High-Altitude Andean H194R HIF2A Allele Is a Hypomorphic Allele

Abstract

For over 10,000 years, Andeans have resided at high altitude where the partial pressure of oxygen challenges human survival. Recent studies have provided evidence for positive selection acting in Andeans on the HIF2A (also known as EPAS1) locus, which encodes for a central transcription factor of the hypoxia-inducible factor pathway. However, the precise mechanism by which this allele might lead to altitude-adaptive phenotypes, if any, is unknown. By analyzing whole genome sequencing data from 46 high-coverage Peruvian Andean genomes, we confirm evidence for positive selection acting on HIF2A and a unique pattern of variation surrounding the Andean-specific single nucleotide variant (SNV), rs570553380, which encodes for an H194R amino acid substitution in HIF-2α. Genotyping the Andean-associated SNV rs570553380 in a group of 299 Peruvian Andeans from Cerro de Pasco, Peru (4,338 m), reveals a positive association with increased fraction of exhaled nitric oxide, a marker of nitric oxide biosynthesis. In vitro assays show that the H194R mutation impairs binding of HIF-2α to its heterodimeric partner, aryl hydrocarbon receptor nuclear translocator. A knockin mouse model bearing the H194R mutation in the Hif2a gene displays decreased levels of hypoxia-induced pulmonary Endothelin-1 transcripts and protection against hypoxia-induced pulmonary hypertension. We conclude the Andean H194R HIF2A allele is a hypomorphic (partial loss of function) allele.

Keywords: EPAS1; HIF2A; Andean evolution; high-altitude adaptation; hypoxia; hypoxia-inducible factor.

Fig. 1.

(A) Global allele frequency of the H194R missense variant depicted for Peruvian Andeans, Mexican Maya, Argentine Collas, and 1000 Genomes Project populations. The ancestral allele (A) and the derived allele (G) are shown. This image was downloaded from the Geography of Genetic Variants Browser (Marcus and Novembre 2017). (B) Violin plots of the natural log of the fractional exhaled NO (FeNO) by H194R genotype. Individual data points for each genotype are shown as filled circles. Mean natural log FeNO values are shown as a single solid central circle for each genotype with the mean values presented for each. H194R genotype was significantly associated with FeNO using an additive model (β = 0.20 [95% CI 0.02–0.38], P = 0.02, R² = 0.13) and a dominant model (β = 0.26 [95% CI 0.04–0.48], P = 0.02, R² = 0.14). Models controlled for sex and population structure.

Fig. 2.

H194R HIF-2α displays impaired heterodimerization with ARNT and decreased transcriptional activity. (A) Top, diagram of ARNT and HIF-2α. Primary site of regulatory prolyl hydroxylation in HIF-2α, Pro-531, is shown. Bottom, comparison of HIF-2α amino acids 184–203 (human numbering) across species. Shading indicates conserved residues. His-194 is indicated by an asterisk and dark shading. (B) Structure of the HIF-2α:ARNT heterodimer (HLH and PAS domains) bound to PT2385 (Wu et al. 2019). HIF-2α residue His-194, near the interface, is highlighted as indicated. PT2385, which binds to a pocket in HIF-2α, is indicated. (C) Coimmunoprecipitation studies of in vitro translated WT or H194R hemagglutinin (HA)-tagged HIF-2α with in vitro translated Flag-ARNT. Positions of molecular weight markers are shown to the left. (D) Quantitation of coimmunoprecipitation experiments shown in (C). n = 3 per group; ***P < 0.001. (E) Anti-GST western blot of GST and GST-ARNT (1–585). Positions of these proteins are shown to the right. Positions of molecular weight markers are shown to the left. (F) GST pull-down assays of WT or H194R HA-HIF-2α (1–356) from transfected HEK293FT cells. Positions of molecular weight markers are shown to the left. (G) Quantitation of GST pull-down experiments shown in (F). n = 3 per group; ***P < 0.001. (H and I) HRE reporter gene assays for WT or H194R HA-HIF-2α. Plasmid doses are indicated at the bottom. In controls, total DNA dose was normalized with pcDNA3. Luciferase activity was normalized to a Renilla luciferase internal transfection control. n = 6 (H) or 3 (I) per group; *P < 0.05, ***P < 0.001. Anti-HA or anti-Flag western blots for the constructs comparing expression are shown in insets. In (D) and (G)–(I), means ± SEM are shown. Data analyzed by unpaired two-tailed t-test in (D) and (G), and by one-way ANOVA and Tukey's post hoc test in (H) and (I).

Fig. 3.

Hif2a ^H194R/H194R mice display protection against hypoxia-induced pulmonary hypertension. (A and B) mRVP and (C and D) Fulton index were measured in mice with the indicated Hif2a genotypes exposed to either (A and C) hypoxia (3 weeks of 10% O₂) or (B and D) maintained under normoxia. Mice were 3–4 months of age. *P < 0.05, ***P < 0.001, ns, not significant. Data analyzed by unpaired two-tailed t-test. Numbers of mice in each group were as follows: (A and C) +/+ = 10 (five males and five females) and H194R/H194R = 9 (four males and five females); (B and D) +/+ = 8 (four males and four females) and H194R/H194R = 10 (five males and five females).

Fig. 4.

Hif2a ^H194R/H194R mice do not display changes in red cell indices. (A and B) Hct, (C and D) Hb concentration, and (E and F) red cell count were measured in mice with the indicated Hif2a genotypes exposed to either (A, C, and E) hypoxia (3 weeks of 10% O₂) or (B, D, and F) maintained under normoxia. Mice were 2–3 months of age. ns, not significant. Data analyzed by two-tailed t-test. Numbers of mice in each group were as follows: (A, C, and E) +/+ = 8 (three males and five females) and H194R/H194R = 8 (three males and five females); (B, D, and F) +/+ = 9 (four males and five females) and H194R/H194R = 5 (two males and three females).

Fig. 5.

Hif2a ^H194R/H194R mice display decreased small pulmonary artery thickness following exposure to hypoxia. (A and B) Photomicrographs of immunoperoxidase-stained sections of lung using antibodies against smooth muscle actin from wild-type and Hif2a^H194R/H194R mice exposed to hypoxia (3 weeks of 10% O₂). Small pulmonary arteries are in the center of each photomicrograph. Bars indicate 50 µm. (C and D) The medial thickness of small pulmonary arteries (expressed as a percentage of the external diameter) was measured in lungs obtained from mice with the indicated Hif2a genotypes exposed to either (C) hypoxia (3 weeks of 10% O₂) or (D) maintained under normoxia. Mice were 3 months of age. *P < 0.05; ns, not significant. Data analyzed by two-tailed t-test. Numbers of mice in each group were as follows: (C) +/+ = 5 (two males and three females) and H194R/H194R = 5 (two males and three females); (D) +/+ = 5 (two males and three females) and H194R/H194R = 5 (two males and three females). Each data point represents a single animal.

Fig. 6.

(A–E) Hif2a^H194R/H194R mice display decreased levels of hypoxia-induced Edn1 mRNA transcripts in the lung. The levels of the indicated mRNAs (relative to 18S rRNA) in the lung were determined by real-time PCR. Mice were 2–3 months of age and were exposed to either normoxia or hypoxia (48 h of 8% O₂). *P < 0.05, **P < 0.01, ***P < 0.001. Data analyzed by one-way ANOVA with Tukey's post hoc test. Numbers of mice in each group were as follows: +/+ normoxia = 6 (four males and two females); H194R/H194R normoxia = 6 (four males and two females); +/+ hypoxia = 6 (four males and two females); H194R/H194R hypoxia = 6 (four males and two females).