Fig 1.
Characterization of the allelic heterogeneity of the 33-bp VNTR in the human TRIB3 promoter.
(A) A schematic of the TRIB3 gene 5’ region showing the location of VNTR, containing 3 copies of the 33-bp repeat as in the human reference genome (GRCh38). (B) PCR analysis of the TRIB3 promoter 33-bp repeat copy number in human cell lines. The DNA of hepatocellular carcinoma (HepG2), colorectal adenocarcinoma (Colo205), non-small cell lung carcinoma (H1299), osteosarcoma (SaOS2) and chronic myelogenous leukemia (K562) was amplified using primers flanking the VNTR and the products were separated by agarose gel electrophoresis. (C) Human TRIB3 promoter 33-bp VNTR allele frequencies in Estonian Biobank participants, based on whole genome sequencing data. (D) PCR analysis of the 33-bp repeat copy number in a selection of Estonian Biobank participants. The DNA was analyzed as in panel B, and HepG2 is shown as a control. Lane M in panels B and D contains a molecular weight marker (100 bp DNA ladder).
Fig 2.
Evolutionary conservation of the TRIB3 promoter 33-bp VNTR area across mammalian species and the geographical distribution of the TRIB3 33-bp VNTR alleles in human populations.
(A) Nucleotide sequence alignment of the human TRIB3 promoter 33-bp VNTR repeat unit and flanking areas with the corresponding sequence in various mammalian genomes. Only one copy of the repeat is shown for modern human and Denisovan and Neanderthal hominins. Alignment was produced using Clustal Omega and Boxshade, with identical nucleotides shown in black and conserved purines in grey. The location of the C/EBP-ATF binding site is highlighted, overlapping a highly conserved sequence element. (B) A map showing the allele frequency distribution of the human TRIB3 promoter 33-bp VNTR repeat in various populations around the world. The results are based on data from the Human Genome Diversity Project (HGDP). (C) Multi-allelic fixation index (FST) for TRIB3 33-bp repeat copy number alleles in HGDP superpopulations.
Fig 3.
Human TRIB3 promoter variants containing more copies of the 33-bp repeat demonstrate elevated transcriptional activity in luciferase reporter assays.
(A) Results from HepG2 cells transfected with human TRIB3 promoter reporter constructs containing two (TRIB3-2x33-pGL3) or five (TRIB3-5x33-pGL3) copies of the 33-bp repeat. Cells were either left untreated or treated with tunicamycin (2 μg/ml, 10 h) or arsenite (20 μM, 10 h), or subjected to glucose deprivation (10 h). Luciferase activities are shown as mean ± SD from 4 independent transfections. *P < 0.05 comparing allelic variants using two-tailed t test followed by Holm–Bonferroni correction. (B) Reporter constructs carrying 1 to 5 copies of the 33-bp repeat were transfected into H1299 cells and their transcriptional activity was assayed under control condition (untreated), oxidative stress (30 μM arsenite, 10 h), and with co-transfection of an ATF4 overexpression plasmid (ATF4-pCG). Four independent transfections were performed, and the relationship between 33-bp copy number and luciferase activity was modeled using linear regression. For both panels, luciferase activities are presented relative to the activity of empty reporter vector (pGL3-Basic without TRIB3 fragment) under the same conditions. The genomic coordinates (GRCh38 human reference genome) of the region studied in the reporter assay are indicated below the plasmids.
Fig 4.
TRIB3 33-bp repeat copy number is positively correlated with the expression level of TRIB3 in genome-wide gene expression profiling.
(A) Validation of a k-mer counting-based method for the estimation of TRIB3 33-bp repeat copy number from whole genome sequencing data. The k-mer based copy number estimates are compared to the results of PCR product gel analysis from the same individuals. The PCR results are shown as the average repeat copy number from the diploid genotype of each individual. (B) Correlation of TRIB3 33-bp copy number to TRIB3 gene expression in the subset of individuals with PCR-validated 33-bp repeat genotypes. Individuals from panel A that further had blood RNA-Seq data available are included. (C) Across all genes, the copy number of the TRIB3 33-bp repeat is correlated specifically with the expression level of TRIB3 mRNA. Correlation between gene expression and the copy number of the TRIB3 33-bp repeat was analyzed for all genes expressed in blood, as determined by RNA-Seq (total 13896 genes). For the panels of this figure, the genotyping samples and the DNA and RNA sequencing data originate from the Estonian Biobank.
Fig 5.
Variation in the TRIB3 33-bp repeat copy number is linked to TRIB3 expression level in various tissues of the body.
(A) Correlation between TRIB3 33-bp repeat copy number and TRIB3 gene expression level in 53 tissues from the Genotype-Tissue Expression (GTEx) v8 data set. The analysis included all tissues with ≥5 donors with RNA-Seq and whole genome sequencing from the same donor. The k-mer counting method was used to estimate the TRIB3 33-bp repeat copy number from genome sequencing reads. False discovery rate was used to correct the p values. (B) Scatter plots from selected tissues from panel A, showing the correlation between TRIB3 33-bp repeat copy number and TRIB3 mRNA expression at the level of individuals donors. (C) The strength of the association between 33-bp repeat copy number and TRIB3 expression level in different tissues (the correlation p value) does not appear to be driven strictly by the available sample size. All tissues from panel A are shown, and a selection of tissues with >400 samples are labeled. (D) Linkage disequilibrium results for SNPs that are TRIB3 eQTLs in multiple tissues. The eQTL tissue counts were obtained from the GTEx v8 data release. LD was calculated from GTEx donor genotypes. (E) Known SNP eQTLs for TRIB3 expression level demonstrate associations with the TRIB3 33-bp repeat copy number. Statistical significance was calculated by linear regression for the SNP allele effect.
Fig 6.
TRIB3 SNPs highlighted by genetic association studies are linked to 33-bp repeat copy number as well as TRIB3 mRNA expression in phenotype-relevant tissues.
SNPs localizing near the TRIB3 33-bp repeat and associated (or suggestive) for human phenotypes based on GWAS, PheWAS or other genetic association studies are shown. Using RNA-Seq and whole genome sequencing data from GTEx v8, the association between the lead SNP genotype and TRIB3 RNA expression level or TRIB3 33-bp repeat copy number was calculated using linear regression for the SNP allele effect. The k-mer counting method was used to estimate the TRIB3 33-bp repeat copy number from genome sequencing reads.