Design, synthesis, and biological evaluation of a scaffold for iGluR ligands based on the structure of (-)-dysiherbaine

Article abstract of DOI:10.1016/j.bmcl.2006.01.047

The design and synthesis of four 2,2-disubstituted dihydrobenzofurans that are structurally related to several glutamate-containing natural products, including (-)-dysiherbaine, is described. Biological evaluation of these analogs shows that one is a KA r

Full text of DOI:10.1016/j.bmcl.2006.01.047

Bioorganic & Medicinal Chemistry Letters 16 (2006) 2189–2194
Design, synthesis, and biological evaluation of a scaffold
for iGluR ligands based on the structure of (ꢀ)-dysiherbaine
Jamie L. Cohen,a Agenor Limon,b Ricardo Miledib and A. Richard Chamberlina,*
aDepartment of Chemistry, University of California at Irvine, Irvine, CA 92697, USA
bDepartment of Neurobiology and Behavior, University of California at Irvine, Irvine, CA 92697, USA
Received 22 December 2005; revised 10 January 2006; accepted 11 January 2006
Available online 7 February 2006
Abstract—The design and synthesis of four 2,2-disubstituted dihydrobenzofurans that are structurally related to several glutamate-
containing natural products, including (ꢀ)-dysiherbaine, is described. Biological evaluation of these analogs shows that one is a
KA receptor antagonist and another is an NMDA receptor agonist.
Ó 2006 Elsevier Ltd. All rights reserved.
1. Introduction
line the first essential steps in this plan, the design and
synthesis of a natural product-based library scaffold
candidate, along with several substituted variants, and
report their activity as iGluR ligands.
Over the past 5 years, significant attention has been
devoted to the chemical synthesis of natural product-like
libraries.1 As Nature has already identified natural prod-
ucts as lead structures of biological relevance, libraries
based on specific substructures found across a class of
natural products are, to an extent, biologically validat-
ed.2 Accordingly, libraries generated from core scaffolds
derived from natural products are proving to be a valu-
able source of biologically active compounds that pro-
vide new probes for molecular targets.3,4
Naturally occurring c,c-disubstituted glutamates such as
the kainate receptor agonists (ꢀ)-dysiherbaine9 (1) and
(ꢀ)-neodysiherbaine10 (2), isolated from the Microne-
sian marine sponge Dysidea herbacea, and (S)-(+)-lyco-
perdic acid11 (3), a non-proteinogenic a-amino acid
isolated from the mushroom Lycoperdon perlatum, have
attracted wide attention among the chemical and biolog-
ical communities due to their distinctive structures
(Fig. 1) and unique biological profiles.12–14 These struc-
turally similar glutamate-containing natural products
were therefore selected as viable, biologically validated
starting points for library design. Inspired by the current
interest in using specific substructures found within a
class of natural products as library scaffolds,4,7 a gluta-
mate-appended oxolane (tetrahydrofuran) ring—the
core structural motif shared by 1, 2, and 3—was selected
as the template for library synthesis. In order to provide
increased opportunities for structural diversity, a phenyl
ring was introduced as a framework upon which to dis-
play diverse functional groups. Thus, the c,c-disubstitut-
ed glutamate-appended dihydrobenzofuran 4 evolved as
the targeted scaffold for the synthesis of the next gener-
ation of glutamate analogs.
Our interest in natural product-based libraries was moti-
vated by our ongoing research efforts to discover novel
ligands for the ionotropic glutamate receptors
(iGluRs).5 Characterizing the iGluRs is a top priority
not only because of their function in normal CNS pro-
cesses such as learning and memory, but also due to
their role in causing damage in neurodegenerative
disorders including stroke, epilepsy, and Alzheimer’s
disease.6 Attempts to delineate the properties of the
iGluR subtypes, however, have been impeded by a lack
of selective ligands. Given the successes of natural prod-
uct-based libraries in the discovery of new ligands for
biological targets,1a,7,8 we envisioned application of this
concept to biologically active natural products known to
interact with glutamate receptors. In this Letter, we out-
Clearly, synthetic accessibility and retention of reason-
able levels of iGluR activity in the core scaffold 4 are
important prerequisites for the success of this plan. In
addition, information concerning which positions on
Keywords: Excitatory amino acid receptors; KA receptor agonist; KA
reeptor antagonist; NMDA receptor agonist; Dysiherbaine; Scaffold.
*
Corresponding author. E-mail: archambe@uci.edu
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.01.047

2190
J. L. Cohen et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2189–2194
OH
This model provided some reassurance that 4 might in-
deed act as an acceptable surrogate for 1–3 as a library
scaffold, and further, it suggested that substitutents
would be better tolerated at C-4 and/or C-5 of the aryl
ring than at C-6 or C-7. As an initial experimental test
of this model, three analogs were envisioned in addition
to the parent compound 4: the C-5 and C-7 methyl
derivatives, and the C-2 epimer (which in effect inter-
changes the oxygen and methylene substituents in the
tetrahydrofuran ring without changing the overall orien-
tation of binding). The calculated complex suggests not
only that a C-5 methyl analog of 4 should be a better li-
gand than its C-7 methyl counterpart, but also that
binding of the C-2 epimer would be unfavorable because
the ring oxygen lone pairs, rather than the methylene
group of the parent structure, would be in contact with
the electron-rich face of the tyrosine-450 phenyl ring
(assuming the overall orientation of the ligand does
not change). To investigate these computational predic-
tions, the analogs in question were synthesized and
submitted to biological evaluation.
NH2CH3
OH
OH
O
HO2C
H3N
O
O2C
O
O
CO2
CO2
H3N
Dysiherbaine, 1
Neodysiherbaine, 2
R
7
O
O
HO2C
O
HO2C
2
5
3
R
H3N
CO2
H3N
CO2
Target Analog, 4 (R = H)
Scaffold (R = diversity elements)
Lycoperdic Acid, 3
Figure 1. Glutamate-containing natural products and simplified
analog.
the phenyl ring might best tolerate substituents would be
useful from a library design perspective. Accordingly,
we conducted preliminary modeling studies in order to
assess computationally whether the proposed core scaf-
fold 4 might be a reasonable ligand for AMPA/KA
iGluR receptors.15 Specifically, 4 was docked into the
glutamate binding cleft of Gouaux’s X-ray structure of
a kainate–iGluR2 construct complex, which at the time
was the most reliable structure available for AMPA/KA
receptor binding sites.16,17 The resultant calculated
structure (Fig. 2) indicated that the aryl ring is well
accommodated in the binding site (contact residues are
shown in lighter teal). The a-amino and a-carboxyl
groups (not visible in the view shown) make contacts
analogous to the corresponding residues in the Gouaux
kainic acid complex upon which this structure was
based, while the methylene group of the
dihydrobenzofuran is favorably wedged against the face
of the phenolic ring of Tyr-450, as shown. In addition,
two of the four C–H bonds of the scaffold’s aryl
ring—in the 4- and 5-positions—are directed toward
the mouth of the binding cleft; in contrast, the corre-
sponding bonds at the 6- and 7-positions are buried.
The development of a stereoselective and scalable syn-
thesis of 4 is paramount to the success of this library
plan. The major synthetic challenge in the preparation
of 4 and related analogs is establishing the stereochem-
istry of the tetrasubstituted carbon center in the gluta-
mate-appended oxolane ring. A synthesis plan that
appeared most appealing in its generality, simplicity,
and potential diastereoselectivity was a three-step pyrog-
lutamate a-annulation consisting of: (1) a-benzylation
of a suitably protected pyroglutamate derivative with
an electrophile carrying a latent phenoxide nucleophile,
(2) diastereoselective a-halogenation, and (3) stereospe-
cific SN2 intramolecular displacement of the halide by
the phenoxide nucleophile.
Accordingly, the synthesis of scaffold 4 commenced with
a-benzylation of the protected pyroglutamate 518 using
an aldol/elimination/reduction sequence19 to afford the
3,5-disubstituted pyrrolidinone 6 in 85% yield over three
steps. The second step of this synthesis plan required the
diastereoselective halogenation of this pyrrolidinone,
which after an exhaustive screening of various halogena-
tion procedures was realized by the sequential treatment
of 6 with Et3N, TMSOTf, and N-bromosuccinimide to
afford a separable mixture of the diastereomeric bro-
mides 7 and 8 in 87% yield. The ratio of diastereomers
was 5:1 as determined by 1H NMR analysis of the crude
reaction mixture, and the stereochemistry of the major
isomer, 7, was established by NOESY studies on the
corresponding demethylated derivative 9. The third
step of the glutamate ring annulation, an SN2 cyclization
of brominated phenol 9, was achieved by treatment
with DBU at room temperature to furnish the
dihydrobenzofuran 10 in 91% yield and establish the ste-
reochemistry at the tetrasubstituted center (vide infra).
With all of the stereogenic centers of the target molecule
secured, sequential N-Boc protection and desilylation
produced pyroglutaminol 11 in 82% yield over two
steps. At that point, the stereochemistry of the tetrasub-
stituted a-center was verified by NOESY experiments,
Figure 2. Docking of 4 into Gouaux’s iGluR2 construct (a-carboxyl
and amino groups are obscured; see text for the significance of labeled
contacts).

J. L. Cohen et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2189–2194
2191
indicating that the SN2 cyclization to form 11 had pro-
ceeded stereospecifically with inversion, as expected, to
give the desired stereoisomer. Completion of the synthe-
sis proceeded with Jones oxidation of pyroglutaminol 12
to the corresponding carboxylic acid followed by imide
hydrolysis and N-Boc deprotection in refluxing 6M
HCl to provide crude 4 as a white solid.20 Purification
by reverse-phase HPLC afforded the pure parent scaf-
fold 4 as the TFA salt. The C-2 epimer 12 was similarly
prepared from the minor bromination diastereomer 8
(Scheme 1), and the C-5 and C-7 methylated analogs,
13 and 14, respectively, were likewise synthesized in an
identical sequence by substituting the appropriate meth-
yl-substituted anisaldehydes for the o-anisaldehyde em-
ployed in the initial benzylation of 5. Overall, this
sequence highlights an efficient and reasonably dia-
stereoselective method for the annulation of a substitut-
ed oxolane ring onto the c-position of glutamic acid,
and it provided in this case ample quantities for biolog-
ical evaluation.
in multiple oocyte preparations (N = 2–7), generally
with a standard error of only a few percent.
In the agonist activity assays (Fig. 3), there was no cur-
rent observed for the parent scaffold 4 or for its 5-methyl
analog 13 at concentrations of 100 lM, while the con-
trol agonist kainate induced currents at concentrations
as low as 10 lM; the C-2 epimer 12 also showed no
activity in this assay (data not shown). On the other
hand, at 100 lM the corresponding 7-methyl analog 14
alone did elicit a weak current that was 6 1% of the
kainate control at the same concentration. Significantly,
this signal was approximately doubled, to 13 4% of
control, in the presence of a 10 lM concentration of
the AMPA-specific potentiator cyclothiazide (CTZ).
These observations suggest that 14, but not 4, 12, or
13, is acting (at least in part) as an agonist—albeit a
rather weak one—at AMPA receptors.
Similar experiments were conducted on the analogs in the
presence of the NMDA-specific co-agonist glycine. Once
again, the parent scaffold 4 showed no activity under these
conditions, nor did 14; however, the 5-methyl analog 13
elicited a significant current at 100 lM, 44 11% of con-
trol, suggesting that it is an NMDA agonist.
The activities of the parent compound 4 and its analogs
(13, 14, and 12) were screened in oocytes injected with
native mRNA from rat cerebral cortex.21 Potential KA
receptor agonist activity was assayed based on the mem-
brane current generated at a concentration of 100 lM
for each analog against 100 lM of kainate as control.
The same compounds were also screened for antagonist
activity by measurements of 100 lM kainate-induced
membrane currents inhibited by the presence of each
analog (100 lM). Experiments were quite reproducible
In direct contrast to the complete inactivity of the parent
scaffold 4 in the agonist assays, it proved to be a reason-
ably potent antagonist that reduces kainate control
currents by approximately one-half (to 55 13% of
control) at a concentration of 100 lM (Fig. 4).
MeO
MeO
MeO
HO2C
O
Br
Br
d
a, b, c
CO2H
+
H3N
TFA
O
N
O
O
O
N
N
N
Boc
6
OTBDPS
Boc
H
H
12 (C-2 epimer)
OTBDPS
OTBDPS
OTBDPS
5
7
8
5 : 1
e-j
MeO
HO
O
HO2C
Br
i,j
g,h
e
f
Br
O
O
O
CO2H
H3N
TFA
O
O
N
N
H
N
H
O
N
H
OH
OTBDPS
OTBDPS
Boc
OTBDPS
4
7
9
10
11
Stereochemical Assignments and Key nOes:
OH
Me
H
H
O
O
H
H
HO2C
HO2C
H
H H
7
H
H
5
Me
H
H
OH
O
O
H
Boc
N
CO2H
CO2H
H3N
TFA
H3N
TFA
N
H
O
OTBDPS
13
14
Br
H
9
11
Scheme 1. Synthesis of parent core scaffold 4 and analogs 12, 13, and 14. Reagents and Conditions: (a)—(i) LiHMDS, THF, ꢀ78 °C (ii) o-
anisaldehyde, BF3 Æ OEt2, 99%; (b) Ph3P, I2, imidazole, CH2Cl2, 0 °C to rt, 92%; (c) H2, Pd/C, EtOAc, 93%; (d) TMSOTf, Et3N, NBS, CH2Cl2, 0 °C,
87%; (e) BBr3, CH2Cl2, 0 °C to rt, 63%; (f) DBU, CH2Cl2, 0 °C, 91%; (g) Boc2O, Et3N, DMAP, CH2Cl2, 0 °C to rt, 90%; (h) TBAF, AcOH, THF,
91%; (i) Jones reagent, acetone, 0 °C, 56%; (j) 6M HCl, reflux, then HPLC, 58%.

2192
J. L. Cohen et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2189–2194
Significantly from the perspective of the binding model,
neither the 7-methyl derivative 14 (95 3% of control)
nor the C-2 epimer 12 (100 3% of control) at
100 lM showed significant activity in antagonizing
100 lM kainate currents. Conversely, the C-5 methyl
analog 13 did block the kainate control currents, albeit
somewhat less potently than the parent compound, to
72 1% of control. For comparison, the very potent
antagonist CNQX was also assayed in this system,
blocking 38 3% of the current at a concentration of
1 lM. None of the analogs (at 100 lM) antagonized
100 lM NMDA + 100 lM glycine currents (data not
shown).
The potency of dysiherbaine as an AMPA/KA receptor
agonist led to the expectation that analogs such as 4
with similar core structures and steric characteristics
might also act as agonists at these same iGluRs;
however, the assays demonstrated that, instead, one
(13) is an NMDA receptor agonist and another (4) is
an AMPA/KA receptor antagonist. Specifically, none
of the four analogs 4, 12, 13, and 14 by itself elicited
strong currents relative to kainate. Indeed only 14
showed any detectable agonist activity, by itself, under
the assay conditions. The fact that the weak current
observed for 14 is doubled in the presence of the
AMPA-specific potentiator CTZ suggests that it is a
weak AMPA receptor agonist. More interestingly, the
5-methyl analog 13 (but not 4 or 14) did elicit a signifi-
cant current, but only in the presence of the NMDA
receptor co-agonist glycine. This result suggests that 13
is a selective NMDA receptor agonist with a potency
somewhat lower than that of NMDA itself (a detailed
characterization of the potency and efficacy of this
compound is underway). Thus, while 14 is a weak
AMPA/KA receptor agonist and 13 is a moderately ac-
tive (and selective) NMDA receptor agonist, the parent
scaffold 4 shows no detectable iGluR agonist activity.
Figure 3. Agonist activity. Comparison of parent dysiherbaine analog
4 to its C-2 epimer 12, and the methylated analogs 13 and 14. Agonist
activity assays were conducted in oocytes injected with native mRNA
from rat cerebral cortex, and the results (blue bars) are expressed as the
percentage of current induced by 100 lM kainate (with and without
the AMPA receptor desensitization inhibitor cyclothiazide (CTZ,
10 lM). NMDA currents (red bars) were measured for each analog in
the presence of NMDA plus the NMDA co-agonist glycine (both at
100 lM), with NMDA plus glycine (both at 100 lM) as the positive
control.
While this result was briefly disappointing in terms of
our original expectations, it assumed favorable signifi-
cance when the antagonist activity assays revealed that
an equimolar concentration 4 reduced kainate-induced
currents by approximately 50%, suggesting that the par-
ent analog 4 is an AMPA/KA receptor antagonist that is
comparable in potency to that of kainate itself.22 The 7-
methylated analog 14 and the C-2 epimer 12 were both
essentially inactive, and the 5-methyl derivative was
intermediate in potency, so that the order of potencies
determined in the antagonist assays is 4 > 13 ꢁ 14, 12.
The observation that NMDA/glycine currents are not
diminished by any of the analogs demonstrates that
the antagonist activity of 4 and 13 is selective for
AMPA/KA receptors. Further, while 13 is also an
NMDA receptor agonist, 4, in contrast, appears to be
strictly an AMPA/KA receptor antagonist.
Figure 4. Antagonist Activity. Comparison of parent dysiherbaine
analog 4 to its methylated counterparts 13 and 14, its C-2 epimer 12,
and the AMPA/KA antagonist CNQX. Antagonist activity assays
were conducted in oocytes injected with native mRNA from rat
cerebral cortex, and the results are expressed as the percentage of
current induced by kainate (100 lM) in the presence of equimolar
amounts of analogs (100 lM; no analog = 100%). In this comparison,
the very potent but non-selective CNQX was present at 1 lM. Inset
shows a typical example of a recording for this measurement (analog 4
in this case), with current on the y-axis and time on the x-axis.
While the antagonist activity of 4 and 13 was unantici-
pated based on the agonist characteristics of the dysiher-
baine core around which the analogs were designed, it is
certainly not unusual to observe such differences in
behavior for closely related compounds.12h In fact, we
recently reported an example in which analogs based

J. L. Cohen et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2189–2194
2193
4. Koch, M.; Waldmann, H. Drug Discovery Today 2005, 10,
471.
5. Shou, X.; Miledi, R.; Chamberlin, A. R. Biorg. Med.
Chem. Lett. 2005, 15, 3942.
6. (a) Doble, A. Pharmacol. Ther. 1999, 81, 163; (b) Loscher,
W. Prog. Neurobiol. 1998, 54, 721; (c) Trist, D. G. Pharm.
Acta Helv. 2000, 74, 221; (d) Bleakman, D.; Lodge, D.
Neuropharmacology 1998, 37, 1187.
7. Nicolaou, K. C.; Pfefferkorn, J. A.; Cao, G. Q.; Barlu-
enga, S.; Mitchell, H. J. J. Am. Chem. Soc. 2000, 122,
9939.
on the structure of the agonist kainate proved to be
antagonists.5 In that case, and in that of 4, we rational-
ize the antagonist behavior to be the result of the phenyl
ring of the respective ligands interfering with the forma-
tion of a critical inter-domain Glu402–Thr686 interac-
tion (residues labeled in yellow, Fig. 2) that develops
during the conformational change leading to channel
opening, as proposed by Gouaux for the known antag-
onist DNQX. Regardless of its mechanistic origins, this
unexpected activity, although serendipitous, provides an
exciting new lead for the development of new, selective
iGluR antagonists.
8. Koehn, F.; Carter, G. Nat. Rev. Drug Disc. 2005, 4, 206.
9. Sakai, R.; Kamiyla, H.; Murata, M.; Shimamoto, K.
J. Am. Chem. Soc. 1997, 119, 4112.
10. Sakai, R.; Koike, T.; Sasaki, M.; Shimamoto, K.; Oiwa,
C.; Yano, A.; Suzuki, K.; Tachibana, K.; Kamiya, H. Org.
Lett. 2001, 3, 1479.
11. (a) Banga, N. R.; Welter, A.; Jadot, J.; Casimir, J.
Phytochemistry 1979, 18, 482; (b) Lamotte, J.; Oleksyn, B.;
Dupont, L.; Didberg, O.; Campsteyn, H.; Vermeire, M.
Acta Crystallogr. 1978, B34, 3635.
While the number of analogs admittedly is small, it was
gratifying to note that the relative potencies of the four
compounds were consistent with the predictions of the
modeled complex, notwithstanding their unanticipated
antagonist behavior. The docked structure of 4 suggest-
ed that (a) the 4- and 5-positions of the aromatic ring
can tolerate additional steric bulk and may be amenable
to elaboration, (b) the 6- and 7-positions cannot be fur-
ther functionalized because they are buried, and (c) the
C-2 stereochemistry cannot be inverted because the re-
sult would be an unfavorable interaction between the
tetrahydrofuran ring oxygen lone pairs and the elec-
tron-rich face of the tyrosine-450 aryl ring. The assay re-
sults for 4, 13, 14, and 12 are consistent with these
predictions. Both 4 and 13 are in harmony with all three
criteria and show good activity, while 14 and 12 are
incongruent with (b) and (c), and are essentially inactive.
Thus supported experimentally, this calculated struc-
ture, in conjunction with the validated synthetic route
described in this Letter, provides a reasonable paradigm
for the design of focused libraries. Furthermore, the pro-
totype analog 4 is significant in its own right as a selec-
tive AMPA/KA receptor antagonist, as is the new
NMDA agonist 13. These results provide a significant
new lead for the development of iGluR antagonists
and agonists, and further motivate the generation of a
library based on 4, which will be reported in due course.
12. Dysiherbaine Total Syntheses: (a) Phillips, D. P.; Cham-
berlin, A. R. J. Org. Chem. 2002, 67, 3194; (b) Snider, B.
B.; Hawryluk, N. A. Org. Lett. 2001, 2, 635; (c) Masaki,
H.; Maeyama, J.; Kamada, K.; Esumi, T.; Iwabuchi, Y.;
Hatakeyama, S. J. Am. Chem. Soc. 2000, 122, 5216; (d)
Sasaki, M.; Koike, T.; Sakia, R.; Tachibana, K. Tetrahe-
dron Lett. 2000, 41, 3923; Dysiherbaine biological activity:
(e) Swanson, G. T.; Green, T.; Sakai, R.; Contractor, A.;
Che, W.; Kamiya, H.; Heinemann, S. Neuron 2002, 34,
589; (f) Sakai, R.; Swanson, G.; Shimamoto, K.; Green,
T.; Contractor, A.; Ghetti, A.; Tamura-Horikawa, Y.;
Owia, C.; Kamiya, H. J. Pharmacol. Exp. Ther 2001, 296,
650; (g) Sasaki, M.; Maruyama, T.; Sakai, R.; Tachibana,
K. Tetrahedron Lett 1999, 40, 3195; (h) Sanders, J.; Ito,
K.; Settimo, L.; Pentikainen, O.; Shoji, M.; Sasaki, M.;
Johnson, M.; Sakai, R.; Swanson, G. J. Pharmacol. Exp.
Ther. 2005, 314, 1068; (i) Shoji, M.; Shiohara, K.; Oikawa,
M.; Sakai, R.; Sasaki, M. Tetrahedron Lett. 2005, 46,
5559; Dysiherbaine synthetic efforts: (j) Kang, S.; Lee, Y.
Synlett 2003, 7, 993; (k) Miyata, O.; Iba, R.; Hashimoto,
J.; Naito, T. Org. Biomol. Chem. 2003, 1, 772; (l) Huang,
J.; Xu, K.; Loh, T. Synthesis 2003, 5, 755; (m) Naito, T.;
Nair, S.; Nishiki, A.; Yamashita, K.; Kiguchi, T. Hetero-
cycles 2000, 53, 2611.
13. Neodysiherbaine synthesis: (a) Lygo, B.; Slack, D.; Wil-
son, C. Tetrahedron Lett. 2005, 46, 6629; (b) See Ref. 10.
14. Lycoperdic acid total syntheses: (a) Makino, K.; Shintani,
K.; Yamatake, T.; Hara, O.; Hatano, K.; Hamada, Y.
Tetrahedron 2002, 58, 9737; (b) Masaki, H.; Mizozoe, T.;
Esumi, T.; Iwabuchi, Y.; Hatakeyama, S. Tetrahedron
Lett. 2000, 41, 4801; (c) Yoshifuji, S.; Kaname, M. Chem.
Pharm. Bull. 1995, 43, 1617; (d) Kaname, M.; Yoshifuji, S.
Tetrahedron Lett. 1992, 33, 8103; (e) Tamura, O.; Shiro,
T.; Ogasawara, M.; Toyao, A.; Ishibashi, H. J. Org. Chem.
2005, 70, 4569.
Acknowledgments
We are grateful for support from the National Institute
of Neurological Disorders and Stroke (NS-27600 to
A.R.C.) and the National Institute of General Medical
Sciences (GM07311, training grant support for J.L.C.).
References and notes
15. Molecular dynamics and minimizations of the ligand/
iGluR2S1S2 construct complex were carried out on a
Silicon Graphics Octane 2 computer using the Discover _3
module on Insight II version 2002. Dynamics were carried
out at 300 K. Only the residues of the S1S2 construct
1. (a) Boldi, A. Curr. Opin. Chem. Biol. 2004, 8, 281; (b)
Wessjohann, L. Curr. Opin. Chem. Biol. 2000, 4, 303; (c)
Nielsen, J. Curr. Opin. Chem. Biol. 2002, 6, 297; (d)
Kulkarni, B.; Roth, G.; Lobkovsky, E.; Porco, J. J. Comb.
Chem. 2002, 4, 56.
˚
within 7.0 A of the docked ligand were allowed to relax
during the minimization and dynamics simulation.
16. Armstrong, M.; Sun, Y.; Chen, G.-Q.; Gouaux, E. Nature
1998, 395, 913.
17. Crystal structures of the iGluR5 and iGluR6 have recently
been reported: (a) Mayer, M. Neuron 2005, 45, 539; (b)
Naur, P.; Vestergaard, B.; Skov, L.; Egebjerg, J.; Gajhede,
M.; Kastrup, J. FEBS Lett. 2005, 579, 1154.
2. (a) Breinbauer, R.; Vetter, I. R.; Waldmann, H. Angew.
Chem. Int. Ed. 2002, 41, 2878; (b) Balamurugan, R.;
Dekker, F.; Waldmann, H. Mol. BioSyst. 2005, 1, 36.
3. Lee, M.-L.; Schneider, G. J. Comb. Chem. 2001, 3, 284;
Shang, S.; Tan, D. Curr. Opin. Chem. Biol. 2005, 9, 248;
Messer, R.; Fuhrer, C.; Haner, R. Curr. Opin. Chem. Biol.
2005, 9, 259.

2194
J. L. Cohen et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2189–2194
18. (a) Otsuka, M.; Masuda, T.; Haupt, A.; Ohno, M.; Shiaki,
T.; Sugiura, Y.; Maeda, K. J. Am. Chem. Soc. 1990, 112,
838; (b) Evans, D.; Lundy, K. J. Am. Chem. Soc. 1992,
114, 1495; (c) Pickering, L.; Malhi, B.; Coe, P.; Walker, R.
Nucleosides Nucleotides 1994, 13, 1493; Shin, C.; Nakam-
ura, Y.; Yamada, Y.; Yonezawa, Y.; Umemura, K.;
Yoshimura, J. Bull. Chem. Soc. Jpn. 1995, 68, 3151;
Barrett, A.; Head, J.; Smith, M.; Stock, N.; White, A.;
Williams, D. J. Org. Chem. 1999, 64, 6005.
J = 7.5, 1H), 6.95 (d, J = 8.1, 1H), 4.16 (ddd, J = 9.4, 3.9,
1.2, 1H), 3.64 (d, J = 16.5, 1H), 3.36 (d, J = 16.5, 1H), 2.92
(dd, J = 15.5, 3.9, 1H), 2.44 (dd, J = 15.5, 9.4, 1H); 13C
NMR (125 MHz, D2O) d 176.4, 172.1, 157.8, 129.1, 125.9,
125.3, 122.5, 110.4, 89.1, 51.6, 41.2, 37.9; HRMS (ESI/
methanol) m/z calcd for C12H13NO5 (M+Na)+ 274.0692,
26
found 274.0681. ½aꢂD ꢀ20.0 (c 1.6, MeOH).
21. Polezani, L.; Woodward, R. M.; Miledi, R. Proc. Natl.
Acad. Sci. U.S.A. 1991, 88, 4318.
19. Ezquerra, J.; Pedregal, C.; Yruretagoyena, B.; Rubio, A.
J. Org. Chem. 1995, 60, 2925.
20. Characterization data for 4: H NMR (500 MHz, D2O) d
7.30 (d, J = 7.3, 1H), 7.25 (app t, J = 7.7, 1H), 7.02 (app t,
22. Kainate is not highly selective and can induce currents
through both AMPA and KA receptors; experiments are
underway with cloned receptor subtypes to determine
which specific iGluRs are involved.
1