Journal of Molecular Graphics and Modelling 70 (2016) 4044

Contents lists available at ScienceDirect

Journal of Molecular Graphics and Modelling

journal homepage: www.elsevier.com/locate/JMGM

Molecular design and validation of halogen bonding orthogonal to

hydrogen bonding in breast cancer MDM2peptide complex
Anzhong Huang 1 , Liang Zhou 1 , Dawei Zhang, Junliang Yao, Yong Zhang
Department of General Surgery, Jinshan Hospital, Fudan University, Shanghai 201508, China

a r t i c l e i n f o a b s t r a c t

Article history: Peptide therapeutics has been raised as an attractive approach for the treatment of breast cancer by target-
Received 11 February 2016 ing the oncogenic protein MDM2 that inactivates p53 tumor suppressor. Here, we performed molecular
Received in revised form 6 September 2016 design of halogen bonding orthogonal to hydrogen bonding at the complex interface of MDM2 protein
Accepted 12 September 2016
with its cognate peptide ligand to improve the peptide binding afnity and specicity. Crystal structure
Available online 13 September 2016
analysis, high-level quantum chemistry (QC) calculations and combined quantum mechanics/molecular
mechanics (QM/MM) modeling revealed that halogen substitution at position 3 of the benzene moiety
Keywords:
of peptide Phe3 residue can constitute a putative halogen bonding, which is shown to be geometrically
Halogen bonding
Hydrogen bonding
perpendicular to and energetically independent of a native hydrogen bonding that share a common car-
MDM2 bonyl oxygen acceptor. The designed halogen bonding was then validated by surface plasmon resonance
Peptide (SPR) assays, that is, substitution with bromine at position 3 can considerably improve peptide afn-
Breast cancer ity by 4-fold, but the peptide binding does not change substantially upon the bromine substitution at
other positions of the Phe3 benzene moiety (the negative controls that are theoretically unable to form
the halogen bonding), indicating that the orthogonal molecular interaction (OMI) system between the
designed halogen bonding and native hydrogen bonding can co-work well at the complex interface of
MDM2 protein with its halogenated peptide ligands.
2016 Elsevier Inc. All rights reserved.

1. Introduction cer by disrupting MDM2p53 interaction [9,10]. Peptide-like drugs

The MDM2 is an oncoprotein that exerts its oncogenic activity
predominantly by inhibiting p53, a tumor suppressor that responds
to cellular stresses such as DNA damage and oncogenic activation
and is directly or indirectly involved in the development, pro-
gression and metastasis of almost all human cancers [1]. Serve as
a major regulator, MDM2 is induced by p53 and acts as a feed-
back inhibitory protein [2]. Over the past decades, the interaction
between peptides derived from the p53 and the MDM2 protein has
been extensively studied [35], and chemical agents have been pro-
posed to abolish this interaction. Over the past decades, a number of
small-molecule inhibitors have been successfully developed to tar-
get MDM2 protein for cancer therapy [6,7]; idasanutlin (RG7388),
an oral potent MDM2 antagonist, for example, is already in a clinical
advanced phase for acute myeloid leukemia [8].
Instead of chemical agents, peptide therapeutics has in recent
years attracted increased interest for the treatment of breast can-
Corresponding author.
E-mail address: zl9957@sina.com (Y. Zhang).
1
These two authors contributed equally to this work.
exhibit high specicity, low toxicity and satisfactory biocompat-
ibility, which are likely to be more suitable candidates to act as
competitive inhibitors of proteinprotein interactions, consider-
ing their similar binding mode [11]. In order to improve peptide
afnity and selectivity, several chemical methods such as hydro-
carbon stapling [12] and cyclization [13] have been utilized to
modify the natural peptide ligands of MDM2 protein. In the present
study, we utilized a new strategy for promoting peptide binding
to MDM2 by molecular design of an orthogonal molecular inter-
action (OMI) between halogen bonding and hydrogen bonding at
the complex interface of MDM2 with a p53 peptide [14]. According
to a previous report of Voth et al. [15], halogen bonding is shown
to be geometrically perpendicular to and energetically indepen-
dent of hydrogen bonding that share a common carbonyl oxygen
acceptor. This orthogonal relationship is accommodated by the
in-plane and out-of-plane electronegative potentials of the oxy-
gen, which are differentially populated by halogen bonding and
hydrogen bonding. The halogen bonding has been termed as -hole
that is geometrically and energetically similar to hydrogen bonding
[1618]. Here, an integration of high-level quantum chemistry (QC)
calculation, combined quantum mechanics/molecular mechanics
(QM/MM) analysis and surface plasmon resonance (SPR) assay was

http://dx.doi.org/10.1016/j.jmgm.2016.09.007
1093-3263/ 2016 Elsevier Inc. All rights reserved.
 A. Huang et al. / Journal of Molecular Graphics and Modelling 70 (2016) 4044
employed to design OMI in the context of MDM2peptide complex
crystal structure and to characterize the structural basis and ener-
getic landscape of the designed OMI in small model system and at
biological complex interface.
2. Materials, methods, and experiments
2.1. QC analysis of small model systems
The small model systems containing OMIs were analyzed at
Mller-Plesset second-order perturbation theory in conjunction
with the Dunnings augmented correlation consistent basis set
(MP2/aug-cc-pVDZ). The MP2/aug-cc-pVDZ was thought to be ade-
quate for reasonably weak nonbonded interactions and was used in
previous theoretical studies of biological adducts involving halogen
bonding and hydrogen bonding [19]. Considering that the Dun-
nings basis set series does not cover iodine, the Lanl2DZ + (df) basis
set, which is augmented by a set of d and f polarization functions
(exponents 0.292 and 0.441, respectively) and s and p diffuse func-
tions (exponents 0.0569 and 0.0330, respectively) from Lanl2DZ
basis set [20], was applied for iodine-containing systems. The inter-
action energy of halogen bonding (Exb ) or hydrogen bonding
(Ehb ) was estimated as the difference between the total energy
of the bonding adduct and the sum of total energies of the two
monomers, and basis set superposition error (BSSE) was corrected
by means of the counterpoise strategy [21].
2.2. QM/MM calculation of biological complex systems
A two-layer ONIOM-based QM/MM approach [22] was
employed to investigate MDM2peptide complexes involving
OMIs. The high-level QM theory of B3LYP/6-31G(d) was applied to
the inner layer that contains the residues that form halogen bond-
ing and hydrogen bonding in the OMIs, while the low-level MM
method of UFF force eld was used to describe the outer layer that
covers rest of the complexes. Hydrogen was used as link atoms
to saturate the dangling bonds at partitioning interface and the
electrostatic potential between inner and outer layers was treated
with mechanical embedding scheme to reduce computational cost.
The complex structures were fully minimized with the two-layer
QM/MM scheme without constraints [23]. The binding free energy
(G) between MDM2 and peptide was calculated via supramolec-
ular approach proposed by Zhou et al. [24] based on the minimized
structures (Fig. 1).
2.3. SPR assays
The natural peptide and its halogenated versions were syn-
thesized by Fmoc-solid phase synthesis protocol; the Fmoc
halogenated amino acids used in the synthesis were obtained from
unnatural amino acids library. The N-terminal domain of human
MDM2 (residues 23111) was expressed in Escherichia coli BL21-
Gold (DE3) with a pDEST-His-MBP vector and then puried by
afnity chromatography in nickel-nitrilotriacetic acid Superow.
The MDM2peptide interaction was measured by SPR using a pro-
tocol modied from previous reports [3,4]. Briey, peptides were
immobilized on the surface of sensor chip at room tempature
in a BIAcore. A total of 100 nM MDM2 was incubated in 10 mM
HEPES buffer (pH 7.4) containing 150 mM NaCl and 0.005% sur-
factant P20 with varying concentrations of tested peptide on the
chip. Nonlinear regression analysis was performed using Graph-
Pad Prism to derive dissociation constants with the equation
KD = [MDM2][peptide]/[MDM2peptide complex].
41
3. Results and discussion
The high-resolution crystal structure of human MDM2 in com-
plex with a 12-mer pDI peptide was retrieved from the protein data
bank (PDB) database (pdb: 3lnz) [5]. The cocrystallized pDI pep-
tide (LTFEHYWAQLTS) was derived from a phage display library,
which exhibited a moderate inhibitory activity against MDM2p53
interaction (IC50 = 550 nM) [5]. By visually dissecting the crys-
tal structure we identied an intramolecular hydrogen bonding
(length = 2.05 ) between the oxygen atom O of Gly58 residue and
the hydrogen atom H of Met62 residue in MDM2 protein, while
the hydrogen atom H at position 3 of the benzene moiety of pep-
tide Phe3 residue is close to the oxygen atom O of MDM2 Gly58
residue with a spatial distance of 2.75 between them. The small
distance between MDM2 Gly58 O and peptide Phe3 H suggests
that they may form a halogen bonding if the hydrogen atom H of
peptide Phe3 residue is replaced by a halogen atom X (X = F, Cl, Br
or I); the putative halogen bonding seems to be roughly orthogonal
to the intramolecular hydrogen bonding Gly58 Met62 of MDM2,
and they can co-dene an OMI system at the MDM2pDI complex
interface.
In order to examine the hypothesis, we separately modied the
H to F, Cl, Br and I in the complex crystal structure; each was
then subjected to a round of QM/MM minimization with default
settings where the MDM2 residues Gly58 and Met62 as well as
the modied residue Phe3 (X) of peptide were included in the
high-level QM layer, while rest of the complex was treated with
low-level MM approach. The equilibrated OMI systems in QM/MM-
minimized complexes of MDM2 protein with different modied
versions of pDI peptide are shown in Fig. 2, and their geometric
and energetic parameters are listed in Table 1. As can be seen,
halogenations can address considerable effects on the geometric
prole and energetic property of the OMIs at MDM2peptide inter-
face. For putative halogen bonding, the equilibrium distance dO X
between the oxygen atom O of MDM2 Gly58 residue and the
halogen atom X of peptide Phe3 residue ranges between 3.08
and 3.35 , which is in line with the bond length of those bio-
logical halogen bonding revealed previously [25]. In addition, the
distance is smaller than the sum of the van der Waals radii [26]
of respective O and X atoms, indicating that there is an attractive
chemical force between the two atoms. The Gly58 Phe3 interac-
tion energy Exb increases signicantly from 1.02 (H) to 2.18
(Cl), 2.64 (Br) and 2.31 (I) kcal/mol upon the halogena-
tions, although the F substitution (Exb = 1.37 kcal/mol) does
not change the energy substantially as compared to the native H
(Exb = 1.02 kcal/mol) this is expected because uorine cannot
form halogen bonding in most cases [27]. Overall, it is suggested
that halogen bonding is indeed presented at the MDM2peptide
complex interface when the hydrogen atom H replaced by halo-
gen atom X. Next, we examined the geometric and energetic effect
of halogenations on the hydrogen bonding between the Gly58 oxy-
gen atom O and Met62 hydrogen atom H of MDM2 protein. The
native length dO H and interaction energy Ehb of the hydrogen
bonding in crystal structure is 2.05 and 2.31 kcal/mol, respec-
tively, which are changed modestly by the halogenations, except
the substitution by iodine atom I, which can moderately inu-
ence the hydrogen bonding. Moreover, the angle  H O X is close
to 90 ( H O X = 71.3 , 82.5 , 84.6 and 71.3 for F, Cl, Br
and I, respectively). In this respect, the putative halogen bonding
is shown to be geometrically perpendicular to and energetically
independent of hydrogen bonding that share a common carbonyl
oxygen acceptor; this is consistent with the OMI dened by Voth
and coworkers [15].
To examine whether the designed OMI system exists at
MDM2peptide interface, the binding afnities of pDI peptide with
Br substitution at different positions of the benzene moiety of
42 A. Huang et al. / Journal of Molecular Graphics and Modelling 70 (2016) 4044

Fig. 1. The crystal structure of human MDM2 in complex with a 12-mer pDI peptide (pdb: 3jzo). There is an intramolecular hydrogen bonding (length = 2.05 ) between
the oxygen atom O of Gly58 residue and the hydrogen atom H of Met62 residue in MDM2 protein, while the hydrogen atom H at position 3 of the benzene moiety of
peptide Phe3 residue is close to the oxygen atom O of MDM2 Gly58 residue with a spatial distance of 2.75 between them.

Fig. 2. The equilibrated OMI systems in QM/MM-minimized complex structures of MDM2 protein with different modied versions of pDI peptide: (a) H, (b) F, (c) Cl,
(d) Br and (e) I.

Table 1
Geometric and energetic parameters of the OMI systems in QM/MM-minimized complex structures of MDM2 protein with different modied versions of pDI peptide.

X Hydrogen bonding (Gly58 Met62) Halogen bonding (Gly58 Phe3) H O X ( )

dO H () Ehb (kcal/mol) dO X () Exb (kcal/mol)

H 2.05 2.31 2.75 (2.72)a 1.02 102.4

F 1.98 2.36 3.08 (2.99) 1.37 73.0
Cl 2.14 2.18 3.20 (3.27) 2.18 82.5
Br 2.17 2.07 3.15 (3.37) 2.64 84.6
I 2.26 1.75 3.35 (3.50) 2.31 71.3
a
the value in parentheses is sum of the van der Waals radii of respective O and X atoms, taken from Ref. [23].

Phe3 residue to recombinant protein of human MDM2 were mea-
sured via SPR approach. According to computational analysis, the
Br substitution at position 3 can constitute a putative halogen
bonding orthogonal to native hydrogen bonding, whereas the sub-
stitution at other positions (2, 2 , 3 and 4) is unable to form the
halogen bonding. Here, we calculated the Gly58 Phe3 interac-
tion energy EGly58 Phe3 and the MDM2peptide binding free
energy GMDM2peptide using QC and QM/MM, respectively, as
well as measured the MDM2peptide binding afnity KD with the
Br systematic substitution at all positions (2, 2 , 3, 3 and 4) of
Phe3 benzene moiety by SPR, and obtained data are tabulated in
Table 2. Evidently, there is a good (positive) correlation between the
EGly58 Phe3 and GMDM2peptide with Pearson correlation coef-
cient rp of 0.84; all these sample points distributed linearly and
only the Br substitution at 3 position exhibits an outlier behav-
ior that deviates slightly from the line (Fig. 3a). In addition, the
QM/MM derived GMDM2peptide seems to have a moderate (neg-
ative) correlation with experimentally measured KD (rp = 0.64)
this is acceptable if considering that some other (minor) factors
such as peptide exibility and entropy penalty were not consid-
ered here that may cause a systematic bias to the computational
binding free energy. The native pDI peptide can bind to MDM2
 A. Huang et al. / Journal of Molecular Graphics and Modelling 70 (2016) 4044
Table 2
The calculated energetic parameters and measured binding afnities of the Br substitution at different positions of the benzene moiety of peptide Phe3 residue.
Br substitution position EGly58 Phe3 (kcal/mol)a
None 1.02
2 0.97
2 1.20
3 2.64
3 1.58
4 1.52
a
the interaction energy between peptide Phe3 residue and MDM2 Gly58 residue calculated by QC.
b
he binding free energy of peptide to MDM2 derived by QM/MM.
c
the dissociation constant determined by SPR.
Fig. 3. Plots of (a) QC-calculated Gly58 Phe3 residue interaction energy (EGly58 Phe3 ) against QM/MM-derived MDM2peptide binding free energy (GMDM2peptide ) and
(b) SPR-measured MDM2peptide binding afnity (pKD ) against QM/MM-derived MDM2peptide binding free energy (GMDM2peptide ) with the Br substitution at different
positions of the benzene moiety of peptide Phe3 residue.
with a moderate afnity (KD = 657 54 nM), which is basically con-
sistent with the inhibitory activity (IC50 = 550 nM) of the peptide
reported by Phan et al. [5]. Given the Br substitution at 3 position
can considerably improve the peptide afnity (KD = 174 12 nM)
by 4-fold as compared to the native peptide, indicating that the
p3-bromine can improve MDM2peptide binding considerably. As
might be expected, the afnity (KD > 350 nM) of peptide with Br
substitution at other positions (negative controls) does not change
substantially relative to the native peptide (KD > 657 54 nM),
demonstrating that only the Br substitution at position 3 can
constitute the designed OMI system at MDM2peptide complex
interface.
Acknowledgements
This work was supported by the Jinshan Hospital Fudan Univer-
sity.
References
[1] M. Wade, Y.C. Li, G.M. Wahl, MDM2, MDM2X and p53 in oncogenesis and
cancer therapy, Nat. Rev. Cancer 13 (2013) 8396.
[2] G.W. Yu, S. Rudiger, D. Veprintsev, S. Freund, M.R. Fernandez-Fernandez, A.R.
Fersht, The central region of HDM2 provides a second binding site for p53,
Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 12271232.
[3] M. Pazgier, M. Liu, G. Zou, W. Yuan, C. Li, C. Li, J. Li, H. Monbo, D. Zella, S.G.
Tarasov, W. Lu, Structural basis for high-afnity peptide inhibition of p53
interactions with MDM2 and MDM2X, Proc. Natl. Acad. Sci. U. S. A. 106 (2009)
46654670.
43
GMDM2peptide (kcal/mol)b KD (nM)c
15.4 657 54
14.2 485 42
16.9 520 48
19.2 174 12
17.6 536 20
18.4 383 46
[4] C. Li, M. Pazgier, C. Li, W. Yuan, M. Liu, G. Wei, W.Y. Lu, W. Lu, Systematic
mutational analysis of peptide inhibition of the p53-MDM2/MDM2X
interactions, J. Mol. Biol. 398 (2010) 200213.
[5] J. Phan, Z. Li, A. Kasprzak, B. Li, S. Sebti, W. Guida, E. Schnbrunn, J. Chen,
Structure-based design of high afnity peptides inhibiting the interaction of
p53 with MDM2 and MDM2X, J. Biol. Chem. 285 (2010) 21742183.
[6] S. Shangary, S. Wang, Small-molecule inhibitors of the MDM2-p53
proteinprotein interaction to reactivate p53 function: a novel approach for
cancer therapy, Ann. Rev. Pharmacol. Toxicol. 49 (2009) 223241.
[7] B. Zhang, B.T. Golding, I.R. Hardcastle, Small-molecule MDM2-p53 inhibitors:
recent advances, Future Med. Chem. 7 (2015) 631645.
[8] Q. Ding, Z. Zhang, J.J. Liu, N. Jiang, J. Zhang, T.M. Ross, X.J. Chu, D. Bartkovitz, F.
Podlaski, C. Janson, C. Tovar, Z.M. Filipovic, B. Higgins, K. Glenn, K. Packman,
L.T. Vassilev, B. Graves, Discovery of RG7388, a potent and selective
p53-MDM2 inhibitor in clinical development, J. Med. Chem. 56 (2013)
59795983.
[9] T.N. Do, R.V. Rosal, L. Drew, A.J. Raffo, J. Michl, M.R. Pincus, F.K. Friedman, D.P.
Petrylak, N. Cassai, J. Szmulewicz, G. Sidhu, R.L. Fine, P.W. Brandt-Rauf,
Preferential induction of necrosis in human breast cancer cells by a p53
peptide derived from the MDM2 binding site, Oncogene 22 (2003)
14311444.
[10] Y. Fang, R. Jin, Y. Gao, J. Gao, J. Wang, Design of p53-derived peptides with
cytotoxicity on breast cancer, Amino Acids 46 (2014) 20152024.
[11] P. Vanhee, A.M. van der Sloot, E. Verschueren, L. Serrano, F. Rousseau, J.
Schymkowitz, Computational design of peptide ligands, Trends Biotechnol. 29
(2011) 231239.
[12] Y.S. Chang, B. Graves, V. Guerlavais, C. Tovar, K. Packman, K.H. To, K.A. Olson,
K. Kesavan, P. Gangurde, Stapled -helical peptide drug development: a
potent dual inhibitor of MDM2 and MDM2X for p53-dependent cancer
therapy, Proc. Natl. Acad. Sci. U. S. A. 110 (2013) E3445E3454.
[13] J.P. Malkinson, M. Zloh, M. Kadom, R. Errington, P.J. Smith, M. Searcey,
Solid-phase synthesis of the cyclic peptide portion of chlorofusin, an inhibitor
of p53-MDM2 interactions, Org. Lett. 5 (2003) 50515054.
[14] Y. Li, X. Yu, Y. Lou, T. Wang, Rational design of MDM2 protein with p53
peptide, Aust. J. Chem. 69 (2016) 735744.
44 A. Huang et al. / Journal of Molecular Graphics and Modelling 70 (2016) 4044
[15] A.R. Voth, P. Khuu, K. Oishi, P.S. Ho, Halogen bonds as orthogonal molecular
interactions to hydrogen bonds, Nat. Chem. 1 (2009) 7479.
[16] K.E. Riley, J.S. Murray, P. Politzer, M.C. Concha, P. Hobza, BrO complexes as
probes of factors affecting halogen bonding: interactions of bromobenzenes
and bromopyrimidines with acetone, J. Chem. Theory Comput. 5 (2009)
155163.
[17] J.S. Murray, K.E. Riley, P. Politer, T. Clark, Directional weak intermolecular
interactions: -hole bonding, Aust. J. Chem. 63 (2010) 15981607.
[18] P. Politzer, J.S. Murray, T. Clark, Halogen bonding and other -hole
interactions: a perspective, Phys. Chem. Chem. Phys. 15 (2013) 1117811189.
[19] L.A. Hardegger, B. Kuhn, B. Spinnler, L. Anselm, R. Ecabert, M. Stihle, B. Gsell, R.
Thoma, J. Diez, J. Benz, J.M. Plancher, G. Hartmann, D.W. Banner, W. Haap, F.
Diederich, Systematic investigation of halogen bonding in protein-ligand
interactions, Angew. Chem. Int. Ed. 50 (2011) 314318.
[20] P. Zhou, Y. Ren, F. Tian, J. Zou, Z. Shang, Halogen-ionic bridges: do they exist in
the biomolecular world, J. Chem. Theory Comput. 6 (2010) 22252241.
[21] S.F. Boys, F. Bernardi, The calculation of small molecular interactions by the
differences of separate total energies. Some procedures with reduced errors,
Mol. Phys. 19 (1970) 553566.
[22] L.W. Chung, W.M. Sameera, R. Ramozzi, A.J. Page, M. Hatanaka, G.P. Petrova,
T.V. Harris, X. Li, Z. Ke, F. Liu, H.B. Li, L. Ding, K. Morokuma, The ONIOM
method and its applications, Chem. Rev. 115 (2015) 56785796.
[23] Y. Lu, T. Shi, Y. Wang, H. Yang, X. Yan, X. Luo, H. Jiang, W. Zhu, Halogen
bonding ??a novel interaction for rational drug design, J. Med. Chem. 52
(2009) 28542862.
[24] P. Zhou, J. Zou, F. Tian, Z. Shang, Fluorine bonding ?how does it work in
protein-ligand interactions, J. Chem. Inf. Model 49 (2009) 23442355.
[25] P. Aufnger, F.A. Hays, E. Westhof, P.S. Ho, Halogen bonds in biological
molecules, Proc. Natl. Acad. Sci. USA 101 (2004) 1678916794.
[26] A. Bondi, Van der waals volumes and radii, J. Phys. Chem. 68 (1964) 441451.
[27] K. Eskandari, M. Lesani, Does uorine participate in halogen bonding,
Chemistry 21 (2015) 47394746.