Methylation of l-trans-2,4-pyrrolidine dicarboxylate converts the glutamate transport inhibitor from a substrate to a non-substrate inhibitor

Article abstract of DOI:10.1016/S0968-0896(02)00250-X

The 4-methyl analogue of the potent inhibitor of CNS L-glutamate neurotransmitter transporters, L-trans-2,4-PDC, was synthesized via a 1,3-dipolar cycloaddition reaction sequence. The bioassays performed not only exhibit increased potency of the methylated derivative over L-trans-2,4-PDC, but also exhibit non-substrate properties at the rat forebrain synaptosomal glutamate transporter while the parent L-trans-2,4-PDC exhibits substrate properties. These results support two hypotheses developed for distinguishing the physiological properties of transport inhibitors based on molecular modeling studies, and are reported here.

Full text of DOI:10.1016/S0968-0896(02)00250-X

Bioorganic & Medicinal Chemistry 10 (2002) 3509–3515
Methylation of L-trans-2,4-Pyrrolidine Dicarboxylate Converts the
Glutamate Transport Inhibitor from a Substrate to a
Non-substrate Inhibitor
C. Sean Esslinger,a,* Jody Titus,b Hans P. Koch,a,y Richard J. Bridgesa
and A. Richard Chamberlinb
aCOBRE Center For Structural and Functional Neuroscience, Department of Pharmaceutical Sciences, University of Montana,
Missoula, MT59812, USA
bDepartment of Chemistry, University of California, Irvine, CA 92717, USA
Received 2 May 2002; accepted 13 June 2002
Abstract—The 4-methyl analogue of the potent inhibitor of CNS l-glutamate neurotransmitter transporters, l-trans-2,4-PDC, was
synthesized via a 1,3-dipolar cycloaddition reaction sequence. The bioassays performed not only exhibit increased potency of the
methylated derivative over l-trans-2,4-PDC, but also exhibit non-substrate properties at the rat forebrain synaptosomal glutamate
transporter while the parent l-trans-2,4-PDC exhibits substrate properties. These results support two hypotheses developed for
distinguishing the physiological properties of transport inhibitors based on molecular modeling studies, and are reported here.
# 2002 Elsevier Science Ltd. All rights reserved.
Introduction
EAA receptors, these excitatory amino acid transporters
(EAATs) are postulated to participate not only in the
maintenance of glutamate concentrations below those
which are excitotoxic, but also in signal termination and
neurotransmitter recycling.5À8 Of the five sodium-
dependent high-affinity EAATs identified thus far, the
EAAT2-subtype (rat homologue: GLT1) present on
glial cells throughout the forebrain is thought to be the
dominant transporter responsible for these functions.9
Recently, a splice variant of EAAT2/GLT1 has also
been reported to be present on neurons.10,11
l-Glutamate is recognized as the major excitatory
neurotransmitter in the mammalian central nervous
system (CNS). Through the activation of a variety of
ionotropic and metabotropic excitatory amino acid
(EAA) receptors, l-glutamate participates not only in
fast excitatory transmission, but also in the higher order
signal processing required in development, learning,
memory and synaptic plasticity (for review see: refs 1
and 2). Of equal interest, over activation of EAA
receptors by excessive levels of l-glutamate can also
induce excitotoxic-mediated neuronal pathology.3,4
High-affinity transporter proteins that translocate l-
glutamate across the cellular membranes of neurons and
glia are thought to play a significant role in regulating
extracellular levels of l-glutamate and maintaining the
balance between its physiological and pathological
actions. As a consequence of the ability to efficiently
clear l-glutamate from the synaptic spaces surrounding
Previous studies identified l-trans-2,4-pyrrolidine dicar-
boxylate (l-trans-2,4-PDC) as a potent competitive
inhibitor of EAAT2 (as well as all of the other EAATs)
that can also act as a substrate of this transporter (Fig.
1).12,13 While more conformationally constrained than
l-glutamate, l-trans-2,4-PDC can, owing to conforma-
tional inter-conversions of the ring, mimic more than
one conformation of l-glutamate. Based upon modeling
comparisons of glutamate with l-trans-2,4-PDC (Fig. 2
A and B, R=H), we initially postulated that A is the
bioactive conformation for binding to the substrate site
on the transporter protein. This hypothesis was pre-
dicated simply upon binding data garnered from com-
petitive inhibition studies; the issue of whether a given
*Corresponding author. Tel.: +1-406-243-4713; fax: +1-406-243-
4972; e-mail: seaness@selway.umt.edu
yCurrent address: Vollum Institute, Oregon Health Sciences Uni-
versity, Portland, OR 97201, USA.
0968-0896/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(02)00250-X

3510
C. S. Esslinger et al. / Bioorg. Med. Chem. 10 (2002) 3509–3515
hypothesis that B0 and
B define the substrate
pharmacophore (i.e., good binding, efficient transport),
while A0 and A represent the equivalent of an non-sub-
strate inhibitor pharmacophore (good binding, poor
transport). A functional analogy may be drawn between
substrates and non-substrate inhibitors of neuro-
transmitter transporters and the respective actions of
agonists and antagonists at neurotransmitter receptors.
This conclusion is consistent with known properties of
l-trans-2,4-PDC (Fig. 2 A and B: R=H), which is
readily transported because it can achieve the ‘trans-
portable’ conformation represented by B/B0.13,17 This
clearly suggests that substituents on l-trans-2,4-PDC
that would favor the folded conformer A should
decrease transport rates without necessarily decreasing
binding affinity. However, it is possible that such struc-
tural modifications, including those present on l-anti,-
endo-3,4-MPDC or 2,4-MPDC, may independently
influence activity as a consequence of the introduction
of steric bulk. Thus, similar structure/activity studies
carried out with a larger library of compounds, the
majority of which were less conformationally con-
strained than either l-anti,endo-3,4-MPDC or 2,4-
MPDC, also suggested that steric arguments can be
made to differentiate the pharmacophores of substrates
and non-substrate inhibitors.13
Figure 1. l-Glutamate and conformationally constrained competitive
inhibitors of EAAT2.
ligand was transported (i.e., translocated across the
membrane) or not was ignored. To further test this
model and ascertain if A might indeed be the active
form, conformationally locked mimics of the two
‘extreme’ conformations of l-trans-2,4-PDC (there are
many closely related conformers as well) were prepared:
l-anti,endo-3,4 - methanopyrrolidine - 3,4 - dicarboxylate
(l-anti,endo-3,4-MPDC, A0) representing the ‘folded’
(pseudo-axial) conformation (A) and 2,4-methano-
pyrrolidine-2,4-dicarboxylate (2,4-MPDC, B0) which
closely aligns with the extended form B (psedo-
equatorial).14À16 Our criterion for good alignment was
an RMS deviation of less than 0.1 A. for atoms of the
carboxylate carbons and ammonium nitrogen atom,
and the two derivatives shown gave the best fits of all
compounds inspected.
Interestingly, both analogues were able to effectively
bind to the transporter, as indicated by the potency with
which they competitively blocked the synaptosomal
uptake of d-[3H]aspartate (e.g., Ki ꢀ5 and 7 mM,
respectively, Table 1). These values are within experi-
mental error for the activity of the endogenous sub-
strate, l-glutamate (Ki ꢀ5 mM). It was suspected,
however, that these results reflected more than just a
binding domain with a broader than expected specificity
range. Indeed, more recent experiments demonstrated
that there is a significant functional difference between
A0 and B0 that inhibition-based assays cannot distin-
guish: B0 is very effectively translocated across the
membrane (i.e.,0 is a substrate comparable to glutamate
itself), while A is not.13 These findings support the
As a further test of the these relationships, we have
synthesized 2S,4R-4-methyl-pyrrolidine-2,4-dicarboxy-
late (4-Me-l-trans-2,4-PDC) 6 as a conformationally
biased analogue of l-trans-2,4-PDC and used it as a
probe with which to further delineate the chemical con-
straints that differentially govern the ability of ana-
logues to bind to and then be (or not be) translocated by
the EAAT2/GLT1 subtype of glutamate transporter.
Specifically, 4-Me-l-trans-2,4-PDC was found to com-
petitively inhibit the uptake of d-[3H]aspartate in rat
forebrain synaptosomes, but did not exchange with
radiolabeled substrate in complementary efflux assays
(Fig. 3, Table 1). Significantly, these results suggest that
the addition of the 4-methyl group essentially converted
the parent compound l-trans-2,4-PDC from an
Figure 2. Pseudo-axial, ‘folded’ (A) and pseudo-equitorial, ‘extended’ (B) conformations of l-trans-2,4-PDC.
Table 1. Glutamate and related pyrrolidine dicarboxylates as transporter inhibitors and substrates
Compd
Ki (mM)
Concd (mM)
Exchange rate (pmol/mg pro/2 min)
Substrate activity (% of Glu)
l-GIu
4.9Æ1.0
3.1Æ0.5
4.5Æ0.9
>50
1.5Æ0.5
6.8Æ3.0
4.9Æ2.0
50
30
45
—
15
75
50
344Æ16
0Æ8
100
0
2(S),4(R)-4-Me-2,4-PDC
(2S,4R)-4-Me-Glu
(2S,4S)-4-Me-Glu
l-trans-2,4-PDCa
2,4-MPDCa
194Æ5
—
56
—
60
81
35
208Æ28
280Æ30
120Æ18
l-anti-endo-3,4-MPDC
aAs reported in Koch et al.13

C. S. Esslinger et al. / Bioorg. Med. Chem. 10 (2002) 3509–3515
3511
stereomers and the need to cleave the C–S and C–N
bonds, this strategy promised to provide a very concise
means of producing the target compound (Scheme 1).
The key cycloaddition was conducted by reaction of four
equivalents of the acrylate 128 with the thiazolium salt 225
to give the diastereomeric tricyclic adducts 3. Since the
separation of the minor isomers was problematic (most
likely due to the labile N,O-acetal), the crude mixture was
treated with AgOTf29,30 to generate the corresponding
imine intermediate 4 and two minor diastereomers in a
ratio of 83:11:6. The desired major product 4 is con-
sistent with regioselective, auxiliary-directed ‘exo’ attack
of the chiral acrylate on the conformation of the dipole
shown, all in accord with literature precedent. These
imines, however, were also inseparable on a preparative
scale. They were therefore reduced as a mixture to the
corresponding pyrrolidines and then protected to give a
separable mixture of trifluoroacetamides, from which 5
could be obtained in pure form. Finally, the tri-
fluoroacetamide 5 was deprotected and purified to give
the target compound 6, which was identical to material
prepared by the more circuitous route from l-hydroxy-
proline via l-trans-2,4-PDC (data not shown). This
regio-, stereo-, and enantioselective route provides the
highly substituted (2S, 4R)-4-methyl-trans-2,4-PDC in
reasonable yield in only five steps.
Figure 3. Representative Lineweaver–Burk plot demonstrating com-
petitive inhibition of d-[3H]aspartate uptake into rat forebrain synap-
tosomes by 4-Me-l-trans-2,4-PDC. The inset shows a replot of Km,app
versus [inhibitor]. The plots shown were generated using the kcat
kinetic program (BioMetallics Inc., Princeton, NJ, USA) with weight-
ing based on constant relative error, and yielded values of: Km=2.8
mM (d aspartate), Ki =3.6 mM, and a Vmax of 2.6 nmol/min/mg pro-
tein. Three such analyses yielded an average Ki=3.1Æ0.5 mM.
alternative substrate into a non-substrate inhibitor. The
development of such analogues, as well as the elucida-
tion of the pharmacological relationships controlling
substrate translocation, are key to understanding the
structural and functional properties of the glutamate
transporter proteins.
Chemistry
We first prepared 2(S),4(R)-4-Methyl-2,4-PDC (6) from
a protected form of 2,4-PDC (which is turn prepared
from l-hydroxyproline12) via methylation of the selec-
tively-formed g-enolate.18 That route was rather
lengthy, however, and we therefore investigated an
alternative strategy employing a chiral auxiliary-direc-
ted 1,3-dipolar cycloaddition (Scheme 1).19,20 In princi-
ple, an acrylate and a simple azomethine ylide might be
employed,21À23 but these reactions proved to be gen-
erally unacceptable, at least with the requisite mono-
activated dipolarophiles. The use of a synthetic equiva-
lent of the azomethine ylide was therefore required. The
thiazolium ylide shown is a known synthon for the
simple azomethine dipole and, fortuitously, undergoes
cycloadditions with higher regioselectivity than the parent
ylide.24,25 In order to control the enantioselectivity of the
reaction, an Evans’ oxazolidinone was incorporated into
the acrylate dipolarophile, as in 1.26 Despite the additional
complexities introduced by utilizing this thiazolium ylide,
including the production of ‘temporary’ N,O-acetal dia-
Results and Discussion
The inhibitory activity of 4-Me-l-trans-2,4-PDC was
assessed by quantifying its ability to block the uptake of
d-[3H]aspartate in rat forebrain synaptosomes, a pre-
paration that exhibits a pharmacological profile most
closely aligned with EAAT2.13 As illustrated in the
Lineweaver–Burk plot depicted in Figure 3, the analo-
gue blocked uptake in a manner consistent with that of
a competitive inhibitor and exhibited a Ki value similar
to that of l-glutamate (e.g., 3.1 mM vs 4.9 mM). To
determine whether or not 4-Me-l-trans-2,4-PDC could
also act as a substrate, the analogue was tested for its
ability to exchange with d-[3H]aspartate that had been
pre-equilibrated into the synaptosomes.13,31 Thus, synap-
tosomes were first incubated with d-[3H]aspartate, reiso-
lated, resuspended (d-[3H]aspartate content=1152Æ65
pmol/mg protein), and then diluted 30-fold into assay
buffer (37 ꢁC) containing potential substrates. To insure
Scheme 1.

3512
C. S. Esslinger et al. / Bioorg. Med. Chem. 10 (2002) 3509–3515
comparable levels of binding, the analogues were inclu-
ded in the efflux assay at a concentration approximately
10-fold greater than the Ki values with which they
inhibited uptake. The extent of the efflux was measured
over a two min time course and corrected for the efflux
that occurred in the absence of inhibitor (228Æ8 pmol/
mg protein/2 min). As summarized in Table 1, known
substrates, such as l-glutamate, l-trans-2,4-PDC and
2,4-MPDC all produced substantial increases in the
synaptosomal efflux of d-[3H]aspartate that are indica-
tive of the process of heteroexchange.13,31 As previously
reported, l-anti-endo-3,4-MPDC proved to be a poor
substrate, producing only about one third of the
amount of exchange observed with l-glutamate.13 Sig-
nificantly, inclusion of the 4-Me-l-trans-2,4-PDC did
not stimulate an efflux of d-[3H]aspartate, consistent
with the action of a non-substrate inhibitor. In contrast,
its acyclic counterpart (2S,4R)-4-Me-glutamate did act
as a partial substrate in the synaptosomal preparation,
exhibiting about half the activity of l-glutamate. Con-
sidering that (2S,4R)-4-Me-Glu has previously been
shown to be a non-substrate inhibitor of EAAT2
expressed in Xenopus oocytes (Vandenberg, 1997
#1271), as has l-anti-endo-3,4-MPDC (M. Kavanaugh,
personal communication), it would suggest that the
synaptosomal preparation likely contains other trans-
porters, possibly EAAT3/EAAC1, at which the com-
pounds act as substrates. The fact that no
heteroexchange was observed in the presence of 4-Me-l-
trans-2,4-PDC indicates that it is not only a non-sub-
strate inhibitor of EAAT2/GLT1, but also of any other
EAATs present in the synaptosomes.
orientation. In a second modeling exercise (lower panels
of Fig. 4), the molecular volumes of the ‘folded’ and
‘extended’ conformations of 4-Me-l-trans-2,4-PDC are
compared with a composite substrate volume (depicted
as the yellow mesh) generated from the additive mole-
cular volumes of the known transportable (alternative
substrates) inhibitors: l-glutamate, 2,4-MPDC, l-trans-
2,4-PDC (axial, A/A0), and 1-aminocyclobutane-1,3-
dicarboxylate (ACBD).13 Previous modeling studies
revealed that several known non-substrate inhibitors of
EAAT2, dihydrokainate (DHK) and b-threo-benzyloxy-
aspartate (TBOA), exhibited substantial deviations
from this volume.13 In panels C and D of Figure 4,
regions of the volumes of 4-Me-l-trans-2,4-PDC that
protrude beyond of the composite substrate volume are
represented by the green mesh. When the more stable
pseudo-axial conformation of 4-Me-l-trans-2,4-PDC is
overlaid onto the analogous conformation of l-trans-
2,4-PDC within the confines of composite substrate
volume, steric differences between the two appear to be
primarily limited to the region occupied by the 4-methyl
group (panel C). Interestingly, this region is in the same
vicinity as the larger volumetric deviations observed
with the other two non-substrate inhibitors, DHK and
TBOA.13 When the less favored conformation of 4-Me-
l-trans-2,4-PDC is similarly compared with the pseudo-
equatorial (B/B0) conformation of l-trans-2,4-PDC,
there are a number of regions that fall outside the com-
posite substrate volume (panel D). As there are pre-
sently no other analogues available with which to
independently assess whether or not the observed volu-
metric differences produced by the methyl group are
sufficient to alter substrate activity, we conclude that
both the original conformational hypothesis and the
steric refinement lead to the same result. Biasing the
conformational preferences of the pyrrolidine ring away
from extended towards folded conformation disfavors
the requirement for efficient transport of the molecule,
as does adding steric bulk in the 4-position. Thus, either
of these factors, or both acting in concert, appears to
significantly influence the ability of these analogues to
act as substrates of EAAT2/GLT1.
These results indicate that the specific addition of a
methyl group to the 4-position of l-trans-2,4-PDC does
not diminish its ability to bind the transporter, yet it
substantially alters its ability to serve as a substrate and
be translocated by the uptake system. While this struc-
ture–activity relationship is consistent with the pro-
posed conformational hypothesis discussed above, the
addition of any new substituent introduces another
variable, that is steric bulk. Accordingly, modeling
studies (see Experimental) were conducted to assess the
potential steric role of the added methyl group in bind-
ing and translocation. Two sets of comparisons were
made using 4-Me-l-trans-2,4-PDC and the known inhi-
bitors l-trans-2,4-PDC, l-anti-endo-3,4-MPDC, and
2,4-MPDC. As depicted in Figure 4, the minimized
pseudo-axial conformation of 4-Me-l-trans-2,4-PDC
(the more stable conformation by ꢂ1.5 kcal/mol, or
approx. a 14:1 ratio, induced by the addition of the
methyl group) is overlaid with the non-transportable
inhibitor l-anti-endo-3,4-MPDC and the pseudo-axial
conformation of l-trans-2,4-PDC, giving a close fit of
overlaid atoms (Panel A, hydrogens are omitted for
clarity,), RMS deviation (both carboxyl C’s and the
ammonium N) of 0.043 A. When the same conforma-
tion of 4-Me-l-trans-2,4-PDC is similarly overlaid with
the alternative substrate inhibitor 2,4-MPDC and the
pseudo-equatorial conformation of l-trans-2,4-PDC,
the result is a poor fit, RMS deviation of 0.331 A. This
suggests the more stable conformation of 4-Me-l-trans-
2,4-PDC more closely resembles the non-transportable
Experimental
Thiazolium salt (2). The salt was prepared according to
the procedure used by Monn and Valli41 except for the
use of acetone as a recrystallization solvent which
afforded a white solid: mp 97–98 ꢁC; IR (CDCl3) umax
3330, 2986, 1748, 1589, 1469, 1396, 1375, 1347, 1225,
1
1096, 1068; H NMR (400 MHz, DMSO-d6) d 10.12 (s,
1H), 5.65 (s, 2H), 5.2 (br s, 1H), 4.25 (q, 2H, J=7.2 Hz),
3.65 (t, 2H, J=5.4 Hz), 3.06 (t, 2H, J=5.4 Hz), 2.39 (s,
3H), 1.26 (t, 3H, J=7.2 Hz); 13C NMR (100 MHz,
DMSO-d6) d 166.1, 159.1, 142.0, 135.3, 62.3, 59.6, 53.0,
29.4, 13.9, 11.1; HRMS C10H16N1S1-Br, calcd 230.0852
found 230.0851.
Imine diastereomers (4–6). (4R)-3-(2-Methyl-2-prope-
noyl)-4-(phenylmethyl)-2-oxazolidinone27 (0.9877 g,
4.03 mmol) was added to 10 mL toluene and 4 A

C. S. Esslinger et al. / Bioorg. Med. Chem. 10 (2002) 3509–3515
3513
Figure 4. Overlay of: (A) the pseudo-axial, ‘folded’ conformation of 4-Me-l-trans-2,4-PDC (green) with pseudo-axial conformation of l-trans-2,4-
PDC and l-anti,endo-3,4-MPDC yields a RMS=0.043 A; and (B) the pseudo-axial, ‘folded’ conformation of 4-Me-l-trans-2,4-PDC (green) with
pseudo-equatorial, ‘extended’ conformation of l-trans-2,4-PDC and 2,4-MPDC yields an RMS=0.331 A. Overlay of: (C) the pseudo-axial, ‘folded’
conformations of 4-Me-l-trans-2,4-PDC and l-trans-2,4-PDC within the confines of a composite substrate shell (yellow mesh) representing added
molecular volumes (see text); and (D) the pseudo-equatorial ‘extended’ conformations of 4-Me-l-trans-2,4-PDC and l-trans-2,4-PDC within the
same composite substrate shell. Regions of the volumes of 4-Me-l-trans-2,4-PDC that protrude beyond of the composite substrate volume are
represented by the green mesh. Modeling studies were carried out as described in the Experimental.
molecular sieves heated to reflux. Diisopropylethyl-
amine (0.35 mL, 2.0 mmol) was added next followed by
the slow dropwise addition of thiazolium salt 2 (0.3076
g, 0.992 mmol) in 4 mL CH2Cl2. An additional 0.6 mL
CH2Cl2 was used to quantititatively transfer the solu-
tion, and the reaction mixture was heated at relux for 5
h. After cooling to room temperature the reaction mix-
ture was diluted with Et2O and partitioned with 0.1 N
HCl. The organic layer was then washed with H2O fol-
lowed by brine and dried over MgSO4. The crude
cycloadducts were isolated in vacuo and then dissolved
in 20 mL of CH2Cl2. The reaction flask was covered
with aluminum foil, AgOTf (0.3757 g, 1.46 mmol) was
added, and the reaction mixture was stirred for 1 h.
Next, 40 mL of pH 7 phosphate buffer was added and
the reaction mixture allowed to stir vigourously for 1 h.
The solution was then washed with H2O followed by
brine. The H2O layers were combined and back extrac-
ted with CH2Cl2. The combined CH2Cl2 layers were
then dried over MgSO4 and purified by flash chromato-
graphy (2:1 hexanes/ethyl acetate gradient eluted to 1:1
hexanes/ethyl acetate) to remove excess (4R)-3-(2-
methyl-2-propenoyl)-4-(phenylmethyl)-2-oxazolidinone,
giving an otherwise pure mixture of imine diastereomers
(m, 2H), 4.78 (ddd, 1H, J=8.8, 6.0, 2.4 Hz), 4.74–4.69 (m,
1H), 4.35–4.13 (m, 4H), 3.27 (dd, 0.2H, J=13.6, 2.8 Hz),
3.21 (dd, 0.8H, J=13.4, 3.2 Hz), 2.87 (dd, 1H, J=13.6, 9.2
Hz), 2.73 (dd, 1H, J=14.4, 8.8 Hz), 2.28 (dd, 1H, J=14,
6), 1.65 (s, 3H), 1.33 (t, 3H, J=7.2 Hz); HRMS
C19H22N2O5+H calcd 359.1608, found 359.1607.
Major trifluoroacetamide (7). The imine mixture (0.0897
g, 0.250 mmol) was transferred to a 15 mL round bot-
tomed flask, concentrated in vacuo, and redissolved in 5
mL CH2Cl2. Next, TFA (0.09 mL, 1.03 mmol) and
Et3SiH (0.16 mL, 1.00 mmol) were added to the reaction
mixture. After stirring for 24 h, the reaction mixture was
diluted with 70 mL of Et2O and washed 3 Â 20 mL H2O.
The aqueous layers were combined, concentrated in vacuo
and azeotroped with PhCH3 to give a 76% combined yield
of the TFA salts, which (0.0896 g, 0.189 mmol) were
transferred to a 10 mL round bottomed flask, con-
centrated in vacuo, redissolved in 2 mL of CH2Cl2 and
cooled to 0 ꢁC. Next, TFAA (0.03 mL, 0.212 mmol) and
pyridine (0.03 mL, 0.371 mmol) were added to 2 mL
CH2Cl2 at 0 ꢁC, into which was cannulated the TFA salt
mixture.The reaction mixture stirred for 1.5 h, and then
diluted with 70 mL Et2O, washed with 2 Â 20 mL H2O
followed by brine, and dried over MgSO4. After con-
centration in vacuo the crude trifluoroacetamides were
separated by flash chromatography (1:1 Et2O/hexanes),
1
in 59% combined yield. H NMR (400 MHz, CDCl3) d
7.94 (d, 0.11H, J=2.4 Hz), 7.91 (d, 0.06H, J=2.4 Hz),
7.89 (d, 0.83H, J=2.4 Hz), 7.54–7.32 (m, 3H), 7.23–7.02

3514
C. S. Esslinger et al. / Bioorg. Med. Chem. 10 (2002) 3509–3515
providing an 83:11:6 ratio of the three diastereomers in
a combined yield of 79%. Major trifluoroacetamide (7,
66% yield): IR (neat) 3081–2852, 1786, 1681, 1452,
1386, 1352, 1324, 1214, 1105, 1048, 981, 914, 757, 700
with an additional 4 mL of ice-cold assay buffer. Filters
were transferred to scintillation fluid (National Diag-
nostics, Atlanta, GA, USA) and the retained radio-
activity was quantified by liquid scintillation counting.
Within each experiment, uptake rates were determined
in duplicate. Nonspecific uptake and/or binding was
corrected for by subtracting the amount of d-
[3H]aspartate accumulated at 4 ꢁC. Previous studies
demonstrated that under these conditions, uptake was
linear with respect to both time and protein content
(data not shown). Protein levels were determined by the
bicinchoninic acid assay (Pierce, Rockford, IL, USA).
1
cmÀ1; H NMR (400 MHz, CDCl3, major rotamer) d
7.72–7.34 (m, 3H), 7.25–7.23 (m, 2H), 4.79–4.70 (m,
2H), 4.51 (app t, 1H, J=17.2 Hz), 4.39–4.26 (m, 4H),
3.64 (d, 1H, J=12.1 Hz), 3.32–3.24 (m, 2H), 2.87 (dd,
1H, 1, J=13.4, 9.4 Hz), 2.00 (dd, 1H, J=13.7, 8.8 Hz),
1.64 (s, 3H), 1.33 (t, 3H, J=11.0 Hz); 1H NMR
(400 MHz, DMSO-d6, 438K) d 7.35–7.22 (m, 5H), 4.85–
4.79 (m, 1H),
(2R,4S)-4-Methyl-2,4-PDC (8). The major tri-
fluoroacetamide (7) (0.1391 g, 0.305 mmol) was
dissolved in 6 mL of THF and cooled to 0 ꢁC. TBAH
(0.5 mL, 1.91 mmol) was added dropwise, and the
reaction mixture was allowed to warm to room tem-
perature and stirred for 40 h. The volatiles were evapo-
rated, H2O was used to transfer the residue to a
separatory funnel, and the aqueous layer was extracted
three times with CH2Cl2. The aqueous layer was then
concentrated in vacuo and chased three times with tol-
uene. The resultant residue was dissolved in a minimal
amount of H2O and loaded onto a column of 3.6 g of
AG1X2 acetate resin. The column was washed with eight
column volumes of H2O, and then the product was eluted
with 1 N HOAc to give a 61% crude yield. The product-
containing fractions were concentrated in vacuo, chased
with H2O followed by toluene, and dried under high
vacuum. The white solid obtained was then triturated
with MeOH to give 44% of the title compound: mp
Heteroexchange-mediated release of d-[3H]aspartate
from synaptosomes was quantified as described by
Koch.13 Essentially, synaptosomes were suspended in
assay buffer at a concentration of 0.45 mg of protein/
mL. Aliquots (10 mL) of this suspension were allowed
to incubate with 2.5 mM d-[3H]aspartate for 15 min at
25 ꢁC. Synaptosomes containing the [3H]-substrates
were reisolated by centrifugation (28,150g, 20 min,
4 ꢁC), rinsed, resuspended to 1 mg of protein/mL of ice-
cold assay buffer, and maintained on ice. The total
content of radiolabel in the synaptosomes was deter-
mined by adding 100 mL of the suspension to 2.9 mL of
ice-cold assay buffer and immediately vacuum filtering
as described above. Radioactivity present in the synap-
tosomes was quantified by liquid scintillation counting.
This value was determined at the beginning, middle, and
end of each experiment to ensure that the synaptosomal
content of d-[3H]aspartate did not change during main-
tenance on ice and that different preparations contained
similar levels of each of the radiolabeled substrates.
Assays quantifying the efflux of d-[3H]aspartate were
initiated by adding a 100-mL aliquot of synaptosomes
containing the radiolabeled compounds to 2.9 mL of
assay buffer pre-equilibrated and maintained at 37 ꢁC.
In those experiments in which heteroexchange was exam-
ined, the indicated compounds were included in the assay
buffer. Assays were terminated 1–5 min later by the addi-
tion of ice-cold assay buffer. The radioactivity remaining
in the synaptosomes was determined as described above.
25
(dec) 255–256 ꢁC; ½ꢀꢃD À54.3ꢁ (c 0.49, H2O), lit.18
À54.8); IR (KBr) br 3422, 3146, 3055, br 1933, 1711,
1617, 1456, 1400, 1344, 1272, 1228, 1156, 1044, 911
1
cmÀ1; H NMR (400 MHz, D2O calibrated to internal
HOD at 4.80) d 4.29 (app t, 1H, J=8.8 Hz), 3.82 (d, 1H,
J=12.0 Hz), 3.26 (d, 1H, J=12.0 Hz), 2.83 (dd, 1H,
J=13.6, 8.4 Hz), 2.07 (dd, 1H, J=13.6, 9.2 Hz), 1.39 (s,
3H); 13C NMR (125 MHz, D2O/CD3CN calibrated to
internal CHCD2CN at 118.20) d 177.4, 172.5, 60.0, 52.5,
48.6, 38.7, 19.9; HRMS C7H11N1O4+H calcd
174.0766 found 174.0762.
Computational analysis
Synaptosomal transport and exchange
The molecular modeling was performed on a Silicon
Graphics Iris computer using the Sybyl 6.2 software by
Tripos Inc. Minimization parameters used were: Tripos
force field, dielectric constant=80, Pullman charge cal-
culation, and Powell minimization method. The mini-
mized (local minima) structures were overlaid using a
three-point comparison: the two carboxylate carbons and
the nitrogen. The volume comparisons (green and yellow
mesh) are representations of the additive Van der Waals
radii of the substrate molecules (yellow) subtracted from
the 2(S),4(R)-4-Me-2,4-PDC Van der Waals radius
(green).
Rat cortical synaptosomes were prepared essentially by
the procedure of Booth and Clark,32 using a dis-
continuous Ficoll/sucrose gradient as described pre-
viously.13 Uptake of d-[3H]aspartate was measured
essentially by the method of Kuhar and Zarbin.33
Synaptosomes were suspended in assay buffer (10 mM
Tris–acetate, 128 mM NaCl, 10 mM d-glucose, 5 mM
KCl, 1.5 mM NaH2PO4, 1 mM MgSO4, 1 mM CaCl2,
pH 7.4) at a concentration of 0.2 mg of protein/mL.
After a 5-min preincubation at 25 ꢁC, uptake assays
were initiated by the addition of d-[3H]aspartate (1–20
mM) to the synaptosomes. In the inhibition experiments,
d-[3H]aspartate and inhibitor were added simulta-
neously. After a 2-min incubation at 25 ꢁC, the assay
was rapidly quenched by the addition of 4 mL of ice-
cold assay buffer. The suspension was quickly filtered
through Whatman GF/F micro-fiber filters and rinsed
Acknowledgements
The authors gratefully acknowledge that this work was
supported in part by the NIH: NINDS NS30570

C. S. Esslinger et al. / Bioorg. Med. Chem. 10 (2002) 3509–3515
3515
(R.J.B.), NS27600 (A.R.C.), NS10156 (C.S.E.) and
NCRR RR15583 (R.J.B., C.S.E.). The authors also
wish to thank M. P. Kavanaugh, C. M. Thompson, and
J. M. Gerdes for their insightful discussions.
15. Bridges, R. J.; Lovering, F. E.; Koch, H.; Cotman, C. W.;
Chamberlin, A. R. Neurosci. Lett. 1994, 174, 193.
16. Esslinger, C.; Koch, H.; Kavanaugh, M.; Phillips, D.;
Chamberlin, A.; Thompson, C.; Bridges, R. Bioorg. Med.
Chem. Lett. 1998, 8, 3101.
17. Arriza, J. L.; Fairman, W. A.; Wadiche, J. I. G. H. M.;
Kavanaugh, M. P.; Amara, S. G. J. Neurosci. 1994, 14, 5559.
18. Hull, C. E. Master’s Thesis, Department of Chemistry,
University of California Irvine, Irvine, CA, 1996.
19. In 1,3-Dipolar Cycloaddition Chemistry; Padwa, A., Ed.;
Wiley-Interscience: City 1984; Vols. 1 and 2.
20. Tsuge, O.; Kanemasa, S. In Advances in Heterocyclic
Chemistry; Academic: Wiley-Interscience, New York, 1989; Vol.
49, p 231.
References and Notes
1. Monaghan, D. T.; Bridges, R. J.; Cotman, C. W. Annu.
Rev. Pharmacol. Toxicol. 1989, 29, 365.
2. Monaghan, D. T.; Howson, P. A.; Jane, D. E.; Bridges, R.
J. In Biological Psychiatry; D’haenen, H., den Boer, J. A.,
Westenberg, H., Willner, P., Eds.; John Wiley and Sons: Chi-
chester, UK. In press.
3. Meldrum, B. S. J. Nutr. 2000, 130, 1007S.
4. Zipfel, G. J.; Babcock, D. J.; Choi, D. W. J. Neurotrauma
2000, 17, 857.
5. Gegelashvili, G.; Schousboe, A. Mol. Pharmacol. 1997, 52, 6.
6. Seal, R. P.; Amara, S. G. Annu. Rev. Pharmacol. Toxicol.
1999, 39, 431.
7. Takahashi, M.; Billups, B.; Rossi, D.; Sarantis, M.;
Hamann, M.; Atwell, D. J. Exp. Biol. 1997, 200, 401.
8. Danbolt, N. C. Prog. Neurobiol. 2001, 65, 1.
9. Robinson, M. B. Neurochem. Int. 1998, 33, 479.
10. Chen, W.; Aoki, C.; Mahadomrongkul, V.; Gruber, C. E.;
Wang, G. J.; Blitzbau, R.; Irwin, N.; Rosenberg, P. A. J.
Neurosci. 2002, 22, 2142.
21. DeSchong, P.; Kell, D. A.; Sidler, D. R. J. Org. Chem.
1985, 50, 2309.
22. Husinec, S.; Porter, A. E.; Roberts, J. S.; Strachan, C. H.
J. Chem. Soc., Perkins Trans 1 1984, 2517.
23. Tsuge, O.; Kanemasa, S.; Masayuki, O.; Yorozu, K.;
Takenaka, S.; Ueno, K. Bull. Chem. Soc. Jpn. 1987, 60,
4067.
24. Collado, I.; Ezquerra, J.; Mateo, A. I. C. P.; Rubio, A. J.
Org. Chem. 1999, 64, 4304.
25. Monn, J. A.; Valli, M. J. J. Org. Chem. 1994, 59, 2773.
26. Desimoni, G.; Faita, G.; Mella, M.; Righetti, P.; Zema,
M. Tetrahedron 1999, 55, 8509.
27. Karlsson, S.; Han, F.; Hogberg, H. E.; Caldirola, P.
Tetrahedron: Asymmetry 1999, 10, 2605.
28. Hosokawa, T.; Yamanaka, T.; Itotani, M.; Murahashi, F.
J. Org. Chem. 1995, 60, 6159.
29. Craig, D.; Munasinghe, R. N. Tet. Lett. 1992, 33, 663.
30. Luhowy, R. R.; Cieciuch, R. F.; Meneghini, F. Tetra-
hedron Lett. 1973, 1285.
11. Schmitt, A.; Asan, E.; Lesch, K. P.; Kugler, P. Neu-
roscience 2002, 109, 45.
12. Bridges, R. J.; Stanley, M. S.; Anderson, M. W.; Cotman,
C. W.; Chamberlin, A. R. J. Med. Chem. 1991, 34, 717.
13. Koch, H. P.; Kavanaugh, M. P.; Esslinger, C. S.; Zer-
angue, N.; Humphrey, J. M.; Amara, S. G.; Chamberlin,
A. R.; Bridges, R. J. Mol. Pharmacol. 1999, 56, 1095.
14. Bridges, R. J.; Lovering, F. E.; Humphrey, J. M.; Stanley,
M. S.; Blakely, T. N.; Cristofaro, M. F.; Chamberlin, A. R.
Bioorg. Med. Clem. Lett. 1993, 3, 115.
31. Koch, H. P.; Chamberlin, A. R.; Bridges, R. J. Mol.
Pharmacol. 1999, 55, 1044.
32. Booth, A. G.; Kenny, A. J. Biochem. J. 1974, 142, 575.
33. Kuhar, M. J.; Zarbin, M. J. Neurochem. 1978, 31, 251.