One-Pot Enantioselective Synthesis of d-Phenylglycines from Racemic Mandelic Acids, Styrenes, or Biobased l-Phenylalanine via Cascade Biocatalysis

Article abstract of DOI:10.1002/adsc.201700956

Enantiopure d-phenylglycine and its derivatives are an important group of chiral amino acids with broad applications in thepharmaceutical industry. However, the existing synthetic methods for d-phenylglycine mainly rely on toxic cyanide chemistry and multistep processes. To provide green and safe alternatives, we envisaged cascade biocatalysis for the one-pot synthesis of d-phenylglycine from racemic mandelic acid, styrene, and biobased l-phenylalanine, respectively. Recombinant Escherichia coli (LZ110) was engineered to coexpress four enzymes to catalyze a 3-step reaction in one pot, transforming mandelic acid (210 mM) to give enantiopure d-phenylglycine in 29.5 g L^?1 (195 mM) with 93% conversion. Using the same whole-cell catalyst, twelve other d-phenylglycine derivatives were also produced from the corresponding mandelic acid derivatives in high conversion (58–94%) and very high ee (93–99%). E. coli (LZ116) expressing seven enzymes was constructed for the transformation of styrene to enantiopure d-phenylglycine in 80% conversion via a one-pot 6-step cascade biotransformation. Twelve substituted d-phenylglycines were also produced from the corresponding styrene derivatives in high conversion (45–90%) and very high ee (92–99%) via the same cascade reactions. A nine-enzymeexpressing E. coli (LZ143) was engineered to transform biobased l-phenylalanine to enantiopure d-phenylglycine in 83% conversion via a one-pot 8-step transformation. Preparative biotransformations were also demonstrated. The high-yielding synthetic methods use cheap and green reagents (ammonia, glucose, and/or oxygen), and E. coli whole-cell catalysts, thus providing green and useful alternative methods for manufacturing d-phenylglycine. (Figure presented.).

Full text of DOI:10.1002/adsc.201700956

Title: One-Pot Enantioselective Synthesis of ꢀ-Phenylglycines from
Racemic Mandelic Acids, Styrenes, or Biobased ʟ-Phenylalanine
via Cascade Biocatalysis
Authors: Yi Zhou, Shuke Wu, and Zhi Li
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Adv. Synth. Catal. 10.1002/adsc.201700956
Link to VoR: http://dx.doi.org/10.1002/adsc.201700956

10.1002/adsc.201700956
Advanced Synthesis & Catalysis
FULL PAPER
DOI: 10.1002/adsc.201((will be filled in by the editorial staff))
One-Pot Enantioselective Synthesis of ᴅ-Phenylglycines from
Racemic Mandelic Acids, Styrenes, or Biobased ʟ-
Phenylalanine via Cascade Biocatalysis
Yi Zhou,a Shuke Wu,a, b and Zhi Lia, b,*
a
Synthetic Biology for Clinical and Technological Innovation (SynCTI), Life Sciences Institute, National University of
Singapore, 28 Medical Drive, Singapore 117456
Department of Chemical and Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4,
b
Singapore 117585
Fax: (+65)-6779-1936; phone: (+65)-6516-8416; e-mail: chelz@nus.edu.sg
Received: ((will be filled in by the editorial staff))
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201######.((Please
delete if not appropriate))
Abstract. Enantiopure ᴅ-phenylglycine and its derivatives the transformation of styrene to enantiopure ᴅ-
are an important group of chiral amino acids with broad phenylglycine in 80% conversion via one-pot 6-step
applications in pharmaceutical industry. However, the cascade biotransformation. 12 Substituted ᴅ-phenylglycines
existing synthetic methods for ᴅ-phenylglycine mainly rely were also produced from the corresponding styrene
on toxic cyanide chemistry and multistep processes. To derivatives in high conversion (45-90%) and very high ee
provide green and safe alternatives, we leveraged cascade (92-99%) via the same cascade reactions. A nine-enzyme-
biocatalysis for one-pot synthesis of ᴅ-phenylglycine from expressing E. coli (LZ143) was engineered to transform
racemic mandelic acid, styrene, and biobased ʟ- biobased ʟ-phenylalanine to enantiopure ᴅ-phenylglycine in
phenylalanine, respectively. Recombinant Escherichia coli 83% conversion via one-pot 8-step transformation.
(LZ110) was engineered to coexpress four enzymes to Preparative biotransformation was also demonstrated. The
catalyze 3-step reactions in one pot, transforming mandelic high-yielding synthetic methods use cheap and green
acid (210 mM) to give enantiopure ᴅ-phenylglycine in 29.5 g reagents (ammonia, glucose, and/or oxygen), and E. coli
L−1 (195 mM) with 93% conversion. Using the same whole- whole-cell catalysts, thus providing green and useful
cell catalyst, other 12 ᴅ-phenylglycine derivatives were also alternative methods for manufacturing ᴅ-phenylglycine.
produced from the corresponding mandelic acid derivatives
in high conversion (58-94%) and very high ee (93-99%). E.
coli (LZ116) expressing seven enzymes was constructed for
Keywords: Amino acids; Biotransformations; Cascade
reactions; Enzyme catalysis; Green chemistry
cyanide,[2a,6]
followed
by
diastereomeric
Introduction
crystallization with ᴅ-camphorsulfonic acid to give
68% yield.[7] Recently, asymmetric Strecker reaction
was reported for the synthesis of enantiopure α-amino
acids from N-substituted imines without the
resolution step, but it requires complicated and costly
protection/deprotection steps with low atom
Enantiopure ᴅ-amino acids are a group of very
important chiral chemicals for manufacturing
pharmaceuticals and other bioactive compounds.[1]
For instance, ᴅ-phenylglycine 1a is consumed at
several thousand tons per year for the synthesis of β-
lactam antibiotics (ampicillin and cephalexin)[2] and
also used for the preparation of an HCV NS5A
inhibitor[3a] and a dual inhibitor of phosphoinositide-
dependent kinase-1 and aurora kinase A.[3b] 4-Fluoro-
ᴅ-phenylglycine 1d is a chiral building block for a
potent h5-HT1D receptor agonist[4a] and HCV NS5B
polymerase inhibitors.[4b] 4-Chloro-ᴅ-phenylglycine
1f is a key chiral synthon for a macrocyclic
efficiency.[8]
hydrocyanation
For
example,
asymmetric
of
N-benzhydryl-protected
benzylimine with a thiourea-derived catalyst gave N-
benzhydryl-protected phenylglycinonitrile in 98%
ee,[8b] but it needs further hydrolysis and deprotection
steps to form ᴅ-phenylglycine.
Recently, enzyme catalysis has become a useful
tool for enantioselective synthesis.[9] Kinetic
resolution of N-acetylphenylglycine or phenylglycine
methyl ester with acylase or lipase to give ᴅ-
phenylglycine are well known but suffering from the
50% maximum yield.[10] Dynamic kinetic resolution
(DKR) of phenylglycine methyl ester with lipase and
Hedgehog pathway inhibitor[5a] and
a
potent
morpholinone MDM2 inhibitor.[5b] The conventional
process to manufacture ᴅ-phenylglycine is via
Strecker reaction of benzaldehyde with highly toxic
1
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700956
Advanced Synthesis & Catalysis
available fermentation product from renewable
feedstock in a large amount, thus its conversion to
(R)-1 could provide sustainable synthesis of new
biobased chemicals. In each of these synthetic routes
(Figure 1a, 2a, 3a), the involved multistep reactions
are performed in one pot, which can avoid the tedious
and costly intermediate isolation and waste
generation.[16] Here, we report the engineering of
three recombinant E. coli strains coexpressing 4-9
enzymes as efficient biocatalysts for the
corresponding purposed one-pot enantioselective
syntheses, and the high-yielding production of
enantiopure ᴅ-phenylglycines from the three
substrates, respectively, by whole-cell based one-pot
cascade biotransformation.
Scheme 1. One-pot synthesis of ᴅ-phenylglycines (R)-1a-
1m from racemic mandelic acids, styrenes, or ʟ-
phenylalanine via cascade biocatalysis with recombinant E.
coli cells containing the necessary enzymes.
Results and Discussion
Enantioselective conversion of racemic mandelic
acids to ᴅ-phenylglycines
pyridoxal gave ᴅ-phenylglycine amide in 88% ee at
85% conversion.[11a] DKR of phenylglycine amide
with racemase and amidase gave ᴅ-phenylglycine in
88% ee at 99% conversion.[11b] DKR of
phenylglycinonitrile with nitrilase at alkaline
condition gave ᴅ-phenylglycine in 95% ee at 81%
conversion.[11c] Clearly, the product ee in these
processes was not high enough for direct application.
On the other hand, the hydantoinase process was
developed to produce enantiopure ᴅ-phenylglycines
in 100% conversion from racemic hydantoins by
hydantoinase, carbamoylase, and racemase.[2a,6,12]
However, the substrate 5-phenylhydantoin has to be
synthesized by Bucherer–Bergs reaction with
benzaldehyde, urea, and highly toxic cyanide, and the
low solubility of 5-phenylhydantoin in aqueous
solution resulted in low overall reaction rate.[13]
Coupling ᴅ-selective α-transaminase and R-selective
ω-transaminase gave enantiopure ᴅ-phenylglycine
from a primary amine and relative expensive
phenylglyoxylic acid.[14] Therefore, highly efficient
green methods for the synthesis of enantiopure ᴅ-
phenylglycine are highly wanted. Moreover, in all
chemical and enzymatic methods described above,
cyanide was unavoidably used for the synthesis of
either substrates or products. Thus, developing a
cyanide-free and sustainable process is of great
importance. In a recent attempt, ᴅ-phenylglycine was
fermentatively produced from glucose, but in a low
concentration mixed with a large amount of ʟ-
phenylalanine, thus requiring difficult and costly
isolation and generating much waste.[15]
Selection of enzymes, engineering of a host, and
construction of recombinant E. coli strains.
Previously, ʟ-phenylglycine was produced from
racemic mandelic acid via a 3-step enzyme cascade
with
mandelate
racemase,
ᴅ-mandelate
dehydrogenase, and ʟ-amino acid dehydrogenase.[17]
However, attempt to produce ᴅ-phenylglycine failed:
initial test of several commercially available ᴅ-amino
acid dehydrogenases could not produce ᴅ-
phenylglycine,[17a] and an engineered broad-range ᴅ-
amino acid dehydrogenase showed no activity for the
formation of ᴅ-phenylglycine.[18] To provide access to
the important ᴅ-phenylglycine (Figure 1a), we
employed a unique ᴅ-phenylglycine aminotransferase
(DpgAT) from Pseudomonas stutzeri ST-201[19] for
the last amination step. The gene of DpgAT was
synthesized and cloned to give pRSF-dpgat plasmid.
E. coli BL21 (DE3) containing pRSF-dpgat was
engineered and the whole cells were used to convert
phenylglyoxylic acid 2a to give ᴅ-phenylglycine 1a
in 70% ee. The product ee is not high probably due to
the endogenous ʟ-α-transaminases in the E. coli BL21
(DE3) host transformed 2a to (S)-1a. A new E. coli
DL39 (DE3) host was engineered by integration of
λDE3 prophage into the E. coli DL39 strain,[20] which
has mutations in three α-transaminase genes ilvE,
tyrB, and aspC. E. coli DL39 (DE3) expressing
DpgAT produced (R)-1a in 99% ee from 2a (Figure
S1 in the Supporting Information). Thus E. coli DL39
(DE3) was used as the expression host throughout
this study.
DpgAT catalyzes the conversion of 2a to (R)-1a by
using ʟ-glutamate as the amine donor to form α-
ketoglutarate. To enhance the productivity, we
utilized the glutamate dehydrogenase (GluDH) to
regenerate ʟ-glutamate from α-ketoglutarate and
ammonia.
For the oxidation of (S)-mandelic acid 3a, a
hydroxy acid oxidase (HMO) was used in previous
studies,[21] yet it was not very efficient and required
We aimed to achieve green and efficient synthesis
of ᴅ-phenylglycine 1 from racemic mandelic acid,
styrene, or ʟ-phenylalanine via one-pot cascade
biocatalysis (Scheme 1). Racemic mandelic acid is a
cheap and easily available substrate, thus its
conversion to (R)-1 could be highly attractive. While
styrene is a very cheap hydrocarbon, its own
synthesis and conversion to (R)-1 are free of cyanide.
On the other hand, ʟ-phenylalanine is an easily
2
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700956
Advanced Synthesis & Catalysis
Figure 1. (a) Cascade biotransformation of rac-3a to (R)-1a with (S)-mandelate dehydrogenase (SMDH), mandelate
racemase (MR), ᴅ-phenylglycine aminotransferase (DpgAT) and glutamate dehydrogenase (GluDH). (b) Engineering of
recombinant E. coli by combination of SMDH-MR and DpgAT-GluDH modules on four compatible plasmids, pACYC,
pCDF, pET and pRSF, respectively. (c) Screening of E. coli strains (LZ101-112) cells (10 g cdw L−1) for one-pot
production of (R)-1a from rac-3a (30 mM) in 2 h. (d) Construction of E. coli (LZ110). (e) Time course of one-pot
production of (R)-1a from rac-3a using E. coli (LZ110) cells (15 g cdw L−1) with step-wise substrate feeding in KP
buffer containing glucose and NH3/NH4Cl. (f) Precipitation of (R)-1a in the reaction buffer during the biotransformation.
Error bars in (c) and (e) represent the standard deviation of triplicated biotransformation.
catalase to remove the toxic H2O2. Here, we cloned a
membrane-associated, FMN-dependent (S)-mandelate
dehydrogenase (SMDH)[22] from Pseudomonas
putida ATCC 12633 for the oxidation of (S)-3a to 2a.
SMDH showed much higher productivity than HMO
(Figure S2 in the Supporting Information). Therefore,
phosphate (KP) buffer containing NH3 and glucose.
The overall cascade reaction requires 1 mole of
NADPH (by GluDH) for producing 1 mole of (R)-1a.
Thus, glucose (2%) was added to regenerate NADPH
via cellular metabolism. As shown in Figure, 1c, all
twelve strains successfully produced (R)-1a in 2 h.
SMDH was chosen for the second step of the reaction. The E. coli (LZ110) containing pACYC-smdh-mr and
The mandelate racemase (MR)[23] from the same P.
putida strain was cloned in E. coli and employed for
racemization of 3a with high efficiency.
pRSF-dpgat-gludh (Figure 1d) gave quantitative
conversion to (R)-1a, thus it was selected for further
investigation.
To achieve the balanced enzyme expression in an
E. coli strain, we divided the four enzymes into two
enzyme modules, SMDH-MR and DpgAT-GluDH,
and each enzyme module was constructed on four
plasmids, pACYC, pCDF, pET, and pRSF (Figure
1b). To obtain the best single E. coli strain as the
catalyst for the cascade biotransformation, the four
plasmids of SMDH-MR and the four plasmids of
DpgAT-GluDH were combinatorially transformed in
E. coli DL39 (DE3) host to give twelve strains E. coli
(LZ101-LZ112), each coexpressing SMDH, MR,
DpgAT, and GluDH.
Biotransformation of rac-3a in 50 mM with E. coli
(LZ110) gave a very low amount of (R)-1a (15%)
with significant accumulation of 2a (85%),
suggesting the inhibition of 2a. Further experiment
using 30-50 mM 2a as substrate revealed that higher
than 40 mM of 2a significantly inhibited the
transamination to (R)-1a (Figure S3 in the Supporting
Information). To tackle this issue, we exploited the
step-wise addition of substrate rac-3a into the
reaction system. The best result was depicted in
Figure 1e: 30 mM rac-3a was fed in every two hours
for 7 times, and 195 mM (29.5 g L−1) of enantiopure
(R)-1a was produced in 24 h, representing 93%
conversion from rac-3a (210 mM). Interestingly, (R)-
1a precipitated in the reaction buffer during the
biotransformation (Figure 1f). Clearly, this in situ
precipitation is highly wanted in bioprocess to avoid
One-pot synthesis of ᴅ-phenylglycine 1a from
racemic mandelic acid 3a with E. coli coexpressing
4 enzymes. The resting cells of the twelve strains E.
coli
(LZ101-LZ112)
were
examined
for
biotransformation of 30 mM rac-3a in potassium
3
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700956
Advanced Synthesis & Catalysis
the product inhibition and facilitate the product
recovery.
Enantioselective conversion of styrenes to ᴅ-
phenylglycines
The one-pot 3-step cascade utilizes easily available
racemic mandelic acid as substrate, cheap and green
O2, NH3, and glucose as reagents, and gives
enantiopure product in high yield. With in situ
precipitation, (R)-1a could be produced in up to 30 g
L−1, representing one of the most efficient examples
of biocatalytic conversion of -OH group to -NH2
group.[24]
Selection of enzymes and construction of E. coli
strains. Styrene is manufactured at a very large scale
and low price by dehydrogenation of ethylbenzene.
Direct production of (R)-1a from styrene 7a could
provide an economic and cyanide-free synthesis.
Based on the above 2-step cascade from (S)-3a to
(R)-1a (Figure 1a), additional four enzymes were
used to convert 7a to (S)-3a (Figure 2a). Styrene
monooxygenase (SMO), epoxide hydrolase (SpEH)
and alcohol dehydrogenase (AlkJ) and aldehyde
dehydrogenase (EcALDH) were employed for the
conversion of styrene 7a to (S)-3a.[21a] The SMDH,
DpgAT, and GluDH were used for the conversion of
(S)-3a to (R)-1a. To achieve the balanced enzyme
expression in E. coli strain, we divided the 7 enzymes
into three enzyme modules, SMO-SpEH, AlkJ-
EcALDH, and DpgAT-GluDH-SMDH, and each
enzyme module was constructed on four plasmids,
pACYC, pCDF, pET, and pRSF (Figure 2b). The
combination of plasmids of SMO-SpEH, AlkJ-
EcALDH, and DpgAT-GluDH-SMDH in E. coli
DL39 (DE3) host gave 24 strains E. coli (LZ113-
LZ136), each coexpressing the 7 enzymes.
One-pot
synthesis
of
substituted
ᴅ-
phenylglycines 1b-1m from substituted racemic
mandelic acids 3a-3m with E. coli (LZ110). To
explore the application scope of the 3-step cascade,
the resting cells of E. coli (LZ110) were applied to
convert 12 different substituted racemic mandelic
acids 3b-3m to (R)-1b-1m. As listed in Table 1, 5
substituted ᴅ-phenylglycines, 1i, 1j, 1k, 1l, 1m, were
produced in enantiopure form (≥ 99% ee) with high
conversion (80-91%). Other 5 ᴅ-phenylglycines, 1b,
1c, 1d, 1f, 1g, were produced in very high ee (95-
98%) and high conversion (73-94%). (R)-1e was
formed in very high ee (97%) with moderate
conversion (58%), probably due to the low activity of
SMDH on (S)-3e. (R)-1h was formed in slightly
lower ee (93%) and high conversion (81%), possibly
due to the less satisfactory enantioselectivity of
DpgAT on 2h. These results demonstrated that the 3-
step enzyme cascade route could be generally
applicable to produce many useful ᴅ-phenylglycine
derivatives from racemic mandelic acids.
Table 2. One-pot synthesis of (R)-1b-1m from 7b-7m via
cascade biocatalysis with recombinant E. coli (LZ116)
cells.[a]
Table 1. One-pot synthesis of (R)-1b-1m from rac-3b-3m
via cascade biocatalysis with recombinant E. coli (LZ110)
cells.[a]
7
X
Sub.
conc.
[mM]
10
20
20
10
20
20
10
10
20
20
10
(R)-1
Conv. Prod.
[%][b]
ee
[%][c]
99
7b
7c
7d
7e
7f
7g
7h
7i
o-F
m-F
p-F
o-Cl
m-Cl
p-Cl
m-Br
p-Br
m-CH3
p-CH3
m-OCH3
p-OCH3
1b
1c
1d
1e
1f
1g
1h
1i
82
88
80
49
75
86
50
45
86
90
87
90
97
99
rac-3
X
Sub.
conc.
[mM]
20
20
20
10
20
10
10
10
20
20
10
(R)-1
Conv. Prod.
[%][b]
ee
[%][c]
98
97
94
3b
3c
3d
3e
3f
3g
3h
3i
o-F
m-F
p-F
o-Cl
m-Cl
p-Cl
m-Br
p-Br
m-CH3
p-CH3
m-OCH3
p-OCH3
1b
1c
1d
1e
1f
1g
1h
1i
89
86
94
58
88
73
81
90
91
90
80
86
99
92
95
97
99
98
97
95
7j
7k
7l
1j
1k
1l
99
97
96
98
93
7m
20
1m
99
99
[a]
Reaction conditions: 7b-7m (10-20 mM) and E. coli
3j
3k
3l
1j
1k
1l
(LZ116) cells (15 g cdw L−1) in two-liquid phase system
consisting of 0.4 mL EO and 2 mL KP buffer (400 mM,
pH 8.0) containing glucose (0.5%) and NH3/NH4Cl (100
mM, pH 8.0) at 30 C for 24 h. Additional glucose (1%)
and NH3/NH4Cl (100 mM) was added at 10 h.
99
99
99
3m
20
1m
o
[a] Reaction conditions: rac-3b-3m (10-20 mM) and E. coli
(LZ110) cells (15 g cdw L−1) in 2mL KP buffer (400 mM,
pH 8.0) containing glucose (2% w/v) and NH3/NH4Cl (200
[b]
[c]
Determined by HPLC analysis.
HPLC analysis.
Measured by chiral
o
[b]
mM, pH 8.0) at 30 C for 24 h.
Determined by HPLC
analysis. [c] Measured by chiral HPLC analysis.
4
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700956
Advanced Synthesis & Catalysis
Figure 2. (a) Cascade biotransformation of 7a to (R)-1a with styrene monooxygenase (SMO), epoxide hydrolase (SpEH),
alcohol dehydrogenase (AlkJ), aldehyde dehydrogenase (EcALDH), SMDH, DpgAT, and GluDH. (b) Engineering of
recombinant E. coli by combination of SMO-SpEH, AlkJ-EcALDH, and DpgAT-GluDH-SMDH modules on four
compatible plasmids, pACYC, pCDF, pET and pRSF, respectively. (c) Screening of E. coli (LZ113-136) cells (10 g cdw
L−1) for one-pot production of (R)-1a from 7a (30 mM) in 24 h. (d) Construction of E. coli (LZ116). (e) Time course of
one-pot production of (R)-1a from 7a (60 mM) using E. coli (LZ116) cells (15 g cdw L−1) in a two-liquid phase system of
EO and KP buffer containing glucose and NH3/NH4Cl. Error bars in (c) and (e) represent the standard deviation of
triplicated biotransformation.
One-pot synthesis of ᴅ-phenylglycine 1a from
styrene 7a with E. coli coexpressing 7 enzymes.
Cascade biotransformation of 30 mM 7a was
performed with resting cells of E. coli (LZ113-
LZ136) in a two-liquid phase system of ethyl oleate
(EO)[25] and KP buffer containing NH3 and glucose.
EO serves as a reservoir for hydrophobic substrates 7
and intermediate 6 to reduce the toxicity. Glucose
was added to regenerate NADPH for the cascade
reaction [1 mole of NADPH required for producing 1
mole of (R)-1a] via cellular metabolism. As shown in
Figure 2c, (R)-1a was produced by all 24 strains in 24
h. Among the strains, the E. coli (LZ116) containing
pACYC-smo-speh, pET-alkj-ecaldh and pRSF-dpgat-
gludh-smdh (Figure 2d) gave the highest conversion
to (R)-1a, thus it was selected as the best catalyst for
this cascade biotransformation. Under the optimized
conditions, 60 mM of 7a was reacted with resting
cells of E. coli (LZ116) to give 48 mM (7.3 g L−1) of
enantiopure (R)-1a in 80% conversion in 24 h (Figure
2e). A small amount of 5a (6 mM) and 2a (2 mM)
were accumulated at the end of the reaction.
and high conversion (80-90%). (R)-1c, (R)-1f, (R)-1l,
and (R)-1m, were also formed in very high ee (94-
97%) and high conversion (75-90%). Moderate
conversion (45-50%) was observed for ortho-chloro
substituted (R)-1e (97% ee) and bromo substituted
(R)-1h (92% ee) and (R)-1i (99% ee), probably due to
the low activity of SMO on the corresponding alkene
substrates (7e, 7h, 7i) with a ortho- or bulky
substituent. These results demonstrated the general
applicability of this 6-step enzyme cascade for
synthesis of ᴅ-phenylglycine derivatives.
In comparison with the 3-step cascade from
mandelic acid, the product concentration of this 6-
step cascade from styrene is lower, probably due to
the moderate efficiency of styrene epoxidation with
SMO and diol oxidation with AlkJ.20a Nevertheless,
the successful production of ᴅ-phenylglycine from
styrene in high conversion by using O2, NH3, and
glucose demonstrated
a
totally new one-pot
enantioselective synthesis of this important class of
chiral compounds without involving the cyanide
chemistry. We believe that cascade biocatalytic
oxidation could be a general tool for cyanide-free
production of carboxy acids.
One-pot
synthesis
of
substituted
ᴅ-
phenylglycines 1b-1m from substituted styrenes
7b-7m with E. coli (LZ116). 12 Substituted styrenes
7b-7m (10-20 mM) were subjected to the resting
cells of E. coli (LZ116) to explore the substrate scope
(Table 2). 5 Substituted ᴅ-phenylglycines, 1b, 1d, 1g,
1j, and 1k, were produced in excellent ee (98-99%)
Enantioselective conversion of ʟ-phenylalanine to
ᴅ-phenylglycine
5
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700956
Advanced Synthesis & Catalysis
Figure 3. (a) Cascade biotransformation of (S)-9a to (R)-1a with phenylalanine ammonia lyase (PAL), pheylacrylic acid
decarboxylase (PAD), SMO, SpEH, AlkJ, EcALDH, SMDH, DpgAT, and GluDH. (b) PAD-PAL module on four
compatible plasmids, pACYC, pCDF, pET and pRSF, respectively. (c) Screening of E. coli (LZ137-160) cells (10 g cdw
L−1) for one-pot production of (R)-1a from (S)-9a (30 mM) in 24h. (d) Construction of E. coli (LZ143). (e) Time course
of one-pot production of (R)-1a from (S)-9a (60 mM) using E. coli (LZ143) cells (15 g cdw L−1) in a two-liquid phase
system of EO and KP buffer containing glucose and NH3/NH4Cl. Error bars in (c) and (e) represent the standard deviation
of triplicated biotransformation.
Selection of enzymes and construction of E. coli
strains. An important trend in chemical synthesis is
to use renewable starting material.[26] While mandelic
acid and styrene are mainly synthesized from
petroleum-based materials, many biogenic ʟ-amino
acids are currently manufactured from sugars by
microbial fermentation. ʟ-Phenylalanine (S)-9a is one
of them, and its cascade biotransformation to (R)-1a
could provide a sustainable synthetic route. Building-
up on the above 6-step cascade from 7a to (R)-1a
(Figure 2a), two more enzymes, phenylalanine
ammonia lyase (PAL) and phenylacrylic acid
decarboxylase (PAD),[21b] were used to convert (S)-9a
to 7a (Figure 3a). The PAL-PAD module was
constructed on the four plasmids (Figure 3b), which
were then transformed into E. coli (LZ112-LZ136)
cells to give 24 strains E. coli (LZ137-LZ160). Each
cascade reaction [1 mole of NADPH required for
producing 1 mole of (R)-1a] via cellular metabolism.
As shown in Figure 3c, (R)-1a was produced from
(S)-9a by all 24 strains in 24 h. The E. coli (LZ143)
with pCDF-pad-pal, pACYC-smo-speh, pET-alkj-
ecaldh, and pRSF-dpgat-gludh-smdh (Figure 3d)
clearly outperformed other 23 strains in conversion to
(R)-1a, thus it was chosen for further study. Under
the optimized reaction conditions, 60 mM of (S)-9a
was transformed by resting cells of E. coli (LZ143) to
give 50 mM (7.5 g L−1) of enantiopure (R)-1a in 83%
conversion in 24 h (Figure 3e). Only a small amount
of 5a (1 mM) and 3a (5 mM) were left at the end of
the reaction. In comparison with the 6-step cascade
from 7a, this 8-step cascade from (S)-9a achieved the
similar high product concentration and conversion,
demonstrating the high efficiency of a long non-
of E. coli (LZ137-LZ160) coexpressed the 9 enzymes, natural enzyme cascade/pathway.
PAD, PAL, SMO, SpEH, AlkJ, EcALDH, DpgAT,
GluDH, and SMDH.
Preparation of ᴅ-phenylglycine 1a and ᴅ-4-
fluorophenylglycine 1d by cascade biocatalysis.
One-pot synthesis of ᴅ-phenylglycine 1a from ʟ-
phenylalanine 9a with E. coli coexpressing 9
enzymes. The 24 E. coli (LZ137-LZ160) strains were
examined for cascade biotransformation of 30 mM
(S)-9a in a two-liquid phase system of KP buffer
containing glucose and EO, which serves as a
reservoir for toxic intermediate 7a. A small amount
of glucose was added to regenerate NADPH for the
To demonstrate the synthetic utility of the three types
of whole-cell based cascade biocatalysis, preparative
biotransformation was performed on a 40-100 mL-
scale in shaking flasks (Table 3). E. coli (LZ110) (20
g cdw L−1) was applied to convert rac-3a (200 mM)
and rac-3d (40 mM) into enantiopure (R)-1a and (R)-
6
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700956
Advanced Synthesis & Catalysis
Table 3. Preparation of (R)-1a and (R)-1d from different substrates via one-pot cascade biocatalysis with E. coli (LZ110),
E. coli (LZ116), or E. coli (LZ143).[a]
Sub.
Sub. Conc.
Biocatalyst
Reaction
volume [mL
40[b]
Prod.
Conv.
[%][d]
93
78
83
Isol. Yield
Prod. ee
[%][f]
99
[mM]
200
60
60
30
[%][e]
71
rac-3a
rac-3d
7a
7d
(S)-9a
E. coli (LZ110)
E. coli (LZ110)
E. coli (LZ116)
E. coli (LZ116)
E. coli (LZ143)
(R)-1a
(R)-1d
(R)-1a
(R)-1d
(R)-1a
40[b]
52
62
46
53
99
99
99
99
120[c]
120[c]
85
81
60
96[c]
[a] General reaction conditions: E. coli resting cells (20 g cdw L−1), glucose, NH3/NH4Cl, 30 C, 24 h (see experiment part
for details). [b] KP buffer (400 mM, pH 8.0). [c] KP buffer (400 mM, pH 8.0) and EO (5: 1). [d] Determined by HPLC
analysis. [e] Simple precipitation without optimization. [f] Measured by chiral HPLC analysis.
o
1d in 93% and 78% conversion in 24 h, respectively.
The products were easily recovered by the following
simple precipitation procedure without optimization:
acidification of the reaction mixture with HCl,
centrifugation to remove the cells, neutralization of
the supernatant with NaOH to precipitate the product,
and filtration. Finally, enantiopure (R)-1a and (R)-1d
were obtained in 71% and 52% isolated yield,
respectively.
For the cascade biocatalysis from styrene, 7a (60
mM) and 7d (30 mM) were transformed with E. coli
(LZ116) (20 g cdw L−1) to enantiopure (R)-1a and
(R)-1d in 83% and 85% conversion in 24 h,
respectively. The products were isolated in 62% and
46% yield after the similar workup, respectively.
For the cascade biotransformation from biobased
ʟ-phenylalanine, enantiopure (R)-1a was produced in
81% conversion from (S)-9a (60 mM) after reaction
with E. coli (LZ143) (20 g cdw L−1) for 24 h.
Following the similar workup procedure, enantiopure
(R)-1a was obtained in 53% isolated yield. These
preparative biotransformations demonstrated the
synthetic potential of these biocatalytic one-pot
cascades.
was developed as an efficient catalyst for this type of
transformation. By using the E. coli cells as catalysts,
enantiopure (R)-1a was successfully produced from
7a in 80% conversion. This one-pot synthesis starting
from styrene and using O2, NH3, and glucose
provides a green and efficient synthesis of ᴅ-
phenylglycine and also a cyanide-free synthesis of
both substrate and product. Enantioselective synthesis
of ᴅ-phenylglycine 1a in high ee and high yield from
biobased ʟ-phenylalanine 9a was demonstrated for
the first time via one-pot 8-step cascade
biotransformation. Engineering of E. coli (LZ143)
strain coexpressing 9 enzymes (PAL, PAD, SMO,
SpEH, AlkJ, EcALDH, SMDH, DpgAT and GluDH)
in 4 plasmids gave a powerful catalyst for the cascade
reactions. Whole-cell based biotransformation
afforded enantiopure (R)-1a in 83% conversion from
biobased (S)-9a. This synthesis starts from a
renewable substrate, uses O2, NH3, and glucose, and
gives rise to a cyanide-free and sustainable synthesis
of ᴅ-phenylglycine. The one-pot 3-step and 6-step
cascade biotransformations were generally applicable
to produce different ᴅ-phenylglycine derivatives (R)-
1b-1h in high ee with high conversion from the
corresponding mandelic acids 3b-3m and styrenes
7b-7m, respectively. Representative preparation
examples [(R)-1a and (R)-1d] with all three one-pot
syntheses were demonstrated by whole-cell
biotransformation, followed by simple product
isolation via precipitation. This study highlights the
great potential of engineering cascade biocatalysis for
one-pot synthesis of important non-natural chiral
Conclusion
Highly efficient, enantioselective, and green synthesis
of ᴅ-phenylglycine 1a in high ee and high yield from
racemic mandelic acid 3a was demonstrated for the
first
time
via
one-pot
3-step
cascade
chemicals in
a
green, safe, sustainable,
biotransformation. An ʟ-α-transaminase-deficient E.
coli (LZ110) strain coexpressing 4 enzymes (MR,
SMDH, DpgAT, and GluDH) in double plasmids was
engineered as a highly productive catalyst for the
enantioselective, and efficient manner.
cascade
reactions.
Whole-cell
based
Experimental Section
biotransformation of easily available rac-3a with
non-toxic reagents (NH3 and glucose) gave
enantiopure (R)-1a in 29.5 g L−1 with 93% conversion,
providing a green and productive method for practical
manufacturing of ᴅ-phenylglycine. Enantioselective
synthesis of ᴅ-phenylglycine 1a in high ee and high
yield from very cheap hydrocarbon styrene 7a was
also demonstrated for the first time via one-pot 6-step
cascade biotransformation. E. coli (LZ116) strain
overexpressing 7 enzymes (SMO, SpEH, AlkJ,
EcALDH, SMDH, DpgAT and GluDH) in 3 plasmids
Genetic engineering of recombinant plasmids
Genetic engineering of plasmids containing dpgat-
gludh: The gene of DpgAT was synthesized and amplified
using primers “ACTGTCATGAGTATTCTGAACGACT
ACAAGC” and “ACTGAGATCTTTAGGATTGATTAC
CACTCAGGTTG” with Phusion DNA polymerase. The
PCR product was double-digested with BspHI and BglII,
and then ligated to the NcoI and BglII digested pRSFduet-
1 vector with T4 DNA ligase. The ligation product was
transformed into E. coli T7 Expression competent cells to
give pRSF-dpgat. The gene of GluDH was amplified from
7
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700956
Advanced Synthesis & Catalysis
the plasmid from previous research[21a] using primers
“ACTGAGATCTTAAGGAGATATATAATGGATCAG
ACATATTCTCTGGAGTC” and “ACTGCTCGAGTTA
AATCACACCCTGCGCCAGCATC” with Phusion DNA
polymerase. The PCR product was double-digested with
BglII and XhoI, and then ligated to the BglII and XhoI
digested pRSF-dpgat with T4 DNA ligase. The ligation
product was transformed into E. coli T7 Expression
competent cells to give pRSF-dpgat-gludh. The dpgat-
gludh was sub-cloned to the other three vectors by the
following procedure. dpgat-gludh was amplified with
primers “ACTGTCATGAGTATTCTGAACGACTACAA
GC” and “ACTGCTCGAGTTAAATCACACCCTGCGC
CAGCATC”, digested with BspHI and XhoI, and then
ligated to NcoI and XhoI digested pACYCduet-1,
pCDFduet-1, and pETduet-1. The transformation of these
products gave pACYC-dpgat-gludh, pCDF-dpgat-gludh,
and pET-dpgat-gludh, respectively.
pACYC-dpgat-gludh-smdh, pCDF-dpgat-gludh-smdh, and
pET-dpgat-gludh-smdh, respectively.
Other plasmids: The plasmids from previous research
were directly used in this study without modification: four
plasmids containing smo-speh (pACYC-smo-speh, pCDF-
smo-speh, pET-smo-speh, pRSF-smo-speh),[21a] four
plasmids containing alkj-ecaldh (pACYC-alkj-ecaldh,
pCDF-alkj-ecaldh,
pET-alkj-ecaldh,
pRSF-alkj-
ecaldh),[21a] and four plasmids containing PAD-PAL
(pACYC-pad-pal, pCDF-pad-pal, pET-pad-pal, pRSF-
pad-pal).[21b]
Genetic Engineering of recombinant E. coli strains
Engineering of E. coli DL39 (DE3) host: E. coli DL39
strain[20] with the genome type λ-, aspC13, fnr-25, rph-1,
ilvE12, tyrB507 was purchased from the Coli Genetic
Stock Center, Yale University (CGSC#: 6913). E. coli
DL39 (DE3) strain was engineered by using a λDE3
lysogenization kit (Merck) according to the standard
protocol of the kit with the following steps: 1. grow the E.
coli DL39 strain on LB supplemented with maltose,
MgSO4; 2. infect the cells with λDE3 phage lysate, helper
phage lysate and selection phage lysate; 3. transfer the
cells to normal LB plate to culture single colonies. 4.
verification of λDE3 lysogens with the T7 tester phage
lysate.
Genetic engineering of the plasmids containing smdh-
mr: SMDH was amplified from the genome of
Pseudomonas putida ATCC 12633[23] using primers
“ACTGTCATGAGCCAGAATCTCTTTA ACGTTG” and
“ACTGGAATTCTCATGCGTGTGTTCCTTTACCAAT
G” with Phusion DNA polymerase. The PCR product was
double-digested with BspHI and EcoRI, and then ligated to
the NcoI and EcoRI digested pRSFduet-1 vector with T4
DNA ligase. The ligation product was transformed into E.
coli T7 Expression competent cells to give pRSF-smdh.
The XhoI restriction site inside smdh was removed by the
following procedure. pRSF-smdh was subjected to
mutagenesis by cyclic PCR with primers “TCGCTTTAG
CTCGTGCAGCAACCAAGG” and “CCTTGGTTGCTG
CACGAGCTAAAGCGA”, then the PCR product was
digested with DpnI and transformed into E. coli T7
Expression competent cells to give pRSF-smdh(A288T).
The gene of MR was amplified using primers “ACTG
GAATTCTAAGGAGATATATAATGAGTGAAGTACT
GATTACCGGCC” and “ACTGCTCGAGTTACACCA
Engineering of E. coli (LZ101-112) strains: each of
smdh-mr plasmids (pACYC-smdh-mr, pCDF-smdh-mr,
pET-smdh-mr, and pRSF-smdh-mr) and each of dpgat-
gludh plasmids (pACYC-dpgat-gludh, pCDF-dpgat-gludh,
pET-dpgat-gludh, and pRSF-dpgat-gludh) were co-
transformed into competent cells of E. coli DL39 (DE3) to
give the 12 E. coli strains coexpressing SMDH, MR,
DpgAT and GluDH. The details were provided in Table S1
Supporting Information.
GATATTTCCCGATTTCT”
with
Phusion
DNA
polymerase. The PCR product was double-digested with
EcoRI and XhoI, and then ligated to the EcoRI and XhoI
digested pRSF-smdh(A288T) with T4 DNA ligase. The
ligation product was transformed into E. coli T7
Expression competent cells to give pRSF-smdh-mr. The
smdh-mr was sub-cloned to the other three vectors by the
following procedure. smdh-mr was amplified with primers
“ACTGTCATGAGCCAGAATCTCTTTAACGTTG” and
“ACTGCTCGAGTTACACCAGATATTTCCCGATTTC
T”, digested with BspHI and XhoI, and then ligated to
NcoI and XhoI digested pACYCduet-1, pCDFduet-1, and
pETduet-1. The transformation of these products gave
pACYC-smdh-mr, pCDF-smdh-mr, and pET-smdh-mr,
respectively.
Engineering of E. coli (LZ113-136) strains: each of smo-
speh plasmids (pACYC-smo-speh, pCDF-smo-speh, pET-
smo-speh, and pRSF-smo-speh), each of alkj-ecaldh
plasmids (pACYC-alkj-ecaldh, pCDF-alkj-ecaldh, pET-
alkj-ecaldh, and pRSF-alkj-ecaldh), and each of dpgat-
gludh-smdh plasmids (pACYC-dpgat-gludh-smdh, pCDF-
dpgat-gludh-smdh, pET-dpgat-gludh-smdh, and pRSF-
dpgat-gludh-smdh) were co-transformed into competent
cells of E. coli DL39 (DE3) to give the 24 E. coli strains
coexpressing SMO, SpEH, AlkJ, EcALDH, SMDH,
DpgAT and GluDH. The details were provided in Table S1
in the Supporting Information.
Engineering of E. coli (LZ137-160) strains: the E. coli
(LZ113-136) strains were cultured and made into
competent cells with CaCl2 treatment. One of pad-pal
plasmids (pACYC-pad-pal, pCDF-pad-pal, pET-pad-pal,
pRSF-pad-pal) was transformed into competent cells of E.
coli (LZ113-136) to give the 24 E. coli strains
coexpressing PAL, PAD, SMO, SpEH, AlkJ, EcALDH,
SMDH, DpgAT and GluDH. The details were provided in
Table S1 Supporting Information.
Genetic engineering of plasmids containing dpgat-
gludh-smdh: dpgat-gludh was amplified with primers
“ACTGTCATGAGTATTCTGAACGACTACAAGC” and
“ACTGGGTACCTTAAATCACACCCTGCGCCAGCAT
C”, digested with BspHI and KpnI, and then ligated to
NcoI and KpnI digested pRSFduet-1. The transformation
of these products gave pRSF-dpgat-gludh(KpnI) with KpnI
and XhoI restriction sites after the gludh. The gene of
SMDH was amplified from the pRSF-smdh(A288T) using
primers “ACTGGGTACCTAAGGAGATATATAATGAG
CCAGAATCTCTTTAACGTTG” and “ACTGCTCGAG
TCATGCGTGTGTTCCTTTACCAATG” with Phusion
DNA polymerase. The PCR product was double-digested
with KpnI and XhoI, and then ligated to the KpnI and XhoI
digested pRSF-dpgat-gludh(KpnI) with T4 DNA ligase.
The ligation product was transformed into E. coli T7
Expression competent cells to give pRSF-dpgat-gludh-
smdh. The dpgat-gludh-smdh was amplified with primers
“ACTGTCATGAGTATTCTGAACGACTACAAGC” and
“ACTGCTCGAGTCATGCGTGTGTTCCTTTACCAAT
G”, digested with BspHI and XhoI, and then ligated to
NcoI and XhoI digested pACYCduet-1, pCDFduet-1, and
pETduet-1. The transformation of these products gave
Procedure for culturing E. coli and biotransformation
General procedure for culturing E. coli cells: E. coli
strain (LZ101-160) was inoculated from the cell stock in
deep freezer to 1 mL LB medium containing a mixture of
antibiotic (50 mg L–1 chloramphenicol, 50 mg L–1
streptomycin, 100 mg L–1 ampicillin, 50 mg L–1
kanamycin) and grew at 37 °C for 8–10 h. The culture was
inoculated into 50 mL M9 medium containing glucose (20
g L–1), yeast extract (6 g L–1), and antibiotics in a 250-mL
tri-baffled flask. The cells continued to grow at 37 °C and
300 rpm for about 2 h to reach an OD600 of 0.6-0.8,
followed by the addition of IPTG to final concentration of
0.5 mM to induce the protein expression. The cells further
grew at 22 °C overnight (12–13 h) to reach late
8
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700956
Advanced Synthesis & Catalysis
exponential phase, and they were harvested by
centrifugation (4000 rpm, 10 min). The cell pellets were
resuspended in KP buffer to the desired density as resting
cells for biotransformation.
3e), 50 μL of EO samples were mixed with EtOAc
(containing 2 mM benzyl alcohol, 950 μL) to prepare the
samples for GC analysis.
Preparation of (2R)-amino(phenyl)acetic acid 1a from
rac-2-hydroxy-2-phenylacetic acid 3a and (2R)-
amino(4-fluorophenyl)acetic acid 1d from rac-(4-
fluorophenyl)(hydroxy)acetic acid 3d by cascade
biotransformation with E. coli (LZ110)
General procedure for biotransformation of rac-2-
hydroxy-2-phenylacetic
acid
3
to
(2R)-
amino(phenyl)acetic acid 1: freshly prepared E. coli
(LZ101-112) cells were resuspended in KP buffer (400
mM, pH 8.0) containing glucose (2%) and NH3/NH4Cl
(200 mM, pH 8.0). Stock solutions of 3a-3m (200 mM) in
the same KP buffer were prepared. Suspension of cells and
solution of 3a-3m were added into a 100-mL flask to form
a 2-mL system with cell density of 10-15 g cdw L−1 and
substrate concentration of 10-30 mM. The reaction mixture
was incubated at 30 °C and 250 rpm. For the time curve of
(R)-1a production (Figure 1e), 0.06 mmol of rac-3a was
added at 0, 2, 4, 6, 8, 10, and 12 h. Additional 0.5% of
glucose and 100 mM of NH3/NH4Cl was fed at 2, 4, 6, 8,
10, and 12 h. 50 μL of reaction samples were taken at 2, 4,
6, 8, 10, 12, and 24 h. For the screening of the best strain
(Figure 1c), 100 μL of reaction samples were taken at 2 h.
For the testing of substrate scope (Table 1), 100 μL of
reaction samples were taken at 24 h. The reaction samples
were mixed with equal volume of 1 M HCl solution (50-
100 μL) and subjected to centrifugation (12000 rpm, 3
min). The supernatant (50-100 μL) was mixed with TFA
solution (0.5%, 450-400 μL) and acetonitrile (containing 2
mM benzyl alcohol, 500 μL) to prepare the samples for
HPLC analysis.
E. coli (LZ110) cells were resuspended in 40 mL of KP
buffer (400 mM, pH 8.0) containing glucose (2%) and
NH3/NH4Cl (200 mM, pH 8.0) to reach a final density of
20 g cdw L−1 in a 250-mL flask with baffles. The reaction
mixture was incubated at 30 °C and 250 rpm with the
following substrate feeding procedure: for the synthesis of
(R)-1a, 1.2 mmol of rac-3a was added at 0, 2, 4, 6, 8, and
10 h, and 0.8 mmol) of rac-3a was added at 12 h (total 8
mmol, 1217 mg, of rac-3a added). Additional 0.5% (w/v)
of glucose and 100 mM of NH3/NH4Cl was fed at 2, 4, 6, 8,
10, and 12 h. For the synthesis of (R)-1d, 1.2 mmol of rac-
3d was added at 0 and 6 h (total 2.4 mmol, 408 mg, of rac-
3d added). Additional 1% of glucose and 100 mM of
NH3/NH4Cl was fed at 6 h. The reaction proceeded for 24
h, and then HCl was added to reach pH of 1 to dissolve the
precipitated (R)-1a and (R)-1d. The mixture was subjected
to centrifugation (3200 g, 15 min), and the clear aqueous
solution was collected, while the cells were washed with
HCl solution (0.01 M). The aqueous solution was
combined and neutralized to pH = 7 with NaOH solution
(10 M). Then, the aqueous mixture was concentrated to
about 10 mL with rotary evaporator. The precipitated
products were separated by filtration and carefully washed
with cold water and ethanol. The products were dried
under vacuum overnight.
General procedure for biotransformation of styrene 7
to (2R)-amino(phenyl)acetic acid 1: freshly prepared E.
coli (LZ113-136) cells were resuspended in KP buffer (400
mM, pH 8.0) containing glucose (0.5%) and NH3/NH4Cl
(100 mM, pH 8.0). Stock solutions of 7a-7m (50-150 mM)
in EO were prepared. Suspension of cells was added into a
100-mL flask to form a 2-mL system with cell density of
10-15 g cdw L−1, and 0.4 mL of EO containing 7a-7m was
added to reach a substrate concentration of 10-30 mM. The
reaction mixture was incubated at 30 °C and 250 rpm.
Additional glucose (1%) and NH3/NH4Cl (100 mM) was
added at 10 h. For the time curve of (R)-1a production
(Figure 2e), 100 μL of reaction samples were taken at 2, 5,
9, and 24 h. For the screening of the best strain (Figure 2c)
and testing of substrate scope (Table 2), 100 μL of reaction
samples were taken at 24 h. The reaction samples were
mixed with equal volume of 1 M HCl solution (100 μL)
and subjected to centrifugation (12000 rpm, 3 min). The
aqueous solution (100 μL) was mixed with TFA solution
(0.5%, 400 μL) and acetonitrile (containing 2 mM benzyl
alcohol, 500 μL) to prepare the samples for HPLC analysis.
For the time curve of (R)-1a production (Figure 2e), 50 μL
of EO samples were mixed with EtOAc (containing 2 mM
benzyl alcohol, 950 μL) to prepare the samples for GC
analysis.
ᴅ-Phenylglycine 1a was isolated as white solid (859 mg,
1
71% yield) in 99% ee. H NMR (400 MHz, in D2O with
2% H2SO4): δ = 7.40–7.35 (m, 5H), 5.07 (s, 1H) ppm. 13C
NMR (100 MHz, in D2O with 2% H2SO4): δ = 170.7,
131.4, 130.3, 129.7, 128.0, 56.5 ppm. [α]D20: −152° (c 1.0,
1 M HCl), {reported [α]D20: −157° (c 1.0, 1 M HCl)[27a]}.
ᴅ-4-Fluorophenylglycine 1d was isolated as white solid
1
(211 mg, 52% yield) in 99% ee. H NMR (400 MHz, in
D2O with 2% H2SO4): δ = 7.39–7.35 (m, 2H), 7.12–7.08
(m, 2H), 5.09 (s, 1H) ppm. 13C NMR (100 MHz, in D2O
with 2% H2SO4): δ = 170.6, 163.4 (d, J = 245.9 Hz), 130.4
(d, J = 9.0 Hz), 127.5 (d, J = 3.4 Hz), 116.5 (d, J = 22.0
Hz), 55.8 ppm. [α]D20: −137° (c 1.0, 1 M HCl), {reported
[α]D22: −135° (c 0.5, 1 M HCl) 98% ee[27b]}.
Preparation of (2R)-amino(phenyl)acetic acid 1a from
styrene 7a and (2R)-amino(4-fluorophenyl)acetic acid
1d from 1-fluoro-4-vinylbenzene 7d by cascade
biotransformation with E. coli (LZ116)
E. coli (LZ116) cells were resuspended in 100 mL of KP
buffer (400 mM, pH 8.0) containing glucose (0.5%) and
NH3/NH4Cl (100 mM, pH 8.0) to reach a final density of
20 g cdw L−1 in a 500-mL flask with baffles. 20 mL of
ethyl oleate (EO) were added to form the second phase. 6
mmol (625 mg) of 7a or 3 mmol (366 mg) of 7d was added
into the system to start the reaction at 30 °C and 250 rpm
for 24 h. Additional 1% of glucose and 200 mM of
NH3/NH4Cl was added at 10 h. At the end of the reaction,
the reaction mixture was acidified with HCl to pH = 1, and
then subjected to centrifugation (3200 g, 15 min). The cells
were washed with HCl solution (0.01 M). The aqueous
solution was combined and washed with cyclohexane (20
mL) to remove the hydrophobic impurities. The aqueous
solution was neutralized to pH = 7 with NaOH solution (10
M), and concentrated to about 20 mL with rotary
evaporator. The precipitated products were separated by
filtration and carefully washed with cold water and ethanol.
The products were dried under vacuum overnight.
General procedure for biotransformation of (S)-2-
amino-3-phenylpropanoic
acid
9a
to
(2R)-
amino(phenyl)acetic acid 1a: freshly prepared E. coli
(LZ137-160) cells were resuspended in KP buffer (400
mM, pH 8.0) containing glucose (0.5%). Stock solutions of
(S)-9a (150 mM) in the same KP buffer was prepared.
Suspension of cells and solution of (S)-9a were added into
a 100-mL flask to form a 2-mL system with cell density of
10-15 g cdw L−1 and substrate concentration of 30-60 mM.
And 0.4 mL of EO was added to the system. The reaction
mixture was incubated at 30 °C and 250 rpm. Additional
glucose (2%) and NH3/NH4Cl (200 mM) was added at 10 h.
For the time curve of (R)-1a production (Figure 3e), 100
μL of reaction samples were taken at 2, 5, 10, and 24 h.
For the screening of the best strain (Figure 3c), 100 μL of
reaction samples were taken at 24 h. The reaction samples
were mixed with equal volume of 1 M HCl solution (100
μL) and subjected to centrifugation (12000 rpm, 3 min).
The aqueous solution (100 μL) was mixed with TFA
solution (0.5%, 400 μL) and acetonitrile (containing 2 mM
benzyl alcohol, 500 μL) to prepare the samples for HPLC
analysis. For the time curve of (R)-1a production (Figure
ᴅ-Phenylglycine 1a was isolated as white solid (562 mg,
1
62% yield) in 99% ee. H NMR (400 MHz, in D2O with
2% H2SO4): δ = 7.41–7.36 (m, 5H), 5.08 (s, 1H) ppm. 13C
NMR (100 MHz, in D2O with 2% H2SO4): δ = 170.8,
9
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700956
Advanced Synthesis & Catalysis
131.4, 130.3, 129.7, 128.0, 56.5 ppm. [α]D20: −150° (c 1.0,
1 M HCl), {reported [α]D20: −157° (c 1.0, 1 M HCl)[27a]}.
ᴅ-4-Fluorophenylglycine 1d was isolated as white solid
Chem. 1999, 42, 4981–5001; b) S. Yan, G. Larson, J. Z.
Wu, T. Appleby, Y. Ding, R. Hamatake, Z. Hong, N.
Yao, Bioorg. Med. Chem. Lett. 2007, 17, 63–67.
1
(233 mg, 46% yield) in 99% ee. H NMR (400 MHz, in
D2O with 2% H2SO4): δ = 7.39–7.36 (m, 2H), 7.13–7.09
(m, 2H), 5.09 (s, 1H) ppm. 13C NMR (100 MHz, in D2O
with 2% H2SO4): δ = 170.6, 163.4 (d, J = 245.7 Hz), 130.4
(d, J = 8.9 Hz), 127.4 (d, J = 3.0 Hz), 116.6 (d, J = 22.1
Hz), 55.8 ppm. [α]D20: −138° (c 1.0, 1 M HCl), {reported
[α]D22: −135° (c 0.5, 1 M HCl) 98% ee[27b]}.
[5] a) C. Dockendorff, M. M. Nagiec, M. Weïwer, S.
Buhrlage, A. Ting, P. P. Nag, A. Germain, H. J. Kim,
W. Youngsaye, C. Scherer, M. Bennion, L. Xue, B. Z.
Stanton, T. A. Lewis, L. MacPherson, M. Palmer, M. A.
Foley, J. R. Perez, S. L. Schreiber, ACS Med. Chem.
Lett. 2012, 3, 808–813; b) A. Z. Gonzalez, J.
Eksterowicz, M. D. Bartberger, H. P. Beck, J. Canon, A.
Chen, D Chow, J. Duquette, B. M. Fox, J. Fu, X.
Huang, J. B. Houze, L. Jin, Y. Li, Z. Li, Y. Ling, M. C.
Lo, A. M. Long, L. R. McGee, J. McIntosh, D. L.
McMinn, J. D. Oliner, T. Osgood, Y. Rew, A. Y. Saiki,
P. Shaffer, S. Wortman, P. Yakowec, X. Yan, Q. Ye, D.
Yu, X. Zhao, J. Zhou, S. H. Olson, J. C. Medina, D.
Sun, J. Med. Chem. 2014, 57, 2472–2488.
Preparation of (2R)-amino(phenyl)acetic acid 1a from
(S)-2-amino-3-phenylpropanoic acid 9a by cascade
biotransformation with E. coli (LZ143)
E. coli (LZ143) cells were resuspended in 80 mL of KP
buffer (400 mM, pH 8.0) containing glucose (0.5%) to
reach a final density of 20 g cdw L−1 in a 500-mL flask
with baffles. 16 mL of EO was added to form the second
phase. 4.8 mmol (793 mg) of (S)-9a was added into the
system to start the reaction at 30 °C and 250 rpm.
Additional 2% of glucose and 200 mM of NH3/NH4Cl was
added at 10 h. The reaction was stopped at 24 h. The same
downstream processing to the preparation of (R)-1a from
7a was carried out.
[6] R. A. Sheldon, J. Chem. Tech. Biotechnol. 1996, 67, 1–
14.
ᴅ-Phenylglycine 1a was isolated as white solid (385 mg,
1
[7] C. Hongo, R. Yoshioka, M. Tohyama, S. Yamada, I.
53% yield) in 99% ee. H NMR (400 MHz, in D2O with
2% H2SO4): δ = 7.40–7.35 (m, 5H), 5.07 (s, 1H) ppm. 13C
NMR (100 MHz, in D2O with 2% H2SO4): δ = 170.8,
131.4, 130.3, 129.7, 128.0, 56.5 ppm. [α]D20: −151° (c 1.0,
1 M HCl), {reported [α]D20: −157° (c 1.0, 1 M HCl)[27a]}.
Chibata, Bull. Chem. Soc. Jpn. 1983, 56, 3744–3747
[8] a) J. Wang, X. Liu, X. Feng, Chem. Rev. 2011, 111,
6947–6983; b) S. J. Zuend, M. P. Coughlin, M. P.
Lalonde, E. N. Jacobsen, Nature 2009, 461, 968–970.
Contents in supporting information
[9] a) U. T. Bornscheuer, G. W. Huisman, R. J. Kazlauskas,
S. Lutz, J. C. Moore, K. Robins, Nature 2012, 485,
185–194; b) M. T. Reetz, J. Am. Chem. Soc. 2013, 135,
12480–12496; c) G. W. Huisman, S. J. Collier, Curr.
Opin. Chem. Biol. 2013, 17, 284–292; d) B. M. Nestl, S.
C. Hammer, B. A. Nebel, B. Hauer, Angew. Chem.
2014, 126, 3132–3158; Angew. Chem. Int. Ed. 2014, 53,
3070–3095; e) Biocatalysis in Organic Synthesis (Eds.:
K. Faber, W.-D. Fessner, N. Turner), Thieme, Stuttgart,
2014.
Chiral HPLC chromatograms (Figure S4-S16), NMR
spectra (Figure S17-S21), list of strains (Table S1),
chemicals, biochemicals, medium, analytical methods,
source and sequence of genes.
Acknowledgements
This work was financially supported by GlaxoSmithKline (GSK)
and Singapore Economic Development Board (EDB) through a
Green and Sustainable Manufacturing Grant (Project No. 279-
000-348-592) and National University of Singapore through
Synthetic Biology for Clinical and Technological Innovation
(SynCTI) Program.
[10] a) G. D. Machado, M. Gomes, O. A. C. Antunes, E. G.
Oestreicher, Process Biochem. 2005, 40, 3186–3189; b)
W. Y. Lou, M. H. Zong, Y. Y. Liu, J. F. Wang, J.
Biotechnol. 2006, 125, 64–74.
References
[11] M. A. Wegman, M. A. P. J. Hacking, J. Rops, P.
Pereira, F. Van Rantwijk, R. A. Sheldon, Tetrahedron:
Asymmetry 1999, 10, 1739–1750; (b) K. Yasukawa, Y.
Asano, Adv. Synth. Catal. 2012, 354, 3327–3332; (c) J.
Qiu, E. Su, W. Wang, D. Wei, Tetrahedron Lett.
2014, 55, 1448–1451.
[1] a) V. A. Soloshonok, K. Izawa in Asymmetric Synthesis
and Application of α-Amino Acids, American Chemical
Society, Washington DC, 2009; b) C. Najera, J. M.
Sansano, Chem. Rev. 2007, 107, 4584–4671.
[2] a) M. A. Wegman, M. H. A. Janssen, F. van Rantwijk,
R. A. Sheldon, Adv. Synth. Catal. 2001, 343, 559–576;
A. Schmid, F. Hollmann, J. B. Park, B. Bühler, Curr.
Opin. Biotechnol. 2002, 13, 359–366.
[12] a) M. Breuer, K. Ditrich, T. Habicher, B. Hauer, M.
Kesseler, R. Stürmer, T. Zelinski, Angew. Chem. 2004,
116, 806–843; Angew. Chem. Int. Ed. 2004, 43, 788–
824; (b) A. I. Martínez-Gómez, S. Martínez-Rodríguez,
J. M. Clemente-Jiménez, J. Pozo-Dengra, F.
Rodríguez-Vico, F. J. Las Heras-Vázquez, Appl.
Environ. Microbiol. 2007, 73, 1525–1531.
[3] a) I. H. Bae, J. K. Choi, C. Chough, S. J. Keum, H.
Kim, S. K. Jang, B. M. Kim, ACS Med. Chem. Lett.
2014, 5, 255–258; b) S. Daniele, S. Sestito, D.
Pietrobono, C. Giacomelli, G. Chiellini, D. Di Maio, L.
Marinelli, E. Novellino, C. Martini, S. Rapposelli, ACS
Chem. Neurosci. 2017, 8, 100–114.
[13] K. Jariyachawalid, P. Laowanapiban, V. Meevootisom,
S. Wiyakrutta, Microb. Cell Fact. 2012, 11, 47.
[14] E. S. Park, J. Y. Dong, J. S. Shin, ChemCatChem
2013, 5, 3538–3542.
[4] a) M. G. Russell, V. G. Matassa, R. R. Pengilley, M. B.
van Niel, B. Sohal, A. P. Watt, L. Hitzel, M. S. Beer, J.
A. Stanton, H. B. Broughton, J. L. Castro, J. Med.
[15] U. Müller, F. Van Assema, M. Gunsior, S. Orf, S.
Kremer, D. Schipper, A. Wagemans, C. A. Townsend,
10
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700956
Advanced Synthesis & Catalysis
T. Sonke, R. Bovenberg, M. Wubbolts, Metab. Eng.
2006, 8, 196–208.
[23] A. Y. Tsou, S. C. Ransom, J. A. Gerlt, D. D. Buechter,
P. C. Babbitt, G. L. Kenyon, Biochemistry 1990, 29,
9856–9862.
[16] a) E. Ricca, B. Brucher, J. H. Schrittwieser, Adv.
Synth. Catal. 2011, 353, 2239–2262; b) J. Muschiol, C.
Peters, N. Oberleitner, M. Mihovilovic, U. Bornscheuer,
F. Rudroff, Chem. Commun. 2015, 51, 5798–5811; c)
V. Köhler, N. J. Turner, Chem. Commun. 2015, 51,
450–464; d) T. Bayer, S. Milker, T. Wiesinger, F.
Rudroff, M. D. Mihovilovic, Adv. Synth. Catal. 2015,
357, 1587–1618; e) S. Wu, J. Liu, Z. Li, Synlett 2016,
27, 2644–2658; f) S. P. France, L. J. Hepworth, N. J.
Turner, S. L. Flitsch, ACS Catal. 2017, 7, 710–724; g) J.
H. Schrittwieser, S. Velikogne, M. Hall, W. Kroutil,
Chem. Rev. 2017, DOI: 10.1021/acs.chemrev.7b00033.
[24] a) J. H. Sattler, M. Fuchs, K. Tauber, F. G. Mutti, K.
Faber, J. Pfeffer, T. Haas, W. Kroutil, Angew. Chem.
2012, 124, 9290–9293; Angew. Chem. Int. Ed. 2012, 51,
9156–9159; b) K. Tauber, M. Fuchs, J. H. Sattler, J.
Pitzer, D. Pressnitz, D. Koszelewski, K. Faber, J.
Pfeffer, T. Haas, W. Kroutil, Chem. Eur. J. 2013, 19,
4030–4035; c) F. G. Mutti, T. Knaus, N. S. Scrutton, M.
Breuer, N. J. Turner, Science 2015, 349, 1525–1529; d)
F. F. Chen, Y. Y. Liu, G. W. Zheng, J. H. Xu,
ChemCatChem 2015, 7, 3838–3841.
[25] D. Kuhn, M. K. Julsing, E. Heinzle, B. Bühler, Green
Chem. 2012, 14, 645–653.
[17] a) V. Resch, W. M. Fabian, W. Kroutil, Adv. Synth.
Catal. 2010, 352, 993–997; b) C. W. Fan, G. C. Xu, B.
D. Ma, Y. P. Bai, J. Zhang, J. H. Xu, J. Biotechnol.
2015, 195, 67–71.
[26] a) U. Biermann, U. Bornscheuer, M. A. Meier, J. O.
Metzger, H. J. Schäfer, Angew. Chem. 2011, 123,
3938–3956; Angew. Chem. Int. Ed. 2011, 50, 3854–
3871; b) P. N. R. Vennestrøm, C. M. Osmundsen, C. H.
Christensen, E. Taarning, Angew. Chem. 2011, 123,
10686–10694; Angew. Chem. Int. Ed. 2011, 50, 10502–
10509; c) C. O. Tuck, E. Pérez, I. T. Horváth, R. A.
Sheldon, M. Poliakoff, Science 2012, 337, 695–699.
[18] K. Vedha-Peters, M. Gunawardana, J. D. Rozzell, S. J.
Novick, J. Am. Chem. Soc. 2006, 128, 10923–10929.
[19] S. Wiyakrutta, V. Meevootisom, J. Biotechnol. 1997,
55, 193–203.
[20] D. M. LeMaster, F. M. Richards, Biochemistry 1988,
27, 142–150.
[27] a) H. Kunz, W. Pfrengle, J. Am. Chem. Soc. 1988, 110,
651–652; b) J. M. Vernier, L. S. Hegedus, D. B. Miller,
J. Org. Chem. 1992, 57, 6914–6920.
[21] a) S. Wu, Y. Zhou, T. Wang, H. P. Too, D. I. C. Wang,
Z. Li, Nat. Commun. 2016, 7, 11917; b) Y. Zhou, S.
Wu, Z. Li, Angew. Chem. 2016, 128, 11819–11822;
Angew. Chem. Int. Ed. 2016, 55, 11647–11650.
[22] B. Mitra, J. A. Gerlt, P. C. Babbitt, C. W. Koo, G. L.
Kenyon, D. Joseph, G. A. Petsko, Biochemistry
1993, 32, 12959–12967.
11
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700956
Advanced Synthesis & Catalysis
FULL PAPER
One-Pot Enantioselective Synthesis of ᴅ-
Phenylglycines from Racemic Mandelic acids,
Styrenes, or Biobased ʟ-Phenylalanine via Cascade
Biocatalysis
Adv. Synth. Catal. Year, Volume, Page – Page
Yi Zhou, Shuke Wu, and Zhi Li*
12
This article is protected by copyright. All rights reserved.