Tertiary structure prediction and identification of druggable pocket in the cancer biomarker - Osteopontin-c

Abstract

Background: Osteopontin (Eta, secreted sialoprotein 1, opn) is secreted from different cell types including cancer cells. Three splice variant forms namely osteopontin-a, osteopontin-b and osteopontin-c have been identified. The main astonishing feature is that osteopontin-c is found to be elevated in almost all types of cancer cells. This was the vital point to consider it for sequence analysis and structure predictions which provide ample chances for prognostic, therapeutic and preventive cancer research.

Methods: Osteopontin-c gene sequence was determined from Breast Cancer sample and was translated to protein sequence. It was then analyzed using various software and web tools for binding pockets, docking and druggability analysis. Due to the lack of homological templates, tertiary structure was predicted using ab-initio method server - I-TASSER and was evaluated after refinement using web tools. Refined structure was compared with known bone sialoprotein electron microscopic structure and docked with CD44 for binding analysis and binding pockets were identified for drug designing.

Results: Signal sequence of about sixteen amino acid residues was identified using signal sequence prediction servers. Due to the absence of known structures of similar proteins, three dimensional structure of osteopontin-c was predicted using I-TASSER server. The predicted structure was refined with the help of SUMMA server and was validated using SAVES server. Molecular dynamic analysis was carried out using GROMACS software. The final model was built and was used for docking with CD44. Druggable pockets were identified using pocket energies.

Conclusions: The tertiary structure of osteopontin-c was predicted successfully using the ab-initio method and the predictions showed that osteopontin-c is of fibrous nature comparable to firbronectin. Docking studies showed the significant similarities of QSAET motif in the interaction of CD44 and osteopontins between the normal and splice variant forms of osteopontins and binding pockets analyses revealed several pockets which paved the way to the identification of a druggable pocket.

Figure 1

Predicted structure of osteopontin-c by I-TASSER. Figure shows five predicted models from I-TASSER Server. Of these five models, model 2 is reported as best model by comparing with the electron microscopic structure of Bone sialoprotein.

Figure 2

Superposition comparison for predicted and refined structure of osteopontin-c. Here all conformations are superimposed with a reference structure (Predicted model) using RMSD fit. Accelry’s Discovery Studio Visualizer 1.7 was utilized for the superimposition. A. predicted structure. B. Refined structure.

Figure 3

The Root mean square deviation plot (Carbon Alpha back bone) obtained from GROMACS tool during molecular dynamics simulation for 10 nanoseconds. Root mean square deviation (RMSD) of osteopontin-c model during the molecular dynamics (MD) simulation. The stabilization of the structure occurred in approximately 10 ns. Generated by GROMACS.

Figure 4

Selected structural domains and their corresponding locations in the human osteopontin-c.

Figure 5

Multiple sequence alignment of osteopontin. Multiple sequence alignment of osteopontin amino acid residues of Bos taurus (cattle), Bubalus bubalis (water buffalo), Homo sapiens (human), Oryctolagus cuniculus (rabbit), Mus musculus (mouse), Rattus norvegicus (Norway rat) and Gallus gallus (chicken). The *(star) in the sequence represents identical residues.

Figure 6

Phylogenetic tree. Phylogenetic tree obtained using a multiple alignment of osteopontin protein sequences from cattle (NP_776612), water buffalo (ABD73011), human (NP_001035149.1), rabbit (NP_001075663), mouse (NP_033289), rat (NP_037013) and chicken (NP_989866). See text for details.

Figure 7

RGD and RSK domains of osteopontin-c and osteopontin-a. RGD and RSK domains are represented in ball and stick model whereas remaining part of the protein is represented in solid ribbon model using Accelry’s Discovery Studio Visualizer 1.7.

Figure 8

N-terminal fragment (A) and C-terminal fragment (B) of osteopontin-c after the action of thrombin. Polyimerized structure (C) of C-terminal fragment of osteopontin-c. RGD domains are represented in ball and stick model in N-terminal fragment. Alternate presence of helices might be seen in C-terminal fragment which might provide the clue for the formation of pore or fibre. Tube and surface model are represented in the Figure C. Polymerization was achieved using ICM Molsoft Tool.

Figure 9

Multiple sequence alignment comparison between osteopontin-a and osteopontin-c with other species with respect to transglutaminase acting stie. Alignment analysis clearly showed the absence of transglutaminase site in osteopontin-c. Transglutaminase-2 acting site is indicated by an open box.

Figure 10

Docked structure of CD44 with osteopontin-c and osteopontin-a. A. Docked structure of CD44 (red) and osteopontin-a (green) and osteopontin-c (purple). B. Interaction site of CD44 (red) and osteopontin-a (blue). C. Interacation site of CD44 (red) and osteopontin-c (blue). It was achieved using Accelry’s Discovery Studio Visualizer 1.7 Tool.

Figure 11

Predicted binding pockets in osteopontin-c and osteopontin-a. PocketFinder and Q-site Finder predicted pockets in osteopontin-a and osteopontin-c are shown as groups.