572.
13. Bala, N.; Chimni, S. S., Recent developments in the asymmetric hydrolytic ring opening of
epoxides catalysed by microbial epoxide hydrolase. Tetrahedron: Asymmetry 2010, 21 (24),
2879ꢀ2898.
14. Mischitz, M.; Kroutil, W.; Wandel, U.; Faber, K., Asymmetric microbial hydrolysis of
epoxides. Tetrahedron: Asymmetry 1995, 6 (6), 1261−1272.
15. For a very recent example see Sun, Z.; Lonsdale, R.; Kong, X.ꢀD.; Xu, J.ꢀH.; Zhou, J.;
Reetz, M. T., Reshaping an Enzyme Binding Pocket for Enhanced and Inverted Stereoselectivity:
Use of Smallest Amino Acid Alphabets in Directed Evolution. Angew. Chem. 2015, 127 (42),
12587−12592.
16. Absolute stereochemistry of the enantioenriched epoxide was determined by conversion to
compound X shown in patent WO2016020785A1. This in turn had been determined by its
preparation from a quarternary aminoacid as described in WO2011137135A1. The absolute
stereochemistry of the quarnernary aminoacid had been determined according to reference 3.
ACS Paragon Plus Environment
28

Page 29 of 31
Organic Process Research & Development
1
2
3
4
5
6
7
8
17. Hellström, H.; Steinreiber, A.; Mayer, S. F.; Faber, K., Bacterial epoxide hydrolaseꢀ
catalyzed resolution of a 2,2ꢀdisubstituted oxirane: optimization and upscaling. Biotechnology
Lett. 2001, 23 (3), 169−173.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
18. Widersten, M.; Gurell, A.; Lindberg, D., Structure–function relationships of epoxide
hydrolases and their potential use in biocatalysis. Biochim. Biophys. Acta, 2010, 1800, 316−326.
19. Protein sequence: MGSSHHHHHH SSGLVPRGSH MKPEPFAIHA EQAMLDELGR
RLQATRWADD IANDDWSSGT NASYLRELAD YWLNHFDWRA QERTINAYPH
FRILLDGMPI
HFIHQKGVGD
VIVPSLPGFG
HAMPLILTHG
YSSPLRQSGV
WPWTFWDFAK
DARVTADRWR
VIGPLTDPAA
KLMVEILGYP
HGGDPADAFD
RFAAHGGDWG AFVTAQLGHK YPGDVLGIHM IGGAPLDCFG KPLPDASSYA
VDEAGWHAKT QAFFATESGY SAEQSTKPQS LAYGLHDSPV GLAAWLVEKR
RGWSDCDGDI ARRFTKDELL
TSIMIYWLTG
TIGTSARYYY
ENRKAGWQPS
HDGTRMVDVP TGCLKLEADV CHWPRSLMEA NFNLQRWTRS AEGGHFAPAE
VPDLVIAEIR DFFRPLRRTQ QEDQ
20. a) Sequence similarity (86.6%) to EH from Sphingomonas sp HXNꢀ200 see: Wu, S.; Li, A.;
Chin, Y. S.; Li, Z., Enantioselective Hydrolysis of Racemic and MesoꢀEpoxides with
Recombinant Escherichia coli Expressing Epoxide Hydrolase from Sphingomonas sp. HXNꢀ200:
Preparation of Epoxides and Vicinal Diols in High ee and High Concentration. ACS Catal. 2013,
3 (4), 752−759; b) Rink, R.; Fennema, M.; Smids, M.; Dehmel, U.; Janssen, D. B., Primary
Structure and Catalytic Mechanism of the Epoxide Hydrolase from Agrobacterium radiobacter
AD1. J. Biol. Chem. 1997, 272 (23), 14650−14657.
ACS Paragon Plus Environment
29