Highly Focused Library-Based Engineering of Candida antarctica Lipase B with (S)-Selectivity Towards sec-Alcohols

Article abstract of DOI:10.1002/adsc.201800711

Candida antarctica lipase B (CALB) is one of the most extensively used biocatalysts in both academia and industry and exhibits remarkable (R)-enantioselectivity for various chiral sec-alcohols. Considering the significance of tailor-made stereoselectivity in organic synthesis, a discovery of enantiocomplementary lipase mutants with high (R)- and (S)-selectivity is valuable and highly desired. Herein, we report a highly efficient directed evolution strategy, using only 4 representative amino acids, namely, alanine (A), leucine (L), lysine (K), tryptophan (W) at each mutated site to create an extremely small library of CALB variants requiring notably less screening. The obtained best mutant with three mutations W104V/A281L/A282K displayed highly reversed (S)-selectivity towards a series of sec-alcohol with E values up to 115 (conv. 50%, ee 94%). Compared with the previously reported (S)-selective CALB variant, W104A, a single mutation provided less selectivity, while the synergistic effects of three mutations in the best variant endow better (S)-selectivity and a broader substrate scope than the W104A variant. Structural analysis and molecular dynamics simulation unveiled the source of reversed enantioselectivity. (Figure presented.).

Full text of DOI:10.1002/adsc.201800711

Title: Highly Focused Library-Based Engineering of Candida antarctica
Lipase B with (S)-Selectivity Towards sec-Alcohols
Authors: Yixin Cen, Danyang Li, Jian Xu, Qiong-Si Wu, Qi Wu, and
Xian-Fu Lin
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Adv. Synth. Catal. 10.1002/adsc.201800711
Link to VoR: http://dx.doi.org/10.1002/adsc.201800711

10.1002/adsc.201800711
Advanced Synthesis & Catalysis
1
FULL PAPER
DOI: 10.1002/adsc.201((will be filled in by the editorial staff))
Highly Focused Library-Based Engineering of Candida
antarctica Lipase B with (S)-Selectivity Towards sec-Alcohols
Yixin Cen,+a Danyang Li,+a Jian Xu,a Qiongsi Wu,a Qi Wua,* and Xianfu Lina,*
a
Department of Chemistry, Zhejiang University, Hangzhou 310027, People's Republic of China
Fax: (+86)-571-8795-2618; e-mail: llc123@zju.edu.cn or xflin@zju.edu.cn
+
These authors contributed equally to this work.
Received: ((will be filled in by the editorial staff))
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201######.((Please
delete if not appropriate))
Abstract. Candida antarctica lipase B (CALB) is one of the sec-alcohol with E values up to 115 (conv. 50%, ee 94%).
most extensively used biocatalysts in both academia and Compared with the previously reported (S)-selective CALB
industry and exhibits remarkable (R)-enantioselectivity for variant, W104A, a single mutation provided less selectivity,
various chiral sec-alcohols. Considering the significance of while the synergistic effects of three mutations in the best
tailor-made stereoselectivity in organic synthesis,
a
variant endow better (S)-selectivity and a broader substrate
discovery of enantiocomplementary lipase mutants with high scope than the W104A variant. Structural analysis and
(R)- and (S)-selectivity is valuable and highly desired. molecular dynamics simulation unveiled the source of
Herein, we report a highly efficient directed evolution reversed enantioselectivity.
strategy, using only 4 representative amino acids, namely,
alanine (A), leucine (L), lysine (K), tryptophan (W) at each
mutated site to create an extremely small library of CALB
variants requiring notably less screening. The obtained best
mutant with three mutations W104V/A281L/A282K
displayed highly reversed (S)-selectivity towards a series of
Keywords: Candida antarctica lipase B; Directed
evolution; Focused library; sec-Alcohols; Hydrolysis
configurational sec-alcohols or hydrolyze their esters.
[10,11]
Introduction
To improve and alter the catalytic properties of
enzymes, protein engineering has emerged as a
powerful means during the past two decades.[12]
Directed evolution and rational design are two main
strategies for protein engineering.[13] ISM (iterative
Chiral sec-alcohols belong to a type of important
intermediates in the pharmaceutical industry for the
production of active pharmaceutical ingredients to
treat health conditions such as cardiovascular
diseases and hypertension.[1-3] For example, (R)- and
(S)-enantiomers of 1-phenylethanol are significant
optically active substances.[4,5] (S)-1-phenylethanol
can be used to synthesize sertraline to treat
depression and can also be a synthetic precursor to
treat asthma and enhance the immune system.[6] On
the other hand, (R)-1-phenylethanol can be used to
saturation
mutagenesis)
based
on
CAST
(combinatorial active site saturation test) proposed by
Reetz and coworkers have been successfully applied
to increase thermostability, activity and selectivity of
various enzymes.[14]
Candida antarctica lipase B (CALB) is one of the
most widely used lipases and has a broad range of
applications in kinetic resolution of racemic alcohols
and amines in both academic and industrial
laboratories on account of its extraordinary catalytic
performance.[15,16] Structurally, CALB has an active
site composed of the catalytic triad Ser105/ His224/
Asp187, an oxyanion hole, and two binding pockets
for the acyl-moiety of the ester and the alcohol
parts.[17-19] Consequently, protein engineering aiming
to improve or reverse the stereoselectivity of CALB
for racemic acids or sec-alcohol substrates mainly
focused on the corresponding acyl or alcohol-binding
pockets, respectively.[20,21] Considering the generally
equivalent significance of two enantiomers of
synthesize
drugs
that
inhibit
cholesterol
absorption.[7,8] Hence, the tailor-made preparation of
optically pure chiral sec-alcohols are particularly
important.
Kinetic resolution (KR) is a key tool in the
synthesis of optically active sec-alcohols. The
common hydrolases show preferential (R)-
stereoselectivity for chiral sec-alcohol in the
hydrolysis or acylation reactions according to the
“Kazlauskas rule”, [9] since large substituents bind to
the large hydrophobic pocket and small ones to the
medium pocket. So far only few hydrolases showed
(S)-stereoselectivity towards sec-alcohols, as in the
case of subtilisin, which can selectively acylate (S)-
resolved
substrates,
discovering
1
This article is protected by copyright. All rights reserved.

10.1002/adsc.201800711
Advanced Synthesis & Catalysis
enantiocomplementary mutants with high (R)- and
(S)-selectivities is valuable and highly desired. For
example, we have previously demonstrated that
CALB acid-binding pocket can be modified through
directed evolution to create enantiocomplementary
The hydrolytic kinetic resolution of racemic 1-
phenylethyl acetate (rac-1a) was chosen as the model
reaction (Figure 1). As mentioned above, WT CALB
had a preference to (R)-sec-alcohol, and the
resolution result of rac-1a was the 99% ee value with
mutants for the kinetic resolution of racemic acids.[20b] 50% yield.
Concerning the inversion of the stereoselectivity of
CALB towards certain sec-alcohols, Hult and
coworkers reported an important CALB mutation,
W104A, which was obtained by a rational design and
being situated in the alcohol-binding pocket, was able
to accommodate much larger groups than the wild-
type (WT) CALB, and this change transformed the
highly (R)-selective WT into an (S)-selective
mutant.[22] However, the enantioselectivity of W104A
was unsatisfactory, the E values were low (2<E<12)
for some typical sec-alcohols, such as phenylethanol,
phenylpropanol and 2-hexanol, in their acylation
reactions, although these E values could be increased
by an appropriate choice of solvents in combination
with a higher temperature and longer alkyl chains of
sec-alcohols.[21-23] Thus, it is highly desirable to
explore other mutations at the W104 hot spot or
supplementary mutations at other important sites
surrounding the alcohol-binding pocket to obtain
CALB mutants with higher (S)-selectivity than
W104A. On the other hand, all previous studies
concerning the (S)-selective mutant W104A were
performed for the acylation reactions in organic
solvents, which leaves uncertainty about the
selectivity of this (S)-selective mutant for the
hydrolytic kinetic resolution of sec-alcohols.
To obtain (S)-selective mutants for rac-1a, which
meant inverting the Kazlauskas rule, the
stereoselective pocket of alcohols should be changed
to accommodate the phenylalkanol substituents,
which are larger than the ethyl group. An early study
[21-23]
has identified W104 as the most important hot
spot in the alcohol-binding pocket of CALB for the
reversed (S)-selectivity of sec-alcohols, and W104
should be replaced by smaller amino acids to enlarge
the volume of the stereoselectivity pocket. According
to the X-ray structure of CALB (PDB code 1TCA,[17]
Figure 2A), we selected four important hydrophobic
residues W104, L278, A281 and A282 (sites A, C,
and B as shown in Figure 2B) surrounding the
alcohol-binding pocket as mutagenesis targets. It is
well-known that highly efficient screening is one of
the biggest bottlenecks in directed evolution of
enantioselective enzymes, especially for multiple-site
saturation mutagenesis. Thus, we proposed to
construct one highly focused library composed of
only a dozen variants. Considering the importance of
the space of the alcohol stereoselective pocket, the
principle of amino acid exchange is based on the
different side chain volumes of the targeting amino
acids.
Considering that large gene libraries created by
random mutagenesis often show low hit ratios, herein
we constructed a highly focused library containing
less than 20 variants using a proposed new strategy.
This strategy was related to iterative saturation
mutagenesis (ISM) at CAST sites proposed by Reetz
and coworkers, [14] while using only 4 representative
amino acids, namely, alanine (A), leucine (L), lysine
(K) and tryptophan (W) at each site to create an
extremely small library requiring notably less
screening. The best mutant obtained in this study
contained a triple W104V/A281L/A282K mutation
and displayed better (S)-selectivity for the hydrolytic
kinetic resolution of a series of sec-alcohol esters
than the previously reported W104A variant.
Figure 2. (A) X-ray structure of WT CALB (PDB:
1TCA)[17] used as a guide in directed evolution. (B)
Mutagenesis sites A (Trp104), B (Ala282/ Ala282), C
(Leu278). (The catalytic triad S105-H224-D187 and the
oxyanion hole T40-Q106 are shown with black sticks, and
the mutant site A is shown with yellow sticks, site B is
blue, site C is pink.) (C) Strategy for the construction of
highly focused CALB library.
Results and Discussion
First, we chose seven residues (alanine (A), glycine
(G), leucine (L), isoleucine (I), cysteine (C), valine
(V) and methionine (M)) for an amino acid exchange
of W104, because of their smaller volume than that of
tryptophan (W), which made up library A. Based on
Figure 1. Stereoselective hydrolysis of 1-phenylethyl
acetate catalyzed by wide type (WT) CALB and mutants.
2
This article is protected by copyright. All rights reserved.

10.1002/adsc.201800711
Advanced Synthesis & Catalysis
the best hit obtained from the "subsaturation"
mutagenesis at site A (W104), three other amino acid
residues at sites B and C were varied iteratively to A,
L, K and W to create a CALB library containing less
than 20 variants (Figure 2C, Table 1). Alanine was
the smallest amino acid except glycine, and glycine
exchange usually impacted the secondary structure of
lipases leading the loss of activity. Leucine and
tryptophan were chosen as amino acids with medium-
to-large side chains, and lysine was a functional
residue with a long carbon chain. This strategy for
selecting amino acids can remarkably reduce the size
of libraries, only require dozens of screening, and
have a high percentage of positive hits. All the
screening results are shown in Table 1.
In library A, most of the W104 variants containing
small side chains showed reversed (S)-selectivity
with moderate E-values, as expected. It is noteworthy
that the replacement with smaller amino acids did not
give rise to better (S)-selectivity of 1-phenylethyl
acetate. For example, the W104G mutant (WB3) was
expectedly inactive, probably due to the disrupted
catalytic structure of the active site of this mutant or
too large of a space in the active site to efficiently
bind the substrate. The W104A mutant (WB2) only
showed moderate (S)-selectivity for rac-1a with E =
18 (eep 78%, conv. 49%). Interestingly, the best
mutant in library A was W104V showing an
acceptable E value (E = 21) with the reversed (S)-
selectivity (WB7, entry 7 in Table 1). The valine side
chain was larger than in alanine and glycine, but
smaller than in other large amino acids, probably
having the most suitable space to accommodate the
phenyl group of rac-1a. The best mutant WB7 was
then used to perform A/L/K/W mutagenesis at site B
(A281 and A282). This library provided an improved
variant W104V/A282K (E = 29) (entry 10, Table 1).
Further mutagenesis at A281 of this variant resulted
in
the
best
variant
WB15
with
W104V/A281L/A282K mutation, providing reversed
94% ee upon 46% conversion (E(S) = 80, entry 15,
Table 1). An attempt to continue to the third-round
evolution at site C (L278) failed to induce further
improvement. The total screening effort involved the
evaluation of only 19 variants.
Table 1. CALB mutants catalyzed kinetic hydrolysis of rac-1a.[a]
[b]
Entry Library Mutant
Sequence
Conversion[b]
(%)
eep
E[c]
(%)
(R)
(S)
1
-
WT
-
50
49
1
99 (R)
78 (S)
0
>200
2
A
WB2
W104A
18
3
WB3
W104G
4
WB4
W104L
9
63 (S)
65 (S)
35 (S)
85 (S)
44 (S)
70 (S)
85 (S)
85 (R)
50 (S)
35 (S)
65 (S)
94 (S)
30 (S)
60 (S)
87 (S)
93 (R)
54 (R)
99 (R)
99 (R)
99 (R)
5
5
WB5
W104I
44
37
38
48
40
48
14
20
14
3
8
6
WB6
W104C
3
7
WB7
W104V
21
4
8
WB8
W104M
9
B
WB9
W104V+A282L
W104V+A282K
W104V+A282W
W104V+A281L
W104V+A281K
W104V+A281W
W104V+A281L+A282K
W104V+A281K+A282K
W104V+A281W+A282K
W104V+L278A+A281L+A282K
W104V+L278K+A281L+A282K
W104V+L278W+A281L+A282K
A281L+A282K
A282K
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
WB10
WB11
WB12
WB13
WB14
WB15
WB16
WB17
WB18
WB19
WB20
WB21
WB22
WB23
29
14
3
2
5
46
25
37
29
9
80
2
6
C
20
30
43
49
49
49
5
-
-
-
>200
>200
>200
A281L
[a] Reactions were performed with rac-1a (0.0125 mmol), acetonitrile (50 μL) and enzymes in PBS (950 μL, 50 mM, pH =
7.5) at 37 °C for several hours. [b] Determined by chiral GC analysis using Agilent CP-Chirasil-Dex CB column and
dodecane as an internal standard. [c] Enantiomeric ratio calculated according to the equation, E = ln[1-C*(1+eep)] / ln[1-
C*(1-eep)], the estimated error of the measurement is less than 3%.
3
This article is protected by copyright. All rights reserved.

10.1002/adsc.201800711
Advanced Synthesis & Catalysis
Table 2. Hydrolytic kinetic resolution of different substrates using CALB mutants originally evolved for rac-1.[a]
Entry
Substrate
Enzymes Time Conversion[b]
eep.[c]
(%)
E
(h)
(%)
(R)
(S)
1
2
3
4
5
6
7
8
9
WT
8
46(42)
>99
>200
WB2
WB15
WT
12
24
48
15
24
48
30
36
49(39)
50(41)
7(-)
92.3 ± 0.1
94.0 ± 0.2
>99
73.1 ± 1.2
115.1 ± 5.1
rac-1b
rac-1c
>200
WB2
WB15
WT
48(39)
23(19)
0(-)
73.6 ± 1.5
96.0 ± 0.8
0
13.3 ± 1.1
65.1 ± 12.5
-
-
WB2
WB15
45(41)
21(15)
54.3 ± 5.2
77.3 ± 2.3
5.3 ± 1.0
9.2 ± 1.7
rac-1d
rac-1e
10
11
12
13
14
15
16
17
18
19
20
21
WT
48
30
36
5
0(-)
0
-
-
WB2
WB15
WT
32(25)
27(20)
49(40)
43(35)
30(20)
49(37)
30(18)
24(18)
49(41)
46(37)
38(26)
81.0 ± 3.1
80.1 ± 2.2
>99
14.2 ± 2.8
12.2 ± 1.6
>200
>200
WB2
WB15
WT
8
39 ± 3
<4
14
5.5
10
14
8
95 ± 2
69.8 ± 30.1
rac-1f
rac-1g
>99
WB2
WB15
WT
35.2 ± 0.4
91 ± 3
<3
31.5 ± 11.5
93.0 ± 0.2
84.5 ± 3.2
96.6 ± 0.6
83.5 ± 2.9
WB2
WB15
13
30
27.0 ± 7.1
109.2 ± 20.1
rac-1h
[a]
Reactions were performed with different substrates (0.0125 mmol), acetonitrile (50 μL) and enzymes in PBS (950 μL,
[b]
50 mM, pH = 7.5) at 37 °C for several hours. Reaction yields were determined by chiral GC analysis using an Agilent
CP-Chirasil-Dex CB column or chiral HPLC analysis using a ChiralPak OJ-H column. The isolation yields (the actual
yields) are given in parentheses (For convenience of isolation, the reaction scale was increased 40 times, thus, the used
substrate were 0.5 mmol). [c] The mean value of 2-3 measurements.
4
This article is protected by copyright. All rights reserved.

10.1002/adsc.201800711
Advanced Synthesis & Catalysis
Deconvolution was further performed to explore
When the alkyl chain length was hexyl, the (S)-
selectivities of both WB2 and WB15 were slightly
improved, which was consistent with the report by
Hult on the influence of the substrate alkyl
substituent length on the enantioselectivity of
the effect of each mutation in the WB15 variant on
the reversed (S)-selectivity of rac-1a. Three single-
point mutants (W104V, A281L, A282K) and three
double-point
mutants
(W104V/A281L,
W104V/A282K, and A281L/A282K) were compared
(entries 7, 23, 22, 12, 10 and 21, Table 1). A281L,
A282K and A281L/A282K mutants showed similar
(R)-selectivity to the WT implying the crucial role
played by W104V mutation for reversing
W104A-catalyzed
transesterifications
in
cyclohexane.[23] Interestingly, when phenyl was
substituted by propyl and butyl, namely, straight
chain alcohols, WB15 showed much higher (S)-
selectivity than WB2 mutant (entries 14-15, 17-18,
Table 2).
stereoselectivity.
W104V
(E(S)
=
21),
W104V/A281L (E(S) = 3) and W104V/A282K (E(S)
= 29) exert synergistic effects in the best variant of
WB15 (E(S) = 80).
As shown in Table 2, the reaction time of different
substrates with WB15 was much longer than WT, this
finding was in full accord with the result of the
kinetic study (Table S1 in the Supporting
Information). To evaluate the thermostability of
Next, the substrate range of the best (S)-selective
mutant WB15 evolved for the model substrate rac-1a
was determined. Moreover, considering that all
previous studies on (S)-selectivity of W104A (WB2)
were carried out for the acylation reaction of racemic
phenylalkanols in nonaqueous media,[21-23] it is
valuable to determine the stereoselectivity of W104A
in the hydrolytic kinetic resolution of phenylalkanol
esters. The comparison results of WT, WB2 and the
best WB15 variants as the catalysts for hydrolytic
kinetic resolution of a series of phenylalkanol esters
are listed in Table 2. As pointed by Hult and
coworkers,[21-23] the largest alkyl group of sec-
alcohols that can smoothly fit into the alcohol-
binding pocket of WT CALB is ethyl. Due to this
constraint, the substrates 1c–1e are not acceptable for
WT CALB, and a drastic drop in both activity and
selectivity is observed (entries 4, 7, 10, Table 2).
W104A (WB2) variant displayed reversed (S)-
selectivity for almost all the tested substrates with
moderate-to-good E values (3-73). Upon employing
the best mutant WB15, remarkably enhanced
selectivity was observed in many cases when
compared with WB2. For example, higher E values
for (S)-selectivity of WB15 were obtained in the KR
of rac-1b-1d and rac-1f-1h than WB2 (entries 1-9,
13-21, Table 2). A CALB mutant with a much higher
reversed (S)-selectivity than the well-known W104A
mutant was successfully obtained.
30
WB15, the T50 values were measured by the 50%
residual activity after a heat treatment for 30 min.
The results showed
a
slightly decreased
thermostability of WB15 (47 °C) when compared
with WT CALB (55 °C). We hypothesized that the
decrease in thermostability of WB15 may be caused
by the introduction of point mutations, which affected
the structural stability of the enzymes.
In addition, the isolation yields were lower than the
GC or HPLC yields in Table 2, the loss of products
may be partly caused by the absorption of samples on
silica gel and the loss of impure products during
column chromatography purification. This situation
may become less pronounced in a scaled-up process.
Therefore, we next performed the scaled-up kinetic
resolution of the substrate rac-1b using crude enzyme
WB15 for the preparation of (S)-2b. The reaction
mixture containing 0.5 g substrate was shaken at
37 °C until the reaction had reached 40-50%
conversion. Then the crude products were extracted
with MTBE. By using flash column chromatography,
174 mg of (S)-2b (94% ee) and 220 mg of (R)-1b
(90% ee) were obtained. The isolated yields reached
up to 45% in this scaled-up process, which was
similar to the yield in the screening reaction
determined by GC (47% conversion). This result
documented the synthetic utility of the reversed
stereoselective kinetic resolution catalyzed by the
mutant WB15. To expand the application range, the
acyl-transfer reactions were also tested for catalysis
with WT and the best mutant WB15 in organic phase.
Rac-2b (2.0 mg) and vinyl acetate (3.0 eq) were
dissolved in 1.0 mL naphthane with 25 mg
immobilized CALB. After shaking at 200 rpm and
50 °C for 12 hours, the reactions were analyzed by
GC. WT CALB provided a good result with 45%
conversion and 99% ee, as expected. To our delight,
the WB15 also exhibited good enantioselectivity of
99% ee with 10% conversion. The low conversion
might be improved by optimizing the immobilization
methods and reaction solvents in further studies.
The length of the substrate alkyl substituent
showed an interesting influence on the
enantioselectivity of the WB2 and WB15. When the
phenylalkanol alkyl chain changed from methyl to
ethyl, the (S)-selectivity of WB2 and WB15 were
improved remarkably, and E values increased from
18 to 73 (WB2) and from 80 to 115 (WB15),
respectively. However, a further increase in the alkyl
chain length of phenylalkanol from ethyl to butyl
resulted in a drastic drop in the selectivity of WB2
and WB15. Although the enlarged stereoselectivity
pockets of the (S)-selective variants WB2 and WB15
were mainly occupied by the phenyl substituents of
these substrates, the increased alkyl chain length of
phenylalkanols
probably
increased
the
To obtain some insight into the source of
complementary stereoselectivity from WT and WB15
CALB, molecular dynamics (MD) computations were
further carried out using the modeling suite
YASARA.[24-28] All WB15 mutations were introduced
hydrophobicity of alkyl substituents, thus, easily
leading the competitive orientation of alkyl
substituents down into the stereoselectivity pocket,
and further resulting in the decreasing selectivity.
5
This article is protected by copyright. All rights reserved.

10.1002/adsc.201800711
Advanced Synthesis & Catalysis
by swapping the specific amino acid residues of WT
CALB, followed by energy minimizations.
substituents represented a starting conformation of
Mode II (Figure S2B and S2D in the Supporting
Information). After MD simulations of these two
modes, we found that the configuration constructed
by Mode I was more dominant for the complex of
(R)-1a with WB15 mutant. The formation energy of
the energy-minimized structure of (R)-TIs in Mode I
was much lower than that in Mode II (-640.6 KJ/mol
vs. -605.7 KJ/mol), probably due to the stronger
hydrophobic interaction of phenyl than of methyl in
the large alcohol-binding pocket surrounded by
hydrophobic residues V104, L278, L281, and I285.
Thus, we chose the Mode I as the binding model for
nonproductive (R)-TI-WB15 complex (Figure 3C).
The binding mode for the productive (S)-TI-WB15
complex is similar as that of (R)-TI-WB15 complex
(Figure 3D). However, their H-bond networks
between the substrate and active sites are different.
The productive (S)-TI-WB15 complex has all
necessary H-bonds for catalysis (Figure 3D), while
the most important H-bond between H224 (HE2
atom) and the bound (R)-TI oxygen (O1 atom) was
lost in the nonproductive (R)-TI-WB15 complex
(Figure 3C, black line). Moreover, the bond distances
(d) between HE2 and O1 atoms and the
corresponding bond angles (α) of NE2--HE2--O1 in
(S)-TI-WB15 complex are also more favorable for the
nucleophilic attack towards the ester bond of 1a than
(R)-TI-WB15 complex (Table 3). These results
clearly implied the reversed (S)-selectivity of WB15
mutant towards sec-alcohols.
The tetrahedral intermediates (TI) of (R)- and (S)-
1a were built on the side-chain O atom of the
catalytic Ser105 in the WT CALB and WB15
variants, respectively. In MD simulations of (R)- and
(S)-TIs, 100 simulation frames (100 ps to 10 ns) were
evaluated and additionally minimized to derive
statistical averages and properties of the
corresponding local minima. Finally, the distances
and bond angles for some important H-bonds were
also compared (Table 3). The energy-minimized
structures of (R)- and (S)-TIs of WT CALB and
mutant WB15 are shown in Figure 3.
For the WT CALB, the space in the alcohol-
binding pocket is very small because tryptophan
occupies position 104, and the largest alkyl group of
sec-alcohols that can be accommodated is ethyl.[21-23]
Thus, as shown in Figure 3A and 3B, both (R)-TI-
CALB complex and (S)-TI-CALB complex have
similar orientations of methyl group located inside
the alcohol-binding pocket. However, they have
completely different H-bonds formed and interactions
between the substrate and active sites. All necessary
H-bonds for catalysis were found in the optimized
pose of the fast-reacting (R)-TI-CALB complex
(Figure 3A), while the most important H-bond
between H224 (HE2 atom) and the bound (S)-TI
oxygen (O1 atom) was lost in the nonproductive (S)-
TI-CALB complex (Figure 3B, black line), leading to
no nucleophilic attack towards the ester bond of 1a.
Additionally, (R)-TI-CALB complex had a shorter
distance between HE2 and O1 atoms (1.812 Å vs.
2.341Å) and larger bond-angles of NE2--HE2--O1
(150.3° vs. 134.4°) (Table 3) than that of
nonproductive (S)-TI-CALB complex. These results
confirmed the high (R)-selectivity of WT CALB and
were also consistent with the previous calculation
studies by Hult.[22]
Table 3. The hydrogen bond distances (d) between His
224 (HE2 atom) and the bound (R)- or (S)-tetrahedral
intermediate (TI) oxygen (O1 atom), the corresponding
bond angles (α) of NE2--HE2--O1, and the possibility of
hydrogen bond formation (P1 and P2) between TI (O atom
of carbonyl) and the oxyanion hole residues (Gln106 and
Thr40).
[b]
[b]
TI-Enzymes
complex
d (Å) [a] α (°) [a]
P1
P2
For the (S)-selective mutant WB15, the W104V
mutation greatly increased the space in the alcohol-
binding pocket, thus, both methyl and phenyl can be
accommodated efficiently, and two binding modes
for the orientations of the alcohol moiety in the active
site of WB15 are possible for (R)-1a or (S)-1a,
respectively.[29] Using (R)-1a as an example, which is
shown in Figure S2 (Supporting Information), Mode I
placed the large substituent (phenyl group) into the
alcohol-binding pocket and the medium substituent
(methyl group) towards the active site entrance
(Figure S2A and S2C in the Supporting Information).
Exchanging the positions of large and medium
(R-1a)-TI-WT
(S-1a)-TI-WT
(R-1a)-TI-WB15 3.745
(S-1a)-TI-WB15 1.798
1.812
2.341
150.3
134.4
142.0
174.1
0.90
1.00
[c]
[c]
-
-
-
-
[c]
[c]
0.91
1.00
[a] Data were obtained from the energy-minimized structure
of the (R/S-1a)-TI-enzyme complex of WT CALB or
[b]
variant WB15.
Data were the average values obtained
from 100 snapshots of the TI-enzyme complex of WT
CALB or variant WB15 during 10 ns MD. [c] Not detected.
6
This article is protected by copyright. All rights reserved.

10.1002/adsc.201800711
Advanced Synthesis & Catalysis
Figure 3. (A) A productive binding mode and a hydrogen bond networks of the (R)-1a complex with WT CALB having
Ser105-His224-Asp187 as the catalytic triad. The H-bond networks for nucleophilic attack and stabilizing the oxyanion are
indicated by purple dotted lines. (B) A nonproductive orientation found for the less reactive (S)-1a-CALB complex,
revealing that the H-bond between H224 (HE2 atom) and the bound ester oxygen no longer exists. (C) and (D) are the
optimized binding modes of the less reactive (R)-1a and the preferential (S)-1a, respectively, in (S)-selective variant WB15.
dynamics simulation unveiled the source of reversed
enantioselectivity. The enlargement of alcohol-
binding pocket caused by point mutation at W104
Conclusion
position and the modification of an active site
To reverse the inherent (R)-enantioselectivity of
entrance at A281/A282 positions allow (S)-
wild-type Candida antarctica lipase B (CALB)
configurational sec-alcohol to form a productive
towards chiral sec-alcohol, an efficient directed
complex. The construction strategy of highly focused
evolution strategy was proposed to construct a highly
library demonstrated herein enables us to engineer
focused library containing only 20 variants, which
various lipases more conveniently and efficiently to
drastically reduced the screening effort. W104 was
obtain the tailor-made biocatalysts and optically pure
selected as the first targeting site to perform
products in further investigation.
"subsaturation" mutagenesis by introducing seven
residues with a volume smaller than that of Trp, as
amino acids exchange. The variant W104V displayed
Experimental Section
better reversed (S)-selectivity than the W104A
mutant, and it was found that an exchange with
Materials and Methods
smaller amino acids at the W104 site is not
necessarily better. Then, further iteration mutagenesis
Unless otherwise specified, all reagents were
at A281 and A282 sites by using only
4
obtained from commercial sources and used without
further purification. Pfu polymerase and Dpn I were
purchased from Thermo Fischer Scientific Inc. The
E.Z.N.A.® Plasmid Midi Kit was supplied by
representative amino acids of A, L, K and W was
adopted. To our great delight, we finally have found
that WB15 (W104V/A281L/A282K) has a good E
value (E = 80) with (S)-selectivity in the hydrolysis of
model substrate and displays broad substrate scope of
1
13
Omega-Bio Tek. The H NMR and C NMR spectra
were recorded with a Bruker AMX400 MHz
spectrometer using TMS as an internal standard. All
spectra were recorded at room temperature in CDCl3,
the following abbreviations were used to describe
peak splitting patterns when appropriate: s = singlet,
d = doublet, t = triplet, q = quartet, m = multiplet.
Coupling constants were reported in Hertz (Hz). The
various
phenylalkanol
esters.
Deconvolution
experiments implied the crucial role played by
W104V mutation for reversing stereoselectivity and
the synergistic effects of these mutations W104V,
W104V/A281L and W104V/A282K in the best
variant WB15. Structural analysis and molecular
7
This article is protected by copyright. All rights reserved.

10.1002/adsc.201800711
Advanced Synthesis & Catalysis
High-Resolution Mass Spectrometer was obtained
from the WATERS GCT Premier with Electron
Impact Ion Source.
of each tube was resuspended in 50 mM Tris-HCl
(pH 7.5) and lysed by sonification (Bandelin, 15×10
sec with 10 sec intervals, at 40% pulse, in a water-ice
bath). The cell debris was removed by centrifugation
for 25 min at 4 °C. The supernatant was stored at -
78 °C.
Chiral GC analysis was performed with a
Shimadzu GC-2014C instrument (Shimadzu, Japan)
equipped with an FID detector. Samples were
analyzed on an Agilent CP-Chirasil-Dex CB column
(25 m × 0.25 mm × 0.1 μm, Teknokroma, Spain),
under the following conditions: carrier gas: N2;
detector temperature, 220 °C. Chiral HPLC was
performed with a Chiralpak OJ-H column (250
mm×4.6 mm, n-hexane/2-propanol as the mobile
phase) and a UV detector (220nm). Detailed chiral
separation conditions are shown in Supporting
Information.
Molecular Modeling and MD Simulation
The molecular modeling and MD simulation were
performed using YASARA structure (version
15.5.31).[25] The AMBER03 force field[26] with
default settings was used for the protein and
AutoSMILES force field assignment for the
substrates and the tetrahedral intermediates.[27]
Energy minimization in the water environment was
performed using the steepest descent and simulated
annealing simulations. The MD simulation was
performed at 298 K using a time-step of 1.25 fs for
inter- and intramolecular forces over 10 ns in an NPT
ensemble using PME.[28]
Models of the enzyme variants with bound 1a were
based on the crystal structure (PDB ID: 1TCA). The
(R)- and (S)-1a substrates were allowed to equilibrate
for 10 ns by a molecular dynamics simulation in WT
and WB15, respectively.
Library Generation
WT-CALB plasmid (pETM11-CALB) was used as
DNA template, and various primers are shown in
Table S2 in the Supporting Information. The 50 μL
PCR mix contained ddH2O (29 μL), Pfu 10X buffer
(5 μL), dNTP (4 μL, 2.5 mM each), forward primers
(5 μL, 2.5 μM each), silent reverse primer (5 μL, 2.5
μM), template plasmid (1 μL, 100 ng/μL) and 1 μL of
Pfu polymerase. PCR reactions were performed at
94 °C for 5 min; followed by denaturation at 94 °C
for 1 min, annealing at 60 °C for 1 min, and
elongation at 72 °C for 14 min repeated 30 times. A
final extension at 72 °C for 10 min ended the reaction.
To ensure elimination of the circular methylated
template plasmid, 20 μL of PCR reaction mixture
were mixed with 2 μL DpnI (10 U/μL) and incubated
overnight at 37 °C. The Dpn I-digested product was
purified with an Omega PCR purification spin-
column, and the plasmid (6 μL) was used to
transform 80 μL of E. coli Origami2
electrocompetent cells (containing chaperone plasmid
pGro7, Takara, Japan). The transformation mixture
was incubated with 1 mL of LB medium at 37 °C
with shaking at 200 rpm and spread on LB-agar
plates containing kanamycin (34 μg/mL) and
chloramphenicol (34 μg/mL).
Acknowledgements
The financial support from National Natural Science Foundation
of China (21472169, 21574113) and the Fundamental Research
Funds for the Central Universities (2018QNA3010) is gratefully
acknowledged.
References
[1] X. Wu, J. Xiao, Chem. Commun. 2007, 38, 2449-2466.
[2] O. Jurcek, M. Wimmerova, Z. Wimmer, Coord. Chem.
Rev. 2008, 252, 767-781.
[3] G. W. Huisman, J. Liang, A. Krebber, Curr. Opin.
Chem. Biol. 2010, 14, 122-129.
[4] S. H. Schöfer, N. Kaftzik, P. Wasserscheid, U. Kragl,
Expression of CALB Mutants
Chem. Commun. 2001, 5, 425-426.
Colonies appeared after cultivation for 12-16 h at
37 °C and were picked into 5.0 mL LB medium
[5] W. D. Fessner, T. Anthonsen, Modern biocatalysis:
stereoselective and environmentally friendly reactions,
Wiley-VCH, Weinheim, 2009.
containing
kanamycin
(34
μg/mL)
and
chloramphenicol (34 μg/mL), and then incubated at
37 °C under shaking at 200 rpm overnight. A fresh
200 mL of TB media was added to 2mL preculture
containing 1.0 mg/mL L-arabinose as the inducer for
expression of chaperone pGro7, kanamycin (34
μg/mL) and chloramphenicol (34 μg/mL). The
cultures were shaken at 37 °C until the optical density
at 600 nm reached 0.6, and then cooled to 4°C for 1h.
Then, 1.0 mM isopropyl β-thiogalactopyranoside
(IPTG) added into the cultures to induce CALB
expression for 48 h at 18 °C. The tubes were
centrifuged at 4000 rpm and 4 °C for 25 min, and
then the supernatants were discarded. The cell pellet
[6] R. N. Patel, Stereoselective Biocatalysis, Chemical
Industry Press, Beijing, 2004.
[7] L. S. Chua, M. R. Sarmidi, J. Mol. Catal. B: Enzym.
2004, 28, 111-119.
[8] A. P. de los Ríos, F. V. Rantwijk, R. A. Sheldon, Green
Chem. 2012, 14, 1584-1588.
[9] a) R. J. Kazlauskas, A. N. E. Weissfloch, A. T.
Rappaport, L. A. Cuccia, J. Org. Chem. 1991, 56,
2656-2665; b) K. E. Jaeger, T. Eggert, Curr. Opin.
Biotechnol. 2002, 13, 390–397; c) A. S. de Miranda, L.
8
This article is protected by copyright. All rights reserved.

10.1002/adsc.201800711
Advanced Synthesis & Catalysis
S. Miranda, R. O. de Souza, Biotechnol. Adv. 2015, 33,
372-393.
[16] S. Naik, A. Basu, R. Saikia, B. Madan, P. Paul, R.
Chaterjee, J. Brask, A. Svendsen, J. Mol. Catal. B:
Enzym. 2010, 65, 18−23.
[10] M. J. Kim, Y. I. Chung, Y. K. Choi, H. K. Lee, D.
Kim, J. Park, J. Am. Chem. Soc. 2003, 125, 11494–
11495.
[17] J. Uppenberg, M. T. Hansen, S. Patkar, T. A. Jones.
Structure. 1994, 2, 293308.
[11] L. Borén, B. Martin-Matute, Y. Xu, A. Córdova, J. E.
[18] J. Uppenberg, N. M. Oehrner, K. Hult, G. J. Kleywegt,
Bäckvall, Chem. Eur. J. 2005, 12, 225-232.
S. Patkar, Biochemistry 1995, 34, 1683816851.
[12] a) G. A. Strohmeier, H. Pichler, O. May, M. Gruber-
Khadjawi, Chem. Rev. 2011, 111, 4141–4164; b) L. G.
Otten, F. Hollmann, I. W. C. E. Arends, Trends
Biotechnol. 2010, 28, 46-54; c) A. Díaz-Rodríguez, B.
G. Davis, Curr. Opin. Chem. Biol. 2011, 15, 211–219;
d) A. S. Bommarius, J. K. Blum, M. J. Abrahamson,
Curr. Opin. Chem. Biol. 2011, 15, 194–200; e) U. T.
Bornscheuer, G. W. Huisman, R. J. Kazlauskas, S.
Lutz, J. C. Moore, K. Robins, Nature 2012, 485, 185-
194; f) C. K. Prier, F. H. Arnold, J. Am. Chem. Soc.
2015, 137, 13992−14006.
[19] R. T. Otto, H. Scheib, U. T. Bornscheuer, J. Pleiss, C.
Syldatk, R. D. Schmid. J. Mol. Catal. B: Enzym. 2000,
8, 201211.
[20] a) P. B. Juhl, K. Doderer, F. Hollman, O. Thum, J.
Pleiss, J. Biotechnol. 2010, 150, 474-480. b) Q. Wu, P.
Soni, M. T. Reetz, J. Am. Chem. Soc. 2013, 135, 1872–
1881.
[21] A. O. Magnusson, J. C. Rotticcimulder, A. Santago-
stino, K. Hult, Chembiochem 2005, 6, 1051-1056.
[22] A. O. Magnusson, M. Takwa, A. Hamberg, K. Hult,
[13] Recent reviews of directed evolution: a) N. J. Turner,
Nat. Chem. Biol. 2009, 5, 567–573; b) C. Jäckel, D.
Hilvert, Curr. Opin. Biotechnol. 2010, 21, 753–759; c)
M. T. Reetz, Angew. Chem., Int. Ed. 2011, 50, 138–
174; d) E. M. Brustad, F. H. Arnold, Curr. Opin. Chem.
Biol. 2011, 15, 201–210; e) A. Currin, N. Swainston, P.
J. Day, D. B. Kell, Chem. Soc. Rev. 2015, 44, 1172–
1239; f) C. A. Denard, H. Ren, H. Zhao, Curr. Opin.
Chem. Biol. 2015, 25, 55–64; g) H. Renata, Z. J. Wang,
F. H. Arnold, Angew. Chem., Int. Ed. 2015, 54, 3351–
3367; h) A. S. Bommarius, Annu. Rev. Chem. Biomol.
Eng. 2015, 6, 319–345; i) S. C. Hammer, A. M. Knight,
F. H. Arnold, Curr. Opin. Green Sustain. Chem. 2017,
7, 23-30.
Angew. Chem. Int. Ed. 2005, 44, 4582−4585.
[23] M. Vallin, P. Syrén, K. Hult, ChemBioChem 2010, 11,
411-416.
[24] E. Krieger, G. Koraimann, G. Vriend, Proteins:
Struct., Funct., Bioinf. 2002, 47, 393-402.
[25] E. Krieger, T. Darden, S. B. Nabuurs, A. Finkelstein,
G. Vriend, Proteins: Struct., Funct., Bioinf. 2004, 57,
678-683.
[26] Y. Duan, C. Wu, S. Chowdhury, M. C. Lee, G. Xiong,
W. Zhang, R. Yang, P. Cieplak, R. Luo, T. Lee, J.
Caldwell, J. Wang, P. Kollman, J. Comput. Chem. 2003,
24, 1999-2012.
[14] M. T. Reetz, J. D. Carballerira, Nat. Protoc. 2007, 2,
[27] A. Jakalian, D. B. Jack, C. I. Bayly, J. Comput. Chem.
2002, 23, 1623-1641.
891-903.
[15] a) U. T. Bornscheuer, R. J. Kazlauskas, Hydrolases in
Organic Synthesis: Regio- and Stereoselective
Biotransformations, Wiley-VCH, Weinheim, 1999; b)
K. E. Jaeger, B. W. Dijkstra, M. T. Reetz, Annu. Rev.
Microbiol. 1999, 53, 315−351; c) O. Kirk, M. W.
Christensen, Org. Prog. Res. Dev. 2002, 6, 446−451; d)
A. Baldessari, L. E. Iglesias, Methods Mol. Biol. 2012,
861, 457−469.
[28] U. Essmann, L. Perera, M. L. Berkowitz, J. Chem.
Phys. 1995, 103, 8577-8593.
[29] C. Orrenius, F. Hbffner, D. Rotticci, N. öhrner, T.
Norin, K. Hult, Biocatal. Biotransform. 1998, 16, 1-15.
9
This article is protected by copyright. All rights reserved.

10.1002/adsc.201800711
Advanced Synthesis & Catalysis
FULL PAPER
Highly Focused Library-
Based Engineering of
Candida antarctica Lipase B
with (S)-Selectivity Towards
sec-Alcohols
Adv. Synth. Catal. 2018,
Volume, Page – Page
Yixin Cen,+a Danyang Li,+a
Jian Xu,a Qiongsi Wu,a Qi
Wua,* and Xianfu Lina,*
10
This article is protected by copyright. All rights reserved.