Exploring the scope of an α/β-aminomutase for the amination of cinnamate epoxides to arylserines and arylisoserines

Article abstract of DOI:10.1021/acscatal.9b01557

Biocatalytic process-development continues to advance toward discovering alternative transformation reactions to synthesize fine chemicals. Here, a 5-methylidene-3,5-dihydro-4H-imidazol-4-one (MIO)-dependent phenylalanine aminomutase from Taxus canadensis (TcPAM) was repurposed to irreversibly biocatalyze an intermolecular amine transfer reaction that converted ring-substituted trans-cinnamate epoxide racemates to their corresponding arylserines. From among 12 substrates, the aminomutase ring-opened 3′-Cl-cinnamate epoxide to 3′-Cl-phenylserine 140 times faster than it opened the 4′-Cl-isomer, which was turned over slowest among all epoxides tested. GC/MS analysis of chiral auxiliary derivatives of the biocatalyzed phenylserine analogues showed that the TcPAM-transamination reaction opened the epoxides enantio- A nd diastereoselectively. Each product mixture contained (2S)+(2R)-anti (erythro) and (2S)+(2R)-syn (threo) pairs with the anti-isomers predominating (-90:10 dr). Integrating the vicinal proton signals in the ¹H NMR spectrum of the enzyme-catalyzed phenylserines and calculating the chemical shift difference (?"?) between the anti and syn proton signals confirmed the diastereomeric ratios and relative stereochemistries. Application of a (2S)-threonine aldolase from E. coli further established the absolute stereochemistry of the chiral derivatives of the diastereomeric enzymatically derived products. The 2R:2S ratio for the biocatalyzed anti-isomers was highest (88:12) for 3′-NO₂-phenylserine and lowest (66:34) for 4′-F-phenylserine. This showed that the stereospecificity of TcPAM is in part directed by the substituent-type on the cinnamate epoxide analogue. The catalyst also converted each cinnamate epoxide analogue to its corresponding isoserine, highlighting a biocatalytic route to arylisoserines, which play a key role in building the pharmacophore seen in anticancer and protease inhibitor drugs.

Full text of DOI:10.1021/acscatal.9b01557

Subscriber access provided by Nottingham Trent University
Article
Exploring the Scope of an #/#-Aminomutase for the Amination
of Cinnamate Epoxides to Arylserines and Arylisoserines
Prakash Kumar Shee, Nishanka Dilini Ratnayake, Tyler
Walter, Olivia Goethe, Edith N Onyeozili, and Kevin D Walker
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.9b01557 • Publication Date (Web): 09 Jul 2019
Downloaded from pubs.acs.org on July 12, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 43
ACS Catalysis
1
2
3
4
Exploring the Scope of an α/β-Aminomutase for the Amination of Cinnamate Epoxides to
5
6
Arylserines and Arylisoserines
7
8
9
Prakash K. Shee,† Nishanka Dilini Ratnayake,† Tyler Walter,† Olivia Goethe,† Edith Ndubuaku
Onyeozili,§ and Kevin D. Walker*,†,‡
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
† Department of Chemistry, Michigan State University, East Lansing, Michigan 48824, United
States
‡ Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing,
Michigan, 48824, United States
§ Department of Chemistry, Florida Agricultural & Mechanical University, Tallahassee, Florida
32307, United States
1
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 43
1
2
3
4
5
6
7
8
9
Abstract: Biocatalytic process-development continues to advance toward discovering alternative
transformation reactions to synthesize fine chemicals. Here, a 5-methylidene-3,5-dihydro-4H-
imidazol-4-one (MIO)-dependent phenylalanine aminomutase from Taxus canadensis (TcPAM)
was repurposed to irreversibly biocatalyze an intermolecular amine transfer reaction that converted
ring-substituted trans-cinnamate epoxide racemates to their corresponding arylserines. From
among twelve substrates, the aminomutase ring-opened 3'-Cl-cinnamate epoxide to 3'-Cl-
phenylserine 140 times faster than it opened the 4'-Cl-isomer, which was turned over slowest
among all epoxides tested. GC/MS analysis of chiral auxiliary derivatives of the biocatalyzed
phenylserine analogues showed that the TcPAM-transamination reaction opened the epoxides
enantio- and diastereoselectively. Each product mixture contained (2S)+(2R)-anti (erythro) and
(2S)+(2R)-syn (threo) pairs with the anti-isomers predominating (~90:10 dr). Integrating the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
vicinal proton signals in the H-NMR spectrum of the enzyme-catalyzed phenylserines and
calculating the chemical shift difference (Δδ) between the anti and syn proton signals confirmed
the diastereomeric ratios and relative stereochemistries. Application of a (2S)-threonine aldolase
from E. coli further established the absolute stereochemistry of the chiral derivatives of the
diastereomeric enzymatically derived products. The 2R:2S ratio for the biocatalyzed anti-isomers
was highest (88:12) for 3'-NO2-phenylserine and lowest (66:34) for 4'-F-phenylserine. This
showed that the stereospecificity of TcPAM is in part directed by the substituent-type on the
cinnamate epoxide analogue. The catalyst also converted each cinnamate epoxide analogue to its
corresponding isoserine, highlighting a biocatalytic route to arylisoserines, which play a key role
in building the pharmacophore seen in anticancer and protease inhibitor drugs.
Keywords: biocatalysis, MIO-aminomutase, phenylisoserine, phenylserine, cinnamate epoxides,
enzyme catalysis
2
ACS Paragon Plus Environment

Page 3 of 43
ACS Catalysis
1
2
3
4
INTRODUCTION
5
6
7
8
Nonproteinogenic amino acids are a specialized class of organic compounds that often have
intrinsic biological activity.1-4 More frequently, these vital amino acids are found in peptides with
antiviral,5 antitumor,6 anti-inflammatory,7 or immunosuppressive activities.8 β-Hydroxy-α-amino
acids reside in an important subclass of nonproteinogenic amino acids, including arylserines,
hydroxyaspartic acid, and hydroxyleucine. These bifunctional compounds introduce two chiral
centers when used as biomolecular building blocks. Incorporation of unique stereochemistry into
the residues of peptide chains can change their properties by, for example, increasing the stability
against peptidases and prolonging bioavailability.9-11 Further, β-hydroxy-α-amino acid scaffolds
(Figure 1) are found in clinically important antibiotics, including glycopeptide vancomycin and its
analogues, ristocetin from Amycolatopsis12-13 and teicoplanins from Actinoplanes,14 the
phenylpropanoid chloramphenicol from Streptomyces,15 and katanosin depsipeptides from
Cytophaga and Lysobacter bacteria.16-17 The pentapeptide gymnangiamide from marine hydroid
Gymnangium regae shows anticancer activity.18 (2S,3R)-3,4-Dihydroxyphenylserine (DOPS,
Droxidopa) is used for hypotension and as a Parkinsonian therapeutic,19-21 while N-arylsulfonyl
derivatives of phenylserine ethyl esters function as non-steroidal anti-inflammatory drugs.22
β‐Hydroxy‐α‐amino acids have also been used recently in protein synthesis by chemical ligation
at N-terminal serine and threonine sites, and this technique can likely be expanded to incorporate
other non‐proteinogenic β‐hydroxy amino acids.23 Additionally, an interesting subclass of non-
natural heterocyclic β-hydroxy-α-amino acid amides are drug leads for their analgesic and
immunostimulant activities.24-27
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 43
1
2
3
4
5
6
7
8
9
HO
OH
H2N
H
NH
O
N
OH
Cl
O
NH
O
O
O
O
O
HN
O
O
O
HO
O
OH
HO
O
OH
O
Cl
O
O
HO
H
H
N
O
NH
O
N
O
N
N
N
HO
NH
O
O
O
O
H
H
H
H
H
HN
O
O
HN
COOH
H2N
N
N
N
H
N
N
H2N
HN
HO
O
NH
N
H
O
O
NH
O
H
OH
HN
NH
NH2
H2N
OH
OH
N
H
HO
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Gymnangiamide [(2S)-syn]
Vancomycin [(2R)-anti]
Katanosin [(2S)-syn]
(anticancer)
(antibiotic)
(antibiotic)
OH
O
OH OH
OH
O
Cl
O
HO
HO
O
O
HN
O
HN
N
O
Cl
N
O
S
O
Br
NH2
O
O
Dihydroxyphenylserine (DOPS)
[(2S)-syn]
Chloramphenicol [(2S)-syn]
N-(p-Br-Benzenesulfonyl)-syn-
4'-nitrophenylserine racemate
(analgesic)
(antibiotic)
(neurotransmitter prodrug)
Figure 1. Synthetic and biosynthetic β-phenyl-β-hydroxy-α-amino acid building blocks in bioactive
compounds. Stereochemical designations in brackets are described in Figure 2.
Given the beneficial pharmacological, chemical, and physical properties of a compound
containing β-hydroxy-α-amino acids, there is considerable interest to make these bifunctional
scaffolds by synthetic approaches or combined with biocatalytic routes.27-29 Various synthetic
strategies have been used to control the regio- and stereochemistry of β-hydroxy-α-amino acids,
including asymmetric aldol condensation, oxy-Michael addition, electrophilic amination,
aminohydroxylation of alkenes, and aza-Claisen rearrangement.19, 30-33 Several of these methods
incorporate protecting group manipulations to direct regiochemistry, add a functionalized chiral
auxiliary to incorporate stereogenic centers, or apply heavy-metal catalysts to promote the
reactions. These synthetic approaches are often highly efficient and stereoselective, but chemical
manufacturers observe that these routes often violate green chemistry principles through
generation of heavy-metal waste, poor atom economy built on the synthesis of intricate chiral-
ligand catalysts, and frequently fall short as sustainable methods.34-35
4
ACS Paragon Plus Environment

Page 5 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
OH
O
OH
O
2

R
R
O
R
O
NH3
NH3
(2S)-syn
[L-threo]
(2R)-syn
[D-threo]
OH
O
OH
O
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
R
O
NH3
NH3
(2S)-anti
[L-erythro]
(2R)-anti
[D-erythro]
Figure 2. Stereoisomerism convention for β-
hydroxy-α-amino acids used herein. The
equivalent, archaic designations are listed in
brackets.
Biocatalysis is emerging as a valuable complementary tool for organic chemists to access regio- and
stereocontrolled chemical transformations that otherwise use complex synthetic chiral ligands and often
proceed through multiple steps in conventional methods.35-38 Biocatalytic routes towards (2S)+(2R)-
(syn/anti)-β-hydroxy-α-amino acids39 (Figure 2) mixtures regularly use threonine aldolases (TAs).
TAs are divided into two groups depending on the (R)- or (S)-stereochemistry at the α-carbon of
threonine where the amino group is attached. While the substrate specificity of TAs depends on
the glycine donor substrate, they have broad specificity for the aldehyde, including non-natural
aryl aldehydes when used to produce arylserines.40 TAs are divided further into four subgroups,
depending on the stereochemistry at the β-carbon of threonine, where the hydroxyl group is
attached. High-specificity (2S)-syn-TAs make only (2S)-syn-threonine analogues, (2S)-anti-TAs
are stereoselective for (2S)-anti-threonine analogues, while low-specificity (2S)-TAs make a
mixture of (2S)-syn- and (2S)-anti-threonine analogues. Only low-specificity (2R)-TAs are known
and make mixtures of (2R)-syn- and (2R)-anti-threonine analogues.29
TA specificity is also used in industrial chemical processes for enantiospecific enzyme
resolution to cleave (via a retro-aldol reaction) one enantiomer of a syn- or anti-β-hydroxy-α-
amino acid pair.29 This process stereospecifically enriches one stereoisomer over the other in a
5
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 43
1
2
3
4
5
6
7
8
9
mixture. For example, a low specificity (2R)-TA stereospecifically catalyzed the retro-aldol
cleavage of the (2R)-isomer in a (2S)+(2R)-syn-racemate to resolve the intact (2S)-syn-(3',4'-
methylenedioxy)phenylserine enantiomer, a precursor of the therapeutic drug (2S)-syn-DOPS (L-
DOPS).41
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Here, we used an irreversible, TA-independent biocatalytic method employing a repurposed 5-
methylidene-3,5-dihydro-4H-imidazol-4-one (MIO)-dependent phenylalanine aminomutase
(TcPAM) from Taxus canadensis plants42-44 to make various ring-substituted phenylserines. In
Taxus plants, TcPAM isomerizes (2S)-α-phenylalanine to (3R)-β-phenylalanine on the
biosynthetic pathway of the anticancer therapeutic paclitaxel (Taxol®) (Scheme 1).42-43 The MIO
moiety is formed post-translationally by active site residues within TcPAM, and this electrophilic
unit transiently becomes N-alkylated by the (2S)-α-phenylalanine substrate. The resulting alkyl
ammonium group is displaced as an NH2-MIO adduct with simultaneous removal of Hβ.45 The
cinnamate intermediate accepts the amino group from the NH2-MIO adduct at Cβ, and Cα is
reprotonated to complete the α- to β-isomerization (Scheme 1).
6
ACS Paragon Plus Environment

Page 7 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
Enz
Tyr
Enz
Tyr
H
H
O
O
H
O
H
N
H
H
O
O
H
O
H
NH2
O
O
NH2
O
NH2
H-Enz
OH
OH
N
Enz
MIO
Enz
N
N
N
N
9
α-Phenylalanine
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
rotation
Enz
Tyr
H
Enz
Tyr
O
O
H
O
H
O
H
O
O
O
H
H
O
3R
H
H
NH2
NH2
NH2
O
OH
O
H
H-Enz
Enz
Enz
N
N
N
N
N
N
β-Phenylalanine
(99.9% ee)
O
Ser
Ala
O
O
OH
OH
O
H
N
N
O
H
O
O
NH
O
N
N
HN
O
O
H
N
H
2 H2O
O
HO
O
O
O
OH
O
NH
O
NH
MIO
Gly
Paclitaxel (Taxol)
Scheme 1. Mechanism of β-phenylalanine formation
catalyzed by TcPAM. Inset left: Anticancer drug
paclitaxel (Taxol). Inset right: Formation of MIO from
the condensation of Ala175, Ser176, and Gly177 of
TcPAM.
An earlier study covalently trapped a cinnamate epoxide in the active site of an MIO tyrosine
aminomutase (SgTAM) crystal to help dissect its enzyme mechanism.46 The ability of SgTAM to
trap an epoxide hinted that the MIO chemistry would be aborted by covalent inhibition. It is
interesting to note that the crystal structures of a phenylalanine aminomutase from Pantoea
agglomerans (PaPAM) in our subsequent study showed dual occupancy of α- and β-
phenylalanines (substrate and product, respectively) covalently linked to the MIO moiety through
their amino groups.47 Thus, we reinterpreted the structures of the previous epoxide inhibition study
to represent β-hydroxy-α-amino adducts, rather than a dihydroxy ether intermediate attached
covalently to the MIO group.46
Later, in a burst phase kinetics study, TcPAM was shown to transfer the amino group from a
surrogate substrate (2S)-styryl-α-alanine to exogenously supplied cinnamates via an NH2-MIO
7
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 43
1
2
3
4
5
6
7
8
9
adduct.48 The transaminase function of TcPAM catalyzed efficient intermolecular exchange of the
amino group to make a mixture of α- and β-amino acids.49 Encouraged by the intermolecular
mechanism of TcPAM and our reinterpretation of the chemistry of MIO enzymes with epoxides,
here, we describe a new application for the catalyst to transfer an amino group to various ring-
substituted cinnamate epoxides to make phenylserine analogues. Michaelis-Menten constants of
TcPAM catalyzing the epoxide ring-opening chemistry were calculated, and the absolute
stereochemistry of each enzymatically derived phenylserine was assessed.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
MATERIALS AND METHODS
General Procedure for Syntheses of Cinnamate Epoxides (2a–2l). The cinnamate epoxides
were synthesized according to a procedure described previously.50 In a 50-mL single-necked round-
bottomed flask, a stirred slurry of a trans-cinnamic acid analogue (1a–1l) (0.75 mmol) in acetone
(515 µL, 7.5 mmol) was treated first with sodium bicarbonate (3.3 mmol) and then with dropwise
addition of water (515 µL). To the resulting thick mixture, a solution of oxone monopersulfate (1.4
mmol, contains 1.8 equiv of KHSO5) in 0.4 mM Na2EDTA solution (1.6 mL) was added dropwise
for 1 h while the temperature was kept at ~25 °C and the pH at 7.5. The mixture was then stirred
an additional 6 h and cooled to –5 °C. The reaction was acidified to pH 2 (12 M HCl) and mixed
with ethyl acetate (5 mL) with rapid stirring. The mixture was then filtered and extracted with
ethyl acetate (3 × 50 mL). The combined organic fractions were washed with saturated NaCl, dried
over anhydrous MgSO4, filtered, and concentrated under vacuum. After all the solvent was
removed, each trans-cinnamic acid epoxide analogue was isolated as an oily residue. The oily
residue was dissolved in ethanol (EtOH) (1 mL), cooled on ice, and treated with a solution of KOH
(3.6 mmol) dissolved in EtOH (1 mL). The resulting thick slurry was filtered, and the residue was
8
ACS Paragon Plus Environment

Page 9 of 43
ACS Catalysis
1
2
3
4
5
6
washed with EtOH and then dried under vacuum to provide the potassium salts of the cinnamate
epoxides as a racemic mixture (Table S1, Figures S1 and S2 of the Supporting Information).
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Expression and Purification of TcPAM. The tcpam cDNA (codon-optimized for expression in
bacteria) was ligated into the expression vector pET28a(+),42 and the recombinant plasmid
encoded an N-terminal His6 tag. Escherichia coli BL21(DE3) cells, transformed to express
TcPAM, were grown at 37 °C for 12 h in 150 mL of Lysogeny broth (LB). Separate aliquots (25
mL) of this inoculum culture were then added to each of six 1-L cultures of LB supplemented with
kanamycin (50 μg/mL). The cells were incubated at 37 °C until OD600 = 0.7. Isopropyl-D-
thiogalactopyranoside (500 μM) was added to the cultures with expression conducted at 16 °C for
16 h. The cells were harvested by centrifugation at 4,650g (15 min), and the pellets were diluted
in resuspension buffer (100 mL of 50 mM sodium phosphate containing 5% (v/v) glycerol and 300
mM NaCl, pH 8.0). The cells were then lysed by brief sonication [one 10-s burst at 60% power
with a 20-s rest interval for 20 cycles on a Misonix Sonicator (Danbury, CT)]. The cellular debris
was removed by centrifugation at 27,200g (20 min) followed by high-speed centrifugation at
142,000g (90 min) to remove light membrane debris. The resultant crude aminomutase in the
soluble fraction was purified by nickel-nitrilotriacetic acid affinity chromatography according to
the protocol described by the manufacturer (Invitrogen, Carlsbad, CA); TcPAM eluted in 50 mL
of 250 mM imidazole dissolved in resuspension buffer. Fractions containing active soluble TcPAM
(76.5 kDa) were combined and loaded onto a size-selective centrifugal filtration unit (30,000
NMWL, Millipore Sigma, Burlington, MA). The protein solution was concentrated and diluted
over several cycles until the imidazole and salt concentrations were <1 µM, and the final volume
was 1 mL, containing ~14 mg of TcPAM. The quantity of TcPAM was measured using a Nanodrop
9
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 43
1
2
3
4
5
6
7
8
9
spectrophotometer (ThermoFisher Scientific, Waltham, MA) and the purity (82%) was assessed
by SDS-PAGE with Coomassie Blue staining using Kodak Gel Logic 100 Imaging System
(version 3.6.3) to integrate the relative intensities of the scanned protein bands (Figure S5 of the
Supporting Information).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Expression and Purification of (2S)-Threonine Aldolase (ltaE). A cDNA (from E. coli,
accession number: P75823) encoding a low specificity (2S)-threonine aldolase (ltaE) was ligated
into a pET28a(+) vector, encoding an appended N-terminal His6-tag, to make a recombinant
plasmid (designated as pKDW014_lsTA) was purchased from GenScript (Piscataway, NJ).
Escherichia coli BL21(DE3) cells were then transformed with pKDW014_lsTA to overexpress
the ltae gene by standard protocols (Millipore Sigma, Burlington, MA). Transformed bacteria were
used to inoculate LB (100 mL) and grown at 37 °C for 12 h. Separate aliquots (15 mL) of this
inoculum culture were added to each of three 1-L cultures of LB supplemented with kanamycin
(50 μg/mL). The cells were incubated at 37 °C until OD600= 0.6, and isopropyl-D-
thiogalactopyranoside (100 μM) was added to induce expression at 16 °C for 16 h. The cells were
harvested by centrifugation at 4,650g (15 min), and the pellets were diluted in resuspension buffer
(at pH 7.0) containing 10 µM PLP and 300 mM NaCl. The cells were then lysed by brief sonication
[one 10-s burst at 60% power with a 20-s rest interval for 20 cycles on a Misonix Sonicator], and
the cellular debris was removed by centrifugation at 27,200g (20 min) followed by high-speed
centrifugation at 142,000g (90 min) to remove light membrane debris. The clarified lysate
containing crude (2S)-TA in the soluble fraction was purified by nickel-nitrilotriacetic acid affinity
chromatography according to the protocol described by the manufacturer (Invitrogen, Carlsbad,
CA); (2S)-TA was eluted in 50 mL of 250 mM imidazole. Fractions containing active soluble (2S)-
10
ACS Paragon Plus Environment

Page 11 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
TA (36.5 kDa) were combined and loaded onto a size-selective centrifugal filtration unit (15,000
NMWL, Millipore Sigma). The protein solution was concentrated and diluted over several cycles
until the imidazole and salt concentrations were <1 µM, and the final volume was 1 mL. The
concentration of (2S)-TA was measured (8.8 mg/mL) using a Nanodrop spectrophotometer
(ThermoFisher Scientific). The purity (99%) was assessed by SDS-PAGE with Coomassie Blue
staining using Kodak Gel Logic 100 Imaging System (version 3.6.3) to integrate the relative
intensities of the scanned protein bands (Figure S5 of the Supporting Information).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Biocatalysis of Phenylserine with TcPAM. A solution of (2S)-styryl-α-alanine (1 mM) in 50 mM
NaH2PO4/Na2HPO4 buffer (pH 8.0, 5% glycerol) was preincubated with TcPAM (100 µg/mL) for
2 min. Cinnamate epoxide (1 mM) was added to the solution, and the assay was mixed at 31 °C
on a rocking shaker for 2.5 h. The reaction was then stopped with 10% formic acid to adjust the
pH to 3.0, and 3'-bromo-α-phenylalanine (50 nM) was added as an internal standard. This reaction
mixture was analyzed using a liquid chromatography-electrospray ionization-multiple reaction
monitoring (LC/ESI-MRM) method.
Kinetic Analysis. To calculate the steady-state enzyme kinetic constants, each epoxide substrate
was varied at intervals from 25 μM up to 1000 μM in triplicate assays containing TcPAM (100
µg/mL). The reactions were terminated with 10% formic acid (pH 3.0) and the resultant ring-
opened biocatalyzed products, without derivatization, were quantified by LC/ESI-MRM. The
apparent kinetic parameters (KM and kcat) were calculated by non-linear regression with Origin Pro
9.0 software (Northampton, MA) (Figures S6 and S7 of the Supporting Information), using the
11
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 43
1
2
3
4
5
6
7
8
9
Michaelis-Menten equation: vo = [Eo]kcat/(KM + [S]). Control assays contained all the
necessary components for an operational assay except either the amine group donor
(2S)-styryl-α-alanine or the enzyme was omitted (Table S3 of the Supporting Information).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
General Procedure for the Syntheses of Phenylserines (4a–4l). The phenylserine analogues
were synthesized according to a procedure described previously.51 Triethylamine (22 mmol) was
added to a solution of glycine (5 mmol) in water (4 mL). To this solution, a benzaldehyde (10
mmol) analogue was added dropwise over 15 min, and the mixture was stirred for 12 h at ~25 °C.
The color of the reaction mixture gradually changed from clear and colorless to yellow-brown. n-
Butanol (3 mL) was added, and the triethylamine was evaporated under vacuum. The butanolic
solution was diluted with water (3 mL), and the mixture was acidified to pH 2 with HCl (6 M).
The acidified solution was stirred at ~25 °C for 3 h and partitioned against ethyl acetate (2 × 5 mL)
to remove the unreacted benzaldehyde. The aqueous layer was separated and neutralized to pH 6.0
with a saturated NaHCO3 solution to precipitate the phenylserine. The mixture was stirred for 1 h
at 0 °C, and the phenylserine product was washed with water (3 mL) and dried under vacuum to
yield the corresponding mixture of phenylserine diastereomers.
General Method for Derivatizing Phenylserines with a Chiral Auxiliary. A mixture of all four
stereoisomers of phenylserine (0.21 mmol) was dissolved in 50 mM NaH2PO4 /Na2HPO4 buffer
(pH 8.0) containing 5% glycerol (1 mL). To this solution were added pyridine (50 µL, 0.62 mmol)
and (2S)-2-methylbutyric anhydride (60 µL, 0.30 mmol), and the reaction mixture was stirred for
20 min at ~25 °C. The solution was adjusted to pH 2 (6 M HCl) to quench the reaction, and the N-
12
ACS Paragon Plus Environment

Page 13 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
protected phenylserine was extracted with ethyl acetate (2 mL). The organic layer was separated
and evaporated under a stream of nitrogen gas, and the resultant residue was dissolved in 3:1
EtOAc/MeOH (v/v) (1 mL). Diazomethane in diethyl ether was added dropwise to obtain the
methyl ester, and the solvent was removed under a stream of nitrogen gas. The resulting methyl
ester was dissolved in dichloromethane (1 mL) to which pyridine (100 µL, 1.24 mmol) and
chlorotrimethylsilane (150 µL, 1.18 mmol) were added, and the solution was stirred for 15 min at
~25 °C. The reaction was quenched with water (1 mL) and the organic fraction was separated and
analyzed by gas chromatography coupled with electron-impact mass spectrometry (GC/EI-MS).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISUSSION
Synthesis of Racemic Epoxide Substrates. Various commercially available cinnamic acids (1a–
1l) were oxidized with Murray's reagent to synthesize the corresponding ring substituted cinnamate
epoxides (2a–2l) as their racemates (see Scheme 2, entry 4 for assignment of substituents a–l).50
The enantiomeric ratios of each racemate methyl ester was ~1:1 by chiral GC/EI-MS analysis
(Figures S1 and S2 of the Supporting Information). Trifluoroacetone was used instead of acetone
to modify Murray's reagent and oxidize the less reactive NO2-cinnamate analogues to epoxides 2g
and 2h. The 4'-OCH3-cinnamate epoxide rapidly hydrolyzed to the dihydroxy compound under
these reaction conditions and thus could not be tested in this study.
Biocatalysis of Arylserines. Each epoxide was incubated separately with the previously
employed amine donor (2S)-styryl-α-alanine (3) and the mutase48 at pH 8.0 to make the
phenylserines. (2S)-Styryl-α-alanine (3) was the amine donor that loaded an NH2 group on the
MIO moiety of TcPAM instead of 6 M ammonium salts at pH 9 used in several previous studies
to convert various cinnamates to α- and β-amino acids.44, 52-53 Another synthesis study informed
13
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 43
1
2
3
4
5
6
7
8
9
that high concentrations of ammonia could non-enzymatically aminate cinnamate epoxide nearly
exclusively at the benzylic carbon (Cβ), converting it to phenylisoserine.54 This selective amination
of cinnamate epoxide at Cβ is directed by the ability of the benzylic carbon to resonance stabilize
the + formed in the transition state. This electronic-based regioselectivity would therefore be
sensitive to the electronic effects of substituents on the aryl ring. Control experiments (with or
without the TcPAM catalyst) in this study showed that after incubating 2 M NH3 (NH4OH in 50
mM phosphate buffer, pH 8) for 2.5 h with the cinnamate epoxide bearing a mesomerically
electron-withdrawing 4'-NO2 (2h), the amination reaction yielded a ~40:60 mixture of isoserine to
serine, with the latter predominating (Figure S3 of the Supporting Information). The 4'-NO2 group
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(σ4'-NO = 0.78) (where σ is the Hammett substituent constant)55 is positioned to pull electron
2
density of the epoxide oxygen toward Cβ, polarize the Cα–O bond, and thus encourage nucleophilic
amination at Cα to form mostly phenylserines as observed. By contrast, cinnamate epoxide (2a)
(σ[H] = 0.0) and its 3'-OCH3 (2b) (σ[3'-OCH ] = 0.12) and 4'-CH3 (2i) (σ[4'-CH ] = -0.17) analogues
3
3
have similar electronic influence on the epoxide ring opening. Substrates 2a, 2b, and 2i favored
intrinsic amination at Cβ to form the arylisoserines at ≥93% compared to the lesser amount of
serine when incubated with 2 M NH3 (Figure S3), as observed in earlier synthetic studies.54, 56
Thus, (2S)-styryl-α-alanine (3) was chosen as a milder amine source to dissect the regiochemistry
and stereoselectivity of TcPAM transamination catalysis.
Assessing the Regiochemistry of the TcPAM-catalyzed Transamination Reaction. The
regiochemistry and relative stereochemistry of the mutase amination reaction were assigned by
derivatizing the products made from enzyme catalysis. For example, authentic racemic (2S)+(2R)-
syn-phenylserine [(2S)+(2R)-syn-4a] and (2R,3S)-syn-phenylisoserine [(2R,3S)-syn-4a] standards
(Scheme 2 insets) were converted to their O-trimethylsilyl N-[(2S)-2-methylbutyryl] methyl esters
14
ACS Paragon Plus Environment

Page 15 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
(Scheme 2A). The diagnostic fragment ions in the GC/EI-MS profile for the derivatized authentic
phenylserine racemate were identical to those of the enzymatically derived phenylserine
diastereomers yet distinct from those of the (2R,3S)-syn-phenylisoserine ((2R,3S)-syn-5a) standard
derivatized identically (Figure 3). GC/EI-MS fragmentation of the other biocatalytically made
phenylserine analogues (5a–5l) was identical to the ion profiles of the corresponding synthetic
phenylserine analogues derivatized identically (Scheme 2B and Figures S8–S18 of the Supporting
Information).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A
B
O
OK
O
OH
O
OH
O
R
1 (a–l)
H
O
O
NH3
NH3
R
R
R
10 mmol
(2R)-syn-4
(2S)-syn-4
a
TMSO
O
O
O
OH
O
OH
O
O
O
O
OH
O
b
b
OK
HN
OH
O
O
O
R
NH3
NH2
R
NH3
NH3
R
R
2 (a–l)
+
R
a
5 mmol
5 (a–l)
4 (a–l)
(2R)-anti-4
(2S)-anti-4
O
NH3
O
Entry
R
anti:syn Entry
R
anti:syn
2
3
O
O
NH
O
O
4a
4b
4c
4d
4e
4f
H
75:25
4g 3′-NO2 96:4
4h 4′-NO2 83:17
NH2
OH
O
3′-OCH3 71:29
3′-CH3 68:32
R
3
OTMS
4 (a–l)
4i
4j
4k
4l
4′-CH3 72:28
4′-F 52:48
R
5 (a–l)
3′-F
71:29
97:3
NH3
2R
3′-Cl
4′-Cl 45:55
4′-Br 58:42
3S
OH
3′-Br 58:42
Scheme 2. A) TcPAM was incubated with (2S)-styryl-α-alanine (3) (1 mM) and separately with each trans-cinnamate
epoxide racemate 2a–2l made from cinnamic acids 1a–1l. Step a) i) TcPAM (1.7 mg/mL) in 50 mM
NaH2PO4/Na2HPO4 (pH 8.0), 5% glycerol, 2 min pre-equilibration, 2a–2l (10 mM), 29 °C, 4 h. Step b) Derivatization
of putative phenylserines and phenylisoserines using a chiral auxiliary for stereoisomeric resolution. i) (2S)-2-
Methylbutyric anhydride, pyridine, rt, 20 min; ii) 6 M HCl, pH 2; iii) CH2N2, EtOAc/MeOH (3:1 v/v), rt, 15 min; and
iv) chlorotrimethylsilane, pyridine, CH2Cl2, rt, 15 min. Insets: stereoisomerism of (2R,3S)-syn-phenylisoserine. B)
Synthesis of the stereoisomers of phenylserine analogues from glycine and a substituted benzaldehyde. a) step i)
triethylamine, n-BuOH and H2O, 12 h, rt; step ii) 6 M HCl, pH 2; and step iii) NaHCO3, pH 6.
TcPAM could theoretically transfer the amino group to either Cα or Cβ of the oxirane substrate
to make phenylserine or phenylisoserine, respectively. We hypothesized that the aryl ring alone
and the electronics of the substituents attached to it would differentially affect the reactivity at Cβ
15
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 43
1
2
3
4
5
6
7
8
9
toward nucleophilic amination, as they did in our control experiments. However, the regiochemical
analysis showed that regardless of the electronic nature of the substituent attached to the aryl ring,
TcPAM aminated predominantly at Cα to open the cinnamate epoxides.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A
B
100
75
50
25
0
100
75
50
25
0
179
179
73
73
57
245
250
245
250
57
336
336
350
50
100
150
200
300
350
50
100
150
200
300
m/z
m/z
C
H
100
75
50
25
0
Si
O
N
Si
O
106
O
O
O
NH
O
O
O
O
O
O
O
HN
Si
HN
O
Si
m/z 336
Panels A + B
m/z 336
Panel C
m/z 245
Panels A + B
m/z 235
Panel C
190
57
73
NH2
Si
O
O
NH
Si
H
235
250
292
300
336
350
m/z 190
Panel C Panels A + B
m/z 179
m/z 106
Panel C
m/z 73
50
100
150
200
m/z
Figure 3. GC/EI-MS spectra of authentic A) (2S)+(2R)-syn-phenylserine from Sigma-Aldrich, B) phenylserine
biocatalyzed from cinnamate epoxide by TcPAM catalysis (see Figure 5 for GC profiles), and of C) (2R,3S)-syn-
phenylisoserine (see Figure S4 of the Supporting Information for GC profile). Each hydroxyamino acid was
derivatized to its O-trimethylsilyl N-[(2S)-2-methylbutyryl] methyl ester. The molecular ion (M•+, m/z 351) was
not observed for either analyte.
1
1
Relative Stereochemistry of TcPAM reaction by H NMR Analysis. A H NMR method was
used to establish the relative syn/anti-stereochemistries of the four synthesized arylserine
diastereomers. The coupling constants for the vicinal protons (Hα and Hβ) were similar (3JHα-H
(anti)
β
3
and JHα-H
≈ 4 Hz) for the underivatized synthetic anti- and syn-phenylserine diastereomers.
(syn)
β
16
ACS Paragon Plus Environment

Page 17 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
Thus, differences in Hα and Hβ coupling constants could not be used to distinguish the anti- and
syn-diastereomers, as done in another study.57 Thus, a small amount of authentic (2S)+(2R))-syn-
phenylserine racemate was added to the synthesized mixture of anti- and syn-phenylserine
diastereomers (Figure 4A). The 1H NMR of the "spiked" sample showed a relative increase of the
doublet pair ("outer doublets") at δ 4.23 and δ 5.41 (Figure 4B). The chemical shift difference (Δδ)
of 1.15 ppm at pH 1.5 for the "outer doublets" assigned these signals to the syn-phenylserine
enantiomers. The "inner doublets" at δ 4.33 and δ 5.37 (Δδanti = 1.04 ppm, pH 1.5) were thus
assigned to the anti-phenylserine enantiomers. We used the trends in the Δδ values to assign the
relative anti/syn-stereochemistries and anti:syn ratios of the remaining synthetic and bicatalytically
derived phenylserine analogues without derivatization (Table 1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A
B
1.00
1.04
8.34
8.55
0.33
0.33
1.00
1.00
5.6 5.4 5.2 5.0 4.8 4.6 4.4 4.2 4.0
5.6 5.4 5.2 5.0 4.8 4.6 4.4 4.2 4.0
 (ppm)
 (ppm)
O
O
H
OH
H
HO
OH
OH
H
H
H3N
H3N
(2S)-anti
(2S)-syn
Figure 4. An expanded 500 MHz 1H NMR of A) synthesized phenylserine (4a),
and B) of 4a containing authentic (2S)-syn-phenylserine (see Figure 2) recorded
in D2O at pH 1.5. Chemical shift values are listed vertically above each peak,
and the relative area under each peak is listed horizontally next to each peak.
Each chemically and enzymatically synthesized arylserine analogue was dissolved separately
in D2O at pH 1.5 and analyzed by 1H NMR. The Δδsyn (~1.15 ppm, pH 1.5) of the "outer doublets"
was consistently larger than the (Δδanti) (~1.05 ppm, pH 1.5) for the "inner doublets" (Figures S19–
17
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 43
1
2
3
4
5
6
7
8
9
S30 of the Supporting Information). The NMR analysis substantiated that TcPAM aminated the
epoxide substrates predominantly at Cα to produce anti-arylserine analogues as the major product,
regardless of the substituent on the aryl ring. A smaller amount (≤9%) of the syn-phenylserines
was also made.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
Table 1. 500 MHz H NMR Data for Synthetically and Biocatalytically Produced
Phenylserine Analogues (4a–4l).
3JHα-H 3JHα-H
3JH –H
b
b
Δδanti
(ppm) (ppm)
Δδsyn
Δδbiocatalytic anti:sync
(ppm)
anti:sync
(GC-MS)
β
β
β
biocatαalyzed
R
anti
(Hz)
syn
(Hz)
(NMR)a
(Hz)
4a
H
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.5
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
1.04
1.02
1.00
1.02
1.03
1.02
1.08
1.06
1.00
1.03
1.03
1.00
1.17
1.15
1.14
1.16
1.18
1.17
1.21
1.22
1.13
1.16
1.16
1.13
4.0
4.0
4.0
4.0
4.0
4.0
3.5
4.0
4.0
4.0
4.0
4.0
1.05
1.02
1.01
1.04
1.01
1.02
1.08
1.09
0.99
1.03
1.00
0.98
93:7
100:0
100:0
98:2
91:9
95:5
96:4
95:5
94:6
92:8
92:8
96:4
95:5
97:3
91:9
95:5
4b 3′-OCH3
4c
4d
4e
4f
3′-CH3
3′-F
3′-Cl
3′-Br
91:9
93:7
4g 3′-NO2
4h 4′-NO2
93:7
97:3
4i
4j
4k
4l
4′-CH3
4′-F
100:0
93:7
4′-Cl
100:0
4′-Br
100:0
a"0" indicates that the syn-isomer was below the limits of detection of the NMR (at ~100 nmol). bΔδanti and
Δδsyn values for the synthetic phenylserine analogues. anti:syn ratio for the biocatalytically produced
phenylserine analogues. n = 3. Standard error < 1%.
c
18
ACS Paragon Plus Environment

Page 19 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
TMSO
O
8.01
TMSO
O
A
7.95
100
O
O
O
O
HN
HN
(2S)-anti
(2R)-anti
8.10
8.13
0
100
TMSO
TMSO
O
O
O
B
C
O
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
HN
(2S)-syn
O
HN
(2R)-syn
0
100
0
100
D
0
7.7 7.8 7.9 8.0 8.1 8.2 8.3 8.4
Time (min)
Figure 5. Gas-chromatography/mass spectrometry
extracted-ion chromatograms with m/z 179 ion
monitoring of A) synthesized phenylserine as shown in
Scheme 2B was derivatized with a chiral auxiliary as per
Scheme 2A. Peaks at 7.95 min and 8.01 min correspond
to (2S)+(2R)-anti-phenylserine isomers, and those at
8.10 and 8.13 min correspond to (2S)+(2R)-syn-
phenylserine isomers; B) derivatized authentic
(2S)+(2R)-syn-phenylserine racemate (Sigma-Aldrich,
contains 13% (2S)+(2R)-anti-phenylserine as an
impurity; C) chiral-auxiliary of biocatalyzed
phenylserine made from the cinnamate epoxide
racemate. Biocatalyzed (2S)+(2R)-syn-phenylserine
stereoisomers were also produced at ~9% (peaks at 8.10
and 8.13 min); and D) Extracted ion (m/z 179)
chromatogram of derivatized phenylserine after treating
with (2S)-TA for 15 min.
Relative Stereochemistry of the TcPAM Reaction by Chiral Auxiliary Derivatization. An
orthogonal GC/MS separation method was used to verify the relative stereochemistry of the
1
biocatalyzed arylserines assigned earlier in this study by a H NMR method. The chiral N-(S)-
methylbutyryl-O-trimethylsilyl methyl ester derivatization of the arylserines described herein were
extended to a diastereomeric mixture of synthesized (2S)+(2R)-syn-4a and (2S)+(2R)-anti-4a
(Figure 5A), commercial grade (2S)+(2R)-syn-4a (Figure 5B), and 4a biocatalyzed by TcPAM
19
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 43
1
2
3
4
5
6
7
8
9
from the cinnamate epoxide racemate (2a) (Figure 5C) (Scheme 2A). GC/MS profiles for the N,O-
derivatives of the commercial (2S)+(2R)-syn-phenylserine had major peaks at retention times 8.10
min and 8.13 min (Figure 5B) and thus established the stereochemistry of the later-eluting isomeric
pair in the mixture. The retention times of these most abundant peaks in the commercial sample
matched those of the lower abundant peaks for the derivatized isomers in the synthetic 4a sample
(Figure 5A). The derivatized enantiomers in the phenylserine diastereomeric mixtures eluting at
7.95 min and 8.01 min were therefore assigned as (2S)+(2R)-anti-phenylserine (Figure 5).
In summary, the derivatives of the anti-phenylserine enantiomers eluted first, and the syn-
phenylserine enantiomers eluted later from the GC/MS. In the biocatalyzed sample, TcPAM
diastereoselectively turned over the racemate of cinnamate epoxide to a mixture of (2R)+(2S)-anti-
phenylserines with one enantiomer predominating (Figure 5C). The arylserines made from
epoxides 2b–2l by TcPAM catalysis were derivatized to 5b–5l and analyzed by GC/MS to assign
the relative stereochemistries (Figures S31–S33). The biocatalyzed products contained (2R)+(2S)-
anti-arylserines as the major stereoisomers, with one enantiomer predominating, while the
(2R)+(2S)-syn-arylserines were minor, at ≤9% relative abundance (Table 1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Absolute Stereochemistry of the TcPAM Reaction by Aldolase Resolution. To assess which
enantiomer was biocatalyzed more abundantly among the (2R)+(2S)-anti-arylserines, we used a
low specificity (2S)-threonine aldolase (TA) from E. coli.58 This enzyme resolution step cleaved
the (2S)-anti diastereomers into a benzaldehyde moiety and glycine, resolving the synthesized
(2R)-anti hydroxy amino acid diastereomers. Each stereoisomeric mixture of synthesized
arylserine (entries 4a–4l) was treated with (2S)-TA. Aliquots were withdrawn at 15 min under
steady-state enzymatic turnover to measure the relative abundances of the arylserine
20
ACS Paragon Plus Environment

Page 21 of 43
ACS Catalysis
1
2
3
4
5
6
diastereomers. The remaining arylserine stereoisomers were treated sequentially with (2S)-2-
methylbutyric anhydride, diazomethane, and chlorotrimethylsilane.
7
8
9
GC/EI-MS analysis of each derivatized arylserine showed that the peak corresponding to the
(2S)-anti isomer diminished relative to that for the (2R)-enantiomer (see Figure 5D and Figures
S46–S49 of the Supporting Information). We could then assign the 2R-anti and 2S-anti-isomers
of each arylserine and calculate the 2R:2S-anti ratios of the biocatalyzed arylserines (Figure 6).
The 2R:2S-anti ratios for the biocatalyzed arylserines ranged from 66:34 (4'-fluorophenylserine,
4j) to 88:12 (3'-nitrophenylserine, 4g) with the (2R)-anti isomer predominating (Figure 6 and
Figures S31–S33 of the Supporting Information). The 2R:2S ratios for each biocatalyzed syn-
product could not be calculated accurately because of their ~10-fold lower abundance than the
anti-isomers and poorer resolution on the GC/MS (see Figure 5B as an example). The anti:syn
ratios calculated from the abundance of derivatized biocatalyzed phenylserines determined by
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
GC/MS analysis agreed with those calculated by the H NMR method for the underivatized
biocatalytically made arylserines (Table 1). No clear trend explained how a combination of
substituent electronic effects, position, and sterics influenced the R:S ratios of the isoserines made.
100
12
15
85
15
85
17
83
20
80
88
23
77
80
60
40
20
0
26
74
29
71
30
70
32
68
33
67
34
66
H
3
3
3
3
3
3
4
4
4
4
4
'
-
'
-
'
-
'
-
'
-
'
-
'
-
'
-
'
-
'
-
'
-
B
O
C
F
C
B
N
N
C
F
C
r
r
H
3
l
O
2
O
2
H
3
l
C
H
3
Figure 6. Distribution of enantiomers (2R)-anti (gray bars)
and (2S)-anti (black bars) in biocatalyzed products for the
substituted phenylserine analogues.
21
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 43
1
2
3
4
A
B
C
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
D
E
Tyr
HO
O
OH
2R
O
O
OH
OH
O
O
O
OH
OH
2S
O
O
O
O
O
O
OH
NH2
H2N
b
a
NH2
anti
H2N
H2N
N
NH2
b
a
syn
O
H
O
H
O
O
N
N
N
N
N
N
N
F
G
Figure 7. A) Low-energy binding poses are shown of (2S,3R)-cinnamate epoxide ((2S,3R)-2a) (orange sticks) and
(2R,3S)-cinnamate epoxide ((2R,3S)-2a) (light-gray sticks) at center of the image. The conformations are
consistent with the active site of TcPAM consisting of residues (shown as yellow sticks), the catalytic Tyr80, a
putative catalytic Tyr322 (light blue sticks), binding contact Arg325 (golden-rod sticks), and the methylidene
imidazolone (MIO) moiety (green sticks). B) (2S,3R)-2a (orange sticks) and C) (2R,3S)-2a (light-gray sticks) are
posed separately in the active site to highlight the nominal distances (<3.5 Å) that substituents on the aryl ring are
from active site residues. Inset) TcPAM in complex with cinnamic acid, based on PDB codes 3NZ4 and 4CQ5.
Printed on the aryl ring, 'm' designates meta-positions (equivalent to the 3'-designation used in the text) and 'p'
designates the para-position (equivalent to the 4'-designation used in the text) per ligand. Heteroatoms are colored
red for oxygen and blue for nitrogen. The images were produced with UCSF Chimera,59 and the docking
conformations were generated with AutoDock Vina60 from PDB code 3NZ4. Numbers are distances in Å. D) a)
Rendering of the mechanism showing attack of the NH2-MIO on the σ* antibonding orbital (gray lobe) at Cα for
the C–O σ bond of (2S,3R)-2a to produce (2R)-anti-phenylserine. b) A double inversion-of-configuration
mechanism is envisioned to access the minor syn-stereoisomer that proceeds through a putative lactone
intermediate. Various viewing angles centering on the surfaces of E) Leu104, F) Leu108, and G) Lys427 deep in
the aryl binding pocket of TcPAM (PDB code 3NZ4) with poses of the docked (2S,3R)-2a (orange sticks) and
(2R,3S)-2a (light-gray sticks).
22
ACS Paragon Plus Environment

Page 23 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
Epoxide Substrate Docking Model of TcPAM. We used solved structures of TcPAM in complex
with cinnamic acid (PDB: 3NZ4 and 4CQ5)45, 52 (Figure 7B, inset) to help interpret whether the
proposed dockings of cinnamate epoxides in the TcPAM active site by the AutoDock program
were reasonable. The cinnamate epoxides (2S,3R)-2a and (2R,3S)-2a docked in the crystal
structure of the TcPAM active site (Figure 7A, B, and C) were consistent with the conformation
of the naturally occurring cinnamate intermediate in the crystal structure (Figure 7B, inset); the
carboxylate group pointed toward Arg325, and the aryl binding region comprised key residues
Ile431, Leu108, Leu104, and Leu227 (Figure 7B and C). The docking information suggested that
(2S,3R)-2a yields the (2R)-anti-phenylserine (Figure 7D). The anti-stereochemistry of the
arylserine products follows a mechanism proceeding through backside nucleophilic attack (SN2)
by the NH2 group at Cα of the epoxide ring. The docked conformation of (2S,3R)-2a places the
epoxide oxygen 3.0 Å from the catalytic Tyr80 of TcPAM. Tyr80 normally functions as a general
acid/base during the natural α- to β-phenylalanine isomerase reaction catalyzed by TcPAM. Here,
we believe Tyr80 also serves as a general acid to promote protonation-initiated epoxide opening
during amination by the NH2-MIO. Conversely, the docked (2R,3S)-2a enantiomer is poised for
making the (2S)-anti-isomer. The oxirane oxygen of (2R,3S)-2a sits 2.9 Å from Tyr322 that has
been identified as a key residue involved in forming the MIO moiety and keeping the NH2-MIO
adduct deprotonated so that it functions as a nucleophile during the normal TcPAM-catalyzed
isomerase reaction (Figure 7A).61 TcPAM likely uses Tyr322 effectively as a surrogate general
acid for protonation-initiated epoxide opening during amination by the NH2-MIO of (2R,3S)-2a.
In this study, the syn-stereoisomers for each epoxide were made at low levels compared to the
anti-isomers, and their resolution as chiral derivatives on GC/MS was poor; thus accurate
measurements of the ratio of the syn-enantiomers could not be made. The (2S)- and (2R)-anti-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
23
ACS Paragon Plus Environment

ACS Catalysis
Page 24 of 43
1
2
3
4
5
6
7
8
9
arylserine isomers are made through SN2 attack by the amino group on the σ* antibonding orbital
of Cα–O bond of the (2S,3R)- and (2R,3S)-epoxide racemates (Figure 7D, route a). However, the
2S- and 2R-syn-arylserine diastereomers cannot be accessed through a similar SN2-type
mechanism unless TcPAM can rotate the substrate into a cisoid conformation that aligns Cα in a
proper orientation for nucleophilic attack.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Earlier mechanistic studies suggest that during its natural reaction, TcPAM rotates the
cinnamate intermediate about a central axis through a bicycle-pedal motion to invert the faces of
the α- and β-carbons 180° from their original positions.45, 61 This rotation enables the NH2 group
of the MIO adduct to add to the opposite face of Cβ from which it was removed from the
phenylalanine substrate (see Scheme 1). The oxirane ring prevents rotations about the axis between
Cα and Cβ. Thus, a double inversion-of-configuration mechanism is envisioned to access the minor
syn-stereoisomers (<10% relative abundance) that proceeds through a putative lactone
intermediate (Figure 7D, route b).62 Through the proposed lactone intermediate pathway, the
(2S,3R)-cinnamate epoxide that produces the more abundant (2R)-anti-phenylserine at 8.01 min
(Figure 5C) likely also makes the 2S-syn serine isomer (Figure 7D) (see peak at 8.10 min), and the
(2R)-syn isomer at 8.13 min (Figure 5C) is likely derived from the (2R,3S)-cinnamate epoxide.
The other ring-substituted epoxides behaved similarly (Figures S31–S33).
Kinetics of Serine Biocatalysis. The Michaelis parameters were measured by quantifying the
underivatized biocatalyzed products by LC/MS-MRM analysis. The apparent KM and kcat of
TcPAM were calculated under steady-state conditions for the analogues of cinnamate epoxide with
the amino donor (3) at 1 mM (Table 2 and see Figure S50 of the Supporting Information showing
the conversion of four example epoxides at 400 µM over 6 h). The catalytic efficiency (kcat/KM) of
TcPAM for 3'-CH3-cinnamate epoxide (2c) was the best among all substrates tested; largely
24
ACS Paragon Plus Environment

Page 25 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
influenced by its relatively lower KM compared to those for the 3'-halo (2d, 2e, and 2f), 3'-NO2
(2g), 4'-NO2 (2h), and 4'-CH3 (2i) substrates. In addition, the catalytic efficiencies of TcPAM for
2c and two other 3'-substituted cinnamate epoxides (3'-F (2d) and 3'-Cl (2e)) were superior to those
of their 4'-substituted counterparts; kcat/KM values for the latter two were principally influenced by
their higher kcat compared to those of their 4'-substituted isomers (2j and 2k).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We note that TcPAM catalyzes its natural α- to β-amino acid isomerization reaction at 3.0
min-1,45 which is ~10-fold faster than the average turnover rate (~0.3 min-1) of TcPAM for the 3'-
substituted cinnamate epoxide substrates used in this study. However, the latter rate is similar to
the average rate at which other MIO-dependent aminomutases catalyze their natural α- to β-amino
acid isomerization reactions involved in specialized metabolism.63 Thus, the turnover of the 3'-
substituted cinnamate epoxides by TcPAM is within the same order of magnitude as mutases that
make β-amino acids in their natural hosts to confer an evolutionary advantage.
Since the calculated docking conformations of the cinnamate epoxides agreed with those of
other MIO-enzymes in complex with their natural phenylpropanoid substrates,45-47, 52 we therefore
used the models to help interpret how the position of the substituents could potentially affect
catalysis due to steric interactions. Substrate docking models show that for each ligand enantiomer
one of the two 3'- ("meta"-) carbons of the aryl ring of the substrate can position its substituent in
steric relief when pointed toward Lys427 (Figure 7B) and Ile431 (Figure 7C and G). This in part
suggests that substrates 2c, 2d, and 2e with relatively smaller 3'-substituents (CH3, F, and Cl,
respectively) are more able to adopt a catalytically competent conformation compared to their 4'-
isomers. The latter place substituents on the 4'- ("para"-) carbon of each ligand in greater steric
conflict with Leu104 (Figure 7B), Leu108, and Leu227 (Figure 7C, E, and F), likely preventing
epoxides 2i, 2j, and 2k from being turned over as efficiently (Table 2) (Figures S51 and S52).
25
ACS Paragon Plus Environment

ACS Catalysis
Page 26 of 43
1
2
3
4
5
6
7
8
9
Table 2. Kinetics of TcPAM for Turnover of Epoxides (2a–2l) to Serines (4a–4l) and
Isoserines (4a–4l).
푺풆풓
푰풔풐
푺풆풓
푰풔풐
풌
풌
풌
풌
KM
kcat/KM
KM
kcat/KM
퐜퐚퐭
퐜퐚퐭
퐜퐚퐭
퐜퐚퐭
Entry
R
Entry
R
(µM) (M-1 s-1)
(µM) (M-1 s-1)
(min-1) (min-1)
0.39 0.003
(min-1) (min-1)
0.43 0.01
2a
H
340
2000
50
19
6.5
113
61
2g 3'-NO2
2h 4'-NO2
2i 4′-CH3
760
610
450
40
10
21
10
17
4.5
15
(0.10)a (<0.001)
(0.07) (0.001)
0.75 0.004
(0.11) (<0.001)
0.17
(0.03) (0.01)
0.04 0.005
(<0.01) (<0.001)
0.01 0.002
(<0.01) (<0.001)
~0.04 0.01
0.46
0.27
(0.14) (0.15)
0.31 0.03
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2b 3′-OCH3
2c 3′-CH3
0.11
(0.01) (0.001)
0.62 0.009
(0.03) (<0.001)
2d
2e
2f
3'-F
3'-Cl
3'-Br
170
600
140
2j
2k 4'-Cl
2l 4'-Br
4'-F
1.3
0.10
41
50
(0.1) (0.01)
0.02 0.002
(<0.01) (<0.001)
2.5
50
(<0.01) (<0.001)
aStandard deviation in parenthesis (n = 3). The "<" symbol indicates that the actual value, estimated to one
푘푆푒푟
푘퐼푠표
is the apparent kcat of
cat
significant figure, is shown.
is the apparent kcat of arylserine production,
cat
arylisoserine production, and kcat is the turnover rate for the production of both arylserine and arylisoserine
by TcPAM. The KM of TcPAM for each substrate is calculated from the production of arylserine and
arylisoserine combined.
Furthermore, the 3'-Cl-cinnamate epoxide (2e) was turned over (kcat) by TcPAM faster than the
other epoxides in this study and interestingly much more superior (>130-fold) than the 4'-Cl isomer
(2k) (Table 2). The other 3'-substituted analogues turned over ~16-fold and ~2-fold faster than
their 4'-substituted counterparts (2j and 2i) were the 3'-F- (2d) and 3'-CH3- (2c) cinnamate
epoxides, respectively. The larger steric 3'-Br- (2f) and 4'-Br- (2l) substrates were turned over with
poorer catalytic efficiency, due mainly to their equally poor kcat values. The KM values of TcPAM
for both Br-substituted substrates were similar yet lower (estimating tighter binding) than for other
substrates such as 2d, 2e, 2g, and 2h that were turned over faster (Table 2). This suggests that a
yet unknown phenomena resulting in part from tighter binding and steric misalignment likely
disrupted the turnover of the bulkier bromo-substituted substrates by TcPAM (Figure S51A and
B, and Figure S52A and B of the Supporting Information).
It was interesting to find that the 3'-OCH3-cinnamate epoxide (2b) with the largest estimated
steric volume64-65 (26.4 Å3) was turned over similar to most 3'-substituted substrates with smaller
substituents, such as H (4.5 Å3), F (7.7 Å3), or Cl (18.8 Å3) (see 2a, 2c, and 2d as examples, Table
26
ACS Paragon Plus Environment

Page 27 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
2). The significantly lower catalytic efficiency, yet similar turnover by TcPAM of 2b compared to
those of 2a, 2c, and 2d with smaller substituents suggested that the larger-size OCH3 group through
a curious pathway affected the KM but not the turnover, like the Br group.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The catalytic efficiency (kcat/KM) of TcPAM for the 4'-NO2-substituted epoxide (2h) was ~2-
fold higher than for the 3'-NO2-isomer (2g), which parallels the difference in their turnover (kcat).
This regiospecificity trend was similar to that described earlier for substrates 3'-Br- (2f) and 4'-Br-
(2l), where the 4'-isomer was turned over slightly faster than the 3'-isomer.
Kinetics of Isoserine Biocatalysis. To make the isoserines, the NH2-MIO adduct needs to align
the phenylpropanoid of the cinnamate epoxide for occasional amino group attack at Cβ instead of
at Cα (see Figure S53 of the Supporting Information for a representative AutoDock modeled
conformation of 2b). We found by LC/ESI-MS/MS analysis that TcPAM converted each
substituted cinnamate epoxide to an arylisoserine (Figures S54–S57). Among the substituted
arylisoserines, the 3'-OCH3-phenylisoserine isomer 4b was made most abundantly at a 37:63 ratio
with its corresponding 3'-OCH3-phenylserine 4b, with the latter predominating (Table 2). The next
most abundant arylisoserines made by TcPAM were the 3'-Cl (4e) and 4'-CH3 (4i) at ~10:90 and
40:60 ratios with their cognate arylserines 4e and 4i, respectively. The remaining cinnamate
epoxides were converted to lesser amounts (≤0.03 min-1) of their isoserines (Table 2).
The sterically demanding 3'-OCH3 substituent likely "nudged" the substrate enough to place Cβ
close to and in alignment with the NH2-MIO moiety to make the 3'-OCH3-phenylisoserine more
abundantly. However, the other epoxide substrates bearing 3'-or 4'-substituents with steric volumes
[NO2 (18.2 Å3), Cl (14.3 Å3), CH3 (21.3 Å3) and Br (24.5 Å3)] equivalent to that of OCH3 (26.4
Å3) were turned over to their isoserines less efficiently. Thus, a combination of the different steric
volumes, electronics, atom geometries, and the 3' and 4'-regiochemistries of the substituents must
27
ACS Paragon Plus Environment

ACS Catalysis
Page 28 of 43
1
2
3
4
5
6
7
8
9
position the carbon skeleton of epoxides into various conformations that unpredictably promote or
preclude NH2-MIO attack at Cβ. Based on the putative docked conformation of 2b (Figure S53 of
the Supporting Information), the anti-stereochemistry between the hydroxyl and amino groups is
expected for isoserines when the epoxide is opened by SN2 amination at Cβ.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The proposed anti-stereoisomerism of the isoserines made enzymatically in this study is unlike
the syn-stereochemistry of the phenylisoserine moiety found in the antineoplastic drug paclitaxel
(see Scheme 1).66 The paclitaxel pathway is presumed to proceed through a β-phenylalanyl taxane,
which is hydroxylated to a penultimate phenylisoserinyl taxane intermediate,67 but further
evidence is needed to confirm this. Another report claims isolation of methyl phenylisoserinate by
methanol extraction of Taxus plants as evidence of a putative, phenylisoserine preassembly
pathway.68 It is unclear, however, whether the phenylisoserinate ester was obtained by
methanolysis of phenylisoserinyl taxanes, such as an advanced paclitaxel precursor, present in the
plant. Nonetheless, if syn-phenylisoserine indeed occurs as a metabolite in plants, its biosynthetic
pathway in Taxus plants remains unknown. While the trans-cinnamates were biocatalytically
converted likely to anti-isoserines in this study, the results highlight cinnamate epoxides as
potential naturally occurring precursors for direct amination to isoserines. We imagine that
naturally occurring cinnamate, derived from phenylalanine by ammonia lyase activity in plants,69
can be converted to cinnamate epoxide by a plant α-ketoglutarate-dependent hydroxylases. This
family of nonheme Fe(II) hydroxylases convert alkenes to epoxides,70 and specifically cinnamate
to cinnamate epoxide in vitro.71
As noted earlier, TcPAM catalyzes the isomerization of α-phenylalanine to β-phenylalanine in
its natural reaction by rotating the carbon skeleton of its reaction intermediate about two bonds
(see Scheme 1).45 This rotation pathway allowed by TcPAM may play a role in torqueing the ring-
28
ACS Paragon Plus Environment

Page 29 of 43
ACS Catalysis
1
2
3
4
5
6
opened trans-epoxide intermediates into a conformation needed for making syn-isoserine. In future
studies, we look to dissect the stereoselectivity of the TcPAM reaction for arylisoserine production.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CONCLUSION
The surrogate activity of TcPAM catalyzed the asymmetric and regioselective transfer of an
NH2 group from a sacrificial amine group donor (2S)-styryl-α-alanine to racemates of cinnamate
epoxide analogues in one step. TcPAM made underivatized anti-stereoisomers of arylserines
predominantly while producing significantly less of the syn-stereoisomer, and these observations
provided valuable insights into an extended mechanism of the natural α- to β-phenylalanine
isomerization reaction. The electronic effects or steric volumes of the substituents on the ring of
the cinnamate epoxide substrates did not predictably change the regioselectivity of the nucleophilic
attack by NH2. Docked poses of the cinnamate epoxide racemate in TcPAM suggested that the
conserved Tyr80 general acid present in all other MIO enzymes72 and proximate to the (2R,3S)-
epoxide purportedly helped catalyze the more abundant (2R)-anti-arylserines. Docking analysis
also identified Tyr322 as a putative general acid that likely facilitated protonation-initiated
amination of the other epoxide antipode to make the lesser abundant (2S)-anti-arylserines. In
addition, (2S)-syn-β-hydroxy amino acid stereoisomers are commonly found in specialized
glycopeptide and depsipeptide natural products. Application of an MIO-aminomutase described
here to biocatalyze enantiodivergent phenylserines with anti-stereoisomerism between the two
chiral centers has industrial and pharmaceutical relevance, allowing the possibility to perform
stereochemical structure/activity studies. For example, replacement of a naturally occurring syn-
arylserine stereoisomer for rarer anti-arylserine stereoisomers in a bioactive drug candidate could
increase drug bioavailability by slowing first-pass metabolism pathways or inhibiting bacterial
29
ACS Paragon Plus Environment