Racemase activity of B. cepacia lipase leads to dual-function asymmetric dynamic kinetic resolution of α-aminonitriles

Article abstract of DOI:10.1002/anie.201007373

Applaudable promiscuity: Racemase-type activity discovered for B. cepacia lipase with N-substituted α-aminonitriles is proposed to involve a C-C bond-breaking/forming mechanism in the hydrolase site of the enzyme, as supported by experimental data and calculations. This promiscuous activity in combination with the transacylation activity of the enzyme enabled the asymmetric synthesis of N-methyl α-aminonitrile amides in high yield (see scheme). Copyright

Full text of DOI:10.1002/anie.201007373

Communications
DOI: 10.1002/anie.201007373
Dynamic Chemistry
Racemase Activity of B. cepacia Lipase Leads to Dual-Function
Asymmetric Dynamic Kinetic Resolution of a-Aminonitriles**
Pornrapee Vongvilai, Mats Linder, Morakot Sakulsombat, Maria Svedendahl Humble,
Per Berglund, Tore Brinck, and Olof Ramstrꢀm*
Chirality is certainly one of the most intriguing phenomena in
nature. It generally leads to specific activities and properties
that differentiate stereoisomers. As a result, there is a high
demand for chiral building blocks, and the search for new and
efficient methods for the synthesis of enantiomerically pure
compounds has been a major area of research in chemistry.[1]
Of the many methods developed, the use of biocatalysts
remains especially attractive, and enzyme-enabled resolution
has become one of the most common ways to prepare
enantiomerically pure compounds on an industrial scale,
owing to high economic efficiency as well as low environ-
mental impact.[2] In this context, the lipase family of enzymes
occupies a privileged position and possesses many advantages
for synthetic applications. These enzymes have high commer-
cial availability, are generally robust and amenable to
recycling, and can accommodate a broad range of sub-
strates.[2–4] Most examples involve secondary alcohols,[5] but
the lipase-mediated asymmetric acylation of primary amines
is becoming increasingly common.[6–10] Surprisingly, however,
only rare examples concerning secondary amines have been
reported.[11–17] We recently addressed this deficiency and
demonstrated a method for the generation and screening of
dynamic N-substituted a-aminonitrile systems through kineti-
cally controlled lipase-mediated amidation.[17] The core a-
aminonitrile (Strecker) structures are synthetically very
versatile. For example, the hydrolysis, reduction, or alkylation
of the nitrile functionality leads to a multitude of building
blocks.[18] Nevertheless, the enzymatic resolution of this class
of compounds has not been much explored.[17,19,20] Accord-
ingly, the information that we gained on substrate selectivity
and stereoselectivity from the screening process in our study
on lipase-mediated amidation prompted us to develop a
practical method for the preparation of optically active N-
methyl a-aminonitrile derivatives.
During these studies, we discovered a novel racemase-
type activity of the lipase.[21] Such catalytic promiscuity of
enzymes has become an important subject for both enzymol-
ogists and organic chemists,[22,23] and could potentially lead to
new synthetic organic methodologies.[24] We also found that
the racemase activity could be used together with the
acylation activity in a coupled process (Scheme 1). Herein,
we describe the use of this dual enzyme functionality for the
asymmetric synthesis of a-aminonitrile amides in good yields
with high enantiomeric purity. A mechanism supported by
experimental and computational studies is also proposed for
the racemization process.
Scheme 1. General concept of the dual-function lipase-mediated
dynamic asymmetric resolution of N-methyl a-aminonitriles. Both the
acylase activity and the racemase activity of the lipase are operating in
the process, in which stereoselective acyl transfer to the secondary
amine is accompanied by racemization of the remaining enantiomer.
Dynamic kinetic resolution (DKR) has proven to be a
powerful method for the preparation of enantiomerically
pure compounds.[25] It circumvents the drawback of tradi-
tional kinetic resolution (KR) to give the desired product in
greater than 50% yield with reduced need for the separation
of unreacted starting materials. The combination of a kinetic
resolving agent, for example, an enzyme or a chemocatalyst,
with in situ racemization, has thus been applied in a variety of
DKR processes.[26] However, the requirement for compati-
bility between the two processes often presents a major
obstacle in the development of such reactions.[27]
[*] Dr. P. Vongvilai, M. Linder,[+] M. Sakulsombat,[+] Prof. T. Brinck,
Prof. O. Ramstrꢀm
Department of Chemistry, KTH—Royal Institute of Technology
Teknikringen 30, 10044 Stockholm (Sweden)
Fax: (+46)8-791-2333
E-mail: ramstrom@kth.se
Dr. M. Svedendahl Humble, Prof. P. Berglund
School of Biotechnology, KTH—Royal Institute of Technology
AlbaNova University Center, 106 91 Stockholm (Sweden)
[+] These authors contributed equally.
In the present study, we initially used compound 1a
(Table 1) to optimize a KR protocol. Reactions were per-
formed at 408C in tert-butyl methyl ether (TBME) with
phenyl acetate as an acyl donor and the lipase from
Burkholderia cepacia (previously Pseudomonas cepacia) as
the resolving agent. The reaction was monitored by 1H NMR
[**] This study was supported by the Swedish Research Council and the
Royal Institute of Technology. We thank Toyo Denka Kogyo Co. and
Dr. Y. Yamashita for providing samples for this study.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201007373.
6592
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 6592 –6595

Table 1: Dynamic kinetic resolution of N-methyl a-aminonitriles 1a–c
enantioselectivity for the formation of the acylated product.
This low enantioselectivity was explained in terms of unex-
pected turnover-related racemization through abstraction of
the a hydrogen atom.[20] We thus tested compound 1c with a
variety of organic acids and bases together with D2O to probe
the potential deuterium exchange at the a position. However,
no exchange was observed under these conditions, even with
strong acids or bases at elevated temperatures.[28] Next, a-
deuterated 1c (d-1c) was prepared and subjected to the DKR
conditions (Scheme 2a). Residual water and/or protein acidic
through the lipase-catalyzed amidation/racemization one-pot process.[a]
Entry
R
Product
t [days]
Yield[b] [%]
ee[d] [%]
(Conv.[c] [%])
1
(t = [days])
2
1
2
3
H
OCH3
F
2a
2b
2c
8 (1.34)
7 (1.13)
13 (2.36)
90 (94)
89 (92)
87 (92)
85
86
88
[a] Reactions were carried out with 1 (0.05 mmol), lipase PS-C I
(400 mg), and phenyl acetate (0.15 mmol) in TBME (1.5 mL) at 408C.
[b] Yield of the isolated product. [c] The conversion was determined by
1H NMR spectroscopy. [d] The ee value was determined by HPLC
analysis on an OD-H or OJ column.
spectroscopy and HPLC. Surprisingly, conversion into the
final product 2a exceeded the maximum KR yield but still
proceeded with high enantioselectivity. Thus, after two days, a
conversion of 60% and an enantiomeric purity of 88% ee
were recorded, whereas the remaining starting material 1a
remained racemic. The process reached completion (94%
conversion into 2a with 85% ee) after a reaction time of eight
days (Table 1, entry 1). This result clearly indicates that the
transformation of compound 1a into 2a occurred through a
DKR process, rather than the expected KR process, under the
conditions used. Other N-methyl a-aminonitrile derivatives,
1b and 1c, were subsequently subjected to the same reaction
conditions. The results for these structures were similar to
those for compound 1a, although the reaction rates were
different. Similar rates of DKR were observed for compounds
Scheme 2. Control experiments supporting the proposed lipase-medi-
ated retro-Strecker racemization mechanism.
1
1
1a (t = = 1.34 days) and 1b (t = = 1.13 days), which contain
2
2
unsubstituted or electron-rich aromatic moieties (Table 1,
entries 1 and 2). In contrast, compound 1c, with an electron-
withdrawing aromatic substituent, showed the slowest rate
hydrogen atoms in the system would then serve as proton
donors. The same yield and enantioselectivity were observed
as for 1c, and no deuterium/proton exchange was observed
for compound d-1c or the corresponding product d-2c.[24] It
has also been proposed that aldehydes may catalyze this
process. Therefore, deuterium-exchange experiments were
carried out in the presence of 4-fluorobenzaldehyde (up to
30%). The presence of the aldehyde did not, however, lead to
any exchange (and hence to racemization) under these
conditions. These results reject the first hypothesis that the
racemization mechanism will pass through abstraction of the
a-proton of the starting Strecker compounds, where neither
enzymatic nor non-enzymatic processes are productive.
To test the second hypothesis, we prepared solutions of
the nonracemic compound 1c (30% ee) in situ and subjected
them to three different potential sets of racemization
conditions (see the Supporting Information for details). The
first experiment was performed under non-enzymatic con-
ditions. The solution of compound 1c (30% ee) was thus
allowed to stand at 408C for up to five days. No racemization
was observed under these conditions during this period
(Scheme 2b). In contrast, when the solution of compound
1c (30% ee) was treated with the B. cepacia lipase at the same
1
(t = = 2.36 days); the reaction required 13 days to reach
2
completion (Table 1, entry 3).[28] A similar longer reaction
time was observed for the corresponding 3-methoxy-substi-
tuted aromatic N-methyl a-aminonitrile derivative, the reac-
tion of which also resulted in lower enantiomeric purity (see
the Supporting Information). The reaction of primary amines
resulted in good conversions, but no enantioselectivity was
observed under these conditions.
In principle, racemization of the starting compounds 1a–c
must be a key step in the DKR process. Two possible
mechanisms were hypothesized for this racemization:
1) abstraction of the a hydrogen atom of the starting N-
methyl a-aminonitrile by a base, or 2) in situ bond cleavage
through a retro-Strecker reaction. We performed various
control experiments to challenge these hypotheses. First, the
stability of the a hydrogen atom of the starting compounds
was examined. Our previous studies demonstrated that the
a hydrogen atom of the final products is sensitive to the
presence of a base and relatively easily abstracted, which
results in racemization.[17] Moreover, previous KR studies
with Candida antarctica lipase B showed generally low
Angew. Chem. Int. Ed. 2011, 50, 6592 –6595
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6593

Communications
temperature, racemization occurred. As expected, the
ee value dropped from 30 to 10% in two days; 0% ee was
reached within three days (Scheme 2c). These results indicate
that the racemization process is exclusively mediated by the
lipase.
To elucidate the importance of the active site of the
enzyme, we performed a third racemization test. The lipase
was thus first inactivated by methyl 4-nitrophenyl hexyl-
phosphonate (MNPHP), an established irreversible inhibitor
of lipases.[5c,29] The resulting enzyme preparation exhibited no
catalytic activity under the acylation conditions used. The
inhibited enzyme was then treated with the solution of
compound 1c (30% ee) at 408C for three days. No racemi-
zation was observed (Scheme 2d), which implies that the
racemization and the transacylation processes take place at
the same site of the enzyme. In addition to these studies,
where a commercial preparation of B. cepacia was used, a
preparation involving a pure lipase sample immobilized on
the same carrier was analyzed.[28] As expected, the same
results were observed with this preparation, and it could be
concluded that the racemase activity emanated from the
lipase. From these control experiments, the racemase activity
could be confirmed. The results also indicate that the
transformation occurs specifically with the starting Strecker
compounds, a conclusion further supported by experiments in
which cyanohydrins were used instead of Strecker compounds
under the DKR conditions; in this case, no racemization was
observed. Other structures based on the N-substituted a-
aminonitrile core structure are, however, suitable substrates,
which extends the scope of the reaction. In addition, the
racemization can also be observed using other lipases, both
immobilized and non-immobilized, not affecting the acylation
selectivity, in contrast to the results for Candida antarctica
lipase B.[20] All these results support the hypothetical lipase-
mediated retro-Strecker racemization mechanism (see Fig-
ure S12 in the Supporting Information for more details).
Additional support for the racemase-type activity was
acquired from molecular dynamics (MD) simulations and
density functional theory (DFT) calculations. MD simulations
were performed for both aminonitrile enantiomers com-
plexed to the B. cepacia lipase.[28] The results showed that
both enantiomers are tightly bound to the oxyanion hole
through hydrogen bonds to the cyano group (Figure 1a).
Strong hydrogen bonding is also observed between the
Ne atom of His286 and the substrate amino group (S3). This
interaction locks the substrates in a rather rigid position in the
active site. Moreover, a fluctuating hydrogen bond observed
between Ser87 Og and the a nitrogen atom (S4) may play a
role in the transition state. The hydrogen-bond network of
each enantiomer is consistent with a general-base-catalyzed
elimination mechanism in which His286 abstracts the proton
of the amino group, and the transition state and the
intermediate cyanide ion are stabilized by the oxyanion hole.
We subsequently performed hybrid DFT calculations to
validate the reaction mechanism.[28] The model system, taken
from the simulation for the R aminonitrile enantiomer, is
displayed in Figure 1b and shows the optimized geometry of
the transition state at the B3LYP/6-31 + G(d) level. The
calculations indicate a concerted mechanism in which proton
Figure 1. a) Snapshot of MD simulations of the S (left) and R amino-
nitriles (right) with selected active-site residues. Hydrogen bonds are
depicted as thin lines. b) Model transition state (R aminonitrile).
Annotated distances [ꢁ]: TS1=2.02, TS2=1.79, TS3=2.89,
TS4=2.05, TS5=1.62, TS6=1.91 (see the Supporting Information for
details).
abstraction from the amino group precedes the leaving of the
cyanide ion. The whole catalytic machinery of the lipase is
active; however, Ser87 evidently acts more like an auxiliary
hydrogen-bond donor than a nucleophile, as predicted by the
MD simulations. The activation energy was calculated to be
21 kcalmolÀ1 for the rate-limiting transition state. This value
is in reasonable agreement with the experimentally observed
rate.
In conclusion, we have discovered that a wild-type lipase
enzyme shows racemase-type activity, specifically for the N-
substituted a-aminonitrile (Strecker) core structure. The
racemization mechanism is proposed to proceed through a
À
C C bond-breaking/forming retro-Strecker/Strecker reaction
catalyzed at the hydrolase active site. Moreover, we have
demonstrated a convenient approach to the asymmetric
synthesis of N-methyl a-aminonitrile derivatives by a lipase-
catalyzed direct one-pot reaction that relies on the combined
racemase activity and normal lipase transesterification activ-
ity of the enzyme. The spontaneous communication between
6594 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 6592 –6595

[8] L. K. Thalꢃn, D. Zhao, J.-B. Sortais, J. Paetzold, C. Hoben, J.-E.
Bꢆckvall, Chem. Eur. J. 2009, 15, 3403 – 3410.
[9] V. Gotor-Fernꢂndez, E. Busto, V. Gotor, Adv. Synth. Catal. 2006,
348, 797 – 812.
[10] F. van Rantwijk, R. A. Sheldon, Tetrahedron 2004, 60, 501 – 519.
[11] G. Asensio, C. Andreu, J. A. Marco, Tetrahedron Lett. 1991, 32,
4197 – 4198.
these two processes through a single catalyst led to the desired
products in good yields with high enantiomeric purity.
Received: November 23, 2010
Revised: April 27, 2011
Published online: June 1, 2011
[12] T.-W. Chiou, C.-C. Chang, C.-T. Lai, D.-F. Tai, Bioorg. Med.
Chem. Lett. 1997, 7, 433 – 436.
[13] G. F. Breen, Tetrahedron: Asymmetry 2004, 15, 1427 – 1430.
[14] V. Gotor-Fernꢂndez, P. Fernꢂndez-Torres, V. Gotor, Tetrahe-
dron: Asymmetry 2006, 17, 2558 – 2564.
[15] A. J. Blacker, M. J. Stirling, M. I. Page, Org. Process Res. Dev.
2007, 11, 642 – 648.
[16] A. Liljeblad, H.-M. Kavenius, P. Taehtinen, L. T. Kanerva,
Tetrahedron: Asymmetry 2007, 18, 181 – 191.
Keywords: dynamic kinetic resolution · enzyme catalysis ·
racemase activity · secondary amines · Strecker reaction
.
[1] R. Noyori, Nat. Chem. 2009, 1, 5 – 6.
[2] M. Breuer, K. Ditrich, T. Habicher, B. Hauer, M. Keßeler, R.
Stꢀrmer, T. Zelinski, Angew. Chem. 2004, 116, 806 – 843; Angew.
Chem. Int. Ed. 2004, 43, 788 – 824.
[3] A. Ghanem, Tetrahedron 2007, 63, 1721 – 1754.
[4] H. E. Schoemaker, D. Mink, M. G. Wubbolts, Science 2003, 299,
1694 – 1697.
[17] P. Vongvilai, O. Ramstrꢄm, J. Am. Chem. Soc. 2009, 131, 14419 –
14425.
[18] H. Grꢄger, Chem. Rev. 2003, 103, 2795 – 2827.
[19] K. Nakai, J. Hiratake, J. i. Oda, Bull. Inst. Chem. Res. 1992, 70,
333 – 337.
[5] For recent accounts, see: a) D. Liu, P. Trodler, S. Eiben, K.
Koschorreck, M. Mꢀller, J. Pleiss, S. C. Maurer, C. Branneby,
R. D. Schmid, B. Hauer, ChemBioChem 2010, 11, 789 – 795;
b) N. Rꢁos-Lombardꢁa, E. Busto, V. Gotor-Fernꢂndez, V. Gotor,
R. Porcar, E. Garcꢁa-Verdugo, S. V. Luis, I. Alfonso, S. Garcꢁa-
Granda, A. Menꢃndez-Velꢂzquez, Chem. Eur. J. 2010, 16, 836 –
847; c) M. Vallin, P.-O. Syrꢃn, K. Hult, ChemBioChem 2010, 11,
411 – 416; d) T. Ema, S. Kamata, M. Takeda, Y. Nakano, T. Sakai,
Chem. Commun. 2010, 46, 5440 – 5442; e) S. Akai, R. Hanada, N.
Fujiwara, Y. Kita, M. Egi, Org. Lett. 2010, 12, 4900 – 4903;
f) L. H. Andrade, T. Barcellos, Org. Lett. 2009, 11, 3052 – 3055;
g) R. O. de Souza, O. A. C. Antunes, W. Kroutil, C. O. Kappe, J.
Org. Chem. 2009, 74, 6157 – 6162; h) P. Vongvilai, R. Larsson, O.
Ramstrꢄm, Adv. Synth. Catal. 2008, 350, 448 – 452; i) M.-J. Kim,
Y. K. Choi, S. Kim, D. Kim, K. Han, S.-B. Ko, J. Park, Org. Lett.
2008, 10, 1295 – 1298; j) P. Vongvilai, M. Angelin, R. Larsson, O.
Ramstrꢄm, Angew. Chem. 2007, 119, 966 – 968; Angew. Chem.
Int. Ed. 2007, 46, 948 – 950; k) N. M. T. Lourenꢅo, C. A. M.
Afonso, Angew. Chem. 2007, 119, 8326; Angew. Chem. Int. Ed.
2007, 46, 8178 – 8181; l) Y. Zhu, K.-L. Fow, G.-K. Chuah, S.
Jaenicke, Chem. Eur. J. 2007, 13, 541 – 547; m) B. A. C. van As, J.
van Buijtenen, T. Mes, A. R. A. Palmans, E. W. Meijer, Chem.
Eur. J. 2007, 13, 8325 – 8332.
[20] P. Lꢇpez-Serrano, J. A. Jongejan, F. van Rantwijk, R. A. Shel-
don, Tetrahedron: Asymmetry 2001, 12, 219 – 228.
[21] Racemization effects were observed for cyanoalkyl acetamides
when preparations of Candida antarctica lipase B were used in
the presence of a zeolite.[20]
[22] U. T. Bornscheuer, R. J. Kazlauskas, Angew. Chem. 2004, 116,
6156 – 6165; Angew. Chem. Int. Ed. 2004, 43, 6032 – 6040.
[23] A. Aharoni, L. Gaidukov, O. Khersonsky, S. M. Gould, C.
Roodveldt, D. S. Tawfik, Nat. Genet. 2005, 37, 73 – 76.
[24] K. Hult, P. Berglund, Trends Biotechnol. 2007, 25, 231 – 238.
[25] H. Pellissier, Tetrahedron 2008, 64, 1563 – 1601.
[26] A. H. Kamaruddin, M. H. Uzir, H. Y. Aboul-Enein, H. N. A.
Halim, Chirality 2009, 21, 449 – 467.
[27] B. Martꢁn-Matute, J.-E. Bꢆckvall, Curr. Opin. Chem. Biol. 2007,
11, 226 – 232.
[28] See the Supporting Information for more details.
[29] a) P. Carlqvist, M. Svedendahl, C. Branneby, K. Hult, T. Brinck,
P. Berglund, ChemBioChem 2005, 6, 331 – 336; b) C. Branneby,
P. Carlqvist, A. Magnusson, K. Hult, T. Brinck, P. Berglund, J.
Am. Chem. Soc. 2003, 125, 874 – 875; c) M. Persson, E. Wehtje, P.
Adlercreutz, ChemBioChem 2002, 3, 566 – 571; d) C. Bovet, R.
Zenobi, Anal. Biochem. 2008, 373, 380 – 382; e) D. Rotticci, T.
Norin, K. Hult, M. Martinelle, Biochim. Biophys. Acta Mol. Cell
Biol. Lipids 2000, 1483, 132 – 140.
[6] M. Rodrꢁguez-Mata, E. Garcꢁa-Urdiales, V. Gotor-Fernꢂndez, V.
Gotor, Adv. Synth. Catal. 2010, 352, 395 – 406.
[7] M. Shakeri, K. Engstrꢄm, A. G. Sandstrꢄm, J.-E. Bꢆckvall,
ChemCatChem 2010, 2, 534 – 538.
Angew. Chem. Int. Ed. 2011, 50, 6592 –6595
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6595