The endothetal Per-Arnt-Sim (PAS) domain protein 1 (EPAS1) gene which encodes hypoxia-inducible-factor-2 alpha (HIF2a) is a transcription factor that is involved in the response to hypoxia. Hypoxia is a major geographical condition associated with high-altitude environments ¹. Hypoxia-inducible-factors (HIFs) are heterodimers consisting of an oxygen-labile HIFa subunit and a stable HIFb subunit ². During hypoxia conditions, three isoforms either HIF1a, HIF2a or HIF3a and HIF1b are activated and function as transcriptional regulators of genes involved with the hypoxia response ^{3, 4, 5}. Genome wide association studies (GWAS) on high-altitude adaptation have implicated several single nucleotide polymorphisms (SNPs) in the regulatory region of the EPAS1 gene which are responsible for the genetic adaptation of high-altitude hypoxia in Tibetans ^{6, 7, 8}. Genetic variation in the regulatory region of the EPAS1 gene may influence gene expression and contribute to changes in biological functions ⁹. EPAS1 is expressed in organs that are involved in oxygen transport and metabolism, such the lung, placenta and vascular endothelium ¹⁰, and is associated with many biological processes and diseases related to metabolism ¹¹, angiogenesis ^{12, 13}, inflammation ^{14, 15} and cancer ^{16, 17, 18}.

The EPAS1 gene maps to human chromosome 2p21 and is about 120 kb in size with a coding region consisting of 15 exons ¹⁹. Four HIF2a SNPs (rs56721780, rs6756667, rs7589621 and rs1868092) have been significantly associated with different levels of high-altitude hypoxia among native Tibetans ²⁰. The rs56721780 SNP in the HIF2a promoter region has also been significantly associated with high-altitude adaptation of Tibetans ⁹ while the rs6756667 SNP from intron two has been significantly associated with susceptibility to acute mountain sickness in individuals unaccustomed to high altitude environments ²¹. The rs1868092 SNP near the HIF2a 3’UTR has been associated with high altitude polycythemia in male Han Chinese at the Qinghai-Tibetan plateau ²².

Single nucleotide changes that affect gene expression by impacting gene regulatory sequences such as promoters, enhances, and silencers are known as regulatory SNPs (rSNPs) ^{23, 24, 25, 26}. A rSNPs within a transcriptional factor binding site (TFBS) can change a transcriptional factor’s (TF) ability to bind its TFBS ^{27, 28, 29, 30} in which case the TF would be unable to effectively regulate its target gene ^{31, 32, 33, 34, 35}. This concept is examined for the abovefour HIF2a rSNPs and their allelic association with TFBS, where computation analyses ^{36, 37, 38, 39} is used to identify TFBS alterations created by the HIF2a rSNPs. In this report, the rSNP associations with changes in potential TFBS are discussed with their possible relationship to disease or sickness in humans.

Methods

The JASPAR CORE database ^{40, 41} and ConSite ⁴² were used to identify the potential STAT4 TFBS in this study. JASPAR is a database of transcription factor DNA-binding preferences used for scanning genomic sequences where ConSite is a web-based tool for finding cis-regulatory elements in genomic sequences. The TFBS and rSNP location within the binding sites have previously been discussed ⁴³. The Vector NTI Advance 11.5 computer program (Invitrogen, Life Technologies) was used to locate theTFBS in theEPAS1 gene (NCBI Ref Seq NM_001430) from 1.6 kb upstream of the transcriptional start site to 539 bps past the 3’UTR which represents a total of 91 kb. The JASPAR CORE database was also used to calculate each nucleotide occurrence (%) within the TFBS, where upper case lettering indicate that the nucleotide occurs 90% or greater and lower case less than 90%. The occurrence of each SNP allele in the TFBS is also computed from the database (Table 3).

Genome-wide association studies (GWAS) over the last decade have identified nearly 6,500 disease or trait-predisposing SNPs where only 7% of these are located in protein-coding regions of the genome ^{44, 45} and the remaining 93% are located within non-coding areas ^{46, 47} such as regulatory or intergenic regions. SNPs which occur in the putative regulatory region of a gene where a single base change in the DNA sequence of a potential TFBS may affect the process of gene expression are drawing more attention ^{23, 25, 48}. A SNP in a TFBS can have multiple consequences. Often the SNP does not change the TFBS interaction nor does it alter gene expression since a transcriptional factor (TF) will usually recognize a number of different binding sites in the gene. In some cases the SNP may increase or decrease the TF binding which results in allele-specific gene expression. In rare cases, a SNP may eliminate the natural binding site or generate a new binding site. In which cases the gene is no longer regulated by the original TF. Therefore, functional rSNPs in TFBS may result in differences in gene expression, phenotypes and susceptibility to environmental exposure ⁴⁸. Examples of rSNPs associated with disease susceptibility are numerous and several reviews have been published ^{48, 49, 50, 51}.

The rs56721780 rSNP EPAS1-G allele ^{G (+ strand) or C (} located in the unique RELA, RUNX1, TFAP2A, B & C TFBS have a 100%, 92% and 94-100% occurrence, respectively in humans (Table 3). Since these binding sites (BS) occur only once in the gene except for the RUNX1 TFBS which occurs twice, the rSNP G allele should have a tremendous impact on gene regulation by these TFs (Table 3). The minor rs56721780 rSNP EPAS1-C allele ^{C (+ strand) or G (} located in the unique FOXP3 and HOXA5 TFBS have a 65% and 31% occurrence, respectively in humans. Since these TFBS have a low occurrence in humans and occur more than once in the gene, the respective TF would not be expected to have much impact on the regulation of the EPAS1 gene (Figure 1, Table 3).

Figure 1. Double stranded DNA from the EPAS1 gene showing the potential TFBS for twenty different TFs which can bind their respective DNA sequence either above (+) or below (-) the duplex (cf. Table 3). The rs56721780 rSNP common EPAS1-G allele is found in each of these TFBS. As shown, this rSNP is located in the promoter region of the EPAS1 gene. Also included with the potential TFBS is their % sequence homology to the duplex.

The rs6756667 rSNP EPAS1-A allele [A (+ strand) or T (- strand) located in the unique CEBPa, FOS::JUN, HIC 1 & 2, HOXA2, NFIA and NRL TFBS has a 78, 94, 95, 98, 90, 100, 86 and 100% occurrence, respectively in humans (Figure 2, Table 3); however, only the CEBPa, NFIA and NRL TFBS occur only once in the gene. Consequently the corresponding three TFs should impact the regulation of the EPAS1 gene. The minor rs6756667 rSNP EPAS1-G allele ^{G (+ strand) or T (} located in the unique ATF7, GMEB2, JDP2, and MGA TFBS have a 99, 100, 99, 100% occurrence, respectively in humans and all only occur once in the EPAS1 gene. Consequently, the TFs for these TFBS could have some impact on the regulation of the EPAS1 gene (Table 2 & Table 3).

Figure 2. Double stranded DNA from the EPAS1 gene showing the potential TFBS for forty eight different TFs which can bind their respective DNA sequence either above (+) or below (-) the duplex (cf. Table 3). The rs7589621 rSNP common EPAS1-G allele is found in each of these TFBS. As shown, this rSNP is located in intron two of the EPAS1 gene. Also included with the potential TFBS is their % sequence homology to the duplex.

The rs7589621 rSNP EPAS1-G allele ^{G (+ strand) or C (} located in the unique NR3C1 & 2 TFBSs have a 100% occurrence in humans and are found only once in the EPAS1 gene (Table 3). Consequently, the corresponding glucocorticoid and mineralocorticoid nuclear receptors which bind their respective BS should have a major impact on the regulation of the EPAS1 gene (Table 2 & Table 3). The minor rs7589621 rSNP EPAS1-A allele ^{A (+ strand) or T (} located in the unique LIN54, POU3F1-4, TEAD1, 3, 4 TFBS has a 100, 95, 95, 90, 99, 92, 100 and 94% occurrence, respectively in humans (Table 3). However, only the POU3F1-3 TFBS occur once in the EPAS1 gene, consequently, their corresponding TFs should have an impact on the EPAS1 gene regulation. The remaining TFBS occur multiple times in the gene and would not be expected to have much impact on gene regulation (Figure 2, Table 2 & Table 3).

The rs1868092 rSNP EPAS1-A allele ^{A (+ strand) or T (} located in the unique NR2C2 and YY1 TFBS have a 81 and 94%% occurrence, respectively in humans and only occur once in the EPAS1 gene (Tables 2 & 3). The NR2C2 orphan nuclear receptor which can occur as a repressor of activator of transcription and the YY1 TF which exhibits both positive and negative control of transcript should have an impact on the regulation of the EPAS1 gene (Table 2). The minor rs1868092 rSNP EPAS1-G allele ^{G (+ strand) or C (} located in the unique HIC2, KLF5, NFIA, and TEAD1 TFBS have a 99, 100, 100 and 95% occurrence, respectively in humans and occur only once in the EPAS1 gene (Table 3). Since the corresponding TFs function as activators, enhancers and repressors, the occurrence of these TFBS should impact the regulation of the EPAS1 gene (Table 2 & Table 3).

Human diseases or conditions can be associated with rSNPs of the EPAS1 gene as illustrated above. What a change in the rSNP alleles can do, is to alter the DNA landscape around the SNP for potential TFs to attach and regulate a gene. As an example, the punitive TFBS associated with the rs56721780 common rSNP EPAS1-G allele from Table 3 as illustrated in Figure 1 as well as the rs7589621 common rSNP STAT4-G allele as illustrated in Figure 2. As can be seen in Table 3, these potential TFBS change when an individual carries the alternate allele. The importance of this has been illustrated in Figure 1 with the punitive TFAP2A, B & C TFBS where the common G allele has binding sites for these TFs and the minor C allele does not. The TFAP2A, B & C TFs act as activators and repressors and are involved in a large spectrum of biological functions such as proper eye, face, body wall, limb and neural tube development (cf. Table 2). Another example would be the punitive NR3C1 & 2 TFBS where the common rs7589621 rSNP G allele has created these binding sites for the glucocorticoid and mineralocorticoid nuclear receptors which regulate carbohydrate, protein and fat metabolism while the minor A allele has eliminated these TFBS (Figure 2, Table 2 & Table 3). Other examples can be found in Table 3.

SNPs that alter the TFBS are not only found in the promoter regions but in the introns, exons and the UTRs of a gene. The nucleus of the cell is where epigenetic alterations occur and TFs operate to convert chromosomes into single stranded DNA for mRNA transcription while it is the cytoplasm where mRNA is processed by separating exons and introns for protein translation. Consequently, it doesn’t matter where TFs bind the DNA in the nucleus because it is only there that TFs function. The SNPs outlined in this report should be considered as rSNPs since they change the DNA landscape for TF binding and have been associated with disease. In this report, examples have been described to illustrate that a change in rSNP alleles in the EPAS1 gene can provide different TFBS which in turn are also associated with human disease or alterations in human health such as adaptions to high altitude. The punitive changes in TFBS created by the four rSNPs could very well influence the significant cline in allele frequencies seen in Tibetans with increasing altitude ²⁰ or the haplotype association with high altitude polycythemia in male Han Chinese ²². As an example, the minor rs7589621 SNP EPAS1-A creates a potential TFBS for the FOXC TF which is an important regulator of cell viability and resistance to oxidative stress. Where oxidative stress is linked to oxygen, hypoxia, heart failure and the hypoxia-inducible factor transcriptional factors ⁵². The potential alterations in TFBS obtained by computational analyses need to be verified by future protein/DNA electrophoretic mobility gel shift assays and gene expression studies.