Search

Bats account for ~20% of all extant mammal species and are considered exceptional given their extraordinary adaptations, including biosonar, true flight, extreme longevity, and unparalleled immune systems. To understand these adaptations, we generated reference-quality genomes of six species representing the key divergent lineages. We assembled these genomes with a novel pipeline incorporating state-of-the-art long-read and long-range sequencing and assembly techniques. The genomes were annotated using a maximal evidence approach, de novo predictions, protein/mRNA alignments, Iso-seq long read and RNA-seq short read transcripts, and gene projections from our new TOGA pipeline, retrieving virtually all (>99%) mammalian BUSCO genes. Phylogenetic analyses of 12,931 protein coding-genes and 10,857 conserved non-coding elements identified across 48 mammalian genomes helped to resolve bats’ closest extant relatives within Laurasiatheria, supporting a basal position for bats within Scrotifera. Genome-wide screens along the bat ancestral branch revealed (a) selection on hearing-involved genes (e.g LRP2, SERPINB6, TJP2), which suggest that laryngeal echolocation is a shared ancestral trait of bats; (b) selection (e.g INAVA, CXCL13, NPSR1) and loss of immunity related proteins (e.g. LRRC70, IL36G), including pro-inflammatory NF-kB signalling; and (c) expansion of the APOBEC family, associated with restricting viral infection, transposon activity and interferon signalling. We also identified unique integrated viruses, indicating that bats have a history of tolerating viral pathogens, lethal to other mammal species. Non-coding RNA analyses identified variant and novel microRNAs, revealing regulatory relationships that may contribute to phenotypic diversity in bats. Together, our reference-quality genomes, high-quality annotations, genome-wide screens and in-vitro tests revealed previously unknown genomic adaptations in bats that may explain their extraordinary traits.

To fulfill their physiological functions, bile acids are conjugated with amino acids. In humans, conjugation is catalyzed by bile acid coenzyme A: amino acid N-acyltransferase (BAAT), an enzyme with a highly conserved catalytic triad in its active site. Interestingly, the conjugated amino acids are highly variable among mammals, with some species conjugating bile acids with both glycine and taurine, whereas others conjugate only taurine. The genetic origin of these bile acid conjugation differences is unknown. Here, we tested whether mutations in BAAT’s catalytic triad could explain bile acid conjugation differences. Our comparative analysis of 118 mammals first revealed that the ancestor of placental mammals and marsupials possessed two genes, BAAT and BAATP1, that arose by a tandem duplication. This duplication was followed by numerous gene losses, including BAATP1 in humans. Losses of either BAAT or BAATP1 largely happened in a reciprocal fashion, suggesting that a single conjugating enzyme is generally sufficient for mammals. In intact BAAT and BAATP1 genes, we observed multiple changes in the catalytic triad between Cys and Ser residues. Surprisingly, although mutagenesis experiments with the human enzyme have shown that replacing Cys for Ser greatly diminishes the glycine-conjugating ability, across mammals we found that this residue provides little power in predicting the experimentally measured amino acids that are conjugated with bile acids. This suggests that the mechanism of BAAT’s enzymatic function is incompletely understood, despite relying on a classic catalytic triad. More generally, our evolutionary analysis indicates that results of mutagenesis experiments may not easily be extrapolatable to other species.

Detecting associations between genomic changes and phenotypic differences is fundamental to understanding how phenotypes evolved. By systematically screening for parallel amino acid substitutions, we detected known as well as novel cases (Strc, Tecta, Cabp2) of parallelism between echolocating bats and toothed whales in proteins that could contribute to high frequency hearing adaptations. Interestingly, our screen also showed that echolocating mammals exhibit an unusually high number of parallel substitutions in fast-twitch muscle fiber proteins. Both bats and dolphins produce an extremely rapid call rate when homing in on their prey, which was shown in bats to be powered by specialized superfast muscles. We show that these genes with parallel substitutions (Casq1, Atp2a1, Myh2, Myl1) are expressed in the superfast sound-producing muscle of bats. Furthermore, we found that the calcium storage protein calsequestrin 1 of bats and dolphins functionally converged in its ability to form calcium-sequestering polymers at lower calcium concentrations, which may contribute to rapid calcium transients required for superfast muscle physiology. The proteins that our genomic screen detected could be involved in the convergent evolution of vocalization in echolocating mammals by potentially contributing to both rapid Ca2+transients and increased shortening velocities in superfast muscles.AbbreviationsSRsarcoplasmic reticulum