Combined 3D-QSAR, Molecular Docking and Molecular Dynamics Study on Derivatives of Peptide Epoxyketone and Tyropeptin-Boronic Acid as Inhibitors Against the β5 Subunit of Human 20S Proteasome.

Abstract

An abnormal ubiquitin-proteasome is found in many human diseases, especially in cancer, and has received extensive attention as a promising therapeutic target in recent years. In this work, several in silico models have been built with two classes of proteasome inhibitors (PIs) by using 3D-QSAR, homology modeling, molecular docking and molecular dynamics (MD) simulations. The study resulted in two types of satisfactory 3D-QSAR models, i.e., the CoMFA model (Q(2) = 0.462, R(2) (pred) = 0.820) for epoxyketone inhibitors (EPK) and the CoMSIA model (Q(2) = 0.622, R(2) (pred) = 0.821) for tyropeptin-boronic acid derivatives (TBA). From the contour maps, some key structural factors responsible for the activity of these two series of PIs are revealed. For EPK inhibitors, the N-cap part should have higher electropositivity; a large substituent such as a benzene ring is favored at the C6-position. In terms of TBA inhibitors, hydrophobic substituents with a larger size anisole group are preferential at the C8-position; higher electropositive substituents like a naphthalene group at the C3-position can enhance the activity of the drug by providing hydrogen bond interaction with the protein target. Molecular docking disclosed that residues Thr60, Thr80, Gly106 and Ser189 play a pivotal role in maintaining the drug-target interactions, which are consistent with the contour maps. MD simulations further indicated that the binding modes of each conformation derived from docking is stable and in accord with the corresponding structure extracted from MD simulation overall. These results can offer useful theoretical references for designing more potent PIs.

Full text

Page 1

Int. J. Mol. Sci. 2011, 12, 1807-1835; doi:10.3390/ijms12031807

International Journal of
Molecular Sciences
ISSN 1422-0067
www.mdpi.com/journal/ijms
Article
Combined 3D-QSAR, Molecular Docking and Molecular
Dynamics Study on Derivatives of Peptide Epoxyketone and
Tyropeptin-Boronic Acid as Inhibitors Against the β5 Subunit of
Human 20S Proteasome
Jianling Liu 1, Hong Zhang 1, Zhengtao Xiao 2, Fangfang Wang 1, Xia Wang 3 and
Yonghua Wang 2,*
1 College of Life Sciences, Northwest University, Xi’an, Shaanxi 710069, China
2 Center of Bioinformatics, Northwest A&F University, Yangling, Shaanxi 712100, China;
E-Mail: dcpwyh@163.com
3 School of Chemical Engineering, Dalian University of Technology, Dalian 116012, China
* Author to whom correspondence should be addressed; E-Mail: yh_wang@nwsuaf.edu.cn;
Tel.: +86-29-87092262; Fax: +86-29-87092262.
Received: 24 January 2011; in revised form: 14 February 2011 / Accepted: 28 February 2011 /
Published: 9 March 2011

Abstract: An abnormal ubiquitin-proteasome is found in many human diseases, especially
in cancer, and has received extensive attention as a promising therapeutic target in recent
years. In this work, several in silico models have been built with two classes of proteasome
inhibitors (PIs) by using 3D-QSAR, homology modeling, molecular docking and molecular
dynamics (MD) simulations. The study resulted in two types of satisfactory 3D-QSAR
models, i.e., the CoMFA model (Q2 = 0.462, R2pred = 0.820) for epoxyketone inhibitors
(EPK) and the CoMSIA model (Q2 = 0.622, R2pred = 0.821) for tyropeptin-boronic acid
derivatives (TBA). From the contour maps, some key structural factors responsible for the
activity of these two series of PIs are revealed. For EPK inhibitors, the N-cap part should
have higher electropositivity; a large substituent such as a benzene ring is favored at the
C6-position. In terms of TBA inhibitors, hydrophobic substituents with a larger size anisole
group are preferential at the C8-position; higher electropositive substituents like a
naphthalene group at the C3-position can enhance the activity of the drug by providing
hydrogen bond interaction with the protein target. Molecular docking disclosed that residues
Thr60, Thr80, Gly106 and Ser189 play a pivotal role in maintaining the drug-target
OPEN ACCESS

Page 2

Int. J. Mol. Sci. 2011, 12

1808
interactions, which are consistent with the contour maps. MD simulations further indicated
that the binding modes of each conformation derived from docking is stable and in accord
with the corresponding structure extracted from MD simulation overall. These results can
offer useful theoretical references for designing more potent PIs.
Keywords: ubiquitin-proteasome; 3D-QSAR; CoMFA; CoMSIA; homology modeling;
molecular docking; molecular dynamics

1. Introduction
Ubiquitin-proteasome system (UPS), as a major factor in regulated intracellular proteolysis in
eukaryotic cells, is essential to maintain intracellular protein homeostasis and control important
signaling pathways [1,2]. In the proteolytic process, the destined degradation protein is first labeled with
a poly-ubiqutin chain through a cascade of enzymes (E1, E2, E3), then the ubiquitylated protein is
recognized and degraded by the 26S proteasome, a multicatalytic multisubunit protease complex [3].
Two types of 26S proteasome have been identified, i.e., constitutive proteasome and immunoproteasome
[4].
The 26S proteasome consists of a 20S core particle (20S CP) and one or two 19S regulatory particles
(19S RP) capped at either (or both) ends of the core [5]. X-ray crystallography studies [6–9] have
demonstrated that 19S RP is built of a ring shaped base and a lid-structure, that regulates the entrance of
substrate to the attached 20S proteasome [10], the 20S CP is a conserved hollow cylinder-shaped
structure which is composed of four homologous rings, arranged in the sequence α7β7β7α7. Each β ring
contains three distinct catalytic activities with three different subunits, namely chymotrypsin-like (β5),
trypsin-like (β2) and the post-glutamyl peptide hydrolyzing, or caspase-like (β1) [11–13]. All three
peptidases exhibit the same catalytic mechanism, in which the N-terminal threonine residue is the active
nucleophile [14,15]. Among them, the importance of individual subunit activities for proteasomal
function is as follows: β5 >> β2 ≥ β1 [16].
Owing to the significance of the proteasome, it is not surprising that aberrations and deregulations
of the proteasome contribute to the pathogenesis of several human diseases. Thus, proteasome
inhibitors (PIs) have become an attractive agent for human diseases therapy, especially for cancer
[17–22]. Their anti-inflammatory and anti-cancer effects are particularly achieved through inhibiting
activation of the transcription factor NFκB and promotion of apoptosis in rapidly dividing cells [23–25].
The most important class of PIs is the peptidic inhibitors, such as aldehydes [1], boronates [26], etc.
Because of the highly reactive functional aldehyde group, aldehydes lack specificity; bortezomib, a
famous PI of the boronate inhibitor family, has been approved for treatment of multiple myeloma and
mantle cell lymphomas by the FDA [27]. However, some side effects of bortezomib have also been
reported [28]. Thus, it is urgent to develop more potent, selective and clinic-friendly PIs. Epoxyketones
(α, β-ketoepoxides) (EPK) are the second-generation PIs found in the late 90s [29,30]. Unlike other PIs,
epoxyketones show potent selectivity to the proteasome, without inhibitory effects on other proteases
such as calpain, trypsin, papain, chymotrypsin and cathepsins [31], due to the fact that epoxyketone
moiety can form a morpholino adduct with the active site amino terminal Thr of the β5 subunit [32].

Page 3

Int. J. Mol. Sci. 2011, 12

1809
Epoxyketone inhibits primarily the CT-L activity of the proteasome [33], but also the T-L and C-L
activities with slower rates [31]. Another attractive class of PIs is boronic acid derivatives of tyropeptin
(TBA). Tyropeptins A, produced by Kitasatospora sp. MK993-dF2, is novel proteasome inhibitor [34].
Takumi Watanabe and coworkers also synthesized a set of TBA derivatives, exhibiting potent inhibitory
against the CT-L activity of human proteasome [35].
Quantitative structure activity relationship (QSAR), which quantitatively correlates the variations in
biological activity with the properties or molecular structures, is one of the most effective approaches for
designing new chemical identities and understanding the action mechanisms of drugs [36–38]. In recent
years, great attention has been paid to discovery and synthesis of novel PIs, studies regarding QSAR of
existing PIs is still relatively insufficient although some 3D-QSAR models of PIs have been reported
[39,40]. The authors offered useful information about the binding mode between the inhibitors and the
proteasome through ligand-based model. However, detailed insights into the active site are still unclear,
since the X-ray crystallographic structure of the human proteasome has not been reported to date. Thus,
in order to reveal the structural features of inhibitors of the β5 subunit of human proteasome, a set of in
silico methods including 3D-QSAR, homology modeling, molecular docking and molecular dynamics
simulations have been conducted on EPK and TBA in the present work. As far as we know, this study
presents the first 3D-QSAR study for these two kinds of PIs, which will provide detailed information for
understanding these two series of compounds and aid screening and design of novel inhibitors.
2. Materials and Methods
2.1. Data Sets
All potent inhibitors of β5 subunit of the human proteasome used in the present study are collected
from recent literatures [35,41]. Discarding compounds with undefined inhibitory activity or unspecified
stereochemistry, 45 compounds of EPK and 41 compounds of TBA are employed in this work. Each
group of compounds is divided into a training set for generating the 3D-QSAR models and a testing set
for evaluating the 3D-QSAR models at a ratio of 4:1. The compounds in the test set have a range of
biological activity values similar to that of the training set. Their IC50 values are converted into pIC50
(i.e., −logIC50) values and used as dependent variables in the CoMFA and CoMSIA calculation. The
pIC50 values of the EPK and TBA compounds cover an interval of three and four log units, respectively.
The structures of these two groups and their IC50 and pIC50 values are given in Appendix.
2.2. Molecular Modeling and Alignment
The 3D-QSAR study and molecular docking is performed using SYBYL 6.9 (Tripos, Inc). The 3D
structures of all compounds are constructed using the Sketch Molecule function with Sybyl software.
The geometry optimization of all compounds is carried out by using the TRIPOS force field with the
Gasteiger-Huckel (GH) charges, and repeated minimization is performed using Powell conjugated
gradient algorithm method with convergence criterion of 0.05 kcal/mol Å. Of note, molecules containing
boron atoms are not supported by SYBYL because it does not provide the force field parameters for the
boron atom in default settings. For this reason, we changed the boron atom in TBA derivatives to a
carbon atom ‘C.3’ in SYBYL according to the strategy adopted in the literatures [42] and colored it with

Page 4

Int. J. Mol. Sci. 2011, 12

1810
‘magenta’ for distinction.
Structure alignment is considered as one of the most critical step in CoMFA and CoMSIA analysis, so
the template molecule selection for alignment and the alignment methods are crucial to the CoMFA and
CoMSIA models. The most active compound of each group is used as a template for superimposition,
which is assumed to represent the most bioactive conformation. The common fragments of EPK inhibitors
and TBA inhibitors shown in the upper left corner of Figure 1 and Figure 2, respectively, is selected for
database alignment rules in SYBYL command. Other molecules in the data set are superimposed on it
(shown in Figure 1 for EPK and Figure 2 for TBA compounds).
Figure 1. Superimposition of EPK compounds in the training and test sets with common
substructure shown in the upper left corner.

Figure 2. Superimposition of TBA compounds in the training and test sets with common
substructure shown in the upper left corner.

2.3. Calculation and Selection of Molecular Descriptors
Molecular descriptors are quantitative representations of the structural and physicochemical feathers
of molecules, which have been extensively used in the SAR studies [43–46]. In the present work, the
DRAGON (http://www.talete.mi.it/index.htm) is employed to compute the molecular descriptors based
on the MOL2 format of all minimized molecules. The program contains scripts for generating 1664

Page 5

Int. J. Mol. Sci. 2011, 12

1811
parameters of different types including: Constitutional, Topological, Geometrical, Charge, GETAWAY
(Geometry, Topology and Atoms-Weighted AssemblY), WHIM (Weighted Holistic Invariant Molecular
descriptors), 3D-MoRSE (3D-Molecular Representation of Structure based on Electron diffraction),
Molecular Walk Counts, BCUT descriptors, RDF (Radial Distribution Functions), 2D Autocorrelations,
Aromaticity Indices, Randic Molecular Profiles, Functional Groups, Atom-Centered Fragments,
Empirical and Properties [47]. The stepwise linear regression method as the variable selection in R
software (www.r-project.org) is used to obtain the best relevant and meaningful descriptors. The
obtained descriptors are further put into the partial least squares analysis for building more reasonable
QSAR models.
2.4. 3D-QSAR Studies
CoMFA and CoMSIA analyses are performed to construct good predictive QSAR models and to
analyze the effect of each field on the activities of PIs. Models of steric and electrostatic fields for
CoMFA are based on both Lennard-Jones and Coulombic potentials [48]. A 3D grid box of 2.0 Å is
created around the aligned molecules. Steric and electrostatic energies are calculated using a sp3 carbon
atom probe with a charge of +1.00 and van der Waals radius of 1.52 Å at each intersection lattice point of
the grid. The CoMFA cutoff values of steric and electrostatic fields are set to 50 kcal/mol for EPK and 30
kcal/mol for TBA, respectively. Other parameters are set as default. The CoMFA fields are automatically
scaled by the CoMFA-STD method in SYBYL.
The CoMSIA method defines explicit hydrophobic, hydrogen bond (H-bond) donor and acceptor
descriptors in addition to the steric and electrostatic used in CoMFA. CoMSIA fields calculations are
performed by constructing the same lattice boxes as those used in CoMFA calculations. A sp3 carbon
probe atom is used to calculate each field with a charge of +1.00, a radius of 1.00 Å, hydrophobicity
+1.00, and H-bond donor and acceptor property +1. The attenuation factor is set to a default value of 0.3
for these two classes of PIs. CoMSIA similarity indices (AF) for molecule j with atom i at grid point q
are calculated by the following formula (1):
??,?
 ???? ? ??ω?????,?ω????????
?
(1)
where k represents the steric, electrostatic, hydrophobic, or hydrogen-bond donor or acceptor descriptor.
A Gaussian type distance dependence is used between the grid point q and each atom i of the molecule.
The partial least squares (PLS) analysis is used to derive the 3D-QSAR models by constructing a
linear correlation between the CoMFA/CoMSIA (independent variables) and the activity values
(dependent variables). To select the best model, the cross-validation (CV) analysis is performed using
the leave-one-out (LOO) method in which one compound is removed from the data set and its activity is
predicted using the model built from rest of the data set [49]. The sample distance PLS (SAMPLS)
algorithm is used for the LOOCV. The optimum number of components used in the final analysis is
identified by the cross-validation method. The Cross-validated coefficient Q2, which as statistical index
of predictive power, is subsequently obtained. To evaluate the real predictive abilities of the CoMFA and
CoMSIA models derived by the training set, biological activities of an external test set is predicted. The
predictive ability of the model is expressed by the predictive correlation coefficient R2pred, which is
calculated by the following formula (2):