Genomic and functional evidence reveals molecular insights into the origin of echolocation in whales

Abstract

Echolocation allows toothed whales to adapt to underwater habitats where vision is ineffective. Because echolocation requires the ability to detect exceptional high-frequency sounds, fossils related to the auditory system can help to pinpoint the origin of echolocation in whales. However, because of conflicting interpretations of archaeocete fossils, when and how whales evolved the high-frequency hearing correlated with echolocation remain unclear. We address these questions at the molecular level by systematically investigating the convergent evolution of 7206 orthologs across 16 mammals and find that convergent genes between the last common ancestor of all whales (LCAW) and echolocating bats are not significantly enriched in functional categories related to hearing, and that convergence in hearing-related proteins between them is not stronger than that between nonecholocating mammalian lineages and echolocating bats. However, these results contrast with those of parallel analyses between the LCA of toothed whales (LCATW) and echolocating bats. Furthermore, we reconstruct the ancestral genes for the hearing protein prestin for the LCAW and LCATW; we show that the LCAW prestin exhibits the same function as that of nonecholocating mammals, but the LCATW prestin shows functional convergence with that of extant echolocating mammals. Mutagenesis shows that functional convergence of prestin is driven by convergent changes in the prestins S392A and L497M in the LCATW and echolocating bats. Our results provide genomic and functional evidence supporting the origin of high-frequency hearing in the LCAW, not the LCATW, and reveal molecular insights into the origin and evolutionary trajectories of echolocation in whales.

Fig. 1. Detection of convergence between echolocating bats and ancestral whales.

(A) Phylogeny of the mammalian species used to detect molecular convergence in this study. Bold lineages indicate toothed whales and echolocating bats with high-quality genomic data, and dashed lineages indicate the nonecholocating baleen whales and Old World fruit bats. The branches labeled I, II, III, IV, and V denote where convergent sites are counted. (B) Different comparisons of C/Ds of hearing-related genes based on the data set containing the inferred amino acids with ≥0.95 posterior probabilities. The numbers of convergent and divergent sites are given on the bars. C/D indicates the ratio of the number of convergent sites to the number of divergent sites. The P values are from two-tailed χ² tests.

Fig. 2. Evolutionary convergence is significantly higher for hearing genes than for nonhearing genes between LCATW and echolocating bats.

Frequency distributions of C/Ds from 104 genes unrelated to hearing for a total of 1000 random sets in the I and II (A), I and III (B), I and IV (C), and I and V (D) comparisons based on the data set containing the inferred amino acids with ≥0.95 posterior probabilities. The arrow in each panel indicates the ratio of the number of convergent sites to the number of divergent sites (C/D) identified from 104 hearing-related genes from different comparisons. C/D indicates the ratio of the number of convergent sites to the number of divergent sites.

Fig. 3. Functional results of prestin in modern whales.

(A) Phylogenetic relationships of whales with prestin sequences. Species names in red indicate echolocating toothed whales, and those in blue denote nonecholocating whales. Underlined names are representative species chosen for the functional examination of their prestin genes. (B) Representative fitting curves of nonlinear capacitance obtained from human embryonic kidney (HEK) 293 cells transfected by prestin. Different line types and colors indicate different species. (C) Comparison of three functional parameters, 1/α, V_1/2, and Q_max/C_lin, between echolocating and nonecholocating whales. All values are presented as means ± SE. *P < 0.05, **P < 0.01, ***P < 0.001. All P values are from Student’s t tests.

Fig. 4. Plot of 1/α versus the frequency of best hearing sensitivity showing significant relationships (R = 0.77, P = 0.015, F test).

Squares represent whale species, and circles represent other mammals. The squares in red indicate echolocating mammals, and those in blue denote nonecholocating mammals.

Fig. 5. Functional tests for resurrected ancestral prestin genes.

(A) Schematic phylogenetic tree showing the nodes where ancestral prestin genes are examined. Representative fitting curves of NLC derived from ancestral prestins are shown. (B) Comparison of 1/α values of prestin among the ancestral and living whales as well as their outgroup. (C) Convergent sites (S392A and L497M) between the LCATW and echolocating bats account for enhancement of the functional parameter 1/α. Both convergent sites have mutations based on the prestin backgrounds of the LCAW and the LCATW, respectively. Values of 1/α significantly increase in the LCAW double mutant and decrease in the LCATW double mutant when compared to their respective wild-type controls. All values are given as means ± SE. **P < 0.01 and ***P < 0.001. All P values are from Student’s t tests.