Disease-associated non-coding variants alter NKX2-5 DNA-binding affinity

Abstract

Genome-wide association studies (GWAS) have mapped over 90 % of disease- or trait-associated variants within the non-coding genome, like cis-regulatory elements (CREs). Non-coding single nucleotide polymorphisms (SNPs) are genomic variants that can change how DNA-binding regulatory proteins, like transcription factors (TFs), interact with the genome and regulate gene expression. NKX2-5 is a TF essential for proper heart development, and mutations affecting its function have been associated with congenital heart diseases (CHDs). However, establishing a causal mechanism between non-coding genomic variants and human disease remains challenging. To address this challenge, we identified 8475 SNPs predicted to alter NKX2-5 DNA-binding using a position weight matrix (PWM)-based predictive model. Five variants were prioritized for in vitro validation; four of them are associated with traits and diseases that impact cardiovascular health. The impact of these variants on NKX2-5 binding was evaluated with electrophoretic mobility shift assay (EMSA) using purified recombinant NKX2-5 homeodomain. Binding curves were constructed to determine changes in binding between variant and reference alleles. Variants rs7350789, rs7719885, rs747334, and rs3892630 increased binding affinity, whereas rs61216514 decreased binding by NKX2-5 when compared to the reference genome. Our findings suggest that differential TF-DNA binding affinity can be key in establishing a causal mechanism of pathogenic variants.

Keywords: Binding affinity; Gene regulation; Non-coding variants; Transcription factors.

Figure 1:

Identification of disease/trait-associated non-coding SNPs affecting NKX2-5 binding. Predicted SNPs from the 1000 Genomes Project were intersected with disease-associated variants from the GWAS catalog.

Figure 2:

Expression and purification of functional NKX2-5 homeodomain. A) SDS-PAGE (left) and Western blot (right) of purified NKX2-5 homeodomain after Ni-NTA affinity chromatography. B) Electrophoretic mobility shift assay (EMSA) of known binding site within the ANF promoter. C) Binding curve analysis of NKX2-5 homeodomain. Data points represent the average value of triplicate measurements and error bars the standard error.

Figure 3:

NKX2-5 binding to reference (ref) and variant (alt) sequences determined through EMSA. (top left) Representative EMSA gel used for binding curve analysis for rs6121514. Binding experiments were performed in triplicates. Binding curves show average X (bound fraction) and error bars are standard error. Parameters of the non-linear regressions are reported in Supplementary Table 1.