3D structure generation, virtual screening and docking of humanRas-associated binding (Rab3A) protein involvedin tumourigenesis

Sharad S. Lodhi • Rohit Farmer • Atul Kumar Singh •

Yogesh K. Jaiswal • Gulshan Wadhwa

Received: 7 May 2013 / Accepted: 11 February 2014

� Springer Science+Business Media Dordrecht 2014

Abstract Rab3A is expressed predominantly in brain and

synaptic vesicles. Rab3A is involved specifically in teth-

ering and docking of synaptic vesicles prior to fusion

which is a critical step in regulated release of neurotrans-

mitters. The precise function of Rab3A is still not known.

However, up-regulation of Rab3A has been reported in

malignant neuroendocrine and breast cancer cells. In the

present study, the structure of Rab3A protein was gener-

ated using MODELLER 9v8 software. The modeled pro-

tein structure was validated and subjected to molecular

docking analyses. Docking with GTP was carried out on

the binding site of Rab3A using GOLD software. The

Rab3A-GTP complex has best GOLD fitness value of

77.73. Ligplot shows hydrogen bondings (S16, S17, V18,

G19, K20, T21, S22, S31, T33, A35, S38, T39 and G65)

and hydrophobic interacting residues (F25, F32, P34, F36,

V37, D62 and A64) with the GTP ligands in the binding

site of Rab3A protein. Here, the ligand molecules of NCI

diversity set II from the ZINC database against the active

site of the Rab3A protein were screened. For this purpose,

the incremental construction algorithm of GLIDE and the

genetic algorithm of GOLD were used. Docking results

were analyzed for top ranking compounds using a con-

sensus scoring function of X-Score to calculate the binding

affinity and Ligplot was used to measure protein–ligand

interactions. Five compounds which possess good inhibi-

tory activity and may act as potential high affinity inhibi-

tors against Rab3A active site were identified. The

top ranking molecule (ZINC13152284) has a Glide score of

-6.65 kcal/mol, X-Score of -3.02 kcal/mol and GOLD

score of 64.54 with 03 hydrogen bonds and 09 hydrophobic

contacts. This compound is thus a good starting point for

further development of strong inhibitors.

Keywords Rab3A � GTP � GOLD Docking �Tumourigenesis

Introduction

The Rab (Ras-associated binding) GTPases comprise the

largest subfamily of the Ras superfamily and function as

regulators of various steps in vesicle trafficking pathways. In

human genome, more than 60 Rab genes are known, out of

which Rab3 is found to be involved in regulation of exocy-

tosis of neurotransmitters and hormones [1–4]. The Rab3

subfamily includes four isoforms: Rab3A, B, C and D, which

are differentially expressed in neuronal and endocrine tissues

[5, 6]. The most abundant expression of Rab3A can be

observed in brain where it is present in almost all synapses,

while Rab3B and Rab3C are only present in a subset of

Electronic supplementary material The online version of thisarticle (doi:10.1007/s11033-014-3263-x) contains supplementarymaterial, which is available to authorized users.

S. S. Lodhi � Y. K. Jaiswal

School of Studies in Biochemistry, Jiwaji University,

Gwalior 474011, India

S. S. Lodhi � G. Wadhwa (&)

Apex Bioinformatics Centre (BTISNET), Department of

Biotechnology, Ministry of Science and Technology, 7th Floor,

Block-2, CGO Complex, New Delhi 110003, India

e-mail: gulshan.dbt@nic.in

R. Farmer

Department of Computational Biology and Bioinformatics, Sam

Higginbottom Institute of Agriculture, Technology and Sciences,

Allahabad 211007, UP, India

A. K. Singh

Centre for Research in Nanotechnology & Science, Indian

Institute of Technology-Bombay, Mumbai 400076, India

123

Mol Biol Rep

DOI 10.1007/s11033-014-3263-x

synapses [7]. Like other Rab-GTPases, Rab3 interacts with

membranes via C-terminal geranylgeranyl moieties and

cycles between a synaptic vesicle-associated GTP-bound

form and a cytosolic GDP-bound form [8, 9]. This cycling is

regulated by three types of regulators: Rab GDP dissociation

inhibitor (GDI), Rab3 GDP/GTP exchange protein, and

Rab3 GAP [10]. Mutations in the genes encoding regulatory

and catalytic subunits of the Rab3 GAP lead to Warburg

Micro and Martsolf syndromes characterized by develop-

mental abnormalities of the eye, nervous system and geni-

talia [11, 12]. Because the cyclical activation is coupled with

membrane association and allows both spatial and temporal

control of Rab3A activity, these regulators are thought to be

important for the proper functioning of Rab3A in synaptic

vesicle transport. The diverse actions of Rab proteins are

mediated via their multiple effectors, which usually interact

with the GTP bound form of Rabs [9]. The known potential

effectors of Rab3A are Rabin3, Rabphilin, RIM1a, RIM2a,

Granuphilin, Noc2, PRA1, Munc18-1, INPP5B, SNAP-29,

Synapsin and Calmodulin [8, 13–16]. However, further

research is needed to determine physiological importance of

these effectors and their mode of coordination to mediate the

diverse actions of Rab3A.

Despite a lot of research being carried out, the involve-

ment of Rab3a in various pathways has been identified but its

precise function has not yet been known. Many studies show

the variation of Rab3A function from species to species.

Rab3A appears to be involved in regulation of targeting and

docking of synaptic vesicle at the active zone. Loss-of

function mutations of Rab3A gene in C. elegans lead to a

significant depletion of synaptic vesicles at presynaptic ter-

minals and a concomitant elevation of vesicles along the

axons [17]. Whereas Rab3A deletion in mice does not appear

to affect synaptic vesicle distribution at the resting state, it

abolishes the activity-dependent recruitment of synaptic

vesicles to the active zone and impairs the replenishment of

docked vesicles after exhaustive stimulation [18]. Rab3A

also seems to regulate the post-docking step in synaptic

vesicle fusion. Electrophysiological studies on Rab3A null

mutant mice show that while the size of the readily releasable

pool of vesicles is unaltered, Ca2?-triggered synaptic vesicle

exocytosis and paired pulse facilitation is increased [19] and

mossy fiber long-term potentiation is abolished [20]. Support

for a role of Rab3A at a late step during Ca2?-triggered

exocytosis is also provided by studies on aplysia neurons

injected with a GTPase-deficient form of Rab3A [21].

Rab3A might coordinate the regulation of coupling between

synaptic vesicle exocytosis and endocytosis through its

putative effector Rabphilin [13]. A recent study also

explored the function of Rab3A at ribbon synapses in the

retina of the tiger salamander (Ambystoma tigrinum), where

the Rab3A mutant blocks synaptic release in an activity-

dependent manner, with more frequent stimulation leading

to more rapid block. The frequency dependence of block by

exogenous Rab3A suggests that it acts competitively with

synaptic vesicles to interfere with their resupply to release

sites. This finding suggests a crucial role of Rab3A in

delivering vesicles to Ca2?-dependent release sites at ribbon

synapses as well [22]. In humans, over expression of Rab3A

has been reported in malignant neuroendocrine cells [23] and

breast cancer cells, which suggests its potential role in the

process of tumourigenesis.

In the present study we aim to explore the Rab3A pro-

tein through molecular modeling and virtual screening of

potential inhibitors. The modeled structure of Rab3A pro-

duced in this study identified new interactions important in

the regulation of GTPase activity.

Materials and methods

3D structure generation

The amino acid sequence of human Rab3A (target Rab3A)

was retrieved from the NCBI sequence database (http://

www.ncbi.nlm.nih.gov) (accession no. AAF67748.1, 220

aa). To identify the templates for homology modeling of

Rab3A, a BLASTP search (http://blast.ncbi.nlm.nih.gov/

Blast.cgi?PAGE0Proteins) was performed against the

Brookhaven Protein Data Bank with the default parameters.

The sequence alignment between target and template were

carried out using the CLUSTAL Wv2 (http://www.ebi.ac.

uk/clustalw) program [24]. The template used was PDB ID

1ZBD which is the crystal structure of the small G protein

Rab3A complexed with the effector domain of rabphilin-3A

from Rattus norvegicu. There was 95 % sequence identity

between template and target with 90 % query coverage. The

academic version of MODELLER9v8 (http://salilab.org/

modeller/), was used for 3D structure generation based on the

information obtained from sequence alignment [25]. Out of

20 models generated by MODELLER, one model central to

cluster was selected and subjected to stereo chemical check

to find the deviations from normal bond length, dihedrals and

non-bonded atom–atom distances. Each model with the

highest G-score of PROCHECK [26], and VERIFY3D [27]

profile was subjected to energy minimization. The energy

minimization was started with side chains and then applied to

main chain of Ca backbone. All calculations were performed

by using ACCELRYS DS modeling 2.5 (Accelrys Inc. San

Diego, CA 92121, USA) software suites. STRIDE [28] was

used for the prediction of secondary structure of the modeled

Rab3A protein. PROSA was used for calculating Z-scores.

The weighted root mean square deviation (RMSD) of the

modeled protein structure was calculated using the combi-

natorial extension algorithm [29]. The modeled structure

was then superimposed on the crystal template without

Mol Biol Rep

123

altering the coordinate systems of atomic position in the

template. The residue profiles of the three-dimensional

models were further checked using VERIFY3D. PRO-

CHECK analysis was performed to assess the stereo-chem-

ical properties of the three-dimensional models and

Ramachandran plots.

Molecular dynamics (MD) simulations

The MD simulations of modeled Rab3A protein were

performed with the GROMACS 4.5.4 software package

[30] using GROMOS 96 force field [31] and the flexible

SPC water model. The initial structure was immersed in a

periodic water box of truncated octahedron shape (0.5 nm

thick). Electrostatic energy was calculated using the par-

ticle mesh Ewald method [32]. Cutoff distance for the

calculation of the Coulomb and Vander Waals interaction

was 1.0. After energy minimization using a steepest des-

cent for 1,000 steps, the system was subjected to equili-

bration at 300 k and normal pressure for 100 ps under the

conditions of position restraints for heavy atoms. LINCS

[33] constraints were performed for all bonds, keeping the

whole protein molecule fixed and allowing only the water

molecule to move to equilibrate with respect to the protein

structure. The system was coupled to the external bath by

the Berendsen pressure and temperature coupling [34]. The

system contained 7 negative charges which were stabilized

by counter 7 NA? ions. The final MD calculations were

performed for 1 ns under the same conditions except that

the position restraints were removed. The results were

analyzed using the standard software provided by the

GROMACS package. An average structure was refined

further using a steepest descent energy minimization.

The modeled structure through various structure vali-

dation and molecular dynamics simulation was subjected

for the prediction of possible active sites using putative

active sites with spheres (PASS) and CastP programs

simultaneously. PASS is a simple computational tool that

uses geometry to characterize regions of buried volume in

proteins and to identify positions likely to represent bind-

ing sites based upon the size, shape, and burial extent of

these volumes [35]. CastP server uses the weighted Dela-

unay triangulation and the alpha complex for shape mea-

surements. It provides identification and measurements of

surface accessible pockets as well as interior inaccessible

cavities, for proteins and other molecules [36].

Virtual screening of NCI diversity set II against Rab3A

protein

The ligand molecules of NCI diversity set II were obtained

for virtual screening in mol2 format from ZINC database,

provided by the Shoichet Laboratory in the Department of

Pharmaceutical Chemistry at the University of California,

San Francisco (UCSF). ZINC database is a central repos-

itory of commercially available compounds for virtual

screening synthesized by several organizations, which

include government organizations such as National Cancer

Institute (NCI) and several private players. It is usually

easier to retrieve data from ZINC in many different formats

and because it gives a unique id to each and every molecule

it is also easier to refer to them later in the analysis.

The NCI diversity set is a small library, ideal for

beginning a screening campaign. NCI diversity set II

consists of a collection of 1,364 synthetic small molecules

selected from the full NCI screening collection. It is a

complete dataset of manageable size with the currently

available computational resources. Moreover, NCI diver-

sity sets are specifically designed for cancer research. This

reduces the efforts for screening large compound libraries,

which may or may not be useful for the purpose.

Therefore, the NCI diversity set II compounds produced

by National Cancer Institute (NCI) and the molecular

database maintained by the ZINC were used in our

analysis.

Protein ligand docking

The ligand library was extracted from the ZINC database

(http://zinc.docking.org/). The shortlisted ligands were

subjected to further predocking preparations where

hydrogens were added followed by minimization and

optimization in OPLS_2005 force field. Finally, 10 con-

formations for each ligand were generated, and ready for

docking. The docking of ligand molecules to Rab3A

structure using GLIDE was performed and cross validation

was done using GOLD. GLIDE uses systematic and sim-

ulation method for searching the poses and ligand flexi-

bility. In a systematic method, it uses incremental

construction for searching, and its output GScore is an

empirical scoring function which is a combination of var-

ious parameters [37] The GScore is calculated in Kcal/mol

as:

G � Score ¼ H bond þ Lipo þ Metalþ Site

þ 0:130 Coulþ 0 : 065 vdW�BuryP�RotB

where H bond = Hydrogen bonds, Lipo = hydrophobic

interactions, Metal = metal-binding term, Site = polar

interactions in the binding site, vdW = Vander-Waals

forces, Coul = columbic forces, Bury P = penalty for

buried polar group, RotB = freezing rotable bonds.

Library of ligands were subjected to glide docking.

Since each ligand has 10 stereoisomers or conformations,

each conformation was first screened through the high

throughput virtual screening module of GLIDE. Finally,

Mol Biol Rep

123

the top five ranked ligands were selected according to

GLIDE and were docked again using GOLD docking

software to obtain consistent and improved results. GOLD

v4.0 [38] is an automated ligand-docking program that uses

a genetic algorithm to explore the full range of ligand

conformational flexibility, namely full acyclic ligand flex-

ibility and partial cyclic ligand flexibility, with partial

flexibility of the protein in the neighbourhood of the pro-

tein active site, and satisfies the fundamental requirement

that the binding of ligand must displace loosely bound

water. In GOLD docking, fifty independent docking runs

were performed for each molecule with default parameters.

Docking result analysis

A molecule was ranked relatively high if it scores well with

these two different methods (or scoring functions). These

methods have different search algorithms and scoring

functions. Hence, it was not possible to compare the fitness

scores of GOLD and GLIDE directly. For comparison and

validation of docking results we used X-Score v1.2.1, [39] a

consensus scoring function. X-Score calculates the negative

logarithm of the dissociation constant of the ligand to the

protein, 2log Kd, as the average of three scoring functions

(HPScore, HMScore and HSScore), and predicts the binding

energy (Kcal/mol) of the ligand. X-Score was reported to

have an accuracy of 62.2 kcal/mol relative to the actual

binding energies. For analysing the interactions of docked

protein–ligand complexes, the Ligplot programme [40] was

used to check the hydrogen bond and hydrophobic interac-

tions between receptor and ligand atoms within a range of

5 A�. Also PyMOL (V-1.3) [41] and Chimera (V-1.4.1) [42]

were used to visualize the interactions and to prepare figures

for top ranked molecules.

Results and discussion

Model building and protein structure validation

The structure of human Rab3A protein was determined by

using homology modeling protocol. BLASTP search was

performed against PDB with default parameters to find suit-

able templates for homology modeling. Based on the maxi-

mum identity with high score and lower e-value 1ZBD chain

A was used as the template for homology modeling. Sequence

alignment between the Rab3A and the template was generated

using CLUSTALW programme. The models generated were

subjected to structure validation and molecular dynamics and

best model was identified for further analysis. The predicted

structure of human Rab3A (Supplementary Fig. 1) shows

typical GTPase fold, comprising six beta strands surrounded

by five alpha helices (Fig. 1).

Following the structure prediction the models generated

were examined for their correctness in terms of stereo

chemical properties and three dimensional residue profile

using PROCHECK and Verfiy 3D respectively. For the

most accurate model it was found that the phi/psi angles of

94.1 % residues fell in the most favored regions (Supple-

mentary Fig. 2). The overall PROCHECK G-factor for the

homology modeled structure was -0.07. These statistics

confirmed good quality of the predicted model. High

quality of model is also confirmed from VERIFY 3D server

as 88.20 % of residues of modeled protein showed a score

higher than 0.2. The structural superimposition of Ca trace

of the target model after MD simulation over template

structure 1ZBD chain A (Fig. 2) resulted in a root mean

square deviation (RMSD) of 0.2 A (Z-score 177) using

DaliLite server. It indicates a valid structure of the model.

Based on these results, it was ascertained that the obtained

structure has reasonably good quality.

Fig. 1 The secondary structure elements in the corresponding region

of the protein sequence indicated

Fig. 2 Superimposition of Ca trace of human Rab3A protein and

1ZBD chain A (template)

Mol Biol Rep

123

Upon verification for physico-chemical properties using

various tools and molecular dynamics simulation the pre-

dicted model showed good enough structural properties for

further analysis. The structure was modeled with a very

high sequence identity with the template that ensures the

correct arrangement of the secondary structures and the

placement of the side chains.

Molecular dynamics simulations

The MD simulation of the modeled Rab3A protein struc-

ture was performed and the resulted trajectory was ana-

lyzed to study the motional properties of the protein. The

time evolution of root mean square deviation (RMSD) was

computed for the modeled structure of the protein by taking

the whole protein as initial structure. It is evident from the

Fig. 3 that RMSD increased slowly up to 300 ps and then

decreases up to 500 ps then again slightly increases up to

600 ps and attained the equilibrium. As the modeled

structure did not undergo any drastic motion beyond

600 psm, the MD simulation was performed only for 1 ns.

Based on intrinsic dynamics, structural stability and

improved relaxation of the modeled structure, the energy

(Fig. 4) of the energy minimized structure was also cal-

culated. The energy and RMSD calculations demonstrated

that the protein is highly conserved in nature i.e. the protein

is not much flexible. Root mean square fluctuation (RMSF)

indicates the flexibility of the protein. Figure 5 under-

standably indicates relatively high fluctuation in few resi-

dues at the N and C-terminal of the protein, these residue

can be noticed in the loop regions as shown in Fig. 1, rest

of the residues are found to be quite stable which are part

of the much conserved secondary structure elements. The

compactness of the model was also confirmed by calcu-

lating radius of gyration (Rg). Rg for the protein backbone

is under continuous decrease over the time indicating rel-

ative compactness of the structure (Supplementary Fig. 3).

Rg value lies between 1.55 and 1.50 nm across the time

scale in ps.

The possible binding sites of modeled Rab3A were pre-

dicted using PASS software and were compared with the

predictions from the CastP webserver and the template

structure. The consensus residues in the comparison were

considered as the active site lining residues. The volume

center of the cavity predicted by PASS was used as the grid

center for the docking analysis where as cavity lining resi-

dues were obtained through CastP results. The cavity lining

residues as found through CastP were SER16, SER17,

GLY19, LYS20, THR21, SER22, PHY25, SER31, PHY32,

THR33, PRO34, ALA35, PHY36, VAL37, SER38, THR39,

THR63, ALA64, GLY65, GLN66, LYS121 and LYS152 for

the most conserved active site of Rab3A protein.

Docking result analysis

We used GLIDE and GOLD docking programs to screen

1,364 molecules from the ZINC database against Rab3A to

identify potential inhibitor molecule as mentioned in the

Methods section. The top five ranking molecules based on

GLIDE scores are listed in Table 1. The GLIDE scores

Fig. 3 Trajectories of the overall Ca (RMSD) of the human Rab3A

protein structure with respect to the starting structure over 1,000 ps

MD simulation. The x axis represents the simulation time in

picoseconds. The y axis represents RMSD in nm unit

Fig. 4 Calculated energy versus time plot using GROMACS soft-

ware. The x axis represents the simulation time in picoseconds. The

y axis represents energy in nm unit. The average energy recorded was

-209,791 kJ/mol

Fig. 5 Residue-wise RMSF profiles of the human Rab3A protein

structure computed after stabilization of the RMSD trajectories. The x

axis represents the residue number. The y axis represents RMSF in nm

Mol Biol Rep

123

and X-Scores of these compounds have ranges of -6.65 to

-2.52 kcal/mol and -4.22 to -2.61 kcal/mol respectively.

The top ranking molecule (ZINC13152284) has a Glide

score of -6.65 kcal/mol and X-Score of -3.02 kcal/mol

with 03 hydrogen bonds and 09 hydrophilic contacts. For

each compound docked using the GLIDE and GOLD

programs, the X-Score (consensus scoring function) pro-

gram was used to calculate binding energies and listed in

Table 1 along with IUPAC name of the compounds and

their respective ZINC ID’s. Interactions (hydrogen and

hydrophobic) for the top five best ranking ligands based on

LIGPLOT showing hydrogen bonding and hydrophobic

contacts between the docked poses of protein and ligands

(Fig. 6).

GTP docking

Docking with GTP was carried out on the binding site of

Rab3A using GOLD software applying the parameters of

standard default setting with 50 genetic algorithm runs,

filtering poses based on GOLD fitness function. The

Rab3A-GTP complex has best GOLD fitness value of

77.73. To substantiate the estimations done by the GOLD

program, we used consensus scoring program X-Score. The

scoring schema used in the software X-Score computes a

binding score for a given protein–ligand complex structure.

The predicted binding energy for the docked complex was

found as 5.51 kcal/mol and predicted average Log Kd as

3.77, using X-Score program. Ligplot shows hydrogen

bonding (S16, S17, V18, G19, K20, T21, S22, S31, T33,

A35, S38, T39 and G65) and hydrophobic interacting

residue (F25, F32, P34, F36, V37, D62 and A64) with the

GTP ligand in the binding site of Rab3A protein (Fig. 7).

Variations in structure have been observed for different

GTPases [43]. There are more than 10 conserved features

in the Rab3 subfamily of proteins. Conserved domain

database was used in identifying critical variations, if any,

in human Rab3A protein [44]. The chemical binding site,

binds GTP and stabilizes the conformation of active Rab.

All the conserved residues in this motif, critical for func-

tion are intact in this motif. In Rabs the residues included in

this site are 32–38, 49, 50, 53, 55–56, 82, 136, 137, 139,

140 and 166–168. The effector interaction site in human

Rab3A is mapped onto the site that includes amino acid

residues 22, 56, 58–60, 75, 77, 84, 85, 88, 92, 94–97,

126–128, 184, 187.

The promotion of GDP bound Rab to GTP bound form

is catalyzed by Guanine nucleotide exchange factors. There

is little sequence similarity in the amino acid sequences of

Rab effectors, GEFs, and GAPs and regulators or effectors

for other GTPases. The putative GEF interaction site in

human Rab is conserved within the family and is formed by

53, 57–64, 71, 73. Guanine nucleotide dissociation inhib-

itors facilitate Rab recycling by masking C-terminal lipid

binding and helping in cytosolic localization. The GDI site,

prediction was based on interaction between S. cervisiae

YPT1 and its guanine nucleotide dissociation inhibitor.

This site includes 56, 57, 59, 77, 78, 85, 87, 89–91.

The Rab subfamily includes surface loops that undergo

conformational changes upon GTP binding, called switch I

and switch II. In Rab3A, the active conformation is further

stabilized by an extensive hydrophobic interface involving

conserved residues in both of switch regions. The switch I

in human Rab3A is represented by 48, 53–61, whereas

switch II is represented by residues 81, 83–93. The G-box

motifs involved in nucleotide binding are conserved in

human Rab3A. The walker A motif includes residues

30–37, whereas walker B motif includes 78–81.

Rab GTPases involved in the regulation of exocytic

vesicle trafficking pathways show a conserved set of amino

acid residues for active conformation of switch regions

[45]. Conserved serine residues along with others are

important for stabilizing the active conformation in Rab’s.

Serine 22 as described earlier is involved in the effector

Table 1 Top ranking ligands (molecules from Zinc database) after virtual screening against Rab3A using GLIDE and GOLD docking programs

S. no. ZINC ID IUPAC name of ligand Glide score

(Kcal/mol)

Glide X-score

(Kcal/mol)

Gold

score

Gold X-score

(Kcal/mol)

1 ZINC13152284 5-phenyl-2-[(2S)-5-phenyl-2,3-dihydro-1,3-benzoxazol-2-yl]

-2, 3-dihydro-1,3-benzoxazole

-6.659 -3.02 64.54 -3.12

2 ZINC13143009 (S)-[(2R)-piperidin-1-ium-2-yl]-[2-(trifluoromethyl)

-6-[4-(trifluoromethyl)phenyl]pyridin-4-yl]methanol

-2.647 -2.62 59.60 -2.48

3 ZINC00247785 3-hydroxy-N-(3-nitrophenyl) naphthalene-2-carboxamide -5.539 -4.22 55.29 -3.92

4 ZINC13143008 (S)-[(2S)-piperidin-1-ium-2-yl]-[2-(trifluoromethyl)

-6-[4-(trifluoromethyl)phenyl]pyridin-4-yl]methanol

-3.509 -2.61 55.70 -3.01

5 ZINC01705919 2-isoquinolin-1-yl-4-(4-methylphenyl)-5-phenyl-1,3-oxazole -2.521 -4.08 64.85 -3.97

cFig. 6 Molecular interaction plots showing interacting cavity lining

residues between human Rab3A receptor and NCI diversity set II

molecules (a) (b) (c) (d) (e)

Mol Biol Rep

123

KeyLigand bond

Non-ligand bond

3.0 Hydrogen bond and its length

His 53 Non-ligand residues involved in hydrophobiccontact(s)

Corresponding atoms involved in hydrophobic contact(s)

KeyLigand bond

Non-ligand bond

3.0 Hydrogen bond and its length

His 53 Non-ligand residues involved in hydrophobiccontact(s)

Corresponding atoms involved in hydrophobic contact(s)

KeyLigand bond

Non-ligand bond

3.0 Hydrogen bond and its length

His 53 Non-ligand residues involved in hydrophobiccontact(s)

Corresponding atoms involved in hydrophobic contact(s)

KeyLigand bond

Non-ligand bond

3.0 Hydrogen bond and its length

His 53 Non-ligand residues involved in hydrophobiccontact(s)

Corresponding atoms involved in hydrophobic contact(s)

KeyLigand bond

Non-ligand bond

3.0 Hydrogen bond and its length

His 53 Non-ligand residues involved in hydrophobiccontact(s)

Corresponding atoms involved in hydrophobic contact(s)

3.05

C8

C9 N1

C5

O1

C20

C10

C11 N2

C12

C19

C18 C17

C16

C13

C15

C14

C25

C21

C24

C23

C22

C4

C6

C3

C2

C1

C7

N

CA

CB

C

OG

O

Phe 25

Phe 36

Thr 33

Gly 19

Thr 21

Val 37

Ser 17

Phe 32Ser 31

A B

C

E

D

Lig 178

Ser 22

2.973.17

2.98

3.03

3.07

3.25

3.24

2.89

3.07

C11

O1 N1

C7

C12

C6

C8

C5

O4

C4

C3

C9

C2

C10

C1

C17 C13

C16

C15 N2

C14

O2

O3

N

CA

CB

C

CG

CD

CE NZ

O

N

CA

CB

C

OG1 CG2

O

N

CA

CB

C

OG

O

N

CA

C

O

N

CA

CB

C

CG1

CG2

O

N

CA

CB

C

OG

O

Phe 25

Phe 36

Thr 33

Lig 178

Lys 20

Thr 21

Ser 17

Gly 19

Val 37

Ser 22

3.11

3.04

2.80

3.30

C7 C8

C12

C4

C9

C10

C11

O1

N1

C13

C14

O2

N2

C20

C15 C19

C16

C18

C17

C21

C26

C22

C25

C24

C23

C3

C5

C2

C1 C6

N

CA

CB C

CG1

CG2

O

N

CA CB

C OG

O

N CA

CB

C

O

Phe 36

Phe 25

Thr 21

Lys 121

Gly 19

Lig 178

Val 37

Ser 22

Ala 35

3.05

3.22 C7

C8

N1

C6

C9 C10

C13

C11 C12

C14

O1 C15

N2

C16

C17

C18

C1

C5 C2

C3 C4

C19

F4

F5

F6

F1

F2

F3

N

CA

CB

C OG

O

N

CA

CB

C

CG

CD

CE

NZ

O

Phe 36

Phe 25

Gly 19

Thr 21

Thr 33

Lig 178

Ser 17

Lys 121

3.07

C7

C8

N1

C6

C9 C10

C13

C11 C12

C14

O1 C15

N2

C16

C17

C18

C1

C5

C2 C3

C4

C19

F4

F5

F6

F1

F2

F3

N

CA CB

C

CG

CD

CE

NZ

O

Phe 36

Phe 25

Ser 17

Gly 19

Thr 21Thr 33

Lig 178

Lys 121

Mol Biol Rep

123

binding site according to sequence homology studies.

Interestingly the molecular interaction plots of 3 NCI

diversity set molecules show involvement of the S22 res-

idue in hydrogen bonding with human Rab3A. The other

hits found in this study indicate potential binding sites in

the G1 box of human Rab3A that spans residues 30–37.

Conclusion

Although much of the physical entity in the protein struc-

ture is conserved in human Rab3A, many non random

substitutions in sequence within functionally important

regions are known. These variations could imply differ-

ences in functional properties that contribute to the speci-

ficity of interactions. With undisputed role in molecular

trafficking and regulation of cell cycle, the structure of

Rab3A provided in this study identifies new interactions

that are predicted to be important in the regulation of

GTPase activity. Apart from structural description of

Rab3A, elaboration of binding sites for inhibitors and

phosphorylation site in Rab3A provides us with an

opportunity to utilize binding pockets in targeted inacti-

vation of this protein. The top ranking molecule

(ZINC13152284) is a good starting point for further

development of strong inhibitors. Rab3A has also been

reported in malignant neuroendocrine cells and breast

cancer cells, which suggests its potential role in tumouri-

genesis. Rab3A could be an important target for cancer

therapeutics and our study can serve as a valuable reference

for structural and functional analysis.

References

1. Schwartz SL, Cao C, Pylypenko O, Rak A, Wandinger-Ness A

(2007) Rab GTPases at a glance. J Cell Sci 120:3905–3910

2. Jordan BA, Fernholz BD, Boussac M, Xu C, Grigorean G, Ziff

EB, Neubert TA (2004) Identification and verification of novel

rodent postsynaptic density proteins. Mol Cell Proteomics

3:857–871

3. Zerial M, McBride H (2001) Rab proteins as membrane orga-

nizers. Nat Rev Mol Cell Biol 2:107–117

4. Pereira-Leal JB, Seabra MC (2000) The mammalian Rab family

of small GTPases: definition of family and subfamily sequence

motifs suggests a mechanism for functional specificity in the Ras

superfamily. J Mol Biol 301:1077–1087

5. Geppert M, Bolshakov VY, Siegelbaum SA, Takei K, De Camilli

P, Hammer RE et al (1994) The role of Rab3A in neurotrans-

mitter release. Nature 369:493–497

6. Darchen F, Goud B (2000) Multiple aspects of Rab protein action

in the secretory pathway: focus on Rab3 and Rab6. Biochimie

82:375–384

7. Fischer von Mollard G, Mignery GA, Baumert M, Perin MS,

Hanson TJ, Burger PM et al (1990) Rab3 is a small GTPbinding

protein exclusively localized to synaptic vesicles. Proc Natl Acad

Sci USA 87:1988–1992

8. Geppert M, Sudhof TC (1998) RAB3 and synaptotagmin: the yin

and yang of synaptic membrane fusion. Annu Rev Neurosci

21:75–95

9. Segev N (2001) Ypt and Rab GTPases: insight into functions

through novel interactions. Curr Opin Cell Biol 13:500–511

10. Takai Y, Sasaki T, Shirataki H, Nakanishi H (1996) Rab3A small

GTP-binding protein in Ca2?-dependent exocytosis. Genes Cells

1:615–632

11. Aligianis IA, Johnson CA, Gissen P, Chen D, Hampshire D,

Hoffmann K, Maina EN et al (2005) Mutations of the catalytic

subunit of RAB3GAP cause Warburg Micro syndrome. Nat

Genet 37:221–223

12. Aligianis IA, Morgan NV, Mione M et al (2006) Mutation in

Rab3 GTPase-activating protein (RAB3GAP) noncatalytic sub-

unit in a Kindred with Martsolf syndrome. Am J Hum Genet

78:702–707

13. Coppola T, Hirling H, Perret-Menoud V, Gattesco S, Catsicas S,

Joberty G et al (2001) Rabphilin dissociated from Rab3 promotes

endocytosis through interaction with Rabaptin-5. J Cell Sci

114:1757–1764

14. Wang Y, Okamoto M, Schmitz F, Hofmann K, Sudhof TC (1997)

Rim is a putative Rab3 effector in regulating synaptic-vesicle

fusion. Nature 388:593–598

15. Haynes LP, Evans GJ, Morgan A, Burgoyne RD (2001) A direct

inhibitory role for the Rab3-specific effector, Noc2, in Ca2?-

regulated exocytosis in neuroendocrine cells. J Biol Chem

276:9726–9732

16. Park JB, Farnsworth CC, Glomset JA (1997) Ca2?/calmodulin

causes Rab3A to dissociate from synaptic membranes. J Biol

Chem 272:20857–20865

Fig. 7 Rab3A and GTP docking

Mol Biol Rep

123

17. Nonet ML, Staunton JE, Kilgard MP, Fergestad T, Hartwieg E,

Horvitz HR et al (1997) Caenorhabditis elegansrab-3 mutant

synapses exhibit impaired function and are partially depleted of

vesicles. J Neurosci 17:8061–8073

18. Leenders AG, Lopes da Silva FH, Ghijsen WE, Verhage M

(2001) Rab3a is involved in transport of synaptic vesicles to the

active zone in mouse brain nerve terminals. Mol Biol Cell

12:3095–3102

19. Geppert M, Goda Y, Stevens CF, Sudhof TC (1997) The small

GTP-binding protein Rab3A regulates a late step in synaptic

vesicle fusion. Nature 387:810–814

20. Castillo PE, Janz R, Sudhof TC, Tzounopoulos T, Malenka RC,

Nicoll RA (1997) Rab3A is essential for mossy fibre long-term

potentiation in the hippocampus. Nature 388:590–593

21. Doussau F, Clabecq A, Henry JP, Darchen F, Poulain B (1998)

Calcium-dependent regulation of rab3 in short-term plasticity.

J Neurosci 18:3147–3157

22. Tian M, Xu CS, Montpetit R, Kramer RH (2012) Rab3A medi-

ates vesicle delivery at photoreceptor ribbon synapses. J Neurosci

32:6931–6936

23. Culine S, Honore N, Tavitian A, Olofsson B (1992) Overex-

pression of the Ras-related Rab2 gene product in peripheral blood

mononuclear cells from patients with hematological and solid

neoplasms. Cancer Res 52:3083–3088

24. Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W:

improving the sensitivity of progressive multiple sequence

alignment through sequence weighting, position-specific gap

penalties and weight matrix choice. Nucleic Acids Res 22:4673–

4680

25. Sali A, Blundell TL (1993) Comparative protein modeling by

satisfaction of spatial restraints. J Mol Biol 234:779–815

26. Laskoswki RA, MacArthur MW, Moss DS, Thornton JM (1993)

PROCHECK: a program to check the sterochemical quality of

protein structures. J Appl Cryst 26:283–291

27. Eisenberg D, Luthy R, Bowie JU (1997) VERIFY3D: assessment

of protein models with three-dimensional profiles. Methods

Enzymol 277:396–404

28. Frishman D, Argos P (1995) Knowledge-based protein secondary

structure assignment. Proteins 23:566–579

29. Shindyalov IN, Bourne PE (1998) Protein structure alignment by

incremental combinatorial extension (CE) of the optimal path.

Protein Eng 11:739–747

30. Hess B, Kutzner C, Spoel D, Lindahl E (2008) GROMACS 4:

algorithms for highly efficient, load-balanced, and scalable

molecular simulation. J Chem Theory Comput 4:435–447

31. Chris O, Alessandra V, Alan E, Wilfred FVG (2004) A biomo-

lecular force field based on the free enthalpy of hydration and

solvation: the GROMOS force–field parameter sets 53A5 and

53A6. J Comput Chem 25:1656–1676

32. Essmann U, Perera L, Berkowitz ML, Darden T, Lee H, Pedersen

LG (1995) A smooth particle mesh Ewald method. J Chem Phys

103:8577–8593

33. Hess B, Bekker H, Berendsen HJC, Fraaije JGEM (1997) LINCS:

a linear constraint solver for molecular simulations. J Comput

Chem 18:1463–1472

34. Berendsen HJC, Postma JPM, Gunsteren WF, DiNola A, Haak JR

(1984) Molecular dynamics with coupling to an external bath.

J Chem Phys 81:3684–3690

35. Patrick BG, Stouten PFW (2000) Fast prediction and visualiza-

tion of protein binding pockets with PASS. J Comput Aided Mol

Design 14:383–401

36. Binkowski TA, Naghibzadeh S, Liang J (2003) CASTp: com-

puted atlas of surface topography of proteins. Nucleic Acids Res

31(13):3352–3355

37. Friesner RA, Banks JL, Murphy RB, Halgren TA, Klicic JJ

(2004) Glide: a new approach for rapid, accurate docking and

scoring. 1. Method and assessment of docking accuracy. J Med

Chem 47:1739–1749

38. Jones G, Willett P, Glen RC, Leach AR, Taylor R (1997)

Development and validation of a genetic algorithm for flexible

docking. J Mol Biol 267:727–748

39. Wang R, Lai L, Wang S (2002) Further development and vali-

dation of empirical scoring functions for structure-based binding

affinity prediction. J Comput Aided Mol Des 16:11–26

40. Wallace AC, Laskowski RA, Thornton JM (1995) LIGPLOT: a

program to generate schematic diagrams of protein–ligand

interactions. Protein Eng 8:127–134

41. DeLano WL (2002) Pymol. DeLano Scientific, South San

Francisco

42. Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt

DM (2004) UCSF chimera—a visualization system for explor-

atory research and analysis. J Comput Chem 25:1605–1612

43. Sprang SR (1997) G protein mechanisms: insights from structural

analysis. Annu Rev Biochem 66:639–678

44. Marchler-Bauer A (2011) CDD: a conserved domain database for

the functional annotation of proteins. Nucleic Acids Res

39:225–229

45. Dumas JJ, Zhu Z, Connolly JL, Lambright DG (1999) Structural

basis of activation and GTP hydrolysis in Rab proteins.

Structure 7:413–423

Mol Biol Rep

123