Identification of ATP2B4 Regulatory Element Containing Functional Genetic Variants Associated with Severe Malaria

Abstract

Genome-wide association studies for severe malaria (SM) have identified 30 genetic variants mostly located in non-coding regions. Here, we aimed to identify potential causal genetic variants located in these loci and demonstrate their functional activity. We systematically investigated the regulatory effect of the SNPs in linkage disequilibrium (LD) with the malaria-associated genetic variants. Annotating and prioritizing genetic variants led to the identification of a regulatory region containing five ATP2B4 SNPs in LD with rs10900585. We found significant associations between SM and rs10900585 and our candidate SNPs (rs11240734, rs1541252, rs1541253, rs1541254, and rs1541255) in a Senegalese population. Then, we demonstrated that both individual SNPs and the combination of SNPs had regulatory effects. Moreover, CRISPR/Cas9-mediated deletion of this region decreased ATP2B4 transcript and protein levels and increased Ca²⁺ intracellular concentration in the K562 cell line. Our data demonstrate that severe malaria-associated genetic variants alter the expression of ATP2B4 encoding a plasma membrane calcium-transporting ATPase 4 (PMCA4) expressed on red blood cells. Altering the activity of this regulatory element affects the risk of SM, likely through calcium concentration effect on parasitaemia.

Keywords: ATP2B4; CRISPR-cas9; SNP; calcium; enhancer; functional genomics; gene reporter; malaria; promoter; regulatory element.

Figure 1

Plot displaying the integrated results for the prioritization of best candidate SNPs. IW scoring method prioritizes functionally relevant noncoding variants. The graph displays the IW score ranks of 125 candidate SNPs. The number of ChIP-seq peaks was extracted from ReMap, which integrates the results of transcription factor ChIP-seq experiments. The graph displays the number of DNA-binding protein ChIP-seq peaks for each candidate SNP. The SNPs with the best IW score rank had the highest number of ChIP-seq peaks.

Figure 2

Visual representation of the ATP2B4 locus and the epigenomic marks of the region containing the five ATP2B4 candidate variants. (a) Transcripts, TSS peaks, and studied SNPs are displayed. The two long transcripts (ATP2B4-203 and ATP2B4-204) and the short transcripts are displayed. The main TSS peaks are visualized. (b) The candidate SNPs (rs11240734, rs1541252, rs1541253, rs1541254, and rs1541255; encircled red) are on chromosome 1, located within a DNaseI hypersensitivity region and peaks of H3K4me3, H3K4me1, and H3K27ac histone marks. The two tagSNPs (rs10900585 and rs4951377; encircled green) are located neither in peaks of ChIP-seq in the ReMap catalog nor in other epigenomic marks. The three additional regulatory variants (rs10751450, rs10751451, and rs10751452; encircled blue), previously identified as functional SNPs, are displayed.

Figure 3

Forest plot displaying meta-analysis for the association between rs10900585 and SM.

Figure 4

Linkage disequilibrium (LD) (r-squared) plot for the ATP2B4 variants in the Senegalese population. (a) LD between the candidate five SNPs and the tagSNP rs10900585 (b) LD between the eight candidate SNPs and the tagSNP rs10900585 displaying r² from 0.72–1.

Figure 5

Description of the genomic regions used or deleted for luciferase gene reporter assay and CRISPR-Cas9 genome editing, respectively. The main ATP2B4 promoter is illustrated in blue, whereas the alternative promoter containing the five SNPs studied is illustrated in red. For the luciferase gene reporter assay, the 601 bp genomic region was cloned first into the promoter position in pGL3-basic vector with the minor or major alleles, and then into the enhancer position in pGL3-Enhancer vector using the 780 bp genomic region as a promoter. For CRISPR-Cas9 editing a 506 bp fragment containing the five SNPs was deleted using gRNA1 and gRNA2. We also deleted a larger region containing eight SNPs (our five SNPs and three SNPs studied by Lessard et al.) using gRNA3 and gRNA2.

Figure 6

Luciferase reporter assay accessing the promoter and enhancer activity of ATP2B4 variants in K562 cells. (a) Promoter activity of the DNA region containing ATP2B4 Major and Minor haplotypes in addition to individual polymorphism. (b) Enhancer activity of ATP2B4 SNPs with Major and Minor haplotypes. SV40 promoter and basic vectors were used as positive and negative controls, respectively. Data presented as the mean ± standard deviation of three independent experiments. **** p < 0.0001, *** p < 0.001, ** p < 0.01, * p < 0.05.

Figure 7

CRISPR-Cas9 mediated genome editing. (a) Principal strategy for the generation of DNA knockouts. gRNAs (gRNA1, gRNA2, and gRNA3) were designed flanking the genomic target to delete the two targeted DNA segments comprising either the five (506 bp) or the eight (1262 bp) candidate SNPs, creating double-strand breaks (DSBs) at 3 bp upstream of the PAM. The resulting DSB is repaired by the NHEJ pathway. The genomic deletion is detected by PCR using primers (F1, R1) and (F2, R2) providing the amplicons of 335 bp and 355 bp, respectively. (b) qPCR analysis of gene expression in wild-type K562 cells and the deleted ATP2B4 clones (for the five candidate SNPs) to quantify either all the transcripts or only the two long transcripts of ATP2B4. (c) qPCR analysis of gene expression in wild-type K562 cells and the deleted ATP2B4 clones (for the eight SNPs) to quantify either all the transcripts or only the two long transcripts of ATP2B4. Error bars illustrate standard deviation (n = 3 independent RNA/cDNA preparations): **** p < 0.0001.

Figure 8

Visualization of calcium concentration and PMCA4 protein expression of K562 cells. (a,b) Visualization of calcium concentration of K562 WT (pink) and deleted cells (blue). An increased intracellular calcium concentration was displayed in clone ΔATP2B4_2 (a) and for clone ΔATP2B4_3 (b), respectively. (c–e) PMCA4 protein expression using flow cytometry analysis. Representative FACs plot illustrating isotype (blue) and JA9 (pink) peaks visualized by an APC conjugated secondary antibody in K562 WT cells (c), ΔATP2B4_2 deleted clone (d), and ΔATP2B4_3 deleted clone (e). These were visualized using FlowJo software.

Figure 9

A model for transcriptional regulation of ATP2B4 gene. The newly identified region containing the five SNPs studied (rs11240734, rs1541252, rs1541253, rs1541254, and rs1541255) corresponds to an alternative promoter with an enhancer function. Its activity is modulated by the combination of SNPs. The major haplotype had a stronger enhancer effect on the main ATP2B4 promoter, increasing transcription of long transcripts, whereas the minor allele haplotype had stronger promoter activity, resulting in an increase in short transcripts.