Enzymatic resolution of n-substituted-β-prolines

Article abstract of DOI:10.1021/op8002097

A general and straightforward strategy for enzymatic resolution of N-substituted-β-proline has been successfully designed and developed in our research laboratories. A first affinity screen is followed by ratio enzyme/substrate optimization to source our

Full text of DOI:10.1021/op8002097

Organic Process Research & Development 2009, 13, 292–296
Enzymatic Resolution of N-Substituted-ꢀ-prolines
,†
,†
†
†
‡
‡
´
Javier Mendiola,* Susana Garc´ıa-Cerrada,* Oscar de Frutos, Mar´ıa Luz de la Puente, Rui Lin Gu, and Vien V. Khau
Centro de InVestigacio´n Lilly S.A. AVda. de la Industria, 30 Alcobendas-Madrid 28108, Spain, and Eli Lilly & Company,
LTC-South, Indianapolis, Indiana 46285-4813, U.S.A.
Abstract:
upon the development of an automated enzymatic screen and
an efficient analytical methodology (that measures yield and
the enantiomeric excess in a single experiment) capable to
efficiently provide us those valuable key intermediates.
A general and straightforward strategy for enzymatic resolution
of N-substituted-ꢀ-proline has been successfully designed and
developed in our research laboratories. A first affinity screen is
followed by ratio enzyme/substrate optimization to source our
chemistry groups N-substituted-ꢀ-proline in kilogram scale in an
efficient and cost-effective way.
2. Results and Discussion
The synthesis of ꢀ-prolines (3a-c) was achieved by adjust-
ing the procedure described in the literature (Scheme 1).3-5 1,3-
Cycloaddition of N-methoxy-N-trimethylsilylbenzylamine (1)
to methyl acrylate in acidic medium yielded pyrrolidine (2) in
high yield.3 Formation of corresponding hydrochloride salt 3a
allowed the easy isolation of highly pure material without any
chromatographic step. Conventional debenzylation by hydro-
genation yielded the desired pyrrolidine (4) in 98% yield (all
attempts with the free amine always provided decomposition
products).4 ꢀ-Proline hydrochloride (4) was used as a key
intermediate to prepare the two different pyrrolidines derivatives
(N-boc 3b5 and N-Cbz 3c6) envisioned to be included in our
enzymatic screen because of their electronic behavior and
synthetic availability.
A set of 13 commercial available hydrolytic enzymes was
selected for a preliminary screen.7 All tests were run in aqueous
medium (phosphate buffer pH 7.5; C, 0.1 M, rt; Scheme 2).The
conversion and optical purity were determined by achiral
reversed-phase HPLC/MS and chiral GC, respectively.
The results for each substrate after 24 h of incubation are
summarized in Table 1. For 3c (Cbz), lipases PS, AY30, PPL,
CCL, proteases P, PS, and acylase overreacted (conversion C
∼100%); lipases AK, AS, FAP15, and M (Table 1, entries 1-4)
showed conversions close to 50% but only lipase M (Mucor
jaVanicus) gave promising results: C, 60%; ee, 92%; and E,8
13 (Table 1, entry 2). When working with substrate 3a,
conversion could not be determined because of the difficulties
found for the isolation of the corresponding carboxylic acid
(zwitterionic behavior). Nevertheless, chiral GC allowed for the
determination of the chiral purity of the remaining ester.
Unfortunately, in all cases, ee was found to be <50% (Table
1, entries 5-17).
1. Introduction
Investigative pharmaceutical agents typically contain multiple
functional groups and one or more stereocenters, and must be
isolated in high chiral as well as chemical purity. Classical
resolution or chiral chromatography can be useful in the
preparation of these drug candidates in enantiomerically pure
form. However, those approaches also show certain drawbacks.
Hence, for instance, resolutions often suffer from low overall
yield and normally require the presence of accessible acidic or
basic functionalities. However, chiral chromatography can be
sometimes solvent and time intensive. A selective synthesis that
yields the desired functionalities in high chemo, enantio, and/
or diastereo control is, therefore, greatly desirable. In this
context, biocatalysis plays an increasing role due to the benefits
associated with those types of processes, such as speed,
selectivity, or ability to quantity.1 Screen of commercially
available enzymes can easily allow the rapid identification of
hits. Further process development involving a more detailed
study to optimize reaction parameters such as temperature,
solvent composition, or rate of addition can then supply the
desired synthetic intermediates in both high yield and high
optical purity.
N-Substituted-ꢀ-prolines are of great interest because of their
biological properties.2 Hence, any valuable technology for the
selective preparation of those type of substrates deserves special
attention. For all these reasons, we decided to focus our efforts
* To whom correspondence should be addressed. E-mail: mendiola_javier@
lilly.com, susana.garcia@lilly.com.
† Centro de Investigacio´n Lilly S.A. Avda. de la Industria.
‡ Eli Lilly & Company.
(1) (a) Patel, R. N. Stereoselective Biocatalysis for Synthesis of Some
Chiral Pharmaceutical Intermediates. In StereoselectiVe Biocatalysis;
CRC Press: Boca Raton, FL, 2000; Vol. 8, pp 7-130. (b) Patel, R. N.
Curr. Org. Chem. 2006, 10, 1289–1321.
(3) Yan, L.; Hale, J. J.; Lynch, C. L.; Budhu, R.; Gentry, A.; Mills, S. G.;
Hajdu, R.; Keohane, C. A.; Rosenbach, M. J.; Milligan, J. A.; Shei,
G.-J. Bioorg. Med. Chem. Lett. 2004, 14, 4861–4866.
(4) Delaney, E. J.; Wood, L. E.; Klotz, I. J. Am. Chem. Soc. 1982, 104,
799–807.
(2) (a) Karoyan, P.; Sagan, S.; Lequin, O.; Quancard, J.; Lavielle, S.;
Chassaing, G. Targets Heterocycl. Syst. 2004, 8, 216–273. (b) Cardillo,
G.; Gentilucci, L.; Tolomelli, A.; Calienni, M.; Qasem, A. R.;
Spampinato, S. Org. Biomol. Chem. 2003, 1, 1498–1502. (c) Hattori,
K.; Yamada, A.; Kuroda, S.; Chiba, T.; Murata, M.; Sakane, K. Bioorg.
Med. Chem. Lett. 2002, 12, 383–386. (d) Wood, S. J.; Wetzel, R.;
Martin, J. D.; Hurle, M. R. Biochemistry 1995, 34, 724–730. (e)
Johnson, G.; Drummond, J. T.; Boxer, P. A.; Bruns, R. F. J. Med.
Chem. 1992, 35, 233–241. (f) Nielsen, L.; Brehm, L.; Krogsgaard-
Larsen, P. J. Med. Chem. 1990, 33, 71–77.
(5) MacLeod, A. M.; Baker, R.; Freedman, S. B.; Patel, S.; Merchant,
K. J. J. Med. Chem. 1990, 33, 2052–2059.
(6) Murphy, J. A.; Tripoli, R.; Khan, T. A.; Mali, U. W. Org. Lett. 2005,
7, 3287–3289.
(7) Lipases AK, AS; FAP15, AY30, R, M, and PGE (Amano Co., Japan);
proteases N, P, and PS (Amano Co., Japan); acylase (Amano Co.,
Japan); lipases PPL, and CCL (Sigma, U.S.A.).
(8) Chen, C. S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am. Chem.
Soc. 1982, 104, 7294.
292
•
Vol. 13, No. 2, 2009 / Organic Process Research & Development
10.1021/op8002097 CCC: $40.75
2009 American Chemical Society
Published on Web 11/12/2008

Scheme 1
Scheme 2
CCL, and acylase (Amano), and proteases PS and P overreacted.
Lipases R, AS, FAP15, PGE, and M, and protease N showed
conversions close to 50%. These six reactions were worked up,
and samples were analyzed by HPLC and chiral GC. Excellent
results were obtained with lipase AS: C, 53%; ee, 96%; and E,
43 (Table 1, entry 19). Compared to 3a and 3c, compound 3b
performed as the best substrate with the popular and inexpensive
lipase AS (Aspergillus niger, Amano Co.) and was then selected
to further refine our enzymatic approach.
Table 1. Screening for affinity study enzyme/substrate (24 h)
conversion
(%)
ee (%)
(isomer favor)
entry substrate enzyme
E
Lipases offer several possibilities for synthetical purposes.
Hence, screening in organic solvents allows enantioselective
esterification of acids, this reaction being a different approach
to also obtain the desired acid in enantiopure form. For this
reason, we envisaged a parallel study of the enzymatic resolution
of the racemic carboxylic acid (5) in an organic solvent9
(MTBE). Required N-Boc-proline (5) was easily obtained by
conventional hydrolysis of the available ester (3b) with NaOH/
MeOH10 (Scheme 3).
Enzymatic resolution screening was run using a set of 24
commercially available hydrolytic enzymes11 in MTBE. The
best results are summarized in Table 2. The three most
promising assays, that showed conversion close to 50% by TLC/
LCMS after 24-120 h in acidic conditions were worked up,
and samples were analyzed by chiral HPLC. Unfortunately, low
selectivity was observed in all cases (Table 2).
1
2
3c
3c
3c
3c
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3b
3b
3b
3b
3b
3b
AK
>50
60
0
M
92 (E2)
15 (E2)
5 (E2)
13
3
AS
16
4
FAP15
Acylase
AK
60
5
12 (E1)
5 (E2)
44 (E2)
2 (E1)
1 (E2)
6
7
AY-30
AS
FAP15
M
N
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
44 (E2)
21 (E2)
18 (E2)
24 (E1)
15 (E2)
1 (E1)
PPL
Prot P
PS
R
PGE
CCL
R
2 (E1)
46 (E2)
84 (E2)
96 (E2)
12 (E2)
74 (E2)
31 (E2)
42 (E2)
53
53
25
47
83
38
19
43
AS
PGE
N
M
FAP15
28
(9) Ghanem, A.; Aboul-Enein, H. Y. Tetrahedron: Asymmetry 2004, 15,
3331–3351.
(10) (a) Boto, A.; Herna´ndez, R.; Leo´n, Y.; Murg´ıa, J. R.; Rodr´ıguez-
Alfonso, A. Eur. J. Org. Chem. 2005, 4, 673–682. (b) Murphy, J. A.;
Tripoli, R.; Khan, T. A.; Mali, U. W. Org. Lett. 2005, 7, 3287–3289.
(11) Hydrolytic enzyme screening kit (24 enzymes), ICR-2400 (Biocata-
lytics Co., U.S.A.).
With substrate 3b (Boc) (Table 1, entries 18-23), all selected
enzymes showed positive reaction. Lipases AK, AY30, PPL,
Vol. 13, No. 2, 2009 / Organic Process Research & Development
•
293

Scheme 3
Table 4. Conversion and optical purity obtained with lipase
AS at different S/E ratios and reaction timesa
time conversion ester ee
acid ee
(%)
entry ratio S/E (h)
(%)
(%)
E
1
2
3
2.5/1
5/1
9/1
24
24
48
65
55
∼40
>98 (E2) 57 (A1) 15
96 (E2) 91 (A1) 83
73 (E2) 94 (A1) 71
a All assays were performed in 20 vol. of buffer at pH 7.5.
Table 2. Results of enzymatic screening for ester formation
Table 5. Conversion and optical purity obtained with lipase
(organic media)
ICR-112 at different S/E ratios and reaction timesa
time conversion
ester
ee (%)
acid
ee (%)
ratio time conversion
ester
acid
entry enzyme
(h)
(%)
E
entry
S/E
(h)
(%)
(ee) (%) (ee) (%)
E
1
2
3
ICR-110
ICR-113 120
ICR-116 24
24
68
39
35
0
0
1
2
2.5/1
2.5/1
4.5
8
59
61
85
47
51
63
29
42
57
18
27
52
55
>98 (E2) 82 (A1) 46
>98 (E2) 74 (A1) 30
>98(E2) 50 (A1)
76 (E2) 86 (A1) 30
>98 (E2) 80 (A1) 40
>98 (E2) 56 (A1) 15
34 (E2) 96 (A1)
80 (E2) 88 (A1) 38
>98 (E2) 74 (A1) 30
14 (E2) 94 (A1)
33 (E1) 20 (A2)
20 (E2) 15 (A1)
3
2.5/1 23
4
5/1
5/1
5/1 24
9/1
9/1
9/1 24
18/1
18/1
18/1 24
18/1 48
4
7
Table 3. Screening results with 24 hydrolytic enzymes
5
6
ester
ee (%)
acid
ee (%)
7
4
7
entry enzyme conversion (%)
E
8
1
2
3
4
5
6
ICR-102
ICR-103
ICR-105
ICR-112
ICR-118
ICR-127
37
35
42
69
45
60
24 (E2) 44 (A1)
36 (E2) 66 (A1)
9
10
11
12
13
4
7
48 (E2) 74 (A1) 11
>98 (E2) 46 (A1) 11
44 (E1) 64 (A2)
34 (E2) 98 (A1)
94 (E2) 84 (A1) 40
>98 (E2) 72 (A1) 27
26 (E1) 20 (A2)
a All assays were performed in 20 vol. of buffer at pH 7.5.
The efforts involved in the simultaneous screen of multiple
enzymes prompted us to take advantage of the TECAN platform
available in our laboratories to enable the automation of both
dispensation of stock solutions/suspensions and collection of
aliquots for monitoring conversion ratio at different times.
Besides this, intensive analytical studies resulted in the develop-
ment of an efficient HPLC/MS method that achieved baseline
resolution of the four isomers (product and reagent) in the same
experiment.12 Furthermore, MS detection assured easy and
accurate identification of each of the components of the mixture.
A, now automated, second screen with the selected 24 hydro-
lytic enzymes11 was then carried out.
Conversion and optical purity were measured every 8 and
24 h hours, respectively. The best results are summarized in
Table 3. Six assays afforded conversion close to 50% (deter-
mined by achiral reversed-phase LCMS under acidic conditions)
and were worked up for chiral HPLC/MS analysis. ICR-112
(CAL-A) showed very promising results for the corresponding
ester, C, 69%; ee >98%; and E, 11 (Table 3, entry 4).
Analysis of all data obtained clearly revealed lipases AS (C,
53%; ee, 96%) and ICR-112 (CAL-A) as the most suitable
hydrolytic enzymes to perform resolution (C,: 69%; ee, >98%).
Nevertheless, and for scale-up purposes, assay conditions
(substrate/enzyme ratio, temperature, buffer concentration, or
reaction time, among others) required further refinement. The
goal was that the selected enzyme yielded conversion close to
50% and optical purity >97% ee in a reaction time of about
24 h. (Shorter reaction times are not convenient to avoid
overreaction due to the impact of time required for analysis
and workup steps.) With lipase AS, the optimization study was
focused on reducing the substrate/enzyme ratio, in order to
improve the workup and therefore decrease the final cost of
the process. In the case of lipase ICR-112 (CAL-A), all efforts
were focused on the adjustment of the observed 69% conversion.
For lipase AS, the ratios selected were 2.5/1, 5/1, and 9/1.
Best results, after incubation at 30 °C and in Na2HPO4/
NaH2PO4, pH 7.5, C, 0.2 M buffer, were obtained with 5/1
S/E ratio (C, 55%; ee, 96%; 24 h; Table 4, entry 2). For ICR-
112 (CAL-A), the conversion obtained in the first screening
was very high (69%); therefore, optimization was done by
adjusting S/E ratios to 2.5/1, 5/1, 9/1, and 18/1. The best ratio
to provide conversion close to 50% in 24 h was 9/1 (Table 5,
entry 9). Other ratios also yielded good conversion and ee (Table
5, entries 1 and 5); however, reaction times were not convenient
for scale-up (<7 h).
After full process optimization, we found that both selected
enzymes provided the ester derivative 3b in high optical purity.
Because of its availability and affordability, lipase AS was
selected for scale-up purposes. Assays were performed on 1 g,
10 g, 100 g, and 1 kg of substrate using 20% w/w (5/1 ratio)
of lipase AS, at 30 °C, and Na2HPO4/NaH2PO4, pH 7.5, C,
0.2 M buffer. Reproducibility of results (Scheme 4) confirmed
the robustness of the process. Configuration of the ester and
acid so isolated was determined to be (S) and (R), respectively,
by comparison of the elution profile of commercially available
materials in the same chiral HPLC conditions selected for the
monitoring of assays.
Final hydrolysis of the methyl ester was performed with
different aqueous bases (NaOH, LiOH) in combination with
organic solvents (MeOH, THF). Reaction with LiOH 2M/THF
yielded the desired acid in 98% yield and >97% ee in a 10 g
scale. It has to be noticed that reaction temperature in this step
must be kept under control (below 0 °C during the addition of
the basic reagent) to fully ensure the optical purity of the
(12) De la Puente, M. L. J. Chromatogr., A 2004, 1055, 55–62.
294
•
Vol. 13, No. 2, 2009 / Organic Process Research & Development

Scheme 4
material. This procedure was reproducible at an 845 g scale,
providing the desired N-Boc-(S)-ꢀ-proline in 74% yield and
>98% ee (Scheme 4).
UV response of target compounds. Mass spectra were recorded
using API-APCI ionization (full scan in positive/negative
modes, simultaneously). All of the four different components
were resolved in the same experiment and easily identified by
ion extraction (positive mode), which facilitated the interpreta-
3. Conclusions
1
We have designed and developed a simple and rapid high-
throughput enzymatic approach for the efficient preparation of
the expensive and low available N-substituted-ꢀ-proline 3b. The
process herein described has proven to be robust and reliable
and has provided this key intermediate for research purposes
in a multigram scale and at 98% ee.
tion of results. H NMR spectra were acquired on a Brucker
Avance DPX 300 MHz spectrometer.
General Procedure for Enzymatic Hydrolytic Screening.
To 10 mg of each enzyme in a test tube was added 0.1 mmol
of substrate dissolved/suspended in 1 mL of 0.1 M phosphate
buffer at pH 7.5 (a stock solution/suspension can be used). The
mixtures were incubated with agitation at 30 °C (sand bath for
hood or Variomag with cooler for TECAN platform), and the
progress of the reaction was monitored by any preferred method
over a 24 h period (HPLC and/or TLC) to check the conversion
of the ester to the carboxylic acid (recommended 8 h, 16 h,
and 24 h). After 24 h, 1 N HCl (1 mL) and MTBE (2 mL)
were added to each mixture. The organic phases were collected
and filtered through a 0.45 nylon filter. The solvent was
removed, and the conversion was monitored by LCMS (acidic
methods).
Procedure for the Scale-Up. Enzymatic Resolution of 1-tert-
Butyl-3-methylpyrrolidine-1,3-dicarboxylate (3b). To 50 L of
reactor with mechanical stirring was added 1-tert-butyl-3-
methylpyrrolidine-1,3-dicarboxylate (930 g, 4 mol) and 18 L
of NaH2PO4/Na2HPO4 buffer (pH 7.5, C 0.2 M). The mixture
was heated at 30 °C. Then, lipase AS (167 g) was added, and
the reaction was stirred for 24 h at 30 °C. The reaction was
checked by chiral chromatography after that time (ee of ester
remaining: >98%). Then, 1 N HCl (5 L) was added. A white
solid was observed. The mixture was extracted with MTBE
twice (20 and 10 L). The organic layers were combined and
filtered through celite. The filtrated was added to a separation
funnel, and the aqueous phase was discarded. The organic layer
was washed with 20% KHCO3 twice (5 and 2 L), dried over
4. Experimental Section
Enzymes and Reagents. All reagents were supplied by
Biocatalytics Co. (U.S.A.) (Hydrolytic enzyme screening kit,
ICR-2400), Amano Co. (Japan), and Sigma (U.S.A.).
Analytical Conditions. HPLC (DAD)/MS was used for the
determination of reaction conversion and enantiomeric excess
of both remaining substrate and resulting product.
All analytical studies were performed on a series 1100 liquid
chromatography/mass selective detector LC/MSD (Agilent,
Waldbronn, Germany) driven by ChemStation software (Rev.
A10.02, Agilent Technologies).
Sample solution of reference materials and those products
yielded by enzymatic screen was prepared by dissolving 2 mg
in 1 mL of a 9/1 hexane/ethanol mixture.
Analyses were conducted using a CHIRALPAK AD unit
from Daicel (Chiral Technologies Europe). Column dimension
was 250 mm × 4.6 mm with the enantioselective phase coated
onto a 10 µm silica-gel substrate. The mobile phase consisted
on a 95/5 mixture of n-hexane (containing TFA at 0.05% v/v)/
ethanol. Flow rate was set at 1 mL/min. All experiments were
carried out at room temperature. The wavelength of UV
detection was monitored from 190 to 500 nm, although
chromatograms were recorded at 205,16 nm due to the poor
Vol. 13, No. 2, 2009 / Organic Process Research & Development
•
295

MgSO4, filtered, and concentrated to give (S)-1-tert-butyl-3-
methylpyrrolidine-1,3-dicarboxylate (399 g, 1.74 mol, 43%, ee
>98%) as an oil. 1H NMR (CDCl3): δ 1.45 (s, 9H), 2.10-2.15
(m, 2H), 3.00-3.10 (m, 1H), 3.28-3.40 (m, 1H), 3.44-3.68
(m, 3H), 3.71 (s, 3H).
The aqueous KHCO3, which contained the acid, was
acidified with 2 N HCl until pH 3. The solid obtained is filtered
and washed with water, and dried to give (R)-1-(tert-butoxy-
carbonyl)pyrrolidine-3-carboxylic acid (481 g, 2.2 mol, 55%,
ee 85%).
the addition of cold 1 N HCl until pH 3.5 (4 L added). The
layers were separated, and the aqueous layer was discarded.
The organic was washed with water (4.25 L) and concentrated
under vacuum (adding toluene until dryness) to give (S)-1-(tert-
butoxycarbonyl)pyrrolidine-3-carboxylic acid (585 g, 2.72 mol,
74%, ee > 98%) as a white solid. H NMR (DMSO-d6): δ
1.39 (s, 9H), 1.88-2.10 (m, 2H), 2.95-3.07 (m, 1H), 3.20-3.45
1
(m, 4H).
Acknowledgment
Hydrolysis of (S)-1-tert-Butyl-3-methylpyrrolidine-1,3-di-
carboxylate. To a 22 L RBF with mechanical agitation,
thermocouple, 2 L addition funnel, and cooling bath was added
(S)-1-tert-butyl-3-methylpyrrolidine-1,3-dicarboxylate (845 g,
3.69 mol, ee >98%) and THF (8.5 L). The mixture was cooled
to -10 °C with an acetone/ice bath. The 2 L addition funnel
was charged with cold LiOH (1.47 L, 2.5 M in water, 3.69
mol), and this solution was added dropwise to the mixture,
maintaining the internal temperature below 0 °C during the
addition (2 h). When the addition was completed, the mixture
was stirred for 2 h at -8 °C. Then cold MTBE (8.5 L) was
added to the mixture, the layers were separated, and the organic
layer was discarded. Then to the aqueous layer was added THF
(4.25 L) and toluene (8.5 L) and cooled to -2 °C, followed by
We thank Drs. Iva´n Collado, Jose´ Miguel M´ınguez, Mer-
cedes Carpintero, Chin Liu, John Masters, Richard Nevill, and
Tony Zhang for their continuous support and for helpful
discussions.
Supporting Information Available
1H NMR characterization for compounds 3a-c, 4, and 5;
standard chiral analysis for 3a and 5. This material is available
free of charge via the Internet at http://pubs.acs.org.
Received for review August 28, 2008.
OP8002097
296
•
Vol. 13, No. 2, 2009 / Organic Process Research & Development