Chemo-enzymatic synthesis of the four stereoisomers of 4-hydroxy-4- methylglutamic acid

Article abstract of DOI:10.1016/S0957-4166(98)00372-3

A chemo-enzymatic synthesis of the four stereoisomers of 4-hydroxy-4- methylglutamic acid is described. As with other glutamate analogues these compounds can serve as probes for the investigation of glutamate receptors.

Full text of DOI:10.1016/S0957-4166(98)00372-3

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 9 (1998) 3855–3861
Chemo-enzymatic synthesis of the four stereoisomers of
4-hydroxy-4-methylglutamic acid
Virgil Helaine and Jean Bolte ∗
Département de Chimie, Laboratoire de Synthèse, Electrosynthèse et Etude de Systèmes d’Intérêt Biologique, associé au
CNRS (UMR 6504), Université Blaise Pascal, 63177 Aubière Cedex, France
Received 7 September 1998; accepted 14 September 1998
Abstract
A chemo-enzymatic synthesis of the four stereoisomers of 4-hydroxy-4-methylglutamic acid is described. As
with other glutamate analogues these compounds can serve as probes for the investigation of glutamate receptors.
© 1998 Elsevier Science Ltd. All rights reserved.
1. Introduction
Glutamic acid plays a crucial role in the central nervous system. As one of the excitatory amino
acids, it is a neurotransmitter for the majority of synapses. Excitatory amino acid receptors appear not
only to mediate normal synaptic transmission, but also to participate in the modification of synaptic
connections associated with brain development and learning or memory process.1,2 Moreover, excessive
release of glutamic acid can result in neurodegenerative disorder such as Huntington’s, Alzheimer’s
and Parkinson’s diseases.3 Two main types of glutamic receptors, called ionotropic and metabotropic,
each with various subgroups, are known.4,5 Glutamic acid analogues appear as important tools for
topology investigation of these receptors and for possible therapeutic effects due to their specificities
for a particular receptor. For example, (2S,4S)-4-methylglutamic acid interacts specifically with the
metabotropic glutamic receptor mGluR1, as an agonist of glutamate, whereas (2S,4R)-4-methylglutamic
acid has an exceptional selectivity for the KA ionotropic receptor.6
Various substituted glutamic acids have been described recently.7 We have already published the syn-
theses of 4-methyl- and 4-hydroxy-L-glutamic acids by enzymatic transamination of the corresponding
α-ketoacid.8 4-Hydroxy-4-methylglutamic acids have not been synthetised so far, although the isomers
of the L series are natural compounds: the (2S,4R) is found in Ledenbergia roseo-aena,9 the (2S,4S) in
∗
Corresponding author. E-mail: jbolte@chimtp.univ-bpclermont.fr
0957-4166/98/$ - see front matter © 1998 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(98)00372-3

3856
V. Helaine, J. Bolte / Tetrahedron: Asymmetry 9 (1998) 3855–3861
Pandanus veitchii10 and both isomers in Phyllitis scolopendrium11 and Caylusea abyssinica.12 In this
paper, we present the synthesis of the four stereoisomers of 4-hydroxy-4-methylglutamic acid 1 and 2.
2. Results and discussion
The synthesis was carried out according to Scheme 1 and is based on the resolution of the dihydroiso-
xazole 3 by Aspergillus oryzae protease as described by Griengl et al.13 Racemic 3 was easily obtained
by cycloaddition of methacrylate and ethoxycarbonylnitrile oxide (generated in situ from ethyl chloro-
oximidoacetate), and then submitted to enzymatic hydrolysis. After 55% conversion, the remaining ester
(R)-3 was isolated in 44% yield and more than 97% ee and the hydrolysed compound, isolated in 55%
yield and 77% enantiomeric excess (ee), was esterified and the ester re-submitted to the enzyme action.
When the hydrolysis reached 80% of conversion, the acid was isolated and esterified to afford (S)-3 in an
overall yield of 40% and an enantiomeric excess of 96%.
Scheme 1. Synthesis of 4-hydroxy-4-methyl glutamic acids: (i) lit.;13 (ii) Aspergillus oryzae protease; (iii) a — SOCl2, EtOH,
∆; b — Aspergillus oryzae protease; c — SOCl2, EtOH, ∆; (iv) H2/Raney Ni, (Bz)2O; (v) HCl 6 N reflux, then LiOH
(S)- and (R)-3 can lead to the desired amino acids as a mixture of separable diastereomers respectively
(2S,4S)/(2R,4S) and (2R,4R)/(2S,4R) by reduction with hydrogen in the presence of Raney nickel.
However, during this experiment, lactamisation occurred. This side reaction was avoided by trapping
the amino acid as an N-acylated form, the reduction being carried out in the presence of a carboxylic
anhydride. We first used acetic anhydride in order to obtain N-acetyl derivatives separable by the action

V. Helaine, J. Bolte / Tetrahedron: Asymmetry 9 (1998) 3855–3861
3857
of acylase. However, acylase was inoperative on these compounds and chromatographic separation was
also inefficient. Finally, the reduction was achieved in the presence of benzoic anhydride. The N-benzoyl
compounds 5 and 6 were formed and have been separated by chromatography on silica gel.
The relative configuration of carbons 2 and 3 was established by careful NMR study. First, since such
an attribution is easier on compounds of restricted conformations, the stereoisomer 5 was transformed
into the lactone 7 (Scheme 2) by heating to reflux in benzene in the presence of para-toluenesulfonic
acid and removing the ethanol formed by azeotropic distillation. All nuclei are well differentiated in
1H NMR spectroscopy, with chemical shifts of δ 1.73, 2.29, 3.10, 4.9 and 7.09 ppm respectively for
C4CH3, C3H, C3H0, C2H, and NH. The stereochemical assignment was based on NOE experiments.
Presaturation of the CH3 (d 1.73 ppm) produced 6% enhancement of the signal at δ=2.29 ppm which
was attributed to the proton in the cis position C3H and 0% for the signal at 3.10 ppm attributed to the
proton in the trans position, C3H0. Presaturation of C3H0 produced 25% enhancement on the vicinal C3H,
12% on C2H and 0% on NH which is consistent with a trans configuration of a planar cycle as shown
in Scheme 2. However, the close values observed (9 and 12 Hz) for the vicinal coupling constants for
HC2C3H and HC2C3H0 seem to indicate a deformation of the cycle which might decrease the difference
in the distances between CH3 and the protons on C3. In order to examine if an ambiguity could remain
at this level, we looked for the more stable conformation by molecular modelling. The dihedral angles
H0C3C2H and HC3C2H appeared to be 21° and 99° respectively, values consistent with the observed
vicinal coupling constants. In this conformation, H is closer to the CH3 group than H0. So, we can
attribute the configuration (2S,4R) and (2R,4S) to the compounds 5 obtained in good yield by reduction of
the enantiomerically pure dihydroisoxazoles 2. Acidic hydrolysis of 5 or 6 leads to 2 or 1 contaminated
by the corresponding lactone as indicated by the characteristic sharp IR band at 1795 cm−1. Thus the
mixture was submitted to hydrolysis in basic conditions (LiOH, pH 12). Then, after pH adjustment to 7
by HCl addition, the solution was poured on a Dowex anion exchange resin OH− form, and the amino
acids 1 and 2 were eluted with acetic acid solution and recovered by evaporation.
Scheme 2. Lactonisation of 5 and NOE experiments on 7
1H and 13C NMR spectra of the four stereoisomers of 4-hydroxy-4-methylglutamic acids are presented
in Table 1 and compared with literature data. Obviously there is a misfit between our results and
literature for the relative configuration of the two stereogenic centers. Our attribution is based on the
specificity of the protease of Aspergillus oryzae in the hydrolysis of 3, which was clearly established
by Griengl and coworkers by chemical correlation13a and by our NMR study of the lactone 7. The
1
previous attribution, by Alderweireldt et al.10 was based on a H NMR study of samples from natural
sources. The protons in C2 and C3 of the two diastereomers have different coupling constants due to the
existence of stabilised conformations. This can allow the attribution of the relative configuration of the

3858
V. Helaine, J. Bolte / Tetrahedron: Asymmetry 9 (1998) 3855–3861
Table 1
1
13C and H NMR chemical shift, δ (ppm) for (2S,4S) and (2S,4R) 4-hydroxy-4-methylglutamic acid
in 0.5 N NaOD in D2O
stereogenic centres provided that these conformations are known. In Alderweireldt’s study, however,
stabilised conformations were proposed without any evidence, on the basis of very simplistic and
empirical arguments. A possible mistake in the attribution of configurations was already suspected by
Kaas and Sorensen,14 since the assignment of the configuration of 4-hydroxy-4-isobutylglutamic acid
1
based on the comparison of its H NMR spectra with those published for 4-hydroxy-4-methylglutamic
acids10 had to be changed after an X-ray structure determination.
3. Experimental section
3.1. General data
Melting points were determined on a Reichert hot-stage apparatus and are reported uncorrected. IR
spectra were recorded on a Perkin–Elmer 801 spectrophotometer. Chromatography was carried out on
220–400 mesh silica gel using the ‘flash’ methodology. Thin-layer chromatography was obtained on
Merck precoated silica gel 60 F254 plates and spots were visualised by UV at 254 nm. 1H (400 MHz) and
13C (100 MHz) experiments were performed on a AC 400 Brucker spectrometer. HPLC analyses were
recorded on a Waters 660 liquid chromatograph at a flow rate of 0.3 ml/min, equipped with a Waters
440 UV detector (λ=220 nm) and a chiral column (Chiralcel OB, 4.6 mm×250 mm). The molecular
modelling study was performed using the SYBYL 6.3 software package15 on a Silicon Graphics R 8000
workstation. The structure was built within SYBYL and maximised by MAXIMIN 2 with the Tripos
force field, in vacuo conditions, to provide reasonable standard geometries. The molecule was deemed
to be minimised, by a conjugate gradient method, when a minimum of less than 0.021 kJ/mol for one
iteration was reached. Semi-empirical AM1 calculations were performed using the MOPAC package16
implemented in SYBYL and involved the singlet state. The conformational space of 7 was explored using

V. Helaine, J. Bolte / Tetrahedron: Asymmetry 9 (1998) 3855–3861
3859
the SYBYL search facility. Torsion angles were defined and a grid search was performed allowing chosen
bonds to rotate with a 360° or 180° revolution by 15° increments. The lowest energy conformers thus
obtained were submitted to AM1 calculations to optimise their geometry.
3.2. Diethyl 5-(RS)-methyl-4,5-dihydroisoxazole-3,5-dicarboxylate (ꢀ)-3
K. Faber and H. Griengl procedure13 provided a 99% yield of 3.
3.3. Diethyl 5-(R)- and 5-(S)-methyl-4,5-dihydroisoxazole-3,5-dicarboxylate 3
By a procedure described by the above mentioned authors, the resolution of 3 with protease from
Aspergillus oryzae (Sigma type XXIII, 3.6 U/mg), was achieved. (R)-3 was obtained in 44% yield (extent
of conversion c=55%, ee>97%, TR=105.96 min, E=35) and (S)-4 in 55% yield (ee=77%).
This last compound was esterified with SOCl2 in refluxing EtOH in quantitative yield, and after a
usual workup was hydrolysed again with the enzyme (c=85%) in order to increase the enantiomeric
purity. Thus, another esterification in the same conditions afforded a 43% overall yield of (S)-3 (ee=96%,
TR=119.62 min).
3.4. General procedure for the synthesis of the diethyl N-benzoyl-4-hydroxy-4-methylglutamates 5 and
6
To a solution of 4.4 mmol of (R)-3 (or (S)-3) in 80 ml of methanol were added 13.2 mmol (3 equiv.) of
benzoic anhydride, 16 ml of phosphate buffer (0.1 N, pH 7) and 0.2 g of Raney nickel. The mixture was
then vigourously stirred in a hydrogen atmosphere at room temperature until starting material was no
longer revealed on TLC with cyclohexane:ethyl acetate (5:5). After filtration of the catalyst, the solution
was evaporated and the residue was dissolved in 50 ml of ethyl acetate, washed twice with a 10% aqueous
NaHCO3 solution, dried over MgSO4 and evaporated in vacuo. The diastereomers were separated and
purified by chromatography on silica gel using diethyl ether:cyclohexane (6:4) as an eluent.
3.4.1. Diethyl (2R,4R)-N-benzoyl-4-hydroxy-4-methylglutamate 6
Oil (419 mg, 28%); 1H NMR (CDCl3) δ 1.15, 1.35 (each t, each 3H, each J=7 Hz, 2CH2CH3), 1.43
(s, 3H, CH3), 2.4 (dd, 1H, J=5, 14 Hz, CH2), 2.75 (dd, 1H, J=5, 14 Hz, CH2), 3.53 (s, 1H, OH), 3.83
(m, 1H, CH2CH3), 4.07 (m, 1H, CH2CH3), 4.25 (q, 2H, J=7 Hz, CH2CH3), 4.7 (q, 1H, J=5 Hz, CH),
7.11 (d, 1H, J=5 Hz, NH), 7.45 (t, 2H, J=8 Hz, Ar meta), 7.51 (t, 1H, J=8 Hz, Ar para), 7.80 (d, 2H,
J=8 Hz, Ar ortho), 13C NMR (CDCl3) δ 14.1 (2CH2CH3), 29.7 (4-CH3), 39.7 (C3), 50.4 (C2), 61.5, 62.1
(2CH2CH3), 73.3 (C4), 127.0, 128.4, 131.7, 133.3 (6CH Ar), 166.8 (COPh), 172.4, 177.2 (2CO2Et).
3.4.2. Diethyl (2S,4R)-N-benzoyl-4-hydroxy-4-methylglutamate 5
Pasty solid (614 mg, 41%), 1H NMR (CDCl3) δ 1.15, 1.32 (each t, each 3H, J=7 Hz, 2CH2CH3), 1.47
(s, 3H, CH3), 2.3 (dd, 1H, J=5, 14 Hz, CH2), 2.45 (dd, 1H J=12, 14 Hz, CH2), 3.99 (m, 2H, CH2CH3),
4.25 (m, 3H, CH2CH3 and OH), 4.97 (m, 1H, CH), 7.05 (d, 1H, J=8 Hz, NH), 7.43 (t, 2H, J=8 Hz, Ar
meta), 7.53 (t, 1H, J=8 Hz, Ar para), 7.80 (d, 2H, J=8 Hz, Ar ortho), 13C NMR (CDCl3) δ 14.0, 14.1
(2CH2CH3), 27.3 (4-CH3), 41.3 (C3), 49.8 (C2), 61.9, 62.2 (2CH2CH3), 73.2 (C4), 127.2, 128.7, 132.1,
133.4 (6CH Ar), 167.6 (COPh), 172.2, 176.2 (2CO2Et).

3860
V. Helaine, J. Bolte / Tetrahedron: Asymmetry 9 (1998) 3855–3861
3.4.3. Diethyl (2S,4S)-N-benzoyl-4-hydroxy-4-methylglutamate 6
Oil (391 mg, 26%); spectroscopic data were identical to those of the diethyl (2R,4R)-N-benzoyl-4-
hydroxy-4-methylglutamate 6.
3.4.4. Diethyl (2R,4S)-N-benzoyl-4-hydroxy-4-methylglutamate 5
Pasty solid (587 mg, 40%); spectroscopic data were identical to those of the diethyl (2S,4R)-N-
benzoyl-4-hydroxy-4-methylglutamate 5.
3.5. General procedure for the synthesis of the 4-hydroxy-4-methylglutamic acids 1 and 2
The so obtained compounds (1.2–1.8 mmol) were dissolved in 6 N HCl (10 ml) and refluxed overnight.
The solution was cooled to room temperature and evaporated under reduced pressure. The residue was
then suspended in water (10 ml) and treated with LiOH until the solution maintained a permanently
alkaline pH of 12. After neutralisation with HCl 1 N, the solution was passed through a strongly basic
ion exchange column (Amberlite IRA-900, OH− form). The amino acid was eluted with acetic acid 1 M.
The solution thus obtained was evaporated under reduced pressure and traces of acetic acid were removed
by azeotropic distillation with toluene.
3.5.1. (2R,4R)-4-Hydroxy-4-methylglutamic acid 1
Solid (215 mg, 98%), IR (KBr) ν 3175, 1740, 1715, 1630, 1595, 1550, 1490, 1410, 1340, 1270, 1185
1
cm−1, H NMR (NaOD) δ 1.43 (s, 3H, CH3), 1.74 (dd, 1H J=10, 14 Hz, CH2), 2.31 (dd, 1H, J=3, 14
Hz, CH2), 3.39 (dd, 1H, J=3, 10 Hz, CH), 13C NMR (NaOD) δ 27.5 (4-CH3), 44.4 (C3), 55.6 (C2), 78.1
(C4), 183.7 (C1), 184.0 (C5).
3.5.2. (2S,4R)-4-Hydroxy-4-methylglutamic acid 2
Solid (312 mg, 97%), IR (KBr) ν 3140, 1695, 1630, 1590, 1525, 1470, 1430, 1410, 1330, 1240, 1140
cm−1, 1H NMR (NaOD) δ 1.42 (s, 3H, CH3), 1.94 (dd, 1H, J=8, 14 Hz, CH2), 2.10 (dd, 1H, J=5, 14 Hz,
CH2), 3.47 (dd, 1H, J=5, 8 Hz, CH), 13C NMR (NaOD) δ 27.0 (4-CH3), 45.8 (C3), 54.0 (C2), 76.0 (C4),
184.1 (C1), 184.3 (C5).
3.5.3. (2S,4S)-4-Hydroxy-4-methylglutamic acid 1
Solid (197 mg, 96%), spectroscopic data were identical to those of the (2R,4R)-4-hydroxy-4-
methylglutamic acid 1.
3.5.4. (2R,4S)-4-Hydroxy-4-methylglutamic acid 2
Solid (302 mg, 98%), spectroscopic data were identical to those of the (2S,4R)-4-hydroxy-4-
methylglutamic acid 2.
3.6. Assignment of the absolute configuration on the 2-carbon of the diethyl (2S,4R)-N-benzoyl-4-
hydroxy-4-methylglutamate 5
They were achieved by considering the NOE effects on the corresponding lactone.
The diethyl (2S,4R)-N-benzoyl-4-hydroxy-4-methylglutamate 5 (0.3 g, 0.9 mmol) was dissolved in 10
ml of benzene. 17 mg (0.09 mmol, 0.1 equiv.) of para-toluenesulfonic acid was added and the mixture
was refluxed overnight. The ethanol formed was then removed using a Dean–Stark system. After cooling
to room temperature, the solution was transferred to a separatory funnel and washed with saturated

V. Helaine, J. Bolte / Tetrahedron: Asymmetry 9 (1998) 3855–3861
3861
sodium bicarbonate solution. The organic layer was dried on MgSO4 and evaporated in vacuo. Column
chromatography on silica gel using cyclohexane:ethyl acetate (6:4) as an eluent gave 0.23 g (89%) of
(2S,4R)-2-benzoylamino-4-ethoxycarbonyl-4-methylbutyrolactone 7 as a white solid (mp=89–90°C). IR
(CHCl3) ν 3432, 3002, 1792, 1742, 1668, 1484, 1307, 1140, 1095, 1020 cm−1, 1H NMR (CDCl3) δ 1.34
(t, 3H, J=7 Hz, CH2CH3), 1.62 (s, 1H, OH), 1.76 (s, 3H, CH3), 2.26 (t, 1H, J=13 Hz, CH2), 3.17 (dd,
1H, J=9, 13 Hz, CH2), 4.30 (m, 2H, CH2CH3), 4.88 (ddd, 1H, J=6, 9, 15 Hz, CH), 6.85 (d, 1H, J=6 Hz,
NH), 7.43 (t, 2H, J=8 Hz, Ar meta), 7.52 (t, 1H, J=8 Hz, Ar para), 7.78 (d, 2H, J=8 Hz, Ar ortho), 13C
NMR (CDCl3) δ 14.1 (CH2CH3), 23.7 (4-CH3), 39.9 (C3), 50.1 (C2), 62.6 (CH2CH3), 82.1 (C4), 127.2,
128.6, 132.1 (6CH Ar), 167.5 (COPh), 171.1, 175.0 (2CO2Et).
Acknowledgements
We are grateful to Dr. D. Gardette for the molecular modelling study.
References
1. Monaghan, D. T.; Bridges, R. J.; Cotman, C. W. Ann. Rev. Pharmacol. Toxicol. 1989, 29, 365–402.
2. Siegelbaum, S. A.; Kandel, E. R. Curr. Opin. Neurobiol. 1991, 1, 113–120.
3. (a) Olney, J. W. Ann. Rev. Pharmacol. Toxicol. 1990, 30, 47–71. (b) Meldrum, B. Brain Res. Rev. 1993, 18, 293–314.
4. (a) Nakanishi, S. Science 1992, 258, 597–603. (b) Cunningham, M. O.; Ferkany, J. W.; Enna. S. J. Life Sci. 1993, 54,
135–148.
5. Acher, F. C.; Tellier, F. J.; Azerad, R.; Brabet, I. N.; Fagny, L.; Pin, J.-P. R. J. Med. Chem. 1997, 40, 3119–3129.
6. (a) Gu, Z.-Q.; Hesson, D. P.; Pelletier, J. C.; Maccaecchini, M.-L. J. Med. Chem. 1995, 38, 2518–2519. (b) Todeshi, N.;
Gharbi-Benarous, J.; Acher, F.; Larue, V.; Pin, J.-P.; Bockaert, J.; Azerad, R.; Girault, J.-P. Bioorg. Med. Chem. 1997, 5,
335.
7. (a) Moody, C. M.; Young, D. W. Tetrahedron Lett. 1994, 35, 7277–7280. (b) Moody, C. M.; Young, D. W. Tetrahedron
Lett. 1993, 34, 4667–4670. (c) Coudert, E.; Acher, F.; Azerad, R. Tetrahedron: Asymmetry 1994, 5, 731. (d) Gu, Z.-Q.;
Hesson, D. P. Tetrahedron: Asymmetry 1995, 6, 2101.
8. (a) Passerat, N.; Bolte, J. Tetrahedron Lett. 1987, 28, 1277–1280. (b) Echalier, F.; Constant, O.; Bolte, J. J. Org. Chem.
1993, 58, 2747–2750.
9. Jadot, J.; Casimir, J.; Loffet, A. Biochim. Biophys. Acta 1967, 136, 79–88.
10. Alderweireldt, F.; Jadot, J.; Casimir, J.; Loffet, A. Biochim. Biophys. Acta 1967, 136, 89–94.
11. Meier, L. K.; Sorensen, J. Phytochemistry 1979, 18, 1173–1175.
12. Olsen, O.; Sorensen, H. Phytochemistry 1980, 19, 1717–1721.
13. Yang, S.; Hayden, W.; Faber, K.; Griengl, H. Synthesis 1992, 365. (b) Yang, S.; Hayden, W.; Faber, K.; Griengl, H.
Monatsschr. Chemie 1994, 125, 469–477.
14. Kaas, K.; Sorensen, H. Acta Chem. Scand. A 1977, 31, 364–368.
15. Tripos Associates, Inc., 1969 S. Harley Rd., Suite 303, St Louis, MO 63144, USA.
16. Merz, K. M.; Besler, B. H. MOPAC version 5.0 ESC, QCPE no. 589, 1990, with AM1 force field parameters.