Stereodivergent Protein Engineering of a Lipase to Access All Possible Stereoisomers of Chiral Esters with Two Stereocenters

Article abstract of DOI:10.1021/jacs.9b02709

Enzymatic stereodivergent synthesis to access all possible product stereoisomers bearing multiple stereocenters is relatively undeveloped, although enzymes are being increasingly used in both academic and industrial areas. When two stereocenters and thus four stereoisomeric products are involved, obtaining stereodivergent enzyme mutants for individually accessing all four stereoisomers would be ideal. Although significant success has been achieved in directed evolution of enzymes in general, stereodivergent engineering of one enzyme into four highly stereocomplementary variants for obtaining the full complement of stereoisomers bearing multiple stereocenters remains a challenge. Using Candida antarctica lipase B (CALB) as a model, we report the protein engineering of this enzyme into four highly stereocomplementary variants needed for obtaining all four stereoisomers in transesterification reactions between racemic acids and racemic alcohols in organic solvents. By generating and screening less than 25 variants of each isomer, we achieved >90% selectivity for all of the four possible stereoisomers in the model reaction. This difficult feat was accomplished by developing a strategy dubbed "focused rational iterative site-specific mutagenesis" (FRISM) at sites lining the enzyme's binding pocket. The accumulation of single mutations by iterative site-specific mutagenesis using a restricted set of rationally chosen amino acids allows the formation of ultrasmall mutant libraries requiring minimal screening for stereoselectivity. The crystal structure of all stereodivergent CALB variants, flanked by MD simulations, uncovered the source of selectivity.

Full text of DOI:10.1021/jacs.9b02709

Subscriber access provided by UNIV OF LOUISIANA
Article
Stereodivergent Protein Engineering of a Lipase to Access All
Possible Stereoisomers of Chiral Esters with Two Stereo-centers
Jian Xu, Yixin Cen, Warispreet Singh, Jiajie Fan, Lian Wu, Xianfu
Lin, Jiahai Zhou, Meilan Huang, Manfred T. Reetz, and Qi Wu
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.9b02709 • Publication Date (Web): 26 Apr 2019
Downloaded from http://pubs.acs.org on April 26, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 15
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Stereodivergent Protein Engineering of a Lipase to Access All
Possible Stereoisomers of Chiral Esters with Two
Stereocenters
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Jian Xu,†, ‡ Yixin Cen,†,||, ‡ Warispreet Singh,§ Jiajie Fan,† Lian Wu,|| Xianfu Lin,† Jiahai Zhou,||,*,
Meilan Huang,§,* Manfred T. Reetz, ¶,#,* & Qi Wu†,*
†Department of Chemistry, Zhejiang University, Hangzhou, 310027, P. R. China;
§School of Chemistry and Chemical Engineering, Queen's University, David Keir Building, Stranmillis Road, Belfast BT9
5AG, Northern Ireland, UK;
||State Key Laboratory of Bio-organic and Natural Products Chemistry, Shanghai Institute of Organic Chemistry, Chinese
Academy of Sciences, Shanghai 200032, China;
¶Max-Planck-Institut für Kohlenforschung, Kaiser-Wilhelm-Platz 1, 45470, Mülheim an der Ruhr, Germany.
#Chemistry Department, Philipps-University, Hans-Meerwein-Str. 4, 35032 Marburg, Germany.
KEYWORDS. lipases; directed evolution; stereodivergent protein engineering; multiple stereocenters; solvent effects.
ABSTRACT: Enzymatic stereodivergent synthesis to access all possible product stereoisomers bearing multiple stereocenters is
relatively undeveloped, although enzymes are being increasingly used in both academic and industrial areas. When two
stereocenters and thus four stereoisomeric products are involved, obtaining stereodivergent enzyme mutants for individually
accessing all four stereoisomers would be ideal. Although significant success has been achieved in directed evolution of enzymes in
general, stereodivergent engineering of one enzyme into four highly stereocomplementary variants for obtaining the full
complement of stereoisomers bearing multiple stereocenters remains a challenge. Using Candida antarctica lipase B (CALB) as a
model, we report the protein engineering of this enzyme into four highly stereocomplementary variants needed for obtaining all
four stereoisomers in transesterification reactions between racemic acids and racemic alcohols in organic solvents. By generating
and screening less than 25 variants each isomer, we achieved >90% selectivity for all of the four possible stereoisomers in the
model reaction. This difficult feat was accomplished by developing a strategy dubbed “focused rational iterative site-specific
mutagenesis” (FRISM) at sites lining the enzyme’s binding pocket. The accumulation of single mutations by iterative site-specific
mutagenesis using a restricted set of rationally chosen amino acids allows the formation of ultra-small mutant libraries requiring
minimal screening for stereoselectivity. The crystal structure of all stereodivergent CALB variants, flanked by MD simulations,
uncovered the source of selectivity.
INTRODUCTION
control of multiple stereogenic centers for obtaining all
possible stereoisomers in catalyst-controlled reactions remains
a challenge.
Different enantiomers of a biologically active molecule
generally have different physiological properties, such as
different degree or different type of biological activities, or
even opposite bioactivity or toxic side effects.1 One of the
most representative examples is thalidomide, which has been
cited many times in discussions of this kind.1,2 The
relationship between the chirality of molecules and their
bioactivity becomes substantially more complex when
compounds contain multiple stereocenters. For example, (2R,
3R)-paclobutrazol is a fungicide, while (2S, 3S)- paclobutrazol
is a plant growth regulator.3 The (S, S)-form of ethambutol is a
front-line anti-tuberculosis drug, while the (R,R)-enantiomer is
completely inactive against mycobacterium tuberculosis, and
the (R, S)-diastereomer is 16 times less effective.4 Thus, it is
necessary to have easy access to all theoretically possible
stereoisomers of a drug candidate in order to evaluate their
bioactivity carefully. Despite remarkable progress in the field
of asymmetric synthesis over the past few decades,5 complete
Asymmetric transformations such as desymmetrization of
meso-epoxides by Jacobsen-catalysts allow access to both of
the expected enantiomers, but not to the respective
diastereomers.6 Some well-designed man-made catalysts for
stereodivergent synthesis methods have emerged which allow
access to some and in rare cases even all stereoisomeric
products with multiple stereocenters from the same set of
precursors. For example, Carreira and coworkers described the
fully stereodivergent synthesis of γ,δ-unsaturated aldehydes
bearing two vicinal quaternary/tertiary stereogenic centers by
an ingenious dual catalytic procedure based on a chiral Ir-
catalyst and a chiral amine co-catalyst for accessing all four
stereoisomers.7 In
a different and likewise impressive
approach, Buchwald and co-workers recently demonstrated
the synthesis of all stereoisomers of a set of 1,3-amino
alcohols bearing up to three contiguous stereocenters through
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 15
sequential, copper-hydride-catalyzed hydrosilylation and
stereodivergent protein engineering of a single wildtype (WT)
hydroamination of readily available enals and enones.8
Diverse reviews have appeared describing asymmetric
catalysis in which more than one stereocenter is involved.9
lipase10 into four highly stereocomplementary variants for
transesterification reactions in organic solvents. Candida
antarctica lipase B (CALB) was chosen as the model enzyme
to be evolved, one of the most extensively used biocatalysts in
both industrial and academic laboratories.10,20 It has two
binding pockets at the active site for the acid (acyl) and
alcohol parts of ester substrates, respectively.20 Based on
previous protein engineering studies and structural knowledge
of the two selectivity pockets of CALB,20b,21,22 we envisioned
that the respective selectivities determined by the acyl or
alcohol-binding pockets of CALB could be combined to form
four highly stereocomplementary variants that exert control
over the transesterification reactions of racemic ester rac-1a
and alcohol rac-2a in organic solvents (Scheme 1). This
means that we aimed to perform the reactions with 50%
conversion in a kind of kinetic resolution (KR).10,23
1
2
3
4
5
6
7
8
Enzymatic stereodivergent synthesis for accessing all
possible stereoisomers bearing multiple stereocenters is a
relatively undeveloped domain, although biocatalysts are
being increasingly used in both academic and industrial
laboratories.10 For example, hydrolytic desymmetrization of
meso-epoxides catalyzed by epoxide hydrolases allows entry
into the two enantiomeric products, but generally not to the
diastereomers.11 The general problem may be due to the
exquisite selectivity of enzymes, which usually causes limited
substrate acceptance and access to only one of the possible
stereoisomeric products. An alternative method utilizes
tandem reactions, in which several enzymes sequentially
control different stereocenters of products.12-13 For example,
we recently reported the combination of P450-BM3 mutants
and appropriate alcohol dehydrogenases as catalysts in 4-step
cascade reactions starting from cyclohexane leading optionally
to either (R,R)-, (S,S)-, or (R,S)-cyclohexane-1,2-diols.12 This
was achieved by applying directed evolution using the
combinatorial active-site saturation test (CAST) and iterative
saturation mutagenesis (ISM) at residues lining the binding
pocket.14 However, developing a simple and rapid method
based on a single enzyme for the complete control of absolute
and relative configuration of products having multiple
stereogenic centers and with selective formation of all
theoretically possible stereoisomers represents a synthetic
challenge in asymmetric biocatalysis.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Stereodivergent Transacylation between 2-
Phenylpropionic Acid p-Nitrophenyl Ester (rac-1a) and 1-
Phenylethanol (rac-2a) Catalyzed by CALB Mutants.
2
1'
2
1'
O
O
O
O
3a
(2R, 1'R)-
3a
(2S, 1'S)-
OH
O
+
O
NO2
1a
rac-
2a
rac-
2
1'
2
1'
O
O
Whereas stereocomplementary enzymes with formation of
products having single stereocenters exist in nature15 or can
evolved by directed evolution,14,16 it is unlikely that those with
divergent stereoselectivity leading to four different products
with two stereogenic centers on an optional basis will be found
without turning to mutagenesis.10 In principle, several different
reaction types can be considered in order to obtain products
having two newly created stereogenic centers, mediated either
by enzymes or man-made catalysts. For example, in P450-
BM3 catalyzed desymmetrization of appropriate achiral
compounds (or meso-compounds) such as methylcyclohexane
with formation of 1-hydroxy-2-methylcyclohexane, only one
of the four stereoisomers proved to be accessible by directed
evolution.17 A different reaction type is cyclopropanation of
styrene, e.g., catalyzed by mutant Fe-heme proteins, in which
two out of the four possible stereoisomers could be produced
with reasonable selectivity.18 It is worth noting that for some
substrates, all four possible stereoisomers in the
cyclopropanation were also obtained by using three different
enzymes.18a To the best of our knowledge, so far few
precedents are available on the directed evolution of an
enzyme for a stereodivergent reaction in which all four
possible stereoisomeric products were accessed. Hilvert and
coworkers reported the directed evolution of an artificial retro-
aldolase which was engineered for promiscuous Michael
reactions leading optionally to four stereocomplementary
products with 80-85% selectivities.19 In this interesting
process, CH-acidic Michael donors such as cyano acetic acid
ethyl ester undergo rapid racemization prior to addition to
prochiral α, ß-unsaturated ketones.
O
O
3a
(2R, 1'S)-
(2S, 1'R)-3a
Although non-aqueous reaction conditions increase the
difficulty of direction evolution, which has been rarely
implemented,24a we succeeded in engineering the selectivities
(exceeding 90%) for each of the four possible stereoisomers in
the model reaction (Scheme 1) by screening less than 100
transformants. The achievement is based on a new protein
engineering strategy which we call “focused rational iterative
site-specific mutagenesis” (FRISM). As will be seen, it is a
logical fusion of rational design and directed evolution.
RESULTS AND DISSCUSSION
Creation of Ultra-Small Focused CALB Mutant
Libraries for Screening in an Organic Solvent. In contrast
to lipase-catalyzed hydrolysis in aqueous medium,
transacylation in an organic solvent can be used more widely,
because of some specific advantages of nonaqueous
enzymology, including the high solubility of organic
substrates in organic solvents and the solvent-dependency of
enzyme properties.25 A special area of enzyme catalysis in
non-aqueous media is dynamic kinetic resolution (DKR) of
sec-alcohols or sec-amines.26 To date, almost all protein
engineering examples of lipases dealt with the hydrolysis of
esters in an aqueous medium using high-throughput screening.
An exception is the directed evolution of Candida antarctica
lipase A (CALA) for enantioselective transacylation of sec-
alcohols in organic solvents.24a It should be noted that the best
mutants obtained from hydrolysis reactions may not perform
equally well for transacylation in organic solvents, since non-
negligible solvent dependency for enzyme selectivity has been
With these challenges in mind, we considered a very
different reaction class and enzyme type. It concerns the
ACS Paragon Plus Environment

Page 3 of 15
Journal of the American Chemical Society
Following a thorough analysis of the structure of the CALB
observed,27 and even reversed enantioselectivity may occur
when changing from water to organic solvents.27b-c
active site into which we docked substrate 1a, seven mutation
sites were selected for generating the respective focused mini-
library (Fig. 1C).
1
2
3
4
5
6
7
8
9
Since the solvent has a great effect on lipase-catalyzed KR,
we first tested the activity and enantioselectivity of WT CALB
(immobilized on acrylic resin) toward the model reaction in 8
different organic solvents at 37 °C. Unfortunately, WT CALB
showed poor diastereoselectivity in all of the tested solvents, a
mixture of (2S, 1’R)-3a and (2R, 1’R)-3a being the main
products (Table S1). Considering the relatively high activity
observed in isopropyl ether (IPE), this solvent was chosen for
subsequent mutagenesis and screening steps.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The difficulty for transacylation-based screening in organic
solvents lies in the laborious purification, lyophilization or
immobilization of each mutant in a large or medium-sized
library. Although a method for microtiter-plate screening in
organic solvents has been reported, it was used for lyophilized
enzymes and not tested for enantioselectivity.28 A seminal
method for screening in organic solvents using nickel-coated
96-well microtiter plates was reported by Bäckvall et al.24a
Alternatively, in order to reduce the costs, time and labor in
the screening step, we set out to devise a mutagenesis method
that allows the rapid creation of ultra-small focused libraries
of high-quality comprising only a few dozen mutants, and
requiring minimal screening.
In order to reach this goal, we developed a strategy, in
analogy to ISM,14 called “focused rational iterative site-
specific mutagenesis” (FRISM) (Fig. 1). Instead of generating
focused saturation mutagenesis libraries at sites lining the
binding pocket in an iterative manner using highly reduced
amino acid alphabets as building blocks as in CAST/ISM14 or
in a manner based on the use saturation mutagenesis
simultaneously at several residues employing one or two
Figure 1. Representation of “focused rational iterative site-
specific mutagenesis” (FRISM) and the X-ray structure of
WT CALB.20b (A) Rationally identifying single amino acid
(AA) residues for site specific mutagenesis. (B) FRISM based
on 2 or 3 mutagenesis single sites, involving 2 or 6 upward
pathways, respectively. (C) X-ray structure of WT CALB and
amino acid residues included in the focused library. The
residues of the catalytic triad Asp187-His224-Ser105 are
shown in pink, and mutagenesis sites are shown in green/red
stick illustrations. Alanine (A), leucine (L), and phenylalanine
(F), regarded as small, medium and large amino acids,
respectively. A+, L+, and F+, regarded as amino acids with
similar steric properties used in the formation of extension
libraries, including G, V/C/I/M and Y/W, respectively.
chosen amino acids as building blocks at each position24a-b
,
FRISM calls for rational decisions on the choice of a highly
limited set of amino acids to be introduced at the chosen
CAST sites by site-specific mutagenesis in an iterative
manner. Saturation mutagenesis is not involved in this
approach. Notice that each new collection of mutants in a
FRISM step constitutes a mini-library, radically smaller in size
than the focused libraries in traditional saturation
mutagenesis.14 Depending upon the particular enzyme and the
catalytic parameter to be evolved,24 different structure-based
choices need to be made. In the present study, we focused
primarily on the steric properties of the amino acids when
applying site-specific mutagenesis, and made rational
decisions in order to reshape selected domains in the CALB
binding pocket iteratively for accepting sterically demanding
substrates.
Firstly, the crystal structure of CALB (PDB ID: 1TCA)20b
was viewed closely for choosing appropriate mutation
residues. The active site of CALB is characterized by the
catalytic triad Asp187-His224-Ser105, an oxyanion hole, and
a binding pocket composed of two domains, one for the acyl-
moiety of the ester and another for the alcohol-part,
respectively. It was logical to assume that the amino acid
residues around the active site of CAL-B are the most
influential ones for improving stereoselectivity and activity,
since many successful examples for manipulating these
enzyme parameters genetically are based on CAST.14
We then proceeded with FRISM in the following manner. A
group of amino acids with different steric properties was
chosen to replace the amino acids individually at the selected
mutation sites: Alanine (A) or glycine (G), leucine(L), and
phenylalanine(F) were regarded as small, medium and large
amino acids, respectively, for modifying the space available in
the two CALB binding pockets. Then, the obtained mutants
were tested in the model reaction (Scheme 1) to identify the
best key sites (“hotspots”) for manipulating stereoselectivity.
This procedure is related to, but not identical to a strategy
exploiting exploratory NNK-based SM at single residues
lining the binding pocket, which enables rational decisions
regarding the choice of reduced amino acid alphabets for
further CAST/ISM at multi-residue randomization sites.14
Since in the present case these simple exploratory experiments
revealed some hits, a few additional mutations with similarly
sized amino acids (A+, L+ or F+) were then introduced
individually at the key sites as an “extension-library” and
screened for enhanced enantioselectivity. A+, L+ and F+
included G, V/C/I/M and Y/W, respectively.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 15
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Focused rational iterative site-specific mutagenesis (FRISM) of CALB to obtain four enantioselective variants (A-D) as
catalysts in the transacylation reactions between 2-phenylpropionic acid p-nitrophenyl ester (rac-1a) and 1-phenylethanol (rac-2a)
providing four different enantio- and diastereomeric products 3a with diastereoselectivies ranging from 91:9 to 95:5 dr and
enantioselectivies >99:1 er. Selectivity factors E are shown separately in Table 1. Abbreviations used: 1st -4th G: 1st to 4th generation
of mutations; Sele.: selectivity, the percentage of the specific enantiomer in four different enantiomer of product 3a. Conversion
data are shown in parentheses after the selectivity data of variants. HPLC conditions: OJ-H, hexane: isopropanol =80:20. The
selectivity or percentage of four different enantiomers of product 3a in the model reaction catalyzed by various CALB mutants
generated in the directed evolution process are listed in Table S2.
In the alcohol binding pocket, residue W104 is known to be
an important site for manipulating the stereoselectivity of
CALB toward racemic secondary alcohols.21a WT CALB
displays excellent (R)-selectivity for many sec-alcohols, while
some mutants such as W104A, first reported by Hult and
coworkers,21a reverses enantioselectivity to (S). Therefore, the
decision was made to explore directed evolution along two
different routes.
the introduction of A(A+), L(L+), and F(F+) mutations,
respectively. To our satisfaction, in this small mutant
collection, Q157L proved to be the best mutant which
enhances reversed (R)-selectivity in favor of (2R, 1’R)-3a with
75% selectivity (Fig. 2B). As a consequence, we decided to
use A281G and Q157L as the templates for further mutations
at other chosen sites to improve the stereoselectivity of
(2S,1’R)-3a and (2R,1’R)-3a, respectively.
In the first route, we kept W104 unchanged to maintain (R)-
selectivity of CALB toward the alcohol part of the substrate,
and then created a single library by modifying the acid-
binding pocket for further manipulating the stereoselectivity
toward the acid part of substrate 1a. Residues A281 and A282
were targeted by site-specific mutagenesis using amino acids
A(A+), L(L+), and F(F+), respectively. Among the mutants of
the first generation, A281G was the best mutant in favor of
(2S, 1’R)-3a with a selectivity of 61% (Fig. 2A). In order to
reverse enantioselectivity of the acid part, residue Q157,
which occupies a large space and points to the oxyanion hole,
was subjected individually to site-specific mutagenesis with
In order to improve selectivity in favor of the (2S,1’R)-3a
enantiomer, the mutants of the second generation were
generated by mutating A282 to G, L, and F. The screening
results showed that A281G/A282L is the best mutant among
this mini-library, improving selectivity from 61% to 66%.
Further mutation of L+ amino acids at A282 position provided
a better mutant, A281G/A282V with 71% selectivity in favor
of (2S,1’R)-3a (Fig. 2A). The same design was applied to the
third position V190 which gives the third generation FRISM
mutants. In this generation, the best mutant was found to be
A281G/A282V/V190C, namely Mutant-SacidRalco, by pathway
ACS Paragon Plus Environment

Page 5 of 15
Journal of the American Chemical Society
281→282→190, with excellent selectivity (95/5 d.r; >99/1 e.r) structurally different racemic acids and alcohols. We first
1
2
3
4
5
6
7
8
(Figs. 2A and 2E).
tested these substrates with WT CALB (Table S3). Most WT
CALB-catalyzed transesterification reactions provided mainly
(2R, 1’R)- and (2S, 1’R)-diastereomers with very low d.r.
values, and no (2R, 1’S)- or (2S, 1’S)-products were detected.
As before, this is not unexpected, because WT CALB is
known to display a remarkably high preference for (R)-sec
alcohols and a relatively low stereoselectivity towards
substrates with chirality centers in the acid part.21 We also
tested the performance of the four best stereocomplementary
mutants in 6 different solvents (Table S4) where WT CALB
showed different activity and selectivity. Of these solvents,
isopropyl ether (IPE) was still the best choice for all the four
best mutants. It is consistent with the result of WT CALB.
In order to improve the selectivity in the formation of
(Racid,Ralco)-3a, mutant Q157L was used as a template for
iterative site-specific mutagenesis at other positions. I189 was
first mutated according to the FRISM strategy. After screening
this focused mini-library comprising I189A/L/F variants,
Q157L/I189A (Mutant-RacidRalco) was identified as an excellent
variant in the formation of (Racid,Ralco)-3a (94/6 d.r; >99/1 e.r)
(Figs. 2B and 2E).
9
In the second route, W104 was first replaced by smaller
amino acids to achieve (S)-selectivity toward the alcohol part
of the substrate. A previous report had shown that mutation
W104A induces only moderate (S)-selectivity in the acylation
of 1-phenylethanol with vinyl butanoate (E = 6.6).21a We
speculated that in our study W104A is not the only variant that
could be useful. Therefore, we created a mini-library at
position 104 containing A, L, and F mutations. To our
surprise, in the formation of the (2R, 1’S)-3a enantiomer, (R)-
selectivity of the acid part was improved significantly using
either W104A or W104L. This shows that the modification of
the alcohol pocket also alters the substrate binding mode in the
acid pocket, indicating the occurrence of cooperative
mutational effects.29 The mutant W104A with moderate
activity was then selected as the template for FRISM-based
site-specific mutagenesis in order to further enhance (2R, 1’S)-
3a-selectivity (Fig. 2C). It is worth noting that the selectivity
manipulation in the formation of (2S, 1’S)-3a is much more
difficult, because both W104A and W104L show very low
selectivity for this particular absolute configuration (Fig. 2D).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The four best stereocomplementary mutants display quite
different results with regard to substrate acceptance and
selectivity (Table 1). By using Mutant-RacidRalco
,
MutantSacidRalco, or MutantRacidSalco, respectively, a number of
aromatic p-nitrophenol esters and sec-alcohols could be
efficiently transformed to the corresponding chiral esters with
remarkably enhanced stereoselectivity. For example, Mutant-
SacidRalco has a broad substrate scope. In the acid domain of the
target products with (2S, 1’R) configuration, different
substituents on the phenyl ring, including methyl, halogen,
isobutyl and different alkyl groups at the α-position such as
methyl and ethyl are all tolerated. Similarly, in the sec-alcohol
moiety of the tested substrates, heterocyclic structures and
various substituents on the phenyl ring are also accepted.
Remarkably, fatty sec-alcohol structures in the target products
((2S, 1’R)-3f, (2S, 1’R)-3g) are efficiently differentiated by
Mutant-SacidRalco, while this cannot be achieved by WT CALB
because it usually cannot differentiate between methyl and
ethyl at the α-position of the sec-alcohol fragment of
substrates.21 Previous reports showed that WT-CALB has very
low activity and selectivity in the KR of ibuprofen esters such
as ibuprofen-PNP ester, which has a similar bulky ester group
as (2S, 1’R)-3r.22b-c Notably, Mutant-SacidRalco also offers a
very effective resolution for rac-ibuprofen, and high optical
purity of the (2S, 1’R)-3r product can be obtained. Mutant-
SacidRalco showed excellent stereoselectivity with 99/1 e.r.
values, and moderate to high distereoselectivity with
64/36∼97/3 d.r. for all the 18 compounds tested. Interestingly,
Mutant-SacidRalco also retains good stereoselectivity in the
conventional KR using substrate rac-1a alone or substrate rac-
2a alone. For example, Mutant-SacidRalco provided the (S)-
product with 21% yield and 99% ee in the transesterification
of rac-1a and benzyl alcohol. Similarly, it can also provide
(R)-product with 49% yield and 99% ee in the
transesterification of rac-2a and vinyl acetate.
For improving the selectivity of (2R, 1’S)-3a-formation, we
therefore used the W104A-template and selected Q157 and
I189 as the mutation points for the second generation FRISM.
The choice of these CAST positions was made in view of their
important effect on (R)-selectivity towards the acid motif of
the substrate. Surprisingly, most tested mutations at Q157
including A, L, and F when added to W104A, caused a notable
reduction of enzyme activity (data not shown), probably due to
the destruction of the active site in these mutants. In contrast,
the second generation FRISM mini-library with mutations at
I189 introduced to the W104A template harbored an excellent
mutant, W104A/I189V (Mutant-RacidSalco), showing high
diastereoselectivity (95/5 d.r) and enantioselectivity (>99/1
e.r) in the formation of (2R, 1’S)-3a (Figs. 2C and 2E). In
parallel work, the selective preference for (2S, 1’S)-3a was
first observed with mutant W104A/I189M (35% selectivity,
Fig. 2D) from the same second generation FRISM mini-library
as Mutant-RacidSalco. Therefore, the decision was made to
continue the evolutionary process using this mutant as the
template to enhance selectivity in the production of (2S, 1’S)-
3a, again by way of FRISM. Initially, this provided mutant
W104A/I189M/V190C, which showed 50% selectivity for
(2S, 1’S)-3a. The best mutant was discovered at the end of
pathway 104→189→190→134. After screening this focused
Compared with Mutant-SacidRalco, Mutant-RacidRalco has some
notable differences in substrate scope. Methyl at the α-position
of the acid section of the target compounds is accepted by
Mutant-RacidRalco, but not ethyl. Moreover, contrary to Mutant-
SacidRalco, fatty sec-alcohol structures in the targeted products
((2R, 1’R)-3f and (2R, 1’R)-3g) are not accepted by Mutant-
RacidRalco (Table 1), probably due to the relative large space in
the alcohol-binding pocket for small alkyl groups. Another
interesting difference was observed in the formation of (2R,
1’R)-3j with 1-substituted
mini
collection
comprising
D134A/L/F
variants,
W104A/I189M/V190C/D134L (91/9 d.r; 99/1 e.r) was
identified as the best mutant for the formation of (2S, 1’S)-3a
enantiomer (Figs. 2D and 2E).
Substrate Scope of the Evolved CALB Mutants. Next, we
wanted to see how the four best stereocomplementary mutants
perform as catalysts in the transesterification of other
Table 1. Substrate Scope of the Four Best FRISM-derived CALB Mutants as Catalysts in Stereoselective Reactions of rac-
ACS Paragon Plus Environment

Journal of the American Chemical Society
Esters and rac-Alcohols According to Scheme 1a.
Page 6 of 15
1
2
Mutant-SacidRalco
3
4
5
O
O
O
O
O
O
O
O
O
O
6
7
8
9
(2S, 1’R)-3a,
(2S, 1’R)-3b,
(2S, 1’R)-3c,
(2S, 1’R)-3d,
(2S, 1’R)-3e,
ton: 1246, yield: 45%,
95/5 d.r., >99/1 e.r.
Emajor >200
ton: 851, yield: 42%,
95/5 d.r.,b >99/1 e.r. c
Emajor >200
ton: 425, yield: 31%,
90/10 d.r., >99/1 e.r.
Emajor >200
ton: 1025, yield: 44%,
92/8 d.r., >99/1 e.r.
Emajor >200
ton:1201, yield: 43%,
83/17 d.r., >99/1 e.r.
Emajor >200
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
S
O
O
O
O
O
O
N
O
N
O
O
O
(2S, 2’R)-3f,
(2S, 2’R)-3g,
(2S, 1’R)-3h,
(2S, 1’R)-3i,
(2S, 1’R)-3j,
ton: 1025, yield: 47%,
92/8 d.r., >99/1 e.r.
Emajor >200
ton: 1210, yield: 49%,
97/3 d.r., >99/1 e.r.
Emajor >200
ton: 814, yield: 41%,
88/12 d.r., >99/1 e.r.
Emajor >200
ton: 941, yield: 44%,
97/3 d.r., >99/1 e.r.
Emajor >200
ton: 910, yield: 41%,
64/36 d.r., >99/1 e.r.
Emajor >200
F
S
O
O
O
O
O
O
Br
O
O
O
O
F
Cl
(2S, 1’R)-3k,
(2S, 1’R)-3l,
(2S, 1’R)-3m,
(2S, 1’R)-3n,
(2S, 1’R)-3o,
ton: 954, yield: 45%,
91/9 d.r., >99/1 e.r.
ton: 514, yield: 35%,
78/22 d.r., >99/1 e.r.
Emajor >200
ton: 1510, yield: 48%,
86/14 d.r., >99/1 e.r.
Emajor >200
ton: 1025, yield: 48%,
89/11 d.r., >99/1 e.r.
Emajor >200
ton: 1014, yield: 47%,
97/3 d.r., >99/1 e.r.
Emajor >200
Emajor >200
Cl
O
O
O
O
O
O
O
O
O
O
CF3
(2S, 1’R)-3p,
(2S, 1’R)-3q,
(2S, 1’R)-3r,
(S)-3v,
(R)-3w,
ton: 321, yield: 31%,
75/25 d.r., >99/1 e.r.
Emajor >200
ton: 984, yield: 43%,
97/3 d.r., >99/1 e.r.
Emajor >200
ton: 1012, yield: 45%,
92/8 d.r., >99/1 e.r.
Emajor >200
yield: 21%,
99.5/0.5 e.r.
Emajor >200
yield: 49%,
99.5/0.5 e.r.
Emajor >200
Mutant-RacidRalco
O
O
O
O
O
O
O
O
O
O
(2R, 1’R)-3a,
(2R, 1’R)-3b, n.d.d
(2R, 1’R)-3c,
(2R, 1’R)-3d,
(2R, 1’R)-3e,
ton: 987, yield: 45%,
94/6 d.r., >99/1 e.r.
Emajor >200
ton: 457, yield: 32%,
89/11 d.r., >99/1 e.r.
Emajor >200
ton: 458, yield: 34%,
92/8 d.r., >99/1 e.r.
Emajor >200
ton: 594, yield: 35%,
86/14 d.r., >99/1 e.r.
Emajor >200
S
S
O
O
O
O
O
O
N
O
O
O
O
(2R, 2’R)-3f, n.d. d
(2R, 2’R)-3g, n.d. d
(2R, 1’R)-3h,
(2R, 1’R)-3j,
(2R, 1’R)-3k,
ton: 531, yield: 38%,
75/25 d.r., 90/10 e.r.
ton: 458, yield: 35%,
97/3 d.r., 96/4 e.r.
ton: 621, yield: 41%,
67/33 d.r., >99/1 e.r.
Emajor >200
Emajor 15
Emajor 39
F
Cl
O
O
O
O
O
O
O
O
Br
O
O
CF3
Cl
(2R, 1’R)-3m,
(2R, 1’R)-3n,
(2R, 1’R)-3o,
(2R, 1’R)-3p,
(2R, 1’R)-3q,
ton: 686, yield: 42%,
97/3 d.r., >99/1 e.r.
ton: 1025, yield: 49%,
88/12 d.r., 94/6 e.r.
ton: 584, yield: 40%,
93/7 d.r., >99/1 e.r.
ton: 451, yield: 35%,
92/8 d.r., >99/1 e.r.
ton: 458, yield: 37%,
95/5 d.r., >99/1 e.r.
ACS Paragon Plus Environment

Page 7 of 15
Journal of the American Chemical Society
Emajor >200
Emajor 42
Emajor >200
Emajor >200
Emajor >200
1
2
3
4
Mutant-RacidSalco
5
6
7
O
O
O
O
O
O
O
O
O
O
8
9
(2R, 1’S)-3a,
(2R, 1’S)-3b,
(2R, 1’S)-3c,
(2R, 1’S)-3d,
(2R, 1’S)-3s,
ton: 968, yield: 43%,
95/5 d.r., >99/1 e.r.
Emajor >200
ton: 476, yield: 31%,
85/15 d.r., b 98/2 e.r.c
Emajor 75
ton: 495, yield: 33%,
83/17 d.r., 97/3 e.r.
Emajor 51
ton: 641, yield: 40%,
93/7 d.r., 98/2 e.r.
Emajor 95
ton: 676, yield: 41%,
86/14 d.r., 96/4 e.r.
Emajor 46
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
O
O
O
O
O
Br
O
O
O
O
10
F
Cl
(2R, 1’S)-3t,
(2R, 1’S)-3u, n.d. d
(2R, 1’S)-3l,
(2R, 1’S)-3m,
(2R, 1’S)-3n,
ton: 632, yield: 37%,
92/8 d.r., 98/2 e.r.
Emajor 87
ton:351, yield: 29%,
80/20 d.r., 80/20 e.r.
ton: 632, yield: 32%,
94/6 d.r., b 98/2 e.r. c
Emajor 77
ton: 1025, yield: 45%,
88/12 d.r., 96/4 e.r.
Emajor 54
Emajor
5
F
O
O
(2R, 1’S)-3o,
ton: 518, yield: 33%,
80/20 d.r., 97/3 e.r.
Emajor 51
Mutant-SacidSalco
O
O
O
O
O
O
O
O
(2S, 1’S)-3a,
(2S, 1’S)-3b, n.d. d
(2S, 1’S)-3c,
(2S, 1’S)-3d,
ton:751, yield: 31%,
91/9 d.r., >99/1 e.r.
Emajor >200
ton: 432, yield: 25%,
71/29 d.r., 96/4 e.r.
Emajor 32
ton: 311, yield: 21%,
78/22 d.r., b 94/6 e.r. c
Emajor 20
a Reaction conditions: 50 mg immobilized lipase (10% enzyme on acrylic resin), rac-1 (50 mM) and rac-2 (45 mM) in 1 ml isopropyl
ether at 37°C for 12 h. Yields and selectivity were determined by chiral HPLC. bd.r. was determined by NMR. c e.r. was determined by
combining HPLC data and NMR. d n.d. not detected.
thiophene in the alcohol moiety, and of (2R, 1’R)-3k with 2-
substrates. Relatively speaking, the substrate scope of Mutant-
SacidSalco is limited, which may be attributed to the specific
active site structure of this mutant.
substituted thiophene. Mutant-RacidRalco shows excellent
distereoselectivity for (2R, 1’R)-3j (97/3 d.r.) and low
selectivity for (2R, 1’R)-3k, while the result is exactly opposite
The reactions can be easily scaled up with good activity and
selectivity, using (2S, 1’R)-3a and (2S, 1’R)-3r as two
examples. The gram-scale reaction mixture containing 2.25
mmol 2a and 2.5 mmol rac-1a (or rac-1r) under the catalysis
of Mutant- SacidRalco provides the corresponding (2S, 1’R)-3a
with 37% isolated yield (211 mg), 96/4 d.r. and 99/1 e.r., and
(2S, 1’R)-3r with 40% isolated yield (279 mg), 93/7 d.r. and
99/1 e.r., respectively. The ester group of the target products
(e.g., (2S, 1’R)-3r) can be
to that of Mutant-SacidRalco
.
Mutant-RacidSalco also shows good substrate scope. A series
of (2R, 1’S)-esters with various substituents in the acid or
alcohol moieties can be prepared with moderate to high d.r.
and e.r. values. Especially, α-alkyl groups with different chain
length (from methyl to butyl) of the alcohol moiety ((2R, 1’S)-
3d, 3s, 3t) are accepted by Mutant-RacidSalco. However, when α-
alkyl was replaced by dodecyl, no activity was observed,
implying that the substrate is too large to fit into the pocket.
Mutant-SacidSalco also increased the target selectivity of several
Table 2. Kinetic Data of WT CALB and Four Different Mutantsa
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 15
Substrate (2R, 1’R)-3a
(2R, 1’S)-3a
(2S, 1’R)-3a
(2S, 1’S)-3a
1
2
3
4
5
6
7
8
kcat
(s-1)
kcat/Km
(M-1 s-1)
428
kcat
(s-1)
-
kcat/Km
(M-1 s-1)
-
kcat
kcat/Km
(M-1 s-1)
674
kcat
(s-1)
kcat/Km
(M-1 s-1)
-
Km
(mM)
Km
(mM)
Km
Km
(mM)
Mutant
(s-1)
(mM)
WT
21.5
2.97
9.21
-
-
32.2
27.3
21.7
-
-
-
Mutant-
RacidRalco
2.93
987
-
-
2.38
87.2
957
-
-
Mutant-
SacidRalco
19.7
14.7
12.0
1.34
68.0
2.72
0.51
-
-
-
16.6
9.87
22.5
15.9
-
-
-
9
Mutant-
RacidSalco
0.040
0.006
3.80
17.1
0.59
0.046
155
2.69
0.003
0.005
0.31
0.22
14.9
3.89
0.018
1.32
1.21
339
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Mutant-
SacidSalco
a Reaction conditions: 5mg immobilized lipase, rac-1a (3-40 mM) and rac-2a (100 mM) in 1 ml isopropyl ether at 37°C. Samples
were taken at different reaction intervals for HPLC analysis.
Figure 3. (A) Cartoon representation of the overall crystal structures of WT (PDB Code 5A7130, gray), Mutant- RacidRalco (PDB
Code 6J1R, red), Mutant- RacidSalco (PDB Code 6J1Q, orange), Mutant- SacidRalco (PDB Code 6J1P, cyan), and Mutant- SacidSalco
(PDB Code 6J1S, blue). Images B-E represent the surface of the binding pocket of these mutants (B: Mutant- RacidRalco; C: Mutant-
SacidRalco; D: Mutant- RacidSalco; E: Mutant- SacidSalco). In images B-E, some representative residues are represented by sticks with
different colors. S105 and H224 in B-E are colored in black and gray, respectively. All mutations in these four mutants are
underlined.
easily hydrolyzed by using well-established methods, to give
optical pure acids and alcohols simultaneously. This result
also documents the synthetic utility of stereodivergent
mutants of CALB in transesterification reactions.
variants was performed. This was achieved by applying site-
directed mutagenesis on WT CALB with creation of a select
series of variants (Table S5).
For Mutant-SacidRalco, three single mutations and the double
mutations derived from their combination proved to be
ineffective in improving the (2S, 1’R)-selectivity relative to
WT, while the triple mutant (V190C/A281G/A282V) was
found to display 95% selectivity toward (2S, 1’R)-
enantiomer, showing a remarkable synergistic effect29 (Table
S5, entries 1-7). Similar cooperative effects were also found
in Mutant-SacidSalco, where all deconvoluted mutants showed
low or even no preference for the (2S, 1’S)-enantiomer,
while in concert their assembly showed 91% selectivity for
the target enantiomer (Table S5, entries 14-27). It is also
worth noting that in Mutant-RacidSalco, the mutation of
W104A in the alcohol-binding pocket resulted in a
These stereocomplementary mutants have largely retained
thermostability relative to WT CALB. The measured T50
value of WT CALB (58.6 °C) and those of the four Mutants
RacidRalco (55.7 °C), SacidRalco (59.2 °C), RacidSalco (53.6 °C)
and SacidSalco (51.3 °C) suggest that the evolved mutations
responsible for manipulated stereoselectivity do not impair
the thermostability of the enzyme to any significant degree.
Deconvolution Experiments. In order to explore the
effect of each amino acid exchange in the four best variants
on the specific selectivity for the transesterification of 2-
phenyl propionic acid p-nitrophenyl ester (1a) and 1-
phenylethanol (2a), partial deconvolution of the four best
ACS Paragon Plus Environment

Page 9 of 15
Journal of the American Chemical Society
significant improvement of (2R, 1’S)-selectivity (Table S5, pocket. Since 3a and 1a are easily hydrolyzed by CALB
1
2
3
4
5
6
7
8
entry 11), completely reversing the (2S, 1’R)-preference of
WT CALB. This result speaks for “remote” mutational
effects.
mutants under most crystallization conditions, 1-
phenethylamine 2-phenyl propionic (3a’) was used as an
analog. Unfortunately, only one complex of Mutant-SacidRalco
with (2S, 1’R)-3a’ at the reaction site crystallized for X-ray
analysis (Fig. S2). As in the case of the apo-protein, two
monomers make up an asymmetric unit in the complexed
structure. Substrate (2S, 1’R)-3a’ was found to be located
slightly next to the catalytic site of the open monomer, with
an occupancy of 91%. As expected, the methyl-group of the
alcohol moiety points to the alcohol-binding pocket,
surrounded by W104, while the phenyl is kept out of the
entrance. Unfortunately, it was not possible to identify the
exact binding mode of the acid part.
Kinetic Studies. Kinetic parameters were measured for
WT-CALB and the four stereocomplementary mutants
(Table 2). It is noteworthy that WT-CALB has similar
activity (kcat/Km) towards the stereoisomers of (2R, 1’R)-3a
and (2S, 1’R)-3a, whereas very low activity was determined
in the case of (2R, 1’S)-3a and (2S, 1’S)-3a. The high
preference of WT-CALB for the (R)-sec alcohols moiety
probably plays a role. For the stereodivergent mutants,
distinct trends were observed. For example, compared to WT
CALB, the kcat/Km of Mutant-RacidSalco for (2R, 1’R)-3a and
(2S, 1’R)-3a decreased from 428 and 674 to 2.7 and 0.3 M-1s-
1, respectively, while the kcat/Km for (2R, 1’S)-3a increased
from essentially 0 to 155 M-1s-1. Similarly, Mutant-SacidSalco
displays high kcat/Km (339M-1s-1) for (2S, 1’R)-3a, while it
hardly transforms other enantiomers. In the case of Mutant-
RacidRalco and SacidRalco, the preference of the respective (2R,
1’R)- and (2S, 1’R)-enantiomers is also clearly indicated by
the correspondingly increased kcat/Km (Table 2).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Unveiling the Source of Differing Stereoselectivity of
CALB Variants. To obtain some insight regarding the
source of differing stereoselectivity of the four highly
stereocomplementary CALB variants, four enantiomeric
ligands of 3a were docked into the binding pockets of
variants RacidRalco
,
SacidRalco
,
RacidSalco and SacidSalco,
respectively, and 100-ns molecular dynamic (MD)
simulations were conducted. For the MD simulated structure,
we measured in particular the distance between the S105
oxygen of the catalytic triad and the carbonyl carbon of the
ester substrate (Fig. S3). In all four cases, the shortest
distance proved to be the one that leads to the observed
stereoselectivity. The binding enthalpies of the four ester
enantiomers in the variants were calculated using the
MM/GBSA method.32 It can be seen that the order of binding
enthalpies correlates with the respective stereoselectivities:
The stronger, the more selective (Table S8). Since the
entropy changes would be similar when the four different
stereoisomeric ligands bind with a same variant, only the
enthalpies were computed and the entropy contributions
were neglected in these calculations (Eq. 2 in supporting
information).
Crystallography. To better understand the relationship
between the manipulated stereoselectivities and specific
protein structures, we obtained X-ray crystal structures of the
four best mutants at 1.60∼1.83-Å resolution (Table S7).
Each mutant contains two chains in one asymmetric unit, and
the two monomers have almost the same structure of the
protein scaffold, except for loop 140-147, which displays
open and closed conformations, respectively (Fig. 3 and Fig.
S1). Different from that found by Stauch et al. (PDB:
5A71),30 our closed conformation is similar to the structure
reported by Uppenberg et al. (PDB: 1TCA).20b In the open
chain conformation, the flexible loop 140-147 is located
away from the entrance of the reaction chamber relative to
WT CALB, thereby exposing the active site. Our B-factor
analysis31 of loop 140-147 in all four mutants indicates
higher flexibility relative to that in CALB-WT (PDB:
1TCA).20b The structural promiscuity of the variants around
the substrate entrance may account for their lower Km values
compared to that of WT, especially toward the preferred
stereoisomers. Remarkably, the four stereodivergent variants
display significantly different structures of the respective
active sites (Fig. 3B-E). Variants with the same selectivity
toward the acid moiety show a similar shape of the acid
pocket, for example Mutant-RacidRalco vs. Mutant-RacidSalco, or
Mutant-SacidRalco vs. Mutant-SacidSalco. The (R)-acid selectivity
pockets of Mutant-RacidRalco and RacidSalco have the same
V190 residue and relatively small 189 residue (A and V),
while the (S)-acid selectivity pockets of Mutant-SacidRalco and
SacidSalco have the same C190 residue and a relatively large
189 residue (I and M). Similarly, (R)-alcohol--selective
Mutants-RacidRalco and SacidRalco have a small space of the
alcohol-binding pockets due to large W104 position, while
Mutants-RacidSalco and SacidSalco have large and deep alcohol-
binding pockets with a preference for (S)-alcohols.
The MD results also show that the binding modes of the
four stereoisomers in their respective stereoselective variants
are divergent. From the MD structure comparison between
the complex of Mutant-RacidRalco (I189A/Q157L) and (2R,
1’R)-3a,
and
the
complex
of
Mutant-SacidRalco
(V190C/A281G/A282V) and (2S, 1’R)-3a (Fig. 4A, C & D),
it was found that the binding modes of the two enantiomers
in their corresponding optimal mutants are completely
different, which are relevant to their different
enantioselectivity. In the Mutant-RacidRalco complex, the
methyl substituent of the acid moiety of (2R, 1’R)-3a points
to L157 position, while the hydrogen atom pointing to I189
at the loop 184-194 (Fig. 4A, D). In Mutant-SacidRalco, the
orientation of methyl and hydrogen of (2S, 1’R)-3a is
reversed (Fig. 4A, C). The benzene group of the acid moiety
in both enantiomers have the similar orientation, pointing to
the entrance of the acid pocket. Moreover, the whole pocket
of Mutant-SacidRalco for binding alcohol and acid is more
closed than that of Mutant-RacidRalco (I189A/Q157L) (Fig.
4A), due to the different movement of
Further, we attempted to get a complex structure which
shows how the substrates or products dock in the reaction
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 15
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. MD-simulated binding modes of various enantiomers in their respective selective CALB variants. (A) Cartoon
representation comparison of (2S, 1’R)-3a enantiomer bound in Mutant-SacidRalco (green) and (2R, 1’R)-3a enantiomer in Mutant-
RacidRalco (purple). (B) Cartoon representation comparison of (2R, 1’R)-3a enantiomer in Mutant-RacidRalco (purple) and (2R, 1’S)-
3a enantiomer in Mutant-RacidSalco (yellow). C-E images are the enlarged binding modes of (2S, 1’R)-3a enantiomer in Mutant-
SacidRalco (C), (2R, 1’R)-3a enantiomer in Mutant-RacidRalco (D) and (2R, 1’S)-3a enantiomer in Mutant-RacidSalco (E). (F) Structure
comparison of the lid 140-149 of four different mutants. (G) The surface of the binding pocket of four different mutants binding
with their respective enantiomers. In images C-E, the mutated residues are represented by orange sphere, and other representative
resides are represented by sphere or sticks with different color. The catalytic triad (S105-H224-D187) in C-E are colored in light
blue. In image G, the mutated residues are represented by orange surface.
the lid 140-149 and the helix 265-285 caused by
corresponding mutations.
Meanwhile, the mutation of I189A provides a relative large
space to accommodate the chiral center containing a large
benzene group of the acid moiety. The benzene of the acid
moiety forms hydrophobic interactions with V154, L157, I285
and A189 (Figs. 4D).
In the case of (2R, 1’R)-3a binding in the Mutant-RacidRalco
(I189A/Q157L), the ligand is exposed to three IPE solvent
molecules (Fig. 4D). Compared the apo variant, the binding of
the ligand pushes the 265-285 helix away from the active site
and I285 moves closer to the benzene group of the acid moiety
(Fig. S6B). Accordingly, L278 moves away from the alcohol
moiety and L277 becomes fully exposed pointing outward the
binding site. As a result, a large entrance to the alcohol pocket
is formed to accommodate the benzene-moiety of the alcohol
moiety (Fig. 4D, 4G and S6B) and the methyl group of the
alcohol portion forms favorable hydrophobic interaction with
W104, which accounts for the high selectivity toward the (R)-
secondary alcohol part. The phenyl group of the alcohol
moiety is nested in a hydrophobic domain formed by H224,
L278, A282 and I285 (Fig. 4D). On the other hand, with the
With the ligand (2S, 1’R)-3a binding, the phenyl group of
the alcohol moiety establishes favorable hydrophobic
interactions with H224, L278 and V282 in Mutant-SacidRalco
(V190C/A281G/A282V) (Fig. 4C, S5A and S6A). The
conformational change around the helix 265-285 induced by
A282V/A281G mutations results in a closed alcohol pocket
(Fig. 4A, 4C and 4G). The methyl group of the alcohol portion
forms favorable hydrophobic interaction with W104 and
thereby accounts for the high selectivity toward the (R)-
secondary alcohol part. On the other hand, the V190C
mutation causes the conformational change of the loop where
it is located. As a result, the acid-selective pocket of CALB is
reshaped and the orientation of methyl and hydrogen of (2S,
1’R)-3a is reversed compared with the complex of Mutant-
RacidRalco accounting for the (S)-selectivity of the acid pocket.
The benzene of acid moiety forms hydrophobic interactions
with I189, as well as T138, Q157 and V154.
ligand bound, the lid region exhibits
a more open
conformation compared to the Mutant-SacidRalco complex. The
incorporation of medium size leucine at position 157 provides
suitable space and hydrophobic interaction with the methyl of
the acid moiety, such that smaller or larger mutations caused
the decrease of target selectivity (Fig.2B). The pointing of
methyl to L157 pushes the ligand closer the helix 265-285.
Similar differences between the orientation of the acid
moiety in (2R, 1’S)-3a and (2S, 1’S)-3a enantiomers in their
ACS Paragon Plus Environment

Page 11 of 15
Journal of the American Chemical Society
optimal mutants are also observed (Fig. S4B and 4G). possible stereoisomers, as in the case of the therapeutic drug
1
2
3
4
5
6
7
8
Interestingly, two mutants with the same acid-selectivity show
the similar conformation of the lid 140-149 (Fig. 4F).
candidates of the ibuprofen-type (Table 1). Moreover, both
chiral alcohols and acids are extensively used in
pharmaceutical industry, and the four stereodivergent
enantiomers can be hydrolyzed to obtain optical pure acids
and alcohols, simultaneously. The best stereocomplementary
mutants display relatively broad substrate scopes, except for
Additionally, the MD structure comparison between the
complex of Mutant-RacidRalco (I189A/Q157L) and (2R, 1’R)-3a,
and the complex of Mutant-RacidSalco (W104A/I189V) and (2R,
1’S)-3a (Fig 4B) offers insight about the source of different
sec-alcohol stereoselectivity of CALB variants. It was found
that both the conformation of the acid pockets of the two
mutants and the orientation of their ligands are very similar,
implying their similar acid-selectivity. The reversion of
alcohol selectivity was mainly caused by the W104A mutation,
which creates extra space for the accommodation of large-
sized benzene group in the alcohol moiety of ligands, and
therefore accounts for the reversed (S)-alcohol-selectivity for
(2R, 1’S)-3a in Mutant-RacidSalco (W104A/I189V) (Fig. 4E and
4G). W104A mutation not only changes the alcohol
selectivity, but also influences the binding mode of the acid
moiety of the ligand. The whole ligand moves closer to the
helix 265-285, due to the sliding of the benzene of the alcohol
moiety into the alcohol-binding pocket (Fig. 4B and 4E). This
movement provides extra space for the methyl substituent
pointing to Q157, accounting for the astonishing increase of
(R)-acid selectivity from essentially 0 (WT) to 81% (W104A)
(Fig. 2C). Moreover the mutation of I189V further improves
this selectivity, similar to that in the case of Mutant-
RacidRalco(Fig. 4G and 2C).
Mutant-SacidSalco. Importantly, a significant tradeoff in
thermostability was not observed in the evolved variants.
Consequently, the range of CALB applications has been
expanded considerably. In order to gain structural insight into
the source of stereodivergent selectivity of the four best
mutants, crystals were grown and studied by X-ray analysis,
flanked by docking studies and molecular dynamics (MD)
simulations.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Finally, we point out that the FRISM method of protein
engineering constitutes a logical fusion of rational design21,24,33
and directed evolution.14,16 The decisions concerning the
choice of the amino acids in the site-specific mutagenesis
steps can be supported by X-ray structural and consensus data
as well as computational aids such as machine learning34 or
Rosetta design.33d We expect that in the future, FRISM will be
used in the protein engineering studies of other enzymes for
enhancing stereo- and regioselectivity as well as activity
and/or thermostability.
MATERIALS AND METHODS
Conclusions and Perspectives. In summary, by way of a
highly focused protein engineering method dubbed “focused
rational iterative site-specific mutagenesis” (FRISM), we have
addressed the challenging issue of stereodivergent engineering
of Candida antarctica lipase (CALB) with generation of four
highly stereocomplementary variants needed to access
catalytically all four stereoisomers bearing two stereocenters
in transesterification reactions between racemic acids and
racemic alcohols in an organic solvent. Although non-aqueous
reaction conditions are known to be challenging in directed
enzyme evolution,16,24a our FRISM strategy required the
generation and screening of less than a total of 100
transformants in the quest to achieve >90% selectivity for all
of the four possible stereoisomers in the model reaction. This
was accomplished by first applying exploratory site-directed
mutagenesis at 7 selected sites lining the binding pocket of
CALB using only 3 representative amino acids, including
alanine (A), leucine (L), and phenylalanine (F). Following the
identification of key mutations at the identified hotspots,
further site-specific mutagenesis was performed iteratively
using similar-sized amino acids (A+: G; L+: V/C/I/M; F+:
Y/W). Saturation mutagenesis with formation of traditional
libraries was not needed.
Library Generation. The PCR (50 μL final volume)
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
(GATGCCGGGAGCAGACAAGCCCGTCAGGGCGC, 5μL,
2.5 μM), template plasmid (pETM11-CALB, 1.0 μL, 100
ng/μL) and 1 μL of Pfu polymerase. PCR conditions used
were 94 °C, 5 min; 30 cycles of (94 °C, 1 min; 60 °C, 1 min;
72 °C, 14 min) and final extension at 72 °C, 10 min. Then, the
PCR mixtures were mixed with 2 μL DpnI (10 U/μL) and
incubated overnight at 37 oC. The product was purified with an
Omega PCR purification spin column, and an aliquot of 15 μL
was used to transform 80 μL of electrocompetent E. coli BL21
(DE3) cells, which contain chaperone plasmid pGro7 (Takara,
Japan). The transformation mixture was incubated with 1 mL
of LB medium at 37 °C under shaking of 200 rpm for 1 hour
and spread on LB-agar plates containing Kanamycin (34
mg/L) and Chloramphenicol (34 mg/L), and the plates were
incubated for 16-24 hours. A single colony was picked and
incubated in 5 mL of LB medium at 37 °C overnight. The
plasmid was extracted with an Omega gel extraction column,
and sequenced by Sangon Biotech (Shanghai, China). Target
mutants were stored with glycerol at -80 °C.
In most previous protein engineering studies of CALB or
other lipases, the respective binding pocket was “divided” into
two parts, the acid (acyl) and the alcohol binding domains.21,22
Mutations were then restricted to the chirality-containing
domain, defined by the nature of the substrate. In the present
study, however, we discovered that mutations in one domain
influence the stereoselectivity of the other, a result of “remote”
effects.
Scaffold Expression. 100 μL stored bacteria were first
inoculated in 5ml LB medium (containing 34 μg/mL
Kanamycin and 34 μg/mL Chloramphenicol), and shaken
overnight as preculture. A fresh 1L of TB medium with 1.0
mg/mL L-arabinose as the inducer for expression of chaperone
pGro7 and the same concentration antibiotics as above, was
inoculated from 5 mL preculture. The cultures were allowed to
grow at 37 °C until OD (600) at 0.6～0.8. After cooling at 4°C
for 1h, 1.0 mM isopropyl β-thiogalactopyranoside (IPTG) was
added to induce CALB expression. The cultures were allowed
to express at 16 °C for 48h with shaking at 200 rpm. Then
The experimental results concerning complete control of
both absolute and relative configuration in compounds having
two stereogenic centers, catalyzed by CALB mutants, are of
synthetic significance because they provide access to all
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 15
o
cells were harvested by centrifugation at 9000 rpm and 4 C
(PDB Code 6J1R), Mutant-RacidSalco (PDB Code 6J1Q),
1
2
3
for 10 min and the supernatants were discarded. The cells
were resuspended in 20mL 50 mM phosphate buffer (pH 7.4)
and stored at -80°C.
Mutant- SacidRalco (PDB Code 6J1P), and Mutant-SacidSalco
(PDB Code 6J1S), and Mutant-SacidRalco in complex with 3a’
(PDB Code 6J1T) have been deposited in the Protein Data
Bank (PDB).
4
5
6
7
8
9
Purification and Immobilization of CALB Mutants. The
cells were subjected to repeated freezing and thawing for 3
times, and then the target proteins were released by sonication.
The cell debris was removed by centrifugation at 12,000 rpm
for 15 min at 4.0 °C. The supernatant was filtered and loaded
on a GE Healthcare HisTrap FF Crude column (5 mL)
preequilibrated with 50 mM phosphate buffer (pH 7.4)
containing 0.5 M NaCl and 5 mM imidazol. Impurity and
chaperone were removed by imidazol at the concentration of
about 70 mM and the enzyme was eluted by 50 mM Tris-HCl
buffer with 0.5 M NaCl and 200 mM imidazol. The proteins
were dialyzed by 50 mM phosphate buffer (pH 7.4) for one
day at 4°C. Then, the purified enzyme solution was mixed
with acrylic resin in 50 mL EP tube and stirred end-over-end
rotation at 20 °C, overnight. Then, the resin was filtered and
freeze-dried in vacuo. The immobilized lipase was used for the
kinetic resolution in organic phase.
ASSOCIATED CONTENT
Supporting Information.
This material is available free of charge via the Internet at
http://pubs.acs.org.” Complete experimental procedures, details of
molecular calculations, all chiral chromatograms and NMR
spectra of products.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
AUTHOR INFORMATION
Corresponding Author
*wuqi1000@163.com, or llc123@zju.edu.cn;
*reetz@mpi-muelheim.mpg.de;
*m.huang@qub.ac.uk;
*jiahai@mail.sioc.ac.cn;
ORCID
General Procedure for Kinetic Resolution of Substrates.
50 mg immobilized CALB (10% enzyme on acrylic resin) was
added to 1 mL anhydrous IPE containing substrates rac-1a
(50mM) and rac-2a (45mM), and this mixture was shaken at
200 rpm and 37 °C for about 12 h. The reaction mixture was
filtered and evaporated in vacuo. Conversion and selectivity
were determined by chiral HPLC using specific conditions as
shown in the section of HPLC chromatograms. The crude
products were purified by silica gel chromatography
(petroleum ether and ethyl acetate).
Qi Wu: 0000-0001-9386-8068
Manfred T. Reetz: 0000-0001-6819-6116
Author Contributions
‡These authors contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENT
Q.W. and X.F.L. thank the National Natural Science Foundation
of China (21574113, 21472169), Natural Science Foundation of
Zhejiang Province (LY19B020014), and the Fundamental
Research Funds for the Central Universities (2018QNA3010) for
funding support. M.H. thanks INVEST NI Research and
Development Programme for funding support. J.Z. thanks the
Strategic Priority Research Program (B) of the Chinese Academy
of Sciences (XDB20000000) for funding support. M.T.R. thanks
the Max-Planck-Society and the Arthur C. Cope Fund for
financial support. We thank the staffs from BL19U1 at Shanghai
Synchrotron Radiation Facility (SSRF, China) for assistance
during X-ray data collection.
Scale-up of Kinetic Resolution of Substrate (2S, 1’R)-3a
and (2S, 1’R)-3r. 150 mg immobilized CALB (10% enzyme
on acrylic resin) was added to 5 mL anhydrous IPE containing
1a (or 1r) (500mM) and 2a (450mM), and this mixture was
shaken at 200 rpm and 37 °C for about 48h. The process was
monitored by chiral HPLC. The mixture was filtered and
evaporated under reduced pressure, and purified by column
chromatography to provide the corresponding (2S, 1’R)-3a
with 37% isolated yield (211 mg), 96/4 d.r. and 99/1 e.r., and
(2S, 1’R)-3r with 40% isolated yield (279 mg), 93/7 d.r. and
99/1 e.r., respectively.
X-ray Structure Determination. Four best variants,
Mutant-RacidRalco, Mutant-SacidRalco, Mutant-SacidSalco, Mutant-
RacidSalco and Mutant-SacidRalco in complex with 3a’ were
crystalized by using the sitting-drop vapor diffusion method at
18 °C. All crystals were mounted in nylon loops and flash-
frozen in liquid nitrogen. All diffraction data was collected at
the wavelength of 0.97775 Å on SSRF beamline 19U1 of the
National Center for Protein Science Shanghai (China). All
data collection was performed at 100 K. All data sets were
indexed, integrated, and scaled using the HKL2000 package.35
The apo-mutant structures were solved by molecular
replacement method using the program PHASER,36 and the
structure of WT-CALB (PDB code 1TCA) as a search model.
The complex structure was solved by the same method using
the structure of Mutant-SacidRalco as a search model. Rounds of
automated refinement were performed with PHENIX,37 and
the models were extended and rebuilt manually with COOT.38
The statistics for data collection and crystallographic
refinement are summarized in supporting information. The
atomic coordinates of four best variants Mutant-RacidRalco
REFERENCES
(1) (a) Drug Stereochemistry: Analytical Methods and
Pharmacology, 3rd ed.; Jóźwiak, K., Lough, W. J., Wainer, I. W.,
Eds.; CRC Press: Boca Raton, FL, 2012. (b) The Practice of
Medicinal Chemistry, 3rd ed.; Wermuth, C. G., Ed.; Elsevier:
Amsterdam, 2008.
(2) Bren, L. Frances Oldham Kelsey. FDA Medical Reviewer
Leaves Her Mark on History. FDA Consum. 2001, 35(2), 24-29.
(3) (a) Zhang, A. P.; Xie, X. M.; Liu, W. P. Enantioselective
Separation and Phytotoxicity on Rice Seedlings of Paclobutrazol. J
Agri Food Chem 2011, 59, 4300-4305.
(4) (a) Wilkinson, R. G.; Shepherd, R. G.; Thomas, J. P.; Baughn,
C. Stereospecificity in a New Type of Synthetic Antituberculous
Agent. J. Am. Chem. Soc. 1961, 83, 2212-2213. (b) Rubin-Preminger,
J. M.; Bernstein, J.; Harris, R. K.; Evans, I. R.; Ghi, P. Y. Variable
Temperature Studies of a Polymorphic System Comprising Two Pairs
of
Enantiotropically
Related
Forms:ꢀ
[S,S]-Ethambutol
Dihydrochloride. Cryst. Growth Des. 2004, 4(3), 431–439.
ACS Paragon Plus Environment

Page 13 of 15
Journal of the American Chemical Society
(5) (a) Gridnev, I. D., P. A. Dub, Enantioselection in Asymmetric
Molecular Basis for Their Enantiopreference, and Prospects for
Mirror-image Biotransformations. Angew. Chem., Int. Ed. 2008, 47,
8782−8793.
Catalysis, CRC Press, 2016. (b) Blaser, H.-U., Schmidt, E. (eds),
Asymmetric Catalysis on Industrial Scale: Challenges, Approaches
and Solutions, Wiley-VCH, Weinheim, 2011. (c) Walsh, P. J.,
Kozlowsky, M. C., Fundamentals of Asymmetric Catalysis,
University Science Books, 2009.
1
2
3
4
5
6
7
8
(16) For selected reviews of directed evolution,14 see: (a) Turner,
N. J. Directed Evolution Drives the Next Generation of Biocatalysts.
Nat. Chem. Biol. 2009, 5, 567–573. (b) Currin, A.; Swainston, N.;
Day, P. J.; Kell, D. B. Synthetic Biology for the Directed Evolution of
Protein Biocatalysts: Navigating Sequence Space Intelligently. Chem.
Soc. Rev. 2015, 44, 1172–1239. (c) Denard, C. A.; Ren, H.; Zhao, H.
Improving and Repurposing Biocatalysts via Directed Evolution.
Curr. Opin. Chem. Biol. 2015, 25, 55–64. (d) Reetz, M. T. Directed
Evolution of Selective Enzymes: Catalysts for Organic Chemistry and
Biotechnology. Wiley-VCH Weinheim, (2016). (e) Hammer, S. C.;
Knight, A. M.; Arnold, F. H. Design and Evolution of Enzymes for
Non-natural Chemistry. Curr. Opin. Green Sustain. Chem. 2017, 7,
23-30. (f) Zeymer, C., Hilvert, D. Directed Evolution of Protein
Catalysts, Annu. Rev. Biochem. 2018, 87, 131-157.
(6) E. N. Jacobsen. Asymmetric catalysis of epoxide ring-opening
reactions, Acc. Chem. Res. 2000, 33, 421-431.
(7) (a) Krautwald, S.; Sarlah, D.; Schafroth, M. A.; Carreira, E. M.
Enantio- and Diastereodivergent Dual Catalysis: α-Allylation of
Branched Aldehydes, Science 2013, 340, 1065-1068. (b) Sandmeier,
T.; Krautwald, S.; Zipfel, H.F.; Carreira, E.M. Stereodivergent Dual
Catalytic α-Allylation of Protected α-Amino- and α-
Hydroxyacetaldehydes, Angew. Chem. Int. Ed. 2015, 54, 14363 –
14367.
(8) Shi, S.-L.; Wong, Z. L.; Buchwald, S. L. Copper-catalysed
Enantioselective Stereodivergent Synthesis of Amino Alcohols,
Nature 2016, 532, 353-356.
(9) Selected reviews of catalysis involving more than one
stereocenter: (a) Krautwald, S.; Carreira, E. M. Stereodivergence in
Asymmetric Catalysis, J. Am. Chem. Soc. 2017, 139, 5627−5639; (b)
Bihani, M.; Zhao, J.C.G. Advances in Asymmetric Diastereodivergent
Catalysis, Adv. Synth. Catal. 2017, 359(4), 534-575. (c) Bhat, V.;
Welin, E.R.; Guo, X.L.; Stoltz, B. M. Advances in Stereoconvergent
Catalysis from 2005 to 2015: Transition-Metal-Mediated
Stereoablative Reactions, Dynamic Kinetic Resolutions, and Dynamic
Kinetic Asymmetric Transformations, Chem. Rev. 2017, 117(5),
4528–4561. (d) Zheng, C.; You, S.L. Catalytic Asymmetric
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(17) Roiban, G.-D., Agudo, R., Reetz, M. T. Cytochrome P450
Catalyzed Oxidative Hydroxylation of Achiral Organic compounds
with Simultaneous Creation of Two Chirality Centers in a Single C-H
Activation Step. Angew. Chem. Int. Ed. 2014, 126, 8803-8807.
(18) (a) Knight, A.M.; Jennifer Kan, S. B.; Lewis, R.D.;
Brandenberg, O. F.; Chen, K.; Arnold, F.H. Diverse Engineered
Heme Proteins Enable Stereodivergent Cyclopropanation of
Unactivated Alkenes. ACS Cent. Sci., 2018, 4, 372–377. (b) Bajaj, P.;
Sreenilayam, G.; Tyagi, V.; Fasan, R. Gram-Scale Synthesis of Chiral
Cyclopropane-Containing Drugs and Drug Precursors with
Engineered Myoglobin Catalysts Featuring Complementary
Stereoselectivity. Angew. Chem., Int. Ed. 2016, 55, 16110−16114. (c)
Tinoco, A.; Steck, V.; Tyagi, V.; Fasan, R. Highly Diastereo- and
Enantioselective Synthesis of Trifluoromethyl Substituted
Dearomatization by Transition-Metal Catalysis:
A Method for
Transformations of Aromatic Compounds, Chem 2016, 1, 830–857.
(e) Trost, B.M.; Rao, M. Development of Chiral Sulfoxide Ligands
for Asymmetric Catalysis, Angew. Chem. Int. Ed. 2015, 54, 5026-
5043. (f) Chen, D.F.; Han, Z.Y.; Zhou, X.L.; Gong, L.Z. Asymmetric
Organocatalysis Combined with Metal Catalysis: Concept, Proof of
Concept, and Beyond, Acc. Chem. Res. 2014, 47, 2365–2377. (g)
Brak, K.; Jacobsen, E.N. Asymmetric Ion-Pairing Catalysis, Angew.
Chem. Int. Ed. 2013, 52, 534-561. (h) Mahlau, M.; List, B.
Asymmetric Counteranion-Directed Catalysis: Concept, Definition,
and Applications, Angew. Chem. Int. Ed. 2013, 52, 518-533. (i) Yoon,
M.; Srirambalaji, R.; Kim, K. Homochiral Metal–Organic
Frameworks for Asymmetric Heterogeneous Catalysis, Chem. Rev.
2012, 112, 1196–1231. (j) Shibasaki, M.; Kanai, M.; Matsunaga, S.;
Kumagai, N. Recent Progress in Asymmetric Bifunctional Catalysis
Using Multimetallic Systems, Acc. Chem. Res. 2009, 42, 1117–1127.
(10) Reviews of enzyme catalysis in organic and pharmaceutical
chemistry: (a) Biocatalysis in the Pharmaceutical and Biotechnology
Industries, Patel, R. N. (ed.), CRC Press, 2006. (b) Goswami, A.,
Stewart, J. D., Organic Synthesis Using Biocatalysts, Academic Press,
2015. (c) Enzyme Catalysis in Organic Synthesis, Drauz, K., Gröger,
H., May, O. (eds.), Wiley-VCH, Weinheim, 2012. (d) Faber, K.,
Biotransformations in Organic Chemistry: A Textbook, Springer,
Stuttgart, 2011. (e) Liese, A., Seelbach, K., Wandrey, C., Industrial
Biotransformations, Wiley-VCH, Weinheim, 2006.
(11) Archelas, A.; Iacazio, G., Kotik, M. Epoxide hydrolases and
their application in organic synthesis, in Green Biocatalysis (Patel, R.
N., ed.), Wiley, 2016: 978-1-118-82229-6.
(12) Li, A.; Ilie, A.; Sun, Z.; Lonsdale, R.; Xu, J.H.; Reetz, M.T.
Whole-Cell-Catalyzed Multiple Regio- and Stereoselective
Functionalizations in Cascade Reactions Enabled by Directed
Evolution. Angew. Chem. Int. Ed. 2016, 55, 12026-12029.
(13) Erdmann, V.; Lichman, B.R.; Zhao, J.; Simon, R. C.; Kroutil,
W.; Ward, J.M.; Hailes, H.C.; Rother, D. Enzymatic and
Chemoenzymatic Three-Step Cascades for the Synthesis of
Cyclopropanes
via
Myoglobin-Catalyzed
Transfer
of
Trifluoromethylcarbene. J. Am. Chem. Soc. 2017, 139, 5293−5296
(19) Garrabou, X.; Macdonald, D. S., Wicky, B.I.M., Hilvert, D.
Stereodivergent Evolution of Artificial Enzymes for the Michael
Reaction. Angew. Chem. Int. Ed. 2018, 57, 5288 –5291
(20) Selected key studies and reviews of CALB: (a) Anderson, E.
M.; Larsson, K. M.; Kirk, O. One Biocatalyst-many Applications: the
Use of Candida Antarctica B-lipase in Organic Synthesis. Biocatal.
1998, 16, 181 – 204. (b) Uppenberg, J.; Hansen, M. T.; Patkar, S.;
Jones, T. A. The Sequence, Crystal Structure Determination and
Refinement of Two Crystal Forms of Lipase B from Candida
Antarctica. Structure 1994, 2, 293-308. (c) Otto, R. T.; Scheib, H.;
Bornscheuer, U. T.; Pleiss, J.; Syldatk, C.; Schmid, R. D. Substrate
Specificity of Lipase B from Candida Antarctica in the Synthesis of
Arylaliphatic Glycolipids. J. Mol. Catal. B: Enzym. 2000, 8, 201211.
(21) Key studies of protein engineering of CALB using rational
design: (a) Magnusson, A. O.; Takwa, M.; Hamberg, A.; Hult, K. An
S-selective Lipase was Created by Rational Redesign and the
Enantioselectivity Increased with Temperature. Angew. Chem. Int.
Ed. 2005, 117, 4658-4661. (b) Engström, K.; Vallin, M.; Syrén, P. O.;
Hult, K.; Bäckvall, J. E. Mutated Variant of Candida Antarctica
Lipase B in (S)-selective Dynamic Kinetic Resolution of Secondary
Alcohols. Org. Biomol. Chem. 2011, 9, 8182. (c) Marton, Z.;
Léonard-Nevers, V.; Syrén, P.-O.; Bauer, C.; Lamare, S.; Hult, K.;
Tranc, V.; Graber, M. Mutations in the Stereospecificity Pocket and at
the Entrance of the Active Site of Candida Antarctica Lipase B
Enhancing Enzyme Enantioselectivity. J. Mol. Catal. B: Enzym. 2010,
65, 1117. (d) Larsen, M. W.; Zielinska, D. F.; Martinelle, M.;
Hidalgo, A.; Juhl Jensen, L.; Bornscheuer, U. T.; Hult, K.
Suppression of Water as a Nucleophile in Candida Antarctica Lipase
B Catalysis. ChemBioChem, 2010, 11, 796801.
(22) Selected studies of protein engineering of CALB using
directed evolution21: (a) Qian, Z.; Fields, C. J.; Lutz, S. Investigating
the Structural and Functional Consequences of Circular Permutation
on Lipase B from Candida Antarctica. ChemBioChem 2007, 8,
19891996. (b) Qin, B.; Liang, P.; Jia, X.; Zhang, X.; Mu, M.; Wang,
X.Y.; Ma, G.Z.; Jin, D.N.; You, S. Directed Evolution of Candida
Antarctica Lipase B for Kinetic Resolution of Profen Esters. Catal.
Commun. 2013, 38, 1–5. (c) Wu, Q.; Soni, P.; Reetz, M.T. Laboratory
Stereochemically
Complementary
Trisubstituted
Tetrahydroisoquinolines. Angew. Chem. Int. Ed. 2017, 56, 12503-
12507.
(14) Reetz, M. T. Laboratory Evolution of Stereoselective
Enzymes: A Prolific Source of Catalysts for Asymmetric Reactions.
Angew. Chem. Int. Ed. 2011, 50, 138-174.
(15) Mugford, P. F.; Wagner, U. G.; Jiang, Y.; Faber, K.;
Kazlauskas, R. J. Enantiocomplementary Enzymes: Classification,
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 15
Evolution of Enantiocomplementary Candida Antarctica Lipase B
Mutants with Broad Substrate Scope. J. Am. Chem. Soc. 2013, 135,
1872−1881.
(33) (a) Pleiss, J. Rational Design of Enzymes. In Enzyme Catalysis
in Organic Synthesis, 3rd. ed., Drauz, K.; Gröger, H.; May, O., eds.,
Wiley-VCH, Weinheim, 2012, pp. 89-117. (b) Ema, T.; Nakano, Y.;
Yoshida, D.; Kamata, S.; Sakai, T. Redesign of enzyme for improving
catalytic activity and enantioselectivity toward poor substrates:
Manipulation of the transition state, Org. Biomol. Chem. 2012, 10,
6299-6308. (c) Steiner, K.; Schwab, H. Recent advances in rational
approaches for enzyme engineering, Comput. Struct. Biotechnol.
2012, 2, e201209010. (d) Richter, F.; Leaver-Fay, A.; Khare, S. D.;
Bjelic, S.; Baker, D. De Novo Enzyme Design Using Rosetta 3. PLOS
ONE 2011, 6, e19230.
1
2
3
4
5
6
7
8
(23) (a) Steinreiber, J.; Faber, K.; Griengl, H. De-racemization of
Enantiomers versus De-epimerization of Diastereomers
-
Classification of Dynamic Kinetic Asymmetric Transformations
(DYKAT). Chem. Eur. J. 2008, 14, 8060-8072. (b) Schober, M.;
Faber, K. Inverting Hydrolases and Their Use in Enantioconvergent
Biotransformations. Trends Biotechnol. 2013, 31, 468-478.
(24) (a) Wikmark, Y.; Svedendahl Humble, M.; Bäckvall, J.-E.
Combinatorial Library Based Engineering of Candida Antarctica
Lipase A for Enantioselective Transacylation of Sec-alcohols in
Organic Solvent. Angew. Chem. Int. Ed. 2015, 54, 4284-4288. (b)
Sandström, A.G.; Wikmark, Y.; Engström, K.; Nyhlén, J.; Bäckvall,
J.-E. Combinatorial Reshaping of the Candida Antarctica Lipase A
Substrate Pocket for Enantioselectivity Using an Extremely
Condensed Library. Proc. Natl. Acad. Sci. USA 2012, 109, 78-83. (c)
Vargas, D.A.; Tinoco, A.; Tyagi, V.; Fasan, R. Myoglobin-Catalyzed
C-H Functionalization of Unprotected Indoles. Angew. Chem. Int. Ed.
2018, 57, 9911–9915.
9
(34) Selected key studies and reviews of machine learning: (a)
Geng, C.; Vangone, A.; Folkers, G.E.; Xue, L.C.; Bonvin, A.M. iSEE:
Interface Structure, Evolution, and Energy-based Machine Learning
Predictor of Binding Affinity Changes Upon Mutations. Proteins.
2019, 87, 110-119. (b) Saito, Y.; Oikawa, M.; Nakazawa, H.; Niide,
T.; Kameda, T.; Tsuda, K.; Umetsu, M. Machine-Learning-Guided
Mutagenesis for Directed Evolution of Fluorescent Proteins. ACS
Synth Biol. 2018, 7(9), 2014-2022. (c) Feng, X.; Sanchis, J.; Reetz, M.
T.; Rabitz, H. Enhancing the Efficiency of Directed Evolution in
Focused Enzyme Libraries by the Adaptive Substituent Reordering
Algorithm, Chem. Eur. J. 2012, 18, 5646-5654. (d) Musdal, Y.;
Govindarajan, S.; Mannervik, B. Exploring Sequence-Function Space
of a Poplar Glutathione Transferase Using Designed Information-rich
Gene variants, Prot. Eng. Des. Sel. 2017, 30, 543-449. (e) Bedbrook,
C. N.; Yang, K. K.; Rice, A. J.; Gradinaru, V.; Arnold, F. H. Machine
Learning to Design Integral Membrane Channelrhodopsins for
Efficient Eukaryotic Expression and Plasma Membrane Localization,
PLOS Comput. Biol. 2017, 13, e1005786. (f) Barley, M. H.; Turner,
N. J.; Goodacre, R. Improved Descriptors for the Quantitative
Structure-Activity Relationship Modeling of Peptides and Proteins, J.
Chem. Inf. Model 2018, 58, 234-243. (g) Cadet, F.; Fontaine, N.; Li,
G.; Sanchis, J.; Ng Fuk Chong, M.; Pandjaitan, R.; Vetrivel, I.;
Offmann, B.; Reetz, M.T. A Machine Learning Approach for Reliable
Prediction of Amino Acid Interactions and Its Application in the
Directed Evolution of Enantioselective enzymes. Sci Rep. 2018, 8,
16757. (h) Li, G.; Dong, Y.: Reetz, M. T. Can Machine Learning
Revolutionize Directed Evolution of Selective Enzymes? Adv. Synth.
Catal. 2019, DOI: 10.1002/adsc.201900149.
(35) Otwinowski, Z.; Minor, W. Processing of X-ray Diffraction
Data Collected in Oscillation Mode. Method. Enzymol. 1997, 276,
307-326.
(36) McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P. D.; Winn,
M. D.; Storoni, L.C.; Read, R. J. Phaser Crystallographic Software. J.
Appl. Crystallogr. 2007, 40, 658-674.
(37) Adams, P. D.; Afonine, P. V.; Bunkoczi, G.; Chen, V. B.;
Davis, I. W.; Echols, N.; Headd, J. J.; Hung, L. W.; Kapral, G. J.;
Grosse-Kunstleve, R. W.; McCoy, A. J.; Moriarty, N. W.; Oeffner,
R.; Read, R. J.; Richardson, D. C.; Richardson, J. S.; Terwilliger, T.
C.; Zwart, P. H. PHENIX: a Comprehensive Python-based System for
Macromolecular Structure Solution. Acta Crystallogr. D. Biol.
Crystallogr. 2010, 66, 213-221.
(38) Emsley, P.; Lohkamp, B.; Scott, W. G.; Cowtan, K. Features
and Development of Coot. Acta Crystallogr. D. Biol. Crystallogr.
2010, 66, 486-501.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(25) Klibanov, A. M. Improving Enzymes by Using Them in
Organic Solvents. Nature, 2001, 409, 241-246.
(26) (a) Martin-Matute, B., Bäckvall, J.-E. Dynamic Kinetic
Resolution Catalyzed by Enzymes and Metals. Curr. Opin. Chem.
Biol. 2007, 11, 226-232. (b) Lee, J. H., Han, K., Kim, M.-J., Park, J.
Chemoenzymatic Dynamic Kinetic Resolution of Alcohols and
Amines. Eur. J. Org. Chem. 2010, 999-121. (c) El-Sepelgy, O.;
Alandini, N.; Rueping, M. Merging Iron Catalysis and Biocatalysis
Iron Carbonyl Complexes as Efficient Hydrogen Autotransfer
Catalysts in Dynamic Kinetic Resolutions. Angew. Chem. Int. Ed.
2016, 55, 13602-13605.
(27) (a) Yang, B.; Lihammar, R.; Bäckvall, J.-E. Investigation of
the Impact of Water on the Enantioselectivity Displayed by CALB in
the Kinetic Resolution of δ-functionalized Alkan-2-ol Derivatives.
Chem. Eur. J. 2014, 20, 13517–13521. (b) Savile, C. K.; Kazlauskas,
R. J. How Substrate Solvation Contributes to the Enantioselectivity of
Subtilisin Toward Secondary Alcohols. J. Am. Chem. Soc. 2005, 127,
12228–12229. (c) Wescott, C. R.; Klibanov, A. M. The Solvent
Dependence of Enzyme Specificity. Biochim. Biophys. Acta 1994,
1206, 1–9.
(28) Konarzycka-Besslor, M.; Bornscheuer, U. T.
A High-
Throughput-screening Method for Determining the Synthetic Activity
of Hydrolases. Angew. Chem. Int. Ed. 2003, 42, 1418–1420.
(29) Reetz, M. T. The Importance of Additive and Non-Additive
Mutational Effects in Protein Engineering. Angew. Chem. Int Ed.
2013, 52, 2658-2666.
(30) Stauch, B.; Fisher, S. J.; Cianci, M. Open and Closed States of
Candida Antarctica Lipase B: Protonation and the Mechanism of
Interfacial Activation. J. Lipid. Res. 2015, 56, 2348-2358.
(31) Sun, Z.; Liu, Q.; Qu, G.; Feng, Y.; Reetz, M.T. Utility of B-
Factors in Protein Science: Interpreting Rigidity, Flexibility, and
Internal Motion and Engineering Thermostability. Chem. Rev. 2019,
119 (3), 1626–1665.
(32) (a) Aqvist, J.; Medina, C.; Samuelsson, J.E. A New Method
for Predicting Binding Affinity in Computer-aided Drug Design.
Protein Eng. 1994, 7(3), 385-391. (b) Godschalk, F.; Genheden, S.;
Söderhjelm, P.; Ryde, U. Comparison of MM/GBSA Calculations
Based on Explicit and Implicit Solvent Simulations. Phys Chem Chem
Phys. 2013, 15(20), 7731-7739.
ACS Paragon Plus Environment

Page 15 of 15
Journal of the American Chemical Society
Insert Table of Contents artwork here
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
15
ACS Paragon Plus Environment