Analogues of the neuroprotective tripeptide Gly-Pro-Glu (GPE): Synthesis and structure-activity relationships

Article abstract of DOI:10.1016/j.bmcl.2005.03.015

A series of GPE analogues, including modifications at the Pro and/or Glu residues, was prepared and evaluated for their NMDA binding and neuroprotective effects. Main results suggest that the pyrrolidine ring puckering of the Pro residue plays a key role

Full text of DOI:10.1016/j.bmcl.2005.03.015

Bioorganic & Medicinal Chemistry Letters 15 (2005) 2279–2283
Analogues of the neuroprotective tripeptide Gly-Pro-Glu
(GPE): synthesis and structure–activity relationships
a
Sergio A. Alonso De Diego, Pilar Mun˜oz, Rosario Gonzalez-Mun˜iz, Rosario Herranz,
a
a
´
Mercedes Martın-Martınez, Edurne Cenarruzabeitia, Diana Frechilla,
a
c
c
´
´
´
Joaquın Del Rıo, M. Luisa Jimeno and M. Teresa Garcıa-Lopez
c
b,
a,
*
*
´
´
´
a
´
´
Instituto de Quımica Medica, Juan de la Cierva, 3, 28006 Madrid, Spain
´
´
Departamento de Farmacologıa-Unidad Asociada al CSIC, Universidad de Navarra, Irunlarrea, 1, 31080 Pamplona, Spain
b
´
Centro de Quımica Organica Manuel-Lora Tamayo (CSIC), Juan de la Cierva, 3, 28006 Madrid, Spain
c
Received 23 December 2004; revised 23 February 2005; accepted 3 March 2005
Abstract—A series of GPE analogues, including modifications at the Pro and/or Glu residues, was prepared and evaluated for their
NMDA binding and neuroprotective effects. Main results suggest that the pyrrolidine ring puckering of the Pro residue plays a key
role in the biological responses, while the preference for cis or trans rotamers around the Gly-Pro peptide bond is not important.
Ó 2005 Elsevier Ltd. All rights reserved.
It has been reported that the tripeptide H-Gly-Pro-Glu-
OH (GPE) (1), naturally cleaved from the N-terminal
sequence of insulin-like growth factor (IGF-1)1,2 shows
neuroprotective properties, both in vitro and in different
animal models of neurodegenerative diseases, such as
Huntington and Parkinson diseases and hypoxia-medi-
ated ischemic brain injury.3–5 This tripeptide displays a
different range of biological actions compared to the
remaining IGF-1 fragment, des(1-3) IGF-1, known as
truncated IGF-1. Thus, truncated IGF-1 acts as a potent
neurotrophic factor via the IGF-1 receptor while GPE
neither binds to this receptor nor has any growth pro-
moting effect.6 Although the mode of action of this tri-
peptide remains unclear, there is evidence that it plays
a neuromodulatory role in the CNS. Initial structural
considerations suggested that GPE could interact with
one or more glutamate receptor types and, in fact, it
was demonstrated that it binds to the N-methyl-D-aspar-
tate (NMDA) receptor but not to the a-amino-3-hydr-
oxy-5-methyl-4-isoxazolepropionic acid (AMPA) or
kainate (KA) receptors.1 This pioneering study indi-
cated that while the C-terminal glutamate is necessary
for NMDA receptor binding, the N-terminal glycine
potentiates this cross-reaction. It has also been shown
that GPE potentiates the potassium-evoked release of
acetylcholine and dopamine through an unknown mech-
anism and via the NMDA receptor, respectively.1,5–7
Our current interest in exploring novel strategies for the
further development of effective neuroprotective drugs
directed our attention to the search of non-peptide
GPE mimics, which could be used as suitable pharmaco-
logical tools for gaining insight into the mode of action
of this tripeptide. To this aim, we have followed the gen-
eral structure/conformation/activity approach towards
peptidomimetics based on the structure of the native
peptide.8 Because of the unique conformational proper-
ties of proline,9 we considered that this amino acid res-
idue should play a key role in the biological activity of
GPE, as it happens with bioactive peptides containing
Xaa-Pro residues.10 In particular, unlike other peptide
bonds which essentially adopt the trans form, the Xaa-
Pro amide bond exists, in unfolded proteins and short
peptides, as a mixture of cis and trans isomers11 which,
in the case of the Gly-Pro dipeptide residue of GPE, cor-
responds to a 19% cis content in 90/10 H2O/D2O solu-
tion, as indicated our preliminary 1H NMR
conformational analysis.12 A number of alkylpro-
lines13,14 and pseudoprolines (WProÕs)15,16 derived from
cysteine, serine and threonine (Xaa[WR1R2pro]) have
been used to alter the cis/trans ratio of the peptide bond
N-terminal to prolyl residues and, hence, to study the
Keywords: Gly-Pro-Glu (GPE); Neuroprotection; Structure–activity
relationships.
*
Corresponding authors. Tel.: +34 915622900; fax: +34 915644853;
e-mail: iqmgl37@iqm.csic.es
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.03.015

2280
S. A. Alonso De Diego et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2279–2283
H-Cys [ΨMe,Mepro]-OH
relationship between cis/trans conformers and peptide
bioactivity.17,18 In this context, the easily synthetically
Boc-Gly-OH
a
accessible
2,2-dimethyl-thiazolidine-4(R)-carboxylic
acid, Cys[WMe,Mepro] is a useful tool to induce up to
100% the cis conformation,19 while the (S)-a-methylpro-
line (PMe) induces preferentially the trans form.20 On
this basis, we have replaced the Pro residue in GPE with
Cys[WMe,Me pro] and PMe, respectively. On the other
hand, it is known that the presence of the sulfur atom
of the thiazolidine moiety alters the ring puckering of
Cys[WR1R2pro] as compared to proline,16,21 which may
result in significant changes in the conformational prop-
erties of peptides incorporating thiazolidine proline
surrogates. Therefore, to determine the influence of the
proline ring puckering on the neuroprotective effect
of GPE, we have also prepared the corresponding
analogues containing Cys[WH,Hpro] and L-5,5-di-
methylproline (dmP).21,22 To further explore the spatial
requirements of GPE for this effect and to establish the
influence of the length of the Glu side chain, stereoiso-
meric GPE analogues incorporating one or both chiral
residues with D-configuration and analogues containing
aspartic or homoglutamic (Hgl) acid have also been syn-
thesized.23 Here, we report the synthesis of the GPE
analogues 2–12 (Fig. 1), and their in vitro biological ef-
fects, in comparison to the native neuroprotective tri-
peptide GPE (1).
Boc-Gly-Cys[ΨMe,Mepro]-OH
H-Glu(OtBu)-OtBu
a
Boc-Gly-Cys[ΨMe,Mepro]-Glu(OtBu)-OtBu
b
2
Scheme 1. Reagents: (a) BOP, DIEA, CH2Cl2; (b) 15% TFA, CH2Cl2.
tative in 90/10 H2O/D2O,25 indicating a drastic increase
in this conformer in comparison to GPE. In the case of
the Cys[WH,Hpro] analogue 3, the fully protected C-ter-
minal dipeptide was firstly prepared by coupling the
commercial Fmoc-WPro derivative with H-Glu(OtBu)-
OtBu (Scheme 2). Fmoc deprotection with piperidine
followed by coupling with Boc-Gly-OH and subsequent
t
removal of the acid labile Boc and Bu groups provided
compound 3 in 56% total yield. As expected, the H
1
NMR spectrum of this compound revealed the presence
of cis and trans conformers around the Gly-[WH,Hpro]
peptide bond in a 23:77 ratio, similar to that of Gly-
Pro in GPE.25,26
The synthesis of the target tripeptides was performed in
solution using BOP/DIEA activation in dichlorometh-
ane for the coupling reactions. Boc- and Bu-protected
t
Gly and Glu were, respectively, used for the synthesis
of the WPro-containing tripeptides 2 and 3 since the
Cys-derived thiazolidine ring system is stable under the
Boc/tBu cleavage conditions.24 As indicated in Scheme
1, Cys[WMe,Mepro], readily obtained by cyclocondensa-
tion of Cys with acetone,16 was directly coupled
with Boc-Gly-OH via BOP/DIEA activation. Similar
coupling of the resulting WPro dipeptide with H-Glu-
(OtBu)-OtBu, followed by concomitant removal of the
For the synthesis of the GPE analogues containing PMe
,
dmP and D-Pro 4–7, the Na-Boc proline derivatives,
Boc-AA1-OH, were coupled with L-or D-glutamic acid
dibenzyl esters H-AA2(OBzl)-OBzl (Scheme 3). Re-
moval of the N-terminal Boc group followed by cou-
pling with Z-Gly-OH led to the fully protected
tripeptides 13–16, which upon hydrogenolysis in the
presence of 10%Pd/C gave compounds 4–7 (63–70% to-
tal yield). Although cis and trans conformers around the
Gly-PMe and Gly-dmP peptide bond were observed in
the 1H NMR spectra of tripeptides 4 and 5, respectively,
the trans content increased by 9% in compound 4, com-
t
Boc and Bu protecting groups led to compound 2 in
25% overall yield. Inspection of the H NMR spectrum
1
of this 2-C dimethylated WPro containing tripeptide
showed that the content of the cis population of the
Gly-Cys [WMe,Mepro] amide bond is practically quanti-
Fmoc-Cys [ΨH,Hpro]-OH
H-Gly-Pro-Glu-OH (GPE)
H-Gly-Cys[ΨMe,Mepro]-Glu-OH
H-Gly-Cys[ΨH,Hpro]-Glu-OH
H-Gly-D-Pro-D-Glu-OH
H-Gly-Pro-D-Glu-OH
1
2
7
8
H-Glu(OtBu)-OtBu
a
H-Gly-Pro-Asp-OH
9
3
Fmoc-Cys[ΨH,Hpro]-Glu(OtBu)-OtBu
H-Gly-PMe-Glu-OH
H-Gly-dmP-Glu-OH
H-Gly-Pro-D-Asp-OH
H-Gly-Pro-Hgl-OH
H-Gly-Pro-D-Hgl-OH
10
11
12
4
1. 20% piperidine
2. Boc-Gly-OH
a
5
H-Gly-D-Pro-Glu-OH
H
6
Boc-Gly-Cys[ΨH,Hpro]-Glu(OtBu)-OtBu
COOH
N3
1
S
4
5
b
Cys[ΨMe,Mepro] =
Me
Me
2
3
Figure 1. GPE analogues studied.
Scheme 2. Reagents: (a) BOP, DIEA, CH2Cl2; (b) 15% TFA, CH2Cl2.

S. A. Alonso De Diego et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2279–2283
2281
the affinity is similar to that exhibited by (R)-AP5, one
of the most potent selective competitive NMDA
antagonists.28
Boc-AA1-OH
H-AA2(OBn)-OBn
a, b
+
The highly trans-populated PMe containing compound 4
and its dmP analogue 5, in which the Gly-dmP bond was
found to mainly adopt a cis conformation, showed very
similar binding affinities. These results seem to suggest
that the Gly-Pro amide conformation is not crucial
for the interaction of GPE with the NMDA receptor
complex. It is interesting to note that neither the
Cys[WMe,Mepro] derivative 2 (>98% cis conforma-
tion) nor the Cys[WH,H pro] analogue 3 (23:77 cis/trans
ratio) were able to bind the NMDA receptor. The pres-
ence of the sulfur atom, that could alter the ring pucker-
ing and the hydrophobicity of the Cys[WMe,Me pro] and
Cys[WH,Hpro] residues, as compared to proline, could
serve to explain the loss of affinity of pseudoprolines 2
and 3 with respect to GPE and to compounds 4 and
5.16 Anyway, the lack of affinity of the stereoisomeric
GPE analogues 6–8 also indicate that the spatial
requirements for interacting GPE with the NMDA
receptor are strict. At this point, it is worth pointing
out that, a survey of literature reports on the struc-
ture–activity relationships of ligands interacting with
the glutamate binding site of the NMDA receptor
reveals the preference for the D-configuration, but a
few exceptions have been described for both agonists
and antagonists including D-Glu which interacts much
less potently with NMDA than does L-Glu.28–30 This
is also the case of H-Gly-Pro-D-Glu-OH (8) which, in
contrast to its diastereoisomer containing L-Glu, GPE,
did not inhibit the L-[3H]glutamate binding at concen-
trations up to 10À4 M. On the other hand, it has also
been found that the a-carboxyl group and the side chain
acidic function are separated by four or six atoms in
most of the competitive NMDA antagonists.28,30 How-
ever, homologation of the Glu residue in GPE to give
the Hgl derivative 11, bearing a side chain of four
carbon atoms linking the two acidic groups, led to the
loss of affinity. In summary, the studies on the struc-
ture–activity described here show the scarce tolerance
of GPE to structural modifications. This makes its mode
of interaction with the NMDA receptor an intriguing
question.
H-AA1-AA2(OBn)-OBn
13-16
Z-Gly-OH
a, c
H-Gly-AA1-AA2-OH
4-7
4, 13 : AA1 = PMe, AA2 = Glu
5, 14 : AA1 = dmP, AA2 = Glu
6, 15 : AA1 = D-Pro, AA2 = Glu
7, 16 : AA1 = D-Pro, AA2 = D-Glu
Scheme 3. Reagents: (a) BOP, DIEA, CH2Cl2; (b) HCl, EtOAc; (c)
10% Pd/C, H2, MeOH.
pared to the model compound GPE, whereas the cis
content increased by 50% in compound 5 with respect
to this model.27
Finally, the GPE analogues modified at the glutamate
residue 8–12 were prepared by coupling of the commer-
cially available Z-Gly-Pro-OH with the corresponding
acidic amino acid dibenzyl ester followed by hydrogeno-
lysis, under the conditions indicated in Scheme 4 (total
yields ranging from 70% to 85%).
Displacement of L-[3H]glutamate from rat brain synap-
tic membranes by all these GPE analogues was deter-
mined and the results were compared to those of GPE
and (2R)-2-amino-5-phosphonopentanoic acid [(R)-
AP5], which were included in the same assay (Table
1). No displacement was found using concentrations
up to 100 lM of compounds 2, 3, 6–12 (data not
shown). However, analogues 4 and 5 incorporating the
PMe and dmP residues, respectively, displayed a notice-
able affinity for GluRs which favourably compares to
that of GPE. Particularly, in the case of compound 5,
Compounds 4 and 5, which were effective in displacing
L-[3H]glutamate from rat membranes, were further
investigated to evaluate their neuroprotective effects in
cultured hippocampal neurons exposed to NMDA
100 lM, or oxygen–glucose deprivation (OGD) follow-
ing described procedures.31,32 As shown in Figure 2A,
Z-Gly-Pro-OH
H-AA3(OBzl)-OBzl
a, b
+
H-Gly-Pro-AA3-OH
Table 1. Displacement of L-[3H]glutamate from rat brain synaptic
membranes
8 : AA3 = D-Glu
9 : AA3 = Asp
Compound
Ki (lM)
10 : AA3 = D-Asp
11 : AA3 = Hgl
L-Glutamate
(R)-AP5
0.87 0.21
3.83 1.76
31.24 15.65
7.96 1.83
3.79 0.53
12 : AA3 = D-Hgl
GPE (H-Gly-Pro-Glu-OH) (1)
H-Gly-PMe-Glu-OH (4)
H-Gly-dmP-Glu-OH (5)
Scheme 4. Reagents: (a) BOP, DIEA, CH2Cl2; (b) 10% Pd/C, H2,
MeOH.

2282
S. A. Alonso De Diego et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2279–2283
References and notes
1. Sara, V. R.; Carlsson-Skwirut, C.; Bergman, T.; Jo¨rnvall,
H.; Roberts, P. J.; Crawford, M.; Hakansson, L. N.;
Civalero, I.; Nordberg, A. Biochem. Biophys. Res. Com-
mun. 1989, 165, 766.
2. Yamamoto, H.; Murphy, L. J. J. Endocrinol. 1995, 146,
141.
3. Alexi, T.; Hughes, P. E.; Van Roon-Mom, W. M. C.;
Faull, R. L. M.; Williams, C. E.; Clark, R. G.; Gluckman,
P. D. Exp. Neurol. 1999, 159, 84.
4. Guan, J.; Krishnamurthi, R.; Waldvogel, H. J.; Faull, R.
L. M.; Clark, R.; Gluckman, P. Brain Res. 2000, 859, 286.
5. Sizonenko, S. V.; Sirimanne, E. V.; Williams, C. E.;
Gluckman, P. D. Brain Res. 2001, 922, 42.
6. Sara, V. R.; Carlsson-Skwirut, C.; Drakemberg, K.;
Giacobini, M. B.; Hakansson, L.; Mirmiran, M.; Nord-
berg, A.; Olson, L.; Reinecke, M.; Stahlbom, P. A.;
Nordqvist, A. C. S. Ann. N.Y. Acad. Sci. 1993, 692, 183.
7. Nilson-Hakansson, L.; Civalero, I.; Zhang, X.; Carlsson-
Skwirut, C.; Sara, V. R.; Nordberg, A. NeuroReport 1993,
4, 111.
8. Hruby, V. J. Nature Rev. Drug Discovery 2002, 1, 847.
9. Schmid, F. X.; Mayr, L. M.; Mucke, M.; Scho¨nbrunner,
¨
E. R. Adv. Prot. Chem. 1993, 44, 25.
10. Yaron, A.; Naider, F. Crit. Rev. Biochem. Mol. Biol. 1993,
28, 31.
11. Wedemeyer, W. J.; Welker, E.; Scheraga, H. A. Biochem-
istry 2002, 41, 14637, and references cited therein.
12. Conformer population ratio was determined by integra-
tion and/or deconvolution of the Gly a-CH2 protons (cis:
3.95 and 3.74 ppm; trans: 4.03 and 3.98 ppm).
13. Beausoleil, E.; Lubell, W. D. J. Am. Chem. Soc. 1996, 118,
12902.
Figure 2. Percentage of protection of rat hippocampal neurons from
NMDA (100 lM) excitotoxicity (A) and oxygen–glucose deprivation
(B). Cell survival was estimated by measuring the activity of
mitochondrial dehydrogenase on the tetrazolium derivative MTT.
Values are the mean SEM of 8–13 experiments. Asterisks indicate
difference from control (*P < 0.05, **P < 0.01, StudentÕs t-test).
14. Beausoleil, E.; Sharma, R.; Michnick, S. W.; Lubell, W.
J. Org. Chem. 1998, 63, 6572.
15. Haack, T.; Mutter, M. Tetrahedron Lett. 1992, 33, 1589.
¨
16. Keller, M.; Sager, C.; Dumy, P.; Schutkowski, M.;
the neuronal death caused by NMDA administration
was significantly prevented on pretreatment with the tri-
peptide analogues 4 and 5, at the two concentrations
assayed, 10 and 100 lM. These neuroprotective effects
appeared to be of similar potency to that of GPE at
the same concentrations in this in vitro assay. Although
neuroprotective effects of i.v. infusion of exogenous
GPE have been described in hypoxic-ischemic rats,33
this tripeptide was not able to increase significantly the
neuronal survival after OGD injury. However, com-
pounds 4 and 5 showed neuroprotective activity, which
was particularly significant in the case of the PMe con-
taining analogue 4, with a protection value of 28.5%.
Fischer, G. S.; Mutter, M. J. Am. Chem. Soc. 1998, 120,
2714.
¨
17. An, S. S. A.; Lester, C. C.; Peng, J. L.; Li, Y. J.; Rothwarf,
D. M.; Welker, E.; Thannhauser, T. W.; Zhang, L. S.;
Tam, J. P.; Scheraga, H. A. J. Am. Chem. Soc. 1999, 121,
11558.
18. Wittelsberger, A.; Keller, M.; Scarpellino, L.; Patiny, L.;
Acha-Orbea, H.; Mutter, M. Angew. Chem., Int. Ed. 2000,
¨
39, 1111.
19. Dumy, P.; Keller, M.; Ryan, D. E.; Rohwedder, T.; Wo¨hr;
Mutter, M. J. Am. Chem. Soc. 1997, 119, 918.
¨
20. Delaney, N. G.; Madison, V. J. J. Am. Chem. Soc. 1982,
104, 6635.
In conclusion, our findings suggest that the pyrrolidine
ring puckering of the Pro residue is a key determinant
for the NMDA binding and the neuroprotective effects
of GPE. Contrastingly, the preference for cis or trans
rotamers around the Gly-Pro peptide bond has no influ-
ence on these biological responses. The above results
offer interesting perspectives for further investigating
on GPE analogues incorporating other proline surro-
gates. This work is now in progress.
21. An, S. S. A.; Lester, C. C.; Peng, J.-L.; Li, Y.-Jin.;
Rothwarf, E. W.; Thannhauser, T. W.; Zhang, L. S.; Tam,
J. P.; Scheraga, H. A. J. Am. Chem. Soc. 1999, 121, 11558.
22. Xia, Q.; Ganem, B. Tetrahedron Lett. 2002, 43, 1597.
23. During the preparation of this manuscript the D-Glu-
containing GPE analogue was also described by others:
Trotter, N. S.; Brimble, M. A.; Harris, P. W. R.; Callis, D.
J.; Sieg, F. Bioorg. Med. Chem. 2005, 13, 501.
24. Wo¨hr, T.; Wahl, F.; Nefzi, A.; Rohwedder, T.; Sato, T.;
Sun, X.; Mutter, M. J. Am. Chem. Soc. 1996, 118, 9218.
¨
25. Although traces of the trans isomer cannot be excluded in
compound 2 (1–2%), no evidence for this form was
found either by H1 or 13C NMR. The cis/trans ratio of
compound 3 was determined in 90/10 H2O/D2O by
integration of the Glu a-H signals (dcis: 4.62 ppm and
Acknowledgements
This work has been supported by the Spanish Ministry
of Education and Science (SAF2003-07207-C02).
S.A.A.D.D. holds a predoctoral fellowship from this
Ministry for Associated Units to CSIC.
dtrans: 4.56 ppm).
26. Kern, D.; Schutkowski, M.; Drakenberg, T. J. Am. Chem.
Soc. 1997, 119, 8403.

S. A. Alonso De Diego et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2279–2283
2283
27. The cis/trans ratio of compounds 4 (10:90) and 5 (69:31)
was determined by integration of the Me and Me2 of PMe
30. Tikhonova, I. G.; Baskin, I. I.; Palyulin, V. A.; Zefirov, N.
S.; Bachurin, S. O. J. Med. Chem. 2002, 45, 3836, and
references cited therein.
31. Snider, B. J.; Moss, J. L.; Revilla, F. J.; Lee, C. S.;
Wheeler, V. C.; MacDonald, M. E.; Choi, D. W.
Neuroscience 2003, 120, 617.
and dmP, respectively (4cis: 1.55 ppm; 4trans: 1.57 ppm; 5cis
:
1.50 ppm; 5trans: 1.43 ppm), in the 1H NMR spectra in
90:10 H2O/D2O.
28. Brauner-Osborne, H.; Egebjerg, J.; Nielsen, E. O.; Mad-
¨
sen, U.; Krogsgaard-Larsen, P. J. Med. Chem. 2000, 14,
2609, and references cited therein.
29. Ornstein, P. L.; Klimkowski, V. J. In Excitatory Amino
Acid Receptors: Design of Agonists and Antagonists;
Krogsgaard-Larsen, P., Hansen, J. J., Eds.; Ellis
Horwood: Chichester, 1992; pp 183–200.
´ ´
32. De Cristobal, J.; Cardenas, A.; Lizasoain, I.; Fernandez-
´
Tome, P.; Lorenzo, P.; Moro, M. A. Stroke 2002, 33,
261.
33. Guan, J.; Thomas, G. B.; Lin, H.; Mathai, S.; Bachelor,
D. C.; George, S.; Gluckman, P. D. Neuropharmacology
2004, 47, 892, and references cited therein.