A computational screening of genetic variants across the solute carrier transporter genes SLC22A1, SLC22A2, and SLC22A3 has identified a prioritized set of deleterious nonsynonymous mutations predicted to impair transporter stability and alter drug disposition. The study, conducted by R. Harini, B. Usha, R. Satish, Pai Akshatha, and T. M. Vijaya kumar, utilized a multilayered bioinformatics pipeline to evaluate exonic single nucleotide polymorphisms that alter amino acid sequences in organic cation transporters. These transport proteins serve as key physiological conduits for the cellular uptake and elimination of metformin, a biguanide agent widely prescribed to manage insulin resistance and anovulation in individuals diagnosed with polycystic ovary syndrome. By integrating sequence homology, evolutionary conservation metrics, and structural modeling, the researchers characterized specific amino acid substitutions that systematically destabilize protein conformation, presenting candidate markers for future clinical and functional validation.
Polycystic ovary syndrome represents a prevalent and heterogeneous endocrine and metabolic disorder affecting approximately ten percent of women of reproductive age. Clinically, the syndrome manifests through elevated androgen levels, irregular or absent ovulation, and polycystic ovarian morphology, frequently accompanied by systemic metabolic disturbances including peripheral insulin resistance, type 2 diabetes mellitus, and cardiovascular risks. While metformin is extensively administered as an off-label insulin sensitizer to suppress hepatic glucose production and enhance reproductive outcomes, therapeutic efficacy remains highly variable across patient populations. Clinical documentation shows that approximately thirty percent of individuals with polycystic ovary syndrome display non-responsive outcomes following standard metformin therapy. Because metformin does not stimulate insulin secretion directly but relies on carrier-mediated uptake to reach intracellular targets and undergo renal excretion, variations in organic cation transporter genes present a plausible molecular explanation for observed inter-individual differences in therapeutic efficacy.
To systematically evaluate the genetic architecture of these transporters, the investigators extracted reference protein sequences from the UniProt Knowledgebase for human organic cation transporters, corresponding to accession numbers O15245 for SLC22A1, O15244 for SLC22A2, and O75751 for SLC22A3. Mapping exonic nonsynonymous single nucleotide polymorphisms through the NCBI dbSNP repository cataloged 576 missense mutations alongside 12,563 intronic variants in SLC22A1, 510 missense mutations alongside 14,226 intronic variants in SLC22A2, and 505 missense mutations alongside 33,392 intronic variants in SLC22A3. Although intronic variants constituted the overwhelming numeric majority of identified polymorphisms across all three loci, the downstream analytical pipeline prioritized missense mutations occurring within coding regions. These nonsynonymous substitutions directly replace native amino acids in the translated polypeptide chain, making them primary candidates for disrupting structural folding, helix packing, and substrate translocation dynamics.
Pathogenicity predictions were executed across a rigorous assembly of five primary computational tools alongside an ensemble consensus meta-predictor. The primary screening deployed Align GVGD to evaluate evolutionary conservation across codon positions, PolyPhen-2 to compute probabilistic damage scores based on naive Bayes classification, PANTHER cSNP using the position-specific evolutionary preservation algorithm, PROVEAN to assess sequence homology clustering against homolog alignments, and PhD-SNP using support vector machine classification trained on disease-associated mutations. Across the cataloged variants, Align GVGD classified over fifty-two percent of variants in each transporter as belonging to the highest functional interference category, Class 65, while PolyPhen-2 designated between thirty-four and forty-three percent as probably damaging. Stringent multi-tool filtering demanded unanimous pathogenic classification across all five algorithms before verification through Meta-SNP, which consolidated predictions into an ensemble score and reliability index.
Consensus evaluation pinpointed a discrete collection of high-risk mutations, comprising L42R, R175C, R175L, and F422S within SLC22A1; R404C, G373D, and E227A within SLC22A2; and R212C, G259R, T275M, R348W, R407C, and P499S within SLC22A3. High-confidence candidates, including R175C, R175L, R404C, R212C, and R348W, achieved Meta-SNP scores exceeding 0.89 alongside reliability index values of seven or eight, demonstrating uniform agreement across sequence and evolutionary models. Conversely, variants such as T275M and G373D demonstrated lower Meta-SNP scores below 0.70 and reliability indices of three, suggesting that while they retain potential modifier effects, they exhibit lower individual pathogenic certainty. This tiered consensus screening effectively eliminated ambiguous variants, establishing an evidence-based subset of mutations possessing high potential for damaging transporter structure.
To understand the thermodynamic consequences of these specific amino acid changes, the research team performed stability evaluations using the machine learning program MuPro. By calculating shifts in Gibbs free energy, designated as Delta Delta G values, the program evaluated whether substituted side chains disrupt native thermodynamic stability. Across all shortlisted nonsynonymous variants, predicted Delta Delta G values were entirely negative, ranging from negative 0.01745 to negative 1.99395. A negative Delta Delta G consistently indicates reduced structural stability, a biophysical state closely correlated with higher susceptibilities to protein misfolding, aggregation, and premature proteasomal degradation. Notably, the E227A substitution in SLC22A2 generated the most severe thermodynamic destabilization with a value of negative 1.99395, followed by the L42R substitution in SLC22A1 and the P499S substitution in SLC22A3, pointing to pronounced structural vulnerabilities in mutated transporter molecules.
Topological and solvent accessibility profiling using NetSurfP-2.0 further revealed how each substitution alters the spatial organization of the transport proteins. Membrane-spanning profiling identified L42R within the first transmembrane helix and F422S within the ninth transmembrane helix of SLC22A1, while R212C and P499S mapped to the third and twelfth transmembrane helices of SLC22A3, respectively. NetSurfP-2.0 computations classified L42R, F422S, R404C, G373D, E227A, G259R, R348W, and R407C as surface-exposed residues with relative solvent accessibility exceeding twenty-five percent. In contrast, R175C, R175L, R212C, T275M, and P499S were classified as buried within the structural interior. Exchanging nonpolar residues for charged residues on exposed faces alters hydration shells and electrostatic interactions, whereas introducing bulky or sulfur-containing residues within buried domains disrupts internal core packing and interferes with conformational cycling required for active substrate translocation.
Three-dimensional protein homology models of native and exposed mutant forms were generated using SWISS-MODEL, selecting structural templates from the UniProtKB proteome spanning between sixty-seven and one hundred percent sequence coverage. Stereochemical validation of the generated three-dimensional coordinate files was confirmed through multiple structural quality assessment algorithms, including MolProbity scores below 1.5, ERRAT overall quality scores exceeding eighty-five percent, and ProSA-web Z-scores ranging between negative 6.8 and negative 8.2. For both wild-type and mutant variants, model number two emerged as optimal for SLC22A1 and SLC22A3, whereas model number one provided optimal stereochemical geometry for SLC22A2. These verified structural models demonstrated how surface-exposed replacements, such as replacing small aliphatic side chains with bulky aromatic or charged residues, modify the external surface topology and local curvature near gating domains.
Functional protein-protein interaction networks and enrichment analyses using the STRING database and Reactome pathways contextualized the physiological integration of these transporters. The functional network of SLC22A1 showed significant associations with the multidrug resistance ABC transporters ABCG2 and ABCB1, as well as the multidrug and toxin extrusion proteins SLC47A1 and SLC47A2. Enriched biological processes for SLC22A1 heavily highlighted xenobiotic transport and detoxification via transmembrane export, with Reactome mapping confirming central involvement in Drug ADME pathways (adjusted p = 1.5e-8). SLC22A2 networks displayed prominent interactions with SLC47A1, SLC47A2, and the cation transporter SLC31A1, reflecting roles in renal elimination of cationic compounds. SLC22A3 networks demonstrated substantial functional connections to the monoamine transporters SLC6A2, SLC6A3, and SLC6A4, showing enriched involvement in norepinephrine and dopamine clearance pathways, pointing to wider systemic and neuromodulatory roles.
Examination of population-scale minor allele frequencies via the Genome Aggregation Database revealed that most identified functional variants represent rare or ultra-rare mutations, displaying frequencies well below 0.001 globally. For instance, the SLC22A1 variants L42R and F422S exhibited global allele frequencies of 7.12e-5 and 2.17e-5, respectively. However, noticeable population-specific variations emerged, with specific variants such as rs768905186 and rs778599126 showing elevated frequencies ranging from 0.004 to 0.006 within South Asian ancestral groups. The authors emphasize that while these in silico findings establish a biophysical and evolutionary basis for altered metformin transport, the investigation is inherently computational. Direct in vitro transport assays, in vivo pharmacokinetic analyses, and clinical cohort sequencing remain necessary to confirm whether these predicted structural disruptions produce measurable clinical pharmacoresistance in patients with polycystic ovary syndrome.
Subject of Research: In silico prioritization and structural modeling of deleterious nonsynonymous variants in human SLC22A1, SLC22A2, and SLC22A3 transporter genes
Article Title: Identification of high-risk SNPs in SLC22A transporter genes: their potential role in PCOS and metformin uptake
Article References: Harini, R., Usha, B., Satish, R., Akshatha, P., & Vijaya kumar, T. M. (2026). Identification of high-risk SNPs in SLC22A transporter genes: their potential role in PCOS and metformin uptake. BMC Genomic Data, 27(1), Article 47. https://doi.org/10.1186/s12863-026-01424-8
Image Credits: AI Generated
DOI: 10.1186/s12863-026-01424-8
Keywords: SLC22A Transporters, Metformin Uptake, Polycystic Ovary Syndrome, Pharmacogenomics, Single Nucleotide Polymorphisms, In Silico Screening, Protein Stability, SWISS-MODEL, Identification, high-risk, SNPs, SLC22A
News Source: Juliet Wilcox. (November 10, 2026). Computational Screening Identifies Deleterious Mutations in SLC22A Transporter Genes Affecting Metformin Uptake. Scienmag.



