One-Pot Asymmetric Synthesis of an Aminodiol Intermediate of Florfenicol Using Engineered Transketolase and Transaminase

Florfenicol Using Engineered Transketolase and Transaminase

Research Article

One-Pot Asymmetric Synthesis of an Aminodiol Intermediate of

Qi Liu, Xinyue Xie, Mancheng Tang, Wentao Tao, Ting Shi, Yuanzhen Zhang, Tingting Huang,

Yilei Zhao, Zixin Deng, and Shuangjun Lin *

Cite This: ACS Catal. 2021, 11, 7477 −7488

Read Online

ACCESS

Metrics & More

Article Recommendations

sı Supporting Information

ABSTRACT: Florfenicol is the 3′-ﬂuoro derivative of thiamphenicol and has been widely used in veterinary medicine for its high

antibacterial activity and safety. However, the development of simpliﬁed and environmentally friendly catalytic methods for the

stereoselective production of ﬂorfenicol is a key challenge. Herein, we established a highly stereoselective enzymatic one-pot reaction

for the synthesis of an aminodiol intermediate of ﬂorfenicol bearing two stereocenters from industrial raw material 4-

(

methylsulfonyl) benzaldehyde by coupling transketolase (TK) and ω-transaminase (TA). The enantioselectivity of TK from E. coli

was converted from (S) (93% ee) to (R) (95% ee), and we also inverted the enantiopreference (E(S) = 9 to E(R) = 12) and ketone/

aldehyde substrate selectivity of TA ATA117 via structure-guided enzyme engineering. Docking calculations and molecular dynamics

simulations of the wild-type and mutant enzymes unveiled the molecular basis for enzymatic stereocontrol. Using the engineered TK

and TA, (1R,2R)-p-methylsulfonyl phenylserinol was biosynthesized with good yield (76%) and high stereoselectivity (96% de and

99% ee). Our work established an enzymatic synthetic route to (1R,2R)-p-methylsulfonyl phenylserinol, facilitating the

development of a chemoenzymatic method for producing ﬂorfenicol.

KEYWORDS: ﬂorfenicol, biocatalysis, TK, TA, stereoselectivity, one-pot reaction

INTRODUCTION

respiratory tract infections in cattle in 1995 and in swine in

■

000.

Florfenicol [D-threo-3-ﬂuoro-2-dichloroacetamido-1-(4-meth-

ylsulfonylphenyl)-1-propanol] is a synthetic, broad-spectrum

ﬂuorinated analogue of thiamphenicol (Figure 1A). Like

chloramphenicol and thiamphenicol, it shows potent activity

against many Gram-positive and Gram-negative bacteria.

Since adverse side eﬀects, in particular dose-independent

Because of the large demand for ﬂorfenicol (4000 tons per

year), there has been considerable attention on the develop-

6−15

ment of diﬀerent strategies for the synthesis of ﬂorfenicol

and the installation of the two adjacent stereocenters of cis-1,2-

11,14

amino alcohols.

However, chemical approaches required

multistep protection−deprotection, making the overall process

,7,9,10,15

irreversible aplastic anemia, have been observed in animals, the

cumbersome.

In addition, high toxicity of some

EU legislation banned chloramphenicol from use in food-

chemical reagents and wastewater pollution raise other

drawbacks in chemical synthesis. Moreover, some routes

suﬀered from the use of transition-metal (e.g., Ru) complexes

with chiral ligands that are toxic and diﬃcult to obtain and

producing animals. Meanwhile, the nitro group (−NO ) in

chloramphenicol is replaced by a methylsulfonyl group

(

−SO CH ) in ﬂorfenicol; therefore, ﬂorfenicol would not

cause dose-unrelated aplastic anemia. In addition, due to the

replacement of the hydroxyl group at the C-3′ position with a

ﬂuorine in ﬂorfenicol, chloramphenicol-resistant strains, in

which resistance is solely based on acetyltransferase activity,

are susceptible to ﬂorfenicol. Multiple studies indicated that

virtually all target bacteria isolated from respiratory tract

remove.

In view of the advantages of enzymatic catalysis, such as high

regio-, chemo-, and stereoselectivity and without the require-

Received: March 17, 2021

Revised: May 28, 2021

Published: June 9, 2021

−5

infections of cattle and pigs are susceptible to ﬂorfenicol.

With its outstanding eﬀectiveness and safety, ﬂorfenicol has

been licensed exclusively for use in veterinary medicine. In the

European Union, it has been approved for the treatment of

2021 American Chemical Society

477

ACS Catal. 2021, 11, 7477−7488

ACS Catalysis

Research Article

hydroxyketone formation by stereoselectively transferring a

ketol unit from polyhydroxylated ketose phosphates as a

“

natural” donor to an aldehyde acceptor. Several successful

applications of TK have already been demonstrated for the

32−36

production of α-hydroxyketones.

α-Hydroxyketones are

of particular value as ﬁne chemicals because of their utility as

building blocks for the production of larger molecules,

particularly for construction of heterocycles required in

pharmaceuticals. Akin to TKs, transaminases (TAs) have

also attracted considerable interest in biocatalysis, both as an

individual biocatalyst and as a part of multienzyme cascades for

38−42

the synthesis of various amines.

TA catalyzes the transfer

of an amino group between an amine donor (diﬀerent amino

acids and amines) and an amine acceptor (a ketone or

aldehyde). Of particular interest are ω-TAs that do not require

the presence of a carboxylic group in substrates, thereby

38,39

accepting a large variety of carbonyl substrates.

Here, we present an enzymatic one-pot two-step reaction for

the synthesis of a stereoselective aminodiol intermediate,

(

1R,2R)-p-methylsulfonyl phenylserinol (3a), from readily

available achiral 4-(methylsulfonyl)benzaldehyde (1) using a

TK and a ω-TA (Figure 1C and Scheme 1). 1 is the raw

Figure 1. Synthesis of chiral intermediates of ﬂorfenicol and

thiamphenicol. (A) Examples of chiral intermediates of ﬂorfenicol

and thiamphenicol. (B) Previous synthetic methods using enzymes.

Scheme 1. One-Pot Two-Step Cascade Reaction to (1R,2R)-

p-Methylsulfonyl Phenylserinol 3a, from the Industrial Raw

Material 4-(Methylsulfonyl)-benzaldehyde 1

(

C) Our enzymatic route based on transketolase and transaminase.

The aminodiol intermediate could be converted to the ﬁnal ﬂorfenicol

with fewer chemical steps than the reported intermediate.

17−21

ment of protection and deprotection steps,

enzymatic

catalysis could be a valuable alternative route for the chemical

synthesis of the chiral ﬂorfenicol intermediate. A ketoreductase

was reported to catalyze a stereoselective (2S,3R)-α-amino-β-

hydroxyl ester formation from the corresponding α-amino-β-

keto ester mediated by dynamic reductive kinetic resolution. In

this approach, the majority of the product must be assembled

in advance, with the enzyme mostly serving to set the

stereochemistry at the end (Figure 1B, entry 1). Besides, L-

threonine aldolase (LTA) from Pseudomonas putida was

reported to catalyze chiral L-threo-p-methylsulfonyl phenyl-

serine (L-threo-MSPS) formation, whereas the diastereoselec-

tivity of the reaction was relatively low, with 53%

material for the industrial production of ﬂorfenicol, suggesting

that our biocatalytic route has potential for industrial

applications. However, there existed three challenges in this

biocatalytic cascade reaction. First, TK family members are

predominantly S-selective. For aromatic aldehyde substrates,

only one work reported an (R)-selective TK mutant that

catalyzed the hydroxyketone formation toward benzaldehyde

diastereomeric excess (de) (Figure 1B, entry 2). The Zhu

group performed enzyme engineering of an LTA from

Pseudomonas sp. and obtained a mutant that catalyzed the

formation of L-threo-MSPS with 71% de (Figure 1B, entry 2).

The Wu group also carried out enzyme engineering of an LTA

with 82% ee, but with a low conversion (10%). Second, the

from Bacillus nealsonii and obtained a mutant that catalyzed the

hydroxyketone intermediate 2 formed by TK undergoes rapid

racemization, resulting in a decrease of optical purity. Third,

formation of L-threo-MSPS with >99% de (Figure 1B, entry

). Recently, the Lin group reported an L-threonine trans-

1 probably serves as a substrate for TK as well as for ω-TA,

and ω-TA may display a higher chemical reactivity toward

aldolase (LTTA) from Pseudomonas sp. Through conditional

optimization, L-threo-MSPS was obtained with 94.6% de

43,44

aldehyde 1 relative to ketone 2.

As a consequence, most of

25,26

catalyzed by PsLTTA (Figure 1B, entry 3).

In general, L-

1 would be converted to the benzylamine side-product in the

one-pot reaction. For addressing the ﬁrst challenge, we

reversed the enantioselectivity of TK from (S) (93% ee) to

(R) (95% ee) via structure-guided enzyme engineering. For

solving the second problem, we inverted the stereopreference

of ω-TA toward chiral hydroxyketone intermediate 2 (E(S) =

9 to E(R) = 12). For addressing the third challenge, we

reversed the aldehyde/ketone substrate selectivity of ω-TA.

Combining the engineered TK and TA, the aminodiol

intermediate (3a) of ﬂorfenicol was accessible with good

yield (76%) and excellent stereoselectivity (96% de and >99%

ee).

threo-MSPS could be biosynthesized with good stereo-

selectivity using LTA or LTTA. To eventually be converted

to ﬂorfenicol, L-threo-MSPS needs to be converted ﬁrst to

(

1R,2R)-p-methylsulfonyl phenylserinol (3a), which requests

two chemical steps comprising esteriﬁcation and reduction.

In order to establish a simpliﬁed synthetic route of ﬂorfenicol,

we attempted to achieve the biobased production of 3a.

Recently, transketolases (TKs) have emerged as attractive

biocatalysts for biotransformation of aliphatic and aromatic

8−36

aldehydes to various chemicals.

TK is a thiamine

pyrophosphate (ThDP)-dependent enzyme that catalyzes α-

478

ACS Catal. 2021, 11, 7477−7488

ACS Catalysis

RESULTS AND DISCUSSION

Choosing Suitable TKs and TAs for the Cascade

Reaction. In the ﬁrst step, β-hydroxypyruvate is decarboxy-

lated and subsequently ligated to 1 yielding 1,3-dihydroxy-1-

using 1 as the substrate. As shown in Figure 2 B, Ec TK1

Research Article

■

(

4-(methylsulfonyl)phenyl)propan-2-one (2). The reaction is

catalyzed by the ThDP-dependent TK. Initially, TKs for

arylated aldehyde substrates were searched from the literature.

Previous studies reported that the TK/D469T/R520Q/S385Y

mutant from E. coli (EcTK1) and TK/L382F/D470S mutant

from Geobacillus stearothermophilus (GsTK1) presented appeal-

ing biocatalytic activities for biotransformations involving

34,35,45

nonpolar aromatic aldehyde.

On the other hand, the

conﬁguration of the hydroxyketone products indicated that

GsTK1 was (S)-stereoselective. Therefore, EcTK1 was tested

Figure 3. Molecular docking modes of aldehyde substrate 1 in EcTK1

variants. Cartoon representation of 1 bound in (A) EcTK1 (cyan) and

(

B) EcTK1_YYH (green). (C) The enlarged binding mode of 1 in

the EcTK1 variant. 1 is shown in wheat. The key active pocket

residues are shown in cyan. (D) The enlarged binding mode of 1 in

the EcTK1_YYH variant. 1 is shown in wheat. The key active pocket

residues are shown in green. TPP is omitted for clarity. (E) The

surface of the binding pocket of 1 bound to the EcTK1 variant. (F)

The surface of the binding pocket of 1 bound to the EcTK1_YYH

variant. 1 is shown in wheat. The key residues are represented by the

yellow surface.

Figure 2. Enzyme engineering of EcTK1 for enantioselective

transketolase reaction. (A) Model reaction converting 1 to 2. With

combining ATA117_AC mutant enzyme, enantiomers 2a and 2b

were converted to diastereomers 3a and 3b, respectively. The ee of 2

was thus determined by investigating the concentrations of 3a and 3b.

(

B) Enantioselectivities (green bar) and enzyme activities (orange

In the second step, 2 was transformed to ﬁnal aminodiol

product 3 catalyzed by the pyridoxal phosphate (PLP)-

bar) of the evolved variants of EcTK1. Mutations in bold were added

in the corresponding round of mutagenesis and screening. Dotted

arrows denote a new round of site-speciﬁc mutagenesis.

47−50

dependent ω-TA. Six diﬀerent (R)-selective ω-TAs

were

screened using 2 as the substrate, and ATA117 (a homolog of

TA from Arthrobacter sp.) gave the most promising result

(

Figure S2 ). ATA117 is one of the most extensively used

catalyzed the conversion of 1 and β-hydroxypyruvate into 2

with acceptable activity (turnover frequency (TOF) ≥10/

min). Using AutoDock 4.2, 1 was docked into the substrate

biocatalysts in both industrial and academic laboratories. It

catalyzes the transamination of ketones with alanine or amines

to produce chiral amines with the (R)-conﬁguration.

The ﬁrst and most prominent example was published by Merck

& Co. and Codexis for the chiral synthesis of (R)-sitagliptin

employing a protein-engineered variant of ATA117.

48,51−56

binding site of the EcTK1 crystal structure (PDB ID: 5hht).

ThDP was converted to the enamine-ThDP intermediate

before docking according to the reaction mechanism of

transketolase. The docking analysis (Figure 3C) indicated

that the R520Q mutation created a space for the

methylsulfonyl moiety of 1. T469 formed a hydrogen bond

to the sulfonyl group, and H26 and H261 formed hydrogen

bonds to the carbonyl group of 1. We also observed that Y385

formed π−π stacking with the benzene ring of 1. As a result of

these mutations, 1 was properly oriented for catalysis (Figure

Determining the Enantioselectivity of EcTK1. After

identifying potential catalyst candidates for the synthesis of 3a,

we attempted to determine the enantioselectivity of EcTK1. In

order to use achiral high-performance liquid chromatography

(HPLC) to screen EcTK1 mutants for their stereoselectivity,

TK products (2a and 2b) were converted into diastereomers

3a and 3b using a ω-TA without (or with very low)

enantiopreference toward the chiral hydroxyketone 2. There-

C).

479

ACS Catal. 2021, 11, 7477−7488

ACS Catalysis

Research Article

fore, the enantiopreference of ATA117 was investigated using

racemic 2 as the substrate. However, ATA117 was determined

to display an obvious kinetic preference to the (S)-enantiomer

with an E value of 8.5 (Figure 4, Figure S3, and Table S3).

Therefore, we attempted to obtain ATA117 mutants that

showed low kinetic selectivity toward 2 through enzyme

engineering.

and start the transamination in the same reaction conditions as

the pretest. As shown in Figure 2 and Table S2, EcTK1

displayed S-selectivity with 93.3% ee, which was in agreement

with a previous study.

Reversing the Enantioselectivity of EcTK1 via Enzyme

Engineering. Since EcTK1 showed acceptable activity toward

, and its crystal structure (PDB ID: 5hht) has been

reported, we selected it as the starting template to reverse

the enantioselectivity through structure-guided enzyme en-

gineering. Using AutoDock 4.2, 1 was docked into the

substrate binding site of EcTK1. The most striking enzyme−

substrate interactions in this model seemed to be the hydrogen

bonds from H26 and H261 to the carbonyl group of 1 (Figure

C). Residue F434 lined on the opposite side. We supposed

that the orientation of the carbonyl group of 1 in the binding

pocket would directly aﬀect the stereoconﬁguration of the

hydroxyketone product. In order to reverse the stereo-

selectivity of TK, we devised a strategy dubbed “Break and

Build”, which is to break the existing hydrogen bonds from

H26 and H261 to the carbonyl group of 1 ﬁrst and then build a

new one between the aldehyde carbonyl group and F434 on

the opposite side via enzyme engineering. Therefore, H26 and

H261 were ﬁrst targeted for single site-saturation mutagenesis

and screening. The enzyme libraries were expressed and

screened as cell lysate catalysts in 24-well plates. To eliminate

the inﬂuence of the tautomerization between (R)- and (S)-2,

excess puriﬁed ATA117_AC enzyme was added to the reaction

mixture to ensure the instantaneous conversion of hydrox-

yketone products, (R)- and (S)-2, into the corresponding

Figure 4. Enzyme engineering of the kinetic enantioselectivity of

ATA117. (A) Model reaction converting racemic ketodiol 2 to

aminodiol diastereomers 3a and 3b. (B) Cumulative improvement in

the kinetic enantioselectivity of ATA117 resulting from the various

iterative evolution steps of the present study. Mutations in bold were

added in the corresponding round of mutagenesis and screening.

Dotted arrows denote a new round of site-speciﬁc mutagenesis. E =

(

1R,2R)- and (1S,2R)-aminodiol diastereomers 3a and 3b,

respectively. The site-saturation mutagenesis libraries of H26

and H261 were screened by HPLC. Among the mutants of the

ﬁrst generation, EcTK1/H26Y (EcTK1_Y) led to a reversion

of enantioselectivity to (R)-2a (4.8% ee) (Figure 2 and Table

S2 ). The switch in enantioselectivity may be attributed to the

breakage of the existing hydrogen bonds from H26 to the

carbonyl group of 1 because the introduction of a tyrosine with

large steric hindrance at this position might alter the

orientation of the carbonyl group of 1 in the active site. In

order to build a new hydrogen bond between aldehyde

carbonyl and F434, rational mutation of F434Y was introduced

into EcTK1_Y to yield the EcTK1/H26Y/F434Y (EcTK1_YY)

mutant. As anticipated, EcTK1_YY led to pronouncedly

inverted stereoselectivity in favor of (R)-2 with 90.8% ee

(Figure 2 and Table S2). In order to further increase the

enantioselectivity of TK, we attempted to ﬁnd more targeted

sites to be mutated. The docking results indicated that residues

L466, H461, Y385, T469, and Q520 also lined in the binding

pocket (Figure 3C). Among them, Y385, T469, and Q520

could not be mutated owing to their contribution in enzymatic

activity toward 1. Therefore, L466 and H461 were targeted for

iterative site-saturation mutagenesis based on the EcTK1_YY

variant. Gratifyingly, the introduction of L466H (EcT-

K1_YYH) and L466F (EcTK1_YYF) exhibited increased

(R)-enantioselectivity, with 95.2% ee and 91.6% ee,

respectively (Figure 2 and Table S2). Based on mutant

EcTK1_YYH, H461 was subjected to site-saturation muta-

genesis. However, no mutant was identiﬁed with enhanced

(R)-selectivity. A homology model of EcTK1_YYH was

constructed, and 1 was docked into the substrate binding

pocket. The docking analysis indicated that aldehyde carbonyl

turned to the opposite side and formed a new hydrogen bond

with the side chain of Y434.

58,59

ln[1 − C× (1 + de)]/ln[1 − C × (1 − de)].

A homology model of ATA117 was constructed using the

structure of ω-TA from Arthrobacter sp. (PDB ID: 3wwh) as a

guide. 2 was docked into its active site with the pyridoxamine-

′-phosphate (PMP) cofactor present using AutoDock 4.2.

V69 and F122 located adjacent to the ligand 2 (Figure S4) in

the small pocket were preferentially targeted for iterative site-

saturation mutagenesis and screening since they were reported

to be key sites for binding the substrate. The variant

ATA117/V69A/F122C (ATA117_AC) was obtained, which

showed low kinetic enantioselectivity with an E value of 1.4

(

Figure 4, Figure S3 , and Table S3 ). Therefore, the ATA117/

V69A/F122C (ATA117_AC) mutant enzyme was selected for

determining the enantioselectivity of TK. In order to test the

accuracy of the method for ee measurement using the

ATA117_AC mutant, we designed and performed a pretest.

The puriﬁed rac-2 ([α]²= 0) was totally converted to almost

equal amounts of 3a and 3b (3a/3b = 1:1.01) in 30 min at 25

C using high loading (200 μM) of puriﬁed ATA117_AC

(Figure S5B ), which manifested that the ratio of 3a and 3b

could accurately reﬂect the ratio of 2a and 2b using

ATA117_AC under these conditions. To determine the

enantioselectivity of TK, we performed transketolase reaction

in a short time (20 min) at 25 °C owing to the tautomerization

between 2a and 2b (Figure S5A ,C,D) while using high

loading (100 μM) of puriﬁed EcTK variant enzyme to ensure a

high value of conversion (>80%). Subsequently, EDTA and

puriﬁed ATA117_AC enzyme were added to the reaction

mixture to quench the transketolase reaction by chelating Mg

480

ACS Catal. 2021, 11, 7477−7488

ACS Catalysis

Research Article

Additionally, the benzene ring of 1 formed π−π stacking

interactions with Y26 and Y385, and the sulfonyl group of 1

formed hydrogen bonds with T469 and R91. H466 might also

form π−π stacking interactions with Y434 since the aromatic

rings of these two residues are coplanar (Figure 3D). These

interactions together with the steric eﬀect exerted by Y26

prompted the carbonyl group of 1 to adopt the inverted

orientation (Figure 3D,F), which demonstrated the molecular

basis of the switch in enantioselectivity from (S) to (R) of TK.

Since the poor solubility of 1 frustrated our ability to carry

out Michaelis−Menten kinetics analysis, we could approximate

the initial rate of hydroxyketone production under the reaction

conditions by measuring the conversion at a short reaction

residues F225, W192, G224, A284, and T282 were targeted for

iterative saturation mutation considering that these residues

were close to the chiral hydroxyl group of 2a in the docking

complex (Figure S6). The introduction of F225H based on the

ATA117_ACH variant (ATA117_ACHH) yielded increased

(R )-enantiopreference with an E value of 11.7 (Figure 4,

Figure S3, and Table S3) and improved ketone/aldehyde

substrate selectivity (3a/MPMA = 1.9:1) (Figure S8). Using

ATA117_ACHH as the template, the site-saturation mutation

of W192, G224, A284, and T282 gave no promising result. In

order to further investigate the kinetic enantiopreference of

ATA117_ACHH, we performed kinetic resolution using 5 mM

rac-2 as the substrate. As shown in Figure S9, 3a was yielded

with good diastereoselectivity (81/19 diastereomer ratio)

when the conversion reached 57% in 4 h.

time. Figure 2B shows that the TOF of EcTK1_YYH was

lower than that of the parent EcTK1 by 0.7-fold. The loss of

enzyme activity might rise from the H26Y mutation. The

introduction of tyrosine with large steric hindrance at this

position might narrow down the binding pocket of the enzyme

To unveil the molecular basis leading to the (R)-

enantiopreference of ATA117_ACHH, 2a and 2b were docked

into the active site of the enzyme with the PMP cofactor

present, respectively. The docked conformations of 2a and 2b

in the active site were picked, where the keto group of the

substrates faces the amino group of PMP allowing a

nucleophilic attack to form to external aldimine. The resulting

complexes (E:PMP + 2a and E:PMP + 2b) were energy-

minimized and equilibrated to serve as starting conformations

(

Figure 3F), thus weakening the binding of the 1 substrate.

Gratifyingly, the TOF of the EcTK1_YYF variant was higher

than that of EcTK1 by 1.7-fold (Figure 2B).

Enhancing the (R)-Enantiopreference and Ketone/

Aldehyde Substrate Selectivity of ATA117 via Enzyme

Engineering. As the biocatalyst for the production of 2a was

achieved, we then turned to engineer ω-TA for the production

of 3a with high stereoselectivity. Therefore, we attempted to

raise the (R)-enantiopreference of ATA117_AC via enzyme

engineering. We constructed a homology model of

ATA117_AC based on the crystal structure of ω-TA (PDB

ID: 3wwh). The 2a enantiomer was placed into the binding

pocket to achieve a plausible nucleophilic attack for identifying

steric clashes that could be alleviated by mutations. The

docking model suggested that the side chain of I157 is

positioned in proximity to the chiral hydroxyl of 2a (2.5 Å),

which is likely clashing with the chiral hydroxyl bound in the

active site (Figure S6 ). So, I157 was chosen for saturation

mutation based on the ATA117_AC variant. Except for

enhancing the (R)-enantiopreference of TA, we attempted to

concurrently promote the ketone/aldehyde substrate selectiv-

ity in order to reduce benzylamine side-product formation in

the one-pot reaction. A challenge in this one-pot reaction is the

fact that the starting material aldehyde serves as a substrate for

TK as well as for the ω-TAs, and ω-TAs in principle possess

60,61

for MD simulations (Figure S10).

According to the

mechanism of the transaminase-catalyzed reaction, the second

half reaction was initiated by the nucleophilic attack on the

60,61

bound substrate by the amino group of PMP.

Therefore,

the distances between the cofactor PMP and the substrate

ketone 2a (d ) and 2b (d ) were measured between the

PMP’s exocyclic nitrogen atom (PMP-NH ) and the carbonyl

carbon atom of the ketone (Figure 5B,C). Figure 5D and

Figure S10D show that 2a stayed in the active site throughout

the 70 ns of simulation at a distance of ≈3.5 Å to PMP. This

stood in stark contrast to the d , which ﬂuctuated markedly

and ranged between 10 and 40 Å (Figure 5D and Figure

S10D ). These ﬁndings suggested that the marginal activity for

2b seems to be relevant to the inability of the enzyme to

initiate the catalytic cycle due to an unfavorable nucleophilic

approach to the bound substrate. Furthermore, the elaborate

analysis of the angular orientation of the bound substrate was

performed, and the dihedral angle (θ_D_H) between the plane

harboring O , C , and N

and the plane harboring O , C ,

PMP

β α

43,44

higher chemical reactivity toward aldehydes than ketones.

and C was deﬁned for the nucleophilic approach (Figure

Therefore, the site-saturation mutagenesis library of I157 was

screened using 1 mM 1 and 2 mM rac-2 (equal to the mix of 1

mM S-2 and 1 mM R-2) as the substrates. ATA117_AC/

I157H (ATA117_ACH) was identiﬁed, and its selectivity was

re-examined using the puriﬁed enzyme. It is worth noting that

the ATA117_ACH variant displayed dramatically increased

5E). The ideal θ_D_Hvalue for the nucleophilic attack on the

carbonyl bond is known to be 90°. The θ_D_Hin the complex

of E:PMP + 2a was concentratedly distributed and approached

75°, which allowed the unhindered and proximal nucleophilic

attack. In contrast, θ_D_Hin the complex of E:PMP + 2b was

dispersed (Figure 5F). Therefore, both the distances of Cα-

N_P_M_Pand θ_D_Hendow the ATA117_ACHH variant with the

(R)-enantiopreference.

R )-enantiopreference with an E value of 8.8 (Figure 4, Figure

S3 , and Table S3), which outperformed the ATA117_AC by 6-

fold. The enhancement in (R)-stereopreference induced by the

I157H mutation could possibly be rationalized by the reduced

steric hindrance and the formation of a new hydrogen bond

between the histidine imidazolyl group and the hydroxyketone

moiety of 2a (Figure S7 ). Moreover, the ATA117_ACH

variant showed reversed ketone/aldehyde substrate selectivity

giving a 1.5:1 ratio of 3a/(4-(methylsulfonyl)phenyl)-

methanamine (MPMA, the amination product derived from

One-Pot Cascade Reaction for 3a Synthesis. After

obtaining the desired enzymes through engineering, the one-

pot simultaneous cascade reaction for the synthesis of 3a was

ﬁrst set up and optimized. Combining EcTK_YYH and

ATA117_ACHH, 3a was biosynthesized with 40% yield,

93% de, and >99% ee (Table 1, entry 2). The loss in the yield

mainly arose from the generation of the benzylamine by-

product. Considering that the starting substrate aldehyde 1 is

the substrate of TK and TA, TK with the high activity may

increase the yield of the ﬁnal product 3a. Therefore,

EcTK_YYH was replaced by EcTK_YYF that has much higher

) compared to the 0.15:1 ratio observed for wild-type

ATA117 (Figure S8). Thereafter, ATA117_ACH was used as

the template for the next round of mutagenesis. Amino acid

481

ACS Catal. 2021, 11, 7477−7488

ACS Catalysis

Research Article

activity (Figure 2B) in the reaction. Indeed, the yield and de

have been improved (Table 1, entry 3).

To further reduce the formation of the benzylamine by-

product and increase the yield and optical purity of 3a, we

performed one-pot cascade reaction in sequential mode

without isolation and puriﬁcation of the intermediate. First,

0 mM 1 was incubated with 60 μM puriﬁed EcTK1_YYH for

h. Subsequently, 50 μM puriﬁed ATA117_ACHH was added

to the reaction mixture and incubated for an extra 3 h.

Additionally, NADH and LDH were used to drive the

reversible equilibrium of transamination through removing

the co-product pyruvic acid (Scheme 1). NADH can be

regenerated using a well-established NADH-regenerating

system, glucose/glucose dehydrogenase (Scheme 1). An

overall yield of 71% and a de value of 96% (Table 1, entry 7

and Figures S12, S13 ) were achieved at the optimized

conditions. As a comparison, we also applied the EcTK1_YYH

and other ATA117 mutants to biosynthesize 3a. The results

showed that the ATA117 variant with higher (R)-enantiopre-

ference would produce 3a with a higher value of de (Table 1,

entries 4−7). As expected, all products had a high

enantiomeric purity of >99% (Table 1 and Figure S14 ).

We further developed a preparative scale (50 mL) of 3a

synthesis in one-pot two-step sequential cascade reaction. The

pretest displayed better results (higher yield and optical purity

of 3a) using EcTK_YYF with much higher activity instead of

EcTK_YYH, so EcTK_YYF was chosen for the preparative

scale. Applying 5 μmol of EcTK1_YYF and 5 μmol of

ATA117_ACHH, 92 mg (0.5 mmol) of 1 was converted to 93

mg of 3a (0.38 mmol, 76% conversion) with a de value of 96%

at 25 °C in 110 min, which corresponds to a space−time

conversion of 24 g/L/day. After the reaction, the protein in the

reaction mixture was precipitated using methanol. The

resulting supernatant was concentrated and puriﬁed by

reversed-phase chromatography with a PRC-ODS column

Figure 5. Statistics of molecular dynamics data. (A) Cartoon

representation comparison of the 2a enantiomer bound in

ATA117_ACHH (orange) and 2b enantiomer in ATA117_ACHH

(

blue). PMP is represented by the white stick. (B) Schematic diagram

of the distance between the cofactor PMP and the substrate 2. (C)

Modeled complex for E:PMP + 2a/2b after molecular dynamics

(

^M^D⁾^sⁱ^m^u^l^a^tⁱ^oⁿ^.^T^h^e^dⁱ^s^t^aⁿ^c^e^s^d²a between the carbonyl carbon

atoms of 2a and the cofactor’s amino group are marked as a yellow

dotted line. The distances d₂_bbetween the carbonyl carbon atoms of

(

20 × 250 mm, pore size 15 μm) from SHIMADZU, and an

b and the cofactor’s amino group are marked as a blue dotted line.

D) Distribution of d₂_a(yellow) and d₂_b(blue) of all conformations

in MD simulations. (E) Modeled complex for E:PMP + 2a/2b after

molecular dynamics (MD) simulation. The dihedral angle (θ_D_H_‑₂_a

between the plane harboring O , C , and N and the plane

0 mg isolated product was obtained as a colorless oil, which

(

corresponds to an isolated yield of 65% compared to the

starting material. The product was conﬁrmed by NMR (Figure

S30 ).

)

PMP

harboring O , C , and C of 2a is marked as a yellow dotted line. The

One-Pot Cascade Reaction for (1R,2R)-Phenylserinol

Derivative Synthesis. Encouraged by these results, we

further investigated the substrate scope of the cascade reaction

using various substituted benzaldehydes. The enzyme cascade

combining EcTK1_YYH and ATA117_ACHH in sequential

mode for the synthesis of (1R,2R)-2-amino-1-(4-

dihedral angle (θ_D_H_‑₂_b) between the plane harboring O , C , and

^NPMP and the plane harboring O , C , and C of 2b is marked as a

blue dotted line. (F) Distribution of θ_D_H_‑₂_a(yellow) and θ_D_H_‑₂_b(blue)

of all conformations in MD simulations.

Table 1. (1R,2R)-p-Methylsulfonyl Phenylserinol Produced by a Panel of Catalysts in One-Pot Cascade Reaction

entry

one-pot reaction mode

simultaneous mode

catalyst

time (h)

yield (%)

de (%)

ee (%)

EcTK1_YYH + ATA117

43.1

93.0

94.4

68.8

75.4

92.1

96.0

>99

EcTK1_YYH + ATA117_ACHH

EcTK1_YYF + ATA117_ACHH

EcTK1_YYH + ATA117

EcTK1_YYH + ATA117_AC

EcTK1_YYH + ATA117_ACH

EcTK1_YYH + ATA117_ACHH

sequential mode

1 + 3

Nonisolated yield [%] = ([product ]/[substrate_t₌₀]) × 100; [product ] was determined using the constructed calibration curves (Figure S11).

482

ACS Catal. 2021, 11, 7477−7488

ACS Catalysis

Research Article

Table 2. (1R,2R)-Phenylserinol Derivatives Produced by a Panel of Catalysts in Sequential One-Pot Cascade Reaction

entry

R group

catalyst

yield (%)

de (%)

ee (%)

CH₃

NO₂

EcTK1_YYH + ATA117_ACHH

EcTK1_YYF + ATA117_ACHH

97.0

96.3

95.3

94.7

96.3

96.6

94.5

>99

Nonisolated yield [%] = ([product ]/[substrate_t₌₀]) × 100; [product ] were determined using the constructed calibration curves (Figure S11).

ﬂuorophenyl)propane-1,3-diol ((1R,2R)-AFPPD) was ﬁrst set

up and optimized. The reaction gave moderate yield (33%) but

very high stereoselectivity with 97% de and >99% ee (Table 2,

entry 1). A potential bottleneck in the transformation of 4-

mutagenesis and rational mutation. Moreover, the kinetic

enantiopreference of TA was also successfully inverted via four

rounds of site-saturation mutagenesis. The evolved

ATA117_ACHH mutant showed higher activity toward

ketone relative to the aldehyde substrate, which is a rare

event for ω-TA. Combining the engineered EcTK1_YYH or

EcTK1_YYF and ATA117_ACHH, 3a was accessible with

71−76% yield and 96% de in sequential one-pot cascade

reaction. Overall, our study established an eﬃcient biocatalytic

route to 3a, a chiral intermediate of ﬂorfenicol, which required

fewer subsequent steps to produce ﬂorfenicol than the

reported intermediates.

(ﬂuoro)benzaldehyde to AFPPD might be associated with the

low activity of EcTK1_YYH toward 4-(ﬂuoro)benzaldehyde.

Considering that EcTK1_YYF showed 2−3-fold higher

enzyme activity than EcTK1_YYH, EcTK1_YYF was thus

combined to synthesize (1R,2R)-AFPPD. As expected, the

cascade using EcTK1_YYF gave a signiﬁcantly increased yield

(

80%) and an equivalent stereoselectivity (96% de and >99%

ee) (Table 2, entry 2 and Figures S15, S16, S22, S28).

Moreover, other para-substituted benzaldehydes with both

electron-donating and electron-withdrawing groups were well

tolerated by the cascade reaction combining EcTK1_YYF and

ATA117_ACHH, enabling the highly diastereo- and enantio-

selective (95−97% de and >99% ee) synthesis of the

corresponding aminodiol products with moderate to good

yields (51 −72%) within 6 h (Table 2, entries 3−7 and Figures

S15 and S17 −S28 ). Notably, the para-substituted phenyl-

serinols with ﬂuorine, chlorine, and bromine groups furnished

in our cascade reaction possessed higher conversion and higher

de values than the reported corresponding para-substituted

EXPERIMENTAL SECTION

■

General Methods and Materials. The (1R,2R)-p-

methylsulfonyl phenylserinol (3a) standard was bought from

Huangshi Yongxin Biotechnology Co. (Huangshi, Hubei,

China). (1S,2S)-2-Amino-1-phenylpropane-1,3-diol and

(

1S,2S)-2-amino-1-(4-nitrophenyl)propane-1,3-diol were

bought from Sigma-Aldrich Co. (Shanghai, China). (1R,2R)-

-Amino-1-phenylpropane-1,3-diol was bought from Jiangsu

Aikang Biomedicine Co. (Nanjing, Jiangsu, China). (1R,2R)-2-

Amino-1-(4-nitrophenyl)propane-1,3-diol was bought from

TCI Co. (Shanghai, China). (1R,2R)-2-Amino-1-(4-

bromophenyl)propane-1,3-diol was bought from Nanjing

Zhicui Biotechnology Co. (Nanjing, Jiangsu, China). (Eryth-

ro)- and (threo)-2-amino-1-(4-ﬂuorophenyl)propane-1,3-diol,

24,26,63

intermediates of ﬂorfenicol,

which suggested that our

cascade route exhibited a broad substrate scope.

Although benzaldehyde derivatives bear diﬀerent para-

substitutions, they gave all products with absolute enantiose-

lectivity (>99% ee) (Tables 1 and 2), which suggested that the

evolved mutations (V69A/F122C/I157H/F225H) of ATA117

responsible for manipulating the enantiopreference toward the

chiral hydroxyketone do not impair the innate (R)-stereo-

speciﬁcity of the enzyme. For the D-amino acid amino-

transferase (Fold Type IV) to which ATA117 belongs, the

active site lysine (e.g., lys188 in ATA117) is positioned on the

re face of the quinonoid intermediate and the si face is solvent-

exposed, and therefore, the nucleophilic attack by amino donor

PMP occurs at the re face of the substrate to exclusively give

(

erythro)- and (threo)-2-amino-1-(4-chlorophenyl)propane-

,3-diol, and (erythro)- and (threo)-2-amino-1-(p-tolyl)-

propane-1,3-diol were bought from Shanghai Kaishu Chemical

Technology Co. (Shanghai, China). Lithium hydroxypyruvate

(

LiHPA) was purchased from Sigma-Aldrich Co. (Shanghai,

China). Thiamine diphosphate (ThDP), pyridoxal-5′-phos-

phate hydrate (PLP), glucose, and glucose dehydrogenase

(

GDH) were purchased from Aladdin Co. (Shanghai, China).

The reduced form of nicotinamide-adenine dinucleotide

NADH) and D-Ala were purchased from TCI Co. (Shanghai,

(

the (R)-conﬁguration.

China). Lactate dehydrogenase (LDH) was purchased from

Sinopharm (Shanghai, China). The materials used for culture

media, including peptone, yeast extract, and agar, were

purchased from OXOID (Shanghai, China). The Phanta

Max Super-Fidelity DNA Polymerase was obtained from

Vazyme Co. (Shanghai, China). The restriction enzyme Dpn I

was bought from Thermo Scientiﬁc (Shanghai, China). The

oligonucleotides were synthesized, and DNA sequencing was

conducted by Shanghai BioSune (Shanghai, China). The PCR

extraction kit and the plasmid mini kit used for extracting

CONCLUSIONS

■

In summary, in order to achieve the biosynthesis of the chiral

aminodiol intermediate of ﬂorfenicol, (1R,2R)-p-methylsulfon-

yl phenylserinol 3a, multiple rounds of structure-guided

iterative site-saturation mutagenesis of TK and TA were

conducted to reverse the enzymatic stereoselectivity. This

study ﬁrstly obtained a highly (R)-selective TK enzyme toward

aromatic aldehyde through three rounds of site-saturation

483

ACS Catal. 2021, 11, 7477−7488

ACS Catalysis

Research Article

plasmids were bought from Vazyme Co. (Shanghai, China).

The lysogeny broth (LB) medium contained 10 g/L tryptone,

static interactions were calculated using the Particle Mesh

Ewald (PME) method. The CPPTRAJ tool implemented

in AMBER was used to conduct the structural analyses of the

molecular dynamics trajectory, such as the distance and

dihedral of atoms in the ligand molecule, root-mean-square

deviation (RMSD), and cluster analysis. The structure ﬁgures

were prepared using PyMOL.

g/L yeast extract, and 10 g/L NaCl. The racemic 1,3-

dihydroxy-1-(4-(methylsulfonyl)-phenyl)-propan-2-one was

prepared by the transketolase reaction of 4-(methylsulfonyl)-

benzaldehyde with hydroxypyruvate catalyzed by EcTK1. The

NMR spectra were measured on a Bruker Avance 600 using

MeOD as the solvent.

Construction of TK and TA Mutant Libraries. Plasmid

pET28a (Novagen) was used as a cloning and expression

vector for TK variants, and plasmid pRSFduet (Novagen) was

used as a cloning and expression vector for TA variants

described in this paper. The saturation mutagenesis libraries of

H26, L466, V69, F122, I157, F225, and etc. were generated

utilizing degenerate oligos (i.e., NNK) to incorporate

mutations at the targeted position. On the basis of the

EcTK1 and ATA117 or related mutant sequence, appropriate

primers were designed. The mutant genes were then

constructed by using PCR ampliﬁcation methods with a

Phanta Max Super-Fidelity DNA Polymerase. The PCR

products were analyzed on agarose gel by electrophoresis.

The PCR products were digested with the Dpn I restriction

enzyme at 37 °C for more than 4 h to remove the template

DNA. After digestion, the products were directly transformed

to commercial E. coli BL21 (DE3) by heat shock. Diluted

transformation mixtures (about 1/7) were spread on agar

plates containing 50 μg/mL kanamycin and grown overnight at

37 °C. The individual clones with the expected mutation

evaluated by DNA sequencing were picked.

Screening of TK and AT Mutant Libraries. E. coli BL21

(DE3) cells carrying the recombinant plasmid were cultivated

in 1 mL of LB medium containing kanamycin (50 μg/mL)

using a 24 shallow-well plate at 37 °C and 220 rpm for 12 h. 40

μL of the overnight culture was transferred to a 24 deep-well

plate with a fresh lysogeny broth medium (4 mL/well) and was

cultivated under the same conditions. When the OD₆₀₀

reached 0.6−0.8, isopropyl β-D-thiogalactoside (IPTG) was

added to a ﬁnal concentration of 0.15 mM. After continuing

cultivation at 20 °C for 18 h, the cell pellets were harvested

and washed with 1 mL of Tris−HCl Buﬀer (50 mM, 150 mM

NaCl, pH 7.5) and centrifuged for 20 min with 4000 rpm at 4

°C. The precipitate was frozen at −80 °C and thawed once.

The cells were resuspended in 250 μL of the same buﬀer and

lysed by ultrasonication (2 s on and 4 s oﬀ; 30 cycles). Cell

debris was removed by centrifugation at 4 °C.

Molecular Docking. The substrates 2 and 2a were docked

into the binding pocket of ATA117 and the homology model

of ATA117 mutants (ATA117_AC and ATA117_ACH),

respectively, using the position of PLP in the bound ligand

of 3wwh as a guide with AutoDock 4.2. The catalytic lysine

(

K188) was deprotonated, and the PLP was converted to PMP

before docking. Postdocking, the conformations obtained were

ﬁltered to derive a plausible attack conformation.

The substrate 1 was docked into the binding pocket of the

EcTK1 structure (PDB ID: 5hht) and the homology model of

EcTK1_YYH. The EcTK1 structure was stripped of all

crystallographic waters. ThDP was converted to the enamine-

ThDP intermediate before docking. The explorable space for

docking was deﬁned as a cube 15 Å in length centered at

3.489, 15.398, and 5.953 including the entire active-site and

omitting the surface hydrophobic pockets. Resulting poses

were analyzed and checked for hydrogen bonding in the

PyMOL Molecular Graphics System. Only the most energeti-

cally favorable pose-clusters were retained. Among them, poses

that were observed to be potentially “catalytically productive”

were selected. Catalytically productive was taken to mean

orientations of the benzene ring that point the reactive

aldehyde group close to the enamine-ThDP intermediate.

MD Simulations. The ligands 2a and 2b were docked into

the homology model of the ATA117_ACHH mutant using

AutoDock 4.2. First, optimization was accomplished using the

Gaussian09 program at the level of b3lyp/6-31 g(d) followed

by calculation of the electrostatic surface potential (ESP)

charge. Subsequently, restrained electrostatic potential (RESP)

charge ﬁtting was applied using the Multiwfn program. The

parameter preparation for PMP and the substrate molecules 2a

and 2b was performed by using the Antechamber package.

Finally, the GPU version of the PMEMD engine integrated

with the Amber 16 package and FF14SB force ﬁeld was used

for the MD simulations. The enzyme−ligand complex was

placed in a solvent box of the TIP3P water model with a shell

of 10 Å thick.

For screening of TK mutant libraries, puriﬁed ATA117_AC

enzyme, which showed low kinetic selectivity toward the chiral

hydroxyketone, was combined to ensure the instantaneous

conversion of hydroxyketone intermediate, which was apt to

The entire systems were ﬁrst subjected to energy

minimization using the 5000-step steepest descent followed

by the conjugate gradient algorithm for 5000 steps. The system

was then subjected to 1400 ps controlled heating using a

undergo racemization. By combining ATA117_AC, (S)- and

(R)-hydroxyketone were converted to the corresponding

(1R,2R)- and (1S,2R)-aminodiol diastereomers 3a and 3b,

respectively. The 100 μL screening system contained 40 μL of

supernatant, 50 mM Tris−HCl buﬀer, 5 mM 1 substrate, 5

−

Langevin thermostat with a collision frequency of 10 ps ,

which requires the temperature to be increased from 0 to 300

K during the ﬁrst 1000 ps. The protein and the ligand

molecules were heated with a harmonic potential restraint of 5

−

kcal·mol ·Å . The system was then equilibrated at 300 K

using the NPT ensemble for 200 ps to facilitate adjustment of

the system to the appropriate density and dimension. Finally,

the molecular dynamics simulations were performed without

any restraint under NPT conditions at 300 K for 100 ns with a

equiv of LiHPA, 2.4 mM ThDP, 9 mM MgCl , 1 mM PLP,

200 mM D-Ala, 5 equiv of NADH, 45 U/mL LDH, and 20 μM

puriﬁed ATA117_AC variant enzyme. Aminodiol diastereomer

products were measured by HPLC using a Super-C18 column.

For screening of TA mutant libraries, equivalent 2a, 2b, and

1 were used as the substrates in order to concurrently promote

the (R)-enantiopreference and ketone/aldehyde substrate

selectivity of TA. Therefore, the 100 μL screening system

contained 40 μL of supernatant, 50 mM Tris−HCl buﬀer, 2

mM rac-2, 1 mM 1, 200 mM D-Ala, and 1 mM PLP. Aminodiol

step time of 2 fs. The Berendsen barostat was used to

maintain the pressure at 1 bar during the equilibration and

production MD phase. The SHAKE algorithm was used to

constrain the lengths of all the bonds involving hydrogen

atoms during the whole procedure. The long-range electro-

484

ACS Catal. 2021, 11, 7477−7488

ACS Catalysis

https://pubs.acs.org/doi/10.1021/acscatal.1c01229.

Research Article

and benzylamine products were measured by HPLC using a

Super-C18 column.

min at 25 °C. 1 was added in 5.6 mM batches every 10 min

considering the insolubility of the aldehyde substrate. The

second step was initiated by adding 20 mL of transamination

reaction mixture. The total 50 mL reaction system contained

100 μM puriﬁed ATA117_ACHH enzyme, 200 mM D-Ala, 2

equiv of NADH, 90 U/mL LDH, 200 mM glucose, and 30 U/

mL GDH. The reaction mixture was incubated at 25 °C for an

additional 75 min.

Measurement of Enantioselectivity for EcTK1 Mutant

Enzymes. The 100 μL reaction system contained 50 mM

Tris−HCl buﬀer (pH 7), 5 mM 1, 25 mM LiHPA, 9 mM

MgCl , and 4.8 mM ThDP. The reaction was initiated by

addition of 100 μM TK enzyme. After reacting 20 min at 25

C, the reaction was quenched by addition of 18 mM EDTA

for chelating Mg that was indispensable for TK activity.

Sequential One-Pot Cascade Reaction to (1R,2R)-

Phenylserinol Derivatives. For (1R,2R)-p-ﬂuorine phenyl-

serinol, p-chlorine phenylserinol, p-methyl phenylserinol, and

phenylserinol biosynthesis, the 100 μL reaction system

contained 100 mM Tris−HCl buﬀer (pH 7.5), 5 mM

Subsequently, 20 μL of reaction mixture was added to a new

00 μL reaction system that contained 200 mM D-Ala, 2 mM

PLP, 20 mM NADH, 90 U/mL LDH, and 200 μM puriﬁed

ATA117_AC enzyme. Thus, chiral hydroxyketone enan-

tiomers yielded at the ﬁrst step were fully converted to

aminodiol diastereomers 3a and 3b in 30 min at 25 °C. The

aminodiol products 3a and 3b were measured by HPLC using

a Super-C18 column.

benzaldehyde derivatives, 15 mM LiHPA, 18 mM MgCl , 4.8

mM ThDP, 50 μM puriﬁed enzyme EcTK1_YYF, 200 mM D-

Ala, 2 mM PLP, 200 mM glucose, 2 equiv of NADH, 90 U/mL

LDH, 30 U/mL GDH, and 50 μM puriﬁed enzyme

ATA117_ACHH. For the ﬁrst transketolase reaction step, 1,

Measurement of E for the ATA117 Mutant Enzyme.

The E values of the puriﬁed ATA117 variants toward the

transamination of rac-2 were determined referring to previous

LiHPA, MgCl , ThDP, and EcTK1_YYF were added into the

reaction vessel and incubated at 25 °C for 1.0, 1.5, 2.0, 2.5, and

3.0 h. For the second transamination step, D-Ala, PLP, glucose,

NADH, LDH, GDH, and ATA117_ACHH were added to the

reaction mixture and incubated at 25 °C for an additional 3 h.

For (1R,2R)-p-bromine phenylserinol biosynthesis, the 100

μL reaction system contained 100 mM Tris−HCl buﬀer (pH

7.5), 6% DMSO, 5 mM 4-bromobenzaldehyde, 15 mM

8,59

research.

contained 50 mM Tris−HCl buﬀer ( pH 7.5), 2 mM rac-2,

00 mM D-Ala, 2 mM PLP, and 5−40 μM ATA117 variants

E was measured in a 100 μL reaction system that

depending on the activity. The concentrations of puriﬁed

enzymes ATA117, ATA117_A, ATA117_AC, ATA117_ACH,

and ATA117_ACHH were set to be 5, 20, 20, 40, and 40 μM,

respectively. After reacting for 1 h at 30 °C, the reaction was

quenched by addition of 200 μL of methanol. Samples were

measured by HPLC using a Super-C18 column.

Simultaneous One-Pot Cascade Reaction to (1R,2R)-

p-Methylsulfonyl Phenylserinol 3a. The 100 μL reaction

system employing EcTK1_YYH and ATA117_ACHH (or

ATA117) contained 50 mM HEPES buﬀer (pH 8), 5 mM 1,

LiHPA, 18 mM MgCl , 4.8 mM ThDP, 50 μM puriﬁed

enzyme EcTK1_YYF, 200 mM D-Ala, 2 mM PLP, 200 mM

glucose, 2 equiv of NADH, 90 U/mL LDH, 30 U/mL GDH,

and 50 μM puriﬁed enzyme ATA117_ACHH. For the ﬁrst

transketolase reaction step, 4-bromobenzaldehyde, LiHPA,

MgCl , ThDP, and EcTK1_YYF were added into the reaction

vessel and incubated at 25 °C. Among them, the 4-

bromobenzaldehyde substrate was added in two batches. The

second half of the 4-bromobenzaldehyde substrate was added

after 1 h. The reaction mixture was then incubated for 1.0, 1.5,

and 2.0 h. For the second transamination step, D-Ala, PLP,

glucose, NADH, LDH, GDH, and ATA117_ACHH were

added to the reaction mixture and incubated at 25 °C for an

additional 3 h.

5 mM LiHPA, 18 mM MgCl , 4.8 mM ThDP, 60 μM puriﬁed

enzyme EcTK1_YYH, 50 mM D-Ala, 2 mM PLP, 3 equiv of

NADH, 90 U/mL LDH, and 30 μM puriﬁed enzyme

ATA117_ACHH (or ATA117). The reaction mixture was

incubated at 25 °C for 4 h.

The 100 μL reaction system employing EcTK1_YYF and

ATA117_ACHH contained 50 mM HEPES buﬀer (pH 8), 5

For (1R,2R)-p-nitro phenylserinol biosynthesis, the 100 μL

reaction system contained 100 mM Tris−HCl buﬀer (pH 7.5),

mM 1, 25 mM LiHPA, 18 mM MgCl , 4.8 mM ThDP, 60 μM

puriﬁed enzyme EcTK1_YYF, 50 mM D-Ala, 2 mM PLP, 3

equiv of NADH, 45 U/mL LDH, and 40 μM puriﬁed enzyme

ATA117_ACHH. The reaction mixture was incubated at 25

% DMSO, 5 mM 4-nitrobenzaldehyde, 20 mM LiHPA, 18

mM MgCl , 4.8 mM ThDP, 60 μM puriﬁed enzyme

EcTK1_YYF, 200 mM D-Ala, 2 mM PLP, 200 mM glucose,

C for 4 h.

equiv of NADH, 90 U/mL LDH, 30 U/mL GDH, and 60

Sequential One-Pot Cascade Reaction to (1R,2R)-p-

μM puriﬁed enzyme ATA117_ACHH. For the ﬁrst trans-

Methylsulfonyl Phenylserinol 3a. The 100 μL reaction

system contained 100 mM Tris−HCl buﬀer (pH 7.5), 5 mM

^k^e^t^o^l^a^s^e^r^e^a^c^tⁱ^oⁿ^s^t^e^p^,⁴^-ⁿⁱ^t^r^o^b^eⁿ^z^a^l^d^e^h^y^d^e^,^Lⁱ^H^P^A^,^M^g^C^l2^,

ThDP, and EcTK1_YYF were added into the reaction vessel

and incubated at 25 °C for 2.7 h. For the second

transamination step, D-Ala, PLP, glucose, NADH, LDH,

GDH, and ATA117_ACHH were added to the reaction

mixture and incubated at 25 °C for an extra 2, 3, 4, and 5 h.

, 15 mM LiHPA, 18 mM MgCl , 4.8 mM ThDP, 30 μM

puriﬁed enzyme EcTK1_YYH, 200 mM D-Ala, 2 mM PLP, 200

mM glucose, 2 equiv of NADH, 90 U/mL LDH, 30 U/mL

GDH, and 50 μM puriﬁed enzyme ATA117 variants. For the

ﬁrst transketolase reaction step, 1, LiHPA, Tris−HCl, MgCl ,

ThDP, and EcTK1_YYH were added into the reaction vessel

and incubated at 25 °C for 1 h. For the second transamination

step, D-Ala, PLP, glucose, NADH, LDH, GDH, and ATA117

variants were added to the reaction mixture and incubated at

ASSOCIATED CONTENT

■

sı

Supporting Information

The Supporting Information is available free of charge at

5 °C for an additional 1−4 h.

Preparative Scale of (1R,2R)-p-Methylsulfonyl Phe-

Experimental section, primers used for site-directed

saturated mutagenesis, transketolase reaction data

presented in Figure 2 of the main text, enantioselective

transamination of rac-2 with ATA117 variants, HPLC

nylserinol 3a Synthesis. A preparative scale (50 mL) of the

one-pot two-step cascade to 3a has been developed. In the ﬁrst

step, 16.7 mM 1 was converted with 3 equiv of LiHPA and 167

μM puriﬁed EcTK1_YYF in a 30 mL reaction system for 35

485

ACS Catal. 2021, 11, 7477−7488

ACS Catalysis

Complete contact information is available at:

Research Article

Biotechnology, Shanghai Jiao Tong University, Shanghai

00240, China

analytical methods of aminodiol derivatives, HPLC

analytical methods for ee measurement of aminodiol

derivatives, LC−MS, screening, HPLC, molecular

docking results, time course data, kinetic resolution

data, mass data, NMR data, and sequence information

https://pubs.acs.org/10.1021/acscatal.1c01229

(PDF)

Notes

The authors declare no competing ﬁnancial interest.

AUTHOR INFORMATION

■

ACKNOWLEDGMENTS

■

Corresponding Author

We thank Mrs. J. Wu at the Instrumental Analysis Center of

SJTU for performing the NMR experiments. This work was

ﬁnancially supported by the National Key Research and

Development Program of China (grant no. 2018YFA0901900)

and the National Science Foundation of China (21632007 and

Shuangjun Lin − Frontiers Science Center for Transformative

Molecules, Shanghai Jiao Tong University, Shanghai 200240,

China; State Key Laboratory of Microbial Metabolism, Joint

International Research Laboratory on Metabolic &

Developmental Sciences, School of Life Sciences &

Biotechnology, Shanghai Jiao Tong University, Shanghai

1425001).

00240, China; orcid.org/0000-0001-9406-9233;

Email: linsj@sjtu.edu.cn

REFERENCES

■

1) Kehrenberg, C.; Schwarz, S. Distribution of Florfenicol

Authors

Resistance Genes fexA and cfr among Chloramphenicol-Resistant

Staphylococcus Isolates. Antimicrob. Agents Chemother. 2006, 50,

1156−1163.

Qi Liu − State Key Laboratory of Microbial Metabolism, Joint

International Research Laboratory on Metabolic &

Developmental Sciences, School of Life Sciences &

(2) Schwarz, S.; Kehrenberg, C.; Doublet, B.; Cloeckaert, A.

Molecular basis of bacterial resistance to chloramphenicol and

florfenicol. FEMS Microbiol. Rev. 2004, 28, 519−542.

Biotechnology, Shanghai Jiao Tong University, Shanghai

00240, China

(

3) Kehrenberg, C.; Mumme, J.; Wallmann, R.; Verspohl, J.; Tegeler,

Xinyue Xie − State Key Laboratory of Microbial Metabolism,

Joint International Research Laboratory on Metabolic &

Developmental Sciences, School of Life Sciences &

R.; Kuhn, T.; Schwarz, S. Monitoring of florfenicol susceptibility

among bovine and porcine respiratory tract pathogens collected in

Germany during the years 2002 and 2003. J. Antimicrob. Chemother.

004, 54, 572−574.

4) Schwarz, S.; Kehrenberg, C.; Salmon, S. A.; Watts, J. L. In vitro

activities of spectinomycin and comparator agents against Pasteurella

multocida and Mannheimia haemolytica from respiratory tract

infections of cattle. J. Antimicrob. Chemother. 2004 , 53 , 379 −382.

(5) Priebe, S.; Schwarz, S. In vitro activities of florfenicol against

bovine and porcine respiratory tract pathogens. Antimicrob. Agents

Chemother. 2003, 47, 2703−2705.

6) Schumacher, D. P.; Clark, J. E.; Murphy, B. L.; Fischer, P. A. An

Efficient Synthesis of Florfenicol. J. Org. Chem. 1990, 55, 5291−5294.

7) Wu, G.; Schumacher, D. P.; Tormos, W.; Clark, J. E.; Murphy, B.

Biotechnology, Shanghai Jiao Tong University, Shanghai

Biotechnology, Shanghai Jiao Tong University, Shanghai

00240, China

Mancheng Tang − State Key Laboratory of Microbial

Metabolism, Joint International Research Laboratory on

Metabolic & Developmental Sciences, School of Life Sciences

00240, China; orcid.org/0000-0001-9418-3239

Wentao Tao − State Key Laboratory of Microbial Metabolism,

Joint International Research Laboratory on Metabolic &

Developmental Sciences, School of Life Sciences &

L. An Improved Industrial Synthesis of Florfenicol plus an

Biotechnology, Shanghai Jiao Tong University, Shanghai

Enantioselective Total Synthesis of Thiamphenicol and Florfenicol.

J. Org. Chem. 1997, 62, 2996−2998.

00240, China

Ting Shi − State Key Laboratory of Microbial Metabolism,

Joint International Research Laboratory on Metabolic &

Developmental Sciences, School of Life Sciences &

(8) Bhaskar, G.; Satish Kumar, V.; Venkateswara Rao, B. A short

stereoselective synthesis of (−)-chloramphenicol and (+)-thiamphe-

nicol. Tetrahedron: Asymmetry 2004, 15, 1279 −1283.

Biotechnology, Shanghai Jiao Tong University, Shanghai

9) George, S.; Narina, S. V.; Sudalai, A. A short enantioselective

00240, China; orcid.org/0000-0003-3921-4412

synthesis of (−)-chloramphenicol and (+)-thiamphenicol using

Yuanzhen Zhang − State Key Laboratory of Microbial

Metabolism, Joint International Research Laboratory on

Metabolic & Developmental Sciences, School of Life Sciences

tethered aminohydroxylation. Tetrahedron 2006 , 62 , 10202 −10207.

10) Lu, W.; Chen, P.; Lin, G. New stereoselective synthesis of

thiamphenicol and florfenicol from enantiomerically pure cyanohy-

Biotechnology, Shanghai Jiao Tong University, Shanghai

drin: a chemo-enzymatic approach. Tetrahedron 2008, 64, 7822−

00240, China

11) Li, F.; Wang, Z.-H.; Zhao, L.; Chen, F.-E. A Facile and Efficient

827.

Tingting Huang − State Key Laboratory of Microbial

Metabolism, Joint International Research Laboratory on

Metabolic & Developmental Sciences, School of Life Sciences

Asymmetric Synthesis of Florfenicol. Synlett 2011, 2011, 2883−2885.

(12) Li, F.; Wang, Z.-H.; Zhao, L.; Xiong, F.-J.; He, Q.-Q.; Chen, F.-

E. An efficient enantioselective synthesis of florfenicol via a vanadium-

catalyzed asymmetric epoxidation. Tetrahedron: Asymmetry 2011, 22,

Biotechnology, Shanghai Jiao Tong University, Shanghai

00240, China

13) Wang, Z.; Li, F.; Zhao, L.; He, Q.; Chen, F.; Zheng, C. An

337−1341.

Yilei Zhao − State Key Laboratory of Microbial Metabolism,

Joint International Research Laboratory on Metabolic &

Developmental Sciences, School of Life Sciences &

efficient enantioselective synthesis of florfenicol via asymmetric

aziridination. Tetrahedron 2011, 67, 9199−9203.

14) Wang, Z.-H.; Zheng, C.; Li, F.; Zhao, L.; Chen, F.-E.; He, Q.-Q.

An Efficient Enantioselective Synthesis of Florfenicol Based on

Sharpless Asymmetric Dihydroxylation. Synthesis 2012, 44, 699−704.

(15) Wang, X.; Xu, L.; Yan, L.; Wang, H.; Han, S.; Wu, Y.; Chen, F.

Catalytic asymmetric transfer hydrogenation/dynamic kinetic reso-

Biotechnology, Shanghai Jiao Tong University, Shanghai

00240, China; orcid.org/0000-0003-4687-7847

Zixin Deng − State Key Laboratory of Microbial Metabolism,

Joint International Research Laboratory on Metabolic &

Developmental Sciences, School of Life Sciences &

486

ACS Catal. 2021, 11, 7477−7488

ACS Catalysis

16) Zou, J.; Ni, G.; Tang, J.; Yu, J.; Jiang, L.; Ju, D.; Zhang, F.;

Research Article

lution: an efficient synthesis of florfenicol. Tetrahedron 2016, 72,

787−1793.

Chen, S. Asymmetric Synthesis of Florfenicol by Dynamic Reductive

Kinetic Resolution with Ketoreductases. Eur. J. Org. Chem. 2018,

(34) Saravanan, T.; Reif, M.-L.; Yi, D.; Lorillier

e, M.; Charmantray,

(

F.; Hecquet, L.; Fessner, W.-D. Engineering a thermostable

transketolase for arylated substrates. Green Chem. 2017, 19, 481−489.

(35) Affaticati, P. E.; Dai, S.-B.; Payongsri, P.; Hailes, H. C.;

Tittmann, K.; Dalby, P. A. Structural Analysis of an Evolved

Transketolase Reveals Divergent Binding Modes. Sci. Rep. 2016, 6,

35716.

18) Wu, S.; Snajdrova, R.; Moore, J. C.; Baldenius, K.; Bornscheuer,

018, 5044−5053.

17) Fryszkowska, A.; Devine, P. N. Biocatalysis in drug discovery

and development. Curr. Opin. Chem. Biol. 2020, 55, 151−160.

(36) Galman, J. L.; Steadman, D.; Bacon, S.; Morris, P.; Smith, M. E.

B.; Ward, J. M.; Dalby, P. A.; Hailes, H. C. α ,α ′-Dihydroxyketone

formation using aromatic and heteroaromatic aldehydes with evolved

transketolase enzymes. Chem. Commun. 2010 , 46 , 7608.

(37) Kochetov, G. A.; Solovjeva, O. N. Structure and functioning

mechanism of transketolase. Biochim. Biophys. Acta, Proteins Proteomics

2014, 1844, 1608−1618.

U. T. Biocatalysis: Enzymatic Synthesis for Industrial Applications.

Angew. Chem., Int. Ed. 2021, 60, 88−119.

19) Sheldon, R. A.; Pereira, P. C. Biocatalysis engineering: the big

picture. Chem. Soc. Rev. 2017, 46, 2678−2691.

20) Su, B.; Xu, L.; Xu, X.; Wang, L.; Li, A.; Lin, J.; Ye, L.; Yu, H.

(

38) Fedorchuk, T. P.; Khusnutdinova, A. N.; Evdokimova, E.; Flick,

Redesign of a short-chain dehydrogenase/reductase for asymmetric

synthesis of ethyl (R)-2-hydroxy-4-phenylbutanoate based on per-

residue free energy decomposition and sequence conservatism

analysis. Green Synth. Catal. 2020, 1, 150−159.

R.; Di Leo, R.; Stogios, P.; Savchenko, A.; Yakunin, A. F. One-Pot

Biocatalytic Transformation of Adipic Acid to 6-Aminocaproic Acid

and 1,6-Hexamethylenediamine Using Carboxylic Acid Reductases

and Transaminases. J. Am. Chem. Soc. 2020, 142, 1038−1048.

39) Slabu, I.; Galman, J. L.; Lloyd, R. C.; Turner, N. J. Discovery,

Engineering, and Synthetic Application of Transaminase Biocatalysts.

(

21) Li, J.; Qu, G.; Shang, N.; Chen, P.; Men, Y.; Liu, W.; Mei, Z.;

Sun, Y.; Sun, Z. Near-perfect control of the regioselective

glucosylation enabled by rational design of glycosyltransferases.

Green Synth. Catal. 2021, 2, 45−53.

ACS Catal. 2017, 7, 8263−8284.

41) Patil, M. D.; Grogan, G.; Bommarius, A.; Yun, H. Recent

40) Guo, F.; Berglund, P. Transaminase biocatalysis: optimization

and application. Green Chem. 2017, 19, 333−360.

Advances in ω -Transaminase-Mediated Biocatalysis for the Enantio-

selective Synthesis of Chiral Amines. Catalysts 2018, 8, 254.

42) Steffen-Munsberg, F.; Vickers, C.; Kohls, H.; Land, H.; Mallin,

(

22) Steinreiber, J.; Fesko, K.; Reisinger, C.; Schurmann, M.; van

Assema, F.; Wolberg, M.; Mink, D.; Griengl, H. Threonine

aldolases an emerging tool for organic synthesis. Tetrahedron

007, 63, 918−926.

23) Liu, Z.; Chen, X.; Chen, Q.; Feng, J.; Wang, M.; Wu, Q.; Zhu,

D. Engineering of l-threonine aldolase for the preparation of 4-

methylsulfonyl)phenylserine, an important intermediate for the

synthesis of florfenicol and thiamphenicol. Enzyme Microb. Technol.

020, 137, 109551.

24) Zheng, W.; Yu, H.; Fang, S.; Chen, K.; Wang, Z.; Cheng, X.;

(

H.; Nobili, A.; Skalden, L.; van den Bergh, T.; Joosten, H.-J.;

Berglund, P.; Höhne, M.; Bornscheuer, U. T. Bioinformatic analysis of

a PLP-dependent enzyme superfamily suitable for biocatalytic

applications. Biotechnol. Adv. 2015, 33, 566−604.

Xu, G.; Yang, L.; Wu, J. Directed Evolution of L -Threonine Aldolase

43) Sehl, T.; Hailes, H. C.; Ward, J. M.; Wardenga, R.; von Lieres,

for the Diastereoselective Synthesis of β -Hydroxy-α -amino Acids.

ACS Catal. 2021, 3198−3205.

E.; Offermann, H.; Westphal, R.; Pohl, M.; Rother, D. Two steps in

one pot: enzyme cascade for the synthesis of nor(pseudo)ephedrine

from inexpensive starting materials. Angew. Chem., Int. Ed. 2013, 52,

25) Xu, L.; Wang, L.-C.; Xu, X.-Q.; Lin, J. Characteristics of L-

threonine transaldolase for asymmetric synthesis of β -hydroxy-α -

772−6775.

amino acids. Catal. Sci. Technol. 2019, 9, 5943−5952.

44) Erdmann, V.; Sehl, T.; Frindi Wosch, I.; Simon, R. C.; Kroutil,

W. Methoxamine Synthesis in a Biocatalytic 1-Pot 2-Step Cascade

26) Xu, L.; Wang, L. C.; Su, B. M.; Xu, X. Q.; Lin, J. Multi-enzyme

Approach. ACS Catal. 2019, 9, 7380−7388.

cascade for improving β -hydroxy-α -amino acids production by

engineering L-threonine transaldolase and combining acetaldehyde

elimination system. Bioresour. Technol. 2020 , 310 , 123439.

45) Payongsri, P.; Steadman, D.; Hailes, H. C.; Dalby, P. A. Second

generation engineering of transketolase for polar aromatic aldehyde

substrates. Enzyme Microb. Technol. 2015, 71, 45−52.

46) Mitschke, L.; Parthier, C.; Schröder-Tittmann, K.; Coy, J.;

27) Li, Q.; Zhang, H.; Li, C.; Xu, P. Stereoselective Synthesis of

−)-Chloramphenicol, (+)-Thiamphenicol and (+)-Sphinganine via

Chiral Tricyclic Iminolactone. Chin. J. Chem. 2013, 31, 149−153.

28) Ingram, C. U.; Bommer, M.; Smith, M. E. B.; Dalby, P. A.;

(

Ludtke, S.; Tittmann, K. The Crystal Structure of Human

Transketolase and New Insights into Its Mode of Action. J. Biol.

Chem. 2010, 285, 31559−31570.

(

Ward, J. M.; Hailes, H. C.; Lye, G. J. One-pot synthesis of amino-

alcohols using a de-novo transketolase and β -alanine: pyruvate

transaminase pathway in Escherichia coli . Biotechnol. Bioeng. 2007,

47) Höhne, M.; Schätzle, S.; Jochens, H.; Robins, K.; Bornscheuer,

U. T. Rational assignment of key motifs for function guides in silico

enzyme identification. Nat. Chem. Biol. 2010, 6, 807−813.

48) Savile, C. K.; Janey, J. M.; Mundorff, E. C.; Moore, J. C.; Tam,

S.; Jarvis, W. R.; Colbeck, J. C.; Krebber, A.; Fleitz, F. J.; Brands, J.;

Devine, P. N.; Huisman, G. W.; Hughes, G. J. Biocatalytic

Asymmetric Synthesis of Chiral Amines from Ketones Applied to

Sitagliptin Manufacture. Science 2010, 329, 305−309.

49) Bea, H.-S.; Park, H.-J.; Lee, S.-H.; Yun, H. Kinetic resolution of

aromatic β -amino acids by ω -transaminase. Chem. Commun. 2011, 47,

29) Sehl, T.; Maugeri, Z.; Rother, D. Multi-step synthesis strategies

6, 559−569.

(

towards 1,2-amino alcohols with special emphasis on phenylpropanol-

amines. J. Mol. Catal. B: Enzym. 2015, 114, 65−71.

(

30) Bawn, M.; Subrizi, F.; Lye, G. J.; Sheppard, T. D.; Hailes, H. C.;

Ward, J. M. One-pot, two-step transaminase and transketolase

synthesis of l-gluco-heptulose from l-arabinose. Enzyme Microb.

Technol. 2018, 116, 16−22.

31) Gupta, P.; Mahajan, N. Biocatalytic approaches towards the

5894.

(

50) Weiß, M. S.; Pavlidis, I. V.; Spurr, P.; Hanlon, S. P.; Wirz, B.;

stereoselective synthesis of vicinal amino alcohols. New J. Chem. 2018,

2, 12296−12327.

32) Casajus, H.; Lagarde, A.; Leremboure, M.; De Dios Miguel, T.;

Iding, H.; Bornscheuer, U. T. Amine Transaminase Engineering for

Spatially Bulky Substrate Acceptance. ChemBioChem 2017, 18, 1022−

1026.

(

Nauton, L.; Thery, V.; Fessner, W. D.; Duguet, N.; Charmantray, F.;

Hecquet, L. Enzymatic Synthesis of Aliphatic Acyloins Catalyzed by

Thermostable Transketolase. ChemCatChem 2020, 12, 5772−5779.

(51) Koszelewski, D.; Lavandera, I.; Clay, D.; Rozzell, D.; Kroutil,

W. Asymmetric Synthesis of Optically Pure Pharmacologically

Relevant Amines Employing ω -Transaminases. Adv. Synth. Catal.

2008, 350, 2761−2766.

(52) Andrade, L. H.; Silva, A. V.; Milani, P.; Koszelewski, D.;

Kroutil, W. ω -Transaminases as efficient biocatalysts to obtain novel

(

33) Hailes, H. C.; Rother, D.; Muller, M.; Westphal, R.; Ward, J.

M.; Pleiss, J.; Vogel, C.; Pohl, M. Engineering stereoselectivity of

ThDP-dependent enzymes. FEBS J. 2013, 280, 6374−6394.

487

ACS Catal. 2021, 11, 7477−7488

ACS Catalysis

53) Wang, B.; Land, H.; Berglund, P. An efficient single-enzymatic

Research Article

chiral selenium-amine ligands for Pd-catalysis. Org. Biomol. Chem.

010, 8, 2043.

cascade for asymmetric synthesis of chiral amines catalyzed by ω -

transaminase. Chem. Commun. 2013, 49, 161−163.

54) Ryan, J.; Siauciulis, M.; Gomm, A.; Maciá, B.; O ’Reilly, E.;

(

Caprio, V. Transaminase Triggered Aza-Michael Approach for the

Enantioselective Synthesis of Piperidine Scaffolds. J. Am. Chem. Soc.

(

016, 138, 15798−15800.

55) O’Reilly, E.; Iglesias, C.; Ghislieri, D.; Hopwood, J.; Galman, J.

L.; Lloyd, R. C.; Turner, N. J. A Regio- and Stereoselective ω -

Transaminase/Monoamine Oxidase Cascade for the Synthesis of

Chiral 2,5-Disubstituted Pyrrolidines. Angew. Chem., Int. Ed. 2014, 53,

(

447−2450.

56) Cheng, F.; Li, J. M.; Zhou, S. P.; Liu, Q.; Jin, L. Q.; Xue, Y. P.;

Zheng, Y. G. A Single-Transaminase-Catalyzed Biocatalytic Cascade

for Efficient Asymmetric Synthesis of l-Phosphinothricin. ChemBio-

Chem 2021, 22, 345−348.

57) Romney, D. K.; Murciano-Calles, J.; Wehrmuller, J. E.; Arnold,

F. H. Unlocking Reactivity of TrpB: A General Biocatalytic Platform

for Synthesis of Tryptophan Analogues. J. Am. Chem. Soc. 2017, 139,

58) Chen, C. S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. Quantitative

0769−10776.

analyses of biochemical kinetic resolutions of enantiomers. J. Am.

Chem. Soc. 1982, 104, 7294−7299.

59) Yilmaz, E.; Can, K.; Sezgin, M.; Yilmaz, M. Immobilization of

Candida rugosa lipase on glass beads for enantioselective hydrolysis of

racemic Naproxen methyl ester. Bioresour. Technol. 2011, 102, 499−

60) Han, S. W.; Kim, J.; Cho, H. S.; Shin, J. S. Active Site

06.

Engineering of ω -Transaminase Guided by Docking Orientation

Analysis and Virtual Activity Screening. ACS Catal. 2017, 7, 3752−

(

762.

61) Voss, M.; Das, D.; Genz, M.; Kumar, A.; Kulkarni, N.; Kustosz,

J.; Kumar, P.; Bornscheuer, U. T.; Höhne, M. In Silico Based

Engineering Approach to Improve Transaminases for the Conversion

of Bulky Substrates. ACS Catal. 2018 , 8 , 11524 −11533.

63) Zheng, W.; Chen, K.; Wang, Z.; Cheng, X.; Xu, G.; Yang, L.;

62) Radisky, E. S.; Koshland, D. E. A clogged gutter mechanism for

protease inhibitors. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 10316−

0321.

(

Wu, J. Construction of a Highly Diastereoselective Aldol Reaction

System with l-Threonine Aldolase by Computer-Assisted Rational

Molecular Modification and Medium Engineering. Org. Lett. 2020, 22,

64) Eliot, A. C.; Kirsch, J. F. Pyridoxal Phosphate Enzymes:

763−5767.

Mechanistic, Structural, and Evolutionary Considerations. Annu. Rev.

Biochem. 2004, 73, 383−415.

(

65) Salomon-Ferrer, R.; Götz, A. W.; Poole, D.; Le Grand, S.;

Walker, R. C. Routine Microsecond Molecular Dynamics Simulations

with AMBER on GPUs. 2. Explicit Solvent Particle Mesh Ewald. J.

Chem. Theory Comput. 2013, 9, 3878−3888.

(

66) Maier, J. A.; Martinez, C.; Kasavajhala, K.; Wickstrom, L.;

Hauser, K. E.; Simmerling, C. ff14SB: Improving the Accuracy of

Protein Side Chain and Backbone Parameters from ff99SB. J. Chem.

Theory Comput. 2015, 11, 3696−3713.

(

67) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.;

DiNola, A.; Haak, J. R. Molecular dynamics with coupling to an

external bath. J. Chem. Phys. 1984, 81, 3684 −3690.

68) Darden, T.; York, D.; Pedersen, L. Particle mesh Ewald: An N ·

log(N) method for Ewald sums in large systems. J. Chem. Phys. 1993,

8, 10089−10092.

69) Roe, D. R.; Cheatham, T. E., III PTRAJ and CPPTRAJ:

Software for Processing and Analysis of Molecular Dynamics

Trajectory Data. J. Chem. Theory Comput. 2013, 9, 3084−3095.

488