Deracemization and Stereoinversion of α-Amino Acids by l-Amino Acid Deaminase

Article abstract of DOI:10.1002/adsc.201700806

Enantiomerically pure α-amino acids are compounds of primary interest for the fine chemical, pharmaceutical, and agrochemical sectors. Amino acid oxidases are used for resolving d,l-amino acids in biocatalysis. We recently demonstrated that l-amino acid deaminase from Proteus myxofaciens (PmaLAAD) shows peculiar features for biotechnological applications, such as a high production level as soluble protein in Escherichia coli and a stable binding with the flavin cofactor. Since l-amino acid deaminases are membrane-bound enzymes, previous applications were mainly based on the use of cell-based methods. Now, taking advantage of the broad substrate specificity of PmaLAAD, a number of natural and synthetic l-amino acids were fully converted by the purified enzyme into the corresponding α-keto acids: the fastest conversion was obtained for 4-nitrophenylalanine. Analogously, starting from racemic solutions, the full resolution (ee >99%) was also achieved. Notably, d,l-1-naphthylalanine was resolved either into the d- or the l-enantiomer by using PmaLAAD or the d-amino acid oxidase variant having a glycine at position 213, respectively, and was fully deracemized when the two enzymes were used jointly. Moreover, the complete stereoinversion of l-4-nitrophenylalanine was achieved using PmaLAAD and a small molar excess of borane tert-butylamine complex. Taken together, recombinant PmaLAAD represents an l-specific amino acid deaminase suitable for producing the pure enantiomers of several natural and synthetic amino acids or the corresponding keto acids, compounds of biotechnological or pharmaceutical relevance. (Figure presented.).

Full text of DOI:10.1002/adsc.201700806

Title: Deracemization and Stereoinversion of α-Amino Acids by L-
Amino Acid Deaminase
Authors: Elena Rosini, Roberta Melis, Gianluca Molla, Davide Tessaro,
and Loredano Pollegioni
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.201700806
Link to VoR: http://dx.doi.org/10.1002/adsc.201700806

10.1002/adsc.201700806
Advanced Synthesis & Catalysis
FULL PAPER
DOI: 10.1002/adsc.201((will be filled in by the editorial staff))
Deracemization and Stereoinversion of α-Amino Acids by L-
Amino Acid Deaminase
Elena Rosinia,b,*, Roberta Melisa,b, Gianluca Mollaa,b, Davide Tessarob,c, and Loredano
Pollegionia,b
a
Department of Biotechnology and Life Sciences, Università degli Studi dell’Insubria, via J.H. Dunant 3, 21100
Varese, Italy; e-mail:elena.rosini@uninsubria.it
The Protein Factory, Politecnico di Milano and Università degli Studi dell’Insubria, via Mancinelli 7, 20131 Milano,
b
Italy
c
Department of Chemistry, Materials and Chemical Engineering “Giulio Natta”, Politecnico di Milano, p.zza
Leonardo da Vinci 32, 20133 Milano, Italy
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
Notably, D,L-1-naphthylalanine was resolved either into the
Abstract. Enantiomerically pure α-amino acids are
D- or the L-enantiomer by using PmaLAAD or the D-amino
acid oxidase variant having a glycine at position 213,
respectively, and was fully deracemized when the two
enzymes were used jointly. Moreover, the complete
stereoinversion of L-4-nitrophenylalanine was achieved
using PmaLAAD and a small molar excess of borane tert-
butylamine complex. Taken together, recombinant
PmaLAAD represents a L-specific amino acid deaminase
suitable for producing the pure enantiomers of several
natural and synthetic amino acids or the corresponding keto
acids, compounds of biotechnological or pharmaceutical
relevance.
compounds of primary interest for the fine chemical,
pharmaceutical, and agrochemical sectors. Amino acid
oxidases are used for resolving D,L-amino acids in
biocatalysis. We recently demonstrated that L-amino acid
deaminase from Proteus myxofaciens (PmaLAAD) shows
peculiar features for biotechnological applications, such as a
high production level as soluble protein in Escherichia coli
and a stable binding with the flavin cofactor. Since L-amino
acid deaminases are membrane-bound enzymes, previous
applications were mainly based on the use of cell-based
methods. Now, taking advantage of the broad substrate
specificity of PmaLAAD, a number of natural and synthetic
L-amino acids were fully converted by the purified enzyme
into the corresponding α-keto acids: the fastest conversion
was obtained for 4-nitrophenylalanine. Analogously, starting
from racemic solutions, the full resolution (ee > 99%) was
also achieved.
Keywords: amino acids; biocatalysis; biotransformation;
deracemization; stereoinversion
Introduction
acids from racemic mixtures by deaminating the D-
isomer to the corresponding imino acid[1] and also by
coupling chemical reduction, aimed at converting
the imino acid back into a racemate, the cyclic
deracemization proposed by Soda[2,3] and Turner.[4]
An analogous approach can be used to produce D-
amino acids when an amino acid oxidase with
reverse stereoselectivity is employed. The
flavoenzyme L-amino acid oxidase (LAAO, EC
1.4.3.2) catalyzes the stereoselective oxidative
deamination of L-amino acids into the corresponding
α-keto acids and ammonia; in the second half of the
reaction, the reduced flavin is reoxidized by O2 to
generate H2O2.[5] In that study, however, the
complications associated with overexpression of
snake venom LAAOs in recombinant hosts and the
Biocatalysis represents a well-suited method to
produce enantiomerically pure amino acids for the
pharmaceutical and agricultural industry. As a
general rule, racemic mixtures of amino acids are
less expensive than the corresponding single
enantiomers. The resolution of racemates into the
single enantiomer can thus be of significant
economic interest. Here, the stereoinversion is also
of economic interest since the D-enantiomer is often
more expensive than the corresponding L-form,
especially in the case of natural amino acids, which
may be extracted as single isomers from natural
sources or produced by fermentation (e.g., D-Phe
costs 4-fold more than L-Phe). In past years, the
FAD-containing enzyme D-amino acid oxidase
(DAAO, EC 1.4.3.3) was used to produce L-amino
narrow
substrate
specificity of
microbial
counterparts precluded using the L-selective
1
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700806
Advanced Synthesis & Catalysis
flavooxidase in biocatalysis.[5] Here, a suitable
alternative is represented by L-amino acid deaminase
consist of two main domains formed by several
noncontiguous sequences: a FAD-binding domain and
a substrate-binding domain. In addition to such regions,
LAADs possess an N-terminal putative transmembrane
-helix (whose removal yields the deleted, fully
soluble variant which is almost completely
inactive),[6,9] and a small subdomain (named “insertion
module”) with an + structure, strictly interconnected
to the substrate-binding domain. Details about the
mode of FAD binding, the geometry of the active site,
and the rationale underlying the substrate preference
and stereoselectivity of LAADs have been recently
reported in detail.[6,10]
(LAAD).
This
membrane-associated,
FAD-
containing enzyme catalyzes the O2-dependent
deamination of L-amino acids similarly to LAAO
(and DAAO on the opposite enantiomer), in this
case also yielding the corresponding α-keto acids
and ammonia; however, the electrons are transferred
from the reduced cofactor to a cytochrome-b-like
acceptor without producing any hydrogen peroxide
(see Scheme 1A).[6]
In this work, we selected PmaLAAD as an optimal
catalyst for deracemization and stereoinversion of
selected L-amino acids of biotechnological relevance.
Overexpressed in E. coli cells, the enzyme was easily
purified by a single chromatographic step, recovered
full activity when incubated with membranes and,
notably, did not require exogenous FAD for maximal
activity. The activity of PmaLAAD is highest at pH
7.5 and at 50 °C; indeed, the enzyme is quite stable.[6]
PmaLAAD prefers bulky and hydrophobic substrates,
such as L-Phe, L-Leu, L-Met, and L-Trp, followed by
the polar amino acid L-Cys. It also shows the ability to
deaminate small apolar or charged amino acids at a
low rate.[6] Furthermore, PmaLAAD was also reported
to be active on synthetic and biotechnologically
Scheme 1. A) Reaction catalyzed by PmaLAAD on L-amino
acids. B) Stereoinversion of a L-amino acid into the
relevant
amino
acids,
such
as
L-3,4-
corresponding D-enantiomer by PmaLAAD and
a
dihydroxyphenylalanine (L-DOPA) and substituted
nonselective reduction.
alanines.[6]
When LAADs are separated from membranes, O2
consumption decreases dramatically: the reduced FAD
cofactor of LAAD from Proteus myxofaciens reacts
very slowly with dioxygen with a rate constant of 0.08
s-1 at air saturation.[6]
Results and Discussion
Kinetic properties and docking analysis
LAADs are monomeric enzymes showing a low
sequence identity with canonical homodimeric LAAOs
or DAAOs.[6] LAADs have been identified in Proteus
bacteria only. Two different types of LAAD are
produced by each Proteus species, sharing high
sequence identity (~ 57%) but differing in substrate
specificity. The type-I LAADs preferentially oxidize
large aliphatic and aromatic amino acids - the most
well-known members are those from P. mirabilis
(PmirLAAD) and P. myxofaciens (PmaLAAD) - while
type-II LAADs – e.g., from P. vulgaris (pvLAAD) and
P. mirabilis (Pm1LAAD) - show substantial activity
towards charged (mainly basic) amino acids and L-Phe,
the only hydrophobic amino acid which is also a good
substrate.[6-9]
The kinetic parameters of PmaLAAD on different
substrates were determined under steady state
conditions by the polarographic assay, at pH 7.5 and
25 °C, and at 21% oxygen saturation, see Table 1. The
corresponding Michaelis-Menten plots are depicted in
Figure 1. The enzyme showed a higher activity on L-
Phe, L-Leu, L-Met, and L-4-nitrophenylalanine (L-4-
Npa); the highest catalytic efficiency and apparent
substrate affinity were observed on L-4-Npa and L-Leu
(up to 4-fold in comparison to the reference substrate
L-Phe), see Table 1.
Docking analysis was performed with Autodock
Vina,[11] using the crystallographic structure of
PmaLAAD in complex with the ligand (2-
aminobenzoate) as receptor (PDB code: 5fjn). Docking
analysis shows that all substrates tested can be
accommodated in the active site in a catalytically
competent conformation (i.e., with the αH pointing
toward the N5 atom of the cofactor, see Supporting
Information, Figure S1).[12]
Very recently, the structures of two LAADs were
solved, namely, the type-I PmaLAAD[6] and the type-II
pvLAAD.[9] These proteins share a very similar tertiary
structure: the root-mean-square deviation is 0.72 Å
when the main-chain atoms are taken into
consideration (370 residues). Canonical LAAOs
2
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700806
Advanced Synthesis & Catalysis
These results demonstrate that there is enough space in
the active site of PmaLAAD to accommodate even large
substrates such as L-1-naphthylalanine (L-1-Nal). Only
the substrate L-4-Npa was not docked in a conformation
resembling the one observed for the canonical substrates
such as L-Phe. Interestingly, a small conformation change
of Phe318 allows binding of L-4-Npa in a conformation
suitable for catalysis. Docking to the PmaLAAD with the
altered Phe318 side chain orientation resulted in an
alternative conformation of the bound substrate
appropriate also for further substrates possessing large
side chains (e.g., L-Phe or L-1-Nal, see Supporting
Information, Figure S1A,B and Figure S1I,J). The
binding energies between the two alternative modes of
binding (as estimated by the Autodock Vina scoring
function) are quite close (e.g., -8.1 and -6.9 kcal/mol for
the L-1-Nal). However, we must point out that Phe318
(and Arg316) are located in a very flexible region of the
active site entrance and, thus, it is plausible that
conformational changes in these residues could improve
binding of large substrates, contributing to the broad
substrate specificity of the enzyme.[6]
When the activity of PmaLAAD on the tested
compound was too low to be assayed by the oxygen-
consumption assay (such as with 1-100 mM L-
penicillamine, not shown), reduction of flavin cofactor
under anaerobic conditions represents a very sensitive
alternative detection method.[13] The absorbance
spectrum in the UV-visible region of ~ 10 μM
PmaLAAD was recorded in anaerobic cuvettes in the
absence of E. coli membranes at 25 °C after adding the
compound. Phenylglycine derivatives are amino acids
of main biotechnological interest: here, 0.25 mM of L-
phenylglycine did not reduce PmaLAAD. In contrast,
the cofactor was slowly reduced by adding 0.25 mM L-
4-methoxy-phenylglycine or L-4-methyl-phenylglycine.
In the presence of L-4-methoxy-phenylglycine the
cofactor was fully reduced in 90 min while in the
presence of L-4-methyl-phenylglycine it was not fully
reduced even after 210 min of incubation (see
Supporting Information, Figure S2A,B). Despite the
absence of any detectable O2 consumption in the
turnover enzymatic assay (up to 100 mM substrate),
fully reduced cofactor was also observed after 2-hour
incubation under anaerobic conditions in the presence
of 10 mM L-penicillamine (Supporting Information,
Figure S2C). Overall, PmaLAAD shows the ability to
deaminate various L-amino acids in addition to the
proteinogenic ones.
Figure 1. Michaelis-Menten plots of the activity values
determined on different compounds. For L-1-Nal and L-
homo-Phe the double-reciprocal plot is shown. Bars
indicate SEM for at least three determinations; where not
shown, the error bar is smaller than the symbol used.
Table 1. Kinetic parameters of PmaLAAD on different
substrates.
Substrate
Vmax,app
(U/mg protein)
3.00 ± 0.04
2.66 ± 0.04
1.36 ± 0.09
1.92 ± 0.09
1.18 ± 0.38
2.73 ± 0.02
2.70 ± 0.09
1.51 ± 0.39
Km,app
Vmax,app/Km,app
(mM)
L-Phe
1.60 ± 0.09
0.36 ± 0.03
5.37 ± 1.05
3.82 ± 0.70
0.79 ± 0.27
0.46 ± 0.02
2.80 ± 0.34
2.54 ± 0.24
1.88
7.40
0.25
0.50
1.50
5.95
0.96
0.60
L-4-Npa
L-DOPAa)
D,L-3py-Ala
D,L-1-Nal
L-Leu
Bioconversions of different natural and synthetic L-
and D,L-amino acids
L-Met
D,L-homo-Phe
The activity was determined by the polarographic assay, at 25 °C, pH
7.5 and air saturation. The activity on the corresponding D-amino acid
was negligible. a) The kinetic parameters were determined at pH 7.0
because of the higher stability of DOPA at this value.
The best reaction conditions for L-amino acid
bioconversions were determined analyzing the time
course of the deamination of 25 mM L-Phe using 0.1
mg/mL (corresponding to 0.3 U/mL) of PmaLAAD at
3
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700806
Advanced Synthesis & Catalysis
25 or 37 °C, see Supporting Information. Accordingly,
all bioconversion reactions were conducted at 25 °C
and pH 7.5 (the only exception was L-DOPA).
A preliminary investigation of the deaminase activity
on different natural and synthetic L-amino acids was
performed spectrophotometrically measuring the
formation of the corresponding ketone, see Supporting
Information, Figure S3. Next, the bioconversion results
were verified by measuring the residual L-amino acid
through the ninhydrin method. Finally, under
optimized reaction conditions, the time course of
conversion of different L-amino acids and the
resolution of different racemic mixtures by the
deaminase enzyme were analyzed by HPLC: the
products of the bioconversion reactions were separated
and quantified by reverse-phase HPLC and the
enantiomeric purity was evaluated by chiral HPLC.
As reported in Table S1, ≥ 90% of the substrate was
converted to the corresponding ketone when L-Phe, L-
4-Npa, and L-DOPA (12.5-25 mM) were incubated
with the recombinant enzyme. Interestingly, the
complete conversion of L-4-Npa was achieved in only
30 min (see Figure 2); the high yield and the high rate
of conversion observed for this amino acid derivative
(i.e., 6.8 µmol product/min x mg enzyme) well
correlate to the highest kinetic efficiency and substrate
affinity determined for the deaminase enzyme among
the tested substrates (see Table 1).
When the racemic mixture was used, the chiral HPLC
analysis of all reactions clearly showed that only the L-
enantiomer disappeared in each case (as shown for
D,L-1-Nal in Figure 3) and that the full conversion of
the L-enantiomer was reached for all the tested
substrates (see Table 2 and Figure 2).
Figure 2. Comparison of the time course of bioconversion
of L-isomer of different compounds using 0.1 mg/mL
PmaLAAD, at 25 °C and air saturation. For L-Phe only, the
bioconversion was also carried out by equilibrating the
reaction mixture with pure oxygen (open symbols). For
experimental conditions, see Supporting Information, Table
S1 and Table 2.
Table 2. Resolution of racemic mixtures of different amino
acids by PmaLAAD.
Conversion
yield of
L-enantiomer
(%)
Product formation
Initial
concentration
(mM)
at 90% conversion
Compound
(µmol/min·mg
enzyme)
-
D,L-3py-Alaa)
50
10
1.2
50
50
10
5
~ 75
> 99
1.16
0.41
1.22
-
D,L-1-Nalb)
D,L-Leua)
D,L-Meta)
> 99
> 99
~ 80
> 99
0.68
0.37
D,L-homo-Phea)
~ 95
The bioconversion was carried out at 25 °C, employing 0.1 mg/mL enzyme,
under optimized conditions (2.5 mL, air saturation). The products of the
bioconversion reactions were analyzed by chiral HPLC analysis (detection
at 210 nm). HPLC conditions: a)H2O/MeOH 4:6; b)H2O/MeOH 2:8.
Figure 3. Chiral HPLC analysis of racemic D,L-1-Nal before
(A) and following conversion by PmaLAAD (B, continuous
line), and the subsequent addition of M213G DAAO variant
(B, dashed line). Conditions: 1.2 mM substrate, 0.1 mg
PmaLAAD/mL and 0.015 mg M213G DAAO/mL, pH 7.5,
25 °C (2.5 mL).
4
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700806
Advanced Synthesis & Catalysis
The highest rate of resolution (i.e. the amount of L- acid/min per mg enzyme were produced with a > 99%
amino acid transformed at 90% conversion per minute, yield (see Table 2).
Table 2) was obtained for D,L-3-pyridylalanine (D,L-
3py-Ala) and D,L-Leu as substrates. Interestingly, the
complete resolution of D,L-3py-Ala and D,L-Met
Exhaustive enzymatic deamination of D,L-1-Nal
mixtures was reached when a lower starting amount of Naphtylalanines are aromatic synthetic amino acids
substrate was used (i.e., 10 mM vs. 50 mM): at the used to produce various drugs, such as the peptide drug
highest substrate concentration, only
~
80% nafarelin.[17] When the racemic mixture of D,L-1-Nal
conversion was obtained for the enantiomeric solution (1.2 mM, a concentration close to its maximal
of 3py-Ala and Met while full conversion was solubility) was incubated with PmaLAAD (0.1
achieved for Leu. This observation is in line with the mg/mL) the L-enantiomer was fully consumed, as
lower substrate affinity observed on the two previous shown by the HPLC chromatogram reported in Figure
substrates (see Table 1 and Table 2). Under optimized 3. Then, the DAAO M213G variant (0.015 mg/mL
conditions, the lowest rate was observed for resolution corresponding to 0.3 U/mL on D-1-Nal) was added to
of D,L-homophenylalanine (D,L-homo-Phe) (i.e., 0.37 the reaction mixture produced by the deaminase
µmol product/min x mg enzyme). The slow conversion enzyme:[18] the peak with a t(R) of 24.04 min
can be explained based on the substrate concentration disappeared showing the complete conversion of the
used: owing to the low substrate solubility, the reaction racemic mixture into the ketone product. This result
was performed using 5 mM of racemic mixture (i.e., demonstrates the possibility to selectively trigger the
2.5 mM of each enantiomer), a value close to the Km resolution of a D,L-1-Nal solution into the L- or the D-
value of PmaLAAD for this substrate (see Table 1).
enantiomer by alternatively using the DAAO M213G
The conversion of L-Phe allows to produce variant or the PmaLAAD enzyme, respectively, as well
phenylpyruvic acid, a compound employed as dietary as the possibility to deaminate both enantiomers using
supplement for patients with kidney diseases to reduce the two flavoenzymes simultaneously. Notably, LAAD
urea accumulation or as flavor enhancer in food was also previously employed in multienzymatic
formulations but whose use is hampered by the high reactions. For example, whole E. coli recombinant
cost of production. PmaLAAD better compares to the cells expressing PmirLAAD were used together with a
previous L-Phe bioconversions based on E. coli cells commercial D-transaminase to resolve a racemic
overexpressing selected LAAD or using purified mixture
enzyme preparations (maximal production of 0.51 yl)propanoic acid, reaching
of
2-amino-3(7-methyl-1-H-indazol-5-
a
79% yield;[19]
μmol/min per mg of enzyme).[14] Notably, the L-Phe phenylalanine ammonia lyase and PmirLAAD were
conversion by recombinant PmaLAAD was further used to convert p-nitrocinnamic acid into p-nitro-D-
increased 2.3-fold by substituting air with pure O2: the Phe, resulting in a 71% conversion;[20] lysed E. coli
full L-Phe deamination was reached in ~ 90 min vs. > cells expressing PmaLAAD were used as the first step
300 min when air was used (see Figure 2). The kinetic to convert L-amino acids into the corresponding α-keto
properties of PmaLAAD on L-Phe are significantly acids, which were then asymmetrically reduced into
better than those of the D165K/F263M/L336M (R)- or (S)-2-hydroxy acids in the following step by
evolved PmirLAAD: the kinetic efficiency is 15.6-fold using L- or D-isocaproate reductase activities or
higher for the previous enzyme.[15]
converted into the D-amino acid by an engineered D-
A good production rate of the metabolic intermediate amino acid dehydrogenase.[21,22]
α-ketoisocaproate, a nitrogen-free substitute for Leu D,L-1-Nal is a good substrate for PmaLAAD but the
and a compound used to treat chronic kidney disease volumetric conversion is hampered by its low
and hepatitis B virus infection, could be reached by solubility in aqueous solution. In order to improve the
using PmaLAAD: starting from 50 mM D,L-Leu, 1.22 conversion yield and taking into consideration the
μmol α-ketoisocaproate/min per mg enzyme was higher solubility of the produced ketone, the reaction
produced with a > 99% conversion (Table 2).
was carried out in aqueous dispersions of D,L-1-Nal
E. coli cells overexpressing pvLAAD, both the wild- corresponding to a nominal concentration of 2.4 or 6
type and the K104R/A337S variant, were previously mM, both of them larger than the substrate solubility.
used to convert L-Met into α-keto-γ-methylthiobutyric HPLC analysis of the reaction mixtures (i.e., the
acid (an indirect inhibitor of tumor cell growth and a soluble part) showed that 1.2 and 2.9 mM ketone was
methionine supplement in livestock feed). The formed after 90-120 min, respectively, thus
maximal conversion (~ 90%) was reached after 24 demonstrating that the substrate virtual concentration
hours; the reaction time course is probably hampered in the reaction mixture can be significantly increased.
by the high apparent Km of this enzyme for L-Met (≥ This is not unusual since the use of a saturated solution
240 mM),[16] a value 100-fold higher than that of was reported on the same substrate using DAAO to
PmaLAAD. With the latter enzyme and using 10 mM oxidize the D-enantiomer[23] as well as when
D,L-Met, 0.68 μmol α-keto-γ-methylthiobutyric PmaLAAD was employed on 200 mM L-Tyr (virtual
5
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700806
Advanced Synthesis & Catalysis
concentration since its solubility in buffer is ~ 2.5 Notably, in stereoinversion experiments we employed
mM).[21]
a low amount of reducing equivalents and a “milder,
cheaper, and water-stable” reductant as compared to
NaBH4, a compound largely used but requiring higher
ratios because of its instability in water.[20,26-28]
Alternative reducing agents resulted in a lower
conversion yield, see Supporting Information, page S6.
Interestingly, no α-keto acid formation was observed
in the L-4-Npa or L-DOPA stereoinversion reaction: a
competition between chemical reduction of the
produced imine and its hydrolysis at increasing
substrate concentrations cannot be excluded. Moreover,
no H2O2 is produced during the cofactor reoxidation.
This feature is of utmost relevance: the enzyme
stability is improved compared with oxidases and no
byproducts stemming from oxidative decarboxylation
are formed.[29]
Chemoenzymatic stereoinversion of L-4-Npa
The generation of D-4-Npa (the starting point for the
synthesis of plasmin inhibitors and anti-fibrinolytic
drug production)[24,25] from the corresponding L-
enantiomer was established by coupling the
deamination of the L-amino acid to the corresponding
achiral imino acid and its nonselective in situ reduction
by the borane tert-butylamine complex (Scheme 1B).
A 12.5 mM L-4-Npa solution was incubated with
PmaLAAD under standard conditions and then a 5-
fold molar excess of the reducing agent was added.
Chiral HPLC analysis shows the formation of the D-
amino acid: as shown in Figure 4, after 2 h of
incubation, the complete conversion of L-4-Npa into
the D-enantiomer was apparent (ee > 99%). The D-
Preparative scale
amino acid production stopped to
~
70%
stereoinversion in presence of 5-fold molar excess of
the borane triethylamine complex as reducing agent.
In order to demonstrate the possibility to employ
PmaLAAD on a preparative scale, 130 mg of L-4-Npa
(12.5 mM, 50 mL) was bioconverted. By using 0.1
mg/mL of enzyme and 62.5 mM tert-butylamine
borane complex, the complete conversion into the D-
enantiomer was apparent after overnight incubation,
corresponding to ee > 99% as evaluated by chiral
chromatography.
Conclusion
Enantiomerically pure amino acids, both natural and
synthetic ones, are of increasing interest for the fine
chemicals and pharmaceutical industries. Analogously,
the demand for single-isomer agrochemicals is also
growing due to environmental pressures. In past years,
the deracemization of a D,L-amino acid solution to the
L-isomer was achieved by using the flavoprotein
DAAO, an enzyme used to generate a plethora of
useful variants by protein engineering studies.[1,30] On
the other hand, the isolation of enantiomerically pure
D-amino acids was hampered by the scarce availability
of suitable LAAO activities.[5] Searching for
biocatalytic methods useful for resolving racemic
amino acid mixtures by eliminating the L-enantiomer,
the recombinant LAAD from P. myxofaciens was
proposed in 2001 by Pantaleone et al.[31] Later on, E.
Figure 4. Stereoinversion of L-4-Npa by PmaLAAD and
nonselective reduction (see Scheme 1B). A) HPLC
chromatogram of 12.5 mM L-4-Npa before (continuous line)
and 120 min after adding 0.1 mg PmaLAAD/mL and 62.5
mM borane tert-butylamine complex reducing agent (dashed
line), at 25 °C and pH 7.5. B) Time course of D-4-Npa ()
production in the enzymatic conversion of L-4-Npa () coli cells expressing recombinant LAAD from various
sources were used on different compounds,[14-16,20,21,26-
reported in panel A. HPLC analysis on a C18 column
28]
confirmed the quantitative conversion of the L- into the D-
isomer and the absence of keto acid.
thus demonstrating this to be a feasible approach.
Here, we set up the bioconversion, the deracemization
and the stereoinversion of natural and synthetic amino
In contrast, starting from 25 mM L-DOPA the D-
acids using the purified, recombinant PmaLAAD. The
enantiomer production stopped to ~ 20% amino acid
use of a purified enzyme preparation allowed a faster
stereoinversion:
investigation, see Supporting Information, Figure S4.
this
result
requires
further
L-Phe conversion than with whole cells, as well as a
faster production of α-ketoisocaproate from L-Leu and
6
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700806
Advanced Synthesis & Catalysis
2.0 g of the previous compound (6.2 mmol) were poured
together with 40 mL of 2 M hydrochloric acid. The mixture
was heated to reflux and left to react for 6 h. Then, active
carbon was added and the solution was hot-filtered through
a Celite plug. The solution was concentrated by means of a
rotary evaporator, and the resulting mixture was left
overnight in the refrigerator to completely crystallize. The
following day, the crystals were collected by filtration on a
Hirsch funnel, washed with ice-cold water, vacuum dried,
of α-keto-γ-methylthiobutyric acid from L-Met. The
faster conversion was also coupled to a higher
conversion yield (> 99%).
In our opinion, novel competitive biocatalytic
processes based on LAAD will largely benefit by
generating further improved enzyme variants with
altered substrate specificity. LAAD activity will also
represent a well-suited tool to employ in multi-step
cascade systems, especially in setting up original
synthetic pathways.[32,33]
and
characterized.
[3-(3,4-dihydroxyphenyl)-2-
oxopropanoic acid: 869 mg, yield 71.2%]: 1H-NMR (400
MHz, d6-DMSO)  ppm: 6.26 (s, 1H), 6.70 (d, J=8 Hz, 1H),
6.95 (dd, J=8 Hz, J=2 Hz), 7.35 (d, J=2 Hz, 1H), 8.72 (bs,
1H), 8.91 (bs, 1H), 9.02 (bs, 1H), 12.9 (bs, 1H).
Experimental Section
Chemicals
Enzymes
L-Phe, phenylpyruvic acid (PPA), L-DOPA, D,L-1-Nal, D,L-
Leu, D,L-Met, D,L-DOPA, D,L-homo-Phe, borane
triethylamine complex, borane morpholine complex,
bis(benzonitrile)palladium(II) chloride, and borane tert-
butylamine complex were purchased from Sigma-Aldrich
(Milano, Italy). L-Penicillamine was purchased from TCI
Europe (Zwijndrecht, Belgium). L-4-Npa and D,L-3py-Ala
were kindly supplied by Flamma S.p.A. (Chignolo d’isola,
Recombinant PmaLAAD (the -00N variant) was produced
as stated in Motta et al.[6] The recombinant enzyme was
purified in
a
single step by HiTrap chelating
chromatography; from 1 L fermentation broth, 16 mg
protein were produced. The amount of protein was estimated
based on the absorbance at 457 nm using an extinction
coefficient of 14.17 mM-1 cm-1.[6] The recombinant enzyme
was isolated as a single band at ~ 55 kDa with > 90% purity
as judged by SDS-PAGE analysis. Following reconstitution
with E. coli membranes, the specific activity was 2.9 U/mg
protein on L-Phe as substrate.
Italy).
L-4-Methoxy-phenylglycine
and
L-4-methyl-
phenylglycine were kindly supplied by Prof. Sandro Ghisla.
All other chemicals were of analytical grade and were used
as received.
Recombinant M213G DAAO variant was produced as stated
in Caligiuri et al.[18] The recombinant enzyme was obtained
3-(3,4-Dihydroxyphenyl)-2-oxopropanoic
synthesis[34]
acid with a yield of 3 mg of protein per liter of culture, with a
specific activity of 20 U/mg protein on D-1-Nal as substrate.
In a 50-mL round-bottomed flask, immersed in an oil bath
and equipped with a magnetic stirrer and a water condenser,
2.76 g of 3,4-dihydroxybenzaldehyde (20 mmol) were
poured together with 2.13 g of anhydrous sodium acetate
(26 mmol), 3.04 g of N-acetylglycine (26 mmol) and 10 mL
acetic anhydride. The resulting mixture was heated and
stirred at 80 °C for 4 h, then at 100 °C for 1 h. The resulting
dark brown solution was left at room temperature overnight.
The next day, 10 mL of ice-cold water were added: the
formed yellow crystals were crushed with a glass rod and
filtered on a Buchner funnel, then rinsed with a few
milliliters of cold water until the washings were no longer
dark. The solid was then dried under vacuum, characterized,
and used in the following steps. [4-((2-methyl-5-oxooxazol-
4(5H)-ylidene)methyl)-1,2-phenylene diacetate: 4.57 g,
yield 75.3%].
In a 50-mL round-bottomed flask, immersed in an oil bath
and equipped with a magnetic stirrer and a water condenser,
3.0 g of the previous compound (9.9 mmol) were poured
together with 10 mL water and 10 mL acetone. The resulting
mixture was heated to reflux and left to react for 4 h. Then
active carbon was added and the solution was hot-filtered
through a Celite plug. The filtrate was left in the refrigerator
overnight to allow complete crystallization of the product,
which was suction-filtered the following day. After drying
under vacuum, the light-yellow crystals were characterized
and used in the following step without further purification.
[(Z)-2-acetamido-3-(3,4-diacetoxyphenyl)acrylic acid: 2.71
g, yield 85.4%].
The enzyme (~ 95% pure) yielded a single band at 40 kDa in
SDS-PAGE analysis; DAAO protein concentration was
determined using an extinction coefficient of 12.6 mM-1cm-1
at 455 nm.[18]
Activity assay and kinetic measurements
The activity of PmaLAAD was determined on L-Phe as
substrate in 50 mM potassium phosphate buffer, pH 7.5,
with an oxygen electrode at 25 °C and at air oxygen
saturation (0.253 mM). The enzyme was added with
exogenous membranes prepared from E. coli cells not
expressing PmaLAAD: membranes isolated from 0.3 mg of
cells were added for each microgram of enzyme and the
sample was sonicated (3 cycles of 15 s each, with a 15-s
interval on ice). The enzymatic activity was determined after
incubation at 25 °C for 30 min and defined as micromoles of
oxygen consumed per minute. The activity and kinetic
parameters on different compounds were similarly
determined using a fixed amount of enzyme (prepared as
stated above) and various substrate concentrations (0-50
mM). Activity vs. substrate concentration data were
analyzed according to the classical Michaelis-Menten
equation or modified to account for a substrate inhibition
effect.[35]
Stability was assessed at 25 or 37 °C by incubating the
enzyme in 50 mM potassium phosphate buffer, pH 7.5. At
different times, samples were withdrawn and the residual
activity was assayed using the polarographic method as
described above.
In a 100-mL round-bottomed flask, immersed in an oil bath
and equipped with a magnetic stirrer and a water condenser,
Bioconversion procedures
7
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700806
Advanced Synthesis & Catalysis
References
The reaction mixtures contained 25 mM L-amino acid (if not
stated otherwise) dissolved in 50 mM potassium phosphate
buffer, pH 7.5, and the PmaLAAD enzyme at 0.1 mg/mL
(corresponding to 0.3 U/mL on L-Phe as substrate) in a final
volume of 2.5 mL. The reaction mixtures were incubated at
25 °C in a rotatory mixer and aliquots were withdrawn at
different times for analysis. The time course of the
bioconversion was determined by the ninhydrin assay (see
below) and by HPLC analysis: 20 µL of reaction mixture
was quenched by adding 180 µL of a 1:1 H2O/MeOH
solution containing 0.1% TFA, then centrifuged; the
supernatant was analyzed by HPLC (see below). Product
formation was also followed spectrophotometrically at 320-
360 nm using a Jasco FP-750 spectrophotometer (Jasco
Europe Srl, Cremella, Italy): an aliquot of the reaction
mixture was transferred into a plastic cuvette containing 3 N
NaOH to stop the reaction and allow the color to develop.
[1] L. Pollegioni, G. Molla, Trends Biotechnol. 2011, 29,
276-283.
[2] J.W. Huh, K. Yokoigawa, N. Esaki, K. Soda, Biosci.
Biotech. Biochem. 1992, 56, 2081-2082.
[3] J.W. Huh, K. Yokoigawa, N. Esaki, K. Soda, J.
Ferment. Bioeng. 1992, 74, 189-190.
[4] N. Turner, Curr. Opin. Chem. Biol. 2004, 8, 114-119.
[5] L. Pollegioni, P. Motta, G. Molla, Appl. Microbiol.
Biotechnol. 2013, 97, 9323-9341.
[6] P. Motta, G. Molla, L. Pollegioni, M. Nardini, J. Biol.
Chem. 2016, 291, 10457-10475.
[7] G. Massad, H. Zhao, H. Mobley, J. Bacteriol. 1995,
177, 5878-5883.
[8] J. Baek, J. Seo, O. Kwon, S. Seong, I. Kim, C. Kim, J.
Basic Microbiol. 2011, 51, 129-135.
[9] Y. Ju, S. Tong, Y. Gao, W. Zhao, Q. Liu, Q. Gu, J. Xu,
L. Niu, M. Teng, H. Zhou, J. Struct. Biol. 2016, 195,
306-315.
[10] G. Molla, R. Melis, L. Pollegioni, Biotechnol. Adv.,
2017, 35, 657-668.
[11]O. Trott, A.J. Olson, J. Comp. Chem. 2010, 31, 455-461.
[12] S. Umhau, L. Pollegioni, G. Molla, K. Diederichs, W.
Welte, M.S. Pilone, S. Ghisla, Proc. Natl. Acad. Sci.
U.S.A. 2000, 97, 12463-12468.
[13] C. Harris, G. Molla, M.S. Pilone, L. Pollegioni, J. Biol.
Chem. 1999, 274, 36233-36240.
Ninhydrin assay
The residual amount of amino acid in the reaction mixtures
was estimated using the ninhydrin assay. The aliquots taken
from the reaction were diluted in 800 µL water and 200 µL
of a 0.8% (w/v) ninhydrin solution in acetone added. The
samples were boiled in the dark for 10 min and then placed
in an ice water bath for 5 min. To each sample 200 µL of
50% (v/v) aqueous ethanol was added, the tubes were
vortexed, and the absorbance value at 570 nm was recorded.
In parallel, control samples (not containing substrate or
enzyme) were also analyzed. The amino acid concentrations
were calculated by referring to the corresponding calibration
curve.
[14] Y. Hou, G. Hossain, J. Li, H. Shin, L. Liu, G. Du, Appl.
Micr. Biotechnol. 2015, 99, 8391-8402.
[15] Y. Hou, G. Hossain, J. Li, H. Shin, L. Liu, G. Du, J.
Chen, PLoS One 2016, 11(11).
[16] G. Hossain, J. Li, H. Shin, G. Du, M. Wang, L. Liu, J.
Chen, PLoS One 2014, 9(12).
[17] P. Taylor, D. Pantaleone, R. Senkpeil, I. Fotheringham,
Trends Biotechnol. 1998, 16, 412-418.
HPLC analyses
HPLC analyses were performed on a Jasco apparatus (Jasco
Europe Srl) equipped with a PDA detector and using a Luna
5-µm C18(2) column, length/internal diameter = 150/4.6
mm (Phenomenex, Castel Maggiore, Italy). A binary system
made of 0.1% TFA in H2O (solvent A) and 0.1% TFA in
MeOH (solvent B) was used, flow rate 1 mL/min, column
temperature 35 °C. The retention times t(R) of L-Phe, PPA, L-
4-Npa, 4-nitro-PPA, L-DOPA, and 3-(3,4-dihydroxyphenyl)-
2-oxopropanoic acid were 3.20, 7.10, 3.39, 8.25, 4.64, and
6.52 min, respectively.
[18] A. Caligiuri, P. D'Arrigo, E. Rosini, D. Tessaro, G.
Molla, S. Servi, L. Pollegioni, Adv. Synth. Catal. 2006,
348, 2183-2190.
[19]R.L. Hanson, B.L. Davis, S.L. Goldberg, R.M. Johnston,
W.L. Parker, T.P. Tully, M.A. Montana, R.N. Patel,
Org. Proc. Res. Develop. 2008, 12, 1119-11129
[20]F. Parmeggiani, S.L. Lovelock, N.J. Weise, S.T Ahmed,
N.J. Turner, Angew. Chemie Int. 2015, 54, 4608-4611.
[21] E. Busto, N. Richter, B. Grischek, W. Kroutil,
Chemistry 2014, 20, 11225-11228.
Chiral HPLC analyses were performed on an Agilent
apparatus (Cernusco sul Naviglio, Italy) equipped with a UV
detector set at 210 nm and fitted with an Astec Chirobiotic
[22] F. Parmeggiani, S.T. Ahmed, M.P. Thompson, N.J.
Weise, J.L. Galman, D. Gahloth, M.S. Dunstan, D. Leys,
N.J. Turner, Adv. Synth. Catal. 2016, 358, 3298-3306.
[23] A. Caligiuri, P. D'Arrigo, T. Gefflaut, G. Molla, L.
Pollegioni, E. Rosini, S. Servi, Biocatal. Biotransf.
2006b, 24, 409-413.
[24] S. Saupe, S. Leubner, M. Betz, G. Klebe, T.
Steinmetzer, J. Med. Chem. 2013, 56, 820-831
[25] S. Hinkes, A. Wuttke, S. Saupe, T. Ivanova, S. Wagner,
A. Knorlein, A. Heine, G. Klebe, T. Steinmetzer, J.
Med. Chem. 2016, 59, 6370-6386.
[26]F. Alexandre, D. Pantaleone, P. Taylor, I. Fotheringham,
D. Ager, N. Turner, Tetrahedron Letters 2002, 43, 707-
710
[27]A. Enright, F.R. Alexandre, G. Roff, I.G. Fotheringham,
M.J. Dawson, N.J. Turner, Chem. Comm. 2003, 20,
2636-2637.
TAG column
5 µm (Sigma-Aldrich), length/internal
diameter = 250/4.6 mm, eluent H2O/MeOH, flow rate 0.8
mL/min, and column temperature 30 °C. The retention times
were as follows: D,L-3py-Ala (L-, 10.38 min; D-, 21.17 min),
D,L-1-Nal (L-, 13.71 min; D-, 21.17 min), D,L-Leu (L-, 5.67
min; D-, 11.18 min), D,L-Met (L-, 6.64 min; D-, 12.78 min),
D,L-homo-Phe (L-, 7.48; D-, 16.84 min), and D,L-DOPA (L-,
6.97 min; D-, 24.04 min).
Acknowledgements
Roberta Melis is a PhD student of the “Biotechnology, Biosciences
and Surgical Technology” course at Università degli Studi
dell'Insubria. We acknowledge the support from Fondo di Ateneo
per la Ricerca.
8
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700806
Advanced Synthesis & Catalysis
[28] T. Beard, N. Turner, Chem. Comm. 2002, 3, 246-247.
[29] Z. Findrik, D. Vasic-Racki, M. Primozic, M. Habulin,
Z. Knez, Biocat. Biotransf. 2005, 23, 315-321.
[30] L. Pollegioni, S. Sacchi, L. Caldinelli, A. Boselli, M.S.
Pilone, L. Piubelli, G. Molla, Curr. Protein Pept. Sci.
2007, 8, 600-618.
[31] D. Pantaleone, A. Geller, P. Taylor, J. Mol. Catal. B-
Enzymatic 2001, 11, 795-803.
[32] D. Tessaro, L. Pollegioni, L. Piubelli, P. D’Arrigo, S.
Servi, ACS Catal. 2015, 5, 1604-1608
[33] W.D. Fessner, New Biotechnol. 2015, 32, 658-664
[34] Q.Z. Zhang, Y. Wang, Y.F. Nan, X.Y. Xiong, Q. Pan,
X.L. Zhang, Asian J. Chem. 2013, 25, 7158-7160.
[35] W.W. Cleland in Contemporary enzyme kinetics and
mechanism (Ed.: D.L. Purich), Academic Press, Inc.,
New York, 1983, pp 253-266.
9
This article is protected by copyright. All rights reserved.

10.1002/adsc.201700806
Advanced Synthesis & Catalysis
FULL PAPER
Deracemization and Stereoinversion of α-Amino
Acids by L-Amino Acid Deaminase
Adv. Synth. Catal. Year, Volume, Page – Page
Elena Rosini*, Roberta Melis, Gianluca Molla,
Davide Tessaro, and Loredano Pollegioni
10
This article is protected by copyright. All rights reserved.