NMR evaluation of interactions between substituted-indole and PDZ1 domain of PSD-95

Article abstract of DOI:10.1016/j.bmcl.2011.04.011

We synthesized small organic molecules designed as PDZ ligands. These indole-based compounds were evaluated for their interaction with the PDZ1 domain of the post-synaptic density 95 (PSD-95) protein. Three molecules were found to interact with the targeted PDZ protein by NMR. One of them showed chemical shift perturbations closely related to the natural ligands.

Full text of DOI:10.1016/j.bmcl.2011.04.011

Bioorganic & Medicinal Chemistry Letters 21 (2011) 3349–3353
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
NMR evaluation of interactions between substituted-indole and PDZ1
domain of PSD-95
Alexandre Vogrig a, Benjamin Boucherle a, Hemantkumar Deokar b, Isabelle Thomas b, Isabelle Ripoche b,
Lu-Yun Lian c, Sylvie Ducki b,
⇑
a Clermont Université, UBP, EA 987, LCHG, BP 10448, F-63000 Clermont-Ferrand, France
b Clermont Université, ENSCCF, EA 987, LCHG, BP 10448, F-63000 Clermont-Ferrand, France
c University of Liverpool, NMR Centre for Structural Biology, UK-L69 3BX Liverpool, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
We synthesized small organic molecules designed as PDZ ligands. These indole-based compounds were
evaluated for their interaction with the PDZ1 domain of the post-synaptic density 95 (PSD-95) protein.
Three molecules were found to interact with the targeted PDZ protein by NMR. One of them showed
chemical shift perturbations closely related to the natural ligands.
Received 23 February 2011
Revised 29 March 2011
Accepted 3 April 2011
Available online 9 April 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
PDZ domain
Protein–protein interaction
PSD-95
Analgesia
Post-synaptic density 95/Disc-Large/Zona ocludens-1 (PDZ) do-
mains (ꢀ90 amino acids) are highly conserved structural modules
which are found in signaling proteins of bacteria, yeast or animals
and mediate protein–protein interactions.1 These interactions in-
volve extreme carboxy-terminal of ligand proteins binding to a
hydrophobic groove between aB helix and bB strand, delimited by
the GLGF loop of the PDZ domain.2
PSD-95 is a three PDZ-domain containing protein. The first PDZ
domain (PDZ1) (Fig. 1) was found to interact with the C-terminal
extremity of the serotonin receptor 5HT2a.3 A disruption of the
interaction between these two proteins was found to reduce
hyperalgesia in rodent models.4 Thus, inhibiting the interaction be-
tween PSD-95 PDZ1 and 5-HT2a could lead to the development of a
novel class of analgesic agents.
Small molecules such as indoles A–C (Fig. 2) interrupting PDZ
interactions, were reported by Fujii and colleagues.5–8 Although
these small molecules were designed as peptido-mimetics of C-ter-
minus of PDZ ligands, little information is available about their
interactions with specific PDZ domains.
Here, we describe the design and the synthesis of novel indolic
analogues based on these previously reported indole-based PDZ
inhibitors. The objective is to carry out an in-depth study of the
chemical space within the ligand binding site and to establish
the relative importance of the different types of interactions within
Figure 1. Cartoon representation of the backbone of the 3D structure of PDZ1
domain of PSD95 (PDB 2KA9).
this binding site. These compounds were evaluated by Nuclear
Magnetic Resonance (NMR) spectrometry in order to identify their
interactions with the PDZ1 domain of PSD-95. A docking study was
subsequently carried out using a previously developed model to
explain the results.9
The substituted indoles were synthesized via o-iodo-anilines 2
and 4 which are easily obtained in three steps from the commer-
cially available nitro-benzoic acids 1 and 3 (Scheme 1).5,6
Palladium-catalyzed annulation10 between o-iodo-anilines 2 or
4 and 2-oxobutanoic acid (R3 = Me) or 2-oxopropanoic acid (R3 = H)
⇑
Corresponding author. Tel.: +33 473407132; fax: +33 473407008.
E-mail address: sylvie.ducki@univ-bpclermont.fr (S. Ducki).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.04.011

3350
A. Vogrig et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3349–3353
Trp(-1)
Ph
Ph
N
Val(0)
O
Val(0)
O
COOH
Asp(-3)
H
n-Bu
N
n-Bu
N
HO
HO
HO
Ile(-3)
OH
O
HO
HO
Ser/Thr(-2)
Thr(-2)
Ph
A
B
C
Figure 2. Fujii’s indoles A–C.
Table 1
O
O
Chemical shift perturbations observed by 15N-1H HSQC for compounds 5, 6, 8–10 and
12–14
O
I
a, b, c
a, b, c
HO
HO
MeO
Entry Cpds R1
R2
R3
R4
Chemical shift perturbation
observed
NO2
NO2
NH2
1
2 (94%)
1
2
3
4
5
6
7
8
5
6
8
CO2Me Me
Me
H
H
H
H
H
No
No
Yes
Yes
No
I
Me
CO2Me Me
CO2H
Me
CO2H
CO2H
H
Me
CO2H
Me
Me
CO2Et
CO2H
Me
Me
H
MeO
9
NH2
10
12
13⁄
14⁄
O
Me Me No
Me
Me
4 (34%)
3
H
H
No
Yes
H
Scheme 1. Synthesis of o-iodoanilines 2 and 4. Reagents and conditions: (a) SOCl2,
MeOH, 70 °C, 4 h; (b) Pd/C, H2 (1 atm), MeOH, rt, 24 h; (c) ICl, CaCO3, MeOH, H2O, rt,
24 h.
*
The synthesis of compounds 13 and 14 is described elsewhere.9
gave the substituted indoles 5–7 in moderate to good yields
(Scheme 2). After saponification, carboxylic acids 8–10 were
obtained.
O
R3
HO
R3
R2
R1
R2
R1
I
O
O
a
Indole 5 was N-methylated5 to afford ester 11 (Scheme 3) which
was quantitatively saponified to give carboxylic acid 12.
To evaluate the interaction between compounds 5, 6, 8–10 and
12 and PDZ domains, 15N labeled PDZ1 domain of PSD-95 was pro-
duced and purified after cloning and expression in BL21 E. coli.11
1H/15N heteronuclear single-quantum coherence NMR experi-
ments were recorded using labeled protein with potential ligands.
Chemical shifts assignment of the 15N-1H HSQC spectrum was
based on Ref. 12. Chemical shift changes in the 15N-1H HSQC were
then used to identify the amino acids whose chemical environ-
ments were perturbed in the presence of the compounds. These
variations may be due to direct interactions with the ligand or
could be induced by conformational changes in the protein tertiary
structure.
N
OH
NH2
H
5-7
2, 4
5, R1 = CO2Me, R2 = Me, R3 = Me (55%)
6, R1 = Me, R2 = CO2Me, R3 = Me (50 %)
7, R1 = CO2Me, R2 = Me, R3 = H (85%)
2, R1 = Me, R2 = CO2Me
4, R1 = CO2Me, R2 = Me
b, c
R3
R2
R1
O
OH
The data (Tables 1 and 2) show that while addition of acids 8, 9
and 14 results in several chemical shifts in the protein NMR spec-
trum, none of the esters 5, 6 and 13 led to a chemical shift pertur-
bation of the NMR spectrum. These data conclude on the
importance of the carboxylic acid moiety on the six-membered
ring of the indole, suggesting it is important for the ligands’ inter-
actions with the protein.
Acids 10 and 12 did not induce variations of chemical shifts.
These inactive analogues of indole 8 demonstrate the importance
of substituents at positions 1 and 3 of the indole. Indeed, addition
of a methyl group at position 1 or suppression of the methyl group
at position 3 led to non-interacting molecules.
Specific chemical shift perturbations were observed upon addi-
tion of indole 14 to the protein: signals for amino acids G20, I27,
H69 and L76 shifted (Fig. 4a). The first two amino acids (G20,
I27) are located on the loop between bBÀ and bCÀsheets while
the other two (H69, L76) are on aBÀhelix (Table 2, entry 1).
Although close, none of these perturbations are consistent with
N
H
8-10
8, R1 = CO2H, R2 = Me, R3 = Me (quant.)
9, R1 = Me, R2 = CO2H, R3 = Me (quant.)
10, R1 = CO2H, R2 = Me, R3 = H (45%)
Scheme 2. Synthesis of substituted indoles 8–10. Reagents and conditions: (a)
Pd(OAc)2, DABCO, DMF, 105 °C, 16 h; (b) KOH (2 M), reflux, 24 h; (c) HCl (1 M).
O
O
O
a
b
5
O
HO
N
N
OH
11
O
O
12
Scheme 3. Synthesis of substituted indole 12. Reagents and conditions: (a) K2CO3,
CH3I, DMF, RT, 12 h (90%); (b) KOH (5 M), reflux, 24 h (quant.).

A. Vogrig et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3349–3353
3351
Table 2
Localization of amino acids showing chemical shift perturbation due to ligand (s.c. = side chain)
Entry
Compounds
bA/bB loop
GLGF
bB
bB/bC loop
bE/aB loop
aB
1
2
3
14
9
8
G20, I27
H69, L76
H69, V73
H69, V73, L76
N11
N11
S17, I18
S17
G21, N24 (s.c.)*, H26, I27
G21, N24, N24 (s.c.)*, H26, I27,
G15, F16
V67
*
s.c. = side chain.
the ligand interacting within the defined groove.12 Thus, indole 14
is probably interacting with PDZ domain at the generic binding site
but not as deep as natural ligands. Perturbations of chemical shifts
observed were moderate, meaning that 14 is a weak binder.
Indole 9 was designed as a more hydrophobic derivative of 14,
by addition of a methyl group at position 6. This methyl moiety
was introduced to mimic hydrophobic residues at position 0 (first
C-term residue) of the natural ligands. Indole 9 was found to per-
turb the chemical environment of more amino acids than 14. Fur-
thermore, many of these amino acids are located within the
targeted pocket (Table 2, entry 2; Fig. 3b). Two amino acids of
bB-sheet (S17, I18) and two of aB-helix are perturbed by interac-
tions between 9 and the PDZ protein. Perturbations induced by
6-methyl indole 9 are related to perturbations described for the
natural ligand. Indeed, due to the additional methyl group, this
compound may be inserting deeper into the groove delimitated
by bB and aB. However, amino acids of the GLGF loop, which are in-
volved in the interaction with terminal acid function of natural li-
gands, did not shift upon addition of compound 9. Furthermore,
the binding is still weak because perturbations of chemical shift
are moderate.
Chemical shifts, measured during 1H/15N HSQC experiments
with PDZ1 of PSD-95 and compound 8, are represented on Fig. 5.
Binding of compound 8 and its structural isomer 9 to the PDZ pro-
tein generated quite similar perturbations (Table 2, entries 2 and 3;
Fig. 4c and d). However, new amino acids located in the GLGF loop
(G15 and F16) showed chemical shift changes upon binding of in-
dole 8.
R3
R2
O
N
R4
R1
OH
Indole 8 was found to interact in a very similar way to natural
ligands because all parts of the binding pocket were involved. Per-
turbations for amino acids located in GLGF loop (G15 and F16), bB
sheet (S17), bB/bC loop (G21, N24, H26 and I27) and aB helix (H69,
V73 and L76) were noted. Despite moderate binding, compound 8
Figure 3. General structure of evaluated compounds.
Figure 4. 3D structure of PDZ1 domain of PSD95 (PDB 2KA9). Yellow indicates GLGF loop; purple indicates amino acids with chemical shift perturbation in presence of: (a)
14, (b) 9, (c) and (d) 8; orange indicates amino acids with chemical shift perturbation in the GLGF loop (purple + yellow).

3352
A. Vogrig et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3349–3353
Figure 5. Chemical shift perturbation of PDZ1domain of PSD95. Blue dots correspond to 1H/15N HSQC of the free protein; red dots correspond to the protein + 10 equiv of
indole 8.
appeared to interact within the targeted groove permitting its eval-
uation as a basis to build competitive inhibitor of natural ligands.
The observation that H69 is involved in binding of the three
compounds 8, 9, and 14 is consistent with an interaction in the
same way as natural ligands. Indeed this histidine is known to be
responsible for class selectivity of PDZ domain by hydrogen bond-
ing to hydroxyl group of amino acid at position -2 (thirs C-term
residue) of the natural ligand. H69 is probably interacting with
the carboxylic function of the 5-membered ring of the indole.
Molecular docking is one of the most important and useful
method for predicting and investigating protein–ligand interac-
tions. We performed molecular docking study using Molecular
Operating Environment software (MOE 2010.10),13 to access bind-
ing affinity between PDZ-1 domain of PSD-95 and substituted in-
doles. The purpose of the docking study was to compare the
interactions determined by NMR for the best interacting com-
pound with the docking results. Previously we had carried out
docking study on indole moiety,9 the same protocol was used for
the present docking study. The indole moiety has been modified
at various positions in order to optimize lead molecule.
The best interacting indole 8 was docked into the PDZ1 domain
of PSD-95 in order to try to explain the NMR results. Several dock-
ing poses were found, and the best pose showed indole 8 bound to
the PDZ domain as anticipated (Fig. 6) with total interaction energy
of 7.41 kcal/mol. When the docking pose was analyzed more clo-
sely, it was found that the 6-COOH made hydrogen bond (H-bond)
interactions with G15, P16 of GLGF loop, and made several other
interactions vdW and columbic interaction with G13, L14, G15,
P16, S17, I18, A19, H69, V73, L76. The other carboxylic group (2-
COOH) made H-bond interaction with I18 and vdW and columbic
interaction with H69, I18. The 3-methyl group interacted with
V73 and L76 through vdW and columbic interactions. The indole
nucleus also showed vdW and columbic interactions with P16,
Figure 6. Docking of compound 8 into PDZ1 domain showing H-bond interactions
(red) and other interactions (green).
I18 and L76.
The docking pose is in agreement with the NMR results ob-
addition of a methyl group at this position (compound 12) proved
detrimental to interaction witht eh PDZ protein. This is probably
explained by the steric hindrance of the N-Me too close to the
bB-sheet of the PDZ domain. Finally, the 5-Me was found important
by NMR but no interaction was detected on the docked pose. This
could be explained by an electronic effect of this group more than a
hydrophobic or steric effect.
tained. Indeed the acid analogues were found to interact while
the esters did not, which is explained by the fact that both carbox-
ylic acid moieties are involved in several interactions with the pro-
tein. The methyl group at position 3 was found to interact with the
protein, which is consistent with the fact that 3-demethylated ana-
logue 10 did not show interactions with the protein. Although, N1
is not involved in any interaction according to the docking pose, an

A. Vogrig et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3349–3353
3353
Overall, interactions observed by docking and NMR data were
consistent with each other. We identified three ligands of the
PSD-95 PDZ1 domain. One of them, compound 8, led to perturba-
tions in chemical shift very closely related to changes induced by
natural ligands. This observation was confirmed by molecular
docking.
experiments are available.) associated with this article can be
found, in the online version, at doi:10.1016/j.bmcl.2011.04.011.
References and notes
1. Ducki, S.; Bennett, E. Curr. Chem. Biol. 2009, 3, 146.
2. Doyle, D. A.; Lee, A.; Lewis, J.; Kim, E.; Sheng, M.; MacKinnon, R. Cell 1996, 85,
1067.
Acknowledgements
3. Bécamel, C.; Gavarini, S.; Chanrion, B.; Alonso, G.; Galéotti, N.; Dumuis, A.;
Bockaert, J.; Marin, P. J. Biol. Chem. 2004, 279, 20257.
4. Pichon, X.; Wattiez, A. S.; Becamel, C.; Ehrlich, I.; Bockaert, J.; Eschalier, A.;
Marin, P.; Courteix, C. Mol. Ther. 2010, 18, 1462.
5. Fujii, N.; Haresco, J. J.; Novak, K. A. P.; Gage, R. M.; Pedemonte, N.; Stokoe, D.;
Kuntz, I. D.; Kiplin Guy, R. Bioorg. Med. Chem. Lett. 2007, 17, 549.
6. Fujii, N.; Haresco, J. J.; Novak, K. A. P.; Stokoe, D.; Kuntz, I. D.; Guy, R. K. J. Am.
Chem. Soc. 2003, 125, 12074.
7. Mahindroo, N.; Punchihewa, C.; Bail, A. M.; Fujii, N. Bioorg. Med. Chem. Lett.
2008, 18, 946.
8. Mayasundari, A.; Ferreira, A. M.; He, L.; Mahindroo, N.; Bashford, D.; Fujii, N.
Bioorg. Med. Chem. Lett. 2008, 18, 942.
We thank Robert Gibson and Mark Tully for their assistance
during protein production/purification and Marie Phelan for her
expertise in NMR. We also would like to thank the French National
Research Agency (ANR) for funding of the project PDZ-CANPAIN
and a postdoctoral fellowship for BB; the Regional Council of Au-
vergne (Conseil Régional d’Auvergne) and the European Fund for
Regional Economical Development (FEDER) for a doctoral scholar-
ship for AV; and the PRES Clermont Université for a postdoctoral
fellowship for HD.
9. Boucherle, B.; Vogrig, A.; Deokar, H. K.; Ripoche, I.; Thomas, I.; Ducki, S. 2011,
submitted for publication.
10. Chen, C.-y.; Lieberman, D. R.; Larsen, R. D.; Verhoeven, T. R.; Reider, P. J. J. Org.
Chem. 1997, 62, 2676.
Supplementary data
11. More information could be found in Supplementary data.
12. Piserchio, A.; Pellegrini, M.; Mehta, S.; Blackman, S. M.; Garcia, E. P.; Marshall,
J.; Mierke, D. F. J. Biol. Chem. 2002, 277, 6967.
13. Chemical Computing Group Inc.: Montreal, Canada, 2008; p. http://
www.chemcomp.com.
Supplementary data (Synthetic procedure, analysis data, proto-
cols for protein production and purification and protocols for NMR