The molecular basis of high-altitude adaptation in deer mice

Abstract

Elucidating genetic mechanisms of adaptation is a goal of central importance in evolutionary biology, yet few empirical studies have succeeded in documenting causal links between molecular variation and organismal fitness in natural populations. Here we report a population genetic analysis of a two-locus alpha-globin polymorphism that underlies physiological adaptation to high-altitude hypoxia in natural populations of deer mice, Peromyscus maniculatus. This system provides a rare opportunity to examine the molecular underpinnings of fitness-related variation in protein function that can be related to a well-defined selection pressure. We surveyed DNA sequence variation in the duplicated alpha-globin genes of P. maniculatus from high- and low-altitude localities (i) to identify the specific mutations that may be responsible for the divergent fine-tuning of hemoglobin function and (ii) to test whether the genes exhibit the expected signature of diversifying selection between populations that inhabit different elevational zones. Results demonstrate that functionally distinct protein alleles are maintained as a long-term balanced polymorphism and that adaptive modifications of hemoglobin function are produced by the independent or joint effects of five amino acid mutations that modulate oxygen-binding affinity.

Figure 1. Structural Alignment of Mammalian α-Globins, including Representative High- and Low-Altitude Protein Alleles (Denoted by H and L, Respectively) from Both Paralogs of P. maniculatus

Figure 2. Plot of Linkage Disequilibrium across the 5′ α-Globin Gene, Showing Pairwise Associations between Informative Nucleotide Polymorphisms in the Total Sample of 74 Chromosomes

In the diagram of α-globin gene structure, the three exons are depicted as boxes, the exonic regions that encode the six α helices (A, B, E, F, G, and H) are shown in black, and nonsynonymous polymorphisms are denoted by dotted lines. In the triangle matrix, p-values from Fisher's exact test are given for each pairwise comparison of informative polymorphisms (the Bonferroni-adjusted α level is 0.0000256). Amino acid polymorphisms at residue positions (CD15)50, (E6)57, (E9)60, (E9)64, (EF1)71, and (GH4)116 (mentioned in the text) are attributable to nonsynonymous mutations at nucleotide positions 280, 301, 310, 322, 342, and 665, respectively.

Figure 3. Altitudinal Patterns of Allele Frequency Variation at Five Amino Acid Replacement Polymorphisms in the 5′ α-Globin Gene that Span the E-Helix Domain of the Encoded Polypeptide

In the pie diagram for each polymorphic site, the frequency of the derived allele is shown in black. From west to east, the sampling localities are 1, Mt. Evans, Clear Creek County, Colorado (4,347 m); 2, Bonny Reservoir, Yuma County, Colorado (1,158 m); and 3, Fort Larned National Monument, Pawnee County, Kansas (620 m).

Figure 4. Homology-Based Structural Model of P. maniculatus Oxyhemoglobin, Showing the Location of Five Amino Acid Replacement Polymorphisms in the 5′ α-Globin Gene that Span the E-Helix Domain of the Encoded α-Chain Polypeptide

The heme group (ferroprotoporphyrin IX) of the α-chain is shown in red, and the bound O₂ molecule is shown in blue. Top right: The five amino acid variants that predominate in the high-altitude sample (50Pro/57Gly/60Ala/64Gly/71Ser) are shown in blue. Bottom right: The five amino acid variants that predominate in the low-altitude samples (50His/57Asp/60Gly/64Asp/71Gly) are shown in red.

Figure 5. Sliding Window Plot Showing Variation in Levels of Silent Site Diversity within and between Functionally Defined Haplotype Classes of the 5′ α-Globin Gene

In comparisons between the high- and low-affinity protein haplotypes, a pronounced peak of silent site divergence is directly centered on the region of exon 2 that harbors the five class-defining replacement polymorphisms (indicated by red arrows). The region of exon 2 that encodes the E-helix domain (nucleotides 285 through 335) is shown in black. The protein haplotype with high oxygen-binding affinity (50Pro/57Gly/60Ala/64Gly/71Ser) predominates in the high-altitude sample, whereas the protein haplotype with low oxygen-binding affinity (50His/57Asp/60Gly/64Asp/71Gly) predominates at low altitude.