Evaluation of several routes to advanced pregabalin intermediates: Synthesis and enantioselective enzymatic reduction using ene-reductases

Article abstract of DOI:10.1021/op4002774

This publication describes the evaluation of four synthetic routes to the advanced pregabalin (Lyrica) intermediate 7. Asymmetric reduction of (E)-7 with an ene-reductase (OPR1 from Lycopersicon esculentum) gave a saturated cyanoester intermediate 5 with the desired S stereocenter in ≥99% ee. OPR1 also catalyzed the reduction of (Z)-7 to (S)-5, but with lower conversion and selectivity.

Full text of DOI:10.1021/op4002774

Subscriber access provided by DUESSELDORF LIBRARIES
Full Paper
Evaluation of several routes to advanced pregabalin intermediates:
Synthesis and enantioselective enzymatic reduction using ene-reductases.
Sebastien Debarge, Paul McDaid, Pat O'Neill, James Frahill, John W Wong, Donncha
Carr, Adam Burrell, Simon Davies, Michael j Karmilowicz, and Jeremy Steflik
Org. Process Res. Dev., Just Accepted Manuscript • Publication Date (Web): 09 Dec 2013
Downloaded from http://pubs.acs.org on December 23, 2013
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 38
Organic Process Research & Development
1
2
3
4
5
6
7
8
Evaluation of several routes to advanced pregabalin
intermediates: Synthesis and enantioselective enzymatic
reduction using eneꢀreductases.
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
Sébastien Debarge,a Paul McDaid,a Pat O’Neill,a James Frahill,a John W. Wong,b,* Donncha Carr,a
Adam Burrell,a Simon Davies,a,*, Mike Karmilowicz,b and Jeremy Steflik.b
a
b
Pfizer Global Process Development Centre, Loughbeg, Co. Cork, Ireland; Pfizer Chemical R&D,
Eastern Point Road, Groton, CT 06340, USA.
ABSTRACT: This publication describes the evaluation of four synthetic routes to an advanced
pregabalin (LyricaTM) intermediate 7. Asymmetric reduction of (E)ꢀ7 with an eneꢀreductase (OPR1 from
Lycopersicon esculentum) gave a saturated cyanoester intermediate 5 with the desired (S)ꢀstereocenter in
>99% e.e.. OPR1 also catalyzed the reduction of (Z)ꢀ7 to (S)ꢀ5, but with lower conversion and
selectivity.
Table of contents graphic:
1
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 38
KEYWORDS: Pregabalin, eneꢀreductase enzymes, cyanoacrylate ester, bioreduction, Doebnerꢀ
Knoevenagel, iminium ions, organocatalysis.
1
2
3
4
5
6
7
8
1.0 Introduction
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
(S)ꢀ(+)ꢀ3ꢀaminomethylꢀ5ꢀmethyl hexanoic acid (pregabalin) 3 is a lipophilic GABA (γꢀaminobutyric
acid) analogue that was developed for the treatment of central nervous system disorders including
epilepsy, neuropathic pain, and anxiety.1,2 During the initial process development of 3 several routes
were examined in detail.3 The first generation manufacturing process (Scheme 1) was selected on the
basis of overall yield, cost and high throughput and executed as a racemic synthesis followed by
resolution with (S)ꢀ(+)ꢀmandelic acid.
Scheme 1. First generation pregabalin manufacturing process.
Increasing production demands for 3 led to the requirement to redesign the pregabalin manufacturing
process. Key drivers were the introduction of an early (rather than late stage) resolution as well as a
recycling step to utilise the undesired enantiomer. Initial route selection focused on the use of
asymmetric hydrogenation4 but this was surpassed in terms of cost effectiveness and operability by the
use of an enzymatic process (Scheme 2).5
Scheme 2. Lipolase process to pregabalin API.
2
ACS Paragon Plus Environment

Page 3 of 38
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
The success of the early stage enzyme resolution led to the identification of further process
development opportunities (Scheme 3). It was envisaged that the use of prochiral input material 7 or 8
(easily accessed by hydrolysis of 7) could be reduced selectively using an eneꢀreductase (ERED)
enzyme6 to access late stage intermediates 5 or 6 in the existing lipolase process to 3.
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. Ene-reductase enzymatic routes to existing pregabalin intermediates.
Preliminary screening of recombinant EREDs in whole cells6a and subsequent screening of purified
EREDs6b indicated feasibility for reduction of (E)ꢀ7 to (S)ꢀ5 with >99% enantiomeric purity. However,
this screening showed lower activity and selectivity in the reduction of (Z)ꢀ7 with EREDs.
βꢀCyanoacids (E)ꢀ and (Z)ꢀ8 were also considered as potential precusors to 3 via asymmetric reduction
to (S)ꢀ6 (Scheme 3) and a separate report detailing this approach is planned.
The development of four synthetic pathways to compound 7 is described. The four routes proceed via
(a). triflate chemistry; (b) cyanohydrin; (c) bromination chemistry and (d) Knoevenagel chemistry
(Figure 1).
3
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 38
Figure 1. Synthetic routes to precursor 7.
1
2
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
Drivers for final route selection were based on process efficiency and obtaining a more (E)ꢀ7 selective
route. We also describe the preparative scale bioreduction of (E)ꢀ7 to (S)ꢀ5 with OPR1.
2.0 Results and discsussion
Four processes (triflate, cyanohydrin, bromination and double Knoevenagel route) were investigated
to the key pregabalin intermediate 7 based on their potential low cost and are represented in Figure 1.
2.1 Selective (E) and (Z)-7 synthesis – Triflate route
This route to selectively synthesise (Z)ꢀ7 and (E)ꢀ7 was previously described.6b
4
ACS Paragon Plus Environment

Page 5 of 38
Organic Process Research & Development
Scheme 4. Selective synthesis of (E) and (Z)-7.
1
2
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
Although both (E)ꢀ7 and (Z)ꢀ7 were obtained in good yield and selectivity, the route was not
considered suitable for large scale manufacturing. This was primarily related to concerns around the cost
and reactivity of triflic anhydride, and the use of a high boiling solvent in the final step which
complicated the isolation strategy. Also, since both products are oils, purification would likely require
chromatography, which is undesirable for large scale manufacturing processes. Nevertheless, this route
was used to synthesise (E)ꢀ and (Z)ꢀ7 selectively on a 500 g scale to supply material for ERED
evaluation.
2.2 Cyanohydrin route
The addition of sodium cyanide to βꢀketo ester 9 was evaluated (Scheme 5).7 Unfortunately,
cyanohydrin 11 could not be obtained in significant yield via direct addition under all conditions
attempted. It is worth mentioning that reaction with trimethylsilyl cyanide8 was not attempted due to the
5
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 38
high cost of the reagent. An alternative option evaluated the preparation of 11 via a twoꢀstep process,
1
2
3
using sodium bisulfite addition and subsequent treatment with sodium cyanide (Scheme 5).9
4
5
6
7
8
Scheme 5. Cyanohydrin route.
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
The synthesis of 11 via bisulfite adduct and treatment with NaCN only gave, at best, a 1 to 4 mixture
of βꢀketo ester 9 and cyanohydrin 11.
10
Derivatisation of this crude mixture with SOCl2/NEt3 or MsCl/NEt3 led to the corresponding
products 12a or 12b, respectively, via cyanide elimination (identified by GCMS) (Scheme 5), and it was
apparent that NEt3 gave cyanide elimination prior to derivatisation.
In contrast, conducting the derivatisation with a weaker base such as pyridine, led to derivatisation
without issue, yielding the products 13a or 13b, respectively, (identified by GCMS). Subsequent
addition of NEt3 led to the preferential elimination of OSOCl or OMs groups to give the desired product
7, in addition to unreacted βꢀketo ester 9. As the reaction failed to give 7 with sufficient conversion and
purity, this route was discontinued.
2.3 Bromination route6a
The Knoevenagel condensation of isovaleraldehyde and monoethyl malonate gave predominantly the
α,βꢀunsaturated ester 14. Bromination of 14 led to the corresponding dibromide 15 which, upon
6
ACS Paragon Plus Environment

Page 7 of 38
Organic Process Research & Development
treatment with base, afforded the unsaturated monobromide 16.11 Finally, treatment with potassium
cyanide gave the desired compound 7 (Scheme 6).
1
2
3
4
5
6
7
8
9
Scheme 6. Bromination route.
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 E:Z ratio of 16 was highly dependent on the reaction conditions. The use of NEt3 in EtOH to
synthesise 16 from 15 via dehydrohalogenation gave an E:Z ratio of 1:1, whereas the formation of the Eꢀ
isomer could be favoured by using K2CO3 in EtOH (E:Z 4:1). A variety of conditions were examined for
the key cyanation step of 16 to form compound 7 as shown in table 1 below.
Table 1. Cyanation step.
Monobromide
+
KCN
(eq.)
NH4 Clꢀ
Temp
(°C)
Conversion
(%)a
7 (E:Z
16
(E:Z ratio)a
1 : 1
Entry
Solvent
Time (h)
(mol%)
ratio)a
1
2
3
4
2
1
2
2
20
ꢀ
EtOH
EtOH
EtOH
EtOH
20
16
16
16
5
14
45
1 : 3
1 : 3
1 : 1
Reflux
Reflux
50
4 : 1
ꢀ
100
70
1 : 2.5
1 : 2.5
4 : 1
ꢀ
a Conversion and ratio determined by 1H NMR analysis of the crude reaction mixture.
7
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 38
Pleasingly, full conversion to the desired product 7 was observed using KCN (2 eq.) in refluxing
EtOH (Table 1, Entry 3), but unfortunately a final E:Z ratio of 1:2.5 was observed and the ratio prior to
cyanation was not preserved. Reduced reaction time and temperature gave a lower reaction conversion,
although the E:Z ratio was unaffected (Table 1, Entry 4).
1
2
3
4
5
6
7
8
9
A oneꢀpot process was carried out by addition of KCN after bromination of 14. The one pot
dehydrohalogenationꢀcyanation of 15 was achieved using KCN (2 eq.) in refluxing EtOH to give
compound 7 in 81 % conversion (E:Z 1:2).
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
Since the bromination process gave (Z)ꢀrich mixtures of 7 the route was discontinued owing to the
higher activity of the eneꢀreductase enzyme towards (E)ꢀ7.
2.4 Double Knoevenagel route.
2.4.1 Retrosynthetic Analysis.
It was envisioned that 7 could be constructed via a DoebnerꢀKnoevenagel12 reaction between 2ꢀcyanoꢀ4ꢀ
methylpentanoic acid 17 and commercially available ethyl glyoxylate 18 (Figure 2). The saturated
cyanoacid 17 can be derived from reduction of the corresponding unsaturated cyanoacid 19. Finally, 19
would derive from a Knoevenagel reaction between isobutyraldehyde 20 and cyanoacetic acid 21, both
of which are cheap, commercially available starting materials. This strategy has significant advantages
over both Wittig and related HornerꢀWadsworthꢀEmmons processes due to excellent atom efficiency, as
only H2O and CO2 are generated as byproducts.
8
ACS Paragon Plus Environment

Page 9 of 38
Organic Process Research & Development
Figure 2. Synthetic plan for synthesis of key compound 7.
1
2
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
2.4.2 Initial Knoevenagel Reaction/Reduction Sequence: Synthesis of 2-Cyano-4-methylpentanoic
Acid 17
The Knoevenagel reaction/reduction sequence outlined in Figure 2 is a well precedented process used in
the synthesis of substituted diesters and cyanoesters.13 Our initial attempts to synthesise 19 via a
Knoevenagel reaction14 between 20 and 21 under standard conditions in pyridine with pyrrolidine
provided the desired product as a white solid in 94% yield (Scheme 7). Hydrogenation of 19 with Pd/C
in EtOAc or THF provided a mixture of the desired product 17 and the corresponding amino acid 22 in a
3:1 ratio (assigned by LCꢀMS). Use of an alcoholic (MeOH or EtOH) solvent resulted in formation of a
mixture of the desired 17 and the corresponding cyanoester 23a or 23b in a 5:1 ratio (formed via an in-
situ esterification reaction). Interestingly, alcohols appear to be able to prevent reduction of the nitrile to
the primary amine 22 presumably by interruption of a hydrogen bonding interaction between the acid
and the cyano group.
It was hypothesised that deprotonation of the carboxylic acid to form the corresponding salt should also
suppress this interaction.
9
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 38
Scheme 7. Hydrogenation conditions
1
2
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
Gratifyingly, formation of the sodium salt of 19 in water prior to hydrogenation provided cleanly
saturated cyanoacid 17 as the sole product after acidic workꢀup.
Knoevenagel condensations have been previously reported in water,15 and although water is energy
intensive to heat and cool, it can be a very desirable solvent for chemical reactions for reasons of cost,
safety and environmental concerns if extremes of temperature are not required. As such, efforts were
directed towards the development of an operationally simple oneꢀpot Knoevenagel condensation /
hydrogenation process in water for the synthesis of compound 17 (Scheme 8, condition (i)).
Hydrogenation of a mixture of 21, 20 and aqueous KOH solution in the presence of palladium on carbon
catalyst (Escatꢀ6) in water provided the desired product 17 which was easily isolated as a clear upper
phase after acidic workꢀup. This process was carried out efficiently on a multiꢀgram scale to provide 17
for the development of the second Knoevenegal reaction to generate the desired compound 7.
A one pot Knoevenagel condensation/transfer hydrogenation was viewed as the optimal process,
eliminating the need for specialised hydrogenation equipment. Transfer hydrogenation was first
investigated and optimised using 19. The sodium salt of 19 was efficiently reduced with complete
conversion using sodium hypophosphite16 (2.5eq) and 5% Pd/C in H2O at room temperature for 4 h
providing 17 in 85% yield.
10
ACS Paragon Plus Environment

Page 11 of 38
Organic Process Research & Development
A oneꢀpot process under transfer hydrogenation conditions was achieved in H2O by reaction of a
mixture of 21, aq. NaOH and 20 in the presence of 5% Pd/C and sodium hypophosphite. The desired
product 17 was obtained in quantitative yield and excellent purity (Scheme 8, conditions (ii)).
Scheme 8. One-pot aqueous Knoevenagel condensation – hydrogenation
1
2
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
With a simple, high yielding and inexpensive oneꢀpot route process to 17 in hand, the subsequent
DoebnerꢀKnoevenagel route was examined.
2.4.3 Second Knoevenagel reaction: Pyridine/Ac2O.
The second Knoevenagel reaction is a decarboxylative Knoevenagel condensation (Doebnerꢀ
Knoevenagel)17 between the 17 and ethyl glyoxylate 18 to prepare (E)ꢀ7 and (Z)ꢀ7 (Scheme 9).
Scheme 9. Knoevenagel reaction between saturated cyanoacid 17 and ethyl glyoxylate 18.
The saturated cyanoacid 17 and 18 (50% solution in toluene) were reacted with pyridine/pyrrolidine18
or DMF/DMAP19 to afford the desired products (E)-7 and (Z)-7 and a diastereomeric mixture of the
addition product 24 (Scheme 9). Compound 24 was originally misidentified by GCMS analysis as the
potential hydroxy intermediate 25 (Scheme 10) due to thermal decarboxylation under the GCMS
conditions. The (E)-7:(Z)-7:25 ratio was typically 1:3:5 (determined by GCMS). In all cases, unreacted
starting materials were also observed by GCMS.
11
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 38
In an attempt to favour the formation of (E)-7 and (Z)-7 by dehydration of what was mistakenly
thought to be 25, the reaction was heated at reflux in toluene or pyridine. In both cases, only minor
conversion to the desired (E/Z)-7 with (E)-7:(Z)-7:25 ratio of 1:2:6.
1
2
3
4
5
6
7
8
9
Activation and elimination of the hydroxyl function in compounds 24 or 25 by treatment with Ac2O (1
eq.) and pyridine was attempted to promote the formation of the desired compounds (E)-7 and (Z)-7.
Addition of Ac2O (1 eq.) to the crude mixture derived from the reaction of 17 and 18 with pyridine
(2eq.) in toluene at room temperature provided full conversion to (E)-7 and (Z)-7 in a 1:1 ratio.
Interestingly, when the reaction mixture was heated at reflux and then subsequently cooled and treated at
room temperature with Ac2O (1 eq.), only minor conversion to (E)-7 and (Z)-7 was achieved, with the
acetylated derivatives 26 (Scheme 10) as the major product ((E)-7:(Z)-7:26 ratio of 1:2:5 along with
some degradation products.
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 10. Impurities identified by GCMS.
It was postulated that heating the reaction promoted the formation of 25 through thermal
decarboxylation of 24, and that acetylation of 25 with pyridine/Ac2O led to the stable compound 26.
Conversion to (E)-7 and (Z)-7 after acetylation is only possible via compound 24 through a concomitant
decarboxylationꢀelimination mechanism (Scheme 12).
To verify this hypothesis, a Knoevenagel reaction between 17 and 18 was carried out with pyridine in
toluene at room temperature. Addition of dicyclohexylamine followed by MTBE to the reaction mixture
resulted in the formation of a white solid. LCMS analysis of the white solid revealed the presence of two
diastereomers with a m/z of 243. The structure of 24 was confirmed by NMR analysis. Using GCMS
analysis, decarboxylation of 24 occurred and two signals with a m/z of 199 were observed
corresponding to the two diastereomers of 25.
12
ACS Paragon Plus Environment

Page 13 of 38
Organic Process Research & Development
The isolated diastereomeric mixture 24 was resubmitted to the reaction conditions in the presence of
pyridine/Ac2O and full conversion to desired compound 7 was observed with an (E):(Z) ratio of 1:1.
Two possible mechanistic pathways are shown in Scheme 12 (Pathways 3 and 4).
The reaction worked equally well when Ac2O (1 eq.) was added at the beginning to the reaction
between 17 and 18 with pyridine in toluene at room temperature. Compounds 25 and 26 were not
observed by GCMS (Scheme 11).
1
2
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
The two step process for the synthesis of (E)-7 and (Z)-7 was achieved in one pot via a sequential
addition of reagents (Scheme 11, Table 2, conditions (1) (i), (ii)) or via addition of all reagents upꢀfront
(Scheme 11, Table 2, conditions (2)).
Scheme 11. (E)-7 and (Z)-7 synthesis with pyridine/Ac2O.
CN
O
+
CO2H
H
CO2Et
17
18
(2)
(1) (i)
CN
CN
HO2C
CN
OH
OAc
CO2Et
+
CO2Et
(1) (ii)
HO2C
CO2Et
(E), (Z)-7
24
27
CN
HO2C
CN
OAc
CO2Et
CO2Et
(E), (Z)-7
27
aReaction conditions: (1) (i) 17 (1eq.), 18 50% solution in toluene (1.1eq.), pyridine (2eq.), rt, 12h. (1)
(ii) Ac2O (1.1eq.), rt, 3h. (2) 17 (1eq.), 18 50% solution in toluene (1.1eq.), pyridine (2eq.), Ac2O
b
c
(1.1eq.), rt, 12h. (E)-7:(Z)-7:24 ratio were determined by gas chromatography (GCMS) analysis.
Isolated yield.
13
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 38
Table 2. (E)ꢀ7 and (Z)-7 synthesis with pyridine/Ac2O.
1
2
3
4
5
6
Products ratiob
(Z)ꢀ7
Conditionsa
Yieldc (%)
(E)ꢀ7
24
7
8
9
(1) (i)
(1) (ii)
(2)
1
1
1
3
5 (1:1)
Not isolated
1
ꢀ
ꢀ
80
80
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
0.8
Based on these results, three different pathways (pathways 1,3 and 4) were postulated for the
decarboxylationꢀelimination mechanism as well as a simple base catalysed decarboxylation to 26
(pathway 2) – (Scheme 12).20
Scheme 12. Decarboxylation pathways
Pathway 1 is the desired pathway leading to (E)-7 and (Z)-7. This can occur via a cyclic mechanism
when the carboxylic acid and hydroxyl groups exist in a syn configuration. In pathway 2, temperatureꢀ
mediated decarboxylation of αꢀcyanoacid 24 occurs without elimination of water to generate 25. In
pathway 3, acetylation of the hydroxyl group gives 27. Decarboxylationꢀelimination of the acetate group
occurs when the acetate group is anti to the carboxylic acid group to give (E)-7 and (Z)-7.
14
ACS Paragon Plus Environment

Page 15 of 38
Organic Process Research & Development
In pathway 4, acetylation of the carboxylic acid provides anhydride 28. Intramolecular attack of the
hydroxyl group, when in a syn configuration, on the anhydride functionality could give 27 which could
then readily decarboxylate to provide (E)-7 and (Z)-7.21
1
2
3
4
5
6
7
8
9
Impurities 26 and 23b (Scheme 10) were generated in the reaction after addition of Ac2O. Impurity 26
was generated by acylation of the decarboxylated product 25 via pathway 2, which occurs even at low
temperature. Impurity 23b was generated via esterification of the mixed anhydride derived from 17 and
Ac2O with EtOH, present in the solution of 18.
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 minimise the formation of the two major impurities and increase the overall yield of (E)-7 and (Z)-
7, a combination of lower reaction temperature and EtOHꢀfree 18 was investigated. As the addition was
exothermic, adding the reactants at lower temperature and/or dosed addition of reactants was evaluated;
however this did not provide any decrease in the formation of these impurities.
A screen of various bases was carried out on a crude reaction mixture enriched with compound 26
(formed by heating the reaction between 17 and 18 at reflux before addition of Ac2O), to investigate the
conditions for elimination of the acetate group. Pyridine, NEt3, pyrrolidine, DBU, DMAP and DABCO
were evaluated but in all cases, no conversion to (E)-7 and (Z)-7 was observed and the product was
stable under these conditions.
The effect of solvents on stereoselectivity and formation of impurities was studied. EtOAc, MTBE,
THF, DMF, pyridine, toluene, and CH3CN gave no significant change in the (E):(Z) ratio but toluene
and pyridine were found to be optimal for conversion and impurity profile.
Pyridineꢀtoluene was selected as the optimal system with Ac2O (1 eq.). Under these conditions, a 1:1
mixture of (E)-7 and (Z)-7 was isolated in 80% yield from 17.
The reaction was carried out on a 50 g scale at room temperature using pyridine (1eq.)/Ac2O (1eq.) in
toluene. A mixture of (E)-7 and (Z)-7 was isolated by distillation at reduced pressure in 68% yield,
representing a 65% overall yield of isolated (E)-7 and (Z)-7 from 21. Separation of (E)-7 and (Z)-7 by
fractional distillation was carried out to supply substrate for evaluation of enzymatic reduction.
15
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 38
Despite the robustness and effectiveness of this reaction, the presence of impurities and the 1:1 ratio
of (E)ꢀ7:(Z)ꢀ7 in the Ac2O/pyridine process meant that a cleaner, more effective and more (E)ꢀ7
selective process was required.
1
2
3
4
5
6
7
8
9
2.4.4 Second Knoevenagel reaction: Iminium ion route.
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
Reviewing the proposed mechanism from the Ac2O/pyridine approach, it appears that the
Knoevenagel condensation between 17 and 18 could ultimately be more efficient if a desirable
decarboxylationꢀelimination of 32 (Scheme 13), similar to 27 (Scheme 12), could be achieved.22
Scheme 13. Mechanism for the Knoevenagel reaction of 17 via iminium ion catalysis with
commercial ethyl glyoxylate: Catalytic cycle.
O
CN
H
CO2Et
18
CO2Et
R
R
N
7
H
+ CO2 (g)
R
R
N
CN
O2C
32
NHR2
H
CO2Et
31
CO2Et
CN
CO2H
17
Direct nucleophilic addition of saturated cyanoacid anion equivalent 17 to iminium ion23 31, (formed
by condensation of the aldehyde form of 18 with a secondary amine), is expected to give desired
intermediate 32. In theory, only a catalytic amount of amine is required since turnover of the amine is
achieved through decarboxylationꢀelimination of 32 to give the desired compound 7.
16
ACS Paragon Plus Environment

Page 17 of 38
Organic Process Research & Development
2.4.5 Catalytic Knoevenagel condensation via iminium ion formation using commercial ethyl
glyoxylate 18.
1
2
3
4
5
6
7
8
Since polymeric ethyl glyoxylate is commercially available as a toluene solution, initial investigations
were performed in toluene. Using 17 (1 eq.), 18 (50% solution in toluene, 1.2 eq.), 10 mol %
pyrrolidine, with or without acid coꢀcatalyst (trifluoroacetic acid), in toluene at room temperature
overnight gave a mixture of (E)ꢀ7 and (Z)ꢀ7 (Table 3) and the two diastereoisomers of 24 (identified by
GCMS as the decarboxylated derivatives 25). This suggested that direct coupling of 17 and 18 was
occurring, bypassing the catalytic pathway via iminium intermediate 31.
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
In an attempt to favour the more reactive iminium pathway, the reaction was carried out in toluene at ꢀ
20°C. In this case only (E)ꢀ7 and (Z)ꢀ7 were obtained, suggesting that the reaction proceeded exclusively
through the iminium ion pathway. However, in this case the yield for compound 7 (determined by
HPLC) was poor (~ 10% product with remaining unreacted starting material). When the reaction was
warmed up to room temperature, some starting material reacted via the aldehyde pathway and compound
25 was visible again (by GCMS) in a similar ratio to the reaction previously conducted at room
temperature.
Using 18 with a catalytic amount of pyrrolidine, two mechanisms are possible: the iminium ion
pathway and the aldehyde pathway.
Attack via the aldehyde route leads to the stable intermediate 24, previously isolated and characterised
by NMR and LCMS.
Following the removal of toluene from the commercial solution of ethyl glyoxylate under reduced
pressure, solvent screening was conducted (Table 3 and 4). Screening showed that the reaction pathway
can be directed in favour of the iminium ion route by using polar solvents such as CH3CN, EtOAc,
iPrOAc, DMAc or THF (Table 3, entry 4, 5, 6, 7, 14). CH3CN was found to be the best solvent albeit
affording 7 in poor yield (~50%) (Table 3, entry 7).
Table 3. Synthesis of compound 7 via iminium pathway
17
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 38
Assay yield of
1
2
3
4
7a (%)
Pyrrolidine
(mol%)
CF3CO2H
(mol%)
Entry
Solvents
T°C
E:Z
5
6
7
8
1
Toluene
Toluene
Toluene
EtOAc
EtOAc
EtOAc
CH3CN
MTBE
CPME
THF
10
10
10
10
10
100
10
10
10
10
10
10
10
10
10
ꢀ
rt
rt
ꢀ20
rt
rt
rt
rt
rt
rt
rt
rt
rt
rt
rt
rt
14 (1.8:1)
13 (1.8:1)
10 (2:1)
2
11
ꢀ
9
3
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
ꢀ
37 (3:1)
5
10
100
11
11
11
11
11
11
11
11
11
38 (3:1)
6
40 (1.5:1)
48 (3.5:1)
30 (1.7:1)
25 (1.7:1)
32 (3.2:1)
8.7 (1:2)
33 (2.9:1)
26 (1.1:1)
33 (1.2:1)
8 (1:2.1)
7
8
9
10
11
12
13
14
15
CH2Cl2
iPrOAc
NMP
DMAc
IPA
a The yield was determined by high performance liquid chromatography (HPLC) analysis.
Assay yields showed that there was no need for an acid coꢀcatalyst to promote iminium ion formation
(Table 3, entry 4 and 5). Furthermore, the reaction was cleaner when a catalytic amount of pyrrolidine
was used (Table 3, entries 5 and 6). However, even under the best conditions, the yields were low,
possibly due to the nature of 18 which is commercially available in its polymeric form in toluene
solution, Efforts to purify this starting material and use it as a monomer24,25 did not improve the reaction
yield and since the commercial ethyl glyoxylate toluene solution could not be used directly in the
reaction (from a cost perspective), further investigations on this strategy were abandoned.
1H NMR of commercial ethyl glyoxylate 18 (after evaporation of toluene in vacuo) in CDCl3 showed
that only ~ 16% of the glyoxylate was present as the free aldehyde form, the remaining material was
18
ACS Paragon Plus Environment

Page 19 of 38
Organic Process Research & Development
present as polymeric glyoxylates. As such, use of an excess of 18 (6 eq.) was investigated, to give 1 eq
of reactive ethyl glyoxylate monomer. Spontaneous decarboxylation, evidenced by CO2(g) evolution
was observed as the reaction proceeded, confirming the catalytic turnover of the secondary amine.
We were pleased to see a dramatic yield increase with 76% assay yield of 7 with an E:Z ratio of 3:1
(Table 4, entry 1). This result confirmed that the presence of a stoichiometric quantity of the reactive
aldehyde form of ethyl glyoxylate was necessary to achieve high yields.
1
2
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
On the basis of this result, further solvent screening was conducted to identify, if possible, an
alternative to acetonitrile due to unpredictable supply and price fluctuation of this solvent.
Under these conditions, acetonitrile was still the best solvent for the reaction providing the desired
product in the highest yield (Table 4, Entry 1).
Table 4. Effect of solvents on yields and (E)/(Z) ratio.a
E:Zb
3:1
Assay yield of 7b (%)
Entry
Solvents
CH3CN
1
2
3
4
5
6
7
8
9
10
76
67
71
49
53
37
30
31
21
61
DMAc
1.1:1
1.1:1
4:1
DMF
EtOAc
iPrOAc
3.8:1
2.2:1
2.9:1
5:1
Toluene
MTBE
THF
2ꢀMeꢀTHF
iPrOAc/DMAc (2/1)
4.6:1
1.4:1
b
aToluene was previously removed under vacuo from ethyl glyoxylate solution. The yield was
determined by high performance liquid chromatography (HPLC) analysis.
Unfortunately, the combination of the cost of 18 and the large excess required (6 eq.) were prohibitive
for this process, therefore we focussed on the reactivity of ethyl glyoxylate hemiacetal synthesised from
cheap glyoxylic acid
19
ACS Paragon Plus Environment

Organic Process Research & Development
Page 20 of 38
2.4.6 Knoevenagel condensation using ethyl glyoxylate hemiacetal.
1
2
3
4
5
6
7
8
9
In ethyl glyoxylate hemiacetal 34 (Scheme 14), the aldehyde is protected, preventing any
polymerisation. As such synthesis of the stable ethyl glyoxylate hemiacetal 3426 from cheap
commercially available 50% w/w aq. glyoxylic acid 33 was achieved via removal of water by distillation
followed by heating of the residue in EtOH at 65˚C overnight with catalytic tosic acid (Scheme 14).27
Scheme 14: Synthesis of ethyl glyoxylate hemiacetal
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 reactivity of 34 was evaluated in the Knoevenagel reaction. Reaction of a stoichiometric amount
of crude 34 with 17 in the presence of pyrrolidine (10 mol%) in CH3CN at room temperature provided
product 7 with an HPLC assay yield of 36% after 5h and 69% after 96h with an (E):(Z) ratio of 3.2:1.
Further optimisation was carried out using purified by distillation hemiacetal 34 and in the presence of
pyrrolidine (10 mol%) in CH3CN at room temperature, the desired compounds 7 were isolated with an
HPLC assay yield of 87% after 5h with an (E):(Z) ratio of 3.2:1.
A further solvent screen was carried out under these best conditions at room temperature (Table 5).
Table 5. Effect of solvents on yield of 7 and (E):(Z) ratio.
E:Za
Assay yield of 7a (%)
Entry
Solvents
CH3CN
1
2
3
4
5
6
87
18
14
44
11
83
3.2:1
3.5:1
3.5:1
1.1:1
3.5:1
3.2:1
EtOAc
iPrOAc
DMF
THF
CH3CN:EtOAc (1:1)
a The yield was determined by high performance liquid chromatography (HPLC) analysis.
20
ACS Paragon Plus Environment

Page 21 of 38
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
The results demonstrated that CH3CN gave by far the best conversion after extended reaction time at
room temperature. A mixed solvent system CH3CNꢀEtOAc 1:1 (10 volume) also gave an excellent
conversion ꢀ very similar to that of neat acetonitrile. By further reducing the proportion of CH3CN in the
mixed solvent system to 25% and 10% the reaction yields were reduced to 69% and 50% respectively.
The Knoevenagel condensation was carried out on a 20 g scale of 17 using hemiacetal 34 and the
CH3CNꢀEtOAc mixedꢀsolvent system at 40°C with pyrrolidine (10 mol%). After 22h, the assay yield of
7 was of 88% and 26 g of crude 7 was obtained after workꢀup which gave 83% yield based on purity by
quantitative NMR.
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
Ethyl glyoxylate acetal 35 (Scheme 14) was found to be unreactive under similar conditions.
Therefore, its formation during the synthesis of hemiacetal 34 from glyoxylic acid 33 is undesirable.
Encouraged by these results, the focus moved to the development of a large scale manufacturing
process for the formation of 34.
2.4.7 Synthesis of ethyl glyoxylate hemiacetal 34.28
As the cost for glyoxylic acid monohydrate was too high, a process for the synthesis of the hemiacetal
34 from the cheaper 50% w/w aqueous solution of glyoxylic acid29 33 was required. The quality of 34
and its impact on the second Knoevenagel reaction also required investigation.
Reaction of glyoxylic acid with alcohols can provide several products, namely the desired hemiacetal
34, the hemiacetal acid, ethyl glyoxylate 18 and the unreactive ethyl diethoxyacetate 35. These reactions
take place even in the absence of an external acid catalyst as glyoxylic acid is a sufficiently strong acid
(pKa 2.32) to self catalyse these reactions. Although the use of an external acid catalyst may be essential
to enhance the reaction rates, it could also promote the formation of undesired acetal 35 in significant
quantities.
Two procedures were investigated for the synthesis of 34:
21
ACS Paragon Plus Environment

Organic Process Research & Development
Page 22 of 38
1. Addition of salt (eg. NaCl) to generate a biphasic solution and allow phase separation post
reaction of a 50% w/w aqueous solution of 33 (with or without an acid catalyst) with EtOH (~ 2
eq.) in MTBE.
1
2
3
4
5
6
7
8
9
2. Water removal from 50% w/w aqueous solution of glyoxylic acid 33 either by azeotrope with
heptanes, hexane or toluene or directly by evaporation in vacuo.
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
Option 1 was preferred as it avoided a distillation in the process to remove the water. After overnight
reaction NaCl was added and the organic / aqueous phases separated. The crude material showed
1
approximately 20% conversion to 34 by H NMR. This material was isolated by distillation under
reduced pressure but 34 could only be isolated in ~ 10% yields. Due to the poor conversion to 34 under
these conditions, we decided to focus on option 2 by distillation.
Water was removed from 50 % w/w glyoxylic acid by distillation under vacuum at 60˚C. The residue
was dissolved in EtOH, treated with catalytic acid, and heated at 40˚C (Table 6). Reactions were
allowed to cool and then treated with NaHCO3 and MgSO4, filtered, washed with EtOH and evaporated.
The residue was taken up in MTBE and the solids filtered off. The MTBE solution was evaporated and
the residue analysed by quantitative NMR in DMSOꢀd6 using maleic acid as internal standard to
quantify hemiacetal 34. The crude product was distilled at reduced pressure by drip feed distillation.
Results are tabulated below (Table 6).
Table 6. Effect of acid catalyst on yields and purity of the hemiacetal 34.
w/w purity
before
Weight
after
distillation
w/w purity
after
Active
hemiacetal
yieldc
glyoxylic Acid,
acid 50% Conditions
Weighta
(g)
Entry
1
distillationb
distillationb
40g
40g
40g
None,
31.0g
31.0g
28.8g
29%
63%
27%
17.5g
24.7g
18.5g
58%
75%
73%
31.5%
57.6%
42.0%
40°C, 12h
HCl,
2
3
40°C, 12h
H3PO4,
40°C, 12h
22
ACS Paragon Plus Environment

Page 23 of 38
Organic Process Research & Development
4
40g
pTsOH,
40°C, 12h
HCl,
31.4g
44%
45%
25.5g
74%
63%
58.7%
47.8%
1
2
3
4
5
200g
146.0g
122.0g
5
6
7
8
9
40°C, 5h,
then rt, 12h
6
200g
200g
HCl
168.0g
166.0g
45%
49%
135.0g
141.0g
73%
73%
61.2%
64.0%
10
11
12
13
14
15
40°C, 5h,
then worked
up instantly
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
pTsOH,
40°C, 12h
a
b
after Ethanol removal and MTBE treatment. analysed by quantitative NMR in DMSOꢀd6 using
maleic acid as the internal standard. The yields are calculated based on glyoxylic acid monohydrate
(MW= 92.0 g/mol).
c
Use of an acid catalyst improved the yield and purity and formed the product in a reduced reaction
time. Of the acids screened, pꢀTsOH and HCl are the most promising giving ~60% yield for both after
distillation (corrected for purity) (Table 6, Entry 2, 4ꢀ7). The hemiacetal process scaled up well with no
observed problems. Degradation of 34 was observed over time when the reaction was left for prolonged
periods under the reaction conditions (Table 6, Entry 5 and 6).
Ethyl glyoxylate hemiacetal 34 prepared by both methods was processed in the second Knoevenagel
reaction using the best conditions.
Table 7. Use-test of hemiacetal 34 prepared by different methods in the second Knoevenagel
reaction.
Product 7a
Entry
Source
hemiacetal
34
of w/w purity
Isolated
w/w purity Calculated
of 34 (mass (%)
used)
compound 7 of 7b
yield of 7
1
2
HCl
73% (1.44g) 89.6%
1.30g
1.10g
74.2%c
75%
catalysed
Phosphoric
73% (1.44g) 85.3%
85.6%
73%
23
ACS Paragon Plus Environment

Organic Process Research & Development
Page 24 of 38
acid cat.
1
2
3
4
3
pTSA
catalysed
74% (1.44g) 86.9%
1.21g
81.0%
76%
5
6
7
8
a
Reaction conditions: 17 (92% w/w by NMR, 1.09 g, 7.1 mmol), 34 (1 eq. based on quantitative
NMR), CH3CN/EtOAc (1:1, 20 ml) and pyrrolidine (50 mg), 40˚C overnight. The yield was determined
b
by HPLC analysis. w/w purity was determined by quantitative NMR using benzyl benzoate as the
internal standard. c Low w/w purity due to presence of residual EtOAc.
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
The results demonstrate that the second Knoevenagel chemistry is very consistent and is not affected
by the source of 34, with in-situ yields of 76% observed.
3.0 Use of ene-reductase (ERED) enzymes.
Since syntheses of 7 resulted in mixtures of (E)ꢀ and (Z)ꢀisomers, various approaches were considered
for efficient conversion of (E/Z)ꢀ7 to the desired (S)ꢀ5. One approach which was considered was
asymmetric reduction with EREDs, which catalyze the reduction of activated olefins yielding products
with up to two chiral centers resulting from the stereoselective addition of hydrogen across the carbonꢀ
carbon double bond. Ideally, this approach would involve the stereoꢀconvergent synthesis of (S)ꢀ5 from
both geometric isomers of 7 with a single enzyme or a mixture of two enzymes (Scheme 3). Feasibility
for reduction of (E)ꢀ7 and (Z)ꢀ7 was evaluated with a selection of EREDs from plant and microbial
sources expressed in recombinant E. coli using stoichiometric quantities of NADPH and NADH.6a The
results from this evaluation showed that (E)ꢀ7 was a substrate for all of the EREDS tested and was
converted to (S)ꢀ5 with complete selectivity and low to moderate conversions. (Z)ꢀ7 was also converted
to (S)ꢀ5, however, only one enzyme, OPR1from Lycopersicon esculentum, showed conversion greater
than 1%. Based on the results of the initial screens, OPR1 was selected for further evaluation.
Reductions of (E)ꢀ7 and (Z)ꢀ7 were carried out with OPR1 under conditions in which nicotinamide
cofactor was recycled by the use of a second enzyme; Lactobacillus brevis alcohol dehydrogenase
(LbADH) for NADP+ and formate dehydrogenase (FDH) for NAD+. The results of these studies (Table
8) showed that OPR1 utilized NADP+ and NAD+ with similar efficiency for (E)ꢀ7, but appeared to show
24
ACS Paragon Plus Environment

Page 25 of 38
Organic Process Research & Development
a preference for NAD+ in the case of (Z)ꢀ7. In addition to lower conversion, reduction of (Z)ꢀ7 with
OPR1 also gave lower selectivity, compared to reduction of (E)ꢀ7. Preparative reduction of (E)ꢀ7 with
recombinant OPR1 expressed in E. coli and LbADH for NADP+ recycle afforded (S)ꢀ5 with >99% e.e.
and 69% yield demonstrating potential scalability. However, much lower activity and lower selectivity
in the reduction of (Z)ꢀ7 with OPR1 indicated the need for enzyme optimization before the enzyme can
be used in an efficient bioreduction process which utilizes a mixture of (E)ꢀ and (Z)ꢀ7. It should be noted
that a previous report on bioreduction of (E)ꢀ and (Z)ꢀβꢀcyanoacrylate esters, including (E)ꢀ7 and (Z)ꢀ7,
gave results consistent with the current study, except that (Z)ꢀ7 was not reported as a substrate for
OPR1.6b (Z)ꢀ7 is clearly a substrate for OPR1, albeit a poor one, and the difference in results between
the previous report and the current study is most likely due to higher enzyme loading and longer
reaction times used in the current study.
1
2
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
Asymmetric reduction of βꢀcyanoacrylate esters (E)ꢀ7 and (Z)ꢀ7 to (S)ꢀ5 was demonstrated with OPR1
thus constituting a novel stereoꢀconvergent synthesis of pregabalin. Preparative reduction of (E)ꢀ7 was
demonstrated for the first time using recombinant OPR1 in E. coli cells, albeit requiring a long reaction
time (96 h) and multiple charges of cells, which suggest that increases in specific activity and stability of
the enzyme could benefit the conversion. Much lower conversions of (Z)ꢀ7 to (S)ꢀ5 with OPR1 indicate
that large improvements in enzyme performance would be needed for preparative reduction of (Z)ꢀ7.
Determination of the kinetic parameters for reduction of (E)ꢀ7 and (Z)ꢀ7 to (S)ꢀ5 with OPR1 were not
carried out for this study, but could be useful to support engineering of OPR1. Preliminary feasibility
for improving the properties of OPR1 by enzyme engineering was shown in the previous report6b, which
showed that two variants of OPR1 with single point mutations (Cys20Tyr and His245Asp) gave
improved conversions of (E)ꢀ7 to (S)ꢀ5 compared to wild type OPR1. No conversion of (Z)ꢀ7 was
observed with the Cys20Tyr and His245Asp variants, as was the case for wild type OPR1, however,
these variants did show improved conversions of the methyl ester analog of (Z)ꢀ7 suggesting that further
enzyme engineering efforts could lead to improvements for conversion of (Z)ꢀ7.
25
ACS Paragon Plus Environment

Organic Process Research & Development
Page 26 of 38
1
2
3
4
5
Table 8. Bioreduction of (E)ꢀ7 and (Z)ꢀ7 with OPR1 and Cofactor Recyclea (Scheme 3)
6
7
8
9
(E)ꢀ7
% ee
(Z)ꢀ7
% ee
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
Conditions
Conversion (%)
64.4±1.5
Conversion (%)
1.8±0.5
NADP+ recycle
NAD+ recycle
>99
>99
nd
72.9±0.2
10.8±2.8
91.4±0.4
aKey: nd= not determined; % ee of product (S)ꢀ5
5.0 Conclusion
Four routes were evaluated for the synthesis of cyanoester 7. Although the selectivity of route 1 could
be tailored to (E)ꢀ7 or (Z)ꢀ7 by choice of base to generate the intermediate triflate, the cost of the
reagents were prohibitive, and therefore, this route was not progressed. More economical processes
were identified ꢀ route 2 (cyanohydrin process) and route 3 (bromination) – however both routes
generated 7 as (Z)ꢀrich mixtures. (Z)ꢀselectivity was a challenge as enzyme screening indicated that the
(E)ꢀisomer was the preferred substrate for ERED enzymes. A fourth route (double Knoevenagel process)
was identified as a process to generate 7 with preferred (E)-selectivity. Process development identified a
simple one pot transfer hydrogenation process to generate the key cyanoacid intermediate 17 in
quantitative yield. Transformation of 17 to 7 via the DoebnerꢀKnoevenagel reaction generated a 1:1
mixture of (E)ꢀ7:(Z)ꢀ7 in 65% yield from cyanoacetic acid 21. Development of the iminium process
generated 7 in up to 74% yield with 3:1 (E)ꢀselectivity using ethyl glyoxylate hemiacetal synthesised
from cheap glyoxylic acid without the cost associated with use of ethyl glyoxylate solutions in
CH3CN/EtOAc mixtures.
Evaluation of EREDs resulted in the identification of an enzyme, OPR1, which could generate
saturated cyano ester 5 with the desired (S)ꢀselectivity from either (E)ꢀ7 or (Z)ꢀ7. However, (E)ꢀ7 was a
26
ACS Paragon Plus Environment

Page 27 of 38
Organic Process Research & Development
much better substrate for OPR1 than (Z)ꢀ7. Preparative bioreduction of (E)ꢀ7 to (S)ꢀ5 with OPR1
1
2
3
demonstrated the feasibility of this approach for the synthesis of pregabalin.
4
5
6
7
8
9
Experimental section.
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
All commercial reagents were used without further purification. Lactobacillus brevis ADH (LbADH)
was obtained from Xꢀzyme.
GC analyses were carried out using a Column HPꢀ5MS; 30m x 0.250mm, 0.25um film thickness,
Catalogue No: 19019Sꢀ433. Chiral GC analyses were carried out using a Chiraldex GꢀTA column; 30m
x 0.25 mm, 0.25 ꢀm film thickness. Mass spectra were obtained at 70eV. High resolution mass spectra
were obtained on a Thermo LTQ FTꢀMS using electrospray ionization in positive mode. LCꢀUV
analyses for assay were carried out using a Column Kinetics PFP 150mm x 4.6mm, 2.6um (or
equivalent) (solvent A = 0.1% formic acid in water; solvent B = acetonitrile). Conditions were as
follows 5% to 28% solvent B in solvent A over 10 min, then 28% to 65% solvent B in solvent A over 28
min, then 65% solvent B in solvent A over 5 min, then 5% solvent B in solvent A, monitored by UV
1
13
absorption at 210 nm or ELSD. H and C NMR spectra were recorded at room temperature (~ 20 °C)
in deuterated chloroform (CDCl3) unless otherwise stated using tetramethylsilane (TMS) as an internal
1
1
standard on a 400 MHz, 500 MHz or in some cases 600 MHz NMR spectrometer. Hꢀ1H and Hꢀ13C
correlations were used to confirm the NMR peak assignments. Chemical shifts were expressed in parts
per million (ppm) and coupling constants (J) in hertz (Hz). High resolution mass spectra were obtained
either on a oaꢀTOF mass spectrometer with direct insertion and electron impact ionization (70eV) or on
a Ion Trap FTꢀMS using electrospray ionization in positive mode or APCI ionization in negative mode.
Thin layer chromatography (TLC) was carried out on precoated silica gel plates (60 PF254).
Visualisation was achieved by UV (254 nm) light detection.
Selective (E)-7 and (Z)-7 synthesis – Triflate route.6b
27
ACS Paragon Plus Environment

Organic Process Research & Development
Page 28 of 38
(E)-Ethyl 5-methyl-3-cyano-2-hexenoate (E)-7. 1H NMR (400 MHz, CDCl3): δ 0.91 (d, J= 6.7 Hz,
6H), 1.24 (t, J= 7.1 Hz, 3H), 1.95 (hept, J= 6.8 Hz, 1H), 2.61 (dd, J= 7.3, 1.1 Hz, 2H), 4.15 (q, J= 7.1
Hz, 2H), 6.37 (t, J= 1.1 Hz, 1H). HRMS (ES+) m/z calcd for C10H16NO2 [M + H]+ 182.1181, found:
182.1173.
1
2
3
4
5
6
7
8
9
1
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
(Z)-Ethyl 5-methyl-3-cyano-2-hexenoate (Z)-7. H NMR (400 MHz, CDCl3): δ 0.90 (d, J= 6.6 Hz,
6H), 1.27 (t, J= 7.2 Hz, 3H), 1.99 (hept, J= 6.8 Hz, 1H), 2.19 (dd, J= 7.2 Hz, 1.2 Hz, 2H), 4.22 (q, J=
7.2 Hz, 2H), 6.21 (t, J= 1.2 Hz, 1H). HRMS (ES+) m/z calcd for C10H16NO2 [M + H]+ 182.1181, found:
182.1175.
Cyanohydrin route.
Ethyl 3-cyano-3-hydroxy-5-methylhexanoate 11. Sodium metabisulfite (3.31g, 1.2eq.) was stirred
in a solution of EtOH/water (10mL)/(5mL). The mixture was then cooled down with an ice bath and βꢀ
ketoester 9 (2.5g, 14.5 mmol) was added and the reaction was stirred at 0°C for 2 hours. Sodium cyanide
(1.42g, 2eq.) was then added in small portions and the reaction was stirred overnight from 0°C to room
temperature. Reaction was quenched with an excess of HCl solution and the product was extracted with
EtOAc. The combined organic phase was dried over MgSO4, filtered and evaporated in vacuo to give a
clear oil (m= 2.65g with a ratio cyanohydrin 11/ βꢀ ketoester 9 (4/1) by 1H NMR, ~70% yield); 1H NMR
(400 MHz, CDCl3): δ 0.89 (d, J= 6.6 Hz, 3H), 1.10 (d, J= 6.6Hz, 3H), 1.26 (t, J= 7.2 Hz, 3H), 1.65ꢀ1.73
(m, 2H), 1.99ꢀ2.18 (m, 1H), 2.69, 2.83 (ABq, J= 20.0 Hz, 2H), 4.12 (q, J= 7.2 Hz, 2H); 13C NMR (125
MHz, CDCl3): δ 14.1 (CH3), 21.0 (CH), 23.7 (CH3), 24.2 (CH3), 44.2 (CH2), 48.1 (CH2), 61.3 (CH2),
68.5 (CH), 120.7 (CN), 171.1 (C).
Ethyl 5-methyl-3-cyano-2-hexenoate 7. The crude mixture of cyanohydrin 11, containing ~20% of
unreacted βꢀketoester 9, (0.30g, 1.50mmol) was stirred with pyridine (0.19mL, 1.5eq.) in solvent
dichloromethane (5mL). MsCl (0.18mL, 1.5eq.) or SOCl2 (0.16mL, 1.5eq.) was added slowly and the
reaction was stirred at room temperature for 3h until completion of the reaction. The reaction was then
cooled down to 0°C before addition of NEt3 (0.31mL, 1.5eq.). The reaction was stirred for 1h while
slowly warming up to room temperature. Aqueous HCl solution (20%) was then added (pH ~4), and the
28
ACS Paragon Plus Environment

Page 29 of 38
Organic Process Research & Development
product was extracted with dichloromethane. The combined organic phase was washed with H2O, dried
over MgSO4, filtered and evaporated in vacuo (m= 0.22g, ~80%). GC, HPLC and NMR analysis
indicate complete conversion of the contained cyanohydrin 11 to the desired product 7 with a ratio 2:1
(Z)/(E)ꢀ7. The balance is unreacted βꢀketoester 9.
1
2
3
4
5
6
7
8
9
Bromination route.6a
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
Ethyl 5-methyl-2-hexenoate (Knoevenagel condensation) 14.6a A solution of ethyl malonate
monoester (10 g, 0.075 mol), isovaleraldehyde (4.35 g, 0.05 mol), and DMAP (0.61 g, 0.005 mol) in
DMF (50 ml) in a 250 ml roundꢀbottomed flask was stirred at room temperature overnight. The mixture
was diluted with MTBE (100 ml) and washed with saturated ammonium chloride (30 ml), H2O (30 ml),
saturated sodium hydrogen carbonate (30 ml) and once again with H2O (30 ml). The organic layer was
dried over magnesium sulfate and the solvent was removed in vacuo to afford the desired 14 product as
clear oil (13.2g, 73%). 1H NMR (400MHz, CDCl3): δ 0.92 (6H, d), 1.29 (3H, t), 1.75 (1H, m), 2.08 (2H,
m), 4.16 (2H, q), 5.81 (1H, d) 6.94 (1H, m).
Ethyl 2, 3-dibromo-5-methylhexanoate 15.6a A solution of bromine (0.98 ml, 0.019 mol) in
dichloromethane (3 ml) was added dropwise to a stirred solution of the α,βꢀunsaturated ester 14 (3 g,
0.019 mol) in dichloromethane (6 ml) at 0°C. The reaction was stirred at room temperature overnight,
then saturated sodium metabisulphite was added dropwise until the mixture turned clear. The mixture
was diluted with dichloromethane (50 ml) and the aqueous layer was removed. The organic layer was
washed with H2O (2 x 25 ml), dried over magnesium sulfate and concentrated to afford the product 15
1
as a clear oil (5.06g, 85%). H NMR (400MHz, CDCl3): δ 0.92 (3H, d), 1.01 (3H, d), 1.31 (3H, t), 1.69
(1H, m), 2.02 (2H, m), 4.42 (4H, m).
Ethyl 2-bromo-5-methyl-2-hexenoate 16.6a Potassium carbonate (1.05 g, 7.59 mmol) was added to a
solution of the dibromide 15 (2 g, 6.33 mmol) in ethanol (10 ml) at room temperature. The mixture was
stirred for 2 hrs, then filtered. The solids were washed with ethanol (10 ml) and the combined filtrates
were concentrated on the rotary evaporator. The residue was dissolved in MTBE (50 ml) and washed
with H2O (2 x 25 ml), then dried over magnesium sulfate and the solvent removed to afford the product
29
ACS Paragon Plus Environment