Ionic-surfactant-coated Burkholderia cepacia lipase as a highly active and enantioselective catalyst for the dynamic kinetic resolution of secondary alcohols

Article abstract of DOI:10.1002/anie.201104141

With a coat for activity: A highly active enzyme was prepared by coating Burkholderia cepacia lipase with an ionic surfactant for use in dynamic kinetic resolution (DKR). Important features of this enzyme include: the fastest DKR of 1-phenylethanol, the h

Full text of DOI:10.1002/anie.201104141

Communications
DOI: 10.1002/anie.201104141
Dynamic Kinetic Resolution
Ionic-Surfactant-Coated Burkholderia cepacia Lipase as a Highly
Active and Enantioselective Catalyst for the Dynamic Kinetic
Resolution of Secondary Alcohols**
Hyunjin Kim, Yoon Kyung Choi, Jusuk Lee, Eungyeong Lee, Jaiwook Park,* and Mahn-
Joo Kim*
Dedicated to Professor Eun Lee on the occasion of his retirement
Dynamic kinetic resolution (DKR) is among the most
practical methods for the transformation of racemates to
single enantiomers.[1] DKR can provide high yields and
excellent enantiopurities, both approaching 100%, which
are difficult to achieve with classical kinetic resolution
(KR).[2] In the past decade, enzymatic resolution coupled
with metal-catalyzed racemization has been explored as a new
strategy for DKR. Now several procedures are available for
the DKR of racemic alcohols[3] and amines.[4] The wider
applications of these DKR procedures in organic synthesis,
however, can be hampered by the narrow specificity, moder-
ate enantioselectivity, and low activity of the enzyme
employed. Therefore, it is of great importance to develop a
highly active enzyme having high enantioselectivity toward a
wide range of substrates for DKR. We herein report that
significantly lower activity than Candida antarctica lipase B
(CALB; brand name: Novozym 435) (Table 1, entries 1–3);
the latter has been most frequently employed in the metallo-
enzymatic DKR. Fortunately, we have developed a practical
approach for enhancing its activity dramatically. In a typical
procedure, buffer-free soluble BCL (24% protein),[5] pre-
pared from its crude preparation (Lipase PS), was lyophilized
in the presence of an ionic surfactant 4 (Scheme 1) to yield
ISCBCL. The activity of ISCBCL was dependent on the
ionic-surfactant-coated
Burkholderia
cepacia
lipase
(ISCBCL) has great potential as such an excellent enzyme.
Commercially available Burkholderia cepacia lipase
(BCL; brand names: Lipase PS and Lipase PS-C) displays
Table 1: Activities of lipases.[a]
Entry
Lipase
Amount
of 4 [wt%]
Activity[b]
Activation
[mMhÀ1
]
Scheme 1. Preparation of ionic surfactant.
1
2
3
4
5
6
Lipase PS[c]
Lipase PS-C[d]
Novozym 435
ISCBCL[e]
–
–
–
13
13
23
0.24
1.1ꢀ10
1.2ꢀ102
6.4ꢀ10
1.4ꢀ102
2.5ꢀ102
2.7ꢀ102
6.0ꢀ102
1.0ꢀ103
amount of 4 added and the type of solvent system used for
lyophilization (Table 1, entries 4–6). The most active ISCBCL
(18% protein), lyophilized in the presence of 0.30 mass equiv
(23 wt%) of 4 in 1:1 (v/v) water–dioxane cosolvent system,
displayed three orders of magnitude higher activity[6] than its
crude counterpart (Table 1, entry 6). It was more active than
Novozym 435 (20% protein).
The ISCBCL showed excellent performance in the DKRs
of 1-phenylethanol with three ruthenium complexes (7–9) as
the racemization catalysts (Table 2). The ISCBCL-catalyzed
DKR with complex 7[7] proceeded most rapidly and reached
completion at room temperature within 1 hour (Table 2,
entry 1). The corresponding DKR with complexes 8[8] and 9[9]
also proceeded well but the reactions were complete in 2 h
(Table 2, entries 2 and 3, respectively). Previously, the fastest
DKR of 1-phenylethanol took 3 h and was performed with
Novozym 435 and 7 or 9.[7,9] It has been known that
Novozym 435 impedes the ruthenium-catalyzed racemiza-
ISCBCL[f]
ISCBCL[f]
[a] The activities were measured with solutions containing 1-phenyl-
ethanol (0.50m), enzyme (1.0 mg), and isopropenyl acetate (1.5 equiv)
in toluene at 258C. Rate measurements were done twice for each enzyme
preparation. [b] Initial rate per mg of enzyme preparation. [c] Crude
enzyme. [d] Ceramic-supported enzyme. [e] Lyophilized in 1:1 (v/v)
water–THF. [f] Lyophilized in 1:1 (v/v) water–dioxane.
[*] H. Kim, Y. K. Choi, J. Lee, E. Lee, J. Park, Prof. M.-J. Kim
Department of Chemistry
Pohang University of Science and Technology (POSTECH)
San-31 Hyojadong, Pohang 790-784 (Republic of Korea)
E-mail: mjkim@postech.ac.kr
[**] This work was supported by the National Research Foundation of
Korea (KRF-2007-511-C00031 and 2009-0074357).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201104141.
10944
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 10944 –10948

Table 2: DKR of 1-phenylethanol with ISCBCL.[a]
Table 3: DKR of secondary alcohols with ISCBCL.
Entry
Substr.
Ru cat.
Prod.
Yield [%][a]
ee [%][b]
1
2
3
4
5
6
7
8
10a
10b
10c
10d
10e
10 f
10g
10h
10i
7
7
7
7
7
7
11a
11b
11c
11d
11e
11 f
11g
11h
11i
96
98
94
90
98
95
96
96
97
97
96
93
95
96
97
94
97
96
97
97
96
93
96
91
98
95
98
97
96
99
98
95
97
96
98
96
98
97
97
>99
82
82
81
96
99
99
99
99
Entry
Ru cat.
t [h]
Yield [%][b]
ee [%][c]
8[c]
8[c]
8[c]
8[c]
8[c]
8[c]
8[c]
8[c]
8[c]
8[c]
7
7
7
7
7
7
7
7
1
2
3
7
8
9
1
95
75 (95)
85 (97)
99
99
99
1 (2)
1 (2)
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
10j
10k
10l
11j
11k
11l
[a] Performed with solutions containing substrate (1.0 mmol), ISCBCL
(23 wt% 4, 10 mg), Ru catalyst (4.0 mol%), base (1 equiv), and
isopropenyl acetate (IPA, 1.5 equiv) in toluene (2.0 mL) at RT. [b] Yield of
isolated product. [c] Determined by HPLC using a chiral stationary
phase.
10m
10n
10o
10p
10q
10r
10s
10t
10u
10v
10w
10x
11m
11n
11o
11p
11q
11r
11s
11t
11u
11v
11w
11x
tion. In contrast, no significant decrease in racemization rate
was observed with ISCBCL. All the results indicate that
ISCBCL is more efficient than Novozym 435 as the enzyme
for DKR.
To see the scope of ISCBCL-catalyzed DKR, 24 different
secondary alcohols were chosen as the substrates (10a–x,
Figure 1). They carry an aliphatic chain and an aromatic ring
at the hydroxymethine center (RCH(OH)Ar). The length of
[a] Yield of isolated product. [b] Determined by HPLC using a chiral
stationary phase. [c] The use of 7 resulted in the formation of a
significant amount of byproducts.
substrates tested are accepted by ISCBCL at synthetically
useful rates and most of them are transformed with high
enantioselectivity. Relatively lower enantiopurities (81–
82% ee) were observed for the DKRs of substrates carrying
a longer aliphatic chain (Table 3, entries 17–19). In contrast to
this, the DKRs of substrates with iPr or tBu at the para
position of phenyl ring gave the highest enantiopurities
(99% ee or greater) regardless of the length of the aliphatic
chain (Table 3, entries 6, 16, and 21–24). In general the
enantioselectivity of ISCBCL decreased with the length of the
aliphatic chain (compare entries 1, 7, 17, and 18) and
increased with increasing bulk of the para substituent on the
phenyl ring (compare entries 18–22).
All the substrates shown in Figure 1, which have a linear
aliphatic side chain and an aromatic ring at the hydroxyme-
thine center, are transformed with the same R enantioselec-
tivity[12] in DKR. It would be nice if the enantioselectivity
could be switched for the preparation of the opposite
enantiomers. We hypothesized that the enantioselectivity
should be changed if the aliphatic side chain is sterically
enlarged relative to the aromatic ring. To test this hypothesis,
a-phenylpropargyl alcohol (12a) and its trimethylsilyl (TMS)
derivative 13a were chosen as model substrates (Figure 2). It
was found that the DKR of 13a[13] followed by desilylation
with tetrabutylammonium fluoride (TBAF) yielded (S)-14a,
the enantiomer of (R)-14a produced by the enzymatic
acylation of 12a (Scheme 2). These results prove that the
Figure 1. Substrates for DKR.
aliphatic chain ranges widely from three-carbon units such as
chloromethyl (10a–f) and allyl (10g–o) to ten-carbon unit
such as n-decyl (10x). To the best of our knowledge, their
DKRs have not been studied previously. It is noted that
Novozym 435 is practically inapplicable to their DKRs owing
to its narrow specificity.[10]
The DKR reactions of 10a–x were carried out with
solutions containing substrate (0.1–0.3 mmol), ISCBCL (10–
30 mgmmolÀ1 substrate), ruthenium complex (7 or 8, 4.0–
5.0 mol%), K2CO3 or Na2CO3 (1 equiv), and isopropenyl
acetate (IPA, 1.5 equiv) in toluene at room temperature
(Table 3). Most of the DKR reactions were complete within
24 h except those of 10e, 10 f, and 10v, which took 48 h. In
most cases excellent yields and high enantiopurities were
obtained (Table 3).[11] These results indicate that all the
Angew. Chem. Int. Ed. 2011, 50, 10944 –10948
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10945

Communications
Figure 2. a-Arylpropargyl alcohols.
Figure 3. Active-site model of ISCBCL. A,C) Binding modes of more
reactive enantiomers. B,D) Binding modes of less reactive enantio-
mers.
substrates can be accepted by ISCBCL with high enantio-
selectivity at synthetically useful rates.
The switchable enantioselectivity of ISCBCL can be
explained using a model of the active site based on the X-
ray structure of BCL[16] (Figure 3). There are three binding
pockets (HA, HB, and HH) at the active site of BCL for
anchoring three substituents at the stereocenter of sub-
strate. Among them, the HH binding pocket seems to play
an essential role in determining which enantiomer can be
accepted more favorably. It has two rooms, a hydrophilic
trench, and its entrance (a space of 4.5 ꢀ in diameter),
which are separated by a contraction.[16a] Small aliphatic
groups such as methyl and ethynyl can be nicely fitted into
the pocket (Figure 3A). The binding of flat aromatic rings
such as benzene and naphthalene into the pocket is possible
(Figure 3B,C) but with some difficulty because their
increased sizes should give rise to some repulsive effects
around the contraction. The binding of branched and bulky
aliphatic groups such as TMS-ethynyl into the same pocket is
expected to be even more difficult owing to severe steric
repulsion around the contraction (Figure 3D). Therefore, the
enantiomers shown in Figure 3A and C react much faster
than their respective antipodes in Figure 3B and D. The
perfect enantioselectivity of ISCBCL for para-isopropyl- and
para-tert-butylphenylalkanols (10 f, 10p, 10u–x) also can be
rationalized using the active-site model. The linear aliphatic
groups can be fitted into the HH pocket without difficulty
while the branched alkyl-substituted aromatic rings are
Scheme 2. KR and DKR of a-arylpropargyl alcohols with ISCBCL.
enantioselectivity was switched between 12a and 13a.[14] The
switchable enantioselectivity was also confirmed between 12b
and 13b (Scheme 2). The DKRs of TMS-protected a-
arylpropargyl alcohols (13a–i) provided excellent enantio-
purities approaching 100%[15] with high yields (Table 4),
indicating that the switching of enantioselectivity is nearly
perfect. The results also imply that sterically demanding
Table 4: DKR of TMS-protected propargyl alcohols with ISCBCL.[a]
Entry
Substr.
Prod.
Yield [%][b]
ee [%][c]
1
2
3
4
5
6
7
8
9
13a
13b
13c
13d
13e
13 f
13g
13h
13i
15a
15b
15c
15d
15e
15 f
15g
15h
15i
94
89
91
90
91
94
92
90
91
>99
98
96
96
96
>99
>99
98
98
Figure 4. Three different types of secondary alcohols (A–C) and their
enantiomers (D–F) reacting more rapidly in the ISCBCL-catalyzed
DKRs than the others (R=aliphatic, Ar=aromatic, L=linear, and
B=branched and bulky).
[a] See the Supporting Information for a representative procedure.
[b] Yield of isolated product. [c] Determined by HPLC using a chiral
stationary phase.
10946 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 10944 –10948

[2] a) Hydrolases in Organic Synthesis, 2nd ed. (Eds.: U. T. Born-
scheuer, R. J. Kazlauskas), Wiley-VCH, Weinheim, 2006;
b) Asymmetric Organic Synthesis with Enzymes (Eds.: V.
Gotor, I. Alfonso, E. Garcia-Urdiales), Wiley-VCH, Weinheim,
2008.
unable to enter the pocket owing to severe steric repulsion
near the contraction.
All the substrates studied in this work can be classified
into two types, A and B, according to the shape of the
aliphatic substituent (Figure 4). The results from their DKRs
and the active-site model in Figure 3 indicate that the
corresponding enantiomers D and E should react more
rapidly than their antipodes in the ISCBCL-catalyzed DKR.
For the third type of substrates C with two aliphatic
substituents at the hydroxymethine center, which have not
been studied in this work, F should be the more reactive
enantiomer. The structural features of D, E, and F can serve
as the rules for predicting the enantioselectivity in the
ISCBCL-catalyzed DKRs of other secondary alcohols. They
also provide guidelines for the rational design of substrates
for highly enantioselective DKR.
In conclusion, we have developed a highly active enzyme
by coating Burkholderia cepacia lipase with the ionic surfac-
tant 4 for use in DKR. With this enzyme (ISCBCL), we were
able to achieve the fastest DKR of 1-phenylethanol, the
highly enantioselective DKR of a wide range of secondary
alcohols (RCH(OH)Ar, R = C3–C10) previously unexplored,
and, for the first time, the switching of lipase enantioselec-
tivity in DKR. The switchable enantioselectivity of ISCBCL
depending on the shape of aliphatic chain (R) makes it
possible to prepare selectively either R or S enantiomers by
DKR. We have also demonstrated that the ISCBCL-cata-
lyzed DKRs provide routes to esters of enantioenriched g-
chlorohydrins and homoallyl and propargyl alcohols, which
are useful as versatile building blocks in asymmetric syn-
thesis.[17] Overall, ISCBCL appears to be highly promising as
the catalyst for the DKR of secondary alcohols.
[3] For DKR of alcohols, see: a) S.-B. Ko, B. Baburaj, M.-J. Kim, J.
Park, J. Org. Chem. 2007, 72, 6860 – 6864; b) M.-J. Kim, Y. K.
Choi, S. Kim, D. Kim, K. Han, S.-B. Ko, J. Park, Org. Lett. 2008,
10, 1295 – 1298; c) R. M. Haak, F. Berthiol, T. Jerphagnon, J. A.
Gayet, C. Tababiono, C. P. Postema, V. Ritleng, M. Pfeffer, D. B.
Janssen, A. J. Minnaard, B. L. Feringa, J. G. De Vries, J. Am.
Chem. Soc. 2008, 130, 13508 – 13509; d) Q. Chen, C. Yuan, Chem.
Commun. 2008, 5333 – 5335; e) Y. Do, I. C. Hwang, M.-J. Kim, J.
Park, J. Org. Chem. 2010, 75, 5740 – 5742; f) S. Akai, R. Hanada,
N. Fujiwara, Y. Kita, M. Egi, Org. Lett. 2010, 12, 4900 – 4903;
g) L. K. Thalꢁn, A. Sumic, K. Bogar, J. Norinder, A. K. A.
Persson, J.-E. Bꢂckvall, J. Org. Chem. 2010, 75, 6842 – 6847.
[4] For DKR of amines, see: a) M. A. J. Veld, K. Hult, A. R. A.
Palmans, E. W. Meijer, Eur. J. Org. Chem. 2007, 5416 – 5421;
b) A. N. Parvulescu, P. A. Jacobs, D. E. De Vos, Chem. Eur. J.
2007, 13, 2034 – 2043; c) M.-J. Kim, W.-H. Kim, K. Han, Y. K.
Choi, J. Park, Org. Lett. 2007, 9, 1157 – 1159; d) A. N. Parvulescu,
P. A. Jacobs, D. E. De Vos, Adv. Synth. Catal. 2008, 350, 113 –
121; e) K. Han, J. Park, M.-J. Kim, J. Org. Chem. 2008, 73, 4302 –
4304; f) L. H. Andrade, A. V. Silva, E. C. Pedrozo, Tetrahedron
Lett. 2009, 50, 4331 – 4334; g) K. Han, Y. Kim, M.-J. Kim,
Tetrahedron Lett. 2010, 51, 3536 – 3537; h) Y. Kim, M.-J. Kim,
Tetrahedron Lett. 2010, 51, 5581 – 5584.
[5] The protein content was determined using the Bradford method.
M. M. Bradford, Anal. Biochem. 1976, 72, 248 – 254.
[6] It is assumed that the high activity of ISCBCL may result from
two effects of the coating: the protection of enzymes against
deactivation during lyophilization and the better dispersion of
enzymes in organic solvent. It may be also speculated that the
coating could influence the enzymatic activity by changing the
pH and water content of enzyme preparation. It was observed
that the changes of the pH and water content by coating were not
significant. The pH of BCL preparation increased slightly from
6.4 before coating to 6.8 after coating while the water content,
determined by the Karl-Fisher titration, decreased from 11–12%
to 7–10%.
Experimental Section
Preparation of ISCBCL:
A suspension of crude lipase (6.0 g,
Lipase PS from Amano) in a 0.1m phosphate buffer (50 mL, pH 7.8)
was stirred for 10 min at room temperature. Insoluble materials were
removed by centrifugation using a COMBI-514R centrifuge at 48C.
[7] J. H. Lee, N. Kim, M.-J. Kim, J. Park, ChemCatChem 2011, 3,
354 – 359.
[8] a) J. H. Choi, Y. H. Kim, S. H. Nam, S. T. Shin, M.-J. Kim, J.
Park, Angew. Chem. 2002, 114, 2479 – 2482; Angew. Chem. Int.
Ed. 2002, 41, 2373 – 2376; b) J. H. Choi, Y. K. Choi, Y. H. Kim,
E. S. Park, E. J. Kim, M.-J. Kim, J. Park, J. Org. Chem. 2004, 69,
1972 – 1977.
The aqueous solution was then dialyzed using
a Spectra/Por
Membrane (MWCO: 10 K) against water for 2 days and then
freeze-dried to give a white enzyme powder ( ꢀ 100 mg). For the
preparation of ISCBCL, the enzyme powder (50 mg) was dissolved in
deionized water (15 mL) and then mixed with 4 (7.5–15 mg, 0.15–0.30
mass equivalent) dissolved in THF or dioxane (15 mL). The resulting
solution was freeze-dried to yield ISCBCL, which was then used for
kinetic and dynamic kinetic resolution.
[9] B. Martꢃn-Matute, M. Edin, K. Bogꢄr, F. B. Kaynak, J.-E.
Bꢂckvall, J. Am. Chem. Soc. 2005, 127, 8817 – 8825.
[10] Novozym 435 accepts a limited range of secondary alcohols
carrying a small and a significantly larger substituent at the
hydroxymethine center with high enantioselectivity. In addition,
the small substituent of secondary alcohol should be smaller than
three-carbon unit for transformation at synthetically useful rate.
[11] It was observed from the NMR spectra of product mixtures that
some of ionic surfactant in ISCBCL went into solution during
the reaction.
Received: June 16, 2011
Published online: September 27, 2011
Keywords: alcohols · catalysts · dynamic kinetic resolution ·
.
lipases · ruthenium
[12] a) R. J. Kazlauskas, A. N. E. Weissfloch, A. T. Rappaport, L. A.
Cuccia, J. Org. Chem. 1991, 56, 2656 – 2665; b) Q. Jing, R. J.
Kazlauskas, Chirality 2008, 20, 724 – 735.
[13] Here, we used ISCBCL lyophilized in the presence of 13 wt% 4
in 1:1 (v/v) water–THF because more active ISCBCLs were
developed later.
[14] This can be also observed with coating-free BCL, but the
reactions are expected to be very slow. In a previous study, the
resolution of 3a with Lipase PS took 320 h to reach 51%
[1] a) J. H. Lee, K. Han, M.-J. Kim, J. Park, Eur. J. Org. Chem. 2010,
999 – 1015; b) Y. Kim, J. Park, M.-J. Kim, ChemCatChem 2011, 3,
271 – 277; c) H. Pellissier, Tetrahedron 2008, 64, 1563 – 1601; H.
Pellissier, Tetrahedron 2011, 67, 3769 – 3802; H. Pellissier, Adv.
Synth. Catal. 2011, 353, 659 – 676.
Angew. Chem. Int. Ed. 2011, 50, 10944 –10948
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10947

Communications
conversion at 258C (P. Allevi, P. Ciuffreda, M. Anastasia,
Tetrahedron: Asymmetry 1997, 8, 93 – 99).
[17] a) S. Denmark, J. Fu, Chem. Rev. 2003, 103, 2763 – 2793; b) L.
Tietze, T. Kinzel, C. C. Brazel, Acc. Chem. Res. 2009, 42, 367 –
378; c) B. M. Trost, A. H. Weiss, Adv. Synth. Catal. 2009, 351,
963 – 983.
[15] This is compared to a low enantioselectivity (E = 16) observed in
the Lipase PS-catalyzed resolution described in Ref. [14].
[16] a) T. Schulz, J. Pleiss, R. D. Schmid, Protein Sci. 2000, 9, 1053 –
1062; b) D. Guieysse, C. Salagnad, P. Monsan, M. Remaud-
Simeon, V. Tran, Tetrahedron: Asymmetry 2003, 14, 1807 – 1817.
10948 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 10944 –10948