Enzymatic synthesis of ω-carboxy-β-hydroxy-(l)-α-amino acids

Article abstract of DOI:10.1016/j.tet.2008.03.070

Commercially available ω-carboxy-aldehydes and glycine have been subjected to the catalytic action of an l-threonine aldolase from Escherichia coli to give the corresponding β-hydroxy-α-(l)-amino acids as a mixture of erythro/threo epimers. Specifically, the reaction with glyoxylic acid (2) gave the epimeric β-hydroxy-(l)-aspartates (t,e)-9 that could be isolated by ion-exchange chromatography in 67% yield. Following esterification and N-Boc protection, the two epimers could be isolated as pure compounds. Similarly, the aldolase-catalyzed addition of glycine to succinic semialdehyde (4) gave the expected mixture of β-hydroxy-l-α-aminoadipic acids (t)-12 and (e)-12 in 34% yield.

Full text of DOI:10.1016/j.tet.2008.03.070

Tetrahedron 64 (2008) 5079–5084
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Enzymatic synthesis of u-carboxy-b-hydroxy-( )-a-amino acids
L
,
,
*
Francesca Sagui a b, Paola Conti b, Gabriella Roda b, Roberto Contestabile c, Sergio Riva a
a Istituto di Chimica del Riconoscimento Molecolare, C.N.R., Via Mario Bianco 9, 20131 Milano, Italy
b
`
Istituto di Chimica Farmaceutica e Tossicologica, Universita degli Studi di Milano, Via Mangiagalli 25, 20133 Milano, Italy
c
`
Dipartimento di Scienze Biochimiche, Universita di Roma ‘La Sapienza’, Piazzale A. Moro 5, 00185 Roma, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Commercially available u-carboxy-aldehydes and glycine have been subjected to the catalytic action of
an -threonine aldolase from Escherichia coli to give the corresponding b-hydroxy-a-( )-amino acids as
a mixture of erythro/threo epimers.
Specifically, the reaction with glyoxylic acid (2) gave the epimeric b-hydroxy-(L)-aspartates (t,e)-9 that
could be isolated by ion-exchange chromatography in 67% yield. Following esterification and N-Boc
protection, the two epimers could be isolated as pure compounds.
Received 13 December 2007
Received in revised form 25 February 2008
Accepted 19 March 2008
L
L
Available online 22 March 2008
Similarly, the aldolase-catalyzed addition of glycine to succinic semialdehyde (4) gave the expected
mixture of b-hydroxy-L-a-aminoadipic acids (t)-12 and (e)-12 in 34% yield.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
NH2 O
HOOC
CHO
n
HO
OH
2 , n = 0
3 , n = 1
4 , n = 2
The amino acid L-glutamate (Glu) is considered to be the major
O
OCH2Ph
mediator of excitatory signals in the mammalian central nervous
system (CNS) and is probably involved both in brain development
and in most aspects of its normal function including cognition,
memory, and learning.1–6 Due to its potential neurotoxicity, the
extracellular concentration of glutamate is tightly controlled by
high-affinity Naþ-dependent excitatory amino acid transporters
(EAATs) that maintain Glu concentration below the excitotoxic
level.5,7 A number of pharmacological agents are capable of inhib-
iting glutamate transport and most of them act as competitive
substrates. At present, attention is focused on non-transportable
inhibitors (or blockers), compounds that, in pathological condi-
tions, could counteract EAAT-mediated glutamate release in the
1
NR'
COOR
O
OH
O
CHO
(t,e)-6 : R = H ; R' = H2
(t,e)-7 : R = Me ; R' = H3+ Cl-
(t,e)-8 : R = Me ; R' = HBoc
5
Generally speaking, b-hydroxy-a-amino acids (i.e.,1) are a group
of interesting compounds that can be found in a number of complex
natural products possessing a wide range of biological activities,
i.e., they can act as antibiotics and immunosuppressants.9 In addi-
tion, these amino acids are useful building blocks in synthetic,
combinatorial, and medicinal chemistries.10 A number of reports
describe the synthesis of these compounds, both as racemates and
as single enantiomers.11 Specifically, optically pure b-hydroxy-
aspartate derivatives (like 1) were prepared with protocols that
required several synthetic steps and the use of toxic reagents, such
as OsO4.8
synaptic cleft. As reported in the literature,8 derivatives of
b-hydroxyaspartate ( -THA) possessing an ethereal bulky sub-
stituent, such as a benzyl or a naphthyl group, function as blockers
at all subtypes of EAATs. In particular, -threo-b-benzyloxy-
aspartate ( -TBOA) is a non-transportable blocker of the gluta-
D,L-threo-
D,L
D,L
D,L
mate transporters, widely recognized as a reference compound for
biological screenings on the physiological roles of these proteins.
The L-isomer is more efficacious than the D-enantiomer and the
threo isomer (1) is more potent as a blocker than its erythro
The biocatalyzed preparation of these compounds has been
suggested as a complement to the existing chemical methods; the
advantages of an enzymatic strategy, usually performed in a single
step process, being the mild conditions and the minimal need for
counterpart.8
substrate protection.12 Accordingly, in recent years,
L- and D-threo-
nine aldolases have been studied for the non-physiological
synthesis of b-hydroxy-a-amino acids using various aldehydes as
* Corresponding author. Tel.: þ39 02 2850 0032; fax: þ39 02 2890 1239.
E-mail address: sergio.riva@icrm.cnr.it (S. Riva).
0040-4020/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2008.03.070

5080
F. Sagui et al. / Tetrahedron 64 (2008) 5079–5084
acceptors and glycine as a donor.13 In order to extend this meth-
odology to the preparation of u-carboxy-b-hydroxy-a-amino acids
optimization of the reaction conditions, since b-hydroxy-a-amino
acids are quite prone to give retroaldolic reactions. The problem
was solved by performing the reaction in a biphasic system in
which the substrate ((t,e)-7) was dissolved in water and the Boc2O
confined in the organic phase. Dropwise addition of a solution of
sodium bicarbonate to the water phase neutralized the hydro-
chloride 7, which was transferred into the organic phase to react
with (Boc)2O producing the expected N-protected derivatives (t,e)-
8. The excess of glycine present in the crude mixture was similarly
derivatized and the new mixture could be easily separated by flash
chromatography on silica gel, allowing the separation of the threo
and erythro b-hydroxy-amino acid derivatives (t,e)-8 from N-Boc-
glycine methyl ester, giving a mixture of the two b-epimers in 16%
isolated yields. The obtained products were characterized by 1H
NMR, which showed the expected presence of the two threo and
erythro diastereoisomers in an approximately 1:1 ratio.
(analogues of 1), we have studied the performances of an
L-threo-
nine aldolase (
L
-TA) from Escherichia coli14 with structurally dif-
ferent aldehydes, and in the following we present the preliminary
results obtained.
2. Results and discussion
Threonine aldolases (TAs) are pyridoxal-50-phosphate (PLP)-
dependent enzymes that catalyze the reversible cleavage of threo-
nine (or allo-threonine) into glycine and acetaldehyde.15 Genes
encoding for these proteins have been found in plants, vertebrates,
and several bacteria, yeasts, and fungi.16
Two types of TAs exist, depending on the configuration of the
newly formed stereocenter at the a-carbon. Both L- and D-TA have
Glyoxylic acid (2) was the first carboxy-aldehyde investigated,
a strict requirement for glycine as a donor, but can accept a wide
range of aliphatic and aromatic aldehydes as electrophilic sub-
as the threo product that could be obtained by
condensation is a direct precursor of the previously cited
(1).8 A methodology capable of producing stereoisomerically pure
-b-hydroxyaspartate in a single step could be very useful to gen-
erate a library of derivatives, differently substituted at the alcoholic
moiety.
L
-TA-catalyzed aldol
L-TBOA
strates.13,17 As shown in Figure 1 for an
L
-TA-catalyzed condensation,
the reactions proceed with complete stereoselective control at the
a-carbon and the outcome is the sole ( )-epimer. On the other hand,
L
L
these enzymes are usually unable to control the stereochemistry at
the b-carbon and produce a mixture of diastereoisomers in mod-
erate yields. Quite recently, both these drawbacks have been over-
come by means of a bi-enzymatic dynamic kinetic asymmetric
transformation based on the consecutive catalysis of a low-specific
Following the previously discussed protocol, glyoxylic acid (2)
was used as a donor in the presence of glycine and, to our satis-
faction, the formation of a more polar product was clearly detected
by TLC. In order to increase the condensation yields, the reaction
was then performed in the presence of an excess (5 equiv) of gly-
cine and stopped, after 48 h, by snap freezing and lyophilized. The
separation of the amino acids from the enzyme and the excess of
glycine present in the crude residue became possible using a strong
ion-exchange resin, by taking advantage of their different pKa
values. The reaction mixture, dissolved in water at neutral pH, was
loaded on a strongly basic ion-exchange column. Glycine was not
kept by the resin and was eluted in the void volume, while,
conversely, the desired products (t,e)-9 were eluted with a 0.1 M
formic acid solution and, following lyophilization, were isolated in
67% yields (Fig. 2).
L
-threonine aldolase and of a highly selective L-tyrosine decarboxy-
lase.18 This methodology allowed the synthesis of a number of
enantiopure amino alcohols.
To the best of our knowledge, threonine aldolases have never
been tested with aldehydes carrying a carboxylic substituent (e.g.,
glyoxylic acid 2, malonic semialdehyde 3, and succinic semi-
aldehyde 4), despite the fact that the enzyme-catalyzed conden-
sation of glycine with these compounds could offer a direct
approach to epimeric b-hydroxy-aspartates, b-hydroxy-glutamates,
and higher homologues.
Threonine aldolases are still not commercially available, and
therefore our investigation started from the production of the
enzyme, fermenting an E. coli strain overexpressing an L-TA and
1H and 13C NMR analyses confirmed the proposed structures
and allowed us to evaluate as 1:1 the threo/erythro ratio of the
products, thus confirming the absence of any b-stereoselectivity in
partially purifying it via ammonium sulfate precipitation and
ion-exchange chromatography (for details see Section 4).
In order to optimize the reaction conditions and the product
recovery, the biocatalyzed reaction was initially tested on benzyl-
oxyacetaldehyde (5), a compound that has been reported to be
O
HO
CHO
H2N
COOH
+
2
a good substrate for these enzymes.13c
L-TA-catalyzed condensation
L-TA
of glycine with 5 gave a product of intermediate polarity ((t,e)-6,
detected by TLC using the ninhydrin reagent). The reaction mixture
was extracted with Et2O to remove the unreacted aldehyde and
lyophilized. The crude residue was further elaborated to new de-
rivatives that could be more easily isolated and characterized. This
aspect was not fully investigated in previous articles as, quite often,
the reported reaction yields and threo/erythro ratios were only
evaluated on the bases of the 1H NMR spectra of the product
mixtures (with or without unreacted glycine). The derivatization
reactions were quite straightforward and consisted in the esterifi-
cation of the carboxylic acid moiety and N-Boc protection of the
O
NH2
COOH
O
NH2
COOH
ion exchange
chromatography
H2N
COOH
HO
HO
+
(unreacted)
OH
OH
(t,e)-9
(t,e)-9 (isolated)
SOCl2 , MeOH
O
NH3+Cl-
COOMe
+ MeOOC
MeO
NH3+Cl-
OH
(t,e)-10
Boc2O
amino group. The latter step, however, required
a careful
(biphasic system)
NaHCO3
NH2
RCHO
+
R
OH
O
NHBoc
COOMe
O
NHBoc
COOMe
L-TA
O
OH
R = alkyl , aryl
MeO
MeO
MeOOC
NHBoc
+
+
H2N
COOH
OH (e)-11
OH (t)-11
Figure 1. L-TA-catalyzed condensation of
b-hydroxy-a-L-amino acids.
a
generic aldehyde and glycine to give
Figure 2. L-TA-catalyzed condensation of glyoxylic acid (3) and glycine, and
subsequent chemical elaboration of the products.

F. Sagui et al. / Tetrahedron 64 (2008) 5079–5084
5081
Figure 3. 1H NMR spectrum of (t,e)-9 in DMSO-d6.
the
L
-TA-catalyzed condensation. The assignment of the signals due
HOOC
HO
H2N
COOH
CHO
+
H
O
COOH
N
to the two epimers (t,e)-9 was made possible by a comparison of
4
the 1H NMR spectrum of the mixture with that due to the com-
L-TA
NH2
mercially available racemic mixture of
(D,L)-threo-b-hydroxy
a-14
NH2
COOH
aspartic acid. The 1H NMR of (t,e)-9 (in DMSO-d6) is reported in
Figure 3. The protons of the threo isomer were downfield shifted
O
HOOC
O
COOH
+
OH
(t,e)-12
HOOC
HO
H
and characterized by a smaller spin-coupling constant (Jthreo
¼
N
13
1.2 Hz) in comparison to the corresponding signals of its epimer
(Jerythro¼8.0 Hz).
O
e-14
As this protocol of purification did not allow the separation of
the two epimers, we transformed the mixture into the corre-
sponding N-Boc-b-hydroxyaspartate methyl esters, (t)-11 and
(e)-11 (Fig. 2), which could be separated by flash chromatography
and characterized by NMR and mass analyses (see Section 4). The
assignment of the absolute configurations of the two products was
made possible by comparison with literature values.8
Figure 4. L-TA-catalyzed condensation of succinic semialdehyde (4) and glycine, and
subsequent chemical elaboration of the products.
Specifically, in this case the threo isomer was formed in slightly
higher amounts, the threo/erythro ratio being 3:2. The assignment
of the absolute configuration was done by comparing the chemical
shifts and the coupling constants of the signals due to these epi-
mers with those of the b-hydroxy-aspartic acid ((t,e)-9). The
structures of the two lactams (a)-14 and (e)-14, formed when the
purified mixture of (t,e)-9 was kept at acid pH, could also be
assigned by mono- and bi-dimensional 1H NMR.
Since b-hydroxy-
never been tested as potential glutamate transporter inhibitors, we
applied the -TA-catalyzed condensation of glycine to succinic
L-a-aminoadipic acids (t)-12 and (e)-12 have
L
semialdehyde 4, which is commercially available. The bio-
transformation was carried out following the usual protocol, using
an excess (5 equiv) of glycine. The reaction mixture was lyophilized
and the crude residue purified from the excess of glycine by ion-
exchange chromatography to give the expected mixture of products
(t,e)-12 in 34% yields.
3. Conclusions
The structural characterization of these compounds was quite
troublesome, as the formed 3-carboxy-b-hydroxy-a-amino acids
(t,e)-12 gave under slightly acidic conditions a mixture of the lac-
tones 13 and of other isomers like the epimeric lactams (a)-14 and
(e)-14, a phenomenon that was already described in the litera-
ture.19 This problem could be avoided by making the solution of the
recovered products alkaline before lyophilization (see Section 4). In
this way, it was possible to confirm by 1H NMR the structure of
compounds (t,e)-12, and the usual lack of b-selectivity (Fig. 4) was
observed.
We have shown that u-carboxy-aldehydes can be used as
substrates in
L-TA-catalyzed aldolic condensations, givingdin one-
stepdu-carboxy-b-hydroxy-
L
-a-amino acids as threo/erythro mix-
tures. The epimeric products could be isolated by ion-exchange
column chromatography and, in the case of the b-hydroxy-aspar-
tates, the two epimers could be also separated after their trans-
formation into more lipophilic derivatives. These compounds can
be used as a starting material for the synthesis of a wide range of
molecules carrying different substituents at the hydroxylic moiety,
which could perform as inhibitors of glutamate transporters.
Work is in progress to extend this protocol to the not com-
mercially available malonic semialdehyde 3 and the results will be
reported in due course.
The ratio between the threo and erythro epimers could be easily
evaluated by comparing the area of the peaks due to the a-protons
of the two isomers, two doublets resonating at 3.29 and 3.16 ppm.

5082
F. Sagui et al. / Tetrahedron 64 (2008) 5079–5084
4. Experimental part
dehydrogenase (32 U) were added into a cuvette to a final volume
of 990 ml. The reactions were started by adding 10 ml of purified -TA
L
4.1. Materials and methods
and monitored by measuring the decrease of absorbance at 340 nm
in a spectrophotometer (3¼6.2ꢁ103 l/mol cm) at 30 ꢀC.
Aldehydes 2, 4, and5 were purchased from Sigma–Aldrich. All the
other reagents and solvents were purchased from commercial sup-
pliers and were used without further purification unless otherwise
One unit of L-TA is the amount of enzyme that catalyzes the
formation of 1 mmol of acetaldehyde (1 mmol of NADH oxidized) per
minute at room temperature.
stated. E. coli transfected with the gene encoding for L-threonine
aldolase was provided by Prof. R. Contestabile, University of Rome.
Fermentations were performed with a 1 l fermentor Inova 4230
and sterilizations were performed using autoclave Steristeam CDL.
Enzymatic activities were monitored using a Jasco V-530 UV/VIS
spectrophotometer. Enzyme purification was partially performed
by HPLC, using a JASCO 880-PU instrument equipped with a Jasco
870 UV detector.
4.4. Synthesis of 2-(S)-amino-3-(R,S)-hydroxy-4-benzyloxy-
butanoic acid (t,e)-6
To a solution of benzyloxyacetaldehyde (5, 685 ml, 5.3 mmol),
glycine (400 mg, 5.3 mmol), pyridoxal phosphate (PLP, 1.7 mg,
0.125 mM), and dithiothreitol (DTT, 24.6 mg, 3 mM) in 50 ml of
H2O, NaOH (1 M) was added up to pH¼7.5.
L-TA (25 U) was re-
Measurements of pH were carried out with a pH-meter Crison
micropH 2000.
covered by centrifugation from the stored suspension, solubilized
in 3 ml of water, and added to the reaction mixture, which was
stirred for 48 h at 45 ꢀC and monitored by TLC (eluent: BuOH–H2O–
AcOH 4:2:1; Rf: (t,e)-6¼0.48, glycine¼0.18, 5¼0.90; detection was
carried out with the ninhydrin reagent).
TLC: silica plates (Merck 60 F254) or reverse phase silica plates
(Merck RP-18254). Substrates and products were visualized at
254 nm and/or by treatment with the ninhydrin reagent (ninhy-
drin, 2 g, in 100 ml MeOH). Flash chromatography silica gel: Merck
60, 40–63 mm.
The reaction mixture was then extracted with Et2O (2ꢁ150 ml)
to remove the unreacted 5 and the water phase lyophylized to give
a yellow powder. The excess of glycine was precipitated by adding
cold MeOH (100 ml), the suspension was filtered, the solvent dried
over Na2SO4, and evaporated to give the epimeric mixture of the
products ((t,e)-6, 146 mg, 13% yields) as a white solid.
1H and 13C NMR spectra were recorded on a Varian Mercury 300
(300 MHz) or on a Bruker AC400 (400 MHz) or on a Bruker AC500
(500 MHz). Mass spectra were recorded with an FT-ICR (Fourier
Transfer Ion Cyclotron Resonance) APEXÔ II model (Bruker Dal-
tonics) equipped with a 4.7 T crio-magnet (Magnex).
1H NMR (300 MHz, DMSO-d6) d (ppm) 3.25 (d, J¼6.0 Hz,
H-2threo), 3.27 (d, J¼3.7 Hz, H-2erythro): total 1H; 3.49 (dd, J1¼
6.0 Hz, J2¼10.1 Hz, H-4a,threo) and 3.53 (dd, J1¼5.8 Hz, J2¼10.1 Hz,
H-4a,erythro): total 1H; 3.57 (dd, J1¼4.1 Hz, J2¼10.1 Hz, H-4b,threo) and
3.59 (dd, J1¼5.8 Hz, J2¼10.1 Hz, H-4b,erythro): total 1H; 3.94 (dt,
J1¼6.0 Hz, J2¼4.1 Hz, H-3threo) and 4.13 (dt, J1¼3.7 Hz, J2¼5.8 Hz,
H-3erythro): total 1H, H-3; 4.53 (2H, s, ArCH2); 7.10–7.40 (5H, m, H–Ar).
4.2. Production and purification of
L-TA from E. coli
L-TA was expressed from E. coli HMS 174 (DE3)/pET 22 b(þ)(
L-TA)
constructs containing the DNA sequence. The cells were cultivated
on LB media (Tripton 10 g/l, yeast extract 5 g/l, NaCl 10 g/l) supple-
mented with 100 mg/l ampicillin and 30 mg/L of vitamin B6. Over-
night cultures (100 ml) in 500 ml flasks were inoculated with single
colonies and grown at 37 ꢀC. The 500 ml main cultures in 2000 ml
flasks were inoculated with 3 ml of the pre-culture and grown at
37 ꢀC to an OD600 of 0.3 and then induced with IPTG at a final con-
centration of 0.05 M. Cultivation was continued overnight at 37 ꢀC
and the cells were harvested by centrifugation for 20 min at
4500 rpm. After resuspension of the pellets (10 ml/g cell) in a lysis
buffer (10 mM Tris/HCl, 1 mM NaCl, 1 mM EDTA, pH 7.6), the cells
were disrupted using a solution of lysozyme (1 mg/g cell) followed
by the precipitation of DNA by adding streptomycin sulfate (10 g/l
dissolved in minimum amount of the same buffer) and leaving the
suspension at 0 ꢀC for 20 min. The crude lysate was cleared by cen-
trifugation at 12,000 rpm for 30 min. The enzyme was precipitated
fromthesupernatant (cellfreeextract) byadding ammoniumsulfate
(80% saturation) and partially purified using an ion-exchange col-
umn (DEAE 650-S, Merck) where the dissolved pellet was loaded in
a phosphate buffer 20 mM pH¼7.5 and the enzyme eluted with
a linear gradient of NaCl, up to 0.4 M. The partially purified enzyme
was stored as a precipitate in ammonium sulfate at 4 ꢀC.
4.5. Synthesis of methyl 2-(S)-tert-butoxycarbonylamino-3-
(R,S)-hydroxy-4-benzyloxy-butanoate (t,e)-8
(a) The crude residue obtained from the previous biocatalyzed
condensation was subjected to the following reaction without
preliminary purification in order to avoid product loss during the
methanolic precipitation of the excess of glycine.
A solution of thionyl chloride (507 ml, 7 mmol) in MeOH (1.1 ml)
was cooled to 0 ꢀC and crude (t,e)-6 (400 mg, 2.3 mmol) was added
in aliquots under stirring. The mixture was left at 0 ꢀC for 1 h and
then at room temperature for 48 h (TLC eluent: BuOH–H2O–ACOH
4:2:1; Rf: (t,e)-6¼0.48, glycine¼0.18, (t,e)-7¼0.60, glycine methyl
ester¼0.25; detection was carried out with the ninhydrin reagent).
Removal of the liquids under vacuum afforded the methyl esters
of (t,e)-6 as a viscous oil, which was subjected to the subsequent
reaction.
(b) A mixture of the methyl esters (t,e)-7 (2.32 mmol) dissolved
in water (1 ml) and EtOAc (4.7 ml) was cooled to 0 ꢀC followed by
the addition of tert-butoxycarbonyl anhydride (Boc2O, 506 mg,
2.32 mmol) under stirring. A saturated solution of NaHCO3 was
added dropwise until the pH of the water phase became slightly
basic. The reaction mixture was allowed to proceed at 0 ꢀC for 2 h
and then at room temperature overnight (TLC eluent: AcOEt–
petroleum ether 4:6; Rf: (t,e)-7¼0.10, (t,e)-8¼0.61, N-Boc-glycine
methyl ester¼0.70; detection was carried out with ninhydrin).
The organic fraction was separated and the water phase
extracted with EtOAc (3ꢁ50 ml). The organic phases were collected
together, washed with brine (2ꢁ20 ml), and dried over Na2SO4. The
solvent was removed in vacuo to give the epimeric mixture (t,e)-8
as a viscous colorless oil that was purified by silica gel flash chro-
matography (eluent: petroleum ether–AcOEt 8:2) to give 80 mg of
(t,e)-8 (0.25 mmol, 16% yields calculated on the amount of 5 used
In order to improve the yields of recovered L-TA, cell growing
was performed in a 1 l capacity fermentor (INFORS AG CH-4103),
following the same protocol developed for the suspended cell
systems, monitoring oxygen supply, pH, temperature, and stirring.
This experiment gave satisfactory results, asdfollowing cells lyses,
ammonium sulfate precipitation, and subsequent dialysisd688 U
of
L-TA were obtained, to be compared with the 125 U obtained
with the previously described flask cultures.
4.3. L-TA activity assay
L-Allo-threonine (50 mM), sodium phosphate buffer (20 mM,
pH 7), PLP (50 mM), NADH (200 mM), and yeast alcohol

F. Sagui et al. / Tetrahedron 64 (2008) 5079–5084
5083
for the biocatalyzed aldolic condensation). The solid was charac-
terized by mono- and bi-dimensional 1H and 13C NMR analyses,
which confirmed the structure and the 1:1 ratio of the two
diastereomers (threo/erythro).
stirring. A saturated solution of NaHCO3 was added dropwise until
the pH of the water phase became slightly basic. The reaction was
allowed to proceed at 0 ꢀC for 2 h and then at room temperature
overnight, following the conversion by TLC (eluent: AcOEt–petrolem
ether 4:6; Rf: (t)-11¼0.35, (e)-11¼0.44; detection was carried out
with ninhydrin). The organic phase was separated and the water
phase extracted with AcOEt (3ꢁ50 ml); the organic phases were
collected, washed with brine (2ꢁ20 ml), and dried over Na2SO4.
The solvent was removed in vacuo to give the epimeric threo/
erythro mixture (t,e)-11 that could be purified by silica gel flash
chromatography (eluent: petroleum ether–AcOEt 85:15, then pe-
troleum ether–AcOEt 70:30) affording the two pure epimers (t)-11
(55 mg, 0.20 mmol) and (e)-11 (60 mg, 0.22 mmol) as viscous
colorless oils.
1H NMR (300 MHz, CDCl3) d (ppm): 1.46 (s, 9H, Boc), 3.53–3.61
(m, 2H, H-4erythro,threo), 3.68 (s, 3H, OMe), 3.75 (s, 3H, OMe), 4,19 (br
m, 1H, H-3threo), 4.25–4.42 (br m, 2H, H-2erythro,threo), 4.42–4.52 (m,
3H, H-3erythro, ArCH2threo), 4.54 (s, 2H, ArCH2erythro), 5.35 (s, 1H, NH),
5.56 (s,1H, NH), 7.10–7.40 (m, 5H, H–Ar). 13C NMR (300 MHz, CDCl3)
d (ppm) 29.0 ((CH3)3C); 52.3 (OCH3); 55.6 and 57.2 (C-2erythro,threo);
70.5 and 71.2 (C-4); 73.8 (ArCH2); 80.4 and 80.7 (C-3); 128.2, 128.8
and 137.9 (C–Ar); 156.2 and 156.4 ((CH3)3C–C]O); 171.2 and 171.5
(COOMe).
4.6. Synthesis of 2-(S)-amino-3-(R,S)-hydroxy-succinic acid
(t,e)-9
The two epimers were characterized by mono- and bi-
dimensional 1H NMR, 13C NMR, and mass analyses, which con-
firmed the proposed structures. The relative configuration (threo/
erythro) of the two products could be assigned by comparison with
the literature values.8
To a solution of glyoxylic acid (2, 250 mg, 2.71 mmol), glycine
(5 equiv, 1018 mg, 13.60 mmol), pyridoxal phosphate (PLP, 2 mg,
0.125 mM), and dithiothreitol (DTT, 32 mg, 3 mM) in 65 ml H2O,
Compound (e)-11. 1H NMR (400 MHz, CDCl3), d (ppm): 1.47 (s,
9H, Boc), 3.82 (s, 3H, OMe), 3.84 (s, 3H, OMe), 4.71 (br s, 1H, H-3),
4.80 (d, 1H, J¼9.2 Hz, H-2), 5.31 (br d, 1H, NH). 13C NMR (300 MHz,
CDCl3), d (ppm,): 28.50 ((CH3)3C); 53.23 (OCH3); 56.51 (OCH3); 71.34
(C-3); 56.44 (C-2); 155.0,170,173.0 (3C]O). FABþ-MS: 278 (MþH)þ.
Compound (t)-11. 1H NMR (400 MHz, CDCl3), d (ppm): 1.44 (s,
9H, Boc), 3.73 (s, 3H, OMe), 3.83 (s, 3H, OMe), 4.52 (br s, 1H, H-3),
4.84 (d, 1H, J¼7.6 Hz, H-2), 5.51 (br d, 1H, NH). 13C NMR (300 MHz,
CDCl3), d (ppm): 28.58 ((CH3)3C); 53.05 (OCH3); 53.25 (OCH3); 57.37
(C-2); 72.42 (C-3); 155.68, 169.18, 172.05 (3C]O). FABþ-MS: 278
(MþH)þ.
NaOH (1 M) was added up to pH¼7.5.
L-TA (60 U) was recovered by
centrifugation from the storing suspension, solubilized in 5 ml
water, and added to the reaction mixture, which was stirred at
45 ꢀC for 48 h and monitored by TLC (eluent: BuOH–H2O–AcOH
4:2:1; Rf: (t,e)-9¼0.15, glycine¼0.18; detection was carried out with
the ninhydrin reagent). The mixture was then lyophilized to give
a yellow powder that was either purified by ion-exchange chro-
matography or derivatized as previously described.
The mixture was purified using a strong basic ion-exchange
resin (Ionac A-540, CIꢂ form) previously conditioned as follows:
a column (diameter 2 cm, height 42 cm) was filled with 40 g of
resin that was previously washed twice with 1 M NaOH, water up to
neutral pH, 2 M AcOH, and water up to neutral pH. The crude
products were then dissolved in water at pH 7.5 and loaded onto
the column. The resin was initially eluted with water to remove the
excess of glycine and then washed with 0.1 M formic acid to elute
the product. The fractions containing the product were collected
and lyophylized to give the epimeric mixture (t,e)-9 as a white
powder (272 mg, 1.83 mmol, 67% yields calculated on the amount
of 2 initially used for the biotransformation).
4.8. Synthesis of 2-(S)-amino-3-(R,S)-hydroxy-6-carboxy-
hexanoic acid (t,e)-12
To a solution of hemisuccinic aldehyde (4, 1.76 ml, 2.71 mmol),
glycine (1.02 g, 13.57 mmol), pyridoxal phosphate (PLP, 4.25 mg,
0.125 mM), and dithiothreitol (DTT, 40 mg, 3 mM) in 13.6 ml of
H2O, NaOH (1 M) was added up to pH¼7.5.
L-TA (40 U) was re-
covered by centrifugation from the storing suspension, solubilized
in 4 ml water, and added to the reaction mixture, which was stirred
for 48 h at 45 ꢀC and monitored by reverse phase TLC (eluent:
MeOH–H2O–AcOH 95:5:0.5; Rf: (t,e)-12¼0.73, glycine¼0.50;
detection was carried out with ninhydrin).
The product was characterized by mono- and bi-dimensional
1H and 13C NMR, which also allowed the evaluation of the
diastereoisomeric threo/erythro ratio (1:1) of (t,e)-9.
1H NMR (400 MHz, DMSO-d6) d (ppm): (e)-9: 3.78 (d, 1H,
J¼8.0 Hz, H-2), 4.06 (d, 1H, J¼8.0 Hz, H-3); (t)-9: 3.98 (d, 1H,
J¼1.2 Hz, H-2), 4.26 (d, 1H, J¼1.2 Hz, H-3). 13C NMR (300 MHz,
DMSO-d6) d (ppm): (e)-9: 55.16 (C-2); 70.71 (C-3); 168.08 and
171.77 (C-1 and C-4). (t)-9: 54.19 (C-2); 67.06 (C-3); 168.63 and
172.67 (C-1 and C-4).
The product mixture was purified using a strong basic ion-
exchange resin (Ionac A-540, Clꢂ form) previously conditioned with
the following procedure: a column (diameter 2 cm, height 42 cm)
was filled with 40 g of resin that was previously washed twice with
1 M NaOH, water up to neutral pH, 2 M AcOH, and water up neutral
pH. The product mixture was dissolved in water at pH 7.5 and
loaded onto the column; the resin was initially eluted with water to
remove the excess of glycine and then washed with 0.1 M formic
acid to elute the product. The recovered solution was lyophilized to
give the epimeric mixture of the desired products (t,e)-12 (163 mg,
0.92 mmol, 34% yields). The material was redissolved in water, the
pH was adjusted to 11 using 1 M NaOH, and then the solution was
lyophilized again to give a white solid that was analyzed by mono-
4.7. Synthesis of dimethyl 2-(S)-tert-butoxycarbonylamino-3-
(R,S)-hydroxy-succinate (t,e)-11
A solution of thionyl chloride (1.45 ml, 19.8 mmol) in MeOH
(3.14 ml) was cooled to 0 ꢀC and the crude residue of the bio-
catalyzed condensation ((t,e)-9, 500 mg, 6.6 mmol) was added in
aliquots under stirring. The reaction was allowed to proceed at 0 ꢀC
for 1 h and then at room temperature for 48 h (TLC eluent: BuOH–
H2O–AcOH 4:2:1; Rf: (t,e)-9¼0.15, dimethyl esters of (t,e)-9¼0.20;
detection was carried out with ninhydrin). Removal of the liquids in
vacuo afforded the dimethyl esters (t,e)-10 as a viscous oil, which
was immediately used for the subsequent Boc protection reaction.
The previously obtained mixture of the dimethyl esters of
(t,e)-10 (6.6 mmol) was dissolved in water (1.5 ml) and AcOEt
(10 ml). The biphasic solution was cooled to 0 ꢀC and tert-butoxy-
carbonyl anhydride (Boc2O, 1.4 g, 6.6 mmol) was added under
1
and bi-dimensional H and 13C NMR.19
Compounds (t,e)-12. 1H NMR (400 MHz, D2O), d (ppm): 1.60–1.80
(m, 4H, CH2-4threo,erythro), 2.20–2.40 (m, 4H, CH2-5threo,erythro), 3.16
(d, 1H, J¼6.5 Hz, H-1threo), 3.29 (d, 1H, J¼6.5 Hz, H-1erythro), 3.75 (m,
1H, H-2erythro), 3.79 (m, 1H, H-2threo). 13C NMR (400 MHz, D2O),
d (ppm): threo: 29.71 (C-4); 33.82 (C-5); 59.81 (C-2); 72.69 (C-3);
180.47 (C-1); 182.48 (C-6); erythro: 27.58 (C-4); 33.68 (C-5); 60.11
(C-2); 72.92 (C-3); 179.55 (C-1); 182.58 (C-6).
The two epimeric lactams ((a)-14, (e)-14), spontaneously
formed in acidic solution, were also characterized by mono- and

5084
F. Sagui et al. / Tetrahedron 64 (2008) 5079–5084
bi-dimensional 1H and 13C NMR. 1H NMR (400 MHz, D2O), d (ppm):
(a)-14: 1.51–1.59 (m, 1H, H-4eq), 1.82–1.87 (m, 1H, H-4ax), 2.21–2.26
(m, 2H, H-5ax,eq), 3.12 (d, 1H, J¼8.9 Hz, H-2), 3.49 (ddd, 1H,
J1¼4.0 Hz, J2¼10.5 Hz, J3¼9.0 Hz, H-3), 6.83 (br s, 1H, NH); (e)-14:
1.61–1.71 (m, 2H, H-4eq,ax), 1.94–2.07 (m, 1H, H-5eq), 2.12–2.19 (m,
1H, H-5ax), 3.41 (q, 1H, J¼1.5 Hz, H-2), 4.05 (q, 1H, J¼4.0 Hz, H-3),
6.63 (br s, 1H, NH). 13C NMR (400 MHz, D2O), d (ppm): (a)-14 and
(e)-14: 27.13, 27.10 (C-4); 28.56, 29.20 (C-5); 58.82, 59.13 (C-2);
63.62, 67.31 (C-3); 168.82, 169.59 (C-1); 172.14, 172.60 (C-6).
8. (a) Shimamoto, K.; Sakai, R.; Takaoka, K.; Yumoto, N.; Nakajima, T.; Amara, S. G.;
Shigeri, Y. Mol. Pharmacol. 2004, 65, 1008–1015; (b) Shimamoto, K.; Shigeri, Y.;
Yasuda-Kamatani, Y.; Lebrun, B.; Yumoto, N.; Nakajima, T. Bioorg. Med. Chem.
Lett. 2000, 10, 2407–2410.
9. Saeed, A.; Young, D. W. Tetrahedron 1992, 48, 2507–2514 and references
therein.
10. Vassilev, V. P.; Uchiyama, T.; Kajimoto, T.; Wong, C. H. Tetrahedron Lett. 1995, 36,
5063–5064.
11. (a) Cardillo, G.; Gentilucci, L.; Tolomelli, A.; Tomasini, C. Synlett 1999, 1727–
1730; (b) Hanessian, S.; Vanasse, B. Can. J. Chem. 1993, 71, 1401–1406; (c) De
Angelis, M.; Campiani, G. Tetrahedron Lett. 2004, 45, 2355–2357.
12. See, for instance: (a) Riva, S. Trends Biotechnol. 2006, 24, 219–226; (b) Monti, D.;
Candido, A.; Cruz Silva, M. M.; Kren, V.; Riva, S.; Danieli, B. Adv. Synth. Catal.
2005, 347, 1168–1174; (c) Riva, S. J. Mol. Catal. B: Enzym. 2002, 19 and 20, 43–54;
(d) Carrea, G.; Riva, S. Angew. Chem., Int. Ed. 2000, 39, 2226–2254 and refer-
ences therein.
Acknowledgements
13. (a) Steinreiber, J.; Fesko, K.; Reisinger, C.; Schurmann, M.; van Assema, F.;
Wolberg, M.; Mink, D.; Griengl, H. Tetrahedron 2007, 63, 918–926; (b) Nishide,
K.; Shibata, K.; Fujita, T.; Kajimoto, T.; Wong, C. H.; Node, M. Heterocycles 2000,
52, 1191; (c) Vassilev, V. P.; Uchiyama, T.; Kajimoto, T.; Wong, C. H. Tetrahedron
Lett. 1995, 36, 4081–4084.
14. Contestabile, R.; Paiardini, A.; Pascarella, S.; Di Salvo, M. L.; D’Aguanno, S.;
Bossa, F. FEBS J. 2001, 268, 6508–6525.
15. Percudani, R.; Peracchi, A. Embo Rep. 2003, 4, 850–854.
We thank Dr. Martino Luigi di Salvo (Rome University),
Dr. Daniela Monti, Dr. Lara Baratto, and Mrs. Federica Loiacono
(ICRM-CNR) for their valuable technical assistance in the pro-
duction of L-TA and in the work-up of the biocatalyzed reactions.
Thanks are also due to Mr. Walter Panzeri (ICRM-CNR) for recording
the NMR spectra.
16. Liu, J. Q.; Dairi, T.; Itoh, N.; Kataoka, M.; Shimizu, S.; Yamada, H. J. Mol. Catal. B:
Enzym. 2000, 10, 107–115.
References and notes
17. (a) Kimura, T.; Vassilev, V. P.; Shen, G. J.; Wong, C. H. J. Am. Chem. Soc. 1997, 119,
11734–11742; (b) Shibata, K.; Shingu, K.; Vassilev, V. P.; Nishide, K.; Fujita, T.;
Node, M.; Kajimoto, T.; Wong, C. H. Tetrahedron Lett. 1996, 37, 2791–2794.
18. (a) Steinreiber, J.; Schurmann, M.; Wolberg, M.; van Assema, F.; Reisinger, C.;
Tesko, K.; Mink, D.; Griengl, H. Angew. Chem., Int. Ed. 2007, 46, 1624–1626; (b)
Steinreiber, J.; Schuermann, M.; van Assema, F.; Wolberg, M.; Fesko, K.;
Reisinger, C.; Mink, D.; Griengl, H. Adv. Synth. Catal. 2007, 349, 1379–1386.
19. Jackson, B. G.; Pedersen, S. W.; Fisher, J. W.; Misner, J. W.; Gardner, J. P.; Staszak,
M. A.; Doecke, C.; Rizzo, J.; Aikins, J.; Farkas, E.; Trinkle, K. L.; Vicenzi, J.;
Reinhard, M.; Kroeff, E. P.; Higginbotham, C. A.; Gazak, R. J.; Zhang, T. Y.
Tetrahedron 2000, 56, 5667–5677.
1. Danbolt, N. C. Prog. Neurobiol. 2001, 65, 1–105.
2. Pellicciari, R.; Costantino, G.; Macchiarulo, A. Pharm. Acta Helv. 2000, 74, 231–
237.
3. Rao, V. R.; Finkbeiner, S. Trends Neurosci. 2007, 30, 284–291.
4. Schkeryantz, J. M.; Kingston, A. E.; Johnson, M. P. J. Med. Chem. 2007, 50, 2563–
2568.
5. Kanai, Y.; Hediger, M. A. Eur. J. Pharmacol. 2003, 479, 237–247.
6. Beart, P. M.; O’Shea, R. D. Br. J. Pharmacol. 2007, 150, 5–17.
7. Levi, G.; Raiteri, M. Trends Neurosci. 1993, 16, 415–419.