Regulatory impact of a 5’UTR resident single nucleotide variant on EDN1 gene expression and RNA-protein interactions

Abstract

Non-coding regulatory variants are gaining a lot of attention in recent years. This study identifies a 5’UTR variant of the Endothelin-1 gene (EDN1) linked to elevated production of Endothelin-1. The potent vasoactive Endothelin-1 (ET-1) peptide is implicated in vascular homeostasis. Its dysregulated production disturbs the balance between vasodilation and vasoconstriction with pathophysiological consequences including cardiovascular diseases. In an observational study with 95 CAD patients two were identified as homozygous for an insertion of ‘A’ at + 139 position of 5’UTR region of EDN1 gene. These patients exhibited elevated median plasma ET-1 levels (16.02 pg/ml), compared to 6.6 pg/ml and 3.87 pg/ml observed in patients without the insertion in the homozygous and heterozygous states, respectively. This clinical observation, uninformative statistically, was verified in vitro through reporter assays that demonstrated a significant increase in ET-1 expression. RNA affinity pull-down coupled with MS/MS was further employed to delineate differential RNA protein interactions. Proteomic profiling identified recruitment of RNA-binding proteins by the insertion variant. Collectively, these findings implicate the + 139 A insertion in the EDN1 5′-UTR as a functional non-coding variant that primarily enhances ET-1 expression through post-transcriptional mechanisms, including translation efficiency and mRNA stability, thereby influencing cardiovascular disease risk.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-47439-0.

Introduction

Non-coding regions like promoters, silencers, insulators, and enhancers form regulatory elements that control gene expression. Likewise, non-coding regions within mRNA known as 5′ and 3′ untranslated regions (UTRs) play an important role in post-transcriptional regulation¹. Variations in 5’ and 3’- UTR regions, including the Single Nucleotide Variations (SNVs), lead to altered gene expression and thus have attracted a lot of attention in recent years for their role in disease pathogenesis². Traditional exome sequencing focuses primarily on protein-coding sequences, often overlooking variations within non-coding genomic regions that have a significant impact on gene regulation. Mutations in protein-coding regions can cause disease, but mutations in noncoding regions, especially untranslated regions (UTRs), also have important regulatory roles that are less well understood. Single nucleotide polymorphisms (SNPs)/ variants (SNVs) are genetic alterations in the genomic DNA that impact just one nucleotide, and might influence an individual’s susceptibility to disease and how they respond to drugs, immunisations, and environmental factors. Emerging evidence highlights the significance of sequence variations within UTRs, particularly the 5′-UTR, in disease pathogenesis and much of biological research is geared towards understanding the molecular mechanisms underlying the role of these non-coding variants in human disease^3,4. In this study we report an SNV occurring due to insertion of a single adenine (‘A’) nucleotide in the 5’ UTR region of the EDN1 gene that codes for the endothelium-derived vasoconstrictor peptide, Endothelin-1, responsible for maintenance of vasomotor tone in the body^5,6. Endothelin is increasingly recognized as a critical mediator in the development of chronic cardiovascular diseases that have a complex aetiology with genetic, environmental and lifestyle related factors playing a role in disease progression⁷. Endothelin receptor antagonists are well-established in the treatment of pulmonary arterial hypertension and are currently being investigated for broader applications, including diabetic nephropathy, peripheral arterial disease, systemic hypertension, heart failure, and various pulmonary disorders^8,9.

The EDN1 gene is transcribed as a long inactive precursor of 212 residues which through a series of proteolytic cleavages results in the mature 21-aa active ET-1 peptide¹⁰. Genetic alterations in endothelin, endothelin converting enzymes, and endothelin receptors have been implicated in or put at risk for a number of illnesses^11–13. Alone EDN1 gene has been reported to contain 252 SNPs¹⁴. One of such gene variation is in the 5’-UTR (rs1800997), where the insertion or deletion of an adenine (A) nucleotide at position + 139 from transcription start, site is linked to changes in blood pressure (Fig. 1)^15–18.

Fig. 1.

Schematic representation of the human EDN1 gene structure, illustrating location of the + 139 ‘A’ insertion polymorphism within the 5′ untranslated region (5′-UTR). EDN1 gene located on chromosome 6 in humans, spans a 6.8 kb of genomic DNA, consisting of 5 exons and 4 introns. Each of the five exons encodes a different portion of the pre-pro ET-1¹⁰. Exon 1 encodes the precursor’s first 22 amino acids and the 5’ untranslated region (268 bp), including the entire signal sequence.

The + 139 ‘A’ polymorphism in the EDN1 5′-UTR is associated with elevated ET-1 levels and contributes to vasoconstriction, hypertension, and endothelial dysfunction in cardiovascular and pulmonary disorders. Studies on patients with chronic obstructive pulmonary disease (COPD) found not only a significant difference in ET-1 levels between individuals with the 3 A/3A (homozygous + 139 del ‘A’) and 3 A/4A (heterozygous + 139 Ins/del ‘A’) genotypes but also increased risk of COPD development in subjects carrying the 3 A/4A and the 4 A/4A (homozygous + 139 ins ‘A’) genotypes¹⁹. Our in vivo study in paediatric cases of congenital heart defects with severe pulmonary hypertension did not establish statistical significance for the homozygous + 139 ‘A’ insertion variant in the 5′UTR of the EDN1 gene and its association with elevated endothelin-1 expression. Consistent with previous findings, this limitation was primarily due to the rarity of the homozygous + 139 ‘A’ insertion variant, and the small sample size. The mortality among cyanotic infants was suggestive of the variant acting as a risk allele, with a potential contributory role in disease pathophysiology²⁰.

In this study we investigated the mechanisms underlying EDN1 gene expression associated with the rare homozygous + 139 ‘A’ insertion in the 5′-UTR. Plasma ET-1 levels were measured in CAD patients carrying the + 139 ins/del polymorphism. Functional effects were further examined in vitro using luciferase reporters driven by heterologous and endogenous EDN1 promoters. RNA pull-down coupled with MS/MS characterized variant specific RNA protein interactions, and molecular docking of a 50-nt 5′-UTR fragment with identified proteins compared binding modes between + 139 ins/del variants.

Results

Occurrence of + 139 ‘A’ ins. variant in the sample population

Sanger sequencing of the PCR-amplified 5′-UTR region of the EDN1 gene successfully identified the + 139 ‘A’ ins/del polymorphism in genomic DNA samples from CAD patients. Sequencing chromatograms were aligned against the EDN1 reference sequence (NCBI RefSeq: NG_016196.2) to assess nucleotide variation at the + 139-position relative to the transcription start site. Out of 95 patient samples analysed: 75 samples (79%) showed the wild-type genotype (3 A/3A), with no insertion at the + 139 position. 18 samples (19%) were heterozygous (3 A/4A) and displayed dual peaks (mixed A) at the + 139 site in the sequencing chromatogram indicating heterozygosity. 2 samples (2%) were homozygous for the insertion (4 A/4A), showing a clear adenine (A) peak at the + 139 position in both forward and reverse reads (Figure S1).

Plasma ET-1 levels and + 139 ‘A’ polymorphism

A quantitative sandwich ELISA was used to measure plasma endothelin-1 (ET-1) levels in 95 CAD patients. The assay demonstrated high sensitivity and reproducibility, with all samples analysed in duplicate and standard curves showing R² > 0.99. ET-1 concentrations varied markedly across genotypes. Individuals carrying the + 139 ‘A’ variant exhibited elevated ET-1 levels compared to wild-type homozygotes (3 A/3A) and heterozygotes (3 A/4A). However, no statistical tests of significance could be attributed to the observed ET-1 levels in individuals with 4 A/4A genotype because of its low frequency of occurrence in the study population. Median ET-1 levels for 3 A/3A, 3 A/4A and 4 A/4A genotypes were 6.6 pg/mL, 3.87 pg/mL and 16.02 pg/mL respectively (Fig. 2). These values revealed a statistically significant (p = 0.0234) difference in observed ET-1 levels between the 3 A/3A homozygous and 3 A/4A heterozygous groups with the heterozygous group showing lower plasma ET-1 levels compared to the wild-type group. The higher p-values of 0.2053 between 3 A/3A and 4 A/4A and 0.3435 between 3 A/4A and 4 A/4A indicate lack of true statistical significance because of only two observational values in the 4 A/4A group. In the absence of sizeable sample in the 4 A/4A group for ascertaining statistical significance of observed ET-1 levels compared to other groups, in vitro experiments were undertaken to further investigate the association, if any, of higher ET-1 levels with the homozygous insertion variant.

Fig. 2.

Plasma endothelin-1 (ET-1) levels measured in patient samples stratified by EDN1 5′-UTR genotypes: 3 A/3A, 3 A/4A, and 4 A/4A. Data represent comparative analysis of ET-1 concentration across different genotype groups. The horizontal bars represent the median values of observed circulating plasma endothelin levels. Statistical comparisons were performed using a non-parametric two-tailed Mann–Whitney U test; p < 0.05 was considered statistically significant.

Post-transcriptional regulation of EDN1 gene expression

To investigate the regulatory impact of the + 139 ‘A’ insertion within the 5′-UTR of the EDN1 gene, eight luciferase reporter constructs were generated. Fragments lacking the endogenous proximal promoter were cloned downstream of the SV40 promoter in the pGL3-control vector, whereas those retaining the native promoter region were inserted into the promoter-less pGL3-basic vector (Fig. 3a). Among these, four constructs represented the wild-type condition (WT) with a ‘3A’ sequence at positions 136–138, and the remaining four carried an additional adenine at position 139, generating a ‘4A’ stretch at positions 136–139, corresponding to the mutant variant (SDM). The results demonstrated that constructs containing the + 139 ‘A’ insertion exhibited significantly higher transcriptional activity compared to their respective wild-type counterparts (Fig. 3b). Compared to the wild-type (WT 5′-UTR), the mutant construct (SDM 5′-UTR) exhibited approximately 4-fold and 6-fold higher luciferase activity when tested with heterologous and endogenous promoters, respectively. Similarly, the mutant construct (SDM Exon 1) displayed a ~ 3-fold and ~ 5-fold increase in activity compared to the wild-type Exon 1 (WT Exon 1) in heterologous and endogenous promoters, respectively. Higher activity in mutant variants, indicates a potential role of the + 139 ‘A’ insertion in enhancing post-transcriptional activity of EDN1.

Fig. 3.

Luciferase reporter assay (a) Schematic representation of WT (3A) and SDM (4A) 5′-UTR constructs of the EDN1 gene cloned into the pGL3-control and pGL3-basic vectors, containing heterologous (SV40) and endogenous (EDN1) promoters, respectively.” (b) Relative luciferase activity of reporter constructs containing WT (3A) and SDM (4A) 5′-UTR variants of the EDN1 gene. Luciferase reporter assays were performed in HEK293T using both heterologous and endogenous promoters. Data represent fold change in luciferase activity normalized to the internal control (Renilla luciferase) and are presented as mean ± SEM from three independent experiments conducted in triplicates. A significant increase in luciferase activity was observed in the ‘4A’ variant compared to the ‘3A’ construct (*p < 0.05, Mann-Whitney test).

Identification of proteins involved in post-transcriptional regulation of EDN1 gene expression

To explore post-transcriptional regulation, RNA pull-down assay was performed with biotinylated oligo designed against the 5′-UTR regions of EDN1 gene. Biotinylated WT- ‘3A’ and SDM- ‘4A’ RNA oligonucleotides were incubated with whole cell lysate extracted from HEK293T and HUVEC cell lines. Biotin labelled 3 A oligo/4A oligo-protein complexes was precipitated by using streptavidin agarose beads. Following the RNA pull-down assay, potential interacting proteins specific to the WT- ‘3A’ and SDM- ‘4A’ constructs were identified via MS/MS and subsequently analysed using MaxQuant. To account for nonspecific binding, beads-only samples (lacking RNA) served as negative controls. Proteins significantly enriched (log2FC ≥ 1.5, p < 0.05) in 3 A/4A pull-downs were considered putative interactors. A total of 150, 238, 138, and 145 proteins were identified in HEK derived-biotinylated 3 A oligo, HEK derived- biotinylated 4 A oligo, HUVEC derived-biotinylated 3 A oligo, and HUVEC derived -biotinylated 4 A oligo RNA pull-down samples, respectively.

In HEK293T cells, 7 proteins were uniquely enriched with the WT-3 A RNA oligo, while 94 proteins were specifically associated with the SDM-4A oligo, indicating a significant shift in the protein interactome due to the insertion. Analysis of the 94 proteins associated with the 4A oligo in HEK293T cells revealed both direct RNA-binding proteins as well as components of larger ribonucleoprotein complexes, such as chaperones, ribonucleoproteins, RNA-binding proteins, isomerases, ribosomal proteins, hydrolases, and helicases, highlighting the complex network of RNA-associated interactions in HEK293T cells. The remaining proteins were shared between both 3 A and 4A pull-downs, suggesting a subset of interactions are independent of the insertion (Fig. 4a). Similarly, in HUVEC cells, 11 proteins interacted exclusively with the 3 A oligo, whereas 18 proteins were uniquely enriched in the 4A pull-downs, again supporting an insertion-induced alteration in the RNA–protein interaction landscape (Fig. 4b).

Fig. 4.

Comparison of RNA-binding proteins pulled down with 3 A and 4 A oligonucleotide probes in HEK293T and HUVEC cell lines. Venn diagrams depict the overlap of interacting proteins between the two oligo variants in (a) HEK293T cells, (b) HUVEC cells, and (c) across both cell lines, highlighting both shared and unique RNA–protein interactions associated with the + 139 ‘A’ insertion.

Comparative analysis of the 3 A oligo-specific interactomes revealed no shared proteins between HEK293T and HUVEC cells, suggesting a high degree of cell-type specificity in the RNA–protein interactions mediated by the wild-type 3 A insertion. A total of 94 unique proteins pulled down from HEK293T cells using the 4 A mutant RNA oligo were compared with 18 unique proteins identified from HUVECs under the same conditions. Among these, 7 were found to be unique shared proteins between the two cell lines (Table 1), indicating a subset of conserved RNA–protein interactions mediated by the 4 A insertion, despite overall cell-type specific differences (Fig. 4c).

Table 1.

Identification of 4 A oligo–bound proteins shared between HEK293T and HUVEC cell lines by RNA pull-down assay.

S. no	Accession no.	Description	Biological process
1	Q08211	ATP-dependent RNA helicase A*	Cell organization and biogenesis, DNA metabolism OR transcription, mRNA stability
2	P00558	Phosphoglycerate kinase 1	Protein metabolism; other metabolic processes; stress response; other biological processes, mRNA stability
3	P06748	Nucleophosmin	Cell cycle OR cell proliferation, cell organization and biogenesis, DNA metabolism,
4	P07237	Protein disulfide isomerase	Protein metabolism; other metabolic processes; stress response; transport; signal transduction; other biological processes
5	P51991	Heterogeneous nuclear ribonucleoprotein A3*	RNA metabolism or transcription
6	Q15084	Protein disulfide isomerase A6	Protein metabolism; other metabolic processes; stress response
7	P07737	Profilin-1	Cell organization and biogenesis; developmental processes; RNA processing and transcription

*Direct RNA binding proteins.

Pathway analysis of unique 4 A oligo-interacting identified proteins

To assess the impact of the + 139 ‘A’ insertion in the 5′-UTR, proteins uniquely identified in the HEK293-4 A and HUVEC-4 A RNA pull-down fractions were analysed. A total of 94 and 18 proteins detected in HEK and HUVEC respectively was subjected to Gene Ontology (GO) and KEGG pathway (KEGG Database Project, Kanehisa Laboratories, Japan Copyright permission Ref. No. 254378) enrichment analyses. The data generated from KEGG pathway has been attached in the supplementary file. The identified proteins were significantly enriched in pathways related to RNA binding/processing and protein folding. To delve deeper into the biological interactions among the identified proteins, molecular function (Gene Ontology) enrichment analysis was established via the STRING database (Fig. 5)²¹. As expected, mostly RNA-binding proteins were enriched, consistent with the use of a 5′-UTR RNA probe. Annotated proteins without nucleotide-binding activity suggested potential indirect interactions via RNA-binding protein complexes. Network analysis revealed two key components: one comprising proteins involved in RNA binding and mRNA processing, and another consisting of proteins associated with protein folding processes. Both are suggestive of modulation of EDN1 gene expression after the initiation of transcription but before translation into a protein by altering RNA–protein binding, complementing post- transcriptional regulation.

Fig. 5.

STRING-based gene set enrichment analysis of unique RNA pull-down proteins interacting with 4 A oligonucleotide probes in (a) HEK293T cell-line - Molecular Function (Gene Ontology) enrichment highlights functional clusters involved in RNA and ribonucleotide binding, mRNA processing, and protein folding. and in (b) HUVEC cell line- Molecular Function (Gene Ontology) enrichment analysis revealed functional clusters associated with ribonucleoprotein complexes and protein disulfide reductase activity. The interaction networks reveal both conserved and cell-type–specific modules associated with the + 139 ‘A’ insertion.

In silico analysis of the identified RNA binding proteins

To ensure biologically relevant interactions, proteins exhibiting RNA-binding capability were prioritized for docking studies. These included ATP-dependent RNA helicase A (DHX9) and Heterogeneous nuclear ribonucleoprotein A3 (HNRPA3) (Accession nos. Q08211 and P51991 respectively in Table 1). 50-nt oligo with + 139 ‘A’ insertion was docked with HNRPA3 and DHX9. The 50-nt oligo was modelled using RNAfold, which predicts the most thermodynamically stable secondary structure based on minimum free energy algorithms. In the predicted secondary structure, the 3 A and 4 A motifs of the modelled oligo were located within the loop region and remained unpaired (Figure S2). Protein–RNA docking was carried out using the HADDOCK 2.4 web server. In the initial rigid-body docking stage, 1000 orientations of the protein–RNA complex were generated. Of these, the 200 best solutions (based on intermolecular energy) were refined through semi-flexible simulated annealing and further refined in explicit solvent. The resulting docking models were clustered based on the fraction of common contacts (FCC), using a cut-off of 0.6. The top-ranking cluster was selected based on the HADDOCK score, which integrates van der Waals, electrostatic, desolvation energies, and buried surface area.

Docking analysis of HNRPA3 with the oligonucleotides identified Cluster 4 (3 A) and Cluster 7 (4 A) as the most favourable solutions, with average HADDOCK scores of 127.3 ± 3.3 kcal/mol and 106.8 ± 13.9 kcal/mol, respectively (Table 2). The HNRPA3–4 A complex demonstrated more favourable energy constraints compared to the HNRPA3–3 A complex. Interface analysis of the HNRPA3-4 A complex revealed residues Arg96, Val104, Lys126, Asp176, and His177 made six hydrogen bonds and three salt bridges with the RNA backbone at positions A’19–A’22. In contrast, the HNRPA3–3 A complex displayed a less extensive interface, consisting of five hydrogen bonds and a single salt bridge mediated by residues Glu106, Lys108, and Thr124 with nucleotides A’19–A’21 (Fig. 6a, b). Docking of DHX9 with the 3 A oligo identified cluster 7 as the top solution, with a HADDOCK score of 133.3 +/-23.2 kcal/mol and a Z-score of -1.7, representing the most reliable complex within the ensemble. In contrast, docking with the 4 A oligo yielded cluster 10 as the best model, with a more favourable HADDOCK score of 123.0 +/- 21.2 kcal/mol and a Z-score of -2.6, indicative of a more stable interaction as compared to DHX9-3 A complex (Table 2). DHX9-4 A complex analysis revealed extensive contacts, where residues Gln158, Lys266, Glu165, Gln240, Asn234, Lys235, Lys236, and Gly601 made five hydrogen bonds and six salt bridges with RNA backbone at positions A’19–A’22. DHX9-3 A complex showed fewer interactions, limited to five hydrogen bonds involving A’20–A’21 with Lys236, Gln240, Lys264, and Lys266, and two salt bridges between Glu165, Glu179 and A’20 (Fig. 6c, d).

Table 2.

Analysis of protein-RNA interaction obtained from HADDOCK calculation.

S. no	Protein-RNA complex	Haddock score	van der Waals energy	Electro-static energy	Desolvation energy	Restraints violation energy	Z score
1.	HNRPA3-3 A	127.3 ± 3.3	− 63.5 ± 7.2	− 215.6 ± 27.9	17.4 ± 1.0	2164.6 ± 110.1	− 1.7
2.	HNRPA3-4 A	106.8 ± 13.9	− 82.1 ± 9.2	− 272.5 ± 20.4	20.0 ± 2.9	2233.4 ± 102.9	− 2.3
3.	DHX9-3 A	133.3 ±23.2	− 62.6 ± 4.2	− 257.3 ± 34.1	19.7 ± 1.9	2277.0 ± 234.9	− 1.7
4.	DHX9-4 A	123.0 ± 21.2	− 58.0 ± 5.8	− 426.5 ± 22.7	23.9 ± 3.2	2424.1 ± 134.3	− 2.6

Fig. 6.

Intermolecular interactions in protein–oligo complexes predicted by HADDOCK. (a) HNRPA3 (blue)-3 A oligo (yellow) (b) HNRPA3(blue) -4 A oligo (Orange) (c) DHX9 (Green)-3 A oligo (yellow) (d) DHX9 (Green)-4 A oligo (Orange).

Discussion

Dysregulated gene expression, including that of ET-1, is a root cause of many diseases. The robust link between raised plasma endothelin-1 levels and patients with hypertension was reported way back in 1990 in New England Journal of Medicine by Imura et al.²². Since then, association of dysregulated increase in plasma endothelin-1 levels have been reported in several diseases⁶ including CVD where ET-1 plays an important role in early initiation of endothelial dysfunction. This study identifies and functionally characterizes a regulatory variant - a single adenine (‘A’) insertion at position + 139 within the 5′-UTR of the EDN1 gene with the possibility of relevance in disease prognosis. The clinical dataset of 95 CAD patients with the identified + 139 ‘A’ ins/del polymorphism in the 5′-UTR of the EDN1 gene revealed that majority of individuals carried the wild-type (3 A/3A) genotype, while a smaller proportion exhibited the heterozygous (3 A/4A) and only two individuals carried the homozygous insertion (4 A/4A) variant. Plasma ET-1 levels were lower in individuals with 3 A/4A compared to the 3 A/3A genotype, which was counterintuitive, perhaps due to allele-specific expression imbalance where one allele 3 A is preferentially transcribed, or the moderating influence of heterozygosity in lowering of threshold endothelin levels^23,24. In addition, the + 139 A insertion may affect mRNA secondary structure or stability, leading to reduced translation efficiency in heterozygotes while permitting higher expression in homozygotes. This pattern may also be influenced by clinical confounders such as medication use, disease severity, or comorbidities, as well as by sampling variability due to small subgroup sizes in the 4 A/4A group. The + 139-insertion homozygous 4 A/4A condition was observed in only two individuals making it difficult to draw inferences in statistical terms but, nevertheless, raising curiosity for probing deeper, since minor variants causing CVD are associated with rare, severe and early onset-conditions²⁵. For better understanding of the observed endothelin levels in the individuals with 4 A/4A genotype, it became crucial to assess the functional effects of the single homozygous insertion of adenine ‘A’ at + 139 position using in vitro luciferase assays.

Experiments using luciferase reporters showed that the + 139 'A' insertion significantly increases the post-transcriptional activity of the EDN1 in 5’UTR constructs with both native and heterologous promoters. The 4A mutant constructs exhibited elevated luciferase activity compared to the wild-type constructs 3A. This indicates that the inserted nucleotide may enhance the transcript’s abundance or translation efficiency, potentially by altering the RNA’s secondary structure or altering RNA-protein interactions. It has been reported that the proximal promoter region of the EDN1 gene is highly conserved across mammalian species, including humans, mice, and other mammals²⁶. This strong conservation shows that the regulatory elements in the region between ~ − 136 and − 96 bp upstream of the transcription start site are crucial for controlling the basal or stimulus induced transcription. The 5′-UTR primarily governs post-transcriptional mechanisms, including translation; however, it may also contain promoter-proximal or regulatory elements that affect transcription initiation. It can indirectly influence transcription through promoter overlap, transcription factor binding, or co-transcriptional RNA structures. At the translational level, the 5′-UTR may affect the secondary structures, upstream open reading frames (uORFs), and interactions with RNA-binding proteins^27,28. Alterations in 5’-UTR regions might disrupt these regulatory elements, which would change the genes expression and lead to disease phenotypes²⁹. The elevated luciferase activity observed in the 5′-UTR mutant variants suggests that the + 139 A/A insertion may enhance EDN1 expression. Considering, the limitation in assessing the statistical significance of this variant for promoter activity in clinical samples, the results of luciferase assay provide more relevant evidence for the regulatory impact of the insertion of a single adenine in 5’UTR of EDN1 gene.

RNA pull-down coupled with MS/MS demonstrated that the + 139 ‘A’ insertion in the EDN1 5′-UTR modifies the RNA-protein interactome, with the 4 A variant recruiting more proteins than the 3 A form in both HEK293T and HUVEC cells. This indicates improved RNA accessibility or the formation of new binding motifs. Functional enrichment analyses indicated involvement in RNA splicing, mRNA processing, and protein folding; crucial regulators of transcript destiny and translation. The identification of DExH-box helicase 9 (DHX9) among the proteins isolated with 4 A 5′-UTR oligo probes underscores its potential function in regulating EDN1 gene expression at the post-transcriptional stage. DHX9 is a helicase that resolves RNA secondary structures like G-quadruplexes and double-stranded RNA, which affects splicing, stability, and translation³⁰. Its preferential binding to the mutant 5′-UTR probe may suggest that the + 139 ‘A’ insertion stabilises or creates distinct RNA secondary structures that improve DHX9 recognition, potentially resulting in enhanced translational efficiency. This is in line with the luciferase reporter data, which showed that constructs with the 4 A variant have higher luciferase activity. This suggests that the changes in RNA-protein interactions have a functional effect. HNRPA3, a member of the hnRNP A/B family, was also found to be an important differential interactor. HNRPA3 is important for processing pre-mRNA, alternative splicing, moving mRNA, and keeping it stable³¹. Its preference for binding to the 5′-UTR probe suggests that it plays a role in regulating the expression of EDN1 after transcription. It is known that hnRNPs bind to UTRs in a way that depends on the sequence and structure. The insertion might prefer motifs that are important for HNRPA3 interaction, which could affect the half-life of mRNA or the start of translation. Also, docking studies predicted higher binding affinity of DHX9 and HNRPA3 with the “4A” oligo as compared to “3A” probe. This is likely because the electrostatic interactions and hydrogen bonding were improved.

The EDN1 gene expression is affected by many factors and a variety of stimuli have been shown to contribute towards it but the specific signalling mechanisms remain inadequately characterised. The capacity of the EDN1 gene to respond to diverse hormonal and environmental cues is essential for ensuring the spatial, temporal, and quantitative fidelity of ET-1 expression. Ultimately, these signalling cascades converge on regulatory elements within the EDN1 promoter and untranslated regions to modulate gene activity. In the present study, we specifically focus on RNA–protein interactions governed by features in 5’UTR region of the EDN1 gene that may perhaps be involved in modulating ET-1 expression levels post-transcriptionally, while not addressing the multitude of other factors that may contribute to altered expression and circulating ET-1 levels in physiological or diseased states. It is our speculation that structural element within the 5’untranslated region of EDN1 mRNA was modulating post-transcriptional expression of ET-1 by interacting with RNA Binding Proteins (RBPs) that may perhaps stabilize or destabilize secondary structures in the 5’UTR which control accessibility by ribosomes. Thus, the RNA-protein interactions may modulate EDN1 mRNA fate and protein production. This exploratory work aims to provide initial insights into the regulatory role of a rare 5′-UTR variant, which may be under negative selection pressure and carry potential pathophysiological significance.

We propose that the + 139 ‘A’ insertion in the EDN1 5′-UTR modifies RNA-protein interactions, increasing mRNA processing and translation, thereby elevating ET-1 levels and cardiovascular risk. While a significant correlation with homozygous + 139 ‘A’ insertion could not be established in CAD patients due to limited sample size, it is conceivable that the differential EDN1 expression established through in vitro assays is associated with this variant may facilitate pharmacogenomic stratification and enable personalised therapeutic interventions in cardiovascular disease. In conclusion, the study highlights the functional significance of the + 139 ‘A’ insertion in EDN1 expression which affect transcript destiny and translational regulation.

Materials and methods

SNP analysis

Peripheral blood samples were collected from 95 patients diagnosed with coronary artery disease (CAD) at Cardiac- thoracic and Neurosciences Centre AIIMS based on clinical and angiographic criteria Table S1. Peripheral blood mononuclear cells (PBMCs) were isolated from whole blood using Ficoll-Paque density gradient centrifugation. Genomic DNA was extracted from PBMCs using the QIAamp DNA Mini Kit (Qiagen, Germany) according to the manufacturer’s protocol. DNA purity and concentration were assessed using a NanoDrop spectrophotometer (Thermo Scientific, USA), and integrity was verified on 1% agarose gel electrophoresis. A fragment of the EDN1 gene encompassing the + 139 site in the 5′-UTR was amplified using gene-specific primers designed via Primer3. Each 25 µL PCR reaction included 50 ng genomic DNA, 0.2 µM forward and reverse primers, 1× PCR buffer, 1.5 mM MgCl₂, 0.2 mM dNTPs, and 1 U of Taq DNA polymerase (Thermo Scientific). The PCR conditions were: initial denaturation at 95 °C for 5 min, 35 cycles of denaturation at 95 °C for 30 s, annealing at primer-specific temperature for 30 s, and extension at 72 °C for 30 s, followed by a final extension at 72 °C for 5 min. Amplified products were visualized on 1.5% agarose gel and purified using the QIAquick PCR Purification Kit. Purified amplicons were subjected to Sanger sequencing, and the + 139 ins/del in the 5′-UTR of EDN1 was determined by aligning the sequences to the EDN1 reference sequence (NCBI RefSeq: NM_001955, NCBI RefSeq: NG_016196.2). The polymorphism within the 5′-UTR of the EDN1 gene with the + 139-insertion allele is designated as ‘4A’ (because of the stretch of 4As at positions 136, 137,138 and 139) and the deletion allele as ‘3A’. These allelic variants can occur in homozygous form (3 A/3–4 A/4A) or in heterozygous form (3 A/4A).

Quantification of plasma endothelin-1 (ET-1) by enzyme immunoassay (EIA)

Plasma endothelin-1 (ET-1) levels were measured using a commercially available competitive enzyme immunoassay (EIA) kit (Invitrogen Cat # EIAET1), which has a sensitivity of 0.579 pg/mL and a detection range of 0.781-100 pg/mL Briefly, the samples and ET-1 standards were added to wells pre-coated with ET-1 antiserum. An ET-1-acetylcholinesterase conjugate was added, and the plate was incubated at overnight at 4 °C. After incubation, plates were washed and developed with Ellman’s reagent. Absorbance was read at 412 nm using a microplate reader and ET-1 concentrations were calculated from the standard curve for all three conditions. All samples, including the 3 A/3A, 3 A/4A, and 4 A/4A genotypes, were analysed in duplicate, and the mean values were used for statistical analysis.

Cloning of EDN1 gene fragments in reporter vectors

The Homo sapiens EDN1 reference sequence on chromosome 6 (accession no. NG_016196.2, coordinates 38,878–45,711) was retrieved from the National Center for Biotechnology Information (NCBI) database (www.ncbi.nlm.nih.gov). A 257 bp region of the EDN1 proximal promoter (coordinates NG_016196.2_38,621–38,878) upstream of the transcription start site (TSS) was included with the 5′-UTR (267 bp) and the complete exon 1 region (333 bp) for the endogenous promoter reporter construct generation into the pGL3-Basic vector. PCR-amplified fragments containing the 5′-UTR (267 bp) and the exon 1 region (333 bp), excluding the endogenous proximal promoter, were cloned downstream of the heterologous SV40 promoter in the pGL3-Control vector (Figure S3). Site-directed mutagenesis (SDM) was then employed to introduce an additional adenine at the + 139 position, generating the mutant (4 A) variants of both the 5′-UTR and exon 1 constructs. These wild-type and mutant constructs were subsequently used in luciferase reporter assays to evaluate the regulatory contribution of the 5′-UTR to EDN1 expression, with and without the + 139 ‘A’ insertion.

Cell culture

Human Embryonic kidney 293T cells (HEK293T cells; ATCC Catalog No. CRL-3216 ™) were revived and grown in monolayer culture in Dulbecco’s modified eagle medium (DMEM) (Lonza Cat no: 12604 F) supplemented with 10% FBS (Thermo Cat: 10270106) with 4.5 g/L glucose, 1mM L-glutamine, and 100 µg/ml antibiotic penicillin-streptomycin (Merck & Co., Cat no:211–500) in a humidified incubator (5% carbon dioxide, 95% air) at 37 °C. The cloned constructs were transfected in HEK293T cells with Lipofectamine 2000 (Thermo Scientific, Madison, USA) as described in manufacturer’s manual. Human umbilical vein endothelial cells (HUVEC) were isolated from umbilical cords obtained with informed consent under sterile conditions. Veins were flushed with PBS and digested using 0.2% collagenase type I/II at 37 °C for 30–60 min. The cell suspension was centrifuged at 300–400 × g for 10 min, and the pellet was cultured in vascular cell basal medium supplemented with 100 µg/ml penicillin-streptomycin at 37 °C in a 5% CO₂ incubator.

Reporter assay

A dual luciferase reporter assay was performed to assess the regulatory activity of the wild-type ‘3A’ and mutant ‘4A’ constructs. HEK293T cells were co-transfected with two plasmids: one containing the firefly luciferase gene under the control of the experimental wild-type WT or mutant SDM 5′-UTR/Exon 1 regions cloned into either pGL3-control or pGL3-basic vectors, and the other carrying the renilla luciferase gene driven by the thymidine kinase promoter (pRL-TK), serving as an internal control. The pGL3-basic vector (a promoter less firefly luciferase reporter) and the pGL3-control vector (expressing firefly luciferase under the SV40 promoter) were also transfected in parallel and served as negative and positive controls, respectively. The pRL-TK vector was co-transfected at a molar ratio of 1:10 (Renilla: Firefly) with all constructs to normalize for transfection efficiency and cellular variability. Following 48 h of incubation, firefly and renilla luciferase activities were quantified using a dual-luciferase reporter assay system (Promega, USA) according to the manufacturer’s instructions. Firefly luciferase activity was normalized against renilla luciferase activity to control for transfection efficiency. Each transfection was carried out in triplicate, and the experiments were independently repeated a minimum of three times to ensure reproducibility.

RNA-protein pull down assay

To investigate the impact of an adenine insertion at position + 139 in the 5′-UTR of the EDN1 gene, 50-nucleotide synthetic RNA oligonucleotides were designed to represent the wild-type (‘3 A’) and mutant (‘4 A’) sequences. Secondary structure predictions and minimum free energy calculations were performed using RNAstructure³² and RNAfold³³ to ensure preservation of native structural features. These oligos were used in RNA pull-down assays to assess differential binding of RNA-binding proteins, providing insight into regulatory mechanisms influencing ET-1 gene expression³⁴. The RNA oligos were biotinylated at the 3′ end using the RNA 3′ End Desthiobiotinylation Kit (Thermo Fisher). Biotinylated wild-type (WT) and mutant (SDM) RNA oligonucleotides were incubated with whole-cell lysates (HEK293T and HUVEC) in binding buffer at 4 °C overnight to facilitate RNA–protein interactions. Streptavidin-coated magnetic beads were then added, and the mixture was further incubated for 4 h at 4 °C to allow efficient capture of RNA–protein complexes. Beads were subsequently washed 2–3 times with wash buffer to remove non-specific interactions. Elution of bound complexes was performed using elution buffer with gentle mixing for 30 min at 4 °C. Eluted protein concentration was measured with Pierce BCA protein assay kit. The precipitated interacting proteins were then analysed by MS/MS. Streptavidin magnetic beads alone were served as negative controls. The resulting peptides were analysed using a mass spectrometer (thermo scientific orbitrap MS). The mass spectrometry data have been deposited to the ProteomeXchange Consortium via the PRIDE^35,36 partner repository with the dataset identifier PXD076288. Raw MS data were analysed using Mascot version 2.4.01 (Matrix Science, London, UK), and protein identifications were obtained by searching against the human RefSeq protein database (NCBI). Biological process annotation and pathway enrichment of proteins interacting with the WT and SDM oligos were performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID, v6.7)^37,38 and the Kyoto Encyclopedia of Genes and Genomes (KEGG)^39–41. Protein–protein interaction networks were generated using STRING (version 10), applying a high-confidence interaction threshold (combined score > 0.9)²¹.

Bioinformatics analysis

Protein–RNA docking was carried out using the HADDOCK2.4 webserver (High Ambiguity Driven Biomolecular DOCKing)⁴². The docking protocol consisted of a standard three-stage workflow: rigid-body docking, semi-flexible refinement, and final refinement in explicit solvent. Initially, 1000 complex orientations were generated through rigid-body energy minimization (it0 stage), guided by Ambiguous Interaction Restraints (AIRs), which included manually defined active and passive interface residues. The 200 best-scoring models based on HADDOCK’s intermolecular energy function were selected for the next stage. These top models underwent semi-flexible simulated annealing refinement, in which interface residues were allowed to move to optimize interface packing and improve interaction energetics. This step enabled local conformational rearrangements around the binding site. In the final stage, the refined models were subjected to explicit solvent (water) refinement (water stage) to further optimize energetics and solvation effects under more realistic conditions. The resulting structures were clustered based on pairwise backbone RMSD calculated over the interface residues. Clusters were evaluated and ranked according to their average HADDOCK scores, which are weighted combinations of van der Waals, electrostatic, desolvation, and AIR restraint energies. The top-ranked cluster, exhibiting the lowest average HADDOCK score, was selected as the final docked model.

Statistical analysis

Statistical analyses were performed using GraphPad Prism (v7.02; GraphPad Software, La Jolla, CA, USA). Plasma levels of ET1 were compared between homozygous 3 A/3A and heterozygous 3 A/4A using the nonparametric Mann–Whitney U test. All tests were two-tailed and unpaired, with statistical significance defined as P < 0.05. Normalized luciferase activities were expressed as median with interquartile range (IQR). Statistical comparisons between wild-type and mutant constructs were performed using the nonparametric two-tailed Mann–Whitney U test.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ET-1

Endothelin-1

EDN1

Endothelin-1 gene

CAD

Coronary artery disease

SNPs/SNVs

Single nucleotide polymorphisms/Single Nucleotide Variants

5′UTR

5′ untranslated region

Ins./del.

Insertion/deletion

HEK293T

Human embryonic kidney (293-T) cells

HUVEC

Human umbilical vein endothelial cells

DHX9

DExH-box helicase 9

HNRPA3

Heterogeneous nuclear ribonucleoprotein A3

Author contributions

E.S. writing-original draft, Figure 3a, data analysis, review, editing, H. Proteomics studies, T.D. Cloning and dual luciferase assays , Figure 1, K.Y. Cell culture and dual luciferase assays, R.G. SNP profiling of the CAD patients , D.P. Bioinformatics analysis , A.G. designed the research, supervision, resources, data analysis, manuscript editing and proofreading.

Funding

The financial support for the study was provided by Indian Council of Medical Research (ICMR) through the grant Project IDs: 5/4/1–1//CVD/2022-NCD-I, 5/4/1–4/07-NCD-II and N-29 Grant for “M.Sc. Medical Biotechnology Teaching Programme” by Department of Biotechnology (DBT), Ministry of Science and Technology, and Government of India. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Data availability

The mass spectrometry data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD076288. The relevant accession numbers used in the current study for DNA and RNA sequences, and the SNP under study are included in the body of the text appropriately.

Declarations

Competing interests

The authors declare no competing interests.

Ethics declarations

This study received ethical approval from the Institute Ethics Committee at The All India Institute of Medical Sciences, New Delhi-110029, INDIA (Approval IEC Ref No.: IEC-383/06.05.2022) The study was conducted in accordance with ICMR and GCP guidelines. Written informed consent was obtained from all participants prior to their participation in the study. Safety Precautions for performing cloning experiments and handling cell-lines in BSL-2 facility received approval from Institutional Committee on Biosafety for Recombinant DNA Research at All India Institute of Medical Sciences, New Delhi-110029, INDIA (Approval Ref. No. IBSC 0422_AG dated 26/04/2022.