Target hopping as a useful tool for the identification of novel EphA2 protein-protein antagonists

Article abstract of DOI:10.1002/cmdc.201300305

Lithocholic acid (LCA), a physiological ligand for the nuclear receptor FXR and the G-protein-coupled receptor TGR5, has been recently described as an antagonist of the EphA2 receptor, a key member of the ephrin signalling system involved in tumour growth. Given the ability of LCA to recognize FXR, TGR5, and EphA2 receptors, we hypothesized that the structural requirements for a small molecule to bind each of these receptors might be similar. We therefore selected a set of commercially available FXR or TGR5 ligands and tested them for their ability to inhibit EphA2 by targeting the EphA2-ephrin-A1 interface. Among the selected compounds, the stilbene carboxylic acid GW4064 was identified as an effective antagonist of EphA2, being able to block EphA2 activation in prostate carcinoma cells, in the micromolar range. This finding proposes the "target hopping" approach as a new effective strategy to discover new protein-protein interaction inhibitors. Target hopping: Given the ability of lithocholic acid to recognize FXR, TGR5 and EphA2 receptors, we hypothesized the structural requirements to bind each of these receptors might be similar. We selected a set of commercially available FXR or TGR5 ligands and tested them for their ability to inhibit EphA2 by targeting the EphA2-ephrin-A1 interface. Moreover, a small panel of GW4064 derivatives was synthesized. Copyright

Full text of DOI:10.1002/cmdc.201300305

CHEMMEDCHEM
COMMUNICATIONS
DOI: 10.1002/cmdc.201300305
Target Hopping as a Useful Tool for the Identification of
Novel EphA2 Protein–Protein Antagonists
Massimiliano Tognolini, Matteo Incerti, Daniele Pala, Simonetta Russo, Riccardo Castelli,
Iftiin Hassan-Mohamed, Carmine Giorgio, and Alessio Lodola*[a]
Lithocholic acid (LCA), a physiological ligand for the nuclear re-
ceptor FXR and the G-protein-coupled receptor TGR5, has
been recently described as an antagonist of the EphA2 recep-
tor, a key member of the ephrin signalling system involved in
tumour growth. Given the ability of LCA to recognize FXR,
TGR5, and EphA2 receptors, we hypothesized that the structur-
al requirements for a small molecule to bind each of these re-
ceptors might be similar. We therefore selected a set of com-
mercially available FXR or TGR5 ligands and tested them for
their ability to inhibit EphA2 by targeting the EphA2-ephrin-A1
interface. Among the selected compounds, the stilbene car-
boxylic acid GW4064 was identified as an effective antagonist
of EphA2, being able to block EphA2 activation in prostate car-
cinoma cells, in the micromolar range. This finding proposes
the “target hopping” approach as a new effective strategy to
discover new protein–protein interaction inhibitors.
and MDM4 regulatory proteins.[8] In addition to its pleiotropic
binding profile, LCA possesses poor physicochemical proper-
ties that concur to liver toxicity[9] and colon cancerogenesis in
vivo.[3]
LCA has been recently identified as an antagonist of the
ephrin type A receptor 2 (EphA2),[10] a member of the Eph–
ephrin signalling system involved in tumour growth and vascu-
lar functions during carcinogenesis.[11] Specifically, LCA was
shown to reversibly disrupt the EphA2–ephrin-A1 protein–pro-
tein interface and to block EphA2 phosphorylation in intact
cells.[10]
Recently, other classes of small-molecule antagonists of
EphA2 have been reported,[12,13] with some of them obtained
by chemical modification of the 3a-hydroxy-5b-cholanic core
of LCA[14] or via condensation with naturally occurring a-amino
acids.[15] Due to the moderate selectivity and the limited chem-
ical flexibility of LCA-based derivatives, we turned our atten-
tion to the search for alternative chemotypes able to block the
activity of the EphA2 receptor by targeting the EphA2–ephrin-
A1 interface.
Lithocholic acid (LCA, Figure 1) is a secondary bile acid pro-
duced by bacterial biotransformation of chenodeoxycholic
acid. Together with other bile acids, LCA participates in the
We applied here the “target-hopping” approach in which
a ligand for one target is used as a starting point to derive
new leads for another target.[16] Given the ability of LCA to rec-
ognize FXR,[17] TGR5[18] and EphA2 receptors, we hypothesized
that the structural requirements for a small molecule to bind
each of these receptors might be similar. We thus selected
a set of commercially available FXR or TGR5 ligands, compris-
ing both natural and synthetic compounds, to be tested on
the EphA2 receptor. Aiming at finding new chemical entities
with better properties than those of LCA, yet able to block the
activity of EphA2, test candidates were first evaluated by
shape similarity to assess their structural diversity, then docked
within the ligand binding domain of the EphA2 receptor[19] to
ascertain the goodness of fit of the resulting EphA2–ligand
complexes.
Figure 1. Chemical structure (left) and three-dimensional representation of
LCA in its minimum energy conformation (right).
control of glucose metabolism and bile acid homeostasis,
mainly by binding to the nuclear farnesoid X receptor (FXR)
and to the G-protein coupled receptor TGR5.[1–4] LCA is also
known for its biological activity on other targets, including the
nuclear pregnane X receptor (PXR)[5] and vitamin D receptor
(VDR),[6] the 20S subunit of the proteasome,[7] and the MDM2
The selected compounds are reported in Figure 2 and in-
clude TGR5 agonists, that is, ciprofloxacin, betulinic and olea-
nolic acids, FXR agonists, that is, 6-ethylchenodeoxycholic acid
(6-ECDCA), ursodeoxycholic acid (UDCA), TTNBP and GW4064,
and one FXR antagonist, namely guggulsterone E.
[a] Dr. M. Tognolini,+ Dr. M. Incerti,+ Dr. D. Pala, Dr. S. Russo, Dr. R. Castelli,
Dr. I. Hassan-Mohamed, Dr. C. Giorgio, Dr. A. Lodola
Dipartimento di Farmacia, Universitꢀ degli Studi di Parma
V. le delle Scienze 27 A, 43124 Parma (Italy)
E-mail: alessio.lodola@unipr.it
[+] These authors equally contributed to the work.
Shape similarity between the selected compounds and LCA
was assessed (Figure 3) using Phase software.[20] With this ap-
proach, each compound of the dataset was flexibly super-
posed on LCA (in its minimum-energy conformation) and a sim-
ilarity score (equal to one for identical structures and to zero
for unrelated configurations) was retrieved from the overall
volume overlap.[21]
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201300305.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 67 – 72 67

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
Figure 2. Compounds selected for testing on the EphA2 receptor. Ciprofloxa-
cin, betulinic and oleanolic acids are known TGR5 agonist, 6-ethyl cheno-
deoxycholic acid (6-ECDCA) ursodeoxycholic acid (UDCA), TTNBP, GW4064
are FXR agonists, whereas guggulsterone E is a FXR antagonist.
Figure 3. Shape-similarity superposition of TGR5 and FXR ligands on LCA
(white carbons in panels a–h): a) ciprofloxacin (magenta carbons); b) betulin-
ic acid (green carbons); c) oleanolic acid (violet carbons); d) UDCA (pink car-
bons); e) 6-ECDCA, (yellow carbons); f) TTNPB (dark green carbons);
g) GW4064 (orange carbons); h) guggulsterone E (cyan carbons). For each
superposition, the van der Waals surface of LCA and the similarity score are
also reported.
The similarity score ranged from 0.898 for UDCA (Figure 3d)
to 0.485 for oleanolic acid (Figure 3c). This large variation in
the three-dimensional shape of the FXR and TGR5 ligands is
a consequence of the selection of five distinct chemical classes,
that is, quinolones (exemplified by ciprofloxacin), triterpenoids
(i.e., betulinic acid), bile acids (i.e., 6-ECDCA), stilbene carboxyl-
ic acids (i.e., GW4064), and steroids (i.e., guggulsterone E).
Despite their structural diversity, all the selected FXR and
TGR5 ligands can be docked within the binding site of the
EphA2 receptor giving scores (i.e., XP Glide Score)[22,23] compa-
rable to that of LCA (table S1, Supporting Information). In gen-
eral, the top-ranked pose for each docked compound was
characterized by the insertion of the central scaffold of the
molecule into the hydrophobic channel of EphA2 and by the
accommodation of the peripheral carboxyl or carbonyl func-
tion close to Arg103 (Figure 4). In the cases of 6-ECDCA, UDCA,
and GW4064, a salt bridge involving the guanidine group of
Arg103 was also observed, which concurred to the high dock-
ing score of these compounds. For betulinic acid only, the car-
boxylate group remained far away from Arg103 (Figure 4b).
Taken together, this computational analysis supports the
notion that some of the selected FXR or TGR5 ligands might
effectively bind the EphA2 receptor.
Considering these promising indications, the compounds
were tested for their ability to prevent EphA2–ephrin-A1 asso-
ciation by an enzyme-linked immunosorbent assay (ELISA), at
50 mm concentration. The FXR agonists GW4064[24] and 6-
ECDCA[25] were able to significantly displace ephrin-A1 binding
from EphA2 at this concentration (Figure 5), whereas the re-
maining compounds were inactive or poorly active. This con-
firms that high shape similarity with LCA is not a stringent re-
quirement for EphA2 binding, and target hopping can provide
new scaffolds with significant chemical diversity. Interestingly,
the docking poses of GW4064 (Figure 4g) and 6-ECDCA (Fig-
ure 4e) were ranked as the best ones by the Glide XP scoring
function (table S1). On the other hand, the same scoring func-
tion was not able to capture the detrimental effect of placing
a protonated group into the EphA2 binding site. This was the
case with ciprofloxacin (Figure 4a) which, despite being well
scored by Glide XP, resulted inactive in the binding assay.
Both GW4064 and 6-ECDCA dose-dependently disrupted the
EphA2–ephrin-A1 complex, giving a half-maximal inhibitory
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 67 – 72 68

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
Figure 5. Percentage of ephrin-A1 binding to EphA2 receptor after co-incu-
bation with TGR5 or FXR ligands at 50 mm concentration. One-way ANOVA
followed by Dunnet’s post-test was performed comparing control to all
other columns (biotinylated ephrin-A1-Fc, not shown=0%).
&
*
Figure 6. Dose–response curves for GW4064 ( ; IC50 =72 mm), 6-ECDCA (
;
!
IC50 =23 mm) and LCA ( ; IC50 =55 mm) towards EphA2–ephrin-A1 interac-
tion. GW4064, 6-ECDCA and LCA displace the binding of ephrin-A1-Fc from
immobilized EphA2-Fc in a dose-dependent manner. IC50 values are the
mean from at least three independent experiments; error bars represent
standard errors.
Figure 4. Docking of TGR5 and FXR ligands on the EphA2: a) ciprofloxacin
(magenta carbons); b) betulinic acid (green carbons); c) oleanolic acid (violet
carbons); d) UDCA (pink carbons); e) 6-ECDCA (yellow carbons); f) TTNPB
(dark green carbons); g) GW4064 (orange carbons); h) guggulsterone E (cyan
carbons). In all panels, LCA in its docked conformation[14] is also reported.
centrations of the compound under study (12, 25 and 50 mm).
As shown in Figure 7, GW4064 dose-dependently inhibited
EphA2 phosphorylation (IC50 =31 mm), thus being slightly more
potent than LCA (IC50 =48 mm).
concentration (IC50) of 23 mm and 72 mm, respectively
(Figure 6). LCA, used as standard compound, gave an IC50
value of 55 mm. Finally, dilution experiments showed that the
binding of GW4064 and 6-ECDCA to EphA2 was reversible, sim-
ilar to what was observed for LCA (figure S1).
We set to synthesize a small panel of GW4064 derivatives to
identify the minimal requirements of the stilbene carboxylic
acid class to bind the EphA2 receptor, following synthetic
routes similar to those described in reference [24] (scheme S1
and S2, Supporting Information). The potency values of the
newly synthesized compounds are shown in Table 1.
We next chose to investigate further the stilbene derivative
GW4064, which emerged as the most active compound in the
ELISA. We tested its ability to disrupt the EphA2–ephrin-A1
complex in a functional test on prostate adenocarcinoma cells
(PC3 cells) naturally overexpressing the EphA2 receptor.[26] The
EphA2-dependent kinase activity of PC3 cells was stimulated
with 0.25 mgmLꢀ1 ephrin-A1-Fc in presence of different con-
We started our exploration by moving the carboxyl group of
GW4064 (1) to the ortho and the para positions, which led to
compounds 2 and 3 (PCM303), respectively (Table 1). Whereas
for compound 2 a significant decrease of potency was ob-
served, an activity comparable to that of GW4064 was ob-
served for compound 3. Esterification of the carboxyl group of
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 67 – 72 69

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
distal phenyl ring can form a salt bridge with Arg103 only
when it is positioned in meta or in para position (Figure 8).
We next investigated the importance of the stilbene moiety
by substituting the olefinic fragment of compound 3 with an
Figure 7. GW4064 dose-dependently inhibits EphA2 phosphorylation. Data
are the means of at least three independent experiments; error bars repre-
sent standard errors. Data are normalized to unstimulated (Fc+1% DMSO)
= 0% and stimulated (0.25 mgmLꢀ1 ephrin-A1-Fc+1% DMSO) = 100%.
One-way ANOVA followed by Dunnet’s post-test was performed comparing
DMSO to all other columns.
Figure 8. Docking of compound 3 (ball and stick representation) in the
ephrin-binding pocket of the EphA2 receptor (white cartoons, white carbon
atoms). The G-H loop of ephrin-A1 is also displayed (dark grey cartoons).
Table 1. IC50 values for GW4064 analogues.
amide (compound 5) or by replacing the vinyl benzoic acid
portion with a simple carboxylic acid, emerging from the cen-
tral phenyl ring (compound 6, Table 1). These compounds re-
sulted inactive in the binding assay, confirming the importance
of both lipophilicity and steric occupancy of the EphA2 bind-
ing site for disrupting the EphA2–ephrin-A1 interaction.
The ability of compound 3 to act as a functional antagonist
of the EphA2 receptor was verified on adenocarcinoma PC3
cells. Similarly to GW4064, compound 3 dose-dependently in-
hibited EphA2 phosphorylation, with an IC50 value of 46 mm,
confirming that stilbene carboxylic acids can act as effective
antagonists of the EphA2 receptor in cancer cells (figure S2,
Supporting Information).
Compd
R1
IC50 [mm][a]
22ꢁ3.0
1 (GW4064)
2
93ꢁ10
The discovery of small molecules able to block protein–pro-
tein interaction is a challenging task due to the plasticity of
protein–protein interfaces but most importantly to the unbal-
ance between available screening libraries and chemical space
of protein–protein antagonists.[27] In the present work, by ap-
plying a “target-hopping” strategy, we discovered the stilbene
carboxylic acid scaffold as a new template for obtaining novel
antagonists of the EphA2 receptor. Although improvements in
potency and selectivity of this class are still required, we here
provide evidence that it is possible to block the activity of the
EphA2 with small molecules which lack the troublesome fea-
tures of previously reported ligands. These include (1) the (5b)-
cholan-24-oic acid scaffold, whose chemical exploration is lim-
ited by the paucity of functional groups, (2) the aryl-2,5-dime-
thylpyrrole moiety which suffers from poor chemical stability,[28]
(3) (poly)phenolic compounds lacking specificity[29,30] (4) pep-
tides,[31,32] which have ADME drawbacks.[33] Taken together,
these observations confirm the usefulness of the target-hop-
ping approach in the discovery of protein–protein inhibitors
and may help the design of the next generation of Eph–ephrin
system antagonists.
3 (PCM303)
35ꢁ5.0
4
Inactive[b]
5
6
>100[c]
>100[c]
[a] Values are means ꢁ standard error from three independent experi-
ments. [b] Inhibition not significant and lower than 20% when tested at
100 mm. [c] Inhibition between 20% and 50% when tested at 100 mm.
3 gave the inactive methyl ester 4, suggesting that the pres-
ence of a negatively charged group with a specific orientation
in the active site is required for biological activity. Docking
studies support this hypothesis as the carboxyl group on the
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 67 – 72 70

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
were seeded in 12-well plates at a concentration of 105 cellsmLꢀ1
,
Experimental Section
1 mL/well, in complete medium until they reached ꢂ70% conflu-
ence and were serum-starved overnight. The day after, cells were
treated with the test compounds, vehicle or standard drug, stimu-
lated with 0.25 mgmLꢀ1 ephrin-A1-Fc, rinsed with sterile PBS and
solubilized in lysis buffer. The lysates were resuspended and
rocked at 48C for 30 min and then centrifuged (14000ꢂg, 5 min).
The protein content in the supernatant was measured with a bicin-
choninic acid (BCA) protein assay kit (Thermo Scientific Pierce) and
standardized to 150 mgmLꢀ1. EphA2 phosphorylation was mea-
sured in cell lysates using a DuoSet IC Sandwich ELISA (R&D Sys-
tems, #DYC4056), following the manufacturer’s protocol.
Molecular modelling: Farnesoid X receptor (FXR) and G-protein
coupled receptor TGR5 ligands were built using Maestro 9.2[34] and
their geometry optimized using OPLS2005[35] force field in combi-
nation with GB/SA solvation model. Shape similarity analysis was
performed with Phase 3.3.[20] Each ligand of the database was flexi-
bly aligned on LCA in its global minimum conformation. Volume
overlap scores were then computed between superposed atom
pairs and a single shape similarity value, proportional to the total
volume overlap was calculated for each ligand. A conformational
search was applied to each ligand of the database during the
alignment process. Fifty alignments were collected for each com-
pound and ranked according to the computed similarity score.
Those showing the highest score are shown in Figure 3.
Acknowledgements
EphA2 structure (PDBID: 3HEI)[19] was prepared for docking studies
using Maestro 9.2, paying particular attention to the protonation
state of titratable residues.[36] During the docking simulations, the
protein structure was kept rigid. The binding site of EphA2 is not
expected to undergo a significant conformational rearrangement
during ligand binding as reported in recent literature.[37]
This work was supported by the Ministero dell’Universitꢀ e della
Ricerca, “Futuro in Ricerca” program (project code: RBFR10FXCP)
and the Associazione Italiana per la Ricerca sul Cancro (AIRC),
“My first AIRC” grant program (project code: 6181).
Docking simulations were then performed using Glide 5.7[22] start-
ing from the minimized structure of the compounds placed in an
arbitrary position within a region centred on the channel of EphA2,
delimited by Arg103, Phe156 and Arg159, using enclosing and
bounding boxes of 20 and 14 ꢁ on each side, respectively. Van der
Waals radii of the protein atoms were not scaled, whereas van der
Waals radii of the ligand atoms with partial atomic charges lower
than j0.15j were scaled by 0.8. Extra precision mode was ap-
plied.[21]
Keywords: drug design
receptors · structure–activity relationships · virtual screening
· protein–protein interactions ·
[1] G. Porez, J. Prawitt, B. Gross, B. Staels, J. Lipid Res. 2012, 53, 1723–1737.
[2] C. Thomas, R. Pellicciari, M. Pruzanski, J. Auwerx, K. Schoonjans, Nat.
Rev. Drug Discovery 2008, 7, 678–693.
[3] B. W. Katona, C. L. Cummins, A. D. Ferguson, T. Li, D. R. Schmidt, D. J.
Mangelsdorf, D. F. Covey, J. Med. Chem. 2007, 50, 6048–6058.
[4] A. Gioiello, E. Rosatelli, R. Nuti, A. Macchiarulo, R. Pellicciari, Expert Opin.
Ther. Pat. 2012, 22, 1399–1414.
[5] J. L. Staudinger, B. Goodwin, S. A. Jones, D. Hawkins-Brown, K. I. MacK-
enzie, A. LaTour, Y. Liu, C. D. Klaassen, K. K. Brown, J. Reinhard, T. M. Will-
son, B. H. Koller, S. A. Kliewer, Proc. Natl. Acad. Sci. USA 2001, 98, 3369–
3374.
[6] M. Makishima, T. T. Lu, W. Xie, G. K. Whitfield, H. Domoto, R. M. Evans,
M. R. Haussler, D. J. Mangelsdorf, Science 2002, 296, 1313–1316.
[7] Z. Dang, A. Lin, P. Ho, D. Soroka, K. H. Lee, L. Huang, C. H. Chen, Bioorg.
Med. Chem. Lett. 2011, 21, 1926–1928.
Chemistry: FXR and TGR5 ligands were purchased from Sigma–Al-
drich. Compounds 2–6 were synthesized according to the proce-
dure reported in the Supporting Information. The final compounds
were characterized by NMR and MS, and their purity was assessed
by elemental analysis for C, H and N. The percentages found were
within ꢁ0.4% of the theoretical values.
[8] S. M. Vogel, M. R. Bauer, A. C. Joerger, R. Wilcken, T. Brandt, D. B. Veprint-
sev, T. J. Rutherford, A. R. Fersht, F. M. Boeckler, Proc. Natl. Acad. Sci. USA
2012, 109, 16906–16910.
[9] A. F. Hofmann, Drug Metab. Rev. 2004, 36, 703–722.
[10] C. Giorgio, I. H. Mohamed, L. Flammini, E. Barocelli, M. Incerti, A. Lodola,
M. Tognolini, PLoS One 2011, 6, e18128.
[11] E. B. Pasquale, Nat. Rev. Cancer 2010, 10, 165–180.
[12] R. Noberini, I. Lamberto, E. B. Pasquale, Semin. Cell Dev. Biol. 2012, 23,
51–57.
[13] A. Petty, E. Myshkin, H. Qin, H. Guo, H. Miao, G. P. Tochtrop, J. T. Hsieh, P.
Page, L. Liu, D. J. Lindner, C. Acharya, A. D. Mackerell, Jr., E. Ficker, J.
Song, B. Wang, PLoS One 2012, 7, e42120.
[14] M. Tognolini, M. Incerti, I. Hassan-Mohamed, C. Giorgio, S. Russo, R.
Bruni, B. Lelli, L. Bracci, R. Noberini, E. B. Pasquale, E. Barocelli, P. Vicini,
M. Mor, A. Lodola, ChemMedChem 2012, 7, 1071–1083.
[15] M. Incerti, M. Tognolini, S. Russo, D. Pala, C. Giorgio, I. Hassan-Mo-
hamed, R. Noberini, E. B. Pasquale, P. Vicini, S. Piersanti, S. Rivara, E. Bar-
ocelli, M. Mor, A. Lodola, J. Med. Chem. 2013, 56, 2936–2947.
[16] C. G. Wermuth, J. Med. Chem. 2004, 47, 1303–1314.
[17] D. J. Parks, S. G. Blanchard, R. K. Bledsoe, G. Chandra, T. G. Consler, S. A.
Kliewer, J. B. Stimmel, T. M. Willson, A. M. Zavacki, D. D. Moore, J. M. Leh-
mann, Science 1999, 284, 1365–1368.
ELISA and Ki/IC50 determination: Enzyme-linked immunosorbent
assays were performed as previously described.[10] Briefly, 96-well
ELISA high-binding plates (Corning Costar, #2592) were incubated
overnight at 48C with 100 mL/well of 1 mgmLꢀ1 EphA2-Fc (R&D Sys-
tems, #639-A2) diluted in sterile phosphate buffered saline (PBS,
0.2 gLꢀ1 KCl, 8.0 gLꢀ1 NaCl, 0.2 KH2PO4, 1.15 gLꢀ1 Na2HPO4, pH 7.4).
The day after, wells were washed with washing buffer (PBS+
0.05% Tween20, pH 7.5) and blocked with blocking solution (PBS+
0.5% BSA) for 1 h at 378C. Compounds were added to the wells at
proper concentration in 1% DMSO and incubated at 378C for 1 h.
Biotinylated ephrin-A1-Fc (R&D Systems, #BT602) was added at
378C for 4 h at 200 ngmLꢀ1. The wells were washed and incubated
with 100 mL/well streptavidin–horseradish peroxidase (Sigma–
Aldrich, #S5512) in blocking solution (0.05 mgmLꢀ1 in PBS+0.5%
BSA, pH 7.4) for 20 min then washed again and incubated with
0.1 mgmLꢀ1 3,3’,5,5’-tetramethylbenzidine (Sigma–Aldrich, #T2885)
reconstituted in stable peroxide buffer (11.3 gLꢀ1 citric acid,
9.7 gLꢀ1 Na2HPO4, pH 5.0) and 0.02% H2O2 (30 wt% in H2O), added
immediately before use. The reaction was stopped with 3n HCl
100 mL/well and the absorbance was measured using an ELISA
plate reader at 450 nm (Sunrise, Tecan, Switzerland).
[18] Y. Kawamata, R. Fujii, M. Hosoya, M. Harada, H. Yoshida, M. Miwa, S. Fu-
kusumi, Y. Habata, T. Itoh, Y. Shintani, S. Hinuma, Y. Fujisawa, M. Fujino,
J. Biol. Chem. 2003, 278, 9435–9440.
Phosphorylation of EphA2 in PC3 cells: PC3 prostate adenocarci-
noma cells, grown in Ham F12 supplemented with 7% fetal bovine
serum (FBS) and 1% antibiotic solution (penicillin/streptomycin),
[19] J. P. Himanen, Y. Goldgur, H. Miao, E. Myshkin, H. Guo, M. Buck, M.
Nguyen, K. R. Rajashankar, B. Wang, D. B. Nikolov, EMBO Rep. 2009, 10,
722–728.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 67 – 72 71

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
[20] Phase, version 3.3; Schrçdinger, LLC, New York, 2011.
[21] G. M. Sastry, S. L. Dixon, W. Sherman, J. Chem. Inf. Model. 2011, 51,
2455–2466.
[22] Glide, version 5.7, Schrçdinger, LLC, New York, NY, 2011.
[23] R. A. Friesner, R. B. Murphy, M. P. Repasky, L. L. Frye, J. R. Greenwood,
T. A. Halgren, P. C. Sanschagrin, D. T. Mainz, J. Med. Chem. 2006, 49,
6177–6196.
[24] P. R. Maloney, D. J. Parks, C. D. Haffner, A. M. Fivush, G. Chandra, K. D.
Plunket, K. L. Creech, L. B. Moore, J. G. Wilson, M. C. Lewis, S. A. Jones,
T. M. Willson, J. Med. Chem. 2000, 43, 2971–2974.
[25] R. Pellicciari, S. Fiorucci, E. Camaioni, C. Clerici, G. Costantino, P. R. Malo-
ney, A. Morelli, D. J. Parks, T. M. Willson, J. Med. Chem. 2002, 45, 3569–
3572.
[29] I. H. Mohamed, C. Giorgio, R. Bruni, L. Flammini, E. Barocelli, D. Rossi, G.
Domenichini, F. Poli, M. Tognolini, Pharmacol. Res. 2011, 64, 464–470.
[30] M. Tognolini, C. Giorgio, I. H. Mohamed, E. Barocelli, L. Calani, E. Rey-
naud, O. Dangles, G. Borges, A. Crozier, F. Brighenti, D. Del Rio, J. Agric.
Food Chem. 2012, 60, 8877–8884.
[31] S. Duggineni, S. Mitra, S. R. Noberini, X. Han, X. N. Lin, Y. Xu, W. Tian, J.
An, E. B. Pasquale, Z. Huang, Biochem. Pharmacol. 2013, 85, 507–513.
[32] B. Wu, Z. Zhang, R. Noberini, E. Barile, M. Giulianotti, C. Pinilla, R. A.
Houghten, E. B. Pasquale, M. Pellecchia, Chem. Biol. 2013, 20, 19–33.
[33] S. Wang, W. J. Placzek, JL. Stebbins, S. Mitra, R. Noberini, R. M. Koolpe,
Z. Zhang, R. Dahl, E. B. Pasquale, M. Pellecchia, J. Med. Chem. 2012, 55,
2427–2436.
[34] Maestro, version 9.2, Schrçdinger, LLC, New York, NY, 2011.
[35] W. L. Jorgensen, D. S. Maxwell, J. Tirado-Rives, J. Am. Chem. Soc. 1996,
118, 11225–11236.
[26] R. Noberini, E. Rubio de La Torre, E. B. Pasquale, Cell Adhes. Migr. 2012,
6, 102–112.
[36] G. E. Kellogg, M. Fornabaio, F. Spyrakis, A. Lodola, P. Cozzini, A. Mozzar-
elli, D. J. Abraham, J. Mol. Graph Model. 2004, 22, 479–486.
[37] J. P. Himanen, Semin. Cell Dev. Biol. 2012, 23, 35–42.
[27] C. M. Labbꢃ, G. Laconde, M.A Kuenemann, B. O. Villoutreix, O, Speran-
dio, Drug Discovery Today 2013, in press; DOI:10.1016/
j.drudis.2013.05.003.
[28] R. Noberini, S. K. De, Z. Zhang, B. Wu, D. Raveendra-Panickar, V. Chen, J.
Vazquez, H. Qin, J. Song, N. D. P. Cosford, M. Pellecchia, E. B. Pasquale,
Chem. Biol. Drug Des. 2011, 78, 667–678.
Received: July 10, 2013
Published online on October 9, 2013
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 67 – 72 72