Chemoenzymatic synthesis of glutamic acid analogues: Substrate specificity and synthetic applications of branched chain aminotransferase from Escherichia coli

Article abstract of DOI:10.1021/jo070805q

(Chemical Equation Presented) A new route to α-keto acids is described, based on the ozonolysis of enol acetates obtained from α-substituted β-keto esters. Escherichia coli branched chain aminotransferase (BCAT) activity toward a variety of substituted 2-oxoglutaric acids was demonstrated analytically. BCAT was shown to have a broad substrate spectrum, complementary to that of aspartate aminotransferase, and to offer access to a variety of glutamic acid analogues. The usefulness of BCAT was demonstrated through the synthesis of several 3- and 4-substituted derivatives.

Full text of DOI:10.1021/jo070805q

Chemoenzymatic Synthesis of Glutamic Acid Analogues: Substrate
Specificity and Synthetic Applications of Branched Chain
Aminotransferase from Escherichia coli
Mo Xian, Se´bastien Alaux, Emmanuelle Sagot, and Thierry Gefflaut*
UMR 6504, UniVersite´ Blaise Pascal, F-63177 Aubie`re Cedex, France
thierry.gefflaut@uniV-bpclermont.fr
ReceiVed April 17, 2007
A new route to R-keto acids is described, based on the ozonolysis of enol acetates obtained from
R-substituted â-keto esters. Escherichia coli branched chain aminotransferase (BCAT) activity toward a
variety of substituted 2-oxoglutaric acids was demonstrated analytically. BCAT was shown to have a
broad substrate spectrum, complementary to that of aspartate aminotransferase, and to offer access to a
variety of glutamic acid analogues. The usefulness of BCAT was demonstrated through the synthesis of
several 3- and 4-substituted derivatives.
Introduction
potent selective agonist of kainate receptors.10 Glu analogues
are also of major interest in biologically active peptide mim-
ics.11,12 Finally, as Glu and pyro-Glu are commonly used as
starting material for enantioselective total syntheses, the addition
of substituents and chiral centers on the Glu skeleton offers
valuable new synthons as exemplified by the recent synthesis
of new thalidomide derivatives.13
In that context, we have been developing a chemoenzymatic
approach for the synthesis of Glu analogues for several years.
It is based on the use of aspartate aminotransferase (AAT) for
the stereoselective conversion into Glu analogues of substituted
R-ketoglutaric acid (KG) readily prepared by chemical synthe-
sis.14-18 AAT has been shown to be a broad substrate spectrum
Glutamic acid is the major excitatory neurotransmitter within
the central nervous system of vertebrates, where numerous Glu
receptors and transporters have been identified.1-5 Glu analogues
behaving as selective ligands are useful tools for elucidating
the specific roles of these receptors. Glu analogues may also
have some therapeutic effects on several mental disorders where
the glutamatergic system is implicated:3,6-10 for example, the
constrained bicyclic LY354740 is an agonist of group II
metabotropic Glu receptors and has been recognized as a lead
structure for the development of anxiolytics.
(4) Bridges, R. J.; Esslinger, C. S. Pharmacol. Ther. 2005, 107, 271-
285.
(5) Shigeri, Y.; Seal, R. P.; Shimamoto, K. Brain Res. ReV. 2004, 45,
250-265.
(6) Foster, A. C.; Kemp, J. A. Curr. Opin. Pharmacol. 2006, 6, 7-17.
(7) Gardoni, F.; Di Luca, M. Eur. J. Pharm. 2006, 545, 2-10.
(8) Bra¨uner-Osborne, H.; Egebjerg, J.; Nielsen, E. O.; Madsen, U.;
Krogsgaard-Larsen, P. J. Med. Chem. 2000, 43, 2609-2645.
(9) Schoepp, D. D.; Jane, D. E.; Monn, J. A. Neuropharmacology 1999,
38, 1431-1476.
(10) Gu, Z.-Q.; Hesson, D. P.; Pelletier, J. C.; Maccecchini, M.-L.; Zhou,
L.-M.; Skolnick, P. J. Med. Chem. 1995, 38, 2518-2520.
(11) Mutucumarana, V. P.; Acher, F.; Straight, D. L.; Jin, D.-Y.; Stafford,
D. W. J. Biol. Chem. 2003, 278, 46488-46493.
Simple changes in Glu structure can afford some binding
selectivity as exemplified by (2S,4R)-4-methyl Glu, which is a
(12) Milne, C.; Powell, A.; Jim, J.; AlNakeeb, M.; Smith, C. P.;
Micklefield, J. J. Am. Chem. Soc. 2006, 128, 11250-11259.
(13) Yamada, T.; Okada, T.; Sakaguchi, K.; Ohfune, Y.; Ueki, H.;
Soloshonok, V. A. Org. Lett. 2006, 8, 5625-5628.
(1) Watkins, J. C.; Jane, D. E. Br. J. Pharmacol. 2006, 147, S100-
S108.
(2) Hertz, L. Neurochem. Int. 2006, 48, 416-425.
(3) Kew, J. N. C.; Kemp, J. A. Psychopharmacology 2005, 179, 4-29.
(14) Passerat, N.; Bolte, J. Tetrahedron Lett. 1987, 28, 1277-1280.
10.1021/jo070805q CCC: $37.00 © 2007 American Chemical Society
Published on Web 08/24/2007
7560
J. Org. Chem. 2007, 72, 7560-7566

Chemoenzymatic Synthesis of Glutamic Acid Analogues
SCHEME 1. AAT-Catalyzed Synthesis of Glutamic Acid
Analogues
SCHEME 2. Synthesis of 3-Phenyl-KG 1m
enzyme allowing the preparation of a variety of 4-substituted
Glu analogues. As depicted in Scheme 1, AAT from pig heart
or E. coli was used for the preparation of both isomers of L-4-
hydroxy Glu,14 of a variety of L-2,4-syn-4-alkyl Glu15 as well
as 4,4-disubstituted derivatives.16 Unfortunately, AAT is inactive
toward 3-substituted derivatives with the exception of the
3-methyl analogue which is a moderately good substrate.
Recently, the scope of AAT was extended to the constrained
cyclobutane Glu analogues (L-CBG): three stereomers L-CBG-
II to IV were prepared with AAT while the fourth isomer
L-CBG-I was obtained using another enzyme: branched chain
aminotransferase (BCAT) from Escherichia coli.17,18 BCAT
accepts Glu as the amino donor in the biosynthesis of the
branched chain amino acids valine, leucine, and isoleucine.
BCAT has also been shown to be active toward phenylalanine
and methionine. It is thus a broad spectrum enzyme and has
been previously used in the preparation of several unnatural
aminoacids19 including L-tert-leucine, present in several biologi-
cally active pseudopeptides.
SCHEME 3. Synthesis of KG Analogues by Ozonolysis of
Enol Acetates
We present here the chemical synthesis of several KG
analogues, an analytical study of BCAT substrate specificity,
and the preparation of Glu analogues using BCAT and AAT.
SCHEME 4. BCAT Activity Assay
Results and Discussion
Synthesis of 3-Substituted 2-Oxoglutaric Acids. 3-Phenyl
KG 1m was first prepared according to Scheme 2 using a
previously described method based on the Claisen-Johnson
rearrangement.15
Allylic alcohol 4 was readily prepared following a described
procedure including a thermodynamically favored rearrangement
of the Baylis-Hillman adduct 3 formed from methyl acrylate
and benzaldehyde.20 The conversion of 4 into 1m was done
with a good overall yield of 56%. This methodology was
previously shown to be very efficient for preparing 4-substituted
KG from various orthoesters.15 However, with the exception
of compound 4, the various allyl alcohols needed for the
synthesis of 3-substituted derivatives are not easily prepared,
and another strategy was therefore developed as shown in
Scheme 3. The Michael addition of acetoacetate onto substituted
acrylates was performed in basic conditions to afford acetyl
glutarates 7n-p. These â-keto esters were then converted to
their enol acetates 8n-p using AcCl and pyridine. Ozonolysis
of the compounds 8 then gave the R-ketoesters 6n-p that were
finally converted to the lithium salts isolated in overall yield of
21-30%.
The formation of the enol acetate intermediates 8 proved to
be necessary: attempts to achieve the direct oxidative cleavage
of the enolates formed from compounds 7 in basic medium
resulted in low yields (<10%) of R-keto ester, while R-hy-
droxylation of the â-ketoesters 7 was always the main reaction
observed (data not shown).
The new synthetic method described here for compounds
1n-p appears very general and should provide access to a
variety of new KG analogues as well as to other R-keto acids
in the near future.
Substrate Specificity of BCAT. Kinetic constants for the
transamination reaction of KG were determined according to
Scheme 4.
(15) Alaux, S.; Kusk, M.; Sagot, E.; Bolte, J.; Jensen, A. A.; Brauner-
Osborne, H.; Gefflaut, T.; Bunch, L. J. Med. Chem. 2005, 48, 7980-7992.
(16) Helaine, V.; Rossi, J.; Gefflaut, T.; Alaux, S.; Bolte, J. AdV. Synth.
Catal. 2001, 343, 692-697.
(17) Gu, X.; Xian, M.; Roy-Faure, S.; Bolte, J.; Aitken, D. J.; Gefflaut,
T. Tetrahedron Lett. 2006, 47, 193-196.
(18) Faure, S.; Jensen, A. A.; Maurat, V.; Gu, X.; Sagot, E.; Aitken, D.
J.; Bolte, J.; Gefflaut, T.; Bunch, L. J. Med. Chem. 2006, 49, 6532-6538.
(19) Hwang, B. Y.; Cho, B. K.; Yun, H.; Koteshwar, K.; Kim, B. G. J.
Mol. Catal. B: Enzym. 2005, 37, 47-55.
Leucine (Leu) was used as the amino donor substrate in quasi-
saturating concentration (40 mM). 4-Methyl-2-oxopentanoic acid
(20) Mason, P. H.; Emslie, N. D. Tetrahedron 1994, 50, 12001.
J. Org. Chem, Vol. 72, No. 20, 2007 7561

Xian et al.
TABLE 1. Kinetic Parameters of BCAT-Catalyzed
Transaminations with KG Analoguesa
(MOPA) was reduced back to Leu in a non-rate-limiting step
in the presence of NADH, NH4+, and leucine dehydrogenase
(LeuDH). The optical density decrease consecutive to NADH
oxidation was measured at 340 nm to monitor the transamination
reaction. Rate measurements at various substrate concentrations
made it possible to estimate an apparent Michaelis constant (Km)
and maximal velocity (Vm ) kcat[BCAT]) in the defined
experimental conditions on the basis of Michaelis-Menten
model for enzyme-catalyzed reactions. In the case of the non-
natural KG analogues, the same method was used, but Glu and
glutamic dehydrogenase (GluDH) were used instead of Leu and
LeuDH. Indeed, GluDH proved to be a much cheaper enzyme
and showed no activity toward any KG analogues. For every
substrate studied, a titration was performed with the same
enzymatic system to provide evidence of a possible enzyme
enantiopreference: NADH consumption corresponding to half
of the theoretical limiting concentration of KG analogue would
indicate that only one enantiomer within the racemate is a good
substrate for BCAT. Obviously, this method was appropriate
only to demonstrate a high enantioselectivity coefficient (E >
50). Results are reported in Table 1.
BCAT displays a markedly high activity with 3-methyl KG
1n. Despite a drop in affinity and kcat when the substituent size
increases, even the bulky 3-phenyl derivative 1m remains a good
substrate with relative kcat of 75%. Enantioselectivity was shown
only for 3-propyl and 3-phenyl KG 1p and 1m, whereas
complete conversion of racemic 3-methyl and 3-ethyl KG 1n
and 1o was observed. 4-Alkyl derivatives are also very good
substrates of BCAT with, again, a negative effect of the
substituent bulkiness. No enantiopreference was observed for
any 4-alkyl substrates. Finally, 4,4-disubstituted KG analogues
1k and 1l are readily converted to Glu analogues, again without
enantioselectivity toward 1l.
These results contrast to those previously obtained with AAT,
which showed good activity only with 4-alkyl KG, stereose-
lectively converted into L-2,4-syn-4-alkyl Glu.15 E. coli BCAT
thus appears as a broader spectrum enzyme allowing access to
various L-Glu analogues including 3-substituted Glu, as well
as L-2,4-anti-4-alkyl Glu or (2S,4R)-4-hydroxy-4-methyl Glu
that could not be prepared previously with AAT. In the case
where BCAT shows no enantiopreference, the diastereomeric
products should be separable, or alternatively, an enantiopure
keto acid can be used as the substrate of the transamination
reaction, as shown in a following paragraph.
Preparation of Glu Analogues. To confirm the analytical
data relative to BCAT activity and enantioselectivity, preparative
scale transamination reactions were done on a millimole scale.
As transamination reactions are characterized by equilibrium
constants near unity, large scale reactions require an equilibrium
shift strategy. AAT offers an opportunity through the use of
cysteine sulfinic acid (CSA), a close analogue of Asp, as the
amino donor substrate: CSA is converted into pyruvyl sulfinic
acid, which spontaneously decomposes into pyruvic acid, which
is not a substrate for AAT.21 As CSA is not a substrate of BCAT,
another equilibrium shift strategy had to be found. Three
different procedures were used with various KG analogues 1.
In all cases, Glu or Leu was used as a catalytic amino donor
substrate (0.1 molar equiv), and an irreversible regeneration
system allowed the equilibrium shift. Glu was preferably
regenerated in a coupled transaminase system using AAT and
a Values and standard errors were calculated from the Hanes-Woolf
plot according to the least-squares method and Gauss’s error propagation
law. b The absolute kcat value measured in our experimental conditions was
345 ( 10 min-1 c Enantioselectivity toward racemic KG substrate.
.
CSA as the primary stoichiometric amino donor (Scheme 5).
Alternatively, GluDH, NH4+, and NADH were also used for
Glu regeneration. NADH itself was used in a catalytic amount
and regenerated using the irreversible oxidation of HCO2- ions
catalyzed by formate dehydrogenase (FDH) as shown in Scheme
6.22 Finally, when used as the catalytic donor, Leu was
regenerated using the same procedure, using LeuDH in place
of GluDH. The choice of the transamination procedure was
dictated by purification constraints. Indeed, 3-methyl or 3-ethyl
(21) Jenkins, W. T.; D’Ari, L. Biochem. Biophys. Res. Commun. 1966,
22, 376-82.
(22) Tishkov, V. I.; Popov, V. O. Biomol. Eng. 2006, 23, 89-110.
7562 J. Org. Chem., Vol. 72, No. 20, 2007

Chemoenzymatic Synthesis of Glutamic Acid Analogues
SCHEME 5. Equilibrium Shift of BCAT Transamination
Using AAT and CSA
The results obtained for the different transamination assays
are reported in Table 2.
In accordance with analytical results, mixtures of diastereo-
mers were obtained in good yields from 1n, 1o, and 1j
substrates. In these three cases, a small scale transamination
reaction was run and stopped near 45% conversion. As judged
by NMR analysis, mixtures of diastereomers were obtained with
de of <1%, 24%, and 31% from 1j, 1n, and 1o respectively.
These results confirmed the low enantiopreference of BCAT
toward these substrates. On the contrary, 3-propyl and 3-phenyl
Glu analogues 2p and 2m were both isolated as a single isomer.
The (2S,3R) configuration was unambiguously attributed to the
3-phenyl analogue 2m in comparison to the physicochemical
properties and spectroscopic data previously described for this
compound.23,24 The new 3-propyl analogue 2p possesses the
L-2,3-syn-configuration. The (2S,3R)-configuration of 2p was
demonstrated by oxidative decarboxylation of the residual
unreacted ketoacid 1p after the BCAT catalyzed transamination.
The known (S)-2-propylsuccinic acid 9 was identified,25 thus
showing that BCAT displays a high enantioselectivity toward
(R)-1p.
SCHEME 6. Equilibrium Shift of BCAT Transamination
with GluDH and FDH
Finally, to exemplify the stereochemical complementarity of
AAT and BCAT, we decided to develop a two-transamination-
step strategy as described in Scheme 7 to prepare separately
both stereomers of 3-methyl, 4-methyl, and 4-hydroxy-4-methyl-
L-Glu 2n, 2b, and 2l.
The racemic ketoacids 1b, 1n, and 1l were first subjected to
an AAT-catalyzed transamination reaction in the presence of
CSA. The reaction was monitored by enzymatic titration of
pyruvic acid formed from CSA using NADH and lactic
dehydrogenase. The reaction was stopped near 50% conversion
to perform the kinetic resolution of the racemic KG substrate.
Indeed, as previously mentioned, AAT was shown to display a
high enantiopreference for (R)-1b and for (S)-1l.16,26 A high
stereoselectivity was also observed in favor of (R)-1n. The
amino acid product was selectively adsorbed on a short column
of strongly acidic resin, and residual KG analogue was then
submitted to a second transamination reaction catalyzed by
Glu 2n and 2o proved difficult to separate from Glu. Leu was
thus preferred as the amino donor for the synthesis of both
derivatives. In contrast, 2p, 2m, and 2j were readily separated
from catalytic Glu by ion exchange chromatography. Indeed,
these analogues showed stronger interactions with basic Dowex-
type resins presumably through hydrophobic interactions. The
AAT-CSA regeneration system was preferred in the case of 2m
and 2p. However, as 1j is a substrate of AAT, the GluDH/
NADH/FDH system and excess ammonium formate were used
for the preparation of 2j.
TABLE 2. Equilibrium-Shifted BCAT-Catalyzed Transamination of KG Analogues
J. Org. Chem, Vol. 72, No. 20, 2007 7563

Xian et al.
SCHEME 7. Two-Step Transamination Process Using AAT and BCAT
propionic acid (0.04 mL, 0.5 mmol). The mixture was heated under
reflux for 1 h before methanol was slowly removed by azeotropic
distillation. The solution was concentrated under reduced pressure.
Flash chromatography (eluent, cyclohexane-AcOEt, 9:1, v/v)
afforded 5 (1.1 g, 80%), isolated as a colorless liquid: IR (neat
film) 1728, 1709, 1633 cm-1; 1H NMR (400 MHz, CDCl3) δ 7.32-
7.20 (5H, m), 6.35 (1H, s), 5.69 (1H, s), 4.47 (1H, t, J ) 8.0 Hz),
4.08 (2H, q, J ) 7.0 Hz), 3.69 (3H, s), 2.93 (1H, dd, J ) 8.0 and
15.5 Hz), 2.81 (1H, dd, J ) 8.0 and 15.5 Hz), 1.18 (3H, t, J ) 7.0
Hz); 13C NMR (100 MHz, CDCl3) δ 171.4, 166.7, 142.5, 141.2,
128.5, 127.9, 126.9, 124.5, 60.5, 52.1, 42.7, 39.6, 14.1. Anal. Calcd
for C15H18O4: C, 68.68, H, 6.92. Found: C, 68.52, H, 6.97.
1-Methyl-5-ethyl 3-Phenyl-2-oxoglutarate 6m. To a solution
of 5 (1 g, 3.8 mmol) in a mixture of CCl4 (10 mL) and CH3CN (10
mL) were added a solution of NaIO4 (3.24 g, 15.2 mmol) in water
(15 mL) and RuO2 (54 mg, 0.4 mmol). The mixture was stirred at
room temperature for 24 h and then filtered on celite. The organic
layer was isolated, and the aqueous phase was extracted with CH2-
Cl2. The combined organic layers were washed with brine, dried
over MgSO4, and concentrated under reduced pressure. Flash
chromatography (eluent, cyclohexane-AcOEt, 8:2, v/v) afforded
6m (710 mg, 70%) isolated as a colorless liquid: IR (neat film)
1731 cm-1 (broad); 1H NMR (400 MHz, CDCl3) δ 7.36-7.26 (5H,
m), 4.98 (1H, dd, J ) 5.0 and 10.5 Hz), 4.12 (2H, m), 3.79 (3H,
s), 3.28 (1H, dd, J ) 10.5 and 17.0 Hz), 2.72 (1H, dd, J ) 5.0 and
17.0 Hz), 1.23 (3H, t, J ) 7.0 Hz); 13C NMR (100 MHz, CDCl3)
δ 191.6, 171.4, 160.7, 134.6, 129.2, 128.8, 128.2, 61.0, 53.0, 49.4,
37.1, 14.1. Anal. Calcd for C14H16O5: C, 63.63, H, 6.10. Found:
C, 63.49, H, 5.97.
Dilithium 3-Phenyl-2-oxoglutarate 1m. To a solution of 6m
(660 mg, 2.5 mmol) in MeOH (13 mL) was added dropwise a 0.4
M aqueous solution of LiOH (13 mL, 5.25 mmol). The mixture
was stirred at room temperature for 3-4 h until complete hydrolysis.
After evaporation of MeOH, the pH of the aqueous solution was
adjusted to 7.6 by the addition of Dowex 50X8 resin (H+ form).
The resin was removed by filtration before evaporation of the water
under reduced pressure. 1m (575 mg, 98%) was isolated as a white
solid: 1H NMR (400 MHz, D2O) δ 7.39-7.25 (5H, m), 4.66 (1H,
t, J ) 7.0 Hz), 2.89 (1H, dd, J ) 7.0 and 15 Hz), 2.56 (1H, dd, J
) 7.0 and 15 Hz); 13C NMR (100 MHz, D2O) δ 204.4, 180.0, 169.9,
136.5, 129.1, 128.7, 127.7, 51.8, 39.1; HRMS (TOF MS ES-) m/z
calcd for C11H9O5 221.0450, found 221.0472.
Dimethyl 3-Methyl-2-oxoglutarate 6n. A solution of 8n (775
mg, 3 mmol) in CH2Cl2 (10 mL) was treated with a mixture of O2
and O3 bubbling at a rate of 10 L/h until saturation (blue coloration
of the solution). After 30 min, the excess ozone was eliminated by
oxygen bubbling. Dimethylsulfure (0.33 mL, 4.5 mmol) was added,
and the reaction mixture was allowed to warm to room temperature.
After dilution with CH2Cl2 (15 mL), the solution was washed with
water (20 mL) and brine (20 mL), dried over MgSO4, and
BCAT. Leu and the LeuDH/NADH/FDH regeneration system
were used in the case of 1n as already described. Glu and AAT/
CSA were used for 1l; in that case, 2l was easily separated from
Glu on cationic Dowex 1 resin. Although the diastereomeric
excess of (2S,4R)-2l isolated after the BCAT catalyzed reaction
was only 80%, this chromatography allowed a partial separation
of diastereoisomers of 2l and an increase of the diastereomeric
excess of (2S,4R)-2l up to 98%. Another strategy was tried with
1b, by using 4-benzyl-Glu 2j as the catalytic amino donor with
the AAT/CSA regeneration system. Indeed, contrary to Glu, 2j
proved to be easily separated from 4-methyl Glu 2b by anion
exchange chromatography on Dowex 1 resin. Practically,
catalytic amounts of 4-benzyl KG 1j, BCAT, and AAT were
added, after pH adjustment, to the crude mixture of CSA and
residual (S)-1b, with neither compound being adsorbed on the
strongly acidic resin. This procedure proved to be very efficient,
and (2S,4S)-2b was thus isolated in good yield and with very
good de.
Conclusion
We have developed a new method for the preparation of the
keto acid substrates of aminotransferases. This method appeared
efficient and general and should allow the synthesis of many
new derivatives in the near future. We have shown that E. coli
BCAT is a useful tool for the stereoselective preparation of
glutamic acid analogues from the corresponding substituted
ketoglutaric acids, especially 3-substituted derivatives that could
not be obtained before with AAT. Furthermore, BCAT stereo-
selectivity is complementary to that of AAT and both enzymes
offer access to both stereomers of a variety of L-glutamic acids
substituted at position 4. These compounds were prepared with
a high purity, suitable for biological evaluations.
Experimental Section
1-Methyl-5-ethyl 2-Methylidene-3-phenyl-glutarate 5. To a
solution in anhydrous toluene (25 mL) of methyl 2-hydroxymethyl-
3-phenylacrylate 4 (1.0 g, 5.2 mmol) prepared following a described
procedure20 were added triethyl orthoacetate (1.3 g, 7.8 mmol) and
(23) Cai, C.; Soloshonok, V. A.; Hruby, V. J. J. Org. Chem. 2001, 66,
1339-1350.
(24) Herdeis, C.; Kelm, B. Tetrahedron 2003, 59, 217-229.
(25) Bettoni, G.; Cellucci, C.; Berardi, F. J. Heterocycl. Chem. 1980,
17, 603-605.
(26) Echalier, F.; Constant, O.; Bolte, J. J. Org. Chem. 1993, 58, 2747-
2750.
7564 J. Org. Chem., Vol. 72, No. 20, 2007

Chemoenzymatic Synthesis of Glutamic Acid Analogues
concentrated under reduced pressure. Flash chromatography (eluent,
The pH of the solution was adjusted to 7.6 with 1 M NaOH, and
the volume was adjusted to 50 mL before E. coli BCAT (2 IU)
and E. coli AAT (10 IU) were added. The reaction was stirred
slowly at room temperature and monitored by titration of pyru-
vate: 5 µL aliquots of the reaction mixture were added to 995 µL
of 0.1 M potassium phosphate buffer, pH 7.6 containing NADH
(0.2 mM), and rabbit muscle lactate dehydrogenase (1 IU). Pyruvate
concentration was calculated from the ∆DO measured at 340 nm
using ꢀNADH ) 6220 M-1 cm-1. When a conversion rate of 40%
was reached, the reaction mixture was rapidly passed through a
column of Dowex 50X8 resin (H+ form, 25 mL). The column was
then washed with water (100 mL) until complete elution of CSA
and then eluted with 1 M NH4OH. The ninhydrin positive fractions
were combined and concentrated to dryness under reduced pressure.
The residue was diluted in water (5 mL), and the pH was adjusted
to 7.0 with 1 M NaOH before adsorption of the product on a column
of Dowex 2X8 resin (200-400 Mesh, AcO- form, 1.5 cm × 12
cm). The column was washed with water (50 mL) and then eluted
with an AcOH gradient (0.1-0.5M). L-Glutamic acid was eluted
first and completely separated from (2S,3R)-2m isolated as a white
solid (56 mg, 38% from rac-1m) and with a diastereomeric excess
over 98% as judged by NMR: mp 160 °C (lit. mp 158.0-
CH2Cl2) afforded 6n (0.33 g, 58%) isolated as a colorless liquid:
1
IR (neat film) 1735 cm-1 (broad); H NMR (400 MHz, CDCl3) δ
3.87 (3H, s), 3.66 (1H, qdd, J ) 5.0, 7.0 and 9.0 Hz),3.63 (3H, s),
2.81 (1H, dd, J ) 9.0 and 17.0 Hz), 2.48 (1H, dd, J ) 5.0 and
17.0 Hz), 1.18 (3H, d, J ) 7.0 Hz); 13C NMR (100 MHz, CDCl3)
δ 195.8, 172.1, 161.2, 53.0, 51.9, 38.2, 36.8, 15.8. Anal. Calcd for
C8H12O5: C, 51.06, H, 6.43. Found: C, 51.34, H, 6.48. HRMS
(TOF MS ES+) m/z calcd for C8H12O5Na 211.0582, found
211.0594;
Dimethyl 2-Acetyl-3-methylglutarate 7n. To a solution of
methyl crotonate (2.0 g, 20 mmol) and methyl acetoacetate (4.65
g, 40 mmol) in MeOH (40 mL) was added a 1 M solution of
MeONa in MeOH (4 mL). The mixture was heated to reflux for 1
day before the addition of the MeONa solution was repeated (4
mL). After 1 more day of refluxing, methyl crotonate was added
again (4.0 g, 40 mmol), and heating was continued for 1 day. After
cooling, the solution was diluted with brine (50 mL) and extracted
with EtOAc (4 × 30 mL). The combined organic layers were dried
over MgSO4 and concentrated under reduced pressure. Flash
chromatography (eluent, cyclohexane-AcOEt, 7:3, v/v) afforded
a 1:1 diastereomeric mixture of 7n (5.7 g, 66%) isolated as a
colorless liquid: IR (neat film) 1735 cm-1 (broad); 1H NMR (400
MHz, CDCl3) δ 3.72 (3H, s), 3.71 (3H, s), 3.66 (2 × 3H, s), 3.56
(1H, d, J ) 8.0 Hz), 3.48 (1H, d, J ) 8.0 Hz), 2.74 (2 × 1H, m),
2.50-2.20 (2 × 2H, m), 2.23 (2 × 3H, s), 0.99 (2 × 3H, d, J )
7.0 Hz); 13C NMR (100 MHz, CDCl3) δ 202.5, 202.4, 172.4, 169.2,
63.69, 63.6, 52.30, 52.26, 51.5, 38.4, 38.2, 29.9, 29.7, 29.6, 17.7,
17.3; HRMS (TOF MS ES+) m/z calcd for C10H16O5Na 239.0895,
found 239.0897.
158.5 °C23, 197 °C24); [R]25D ) +16.9° (c 0.5, 6 N HCl) (lit. [R]25
D
) +16.7° (c 1.36, 6 N HCl)23; +16.7° (c 0.6, 6 N HCl)24); H
1
NMR (400 MHz, D2O, LiOH) δ 7.32-7.20 (5H, m), 3.67 (1H, d,
J ) 6.5 Hz), 3.46 (1H, m), 2.62 (2H, m); 13C NMR (100 MHz,
D2O) δ 179.6, 174.6, 138.0, 128.9, 128.3, 127.8, 59.9, 44.1, 40.3.
Anal. Calcd for C11H13NO4, 1.2 H2O: C, 53.96, H, 6.34, N, 5.72.
Found: C, 54.14, H, 6.28, N, 5.66.
(2S,3R)- and (2S,3S)-3-Methylglutamic Acid 2n. Method A:
Using E. coli BCAT. To a mixture of 1n (50 mg, 0.29 mmol),
L-leucine (4 mg, 0.03 mmol), NADH (2.0 mg, 0.003 mmol), and
ammonium formate (36 mg, 0.58 mmol) in H2O (10 mL) at pH
7.6 were added a 10 mM solution of pyridoxal phosphate (0.1 mL,
0.1 µmol), BCAT (1 IU), leucine dehydrogenase (2 IU), and formate
dehydrogenase (3 IU). The solution was stirred at room temperature
and monitored by titration of 1n: 5 µL aliquots of the reaction
mixture were added to 995 µL of 0.1 M potassium phosphate buffer,
pH 7.6 containing L-glutamic acid (40 mM), (NH4)2SO4 (50 mM),
NADH (0.2 mM), BCAT (0.04 IU), and glutamic dehydrogenase
(2 IU). The concentration of 1n was calculated from the ∆DO
measured at 340 nm using ꢀNADH ) 6220 M-1 cm-1. After 24 h
the reaction mixture was passed through a column of Dowex 50X8
resin (H+ form, 10 mL). The column was washed with water
(30 mL) and then eluted with 1 M NH4OH. The ninhydrin positive
fractions were combined and concentrated to dryness under reduced
pressure. The residue was diluted with water (5 mL), and the pH
was adjusted to 7.0 with 1 M NaOH before adsorption of the
product on a column of Dowex 2X8 resin (200-400 mesh,
AcO- form, 1.5 cm × 12 cm). The column was washed with water
(50 mL) and then eluted with 0.2 M AcOH. 2n was isolated as a
white solid (34 mg, 73%) as a 48:52 mixture of (2S,3R) and (2S,3S)
Dimethyl 2-(1-Acetoxyethylidene)-3-methylglutarate 8n. To
a solution of 7n (1.1 g, 5 mmol) in anhydrous pyridine (20 mL)
was added acetyl chloride (1.1 mL, 15 mmol). The mixture was
stirred at 45 °C for 3 days. Addition of acetyl chloride (0.55 mL,
7.5 mmol) was repeated at 24 and 48 h. After cooling, the solution
was diluted with Et2O (80 mL), washed with water (50 mL) and
with an aqueous saturated solution of CuSO4 (50 mL), dried over
MgSO4, and concentrated under reduced pressure. Flash chroma-
tography (eluent, cyclohexane-AcOEt, 8:2, v/v) afforded 8n (0.83
g, 64%), isolated as a colorless liquid: IR (neat film) 1762, 1735,
1
1658 cm-1; H NMR (400 MHz, CDCl3) δ 3.77 (3H, s), 3.65
(3H, s), 3.24 (1H, hex, J ) 7.0 Hz), 2.54 (1H, dd, J ) 6.0 and
15.0 Hz), 2.48 (1H, dd, J ) 7.0 and 15.0 Hz), 2.20 (3H, s), 2.11
(3H, s), 1.09 (3H, d, J ) 7.0 Hz); 13C NMR (100 MHz, CDCl3) δ
172.9, 168.2, 167.8, 153.1, 125.6, 51.5, 39.6, 29.1, 20.8, 19.1, 18.5;
HRMS (TOF MS ES+) m/z calcd for C12H18O6Na 281.1001, found
281.0997.
Dimethyl 3-Methyl-2-oxoglutarate 6n. A solution of 8n (775
mg, 3 mmol) in CH2Cl2 (10 mL) was treated with a mixture of O2
and O3 bubbling at a rate of 10 L/h until saturation (blue coloration
of the solution). After 30 min, the excess ozone was eliminated by
oxygen bubbling. Dimethylsulfure (0.33 mL, 4.5 mmol) was added,
and the reaction mixture was allowed to warm to room temperature.
After dilution with CH2Cl2 (15 mL), the solution was washed with
water (20 mL) and brine (20 mL), dried over MgSO4, and
concentrated under reduced pressure. Flash chromatography (eluent,
CH2Cl2) afforded 6n (0.33 g, 58%) isolated as a colorless liquid:
1
diastereomers as determined by H NMR.
Method B: Using AAT and BCAT. To a mixture of 1n (100
mg, 0.58 mmol) and CSA (90 mg, 0.58 mmol) in water (20 mL) at
pH 7.6 was added AAT (50 IU). The reaction mixture was stirred
at room temperature and monitored by titration of 1n as previously
described in method A. When a conversion rate near 50% was
reached in 32 h, the reaction mixture was passed through a column
of Dowex 50X8 resin (H+ form, 10 mL). The column was washed
with water (30 mL), and the fractions containing unreacted (S)-1n
were pooled and concentrated under reduced pressure after adjusting
the pH to 7.6 with 1 M NaOH. The column was then eluted with
1 M NH4OH, and (2S,3R)-2n was further purified and isolated as
previously described in method A (43 mg, 46%) with a diastere-
omeric excess of 96% as judged by 1H NMR. Crude residual
(S)-1n was used in a BCAT-catalyzed transamination reaction as
described previously in method A. (2S,3S)-2n was isolated as a
white solid (25 mg, 27%) with a diastereomeric excess of 86%.
1
IR (neat film) 1735 cm-1 (broad); H NMR (400 MHz, CDCl3) δ
3.87 (3H, s), 3.66 (1H, qdd, J ) 5.0, 7.0 and 9.0 Hz),3.63 (3H, s),
2.81 (1H, dd, J ) 9.0 and 17.0 Hz), 2.48 (1H, dd, J ) 5.0 and
17.0 Hz), 1.18 (3H, d, J ) 7.0 Hz); 13C NMR (100 MHz, CDCl3)
δ 195.8, 172.1, 161.2, 53.0, 51.9, 38.2, 36.8, 15.8. Anal. Calcd for
C8H12O5: C, 51.06, H, 6.43. Found: C, 51.34, H, 6.48. HRMS
(TOF MS ES+) m/z calcd for C8H12O5Na 211.0582, found
211.0594;
(2S,3R)-3-Phenylglutamic Acid 2m. To a solution of 1m (0.154
g, 0.66 mmol) in water (45 mL) were added cysteine sulfinic acid
(CSA) (100 mg, 0.66 mmol), L-glutamic acid (19 mg, 0.13 mmol),
and a 10 mM solution of pyridoxal phosphate (0.5 mL, 0.5 µmol).
J. Org. Chem, Vol. 72, No. 20, 2007 7565

Xian et al.
Data for (2S,3R)-2n: mp 161-163 °C (lit. mp 166.0-168 °C27,
Dowex 50X8 resin (H+ form, 25 mL). The column was then washed
with water (100 mL) until complete elution of excess CSA and
then eluted with 1 M NH4OH. The ninhydrin positive fractions were
combined and concentrated to dryness under reduced pressure. The
residue was diluted in water (5 mL), and the pH was adjusted to
7.0 with 1 M NaOH before adsorption of the product on a column
of Dowex 2X8 resin (200-400 mesh, AcO- form, 1.5 cm × 12
cm). The column was washed with water (50 mL) and then eluted
with 0.2 M AcOH. The ninhydrin positive fractions were concen-
trated under reduced pressure to give (2S,4S)-2b (123 mg, 44%)
isolated as a white solid and with a diastereomeric excess over 98%.
Physical and spectroscopic properties of (2S,4S)-2b were consistent
with the previously reported data.26 Anal. Calcd for C6H11NO4, 0.33
H2O: C, 43.13, H, 7.03, N, 8.38. Found: C, 42.77, H, 6.91,N,
8.83. Consecutive elution of the column with 2 M AcOH afforded
an equimolar diastereomeric mixture of 2j (20 mg).
(2S,4R)- and (2S,4S)-4-Benzylglutamic Acid 2j. To a mixture
of 1j (50 mg, 0.2 mmol), l-glutamic acid (6 mg, 40 µmol), NADH
(1.5 mg, 2 µmol), and ammonium formate (32 mg, 0.4 mmol) in
H2O (10 mL) at pH 7.6 were added a 10 mM solution of pyridoxal
phosphate (0.1 mL, 0.1 µmol), BCAT (2 IU), glutamic dehydro-
genase (3 IU), and formate dehydrogenase (4 IU). The solution
was stirred at room temperature and monitored by titration of 1j
as previously described for the synthesis of 2n, method A. After 3
h, the reaction was stopped ,and 2j was purified as described for
2m; 2j was isolated as a white solid (28 mg, 59%) and as a 1:1
mixture of (2S,4R) and (2S,4S) stereomers: 1H NMR (400 MHz,
D2O) δ 7.40-7.20 (2 × 5H, m), 3.72 (1H, dd, J ) 5.5 and 8.5
Hz), 3.68 (1H, dd, J ) 6.0 and 8.5 Hz), 3.00-2.80 (2 × 3H, m),
2.22 (1H, ddd, J ) 5.5, 9.5 and 15.0 Hz), 2.08 (1H, ddd, J ) 6.5,
10.0 and 15.0 Hz), 1.96 (1H, ddd, J ) 4.5, 8.0 and 14.5 Hz), 1.92
(1H, ddd, J ) 4.5, 9.0 and 15.0 Hz); 13C NMR (100 MHz, D2O ,
LiOH) δ 183.6, 183.3, 140.3, 140.2, 128.9, 128.4, 126.2, 48.2, 48.0,
38.8, 38.3, 37.4, 37.0. Anal. Calcd for C12H15NO4, 0.33 H2O: C,
59.25, H, 6.49, N, 5.76. Found: C, 59.20, H, 6.46, N, 5.60.
170-173 °C24); [R]25 ) +18.2 (c 0.9, 6 N HCl) (lit. [R]25
)
D
D
+22.6 (c 1.03, 6 N HCl)27); 1H NMR (400 MHz, D2O, 1 M NaOH)
δ 3.14 (1H, d, J ) 4.0 Hz), 2.22 (1H, dd, J ) 5.0 and 13.0 Hz),
2.18 (1H, m), 1.98 (1H, dd, J ) 9.5 and 13.0 Hz), 0.79 (3H, d, J
) 7.0 Hz); 13C NMR (100 MHz, D2O) δ 182.4, 182.3, 60.1, 42.5,
34.9, 13.7. Anal. Calcd for C6H11NO4, 0.33 H2O: C, 43.13, H,
7.03, N, 8.38. Found: C, 43.01, H, 6.99, N, 8.25.
Data for (2S,3S)-2n: mp 171-173 °C (lit. mp 169.5.0-170 °C28,
169-171 °C27, 182-184 °C24); [R]25 ) +42.0 (c 0.9, 6 N HCl)
D
(lit. [R]25D ) +41.9 (c 1.03, 6 N HCl),24 +42.8 (c 1.0, 6 N HCl),28
+36.8 (c 1.0, 6 N HCl)27); 1H NMR (400 MHz, D2O, 1 M NaOH)
δ 2.96 (1H, d, J ) 6.0 Hz), 2.26 (1H, dd, J ) 4.0 and 13.5 Hz),
1.96 (1H, m), 1.79 (1H, dd, J ) 11.0 and 13.0 Hz), 0.81 (3H, d, J
) 7.0 Hz); 13C NMR (100 MHz, D2O) δ 182.6, 182.3, 61.3, 41.0,
35.5, 16.0. Anal. Calcd for C6H11NO4, 0.66 H2O: C, 41.64, H,
7.18,N, 8.09. Found: C, 41.42, H, 6.72,N, 7.86.
(2S,4R)- and (2S,4S)-4-Methylglutamic Acid 2b. To a mixture
of 1b (300 mg, 1.74 mmol) and CSA (266 mg, 1.74 mmol) in water
(90 mL) at pH 7.6 was added AAT (40 IU). The reaction mixture
was stirred at room temperature and monitored by titration of
pyruvate as previously described for the synthesis of 1m. When a
conversion rate near 50% was reached in 2 h, the reaction mixture
was passed through a column of Dowex 50X8 resin (H+ form, 10
mL). The column was washed with water until complete elution
of excess CSA. The fractions containing unreacted (S)-1b and CSA
were pooled and concentrated under reduced pressure after adjusting
the pH to 7.6 with 1 M NaOH. The column was then eluted with
1 M NH4OH, and (2S,4R)-2b was isolated and further purified on
Dowex 2 resin as previously described for 2n. (2S,4R)-2b was
isolated as a white solid (132 mg, 47%) with a diastereomeric excess
of 88% as judged by 1H NMR. Physical and spectroscopic properties
were consistent with the previously reported data.15 Anal. Calcd
for C6H11NO4, 0.33 H2O: C, 43.13, H, 7.03, N, 8.38. Found: C,
43.24, H, 6.89, N, 8.75.
To the solution of residual (S)-1b and CSA in water (30 mL)
were added cysteine sulfinic acid (CSA) (75 mg, 0.5 mmol),
dilithium 4-benzyl-2-oxoglutarate 1j (25 mg, 0.1 mmol), and a 10
mM solution of pyridoxal phosphate (0.45 mL, 4.5 µmol). The pH
of the solution was adjusted to 7.6 with 1 M NaOH, and the volume
was adjusted to 45 mL before BCAT (5 IU) and AAT (10 IU)
were added. The reaction was stirred slowly at room temperature
for 15 h. The reaction mixture was passed through a column of
Acknowledgment. We are grateful to the CNRS for financial
support and to Professor Kagamiyama’s group (Osaka medical
college) for providing us with E. coli strains which overexpress
AAT and BCAT.
Supporting Information Available: General experimental
information, enzymatic assays procedures, characterization of
compounds 1n-p, 2o,p, 6o,p, 7o,p, 8o,p, and 9, copies of 1H and
13C NMR spectra for compounds 1m-p, 2b-p, 5, 6m-p, 7n-p,
8n-p, and 9. This material is free of charge via the Internet at
http://pubs.acs.org.
(27) Hartzoulakis, B.; Gani, D. J. Chem. Soc., Perkin Trans. 1 1994,
2525-2531.
(28) Soloshonok, V. A.; Cai, C.; Hruby, V. J.; Van, Meervelt, L.;
Mischenko, N. Tetrahedron 1999, 55, 12031-12044.
JO070805Q
7566 J. Org. Chem., Vol. 72, No. 20, 2007