Network Biology Approaches to Identify the Drug Lead Molecule for Neurodevelopmental Disorders in Human

2.1. Data Collection

For Autism 494 genes, for Intellectual Disability (ID) 399 genes, and 137 genes for Attention Deficit Hyperactivity Disorder (ADHD) were reported from NCBI.

2.2. Construction of PPI Network

STRING [15] allows a detailed study of network interaction by providing several features. PPI network constructed using of STRING 10.0 (Search Tool for the Retrieval of Interacting Genes) [16] by 494 genes for Autism, 399 genes for Intellectual Disability (ID), and 137 genes for Attention Deficit Hyperactivity Disorder (ADHD) as input and based on highest confidence score 0.7 obtained from experimental data [17]. Further PPI network was analyzed and visualized by Cytoscape 3.3.0 [18, 19].

2.3. Analysis of PPI Network

Analysis of the protein-protein interaction (PPI) network was done by using Cytoscape 3.3.0. Betweenness Centrality (BC) and node degree are the parameters included in Cytoscape Plug-in Network Analyzer. Node is represented by gene and interaction is represented by edges. Edges' number specifies node degree [20]. Betweenness centrality explains the significance of the node primarily based on the quantity of the shortest ways that undergo every node. In the study, the PPI network was analyzed based on these parameters.

2.4. Protein Modeling and Structure Validation

Obtained drug target from PPI network analysis for neurodevelopmental disorders: Autism, Intellectual Disability (ID), and AHDH. For the identification of the target sequence, the BLAST program was performed for the search of a structurally similar sequence with the protein data bank (PDB). Based on the structure similarity parameter, SNAP25 and DRD4 were modeled using on-line modeling server SWISS-MODEL. The structure of target protein 4EJN (PDB-ID) for Autism was retrieved from PDB. Further, the modeled structures were validated through the Ramachandran Plot and Errat using the SAVES server.

2.5. Ligand Identification

The structures of twenty-seven possible phytochemical compounds having neuroprotective activity were selected for study based on various literature reviews (Table 1). The 2D & 3D structures of compounds were retrieved from the PubChem database. These compounds were downloaded in (sdf) file format & converted into (PDB) format using Chimera software. Prediction of Molecular & bioactivity analysis of leads was performed by using molinspiration chemoinformatics software that satisfied them as potential inhibitors.

2.6. Molecular Docking

Rigid molecular docking was performed between the target protein and ligand using Autodock 1.5.6 tool to predict the binding affinity, ligand efficiency, and inhibitory constant. The criteria of binding energy were taken into consideration (based on the lowest binding energy score) to select the best protein-ligand complex. Binding energy was calculated using a semi-empirical free energy force field with charge-based desolvation and grid-based docking.

3.1. Network Analysis and Identification of Target

The PPI networks were constructed inputting 494 genes for Autism, 399 genes for Intellectual Disability (ID), and 137 genes for Attention Deficit Hyperactivity Disorder (ADHD) using the STRING database. Obtained networks were analyzed and visualized using Cytoscape 3.3.0 plug-in Network Analyzer. Based on a topological parameter BC and node degree: AKT serine/threonine kinase 1 (AKT1) for Autism, Synaptosomal nerve-associated protein 25 (SNAP25) for Intellectual Disability (ID), and Dopamine receptor D4 (DRD4) for Attention Deficit Hyperactivity Disorder (ADHD) are selected as a key drug target with highest BC value and degree node Figs. (1, 2 & 3).

Various literature finding suggests the active role of AKT1 gene as a key modulator of the AKT-mTOR signaling pathway controlling the tempo of the process of newborn neurons integration during adult neurogenesis, including correct neuron positioning, dendritic development, and synapse formation. Ile67Asn variation in SNAP-25 isoform B inhibits exocytosis of synaptic vesicles by altering the α-helical coiled structure of the SNARE complex. This mutation is a cause disorder for Intellectual Disability(ID). SNAP25 is substantially involved in many major neurological disorders which makes it a primary target for diagnosis and treatment of threatening diseases [21]. Dopamine D4 receptor gene (DRD4) has an important role in the pathogenesis of ADHD and one of the most important genes and is held to be responsible for neuronal signaling in the mesolimbic system of the brain area [22].

3.2. Protein Modeling and Structure Validation

The total length of AKT1 protein is 480 amino acids, SNAP25 protein is 206 amino acids, and DRD4 protein is 467 amino acids. The 3D structure of target protein 4EJN for Autism was retrieved from PDB and the 3-D structures of SNAP25 and DRD4 were built using SWISS-MODEL. The model structure is further validated through online server Rampage. The position of amino acid residues in allowed and disallowed region and overall Stereochemical property of the protein was accessed by Ramachandran plot. The Ramachandran plot indicates 91% for Autism, 89.8% for Intellectual Disability (ID), and 93.9% for Attention Deficit Hyperactivity Disorder (ADHD) in the favored region that reflects the good quality model as the expected value is above 85% in the favored region (Figs. 4, 5, & 6).

3.3. Molecular Docking Analysis

The molecular docking analysis is the probably most suitable method to identify best-fit interaction between protein and ligand. Protein-ligand interaction plays a key role in drug discovery. Docking uses the computational simulation to predict the preferred orientation of ligand to a receptor when interacted with each other to form a higher stability complex.

Molecular docking was performed for best 5 phytochemicals having neuroprotective activity reported in various literature searches, against AKT1, SNAP25, and DRD4 proteins using Autodock 1.5.6 tool. The protein PDB file was prepared by adding essential hydrogen atoms, Kollman charges, and solvation parameters and converting it into PDBQT format by using the Auto Dock Tool (ADT). The ligand file was prepared by detecting root and torsion and converting it into PDBQT format. The grid map was set of 60 x 60 x 60-angstrom grid points that were generated using the Auto grid program. The docking was performed using a Lamarckian Genetic Algorithm (LGA) parameters and run on the Cygwin platform. Docking results were analysed based on the lowest binding free energy (BE) and optimal docking energy conformation (Tables 2, 3 and 4 & Figs. 7, 8 and 9).