The adaptive significance of unproductive alternative splicing in primates

Abstract

Alternative gene splicing is pervasive in metazoa, particularly in humans, where the majority of genes generate splice variant transcripts. Characterizing the biological significance of alternative transcripts is methodologically difficult since it is impractical to assess thousands of splice variants as to whether they actually encode proteins, whether these proteins are functional, or whether transcripts have a function independent of protein synthesis. Consequently, to elucidate the functional significance of splice variants and to investigate mechanisms underlying the fidelity of mRNA splicing, we used an indirect approach based on analyzing the evolutionary conservation of splice variants among species. Using DNA polymerase β as an indicator locus, we cloned and characterized the types and frequencies of transcripts generated in primary cell lines of five primate species. Overall, we found that in addition to the canonical DNA polymerase β transcript, there were 25 alternative transcripts generated, most containing premature terminating codons. We used a statistical method borrowed from community ecology to show that there is significant diversity and little conservation in alternative splicing patterns among species, despite high sequence similarity in the underlying genomic (exonic) sequences. However, the frequency of alternative splicing at this locus correlates well with life history parameters such as the maximal longevity of each species, indicating that the alternative splicing of unproductive splice variants may have adaptive significance, even if the specific RNA transcripts themselves have no function. These results demonstrate the validity of the phylogenetic conservation approach in elucidating the biological significance of alternative splicing.

FIGURE 1.

Phylogenetic relationships among the five primate species in this study, with numbers indicating statistical comparisons of the proportion of splice variants at bifurcation points. All four comparisons resulted in significant differences, each pair of branches (statistics in text). For instance, node 2 indicates a significant difference in splice variant frequency between Gorilla and Pan + Homo.

FIGURE 2.

Estimated maximum number of POLB transcripts in the five primate species studied, based on eight different estimation methods implemented in EstimateS diversity software. The dashed line indicates the observed number of transcripts (27). Error bars represent the standard deviation for each predictor based on 100 sampling replicates.

FIGURE 3.

Correlation of life history parameters and splicing frequency (white squares) and diversity (gray diamonds). Lines and Pearson r statistics represent correlations based on splicing frequency, but qualitatively similar results are obtained from correlations based on SV diversity, as frequency and diversity are strongly correlated.