Chemoenzymatic approaches to the dynamic kinetic asymmetric synthesis of aromatic amino acids

Article abstract of DOI:10.1016/j.tetasy.2004.07.060

Enzymatic approaches for the production of amino acids by nitrilases are described. Dynamic kinetic asymmetric synthesis conditions were established for the aromatic aminonitriles, phenylglycinonitrile and 4- fluorophenylglycinonitrile, at high pH to produce the corresponding amino acid products in high enantiomeric excess. N-Acylation of aromatic aminonitriles led to spontaneous racemization at pH 8, allowing preferential enzymatic hydrolysis of the (R)-enantiomer to afford the product N-acylamino acids in up to 99% enantiomeric excess (ee).

Full text of DOI:10.1016/j.tetasy.2004.07.060

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 15 (2004) 2793–2796
Chemoenzymatic approaches to the dynamic kinetic
asymmetric synthesis of aromatic amino acids
Jennifer A. Chaplin,a,* Michael D. Levin,a Brian Morgan,a Nancy Farid,a
Jen Li,a Zuolin Zhu,a Jeff McQuaid,a Lawrence W. Nicholson,b
Cynthia A. Randb and Mark J. Burka
aDiversa Corporation, 4955 Directors Place, San Diego, CA 92121, USA
bThe Dow Chemical Company, 1897Building, Midland, MI 48667, USA
Received 2 June 2004; accepted 13 July 2004
Abstract—Enzymatic approaches for the production of amino acids by nitrilases are described. Dynamic kinetic asymmetric syn-
thesis conditions were established for the aromatic aminonitriles, phenylglycinonitrile and 4-fluorophenylglycinonitrile, at high
pH to produce the corresponding amino acid products in high enantiomeric excess. N-Acylation of aromatic aminonitriles led to
spontaneous racemization at pH8, allowing preferential enzymatic hydrolysis of the (R)-enantiomer to afford the product N-acyl-
amino acids in up to 99% enantiomeric excess (ee).
Ó 2004 Elsevier Ltd. All rights reserved.
1. Introduction
acids, having the advantages of cost effectiveness and
5
readily available rawmaterials. Asymmetric Strecker
Amino acids have played a significant role in the fine
chemical and pharmaceutical industries for several de-
cades. They are used in a wide range of applications,
including building blocks for the pharmaceutical indus-
try, feed additives and human nutrition.1 While fermen-
tation-based routes have been more economical for the
production of the naturally occurring proteinogenic
L-amino acids, a niche exists for the use of chemoenzy-
matic methods in the production of unnatural D- and/
or L-amino acids.1 Chemoenzymatic routes include the
use of aminoacylases,2,3 hydantoinases,4a aminotrans-
ferases4b and amino acid dehydrogenases.4c Many of
these methods are based upon kinetic resolution of race-
mic mixtures, requiring recovery and racemization of
the undesired enantiomer. There is a distinct require-
ment for economic and efficient dynamic kinetic asym-
metric synthesis methods for the production of these
versatile compounds.
reactions are limited in many cases to the preparation
of N-substituted aminonitriles and require expensive or
difficult to obtain chiral catalysts6,7 or chiral auxiliaries,
while at the same time requiring harsh conditions of pH
and temperature to hydrolyze the nitrile and remove the
N-substituent. The stereoselective hydrolysis of amino-
nitriles by nitrilases represents an attractive approach
for the production of amino acids. We have previously
reported the use of nitrilases for the production of man-
delic acid, phenyllactic acid and substituted carboxylic
acids with very high ee (>97%).8 These high enantio-
meric purities are facilitated by reversible HCN loss
and racemization of cyanohydrins at moderate pH and
temperature. Racemization of aminonitriles turns out
to be considerably more challenging.
There are fewreports of amino acid production using
nitrilases.9–12 Our work focuses on determining condi-
tions under which aromatic a-aminonitriles racemize
and combining these conditions with a nitrilase-cata-
lyzed hydrolysis of the aminonitrile, resulting in
dynamic kinetic asymmetric synthesis.13 Herein we
report two approaches to nitrilase-catalyzed dynamic
kinetic asymmetric synthesis: (i) hydrolysis of aromatic
aminonitriles at high pH and (ii) hydrolysis of N-acyl
The Strecker synthesis of a-aminonitriles, followed by
acid- or base-catalyzed nitrile hydrolysis, is one of the
oldest and most well-known routes to racemic amino
*
Corresponding author. Tel.: +1 858 526 5310; fax: +1 858 526
5810; e-mail: jchaplin@diversa.com
0957-4166/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2004.07.060

2794
J. A. Chaplin et al. / Tetrahedron: Asymmetry 15 (2004) 2793–2796
aminonitriles at pH8. Much of the research was carried
out using phenylglycinonitrile 1a as a model aromatic
aminonitrile, with further development on the target
compound, 4-fluorophenylglycinonitrile 1b, an impor-
tant chiral building block. These approaches demon-
strate the feasibility of dynamic kinetic asymmetric
synthesis for the production of amino acids using
nitrilases.
sic conditions, where the enzymatic reaction might oper-
ate. Initial results also indicated that the addition of
cyanide and/or ammonia to solutions of 1a resulted in
rapid racemization. However, subsequent experiments
showed that the rate enhancement was actually due to
an increase in the pH of the reaction mixture. Further
experimentation established that pH values >10 were
required for the rapid racemization of aromatic amino-
nitriles (Fig. 1); t1/2 for the racemization of 1a at pH10.5
was estimated to be 60min compared to 4h at pH9.5.
While the addition of cyanide did not enhance racemiza-
tion, it did contribute to the stability of the aminonitrile
at the high pH necessary for rapid racemization (Fig. 1),
presumably by shifting the imine/aminonitrile equilib-
rium to the right and preventing decomposition by a
retro-Strecker pathway. Similar results established the
requirement for high pH in the racemization of 1b (data
not shown).
2. Results and discussion
2.1. Chemical approaches to enhance racemization
Enantiomerically enriched aromatic aminonitrile salts
are stable compounds. However, the corresponding free
bases are configurationally labile and can racemize and
decompose via a number of pathways, which include
deprotonation, cyanide elimination and Schiff base for-
mation (Scheme 1).14 The challenge associated with
developing a nitrilase-catalyzed dynamic asymmetric
synthesis was to identify conditions under which racemi-
zation of the aminonitrile (krac) was faster than the nitri-
lase-catalyzed hydrolysis, and that both reactions occur
under conditions, which minimize decomposition
pathways.
2.2. Nitrilase-catalyzed dynamic resolution at high pH
Our early results with Nitrilase 5086 had shown that the
enzyme retained a considerable amount of activity at
pH9 (unpublished results). We investigated the effect
of increasing pH on this enzyme and some of our other
nitrilases. At pH10.2, seven nitrilases were active,
including Nitrilase 5086. Further exploration of this
space showed that Nitrilase 5086 was still active at
pH10.8, which would be sufficient for racemization of
1a. While activity up to pH13 has been reported by a
nitrile hydratase from Rhodococcus sp ATCC 39484,20
we believe that this is the first report of an alkalophilic
nitrilase. Further development of the hydrolysis of 1a
by this nitrilase led to reaction conditions, which
afforded the desired product, (R)-2a in 91% ee and a
yield of 60% at pH10.6.à
The initial screen of our nitrilase collection involving 1a
at pH8 identified a number of enzymes, which formed
(R)-phenylglycine (R-2a) in high yield but with only
moderate ee (55–65%). High conversions coupled with
unreacted starting material of lowee indicated that par-
tial racemization of the aminonitrile had occurred, but
that the rate of racemization was significantly less than
that of the enzymatic hydrolysis. The t1/2 for racemiza-
tion of (R)-1a was subsequently found to be 13h at
pH8. Therefore, we set out to investigate conditions to
enhance the rate of racemization. Enantiomerically
enriched 1a was used for surveying racemization condi-
tions since both enantiomers can be easily prepared in
high enantiomeric purity.15
Unfortunately, the same enzyme was not successful with
substrate 1b at these pH levels. Our available collection
of nitrilases was screened (total screened: 177) to deter-
mine the existence of other alkalophilic enzymes. By
increasing the pH, we identified four enzymes that were
active on the substrate at pH11. Hydrolysis of 1b by
Nitrilase 5275 under controlled conditions at pH10.8
was then investigated. Initially, lower yields (up to
67%) were encountered with two possible reasons postu-
lated: (i) decomposition of the substrate by a retro-
Strecker process, or (ii) high levels of 4-fluorobenzalde-
hyde 3b, a possible enzyme inhibitor. Addition of 1equiv
of cyanide to the reaction enhanced the yield and low-
ered the level of aldehyde, through stabilization of the
aminonitrile. The reaction temperature was also found
to be significant, with lower temperatures (18°C) pro-
viding higher yields. These conditions led to the achieve-
ment of 79.5% yield of (R)-2b with an ee of 96.3%. The
addition of enzyme in two batches was found to be
essential for: (i) maintaining the balance between the
rates of hydrolysis and racemization that lead to high
Racemization of amino acid derivatives using catalytic
amounts of aldehydes or ketones has been coupled with
a resolution process (enzymatic16 or crystallization17) to
effect high yielding dynamic kinetic asymmetric synthe-
ses.18 However, the addition of various aldehydes (benz-
aldehyde, salicylaldehyde, pyridoxal) to an aqueous
solution of (S)-1a at pH9.5 did not result in an acceler-
ation of aminonitrile racemization. Treating (S)-1a
under nonaqueous conditions with a series of eight
different organic bases resulted in rapid racemization
only in the case of DBU (pKa 13.2);19 the other bases
(pKa 8–11) showed no sign of racemization. DBU was
not observed to racemize the aminonitrile under bipha-
Enantiomerically enriched 1a was isolated and stored as its tartrate
salt and the free base generated before use. Compound 1b gave
inconsistent results when resolved under the same conditions. All
newcompounds were characterized by NMR and high resolution
mass spectrometry. Ee was determined by chiral HPLC (i) Crownpak
CR (MeOH–HClO4) for compounds 1and 2 and (ii) Chiralcel OD-
RH (MeCN–HClO4) for 8b.
à Unless otherwise noted, typical reaction procedures are described in
Ref. 8; for the high pH reactions, 0.1M carbonate buffer was used.

J. A. Chaplin et al. / Tetrahedron: Asymmetry 15 (2004) 2793–2796
2795
O
NH2
NH2
COOH
H
CN
kR
H
R
R
R
+CN-
-NH3
+CN-
3
(R)-1
+H+
(R)-2
OH
CN
NH
NH2
CN
krac
H
a: R = H
b: R = F
R
R
R
Nitrilase
R
4
- H+
-CN-
OH
COOH
NH2
CN
NH2
COOH
kS
H
R
R
5
(S)-1
(S)-2
Scheme 1.
2.3. Hydrolysis of aromatic N-acylaminonitriles
100
80
60
40
20
0
Unlike their parents, N-acylated aromatic aminonitriles
are stable compounds, which racemize easily under mild
conditions. In fact, (S)-N-acetyl phenylglycinonitrile 6a
racemizes with t1/2 = 40min at pH8. The nitrilase-cata-
lyzed dynamic kinetic asymmetric hydrolysis of N-acyl
aminonitriles (Scheme 2) is an attractive strategy, since
N-acyl aminoacids are useful intermediates in peptide
synthesis; the acyl group can be conveniently removed
by a variety of methods, including an aminoacylase to
yield the unprotected amino acids.21
0
100
200
Time (min)
300
DiversaÕs collection of nitrilases was screened for the
hydrolysis of N-acetyl 4-fluorophenylglycinonitrile 7a.
The initial screen provided 19 active enzymes, of which
Nitrilase 5086 was the most promising. Satisfactory
results were obtained at 25mM substrate concentration
(95% yield of 8a, 91% ee). However, in order to achieve
high volumetric productivities, higher substrate solubil-
ity was required.
Figure 1. The pH dependence of (S)-1a racemization in the presence
and absence of added cyanide. -ꢀ- pH9.6; -r- pH10.4; -j- pH10.9;
-}- pH10.4, 3equiv CN; -Ã- pH10.9, 3equiv CN.
enantioselectivity, and (ii) to supplement enzyme lost
through inactivation at high pH.
Following the relatively successful results reported
above, the substrate concentration was scaled up to
100mM. A continuous flowof enzyme into the reaction
was maintained (0.23mg/h decreasing to 0.15mg/h at
6.5h). Under these conditions, (R)-2b was isolated in
70% yield and 94.4% ee. While we believed that this
approach showed significant promise, the lability of
the aminonitrile substrate resulted in suboptimal yields
while the purity of the product was compromised by
lowlevels of the byproducts ( 3–5) shown in Scheme 1.
Therefore, we decided to pursue the N-acyl aminonitrile
approach described below.
In general, our nitrilases had shown better activity on
aminonitriles than on the corresponding N-acetyl amino-
nitrile. We, therefore, considered hydrolysis of a less
sterically hindered derivative, N-formyl 4-fluorophenyl-
glycinonitrile 7b. This compound is more water soluble
than the corresponding N-acetyl derivative while its
stronger electron-withdrawing properties should
enhance racemization. The 19 nitrilases that were active
on 7a were screened on 7b. The screen identified three
enzymes with high enantioselectivity. Further experi-
mentation showed that one enzyme produced unaccept-
able levels of amide, while another showed low
O
O
HN
R1
COOH
O
NH2
CN
HN
R1
CN
KCN/NH4Cl/
NH4OH
Nitrilase
Ac2O/TEA
or
H
pH 8.5, rt
0-rt overnight
R
R
HCOOEt, reflux
R
R
8a: R = F, R1= CH3
8b: R = F, R1= H
1a: R = H.
1b: R = F
6a: R = H, R1 = CH3
7a: R = F, R1= CH3
7b: R = F, R1= H
3a: R = H
3b: R = F
Scheme 2.

2796
J. A. Chaplin et al. / Tetrahedron: Asymmetry 15 (2004) 2793–2796
2. Sato, T.; Tosa, T. Bioprocess Technol. 1993, 16, 3–14.
3. Bommarius, A.; Drauz, K.; Gunther, K.; Knaup, G.;
¨
conversions at higher substrate concentration
(>25mM). Nitrilase 5086 was the most promising, and
high conversion was achieved at 100mM substrate
through the addition of two aliquots of enzyme. The
presence of a conserved cysteine in the putative catalytic
triad22 suggested that the addition of a reducing agent
might enhance enzyme stability. In the presence of
1mM dithiothreitol (DTT), a single addition of enzyme
at 1mg/mL was possible; the level of DTT could be
reduced to 0.1mM when the reaction was performed
under nitrogen. Further improvement of enzyme per-
formance resulted from the addition of 15% MeOH
and lowering the reaction temperature from 37°C to
ambient. Having established these conditions, the reac-
tion was reproducibly scaled up to 100mM (1g of 7b;
1mg/mL lyophilized lysate of Nitrilase 5086 in 0.1M
Tris–HCl pH8.5). Complete conversion of the substrate
to (R)-N-formyl-4-fluorophenylglycine 8b was typically
observed in 24h. After standard work-up, the product
was isolated as a crude colorless solid, affording a yield
of 87% and ee of 98–99%.
Schwarm, M. Tetrahedron: Asymmetry 1997, 8, 3197–
3200.
4. (a) Pietzsch, M.; Syldatck, C. In Enzyme Catalysis in
Organic Synthesis, 2nd ed.; Drauz, K., Waldmann, H., Eds.,;
Wiley-VCH Verlag GmbH: Weinheim, 2002; pp 761–798;
(b) Rozzell, D.; Bommarius, A. In Enzyme Catalysis in
Organic Synthesis; Drauz, K., Waldmann, H., Eds.,
2nd ed.; Wiley-VCH Verlag GmbH: Weinheim, 2002; pp
873–894; (c) Bommarius, A. In Enzyme Catalysis in
Organic Synthesis, 2nd ed.; Drauz, K., Waldmann, H.,
Eds.; Wiley-VCH Verlag GmbH: Weinheim, 2002; pp
1047–1063.
5. Boesten, W. H. J.; Seerden, J.-P.; de Lange, B.; Dielemans,
H. J. A.; Elsenberg, H. L. M.; Kaptein, B.; Moody, H. M.;
Kellogg, R. M.; Broxterman, Q. B. Org. Lett. 2001, 3,
1121–1124.
6. Ishitani, H.; Hasegawa, Y.; Komiyama, S.; Kobayashi, S.
J. Am. Chem. Soc. 2000, 122, 762–766.
7. Corey, E. J.; Grogan, M. J. Org. Lett. 1999, 1, 159–
160.
8. DeSantis, G.; Zhu, Z.; Greenberg, W. A.; Wong, K.;
Chaplin, J. A.; Hanson, S. R.; Farwell, B.; Nicholson, L.
W.; Rand, C. L.; Weiner, D. P.; Robertson, D. E.; Burk,
M. J. J. Am. Chem. Soc. 2002, 124, 9024–9025.
9. Macadam, A. M.; Knowles, C. J. Biotechnol. Lett. 1985, 7,
865–870.
3. Conclusion
Two novel approaches to the chemoenzymatic produc-
tion of aromatic amino acids have been described. The
ability of nitrilases to function at high pH has been estab-
lished, with phenylglycine and 4-fluorophenylglycine
produced with high enantioselectivity. In order to in-
crease the yields, N-acyl aminonitriles were explored
and found to allow krac to exceed enzymatic hydrolysis,
thereby establishing conditions for dynamic kinetic
asymmetric synthesis. In one example, N-formyl-4-fluoro-
phenylglycine was obtained in 87% yield and 99% ee. The
versatility of our nitrilase collection has been further
extended by demonstrating function at high pH and high
specificity for derivatives of phenylglycinonitrile.
10. Bhalla, T. C.; Muira, A.; Wakamoto, A.; Ohba, Y.;
Furuhashi, K. Appl. Microbiol. Biotechnol. 1992, 37, 184–
190.
11. Brady, D.; Beeton, A.; Zeevaart, J.; Kgaje, C.; van
Rantwijk, F.; Sheldon, R. Appl. Microbiol. Biotechnol.
2004, 64, 76–85.
12. Choi, S. Y.; Goo, Y. M. Arch. Pharm. Res. 1986, 9, 45–47.
13. Ward, R. S. Tetrahedron: Asymmetry 1995, 6, 1475–1490.
14. Yong, Y. F. Ph.D. Thesis. Purdue University, 1988.
15. Marshall, G. C.; Wright, G. D. Biochem. Biophys. Res.
Commun. 1996, 219, 580–583.
16. Wegman, M. A.; Hacking, M. A. P. J.; Rops, J.; Pereira,
P.; van Rantwijk, F.; Sheldon, R. A. Tetrahedron:
Asymmetry 2001, 10, 1739–1750.
17. Hylton, T. A.; Shenton, F. L. U.S. Patent 4,072,698, 1978.
18. Ebbers, E. J.; Ariaans, G. J. A.; Houbiers, J. P. M.;
Bruggink, A.; Zwanenburg, B. Tetrahedron 1997, 53,
9417–9476.
Acknowledgements
The authors would like to thank the following for tech-
nical support and helpful discussions: Dan Robertson,
David Weiner, Iris Saliba, Kevin Keegan, Pei Chen,
Hongjun Huang, Tamara Bowden, Weixing Chen,
Grace DeSantis, Roger Gaudette and Jay Short.
19. Anderson, N. G. In Practical Process Research and
Development; Academic: San Diego, 2000; p 64.
20. Bauer, A.; Layh, N.; Syldatck, C.; Willets, A. J. 1996, 18, ,
pp 343–348.
21. (a) Beller, M.; Eckert, M.; Geissler, H.; Napierski, B.;
Rebenstock, H.-P.; Holla, E. W. Chem. Eur. J. 1998, 4,
935–941; (b) Sonke, T.; Kaptein, B.; Wagner, A. F. V.;
Quaedflieg, P. J. L. M.; Schultz, S.; Ernste, S.; Schepers,
A.; Mommers, J. H. M.; Broxterman, Q. B. J. Mol. Catal.
B: Enzym. 2004, 29, 265–277.
References
1. Kamphuis, J.; Boesten, W. H. J.; Broxterman, Q. B.;
Hermes, H. F. M.; van Balken, J. A. M.; Meijer, E. M.;
Schoemaker, H. E. Adv. Biochem. Eng. Biotechnol. 1990,
42, 134–186.
22. Nakai, T.; Hasegawa, T.; Yamahsita, E.; Yamamoto, M.;
Kumasaka, T.; Ueki, T.; Nanba, H.; Ikenaka, Y.;
Takahashi, S.; Sato, M.; Tsukihara, T. Structure 2000, 8,
729–739.