Relaxed functional constraints on triplicate α-globin gene in the bank vole suggest a different evolutionary history from other rodents

Abstract

Gene duplication plays an important role in the origin of evolutionary novelties, but the mechanisms responsible for the retention and functional divergence of the duplicated copy are not fully understood. The α-globin genes provide an example of a gene family with different numbers of gene duplicates among rodents. Whereas Rattus and Peromyscus each have three adult α-globin genes (HBA-T1, HBA-T2 and HBA-T3), Mus has only two copies. High rates of amino acid evolution in the independently derived HBA-T3 genes of Peromyscus and Rattus have been attributed to positive selection. Using RACE PCR, reverse transcription-PCR (RT-PCR) and RNA-seq, we show that another rodent, the bank vole Clethrionomys glareolus, possesses three transcriptionally active α-globin genes. The bank vole HBA-T3 gene is distinguished from each HBA-T1 and HBA-T2 by 20 amino acids and is transcribed 23- and 4-fold lower than HBA-T1 and HBA-T2, respectively. Polypeptides corresponding to all three genes are detected by electrophoresis, demonstrating that the translated products of HBA-T3 are present in adult erythrocytes. Patterns of codon substitution and the presence of low-frequency null alleles suggest a postduplication relaxation of purifying selection on bank vole HBA-T3.

Figure 1

Deduced amino acid sequences of bank vole (Clethrionomys) α-globins aligned with those from other rodents. Dots represent amino acids identical to those at corresponding positions in Cavia sequence. Residues distinguishing bank vole HBA-T3 from bank vole HBA-T1 and HBA-T2 are shaded in grey, with residues distinguishing HBA-T3 from only one of the two genes in dark grey. Arrows indicate exon–exon junctions. The position of single nonsense mutations in two bank vole HBA-T3 alleles that generate in-frame PTCs are marked by a dot. The transcript of the PTC1 allele is most likely subject to the nonsense-mediated mRNA decay, whereas the PTC2 allele appears capable of producing a truncated α-globin; the read-through sequences are provided for comparison.

Figure 2

RT–PCR and protein expression analysis of the bank vole HBA-T3 globin gene. (a) Genomic structure of the gene. Alleles containing PTCs in exons 2 and 3, referred to as, respectively, the B allele and C allele, are aligned with a non-PTC allele referred to as the A allele. (b) RT–PCR and PCR assays followed by gel electrophoresis were performed to assess the amplification of HBA-T3 cDNA and genomic DNA (gDNA) using paralogue-specific primers spanning the entire coding region. Data represent HBA-T3 cDNA amplification from mRNA of all genotypes except of the PTC1 homozygote. The HBA-T1 cDNA served as a RT–PCR control and was amplified in all individuals. (c) Urea cellulose acetate electrophoresis of globin polypeptides demonstrating the absence of a HBA-T3 translation product in the homozygotes for both PTC-containing alleles, whereas the heterozygotes and the individuals with non-PTC genotypes show a detectable (although weakly staining) band (arrows).

Figure 3

RNA-seq expression profiling of the bank vole α-globin genes. (a) The mean RPKM of each HBA paralogue with error bars representing s.e.m. (b) Fold difference in transcript abundance between two alleles for each HBA paralogue (that is, AI). In the first three columns, the error bar represents s.e.m. The rightmost column (hatched) shows the result for an individual heterozygous (Het) for the allele containing a premature stop codon in exon 2 (PTC1). (c) RNA-seq depth-of-coverage profiles of an allele with an uninterrupted open reading frame and of the PTC1-containing allele from the heterozygote.

Figure 4

Phylogenetic relationships among rodent α-globin genes. (a) ML tree inferred from coding sequences, with statistical support for tree bipartitions expressed as percentage bootstrap values from ML/NJ analyses and as the Bayesian posterior probabilities (values <50% not shown). (b) Variation in selection pressure across branches as inferred by the genetic algorithm branch analysis. Branch labels represent model-averaged probabilities of ω>1 with the inferred number of synonymous and nonsynonymous substitutions in parentheses (NS/NN).

Figure 5

Homology-based structural model of the bank vole α-globin chain. The locations of 19 amino acids distinguishing the product of HBA-T3 from the products of HBA-T1 and HBA-T2 paralogues are shown. Residues are coloured according to their ConSurf evolutionary conservation score.