Skip to main content
NCBI home page
3 Biotech logoLink to 3 Biotech
. 2024 Mar 11;14(4):108. doi: 10.1007/s13205-024-03955-2

Ab initio modeling of human IRS1 protein to find novel target to dock with drug MH to mitigate T2DM diabetes by insulin signaling

Ritika Kumari Singh 1, Avinash Kumar Chaurasiya 1, Arvind Kumar 1,✉
PMCID: PMC10925585  PMID: 38476643

Abstract

IRS1 is a cytoplasmic adaptor protein that helps in cellular growth, glucose metabolism, proliferation, and differentiation. Highly disordered (insulin receptor substrate 1) IRS1 protein sequence (mol.wt- 131,590.97 da) has been used to develop model using ab initio modeling technique by I-Tassar tool and Discovery Studio/ DogSite Server to decipher a novel active site. The constructed protein model has been submitted with PMDB Id- PM0082210. GRAVY index of IRS1 model ( − 0.675) indicated surface protein–water interaction. Protparam tool instability index (75.22) demonstrated disorderedness combined with loops owing to prolines/glycines. After refinement, the Ramachandran plot showed that 88 percent of AAs were present in the allowed region and only 0.5% in the disallowed region. Novel IRS1 model protein has 10 α-helices, 22 β-sheets, 20 β-hairpins, 5 β-bulges, 47 strands, 105 β-turns, and 8 γ-turns. Docking of IRS1 with drug MH demonstrated interaction of Ser-70, Thr-18, and Pro-69 with C–H bonds; Gln-71, and Glu-113 with hydrogen bonds; while both Glu-114 and Glu-113 with salt-bridge connection. Permissible 1.0–1.5 Å range of RMSD fluctuation between 20 and 45 ns was obtained in simulation of IRS1 and IRS1-met complex confirmed that both complexes were stable during whole simulation process. RMSF result showed that except positions 57AA and 114AA, the binding of drug had no severe effects on the flexibility of the IRS1 and IRS1-met complex. The RoG value of compactness and rigidity showed little change in IRS1 protein. SASA value of IRS1 indicated non-significant fluctuation between IRS1 and drug MH means ligand (drug) and IRS1 receptor form stable structure. Hydrogen bond strength of IRS1 and IRS1-met was 81.2 and 76.4, respectively, which suggested stable interaction.

Keywords: IRS1signaling, T2DM, In silico modeling, Ab initio technique, Docking, Insulin

Introduction

IRSs (insulin receptor substrates) are key proteins among several cytoplasmic adaptor protein families that connect signaling from upstream receptors to multiple downstream effectors. IRSs have an important role in the control of normal growth, metabolism, proliferation, and differentiation of cells mediating through several signaling cascades (Du and Uversky 2017). The activation of the intrinsic tyrosine kinase activity of the Insulin receptor is initiated by the binding of the insulin hormone to its receptor IR (insulin receptor). This binding event leads to the auto-phosphorylation of the insulin receptor, as well as the phosphorylation of tyrosine residues on the IRS1protein by various kinases. Apart from IRS1, the IRS2, IRS3, and IRS4 are the main members of the IRS family including two other forms, namely IRS5 and IRS6 (Cai et al. 2003; Kim and Novak 2007). IRS1 and IRS2 are extensively expressed in several human organs that regulate glucose metabolism (White 2002). IRS4 is found predominantly in the brain, kidney, thymus, and liver cells of human beings; while, IRS3 has been reported only in rodents till now (Lavan et al. 1997; Björnholm et al. 2002).

IRSs are highly disordered proteins, due to which CDRFIT (the content of disordered residues predicted PONDR FIT) (Du and Uversky 2017) values have been utilized for evaluation and prediction of the percent disorderedness of these proteins. CDRFIT values of IRS1, IRS2, and IRS4 in percentage are 70.7, 75.6, and 64.4% respectively. These CDRFIT values significantly expressed hyper disorderedness of IRSs proteins in the KEGG (Kyoto Encyclopedia of Genes and Genomes) data set of proteins associated with T2DM (Type 2 Diabetes Mellitus) (Du and Uversky 2017). IRSs are cytoplasmic skeleton proteins that facilitate signaling between multiple receptor tyrosine kinase (RTK) pathways that include insulin growth factor1 (IGF1), insulin receptor, and SH2 domain-containing proteins (Dearth 2007; Taniguchi et al. 2006). So far, at least 50 proteins have been recorded which showed association with IRSs, together with proteins controlling tyrosine phosphorylation of IRSs by several receptor kinases. These associated proteins control stability, govern intracellular localization, and regulate insulin-like activities of the IRS proteins (Hakuno et al. 2015). The intrinsic disorderedness possibly allows IRSs to fold at binding regions as well as facilitate a stable platform to connect with several interacting partners (Du and Uversky 2017). IRS1 protein, a 131 kDa adaptor protein has 1242 AAs (IRS1—Insulin receptor substrate 1—Homo sapiens (Human)—IRS1 gene and protein www.uniprot.org. Retrieved 2016). This protein is an intermediate molecule of the RTK pathway where one domain of IRS has been found associated with trans-membrane protein acts as an insulin receptor (IR). IRS1 protein contains a single pleckstrin homology (PH) domain at the N-terminus which is associated with the phospho-tyrosine binding (PTB) domain located approximately 40 residues downstream of the PH domain region, followed by a poorly conserved C-terminus tail (Copps et al. 2012). The PH domains are involved in the attachment of proteins to membranes either directly binding to phospholipids or by protein–protein interactions. In the case of the IRS1 protein, the PH domain has both phospho-lipids binding as well as protein binding characteristics at its N-terminal (Takeuchi et al. 1998). The PTB domain of the IRS1 protein binds with the phospho-tyrosine motif found within the juxta-membrane region of the insulin receptor’s motif NPXY (asparagine, proline, any AA, and tyrosine) (Eck et al. 1996.). However, the binding specificities of these two regions are quite distinct, still; overall, they share a similar structure containing a β-sandwich formed by two nearly anti-parallel β-sheets of 4 and 3 strands, respectively. These two binding domains may act to localize IRS1 near the membrane and also to facilitate association with the insulin receptor. Further, insulin receptor causes multiple tyrosine phosphorylation of IRS1 and activate signaling pathways. IRS1 has a role in several metabolic pathways, especially in the insulin-regulated RTK pathway. Insulin (ligand) after binding with its IR, initiates the RTK signaling pathway by activating the intrinsic tyrosine kinase activity of the receptor, which leads to its auto-phosphorylation as well as phosphorylation of a series of tyrosine residues present on IRS1 proteins (Liuet al. 2004). Eventually, the phospho-tyrosine motifs present on these trans-membrane proteins recruit signaling proteins and initiate intracellular pathways such as PI3K/Akt and Erk-MAP kinase pathway that ultimately influence glucose metabolism (Fig. 1) (Dupont et al. 2009; Kim and Novak 2007). Insulin/IGF signaling may primarily that got activated in the case of HCC leading to uncontrolled growth. Therefore, the IRS1 protein biomolecule needed to be well characterized by available techniques. The X-ray crystallographic structure of IRS1 has not been yet illustrated, as well as no substantial information about its active site, catalytic site, super secondary structures (motifs), and conserved regions have been reported by any research group. Several computational tools have evidenced the usefulness for the analysis of the physiochemical and structural properties of IRS1 protein molecules.

Fig. 1.

Fig. 1

Open in a new tab

PI3K/Akt and Erk pathway (RTK-pathway) regulated by Insulin hormone in T2DM signaling cascade (Based on Kim and Novak 2007)

Type II diabetes mellitus (T2DM) is a complex metabolic disorder characterized by cellular resistance to the hormone insulin (Martín-Peláez et al. 2020). For the treatment of diabetes, metformin hydrochloride (MH) is among several recommended first-choice drugs. The World Health Organization (WHO) of the United Nations included this oral anti-diabetic in their 2010 “essential list” of diabetes treatments. MH plays a crucial role in a variety of processes, including lowering the glucose levels in the liver and increasing the cellular absorption of glucose in response to insulin. A small percentage of people (less than 1%) may experience an adverse effect after using MH, such as induction of lactic acidosis or hepatotoxicity. This drug also has an ability to improve insulin sensitivity and adenosine monophosphate-dependent kinase activity; overall, the current work would be helpful in the discovery of new drugs (ligands) that could be binding to the reported novel active site of the IRS1 protein model (Ossai et al. 2021; Polavarapu et al. 2020).

Methods

Retrieval and analysis of AAs sequence of IRS1_Homo sapiens

Insulin receptor substrate 1(IRS1) protein AAs sequence of 1242 has been retrieved from the NCBI database with accession number NP_005535.1. Human IRS1_ protein. The retrieved IRS1AAs sequence has been used for the prediction of complete sequence as well as structure analysis, in addition to its 3D structure modeling. ExPasy’sProtparam server has been used to explore the primary structure of the IRS1 protein (Gasteiger et al. 2005). Besides Protparam, the NetPhos 3.1 server has also been used to illustrate all possible phosphorylating sites (serine, threonine, or tyrosine) utilized by different kinases using an ensemble of neural networks (Blom et al. 1999).

Modeling of a robust protein structureIRS1

Modeling a 3D structure of an IRS1 protein was not possible with its native form because its available homologous template sequence showed less than 25% similarity which could not be used as a template to model this protein (Guex and Peitsch 1997). Subsequently, there was a need for an ab initio (either template-independent or threading-based) approach for the modeling of protein (Gibbs and Clarke 2001). The ab initio modeling has been done using the I-Tasser modeling tool, which is a threading-based server that uses an in-built Lomet function based on a vigorous algorithm to simultaneously identify several homologous templates to construct a robust model (Roy et al. 2010; Yang et al. 2015). IRS1_Homo sapiens protein sequence has been submitted to the I-Tasser server. To study the structural and functional aspects including its 3D structure determination, active site prediction and quality assessment was done subsequently.

Model quality assessment

To improve structural integrity and build energetically stable model protein, refinement was necessary. Model refinement has been done using the Mod-Refiner tool which is based on an algorithm and that works at the atomic level for high-resolution protein structure refinement. Refinement has been done in two steps for constructing a full-atom model from the initial C-alpha trace. In the first step; the tool builds the initial backbone of atoms, and in the second step, energy stabilization of the model was done through atomic-level energy minimization as well as by minimizing physical constraints (Xu and Zhang 2011). Further, its quality was checked on the Procheck server to obtain a Ramachandran plot to analyze the IRS1 protein model. The Procheck server was used to access the stereo-chemical quality of modeled protein by producing several PostScript plots consisting of overall and residue-by-residue geometry (Laskowski et al. 1993). Subsequently, refinement of the protein model was done manually with Swiss PDB Viewer and ModLoop server (Fiser and Sali 2003; Johansson et al. 2012). Swiss PDB Viewer is a robust Swiss tool that facilitates an interactive visual interface and gives detail information about structural motifs, makes rapid search of very large collections of structural databases for motifs with AAs present out of the core region of the IRS1 model protein. The refinement of model AAs presented out of the core region has been done using ModLoop Server which stabilizes the structure through shifting of AAs in the core region by forming loops.

Secondary structures prediction in a model protein

PDBsum server has been used to predict the stretch of AAs sequence those were involved in the formation of secondary structures like α-helices, β-sheets coils, and loops. These secondary structures illustrated structural and functional aspects as well as the robustness of the model protein (Laskowski et al 1997). PDBsum server emphasized structural constraints regarding helices turn and coils present in the IRS1 model protein (Laskowski et al. 1993).

Domain prediction of IRS1_Homo sapiens protein

An online protein domain analyzing server Pfam has been used for the annotation of the IRS1 protein. The protein sequence has been submitted to the Pfam server to get different domains. Pfam is a database of annotated protein families. Each family is described by two alignments and a profile hidden Markov model (HMM). Profile HMMs are probabilistic prototypes used for the statistical interpretation of homologs (Krogh et al. 1994) built from an aligned set of annotated family-representative sequences. A high-quality short read alignment is essential, as it specifies the basis for the position-specific amino acid occurrence, gap, and length parameters in the profile HMM (Finn et al. 2013).

Active site prediction of model protein

IRS1 protein model’s active site prediction is an important aspect that provides a binding place for the ligands/signaling molecules to initiate and activate the signaling cascade inside the cell to activate specific genes to regulate specific functions. Active site prediction also provides information for drug designing either to inhibit or activate the specific function of a particular protein (Singh and Chaube 2014). Two different tools Discovery Studio (DassaultSystèmes BIOVIA, Discovery Studio Modeling Environment, Release 2017) that performs lots of functions like simulation (of small molecule, macromolecule), ligand design, pharmacophore-modeling, and structure-based design, and DoGSiteScorer server have been used to predict active sites of the IRS1 protein model with detailed of interacting AAs. DoGSite scorer is based on a grid-fashioned function prediction method that exhausted a difference of Gaussian filter to distinguish potential binding pockets, to unfold the size, shape, and chemical descriptions of the predicted active site (Volkamer et al. 2012). The default background provides a simple draggability score for each (sub) pocket (active site), which is a linear combination of defined volume, hydrophobicity and hydrophilicity. The Draggability score limit should be between zero and one. A higher score, i.e., one, reflects a more druggable pocket (Volkamer et al. 2012; Yang et al. 2013). Discovery Studio is an assembly of software; concurrently, QSAR analysis and ADME analysis are done to predict toxicity of the binding ligands/drugs.

Prediction of ligand (drug) interacting AAs, in active site milieu

COACH server is an integral algorithm of the I-Tasser protein modeling tool, that facilitates consensus ligand binding residue to the active site of the model protein. COACH is used to predict protein–ligand interacting and binding sites. Using target protein structure COACH server generates corresponding ligand binding sites, which are based on two comparative methods TM-site and S-site to identify ligand binding templates from the BioLiP database. (Yang et al. 2013, 2012).

Preparation and docking of protein (IRS1) with ligand (MH)

Predicted overall AAs of the active site have been used for structure modeling through I-Tasser tool. The model structure was then improved using the Mod-Refiner tool to reduce the structure's energy. The three-dimensional structure of ligand, MH (Compound CID: 14,219) has been retrieved from PubChem Database in SDF format. The crystallized 3D structure of the ligand was processed using the Discovery Studio tool to commute in a separate format. The ligand (MH) and IRS1 modeled protein both were subjected to Autodock.4.2 tool to produce the docking protocol. During the preparation of the protein and ligand for docking purposes, polar hydrogens were added and then Gasteiger partial charges were allocated to all atoms. All rotatable bonds were configured to rotate when torsion angles were applied to the ligand. The protein was set to rigid mode, while the ligand was made flexible. PDBQT files were generated for the IRS1protein and ligand (MH), later used as input files for docking purposes in the next step. These files provided records of partial charges and torsion angle along with information about the coordinating atoms to build grid parameters. The binding affinities were also computed using PDBQT files. The amino acid residues present inside the active site of a protein interacting with the ligand/drug were recognized. The best position of the ligand was chosen based on its binding affinity with the different active sites. The 2D interactions between the ligand and protein have been visualized using the Discovery Studio Visualizer tool (Abdelwahab et al. 2022; Adeniji et al. 2020; Tripathi et al. 2019).

Molecular dynamic simulation

The ligand MH and IRS1 protein docked structure was subjected to MD simulations using the Web-Grow tool (Kalimuthu et al. 2021) Ligand topologies were obtained from the PRODRUG server (Van Aalten et al. 1996). To achieve neutralization, the water molecule and Na + ion were combined at the temperature of 300 K. Stability and conformational changes were obtained via the Desmond Dynamics software (Van Aalten et al. 1996) with time (Kumari and Kumar 2014). The MD simulation time was set to 100 ns for the suggested compound's root mean standard deviation (RMSD) and root mean standard fluctuation (RMSF), radius of gyration (RoG), and also for solvent accessible surface area (SASA). To determine the binding free energy of the protein–ligand complex, Molecular Mechanics Poisson-Boltzmann Surface Area (MMPBSA) analysis was used (Kumari and Kumar 2014).

Results and discussion

Various bioinformatics tools have been used to design, characterize, and decipher of structural details. The prediction of the molecular level structure of active sites and ligand/drug (MH) interacting with the active site of IRS1 protein in current research work was done for docking purposes to treat or mitigate Diabetes particularly Type II.

Sequence IRS1_Homo sapiens protein

 > NP_005535.1 insulin receptor substrate 1 [Homo sapiens] was obtained and used for further experiments.

Primary structure prediction and physiochemical analysis of IRS1_Homo sapiens protein

The basic physical parameters were studied to know the primary structure of any unknown protein by the Protparam tool supported by the Expasy platform. The Protparam tool provided details of IRS1 protein such as sequence length of 1242 AAs, molecular weight 131,590.97 Da, and Isoelectric point (pI) 8.83. pI is normally used for high-resolution protein-separating technique (Ciborowski and Silberring 2016) known as IEF (Isoelectric Focusing). The extinction coefficient of the model protein has been calculated at 280 nm by the same tool with value 86180 M−1 cm−1 which reflects the presence of sufficient numbers of aromatic tryptophan and tyrosine residues in a IRS1 model protein in a specific buffer. The lower aliphatic index value of protein (53.86) indicated that protein could not be stable at varied temperatures (Singh and Chaube 2014). The GRAVY (Grand Average of Hydropathy) index is the measure of interactions of a particular protein with water (Kyte, and Doolittle 1982). If protein would be more hydrophilic, that means its hydropathy index will be low. The GRAVY index value of IRS1 was -0.675, which is very low and indicated healthier interaction of the surface protein with water. Presence of several hydrophilic AAs on the protein surface, and a smaller number of hydrophobic AAs inside the core region prevented the formation of secondary structures of α-helices. In addition, the GRAVY index value also suggested that the IRS1 protein is cytoplasmic rather than a membrane-embedded protein (because it contains smaller number of hydrophobic AAs in the core region) and interacts with insulin receptor (IR) a trans-membrane protein (Fig. 1). IR activates two pathways, either through PI3K or MAPkinase pathway. PI3K and MAPkinase both pathways culminated into cell growth and differentiation leading to HCC via the nucleus. Protparam tool also given information about the stability of the protein in the terms of instability index which is 75.22 showing the disorderedness of the structure. IRS1 protein was constituted of a relatively higher percentage of proline and glycine AAs in addition to serine (Fig. 2) resulting in longer loop regions and restricted natively folded domains which divulge and unstable the nature of the modeled IRS1 protein (Ahuja et al. 2016). However, due to the presence of proline residues in the generated model of IRS1 protein, we got several unfolding in the IRS1 3D structure including the presence of higher number of glycine resulted in generation of more and more loops. The most prevalent AA serine made the IRS1 protein more vulnerable to phosphorylation during activation of the signaling cascade in the RTK pathway as reflected by the NetPhos 3.1 server (Blom et al. 2004). Phosphorylation resulted due to different kinases like GSK3, CDK5, cdc2, PKA, PKC PKG, CKII, RSK, etc. involved in the phosphorylation of the model protein.

Fig. 2.

Fig. 2

Open in a new tab

Percent distribution of all amino acids (AAs) present in the model IRS1 protein predicted by the Protparam tool

The phosphorylating potential values greater than 0.5 would have been considered essential for the phosphorylation of all AAs particularly serine to form cascade and generate signal. Here almost all values for serine phosphorylation are higher than threshold phosphorylating potential values (score ≥ 0.5), means phosphorylation of present serine has occurred. The significant number of serine has been phosphorylated in the predicted region between 250 and 600AAs residues. Other AAs that got phosphorylated are threonine and tyrosine with threshold values for their respective phosphorylation potential either more than or equal to 0.5(score ≥ 0.5). Maximum number of threonine and tyrosine phosphorylation have been predicted in the region of 200–600 and 440–1000 AAs residues, respectively (Fig. 3).

Fig. 3.

Fig. 3

Open in a new tab

Phosphorylating potential (score ≥ 0.5) of Serine, Threonine, and Tyrosine Amino Acids (AAs) predicted in the AAs residues region by the NetPhos 3.1 server

Secondary structure prediction

PDBsum tool has been used for secondary structure prediction of the model protein. Different secondary structures (α-helix, β-sheet, loops, β-hairpin) have been deciphered in the entire protein model (Fig. 4). IRS1 protein model is majorly constituted of β-sheet, loops, and β-hairpin but a very less percentage of helices because the ubiquitous presence of proline and glycine residues makes this protein vulnerable to full of loops and turns which is pointing hindrances toward the stability of the secondary structure (Perálvarez-Marín et al. 2008). The frequent occurrence of proline and/or glycine is a limiting factor in the native folding of all proteins by limiting helix conformations (Yun et al. 1991; Fu et al. 2009). The finding of less and small stretch (AAs less than 16 residues) of helix indicated that IRS1_Homo sapiens protein is not a trans-membrane protein.

Fig. 4.

Fig. 4

Open in a new tab

Number and location of secondary structures with α-helix, β-sheets, loops, and β-hairpins of IRS1 protein predicted by PDBsum tool

Composition of all the secondary structures present in IRS1 protein

PDBsum tool results revealed that the secondary structure of the IRS1 protein contains 10 α-helices, 22 β-sheets, 20 β-hairpins, 5 β-bulges, 47 strands, 105 β-turns, and 8 γ-turns, overall 3.2% AA constituted α-helix, 9.9% AA constituted strands, 0.4% AA constituted 3–10 normally written as 310 helix which is part of 3.2% α-helices, and 86.5% AA for others, e.g., hairpins, bulges, etc. The predicted secondary structure showed that turns and loops were major part of the IRS1 model protein construct.

Modeling of IRS1_Homo sapiens protein by I-Tasser server

IRS1 protein structure has been modeled using threading-based robust tool the I-Tasser server (Fig. 5). The I-Tasser utilized a specific matching template for each different fold of the model protein to build a complete model of a selected protein. The Lomet tool of the I-Tasser identified a sequence-specific template from the PDB library and used the identified homologous sequence as a template in the threading alignment by measuring comparable Z scores. The ten best templates were selected from Lomet’s threading algorithm. Out of ten, each template has been selected for the threading program to prepare a many folds of model to assemble the complete protein model of IRS1_Homo sapiens protein (Yang and Zhang 2015).

Fig. 5.

Fig. 5

Open in a new tab

Model of IRS1 protein predicted using the I-Tasser tool along with its Ramachandran plot, showing 88.0% AAs in the favored region; 10.4% AAs in the allowed region and 1.1% AAs in generously allowed regions

Model assessment of IRS1_Homo sapiens protein

The prepared protein model has been used for further analysis based on energy minimization value followed by an evaluation of the quality. Energy minimization provides a vision toward the stability of the model protein considering all AAs residues and existing side chains for further refinement. The refinement of the model has been done using the MODREFINER tool. The structures acquired after the energy minimization and refinement were very close to its native structure. For quality validation of the modeled protein, the Procheck server has been used to acquire data from the Ramachandran plot. The data obtained reflected the overall quality of the modeled protein. Initially, 62.4% of AAs were present in the favored region, 33.1% of AAs were in the allowed region, and the remaining 4.4% of AAs were in the disallowed region (figure not given). The data indicated that the model has not been yet stereo-chemically stable and hence some loops were amalgamated using a Mod-loop server. The Swiss PDB Viewer was used to evaluate the shifting of AAs from its disallowed region to the favored region after loop modeling of the prepared model. Again, the Procheck server has been exercised to mark the location of all the AAs in the Ramachandran plot to assess and confirm model quality (Laskowski et al. 1993). The Ramachandran plot data obtained after refinement reflected 88.0% AAs in the favored region, 10.4% AAs in the allowed region, 1.1% AAs in the generously allowed region and only 0.5% AAs in the disallowed region (Yang et al. 2015) after removal of physical constraints. The prepared protein model has been submitted to PMDB (Protein modeling database) with Id–PM0082210, for further in silico experiments based on earlier reports (Fig. 5) (Castrignanoet al. 2006).

Topology of the predicted active site in IRS1 protein

The topology of the IRS1 protein has been predicted with the PDBsum server. The protein showed a smaller number of helices, lots of β-sheets and loops that required to stabilize the structure (Fig. 6) as reported by scientific groups (Fu et al. 2009). There were total 10 small helices followed by some AAs sequence that span other present helices. In between these helices, several β-sheets were connected with β-strands and loops. The predicted active site range was between approximately 472 and 640 AAs consisting of one helix (472–478 AAs) and five β-sheets (482–484, 488–492, 498–500, 507–511, and 524–526 AAs) showed in the topological view. Secondary structures and, respectively, the larger number of loops were engaged actively to form active pockets to facilitate interaction with the ligands. Loops provide and create ridges and bulges those are involved in the formation of stereo-chemical space to facilitate significant flexibility in the ligand binding pocket (Venkitakrishnanet al. 2004).

Fig. 6.

Fig. 6

Open in a new tab

Topological views with folding dynamics of model IRS1 protein showing helices, β-sheets and loops predicted by PDBsum server

Different domains present in IRS1 protein

Two main domains of IRS1 protein have been found with curative server CDART (Conserve Domain Architecture Retrieval Tool).The first domain is PH domain envelope which exists between 7th AA and 114th AA and the second is the PTB domain envelope which exist between 160th AA and 262nd (Fig. 7) (PDB ID-1QQG). Active sites of IRS1 have not been predicted yet by the curative server. However, the structural component with a short peptide (891AAs–902AAs) concomitant to the GRB2-binding region which was complexed with the Insulin like growth factor 1 (IGF1) receptor (PDB ID-1K3A), and an additional short peptide (731AAs–736 AAs) complexed with the Insulin Receptor (IR) (PDB ID-2Z8C) has been resolved (Marchler-Bauer et al. 2016; Du and Uversky 2017).

Fig. 7.

Fig. 7

Open in a new tab

Two functional domains PH (7 to 114 AAs) and PTB (160 to 262 AAs) of the IRS1 protein predicted by the CDART tool

Active site prediction using two robust tools: DoGSite server and discovery studio

The promising binding pockets and sub-pockets have been predicted by DoGSite Scorer Server and Discovery Studio tool. These pockets would be further analyzed using their geometrical and physicochemical properties. We used DoGSite Scorer Server, used earlier by Sahu et al. (2017) and predicted 36 pockets in the IRS1 protein model. The predicted key binding pocket P0 has been selected as a potential active site for the binding of ligand (8-A, B, and C) amid all predicted binding pocket. Explanations for AA compositions present in this active site have been specified in terms of their a-polar, polar, positive, and negative AAs ratio which is 0.29, 0.44, 0.20, and 0.07, respectively. The binding site contains AAs residue: Ala, Arg, Asp, Gln, Glu, Gly, His, Ile, Leu, Lys, Met, Phe, Pro, Ser, Thr, Tyr, and Val in the modeled IRS1 protein. The residue range of this active site is approximately 472nd–690th AAs long (Fig. 8-A, B, C and A1, B1, C1) predicted by both of the visualization tools respectively.

Fig. 8.

Fig. 8

Open in a new tab

Different resolutions of the active pocket (active site of 472 to 690 AAs) with largest cavity A, B and C, and A1, B1 and C1, predicted respectively by the DoGsite server and Discovery Studio tool

Predicted active site residues

Active site residues predicted by the DoGsite server and Discovery Studio tool as well found almost similar. All important AAs that are normally involved in shaping of the active sites were deciphered of the modeled IRS1 protein. Though proline is ubiquitously present in IRS1 protein, it is also a part of the predicted active site (Singh and Chaube 2014). This remarkable predicted active site would be further used for other potential drugs. The AAs present in the predicted active site are as follows:

ARG_524, THR_525, HIS_526, SER_527, GLY_529, THR_530, SER_531, PRO_532, ILE_534, THR_535, HIS_536, GLN_537, THR_539, PRO_540, GLN_542, SER_543, SER_544, VAL_545, ALA_546, SER_547, ILE_548, GLU_549, GLU_550, TYR_551, PRO_556, ALA_557, TYR_558, PRO_559, GLY_563, SER_564, GLY_565, GLY_566, ARG_567, LEU_568, PRO_569, GLY_570, HIS_571, ARG_572, HIS_573, SER_574, PHE_576, VAL_577, PRO_578, THR_579, ARG_580, SER_581, PRO_583, GLU_585, GLU_588, ARG_594, GLY_597, HIS_598, HIS_599, ARG_600, PRO_601, ASP_602, SER_603, SER_604, THR_605, LEU_606, HIS_607, THR_608, ASP_610, GLY_611, TYR_612, VAL_622, PRO_623, SER_624, GLY_625, ARG_626, LYS_627, GLY_628, SER_629, MET_635, SER_636, PRO_637, LYS_638, VAL_640, HIS_590.

Further COACH server is used to anticipate consensus ligand binding residues residing in the predicted active site. Highlighted AAs residues (525, 528, 538, 542, 557, 559, 564, 565, 566, 568, 570, 571, 580, 581, 603, and 608) were showing interaction with different ligands predicted by the COACH server Fig. 9 (Ciborowski and Silberring 2016; Yang et al. 2012). Some AAs were not the part of active site but protruded toward the active site were sterically stable to interact with the ligands. These residues are 472_Glycine, 476_Leucine, and 538_Lysine (Fig. 9).

Fig. 9.

Fig. 9

Open in a new tab

Stretch of ligand binding residues in blue and magenta colours showing AAs interacting with drug in predicted active site region

Docking results

The goal of the current in silico study was to use computational analysis to comprehend the molecular relationship between ligand MH and active site of model IRS1 protein. The strong molecular interaction between the AA constituents of MH and the human IRS1 protein has not been reported in detail. The widespread usage of MH in case of T2DM disease shows the significance of this computational work needed to be done for their interaction details.

All binding parameters of MH have been obtained after docking with the active site of IRS1 protein model in Auto-Dock 4.2 tool. The predicted highest free energy of binding of the ligand was –5.07 kcal/mol, while the estimated cluster RMSD was 0.00 (zero). Estimated inhibition constant Ki was 193.31 µM (micromolar) and temperature obtained was 298.15 K. Energy value of different interactions Van der Waals forces, Hydrogen bond, dissolved Energy was -2.93 kcal/mol and electrostatic Energy was -2.13 kcal/mol (Sahu et al. 2017; Khoba et al. 2023; Corbo et al. 2022) (Fig. 10-A, B, and C). The docked structure demonstrates SER-70, THR-68, and PRO-AAs that interact with their ligands by carbon–hydrogen bonds; in contrast to GLN-71 and GLU-113 AAs that interact with their ligands by typical hydrogen bonds. The AAs GLU-114 and GLU-113 both interact with the ligand (MH) by a salt-bridge connection (Fig. 10-D).

Fig. 10.

Fig. 10

Open in a new tab

A. Docking pose of ligand (MH) with IRS1 protein in ribbon form, B. Docking pose in sphere form, C. All types of interacting bonds along with their length and D. all interacting bonds and interacting amino acids

MD simulations studies

Although protein–ligand docking was widely used and had significant applications in drug discovery and therapeutic relevance, it provided a static and interactive view of the ligand's binding pose in the receptor's active site. MD is an effective computational tool for understanding the physical basis of biological macromolecule structure and function since it deals with atomic-level interactions. The receptor–ligand complexes formed from the docking experiments used for MD simulations up to timing 100 ns (nano seconds), and later the complexes were analyzed to know their stability and fluctuations. With the aid of the RMSD (root mean standard deviation), RMSF (root mean square fluctuation), RoG (radius of gyration), and SASA (solvent accessible surface area) of receptor atoms, MD trajectories analysis was carried out to determine the stability and fluctuations of these complexes (Lindorff‐Larsen et al. 2010; Vishvakarma et al. 2022; Aljarbaet al. 2022; Osigbemhe et al. 2022; Beema Shafreen et al. 2022).

Evaluation of the stability of the complexes using RMSD computation is one of the refine parameters in the molecular dynamics simulation program. The RMSD values of protein backbone atoms in angstrom (Å) versus time in nano-second (ns) were used to build a graph to get data to evaluate how the protein's structural conformations altered over time. If the fluctuation in RMSD values of IRS1 and IRS1-met lies in the permissible range of 1–1.5 Å which appears less than 2 Å (< 2 Å) between 20 and 45 ns. This trajectory imaging demonstrated that the IRS1 and IRS1–metprotein complex finds equilibrium or homology after 45 ns till the end of the simulation run. Non-significant deviation was observed after that. The non-significant conformational changes in RMSD values between the starting point to the end point of the simulation run were observed. Hence, one can conclude that up to 100 ns as indicated, the structure of the IRS1 protein was stabilized during the simulation process (Fig. 11) (Das et al. 2022).

Fig. 11.

Fig. 11

Open in a new tab

Study plot of RMSD (Å) of IRS1 and IRS1-met complexes against time up to 100 ns by MD simulation

The flexibility of the target protein's AA residues upon binding to a ligand was assessed using RMSF analysis for IRS1 and IRS1-met complex protein form. The RMSF values for the protein's C-alpha atoms were calculated and displayed against the AA residues (Fig. 12). During the simulation, the AA residues in the IRS1 protein complex showed modest alterations at positions 35, 53, 63, 92, and 124 and considerable variance at position 129. The best part of the simulation is that, except for positions 57 and 114, the binding of the ligands had no severe effect on the flexibility of the protein residues (Fig. 12) (Das et al. 2022).

Fig. 12.

Fig. 12

Open in a new tab

Study plot of RMSF (Å) of IRS1 and IRS1-met complexes against the AAs residues by MD simulation

Additionally, the RoG (radius of gyration analysis) of the complexes was examined. The RoG stands for the root mean square distance between the protein atoms and the rotational axis. It is one of the useful structural factors for examining the protein's compactness and rigidity during the simulation. During the simulation period, the RoG values of the amino acid backbone were plotted against time (Fig. 13-A). In terms of protein compactness and rigidity, IRS1 showed a little change, but the IRS1-met protein complex had no noticeable effect on the protein's structural integrity, as shown by the results (Das et al. 2022).

Fig. 13.

Fig. 13

Open in a new tab

A Radius of gyration (RoG) analysis in Å, B Solvent accessible surface area (SASA) analysis in Å, of IRS1 and IRS1-met complexes against time (nano-seconds) by MD simulation

Both complexes were analyzed for their solvent accessible surface area (SASA). SASA is a significant parameter for determining how much of the receptor is exposed to the surrounding solvent molecules during simulation for binding or interacting with the ligand. If the ligand causes structural changes in the receptor, the area that comes into contact with the solvent will also change. Protein surface area variations were estimated using SASA values displayed against time (nano-second). Despite IRS1 has having very little variation, the trajectory for SASA analysis indicated no significant fluctuation in the IRS1-met protein complex, implying that ligand binding to the receptor forms a more stable structure (Fig. 13-B) (Kalimuthu et al. 2021).

We can learn about protein stability and conformation by doing the study of the formation of hydrogen bonds. Hydrogen bonds are critical for understanding protein structural integrity, the catalytic region, and protein–ligand interactions. We observed that the hydrogen bond interactions in the proteins of both complexes have not changed significantly (Fig. 14). The IRS1 and IRS1-met complexes have average hydrogen bond strengths of 81.2 and 76.4, respectively. Hydrogen bonds between protein and ligand complex structures were also counted on pairs close to each other. The results demonstrated that ligands could generate hydrogen bonds with a high frequency. In addition, the non-covalent interactions should have also played a role in the binding of molecules (Gorai et al. 2022).

Fig. 14.

Fig. 14

Open in a new tab

Comparative analysis of intermolecular hydrogen bond numbers of IRS1 and IRS1-met complex against time (nano-seconds) by MD simulation

Conclusion

Current work is focused on the computational analysis of the highly intrinsically disordered IRS1 protein, which is involved in T2DM disease genesis and could be targeted to mitigate or cure the disease. Many scientific groups are trying to find a connection between two diseases T2DM and HCC as we hypothesized that they are generated by the activation of similar pathways and interconnected to each other. In the present work, we observed one signaling molecule that regulate both of these disease conditions through a common IRS1 protein mediated signal transduction pathway as reflected in the Fig. 1. This shows the significance and the most important characteristic feature of the IRS1 protein molecule in the genesis of T2DM. To achieve the main focus of the current study, human IRS1 protein modeling has been done to generate a reliable and robust structure with an emphasis on several aspects of the protein's model stability. Further, active site prediction has been done and consequently the active site was used for docking purposes with ligand MH. Getting human IRS1protein stable model was of prime importance because its full-length model was not available to access for docking with branded anti-diabetic drugs (like MH). Here we reported the first ab initio model of the full length of human IRS1 protein with greater stability in terms of a smaller non-significant percentage of AA were present in the disallowed region along with deciphering of suitable novel active sites with AAs details. A current modeling technique which is ab initio modeling technique differs from the traditional homology modeling based on similarity of the available templates. Current research demonstrated that the IRS1 protein has a lot of potential phosphorylating AAs such as serine and threonine, and tyrosine along with plenty of kink and loop created by proline and glycine AA residues, respectively. Distortions (loop) present in the model protein IRS1, facilitates the association of various important T2DM signaling proteins with IRS1 target protein to signify IRS1 role in the transmission of Insulin signals. A better understanding of the structure provides a clear vision to make IRS1 protein as a target molecule to dock with MH to treat a metabolically disordered disease T2DM.The molecular docking studies show that when drug MH is combined with the receptor molecule IRS1, it improves the chances of relaying the insulin signaling cascade and makes IRS1 a better target molecule for treating T2DM disease. In addition, the protein–ligand complex's stability and resilience can be gleaned from the several results of MD simulations. Permissible RMSD values are less than < 2 Å (approx 1 to 1.5 Å) which gives the impression that the structure is toward stable conformation as shown in Fig. 11 (Vijayan et al. 2022). Although the target site of MH had been predicted earlier working groups, but it is still important to look for another better suitable target sites in the IRS1 protein that could be targeted to find a comparatively better cure options for T2DM diseases, which is still a worldwide threat. Current research deciphers possible target site prediction on adapter protein IRS1 which docks with authenticated anti-diabetic drug MH normally used to treat T2DM. To find out how well MH works against T2DM targeting the IRS1 protein, more in vitro and in vivo animal model research work required to be done.

Acknowledgements

The financial support to the School of Biotechnology, Institute of Science from DBT, DBT-SAP, UGC-UPE, and DST-PURSE program of Govt. of India, New Delhi, is duly acknowledged.

Data Availability

Not available.

Declarations

Conflict of interest

There is no financial and/or non-financial conflict of interests including authorship sequence among the authors.

Ethical approval

Experiments were performed in silico conditions and no animals/humans were required.

References

  1. Abdelwahab SI, Farasani A, Jerah A, Elhassan Taha MM, Bidwai A. Molecular docking of amphetamine, cathine and cathinone with dihydrofolate reductase: a computational analysis of inhibition of dihydrofolate reductase by khat alkaloids. Toxicol Commun. 2022;4(2):8. doi: 10.53388/2022020208. [DOI] [Google Scholar]
  2. Adeniji SE, Arthur DE, Oluwaseye A. Computational modeling of 4-phenoxynicotinamide and 4-phenoxypyrimidine-5-carboxamide derivatives as potent anti-diabetic agent against TGR5 receptor. J King Saud Univ-Sci. 2020;32(1):102–115. doi: 10.1016/j.jksus.2018.03.007. [DOI] [Google Scholar]
  3. Ahuja P, Cantrelle FX, Huvent I, Hanoulle X, Lopez J, Smet C, Wieruszeski JM, Landrieu I, Lippens G. Proline conformation in a functional tau fragment. J Mol Biol. 2016;428(1):79–91. doi: 10.1016/j.jmb.2015.11.023. [DOI] [PubMed] [Google Scholar]
  4. Aljarba NH, Hasnain MS, Bin-Meferij MM, Alkahtani S. An In-silico investigation of potential natural polyphenols for the targeting of COVID main protease inhibitor. J King Saud Univ-Sci. 2022 doi: 10.1016/j.jksus.2022.102214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Beema Shafreen RM, Seema S, Alagu Lakshmi S, Srivathsan A, Tamilmuhilan K, Shrestha A, Balasubramanian B, Dhandapani R, Paramasivam R, Al Obaid S, Salmen SH. In vitro and in vivo antibiofilm potential of eicosane against Candida albicans. Appl Biochem Biotechnol. 2022 doi: 10.1007/s12010-022-03984-8. [DOI] [PubMed] [Google Scholar]
  6. Björnholm M, He A, Attersand A, Lake S, Liu S, Lienhard G, Taylor S, Arner P, Zierath J. Absence of functional insulin receptor substrate-3 (IRS-3) gene in humans. Diabetologia. 2002;45:1697–1702. doi: 10.1007/s00125-002-0945-z. [DOI] [PubMed] [Google Scholar]
  7. Blom N, Gammeltoft S, Brunak S. Sequence and structure-based prediction of eukaryotic protein phosphorylation sites. J Mol Biol. 1999;294(5):1351–1362. doi: 10.1006/jmbi.1999.3310. [DOI] [PubMed] [Google Scholar]
  8. Blom N, Sicheritz-Pontén T, Gupta R, Gammeltoft S, Brunak S. Prediction of post-translational glycosylation and phosphorylation of proteins from the amino acid sequence. Proteomics. 2004;4(6):1633–1649. doi: 10.1002/pmic.200300771. [DOI] [PubMed] [Google Scholar]
  9. Cai D, Dhe-Paganon S, Melendez PA, Lee J, Shoelson SE. Two new substrates in insulin signaling, IRS5/DOK4 and IRS6/DOK5. J Biol Chem. 2003;278(28):25323–25330. doi: 10.1074/jbc.M212430200. [DOI] [PubMed] [Google Scholar]
  10. Castrignano T, De Meo PDO, Cozzetto D, Talamo IG, Tramontano A. The PMDB protei n model database. Nucl Acids Res . 2006;34(1):D306–D309. doi: 10.1093/nar/gkj105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Ciborowski P, Silberring J, editors. Proteomic profiling and analytical chemistry: the crossroads. Elsevier; 2016. [Google Scholar]
  12. Copps KD, White MF. Regulation of insulin sensitivity by serine/threonine phosphorylation of insulin receptor substrate proteins IRS1 and IRS2. Diabetologia. 2012;55:2565–2582. doi: 10.1007/s00125-012-2644-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Corbo T, Kalajdzic A, Delic D, Suleiman S, Pojskic N. In silico prediction suggests inhibitory effect of halogenated boroxine on human catalase and carbonic anhydrase. J Genet Eng Biotechnol. 2022;20(1):1–11. doi: 10.1186/s43141-022-00437-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Das C, Das D, Mattaparthi VSK. Computational Investigation on the efficiency of small molecule inhibitors identified from indian spices against SARS-CoV-2 Mpro. Biointerface Resin Appl Chem. 2022 doi: 10.33263/BRIAC133.235. [DOI] [Google Scholar]
  15. DassaultSystèmes BIOVIA, Discovery studio modeling environment, release 2017, San Diego: DassaultSystèmes, 2016
  16. Dearth RK, Cui X, Kim HJ, Hadsell DL, Lee AV. Oncogenic transformation by the signaling adaptor proteins insulin receptor substrate (IRS)-1 and IRS-2. Cell Cycle. 2007;6(6):705–713. doi: 10.4161/cc.6.6.4035. [DOI] [PubMed] [Google Scholar]
  17. Du Z, Uversky VN. A comprehensive survey of the roles of highly disordered proteins in type 2 diabetes. Int J Mol Sci. 2017;18(10):2010. doi: 10.3390/ijms18102010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Dupont J, Tesseraud S, Simon J. Insulin signaling in chicken liver and muscle. Gen Comp Endocrinol. 2009;163(1–2):52–57. doi: 10.1016/j.ygcen.2008.10.016. [DOI] [PubMed] [Google Scholar]
  19. Eck MJ, Dhe-Paganon S, Trüb T, Nolte RT, Shoelson SE. Structure of the IRS-1 PTB domain bound to the juxtamembrane region of the insulin receptor. Cell. 1996;85(5):695–705. doi: 10.1016/S0092-8674(00)81236-2. [DOI] [PubMed] [Google Scholar]
  20. Finn RD, Bateman A, Clements J, Coggill P, Eberhardt RY, Eddy SR, Heger A, Hetherington K, Holm L, Mistry J, Sonnhammer EL. Pfam: the protein families database. Nucl Acids Res. 2013;42(D1):D222–D230. doi: 10.1093/nar/gkt1223. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Fiser A, Sali A. ModLoop: automated modeling of loops in protein structures. Bioinformatics. 2003;19(18):2500–2501. doi: 10.1093/bioinformatics/btg362. [DOI] [PubMed] [Google Scholar]
  22. Fu H, Grimsley GR, Razvi A, Scholtz JM, Pace CN. Increasing protein stability by improving β turns. Proteins Struct Funct Bioinform. 2009;77(3):491–498. doi: 10.1002/prot.22509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Gasteiger E, Hoogland C, Gattiker A, Wilkins MR, Appel RD, Bairoch A. Protein identification and analysis tools on the ExPASy server. In: Walker JM, editor. The proteomics protocols handbook. Humana Press; 2005. pp. 571–607. [Google Scholar]
  24. Gibbs N, Clarke AR, Sessions RB. Ab initio protein structure prediction using physicochemical potentials and a simplified off-lattice model. Proteins Struct Funct Genet. 2001;43(2):186–202. doi: 10.1002/1097-0134(20010501)43:2&#x0003c;186::AID-PROT1030&#x0003e;3.0.CO;2-L. [DOI] [PubMed] [Google Scholar]
  25. Gorai S, Junghare V, Kundu K, Gharui S, Kumar M, Patro BS, Nayak SK, Hazra S, Mula S. Synthesis of Dihydrobenzofuro [3, 2-b] chromenes as potential 3CLpro Inhibitors of SARS-CoV-2: a molecular docking and molecular dynamics study. ChemMedChem. 2022;17(8):e202100782. doi: 10.1002/cmdc.202100782. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Guex N, Peitsch MC. SWISS-MODEL and the Swiss-Pdb Viewer: an environment for comparative protein modeling. Electrophoresis. 1997;18(15):2714–2723. doi: 10.1002/elps.1150181505. [DOI] [PubMed] [Google Scholar]
  27. Hakuno F, Fukushima T, Yoneyama Y, Kamei H, Ozoe A, Yoshihara H, Yamanaka D, Shibano T, Sone-Yonezawa M, Yu BC, et al. The novel functions of high-molecular-mass complexes containing insulin receptor substrates in mediation and modulation of insulin-like activities: emerging concept of diverse functions by IRS-associated proteins. Front Endocrinol (lausanne) 2015;6:73. doi: 10.3389/fendo.2015.00073. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. IRS1-Insulin receptor substrate 1 - Homo Sapiens (Human) - IRS1 gene & protein. www.uniprot.org. Retrieved 2016–04–21
  29. Johansson MU, Zoete V, Michielin O, Guex N. Defining and searching for structural motifs using DeepView/Swiss-PdbViewer. BMC Bioinform. 2012;13(1):173. doi: 10.1186/1471-2105-13-173. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kalimuthu AK, Panneerselvam T, Pavadai P, Pandian SRK, Sundar K, Murugesan S, Ammunje DN, Kumar S, Arunachalam S, Kunjiappan S. Pharmacoinformatics-based investigation of bioactive compounds of Rasam(South Indian recipe) against human cancer. Sci Rep. 2021;11(1):1–19. doi: 10.1038/s41598-021-01008-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Khoba K, Kumar S, Chatterjee S, Purty RS. Isolation, characterization, and in silico interaction studies of bioactive compounds from Caesalpinia bonducella with target proteins involved in Alzheimer’s disease. Appl Biochem Biotechnol. 2023 doi: 10.1007/s12010-022-03937-1. [DOI] [PubMed] [Google Scholar]
  32. Kim SK, Novak RF. The role of intracellular signaling in insulin-mediated regulation of drugmetabolizing enzyme gene and protein expression. Pharmacol Ther. 2007;113(1):88–120. doi: 10.1016/j.pharmthera.2006.07.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Krogh A, Brown M, Mian IS, Sjölander K, Haussler D. Hidden Markov models in computational biology: applications to protein modeling. J Mol Biol. 1994;235(5):1501–1531. doi: 10.1006/jmbi.1994.1104. [DOI] [PubMed] [Google Scholar]
  34. Kumari R, Kumar R, Lynn A. Open-source drug discovery. Lynn J Chem Inf Model. 2014;54(1951):10–1021. doi: 10.1021/ci500020m. [DOI] [Google Scholar]
  35. Kyte J, Doolittle RF. A simple method for displaying the hydropathic character of a protein. J Mol Biol. 1982;157(1):105–132. doi: 10.1016/0022-2836(82)90515-0. [DOI] [PubMed] [Google Scholar]
  36. Laskowski RA, MacArthur MW, Moss DS, Thornton JM. PROCHECK: a program to check the stereochemical quality of protein structures. J Appl Crystallogr. 1993;26(2):283–291. doi: 10.1107/S0021889892009944. [DOI] [Google Scholar]
  37. Laskowski RA, Hutchinson EG, Michie AD, Wallace AC, Jones ML, Thornton JM. PDBsum: a web-based database of summaries and analyses of all PDB structures. Trends Biochem Sci. 1997;22(12):488–490. doi: 10.1016/S0968-0004(97)01140-7. [DOI] [PubMed] [Google Scholar]
  38. Lavan BE, Fantin VR, Chang ET, Lane WS, Keller SR, Lienhard GE. A novel 160-kDa phosphotyrosine protein in insulin-treated embryonic kidney cells is a new member of the insulin receptor substrate family. J Biol Chem. 1997;272(34):21403–21407. doi: 10.1074/jbc.272.34.21403. [DOI] [PubMed] [Google Scholar]
  39. Lindorff-Larsen K, Piana S, Palmo K, Maragakis P, Klepeis JL, Dror RO, Shaw DE. Improved side-chain torsion potentials for the Amber ff99SB protein force field. Proteins Struct Funct Bioinform. 2010;78(8):1950–1958. doi: 10.1002/prot.22711. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Liu YF, Herschkovitz A, Boura-Halfon S, Ronen D, Paz K, LeRoith D, Zick Y. Serine phosphorylation proximal to its phosphotyrosine binding domain inhibits insulin receptor substrate 1 function and promotes insulin resistance. Mol Cell Biol. 2004;24(21):9668–9681. doi: 10.1128/MCB.24.21.9668-9681.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Marchler-Bauer A, Bo Y, Han L, He J, Lanczycki CJ, Lu S, Chitsaz F, Derbyshire MK, Geer RC, Gonzales NR, Gwadz M. CDD/SPARCLE: functional classification of proteins via subfamily domain architectures. Nucl Acids Res. 2016;45(D1):D200–D203. doi: 10.1093/nar/gkw1129. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Martín-Peláez S, Fito M, Castaner O. Mediterranean diet effects on type 2 diabetes prevention, disease progression, and related mechanisms. A review. Nutrients. 2020;12(8):2236. doi: 10.3390/nu12082236. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Osigbemhe IG, Louis H, Khan EM, Etim EE, Oyo-Ita EE, Oviawe AP, Edet HO, Obuye F. Antibacterial potential of 2-(-(2-Hydroxyphenyl)-methylidene)-amino) nicotinic acid: experimental, DFT studies, and molecular docking approach. Appl Biochem Biotechnol. 2022;194(12):5680–5701. doi: 10.1007/s12010-022-04054-9. [DOI] [PubMed] [Google Scholar]
  44. Ossai EC, Madueke AC, Amadi BE, Ogugofor MO, Momoh AM, Okpala COR, Anosike CA, Njoku OU. Potential enhancement of metformin hydrochloride in lipid vesicles targeting therapeutic efficacy in diabetic treatment. Int J Molecular Sci. 2021;22(6):2852. doi: 10.3390/ijms22062852. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Perálvarez-Marín A, Lórenz-Fonfría VA, Simón-Vázquez R, Gomariz M, Meseguer I, Querol E, Padrós E. Influence of proline on the thermostability of the active site and membrane arrangement of transmembrane proteins. Biophys J. 2008;95(9):4384–4395. doi: 10.1529/biophysj.108.136747. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Polavarapu NK, Kale R, Sethi B, Sahay RK, Phadke U, Ramakrishnan S, Mane A, Mehta S, Shah S. Effect of gliclazide or gliclazide plus metformin combination on glycemic control in patients with T2DM in India: a real-world, retrospective, longitudinal, observational study from electronic medical records. Drugs-Real World Outcomes. 2020;7(4):271–279. doi: 10.1007/s40801-020-00206-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Roy A, Kucukural A, Zhang Y. I-TASSER: a unified platform for automated protein structure and function prediction. Nat Protoc. 2010;5(4):725. doi: 10.1038/nprot.2010.5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Sahu A, Patra PK, Yadav MK, Varma M. Identification and characterization of ErbB4 kinase inhibitors for effective breast cancer therapy. J Receptors Signal Transd. 2017;37(5):470–480. doi: 10.1080/10799893.2017.1342129. [DOI] [PubMed] [Google Scholar]
  49. Singh A, Chaube R. Bioinformatic analysis, structure modeling, and active site prediction of aquaporin protein from Catfish Heteropneustes fossilis. Int J Recent Innov Trends Comput Commun. 2014;2(10):3208–3215. [Google Scholar]
  50. Takeuchi H, Matsuda M, Yamamoto TA, Kanematsu T, Kikkawa U, Yagisawa H, Watanabe Y, Hirata M. PTB domain of insulin receptor substrate-1 binds inositol compounds. Biochem J. 1998;334(1):211–218. doi: 10.1042/bj3340211. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Taniguchi CM, Emanuelli B, Kahn CR. Critical nodes in signaling pathways: Insights into insulin action. Nat Rev Mol Cell Biol. 2006;7(2):85–96. doi: 10.1038/nrm1837. [DOI] [PubMed] [Google Scholar]
  52. Tripathi A, Shrinet K, Singh VK, Kumar A. Molecular modelling and docking of Mus musculus HMGB1 inflammatory protein with CGA. Bioinformation. 2019;15(7):467–473. doi: 10.6026/97320630015467. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Van Aalten DM, Bywater R, Findlay JB, Hendlich M, Hooft RW, Vriend G. PRODRG, a program for generating molecular topologies and unique molecular descriptors from coordinates of small molecules. J Comput Aided Mol Des. 1996;10(3):255–262. doi: 10.1007/BF00355047. [DOI] [PubMed] [Google Scholar]
  54. Venkitakrishnan RP, Zaborowski E, McElheny D, Benkovic SJ, Dyson HJ, Wright PE. Conformational changes in the active site loops of dihydrofolate reductase during the catalytic cycle. Biochemistry. 2004;43(51):16046–16055. doi: 10.1021/bi048119y. [DOI] [PubMed] [Google Scholar]
  55. Vijayan S, Loganathan C, Sakayanathan P, Thayumanavan P. Synthesis and characterization of plumbagin S-allyl cysteine ester: determination of anticancer activity in-silico and in vitro. Appl Biochem Biotechnol. 2022;194(12):5827–5847. doi: 10.1007/s12010-022-04079-0. [DOI] [PubMed] [Google Scholar]
  56. Vishvakarma VK, Singh MB, Jain P, Kumari K, Singh P. Hunting the main protease of SARS-CoV-2 by plitidepsin: molecular docking and temperature-dependent molecular dynamics simulations. Amino Acids. 2022;54(2):205–213. doi: 10.1007/s00726-021-03098-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Volkamer A, Kuhn D, Rippmann F, Rarey M. DoGSiteScorer: a web server for automatic binding site prediction, analysis, and druggability assessment. Bioinformatics. 2012;28(15):2074–2075. doi: 10.1093/bioinformatics/bts310. [DOI] [PubMed] [Google Scholar]
  58. White MF. IRS proteins and the common path to diabetes. Am J Phys Endocrinol Metab. 2002;283(3):E413–E422. doi: 10.1152/ajpendo.00514.2001. [DOI] [PubMed] [Google Scholar]
  59. Xu D, Zhang Y. Improving protein models' physical realism and structural accuracy by a two-step atomic-level energy minimization. Biophys J. 2011;101(10):2525–2534. doi: 10.1016/j.bpj.2011.10.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Yang J, Zhang Y. I-TASSER server: new development for protein structure and function predictions. Nucl Acids Res. 2015;43(W1):W174–W181. doi: 10.1093/nar/gkv342. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Yang J, Roy A, Zhang Y. BioLiP: a semi-manually curated database for biologically relevant ligand–protein interactions. Nucl Acids Res. 2012;41(D1):D1096–D1103. doi: 10.1093/nar/gks966. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Yang J, Roy A, Zhang Y. Protein-ligand binding site recognition using complementary binding-specific substructure comparison and sequence profile alignment. Bioinformatics. 2013;29(20):2588–2595. doi: 10.1093/bioinformatics/btt447. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Yang J, Yan R, Roy A, Xu D, Poisson J, Zhang Y. The I-TASSER Suite: protein structure and function prediction. Nat Methods. 2015;12(1):7. doi: 10.1038/nmeth.3213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Yun RH, Anderson A, Hermans J. Proline in α helix: Stability and conformation studied by dynamics simulation. Proteins Struct Funct Bioinform. 1991;10(3):219–228. doi: 10.1002/prot.340100306. [DOI] [PubMed] [Google Scholar]
  65. Du Z, Uversky VN. A comprehensive survey of the roles of highly disordered proteins in type 2 diabetes. Int J Mol Sci. 2017;18(10):2010. doi: 10.3390/ijms18102010. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Not available.

Articles from 3 Biotech are provided here courtesy of Springer

RESOURCES