The most dominant forms of neurological conditions are Alzheimer's disease (AD) and Parkinson's disease (PD) in the world's population. Parkinson’s disease affects roughly 6 million individuals, while Alzheimer’s disease affects over 50 million people [1, 2].  In the year 2050, the projected population is expected to reach 152 million, with an annual growth rate of 5% [3, 4]. PD is the 2nd prevalent dementia after dementia, affecting 1-2% of persons over 65 and 4-5% of those over 85 [5]. Lewy bodies are a neuropathological hallmark of PD and are predominantly made up of the protein SNCA. The formation and deposit of beta-amyloid peptide into lesions in the fluid brain tissue, followed by hyperphosphorylated or tau in intracellular neurofibrillary tangles [6], named after the German psychiatrist Alois Alzheimer, AD is observed [7]. Although PD is referred to as another age-related neurotoxic disease, it is recognized clinically by rigidity, tremor, and loss of activity, as well as neuropathologically by gradual neuronal loss of the spinal nerve [8]. Mutations in the PRKN gene represent the primary cause of autosomal-recessive juvenile PD [9]. The ABCA1, APOE, MAPT, and SNCA genes are typically related to AD and are related to cognitive function in PD pathogenesis, and SNCA mutations are identified in familial forms of the disease [10].  Globally, dementia is expected to cost $1 trillion annually, placing a significant strain on people, families, and communities [11, 12]. While there is no FDA-approved treatment for AD, there are drugs that can help with the symptoms [13]. A genetic mutation in the LRRK2, PARK2, PARK7, PINK1, or SNCA gene can cause family-linked PD, which affects around 15% of people. The underlying causes of PD involve the presence of intracellular Lewy body inclusions within neurons and the loss of dopaminergic neurons in the substantia nigra pars compacta (SNpc) [10]. Protein aggregation and gene mutation are major causes of dopamine neuron degeneration. However, the exact cause of dopamine degradation is unknown, and several atypical PD family types are linked to specific genetic loci, knowing the disease's biology [14]. Certain genes, such as the Apolipoprotein E (APOE) gene, affect later-onset dementia, which is more common. An increased risk for this disorder varies with APOE gene variation, notably APOE4 [15]. Familial PD is related to specific mutations in genes such as Synuclein Alpha (SNCA) [16, 17] and Parkin protein (PRKN) [18].  The APOE gene encodes a 317-aminoacid protein that is found on chromosome 19q13.32 with total exons 6.  The three common variations, ?2, ?3, and ?4, each produce a single amino acid change in the APOE protein. Specifically, APOE ?4 is linked with a higher risk of dementia, whereas APOE ?2 with lower risk factors [19]. The variance in the APOE genotype affects beta-amyloid buildup in the brain's cortex and its blood vessels appears to be the basis of the APOE genotype's impact on the emergence of various disorders [20]. Significant evidence ties the APOE gene to the risk of dementia and has been studied in various neurological disorders [21, 22]. The SNCA gene contains alpha-synuclein and is found on chromosome 4q22.1, with 12 exons [16]. The initial association between ?-synuclein and PD was established in 1997 when a coding SNP A to T at position 53 in SNCA was linked to autosoma dominant Parkinsonism [23].  Various researchers found that large SNCA aggregates led to the etiology of an array of neurological conditions, including PD, dementia (AD), multiple systems atrophy (MSA), and dementia with Lewy bodies (DLB) [24]. The protein alpha-synuclein, which has 140 amino acids, is a member of the same protein family as the ?- and ?-synucleins [25].  The first mutation in the SNCA gene, which codes for the ASN protein, and the first mutation found in PD, was reported in 1990 [26]. The first mutation in the Parkin protein (PRKN) gene discovered in 1998 [10, 27]. Biallelic PKRN gene variants are known to be responsible for autosomal recessive Parkinson's disease (PD). However, the role of monoallelic PRKN alterations as a risk factor for PD is still uncertain [28]. Parkin, a 465 amino acid autoinhibited E3 ubiquitin ligase, is encoded by PRKN comprising13 exons and is mapped on human chromosome 6q26 [29, 30]. A global study is being carried out on the genetics of the neurological condition, and reports of variants are emerging. The two most often found genes, PINK1 and PARK2, are both related to autosomal recessive forms of PD [31].  Moreover, In silico studies have a substantial impact on the recognition of potential SNPs because they are quick, economical, and can aid with future genetic research [32]. The objective of this study is to forecast the 3D structures of the APOE, SNCA, and PRKN proteins using the SWISS model, which relies on PyMOL serves to display molecular crystal structures. In this study, computational methods including GeneMANIA and STITCH analysis were employed to find the gene-gene and PPI interactions. Furthermore, protein-ligand docking was used to explore the associations of APOE, SNCA, and PRKN with ligands. Despite its drawbacks, our study generated insights that will be useful for future lab-based and animal studies regarding gene-associated neurological disorders.

MATERIALS AND METHOD

The comprehensive methodology of this current

paper is illustrated in Figure 1.

Analysis of variants from the GnomAD database

The protein SNP data for the APOE, SNCA, and PRKN genes acquired from GnomAD are available at https://gnomad.broadinstitute.org. A worldwide collaboration of researchers established the Genome Aggregation Database Consortium, commonly referred to as GnomAD. Their goal was to collect and standardize exome and genome sequencing data obtained from various extensive sequencing initiatives and furnish data summaries to the scientific community [33, 34]. Previously, this initiative was recognized as the Exome Aggregation Consortium (ExAC). The reference genome employed for sequence alignment was GRCh37/hg19, and the alignment process was executed using the GATK tool [35].

The APOE, SNCA, and PRKN sequences were obtained in the FASTA format from the Universal Protein Resource (UniProt) at http://www.uniprot.org. The APOE from Homo sapiens (P02649), SNCA (P37840), and PRKN (O60260) are about of these proteins [36]. The APOE, SNCA, and PRKN gene data information was retrieved from the NCBI database at https://www.ncbi.nlm.nih.gov [5]. Utilizing computational modeling methods, the sequences of the aforementioned proteins were analyzed, and 3D models were built.  The protein data taken into account for this investigation is displayed in Table 1.

Biophysical and Visualization Analysis

GeneMANIA available at https://genemania.org is a web server that develops hypotheses on gene function, aspects into gene lists, and ranks genes for functional testing [37]. The website GeneMANIA may identify new genes linked to a set of input genes by utilizing a vast repository of biological correlation information. Protein interactions with genes, coexpression, methods, and colocalization are all examples of interaction statistics [38, 39]. To study how proteins interact, the STITCH server, sited at http://stitch.embl.de was used. For quick access to verified theoretical and experimental interactions of the required protein, this database offers an essential evaluation and integration of protein-protein interactions [40, 41].

Prediction of protein models

The protein sequences that were obtained were used as query sequences in comparative modeling. The 3D structures of APOE, SNCA, and PRKN were predicted using SWISS-MODEL are given at http://swissmodel.expasy.org, which also included integrated external data sources like UniProt, STRING analysis and used to predict the protein crystal templates [42]. The refinement is done by Galaxy Refine available at https://galaxy.seoklab.org/cgi-bin/submit.cgi?type=REFINE. UCSF Chimera is installed from http://www.cgl.ucsf.edu/chimera [43], Discovery Studio  was downloaded from https://www.3ds.com/products-services/biovia/products/molecular-modeling-simulation/biovia-discovery-studio [44], PyMOL available at https://pymol.org  [45] is a visualization tool of 3D structure.

Model evaluation by SAVES server

The core stability and reliability of the modeled structures of the APOE, SNCA, and PRKN were assessed using several ways. By counting the residues in the permitted zones of the Ramachandran plot, the stereochemical quality of the model was evaluated using SAVES algorithms available at https://saves.mbi.ucla.edu [46]. The derived protein structure was reevaluated using QMEAN Z-scores from the QMEAN tools available at http://swissmodel.expasy.org/docs/structure_assessment for its dependability and model quality. When evaluating a protein's 3D structure, SAVES employs six modules, including PROCHECK, 3D Verify, and ERRAT, to ensure the quality of the entire 3D model [47, 48]. A Ramachandran plot of psi-phi torsion angles are produced by the PROCHECK, with over 90% of residue in the most favored regions likely to be in acceptable quality.  The Ramachandran plot, which displays how residues behave in zones where they are allowed and not allowed [49]. PROCHECK analyzes both the total structural geometry and the geometry of each distinct residue to ascertain the stereochemical quality of a protein structure [50]. ERRAT evaluates the overall model quality based on the probability distribution of non-bonded interactions among a variety of atoms relying on shared atomic interactions [48, 51, 52].

Virtual screening of protein-ligand docking

The SDF files for the identified ligands, including Amphimedine, Benazepril, Captopril, Curcumin, Deoxyamphimedine, Emetine, Fibrinogen, Hyaluronic Acid, Lacosamide, and Quercetin, were made available by the Protein Data Bank (PDB), PubChem, and EMBL-EBI databases. More details regarding these substances and their biological features may be found in publications. PYMOL was used to convert the SDF file containing the ligand structures to the PDB format [45]. The variety of potential targets for drug discovery has increased for the study of the human genome. High-throughput crystallography, protein purification, and NMR spectroscopy (nuclear magnetic resonance) were all used to offer structural data on protein complexes [53, 54].  Such advances produced molecular docking also referred to as computer-assisted drug design (CAD). Molecular docking employs both ligand-based and structure-based approaches [55].  In this study, molecular docking of known organic antagonists was done to identify the most effective ligand, which may one day serve as the basis for developing drugs against psychiatric diseases caused by APOE, SNCA, and PRKN. Several algorithmic methods were developed and are frequently used for protein-ligand docking, especially if looking for drugs based on small, well-structured molecules. The objectives of molecular docking were to identify ligand-protein interactions and possible ligands [56]. Using the PyRx tool available at https://pyrx.sourceforge.io  we docked all of the chosen ligands with targeted proteins. For virtual ligand screening, the Lamarckian genetic algorithm (LGA), which integrates AutoDock and AutoDock Vina, was used [57-59]. Each ligand's top ten exclusive values were calculated, with the active parameters set to the center's (XYZ axis) grid size. The PDB files were transformed into PDBQT format using AutoDock tools, following which the binding affinities were computed [60]. For virtual screening, the 2D and 3D interactions of ligands with proteins were visualized using Discovery Studio. They displayed a docked ligand's bonding lengths, hydrogen bond interactions, hydrophobic interactions, and the size and position of its bonding sites [61, 62].

SNP analysis of the target gene

This study used the gnomAD database to observe polymorphic variants of the APOE, SNCA, and PRKN genes and examine potential risk factors. The findings showed typical mutations, 188 of which are missense, 87 synonymous regions, and 6 in pLoF were detected in the APOE gene. The data revealed that 61 are missense mutations, 31 are in the synonymous region in the SNCA gene and 310 are missense, 124 are synonymous and 14 pLoF in the PRKN gene. However, current research indicates such variations are most likely linked to neurological and mental disorders. To ascertain the importance of these changes in disease progression, analyses of function are thus required. Figure 2 shows that mutations in the APOE, SNCA, and PRKN genes of Homo sapiens have been found.

Retrieval of FASTA sequence

The selection of the APOE, SNCA, and PRKN genes for this study was based on their involvement in the initial stages of neurodegenerative diseases, specifically Alzheimer's disease and Parkinson's disease. Sequences in FASTA format were obtained from UniProt and subsequently employed for in-depth analysis. The protein sequences under evaluation are outlined in Table 1.

Gene-gene interaction

The APOE gene relates to 20 different genes, namely APOC1, LIPC, APOA1, APOA2, APOA4, APOA5, VLDLR, LDLR, APOC2, APOC3, APOB, LRP1, LRP8, SORL1, APP, LDLRAP1, CNTF, ZIC1, MAPT, and TMCC2. APOE co-expresses 17 genes, including APOC1, APOA1, APOC3, APOC2, APOA2, APOB, LIPC, APOA4, APP, APOA5, CNTF, MAPT, SORL1, LRP1, LDLR, and LRP8. APOE also shares a domain with eight other genes; VLDLR, LDLR, LRP8, LRP1, SORL1, APOA5, APOA1, and APOA4.  These data indicate that APOE and the associated genes may be functionally related and may share 14 similar biological pathways. These findings provide a foundation for further research into the biological activities of APOE and its possible involvement in related pathways, as seen in Figure 3 by showing that the SNCA gene interacts with several Synuclein families, including CHMP5, SNCAIP, MARK1, AK1, KLK6, TUBB3, SNCG, CNP, MYO5A, TPPP, PPP2CA, STX1A, SNCB, HSPA9, MT-ND5, PARK7, APCS, UCHL-1, SLC6A3, and SRCIN1. Figure 4 depicts the 14 coexpression genes and their contributions to achieving comparable activities or sharing four similar protein domains. 

Protein-Protein Interaction

 String discovered that APOE, SNCA, and PRKN proteins interact with other proteins involved in neuroglial endeavors, and Figure 6 depicts other interactions with proteins. To anticipate interactions and elucidate their molecular consequences, the STITCH and STRING databases were employed to build a protein-protein interaction network (PPIN) and to incorporate chemical compounds into the human protein network. In this network, edges represent different types of interactions, chemicals are depicted as tablets, proteins are represented as circles, and the numerical value "degree" indicates the level of activity. Larger nodes correspond to higher degree values, and as the degree value increases, the color shifts from light green to yellow. The network, comprising nodes and edges, is presented in Table 2. The network nodes act as proxies for the genes that code for the protein targets. The connections between the different nodes show how the relevant proteins interact with the bioactive substances.

Figure 6. The STITCH 5.0 displayed a chain of associations between bioactive molecules and the protein targets a) APOE, b) SNCA, and c) PRKN. The circles stand in for proteins, whereas pill-shaped nodes represent compounds.

Table 2. Protein-protein network stats by STITCH analysis

Homology protein modeling by SWISS Model

For comparative homology modeling the generated sequences were selected of at least >30% similarities and identities. The protein modeling is achieved by the SWISS model of target protein APOE, SNCA, and PRKN. The amino acid of APOE was 317, we got 15 templates with 93.06% STML ID P0DKW7.1. A (Range:1-317; coverage: 1.00), Alphafold DB model of APOE_PLEMO (Organism: Dusky titi monkey) identification for the query sequence. For the 140 amino acid of SNCA, we got 10 templates with 99.29% STML ID P61139.1.A (Range:1-140; coverage: 1.00), Alphafold DB model of SYUA_ERYPA (Organism: Red guenon) identification for the query sequence. For the 465 amino acids of PRKN, we got 50 templates with 99.01% STML ID 5c1z.1.A (Range:1-465; coverage: 0.87), model of Parkin (Organism: UbIRORBR) identification for the query sequence.  To learn how mutations drastically alter the stability of proteins, we modeled the three-dimensional structure of the APOE, SNCA and PRKN protein using the template PDB ID.  The results of using a template with quality and PyMoL to create a model are displayed in Figure 7. UCSF Chimera [43], Discovery Studio  [44], PyMOL [45] is a visualization tool of 3D structure. Validation of the modeled framework was performed by SAVES and RAMACHANDRAN plot evaluation was used to examine the secondary structure. All constraints imposed by potential energy calculations were respected by the resulting structure. In the RAMACHANDRAN plot, a significant portion of the amino acid residues within the target proteins were located within a highly favorable region, as visualized in Figure 7. You can access the detailed predicted results in Table 3.

Figure 7. Predicted models of APOE, SNCA and PRKN gene, validated by QMEAN and Ramachandran plot.

Molecular Docking by PyRx

To discover ligand-protein interactions, molecular docking was used by PyRx tool, we docked all of the selected ligands with APOE, SNCA, and PRKN proteins. They created ten distinct conformations for each ligand, which are characterized by binding affinity (-Kcal/mol).

The docking results of ligands indicate that these binding affinities relate to their level of activity, and all 10 small molecules with binding affinities are given in Table 4 and 3D and 2D representations of docking scores are visualized by Discovery Studio as shown in Figure 8.

Table 3. Statistics of Ramachandran plot and analysis of the models from Swiss model

We selected 3 compounds with strong binding affinities, notably Amphimedine, Deoxyamphimedine and Quercetin, and docked them with all 3 of our native protein complexes such as APOE, SNCA and PRKN further to investigate their interactions as depicted in Table 5.

Table 4.

A Deoxyamphimedine and quercetin ligand was fixed in the APOE binding pocket sites by forming the conventional hydrogen bond with residues GLU 106, TYR92, GLN181; and hydrophobic interactions (Alkyl, pi-alkyl, cation, anion, halogens) with ALA 109, LYS113, ARG110, GLU114, LEU177, ALA178, GLN174, TRP38, THR36, GLU37, SER40, and GLN42.  The SNCA docked with Amphimedine and Deoxyamphimedine by showing hydrogen bond with residues GLU20; as hydrophobic interactions with GLU13, ALA17, VAL16, LYS12, ILE88, ALA85, THR92, VAL95, LYS96, PHE94, and ALA89. Moreover, PRKN protein shows PHE264, HIS265, LEU272, GLU321, and GLY338 are vander wall forces; and LEU307, ARG271, LEU287, PHE277, THR270, TYR267, VAL269, LEU266, ALA339, LEU358, GLN317, LEU331, SER10, SER9, ASN8, and ARG314 residues showing hydrophobic interactions as depicted in Table 5. The interacting residues obtained from docking are presented in Figure 8. According to the score, polymorphisms not only alter the protein's conformation because of modifications to the number of interacting residues but also because of alterations in hydrogen bonds and hydrophilic associations.

Table 5. Interacting residues of APOE, SNCA and PRKN with ligands including their binding residues and hydrophobic interactions.