https://orcid.org/0000-0003-0473-7803
Figures
Abstract
GRIN2A is a gene that encodes NMDA receptors found in the central nervous system and plays a pivotal role in excitatory synaptic transmission, plasticity and excitotoxicity in the mammalian central nervous system. Changes in this gene have been associated with a spectrum of neurodevelopmental disorders such as epilepsy. Previous studies on GRIN2A suggest that non-synonymous single nucleotide polymorphisms (nsSNPs) can alter the protein’s structure and function. To gain a better understanding of the impact of potentially deleterious variants of GRIN2A, a range of bioinformatics tools were employed in this study. Out of 1320 nsSNPs retrieved from the NCBI database, initially 16 were predicted as deleterious by 9 tools. Further assessment of their domain association, conservation profile, homology models, interatomic interaction, and Molecular Dynamic Simulation revealed that the variant I463S is likely to be the most deleterious for the structure and function of the protein. Despite the limitations of computational algorithms, our analyses have provided insights that can be a valuable resource for further in vitro and in vivo research on GRIN2A-associated diseases.
Citation: Ahammad I, Jamal TB, Bhattacharjee A, Chowdhury ZM, Rahman S, Hassan MR, et al. (2023) Impact of highly deleterious non-synonymous polymorphisms on GRIN2A protein’s structure and function. PLoS ONE 18(6): e0286917. https://doi.org/10.1371/journal.pone.0286917
Editor: Yang Zhang, University of Michigan, UNITED STATES
Received: February 27, 2023; Accepted: May 25, 2023; Published: June 15, 2023
Copyright: © 2023 Ahammad et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The goal of human genetics research is to comprehend how frequent genetic variations affect the probability of getting common illnesses [1]. Single Nucleotide Polymorphisms (SNPs) which refer to single base substitution in the alleles are such common sources of variance in the genome that can significantly influence an individual’s physical characteristics and their chance of developing certain diseases [2]. SNPs in the coding regions of genes (cSNPs) or regulatory regions are more likely to cause functional differences than SNPs elsewhere. About 90% of the variation in the human genome is due to SNPs, making them the most common type of genetic alteration [3].
Various genetic tools and related genomics have been used for the large-scale extraction of SNPs over the years. Over 10 million SNPs have been found, covering the most common types of genetic variations [4]. Since the human genome covers a vast number of genetic polymorphisms, it is necessary to conduct extensive research to understand the significance of each one and how they may contribute to disease susceptibility and personalized drug development [1]. To scale back the amount of effort required, various computational methods have been developed to identify potential variants before testing them in laboratory conditions [5]. In this circumstance, in silico approach is an efficient way to determine which SNPs are harmful and which are not, using particular processes [6]. Besides, integrative analysis of various algorithms improves the accuracy of the predicted effects of particular mutations. Moreover, cutting-edge approaches, such as molecular dynamics simulation allow for precise assessment of changes in protein properties, including structure, chemical properties, and interactions, within a simulated environment [7–9].
SNPs that occur in the coding region can be either synonymous, meaning they do not cause any change in the amino acid sequence, or non-synonymous (nsSNPS), meaning the amino acid sequence is altered [10]. It is usually assumed that synonymous SNPs are harmless, as the protein’s primary sequence is not changed [11]. On the other hand, nsSNPs may subsequently affect protein structure and protein-protein interactions and exert possible functional effects. However, not all nsSNPs that cause structural and functional changes are potentially harmful. Some nsSNPs affect the structure of a protein, while others create functional consequences. Moreover, some nsSNPs may be linked to disease, while others are considered neutral and do not have any association with disease [12–15]. Therefore, it is crucial to differentiate between harmful and neutral nsSNPs. In this work, we aimed to determine the most harmful nsSNPs within the GRIN2A gene by evaluating their impact on the structure and function of the GluN2A protein.
GRIN2A gene encodes for the GluN2A subunit of the N- methyl-D-aspartate (NMDA) glutamatergic receptor that is expressed throughout the brain and is important in the function of all neuron types. This receptor is activated through simultaneous binding by Glu and glycine and is formed by combinations of NR1, NR2, and NR3 subunits, with NR1-NR2 heterodimers forming the basic functional structure. The NR1 subunit has glycine binding sites, while the NR2 subunit has glutamate binding sites. NR3 subunits are regulatory subunits that decrease ionic currents generated by NR1/NR2 heteromers and are probably involved in activating silent NMDA-alone synapses. NMDA receptors are essential for synaptic transmission, learning, and memory as they function as ion channels that allow positively charged ions to flow through the neuron’s membrane [16]. Any mutations in GRIN2A gene weaken ion flow through the receptor, resulting in abnormal neuron function, epilepsy, and related developmental differences [17,18].
GRIN2A-related speech disorders and epilepsy affects individuals from a young age. All affected ones display some degree of speech disorder, with severe cases including dysarthria and speech dyspraxia, and both receptive and expressive language delay or regression. In addition to speech disorders, about 90% of affected individuals also experience epilepsy, with seizure onset typically occurring between the ages of three and six. The seizures associated with GRIN2A mutations can take various forms, including seizures with an aura of perioral paresthesia, focal or focal motor seizures, and atypical absence seizures. Different epilepsy syndromes including Landau-Kleffner syndrome, epileptic encephalopathy, childhood epilepsy, autosomal dominant rolandic epilepsy and infantile-onset epileptic encephalopathy have been linked to GRIN2A mutations [19]. However, the proportion of the abnormalities caused by the GRIN2A pathogenic variants is yet to be identified. The abundance of SNPs present in the GRIN2A makes laboratory experimentations challenging, seeking to investigate the functional effects of these variations, as such experiments can be both expensive and time-consuming. Therefore a computational screening of SNPs to minimize the number of potential pathogenic ones is a must before experimental mutation analysis. Considering this fact, in the current study we aimed to predict the consequences of the most damaging nsSNPs that occur in the GRIN2A coding region as reported in the dbSNP database (https://www.ncbi.nlm.nih.gov/snp/).
Materials and methods
The complete workflow employed in this study is outlined in Fig 1.
The overall process can be summarized as a series of steps that involve filtering and identifying the most damaging nsSNPs of GRIN2A, followed by a subsequent in-depth analysis of one particular nsSNP that proved to be the most deleterious.
Retrieval of SNPs
Single Nucleotide Polymorphism (SNP) variants recorded for the human GRIN2A gene were retrieved from NCBI dbSNP [20]. Only missense variants were selected for our study. The representative protein sequence of this gene was obtained from UniProt (https://www.uniprot.org/) (UniProt ID: Q12879) [21].
Functional evaluation of the nsSNPs
Five different tools were used to predict the impact of nsSNPs on the GRIN2A. These tools include PolyPhen-2, PANTHER, SNPs & GO, PhD-SNP, and PredictSNP2.
The functional and structural effects of certain missense variants were analyzed using several online tools. At first, the variants were submitted to the PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/) server, which is used to predict the effects of human nsSNPs on the function of protein molecules. This server, which is based on sequence homology, can help identify potentially damaging missense mutations and analyze the potential effects of an amino acid substitution on the structure and function of a human protein [22]. The rsIDs (Example: chr16:10274423 A/G) retrieved from the dbSNP database were used as input queries.
PANTHER (http://pantherdb.org/tools/csnpScoreForm.jsp) is a software that is used to predict non-synonymous genetic variants that may be related to human disease. This tool uses position- specific evolutionary conservation (PSEC) scores to make these predictions [23].
SNPs&GO (https://snps.biofold.org/snps-and-go/pages/method.html) is an online server that combines data from protein sequences, evolutionary information, and functions encoded in Gene Ontology terms to predict the effects of SNPs. This server is considered to be more accurate than other similar methods available [24].
PHD-SNP (https://snps.biofold.org/phd-snp/phd-snp.html) is a tool for predicting the effects of deleterious single nucleotide polymorphisms in humans, which is based on a Support Vector Machine (SVM) classifier. This tool can take the protein sequence, position of the mutation, or the mutated residue as input query. PMut (http://mmb.irbbarcelona.org/PMut/) is another web- based tool for predicting the effects of nsSNPs on protein function. A score greater than 0.5 obtained from this tool indicates that the nsSNPs have a damaging impact on protein function [25].
PredictSNP2 (https://loschmidt.chemi.muni.cz/predictsnp2/) is an online interface that provides easy access to predictions from five different tools and their consensus scores in a user-friendly format. It is tailored to the specific features of various categories of variations. The predictions are also accompanied by annotations from relevant databases to enable a comprehensive evaluation of variants [26].
Structural impact prediction
Four different structural impact prediction tools (NetSurfP-2.0, MUpro, Mutation3D, and HOPE) were used to find out the effect of SNPs on the protein structure.
NetSurfP-2.0 (https://services.healthtech.dtu.dk/service.php?NetSurfP-2.0) is a relatively new sequence-based tool that can accurately predict important local structural features of proteins with fast computation time. This server uses a neural network architecture that includes convolutional and long short-term memory layers, trained on previously solved protein structures [27]. The FASTA sequence of the GRIN2A protein was submitted to NetSurfP, and the output of the server revealed the buried and exposed regions in the protein structure [28].
With a prediction accuracy of 84%, MUpro (https://www.ics.uci.edu/~baldig/mutation.html) utilizes both Support Vector Machines and Neural Networks to calculate the impact of point mutations on protein stability. Similarly to NetSurfP2.0, it requires the FASTA format of the protein sequence as input [29].
Mutation3D (http://mutation3d.org) is a web server that aims to identify driver genes in cancer by detecting clusters of amino acid substitutions within tertiary protein structures [30]. The 3D structure of GRIN2A protein can be predicted and evaluated using mutation3D. To find out whether a mutation falls inside a domain, Mutation3D can be utilized.
Have (y) Our Protein Explained (HOPE) server (https://www3.cmbi.umcn.nl/hope/about/) uses a comprehensive approach to analyze a specific mutant protein. It gathers data from various sources, including 3D coordinate calculations of the protein. Before conducting analysis, this server requires the native protein sequence and information on the specific location and type of mutation to be studied [31].
Conservation profile analysis
To unravel the evolutionarily conserved sites of the human GRIN2A protein, the Consurf web server (http://consurf.tau.ac.il) was used. This server, which has been in operation for over 15 years, uses evolutionary patterns of amino/nucleic acids to identify regions that are critical for structure and/or function [32].
Identification of domain
For domain identification the Pfam server (http://pfam.xfam.org/) was employed. This is a widely used tool for analyzing protein function that features a collection of curated protein families, each defined by two alignments and a profile hidden Markov model (HMM). The profile HMMs, generated by aligning a set of family-representative sequences, are statistical models used to search for similarities among proteins. The Pfam website provides several methods to access its database content, including graphical representations and interactive data access [33].
Homology modeling and verification of native and mutant forms of GRIN2A protein
HHpred (https://toolkit.tuebingen.mpg.de/tools/hhpred) is a web-based tool for detecting remote protein homology and predicting structures that extracts homology information from HH-suite (open-source software for searching sensitive protein sequences). It is one of the first servers to effectuate pairwise comparison of profile hidden Markov models (HMMs). It accepts a single query sequence and multiple alignments as input depending on the user’s preference [34]. The protein FASTA sequences were submitted as input to generate the homology model of the protein. Later, the structure quality was exposed to verification by PROCHECK. The PROCHECK tool (https://servicesn.mbi.ucla.edu/PROCHECK/) evaluates the structural integrity of a protein by looking into the geometry of individual residues and the overall structure of the protein [35]. The model structure was further verified using the SAVES web server, and the quality of the structure was confirmed by analyzing the Ramachandran plot. Usually, a protein structure is considered good if more than 90% residues are in the favored region of Ramachandran plot. These techniques play a fundamental role in understanding the 3D models of proteins.
Refinement of the 3D structure of GRIN2A
To improve the accuracy of the 3D model for the GRIN2A protein, GalaxyRefine (https://galaxy.seoklab.org/cgibin/submit.cgi?type=REFINE) server was employed. This is a web-based tool for predicting and refining protein structures along with other related methods. It processes protein structure by performing repeated perturbation and structural relaxation by molecular dynamics simulation. The mechanism starts with the reconstruction of side chains and continues with side-chain repacking and overall structure relaxation [36].
Interatomic interactions prediction
The nsSNPs selected from the upstream analysis were evaluated using the DynaMut2 server. DynaMut2 (https://biosig.lab.uq.edu.au/dynamut2/) is an all-inclusive tool for analyzing protein motion and flexibility. It integrates optimized graph-based signatures with normal mode parameters to estimate the impact of point mutations on protein stability. Not only that but, it also predicts the effect of missense variations on protein stability and dynamics which is critical for understanding the link between protein structure and function and their role in associated diseases [37].
Molecular dynamic simulation
The wild-type GRIN2A and selected mutant protein models were subjected to a 100 ns molecular dynamics simulation using the GROMACS (version 2020.6) simulation software [38]. The force field utilized in the simulation was GROMOS96 43a1. A waterbox with edges 0.5 nm from the protein surface was constructed using the spc216 water model. Appropriate ions were used to neutralize the systems. A 100 ns molecular dynamic simulation was performed utilizing periodic boundary conditions following energy minimization, isothermal-isochoric (NVT) equilibration, and isobaric (NPT) equilibration of the system. After the simulation was complete, the root mean square deviation (RMSD), root mean square fluctuation (RMSF), the radius of gyration (Rg), and solvent accessible surface area (SASA) analyses were performed using the RMSD, RMSF, Rg, and SASA modules integrated into the GROMACS software. The plots for each of these studies were produced using the ggplot2 package in RStudio.
Results
SNP retrieval
Single nucleotide Polymorphism (SNP) of the human GRIN2A gene (Uniprot ID Q12879) was retrieved from the NCBI database (https://www.https://www.ncbi.nlm.nih.gov/snp/). Out of the 179,447 SNPs covered, 1,320 were missense (nsSNPs).
Assessment of the functional effect of GRIN2A nsSNPs
To evaluate the functional effects of nsSNPs on GRIN2A gene product, five tools were used: PolyPhen 2, PANTHER, SNPs&GO, PhD-SNP, and Predict SNP2. Out of the total 1320 nsSNPs, PolyPhen 2 server predicted 832 nsSNPs to be functionally damaging with 490 labeled as "probably damaging" and 275 as "possibly damaging." To further investigate these SNPs, additional analysis was conducted using SNPs&GO, PANTHER, PhD-SNP, and Predict-SNP2. The number of detrimental ones predicted by Sthe NPs&GO server was 154 while PANTHER, PhD-SNP, and Predict-SNP2 predicted 141, 451, and 692 nsSNPs as disease-associated, respectively.
Based on the outputs, 16 most highly deleterious nsSNPs were selected that were labeled as damaging, possibly/probably damaging, or deleterious by these tools. These 16 functionally significant nsSNPs were further taken into consideration for the next stage of filtering. The results from PolyPhen 2, PANTHER, SNPs&GO, PhD-SNP, and Predict SNP2 are presented in Table 1.
Structural impact prediction of GRIN2A nsSNPs
Sixteen nsSNPs that had been predicted to be damaging by 5 different tools then underwent structural impact analysis. For this purpose, four servers were employed namely, Mutation3D, NetSurfP-2.0, MUpro, and HOPE.
Three-dimensional visualization of GRIN2A mutations
For visualizing the locations of 16 nsSNPs of GRIN2A on the protein, outputs from the Mutation 3D server were loaded onto the PyMol software. The result was a three-dimensional representation of the human GRIN2A protein (Fig 2), with the mutated residues highlighted in red.
Surface accessibility of native and mutant proteins
Native and mutant protein accessibility and stability for the 16 variants were assessed by NetsurfP-2.0. NetSurfP-2.0 predicts the degree to which a residue is buried or exposed within the protein structure, as a percentage. These predictions were substantiated using multiple distinct test datasets. Out of 16 deleterious nsSNPs, 3 (R52Q, A643T, E656K) were found to be exposed in the wild type, while the remaining 13 (G97C, G97S, Q146P, L361Q, G376S, D462E, I463S, V471M, M507T, P527S, L642M, S644G, A778D) were buried. On the other hand, 2 mutants (R52Q, E656K) remained exposed while the rest of the 14 (G97C, G97S, Q146P, L361Q, G376S, D462E, I463S, V471M, M507T, P527S, L642M, A643T, S644G, A778D) were buried (Table 2).
Structural impact of the SNPs on GRIN2A gene product
Out of 16 nsSNPs, the MUpro server predicted that 14 nsSNPs would have a decreasing effect on protein stability. The Project HOPE server was used to investigate the effect of mutations on various aspects of protein structure and function such as physico-chemical properties, hydrophobicity, intermolecular interactions, and structural and functional changes. In accordance with the results, 10 mutant residues (G97C, G97S, L361Q, G376S, D462E, V471M, L642M, A643T, E656K, and A778D) were larger than the wild type, while 6 mutant residues (R52Q, Q146P, I463S, M507T, P527S, and S644G) were smaller than the wild type. Out of the 16 nsSNPs analyzed, 14 were predicted to have a probable damaging effect, one as a possibly damaging effect and one as a damaging effect on the protein structure according to the HOPE server.
Identification of conserved sequence in GRIN2A
To further examine the potential impact of the shortlisted 16 nsSNPs, the ConSurf web tool was used to calculate the evolutionary conservation of amino acid residues in the GRIN2A protein. The results were presented as a structural representation of the protein sequence, with the putative structural and functional residues highlighted.
According to ConSurf output, 12 of the 16 nsSNPs (Q146P, G376S, D462E, I463S, V471M, M507T, P527S, L642M, A643T, S644G, E656K, and A778D) were found to be highly conserved residues with a conservation score of 9. Three variants (G97C, G97S, and L361Q) were predicted as moderately conserved (conservation score of 8) while only one variant (R52Q) was predicted as a variable with a conservation score of 3. Variants located in these conserved regions are considered to be highly damaging to the protein, as referred to as those located in non-conserved sites. The deleterious predictions for each SNP by ConSurf are summarized in Fig 3. After careful analysis, we identified 9 out of 16 mutants as having the most significant impact on the structure of the GRIN2A protein. We subsequently selected specific nsSNPs (listed in Table 3) from the GRIN2A protein for further investigation.
For the identification of the most deleterious nsSNPs, ConSurf, MUpro, and HOPE outputs were analyzed and scrutinized. Finally, 9 out of the 16 mutants were considered the most impactful for the structure of the GRIN2A protein and selected for further analysis.
Identification of domain
Three domains, namely- ANF receptor, Lig chan-Glu, and Lig chan were identified within the GRIN2A protein using the Pfam server. The ANF receptor domain contained 1 nsSNP (Q146P), while the Lig chan-Glu domain had 4 nsSNPs (I463S, V471M, M507T, P527S) and the Lig chan domain had 4 nsSNPs (A643T, S644G, E656K, A778D).
Homology modeling and validation of the GRIN2A structure
The homology model of the protein was generated in PDB Format using the HHPred server. The quality of the model was validated by inspecting the Ramachandran plot generated by the PROCHECK web server. The findings indicated that over 90% of the residues of the homology model belonged to the most favored regions for both the wild and mutant version (S1 Fig).
Refinement of the 3D structure
The 3D model of the protein of interest was refined using the GalaxyRefne server. The input for the refinement process was in PDB file format. The server provided information on the 5 best models in terms of RMSD, Clash score, Rama favored, and MolProbity. The model with the lowest MolProbity score and the highest Rama favoured score was selected (Table 4).
Interatomic interactions analysis
The nsSNPs selected so far were analyzed using the DynaMut2 server (Fig 4, Table 5). DynaMut2 provides the change in stability (ΔΔG) in terms of kcal/mol. A negative value indicates a destabilizing effect while a positive value indicates a stabilizing effect. The more negative the ΔΔG value the more destabilizing the mutation is for the protein. The analysis identified 8 residues (Q146P, I463S, V471M, M507T, P527S, S644G, E656K, A778D) as destabilizing for the protein structure, while 1 residue (A643T) had a stabilizing effect. The SNP, I463S exhibited the highest destabilizing effect on the protein (-2.61 kcal/mol). Therefore, it was selected for sophisticated molecular dynamics simulation in the next step.
Altered interatomic interactions between the wild and the mutant residues with their neighboring atoms have been observed.
Molecular dynamics simulation
Root Mean Square Deviation (RMSD) calculation is performed in to evaluate the stability of the systems. A higher RMSD value corresponds to instability in the protein. The RMSD for the wild GRIN2A stabilized fairly quickly after 10 ns. Since then, its value remained stable at around 0.7 nm. In contrast, the I463S mutant GRIN2A possessed a higher RMSD profile throughout the simulation. Its value ranged above 0.9 nm from 10 ns till the end of the simulation (Fig 5).
Room Mean Square Fluctuation (RMSF) is used to determine the regional flexibility of the protein. The higher the RMSF, the higher is the flexibility of a given amino acid position. Except for a few in the middle, almost all residues in the mutant I463S exhibited higher flexibility compared to the wild-type GRIN2A residues (Fig 6).
The radius of gyration is a measure to determine its degree of compactness. A relatively steady value of the radius of gyration means stable folding of a protein. Fluctuation of the radius of gyration implies the unfolding of the protein. The radius of gyration analysis indicated that the wild type GRIN2A was relatively more compact than the I463S mutant GRIN2A as the former reached a stable value shortly after 35 ns while the latter reached this state after 65 ns (Fig 7).
Solvent Accessible Surface Area (SASA) is used in MD simulations to predict the hydrophobic core stability of proteins. The higher the SASA value, the higher the chance of destabilization of the protein due to solvent accessibility. SASA values for the wild type and the I463S mutant GRIN2A remained close. However, the wild type displayed a slightly lower SASA profile compared to the other (Fig 8).
Discussion
Almost half of the genetic variations that cause hereditary diseases come from nsSNPs [39]. Thorough research about the effect of these nsSNPs on disease-associated proteins can aid in the creation of more tailored, personalized treatments for the affected ones [40,41]. Although it can be laborious to identify which nsSNPs are responsible for specific symptoms due to the tedious and expensive nature of molecular approaches, bioinformatics can be used to predict which nsSNPs are pathogenic and prioritize them for future investigation [42,43]. This can help improve our understanding of the structure and function of proteins involved in these diseases.
Among the four genes (GRIN1, GRIN2A, GRIN2B, and GRIN2D) that encode NMDAR subunits that have so far been linked to human disease, GRIN2A appears to have the most extensive and well-characterized range of phenotypic effects [44]. Functional investigations of disease-associated GRIN2A missense variants have revealed various gain- or loss-of-function effects [45]. In a study by Endele et al. (2010), a de novo GRIN2A variant was found in a child with early-onset epileptic encephalopathy [46]. Another study by Gao et al., identified a new missense mutation in the GRIN2A gene in a patient with childhood focal epilepsy and acquired epileptic aphasia. This particular mutation was found to lower NMDAR activation implying that the decreased function of NMDAR could potentially be involved in the development of epilepsy [47]. Therefore, a detailed analysis of GRIN2A related nsSNPs is essential before starting targeted treatment options. In this study, we conducted intensive in silico analyses to identify the pathogenic nsSNPs of the GRIN2A gene. For this purpose, a wide range of computational tools was used to get a detailed understanding of the potential impact of these SNPs on the target gene.
To conduct our study, we retrieved all the available nsSNPs of the GRIN2A gene and annotated them using multiple computational tools with an aim to distinguish between the functional and neutral variants. The combination of these tools allowed us to produce a cohesive image of the potential pathogenic SNPs of the GRIN2A gene. To identify the high-risk SNPs, we isolated 16 nsSNPs that were predicted to be deleterious by all of the prediction algorithms, as these were considered to have a greater chance of being pathogenic.
The function, activity, and regulation of a protein are closely tied to its structural stability. When stability decreases, proteins undergo degradation, misfolding, and clumping, leading to eventual dysfunction [48]. To evaluate the impact of the 16 harmful nsSNPs on the stability of the GRIN2A protein, the nsSNPs were first analyzed using 3D structures of the mutated domains. Homology modeling was performed in order to obtain the 3D structure of the human GRIN2A protein due to the lack of existing X-ray crystallographic or NMR structure. Additionally, to determine the deleteriousness of the functional nsSNPs, the nsSNPs located in the conserved domains of the proteins were selected, as nsSNPs in highly conserved regions are more likely to cause harm than those located in the variable regions [49]. After undergoing homology modeling, model verification, secondary structure analysis, and interatomic interaction prediction, the list of mutations was narrowed down to I463S which exhibited the highest destabilizing effect on the protein.
Moreover, 100 ns long molecular dynamics simulation of the wild-type GRIN2A and I463S mutant GRIN2A proteins revealed that the mutation imposed a destabilizing effect on the structure of the GRIN2A protein under physiological conditions. Due to such destabilizing effects, protein functioning is also likely to be interrupted.
This study has reduced our knowledge gap about the polymorphisms of the GRIN2A gene and brought to our attention the probable presence of similar effect-causing variants. Consequently, a broader scale of human in vivo molecular imaging studies may be necessary to pinpoint the most impactful variants. In general, this study highlights the importance of continued research in this field to facilitate the development of customized treatment options for individuals with abnormalities related to GRIN2A. Besides, the insights obtained from this study may aid in developing novel drug-targeting strategies and biomarkers for those affected by GRIN2A-linked disorders.
Conclusion
The identification of a potential high-risk variant of the GRIN2A gene in our study paves the way for further investigations into the impact of nsSNPs on the function of GRIN2A protein and their correlation with disease. In vivo models, genome-wide association studies, and clinical inspections can provide valuable insights into the biological mechanisms underlying these disorders and aid in the development of improved diagnostic and therapeutic strategies.
Supporting information
S1 Fig. Ramachandran plot of wild and mutant GRIN2A protein derived from PROCHECK.
The plot shows the distribution of the residues within the most favored, additional allowed, generously allowed, and disallowed regions.
https://doi.org/10.1371/journal.pone.0286917.s001
(PNG)
References
- 1. Cargill M, Altshuler D, Ireland J, Sklar P, Ardlie K, Patil N, et al. Characterization of single-nucleotide polymorphisms in coding regions of human genes. Nat Genet. 1999 Jul;22(3):231–8. pmid:10391209
- 2. Lira SS, Ahammad I. A comprehensive in silico investigation into the nsSNPs of Drd2 gene predicts significant functional consequences in dopamine signaling and pharmacotherapy. Sci Rep. 2021 Dec 1;11(1):23212. pmid:34853389
- 3. Collins FS, Brooks LD, Chakravarti A. A DNA Polymorphism Discovery Resource for Research on Human Genetic Variation: Table 1. Genome Res. 1998 Dec 1;8(12):1229–31.
- 4. Ronaghi M, Langaee T. Single nucleotide polymorphisms: discovery, detection and analysis. Pers Med. 2005 May;2(2):111–25. pmid:29788584
- 5. Marín-Martín FR, Soler-Rivas C, Martín-Hernández R, Rodriguez-Casado A. A Comprehensive In Silico Analysis of the Functional and Structural Impact of Nonsynonymous SNPs in the ABCA1 Transporter Gene. Cholesterol. 2014 Aug 19;2014:1–19.
- 6. Reza MN, Ferdous N, Emon MdTH, Islam MdS, Mohiuddin AKM, Hossain MU. Pathogenic genetic variants from highly connected cancer susceptibility genes confer the loss of structural stability. Sci Rep. 2021 Sep 28;11(1):19264. pmid:34584144
- 7. George Priya Doss C, Rajasekaran R, Sudandiradoss C, Ramanathan K, Purohit R, Sethumadhavan R. A novel computational and structural analysis of nsSNPs in CFTR gene. Genomic Med. 2008 Jan;2(1–2):23–32. pmid:18716917
- 8. Kumar A, Purohit R. Use of Long Term Molecular Dynamics Simulation in Predicting Cancer Associated SNPs. MacKerell AD, editor. PLoS Comput Biol. 2014 Apr 10;10(4):e1003318.
- 9. Salo-Ahen OMH, Alanko I, Bhadane R, Bonvin AMJJ, Honorato RV, Hossain S, et al. Molecular Dynamics Simulations in Drug Discovery and Pharmaceutical Development. Processes. 2020 Dec 30;9(1):71.
- 10. Shastry BS. SNPs in disease gene mapping, medicinal drug development and evolution. J Hum Genet. 2007 Nov;52(11):871–80. pmid:17928948
- 11.
Hunt R, Sauna ZE, Ambudkar SV, Gottesman MM, Kimchi-Sarfaty C. Silent (Synonymous) SNPs: Should We Care About Them? In: Komar AA, editor. Single Nucleotide Polymorphisms [Internet]. Totowa, NJ: Humana Press; 2009 [cited 2023 Apr 29]. p. 23–39. (Methods in Molecular Biology; vol. 578). Available from: http://link.springer.com/10.1007/978-1-60327-411-1_2 pmid:19768585
- 12. Bromberg Y, Rost B. SNAP: predict effect of non-synonymous polymorphisms on function. Nucleic Acids Res. 2007 May 7;35(11):3823–35. pmid:17526529
- 13. Yates CM, Sternberg MJE. The Effects of Non-Synonymous Single Nucleotide Polymorphisms (nsSNPs) on Protein–Protein Interactions. J Mol Biol. 2013 Nov;425(21):3949–63. pmid:23867278
- 14. Datta A, Mazumder MdHH, Chowdhury AS, Hasan MdA. Functional and Structural Consequences of Damaging Single Nucleotide Polymorphisms in Human Prostate Cancer Predisposition Gene RNASEL. BioMed Res Int. 2015;2015:1–15. pmid:26236721
- 15. Dakal TC, Kala D, Dhiman G, Yadav V, Krokhotin A, Dokholyan NV. Predicting the functional consequences of non-synonymous single nucleotide polymorphisms in IL8 gene. Sci Rep. 2017 Jul 26;7(1):6525. pmid:28747718
- 16. Flores-Soto ME, Chaparro-Huerta V, Escoto-Delgadillo M, Vazquez-Valls E, González-Castañeda RE, Beas-Zarate C. Structure and function of NMDA-type glutamate receptor subunits. Neurol Engl Ed. 2012 Jun;27(5):301–10.
- 17.
Philadelphia TCH of. GRIN2A-Related Disorders [Internet]. The Children’s Hospital of Philadelphia; 2020 [cited 2023 Apr 29]. Available from: https://www.chop.edu/conditions-diseases/grin2a-related-disorders.
- 18. Liu XR, Xu XX, Lin SM, Fan CY, Ye TT, Tang B, et al. GRIN2A Variants Associated With Idiopathic Generalized Epilepsies. Front Mol Neurosci. 2021 Oct 14;14:720984. pmid:34720871
- 19.
Myers KA, Scheffer IE. GRIN2A-Related Speech Disorders and Epilepsy. In: Adam MP, Mirzaa GM, Pagon RA, Wallace SE, Bean LJ, Gripp KW, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993 [cited 2023 Apr 29]. Available from: http://www.ncbi.nlm.nih.gov/books/NBK385627/.
- 20. Sherry ST. dbSNP: the NCBI database of genetic variation. Nucleic Acids Res. 2001 Jan 1;29(1):308–11. pmid:11125122
- 21. The UniProt Consortium. The Universal Protein Resource (UniProt) in 2010. Nucleic Acids Res. 2010 Jan 1;38(suppl_1):D142–8.
- 22. Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, Bork P, et al. A method and server for predicting damaging missense mutations. Nat Methods. 2010 Apr;7(4):248–9. pmid:20354512
- 23. Tang H, Thomas PD. PANTHER-PSEP: predicting disease-causing genetic variants using position-specific evolutionary preservation. Bioinformatics. 2016 Jul 15;32(14):2230–2. pmid:27193693
- 24. Calabrese R, Capriotti E, Fariselli P, Martelli PL, Casadio R. Functional annotations improve the predictive score of human disease-related mutations in proteins. Hum Mutat. 2009 Aug;30(8):1237–44. pmid:19514061
- 25. Ferrer-Costa C, Gelpi JL, Zamakola L, Parraga I, de la Cruz X, Orozco M. PMUT: a web-based tool for the annotation of pathological mutations on proteins. Bioinformatics. 2005 Jul 15;21(14):3176–8. pmid:15879453
- 26. Bendl J, Musil M, Štourač J, Zendulka J, Damborský J, Brezovský J. PredictSNP2: A Unified Platform for Accurately Evaluating SNP Effects by Exploiting the Different Characteristics of Variants in Distinct Genomic Regions. Gardner PP, editor. PLOS Comput Biol. 2016 May 25;12(5):e1004962.
- 27. Klausen MS, Jespersen MC, Nielsen H, Jensen KK, Jurtz VI, Sønderby CK, et al. NetSurfP‐2.0: Improved prediction of protein structural features by integrated deep learning. Proteins Struct Funct Bioinforma. 2019 Jun;87(6):520–7. pmid:30785653
- 28. Petersen B, Petersen TN, Andersen P, Nielsen M, Lundegaard C. A generic method for assignment of reliability scores applied to solvent accessibility predictions. BMC Struct Biol. 2009 Dec;9(1):51. pmid:19646261
- 29. Cheng J, Randall A, Baldi P. Prediction of protein stability changes for single-site mutations using support vector machines. Proteins Struct Funct Bioinforma. 2005 Dec 21;62(4):1125–32.
- 30. Meyer MJ, Lapcevic R, Romero AE, Yoon M, Das J, Beltrán JF, et al. mutation3D: Cancer Gene Prediction Through Atomic Clustering of Coding Variants in the Structural Proteome. Hum Mutat. 2016 May;37(5):447–56. pmid:26841357
- 31. Venselaar H, te Beek TA, Kuipers RK, Hekkelman ML, Vriend G. Protein structure analysis of mutations causing inheritable diseases. An e-Science approach with life scientist friendly interfaces. BMC Bioinformatics. 2010 Dec;11(1):548.
- 32. Ashkenazy H, Abadi S, Martz E, Chay O, Mayrose I, Pupko T, et al. ConSurf 2016: an improved methodology to estimate and visualize evolutionary conservation in macromolecules. Nucleic Acids Res. 2016 Jul 8;44(W1):W344–50.
- 33. Finn RD, Bateman A, Clements J, Coggill P, Eberhardt RY, Eddy SR, et al. Pfam: the protein families database. Nucleic Acids Res. 2014 Jan;42(D1):D222–30. pmid:24288371
- 34. Soding J, Biegert A, Lupas AN. The HHpred interactive server for protein homology detection and structure prediction. Nucleic Acids Res. 2005 Jul 1;33(Web Server):W244–8. pmid:15980461
- 35. Laskowski RA, MacArthur MW, Moss DS, Thornton JM. PROCHECK: a program to check the stereochemical quality of protein structures. J Appl Crystallogr. 1993 Apr 1;26(2):283–91.
- 36. Heo L, Park H, Seok C. GalaxyRefine: protein structure refinement driven by side-chain repacking. Nucleic Acids Res. 2013 Jul 1;41(W1):W384–8. pmid:23737448
- 37. Rodrigues CHM, Pires DEV, Ascher DB. DynaMut2: Assessing changes in stability and flexibility upon single and multiple point missense mutations. Protein Sci. 2021 Jan;30(1):60–9. pmid:32881105
- 38. Abraham MJ, Murtola T, Schulz R, Páll S, Smith JC, Hess B, et al. GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX. 2015 Sep;1–2:19–25.
- 39. Ramensky V. Human non-synonymous SNPs: server and survey. Nucleic Acids Res. 2002 Sep 1;30(17):3894–900. pmid:12202775
- 40. Gray IC. Single nucleotide polymorphisms as tools in human genetics. Hum Mol Genet. 2000 Oct 1;9(16):2403–8. pmid:11005795
- 41. Vallejos-Vidal E, Reyes-Cerpa S, Rivas-Pardo JA, Maisey K, Yáñez JM, Valenzuela H, et al. Single-Nucleotide Polymorphisms (SNP) Mining and Their Effect on the Tridimensional Protein Structure Prediction in a Set of Immunity-Related Expressed Sequence Tags (EST) in Atlantic Salmon (Salmo salar). Front Genet [Internet]. 2020 [cited 2023 Apr 29];10. Available from: https://www.frontiersin.org/articles/10.3389/fgene.2019.01406.
- 42. Mah JTL, Low ESH, Lee E. In silico SNP analysis and bioinformatics tools: a review of the state of the art to aid drug discovery. Drug Discov Today. 2011 Sep;16(17–18):800–9. pmid:21803170
- 43. Brown DK, Tastan Bishop Ö. Role of Structural Bioinformatics in Drug Discovery by Computational SNP Analysis. Glob Heart. 2017 Jun 1;12(2):151.
- 44. Lemke JR, Lal D, Reinthaler EM, Steiner I, Nothnagel M, Alber M, et al. Mutations in GRIN2A cause idiopathic focal epilepsy with rolandic spikes. Nat Genet. 2013 Sep;45(9):1067–72. pmid:23933819
- 45. Strehlow V, Heyne HO, Vlaskamp DRM, Marwick KFM, Rudolf G, de Bellescize J, et al. GRIN2A -related disorders: genotype and functional consequence predict phenotype. Brain. 2019 Jan 1;142(1):80–92.
- 46. Endele S, Rosenberger G, Geider K, Popp B, Tamer C, Stefanova I, et al. Mutations in GRIN2A and GRIN2B encoding regulatory subunits of NMDA receptors cause variable neurodevelopmental phenotypes. Nat Genet. 2010 Nov;42(11):1021–6. pmid:20890276
- 47. Gao K, Tankovic A, Zhang Y, Kusumoto H, Zhang J, Chen W, et al. A de novo loss-of-function GRIN2A mutation associated with childhood focal epilepsy and acquired epileptic aphasia. Mothet JP, editor. PLOS ONE. 2017 Feb 9;12(2):e0170818.
- 48. Rozario LT, Sharker T, Nila TA. In silico analysis of deleterious SNPs of human MTUS1 gene and their impacts on subsequent protein structure and function. Wang J, editor. PLOS ONE. 2021 Jun 14;16(6):e0252932. pmid:34125870
- 49. Miller MP. Understanding human disease mutations through the use of interspecific genetic variation. Hum Mol Genet. 2001 Oct 1;10(21):2319–28. pmid:11689479