Research ArticleEVOLUTIONARY BIOLOGY

Molecular parallelism in fast-twitch muscle proteins in echolocating mammals

See allHide authors and affiliations

Science Advances  26 Sep 2018:
Vol. 4, no. 9, eaat9660
DOI: 10.1126/sciadv.aat9660
  • Fig. 1 Screen for parallel molecular evolution in mammals.

    (A) Illustration of parallel amino acid substitutions, defined as the same amino acid substitution from an identical ancestral state. Inferred ancestral amino acids are in gray font. Red branches have a parallel substitution from E (Glu) to K (Lys) at this alignment position. (B) Functional enrichments of 409 (416) proteins, where maximum likelihood (Bayesian) ancestral reconstruction infers parallel substitutions between the microbat (little brown bat) and the bottlenose dolphin, using GeneTrail2 (https://genetrail2.bioinf.uni-sb.de/). Ontology terms relating to muscle fiber components are in red font. Only the top 10 enrichments are shown for cellular component (65 significant enrichments). Muscle-related and other non–muscle-related enrichment terms for GO biological process and molecular function are shown in table S2. (C) Histograms showing the number of radical parallel amino acid substitutions in conserved positions of fast-twitch versus slow-twitch Ca2+ signaling and sarcomeric proteins for all 1296 independent pairs of branches. Ancestral sequence reconstruction with maximum likelihood (left) and a Bayesian approach (right) consistently shows that microbat and dolphin have seven parallel substitutions exclusively in fast-twitch fiber proteins (inset) and that no other branch pair has an excess of seven or more substitutions in the fast-twitch fiber set.

  • Fig. 2 Parallel molecular evolution in fast-twitch muscle fiber proteins in echolocating mammals.

    In addition to species included in the genome-wide screen (marked by asterisk), the alignments contain 46 additional mammals, which shows that parallel amino acid substitutions also occur in other echolocating mammals that emit calls at a high repetition rate (big brown bat, other Myotis species, killer whale, and baiji) but not exclusively in them. Sequences of additional mammals were extracted from a recent multiple genome alignment (46). Gray box, missing sequence.

  • Fig. 3 Simplified diagram showing the skeletal muscle contraction-relaxation cycle and 3D structures of proteins with parallel substitutions.

    Ca2+ bound to Casq1 is released from the SR through ryanodine receptors into the sarcoplasm to initiate the cross-bridge cycle during which myosin complexes generate mechanical work. For relaxation, Ca2+ is actively pumped back into the SR by Atp2a1, where majority of Ca2+ ions are bound again to polymerized Casq1. The 3D protein structures of Casq1 (Protein Data Bank: 3TRP; residues that bind Ca2+ with high affinity are in light blue), Atp2a1 (4NAB), and Myh2 (2MYS; Q863E is in the myosin light chain binding region but not part of this structure) are shown. Parallel substitutions are indicated in red. Residues involved in Casq1 dimerization and Ca2+ binding (21) and the Atp2a1 residue K400 that binds the inhibitor phospholamban (47) are labeled. The Myh2 relay domain that determines the actin sliding velocity (39) by communicating conformational changes between the converter domain, nucleotide-binding site, and actin-binding site (38, 48) is indicated. K49Q is not part of any available Myl1 structure.

  • Fig. 4 Comparison of gene expression of the four fast-twitch muscle fiber proteins relative to their slow-twitch paralogs.

    Expression level of calsequestrin (A), Ca2+ ATPase (B), myosin heavy chain (C), and myosin light chain (D) genes comparing fast-twitch (red font) and slow-twitch (black font) muscle fiber components in three biological replicates of P. parnellii anterior cricothyroid muscle (yellow) and breast muscle (blue). Horizontal line is the median. Genes with parallel substitutions are marked by an asterisk. Expression ratios and P values of a two-sided t test are shown at the bottom and in table S6. n.s., not significant.

  • Fig. 5 Microbat and dolphin Casq1 functionally converged in the ability to form Ca2+-sequestering polymers at lower Ca2+ concentrations.

    (A) Ca2+-dependent turbidity shows that microbat and dolphin Casq1 oligomerizes at lower Ca2+ concentrations. Error bars represent standard deviations of triplicate measurements. (B) Multiangle light scattering experiments show that microbat and dolphin Casq1 dimerizes at lower Ca2+ concentrations compared to mouse Casq1. While Casq1 from all three species is monomeric at a Ca2+ concentration of 0 mM, both microbat and dolphin Casq1 dimerizes at a concentration of 1 mM, in contrast to mouse Casq1 that remains monomeric.

  • Fig. 6 Both microbat/dolphin Myh2 and the fastest Drosophila myosin exhibit a negatively charged residue at position 512.

    The amino acid sequence of exon 15 of the mammalian Myh2 and the homologous alternative exons 9a/9b/9c of the Drosophila Myh are shown. In contrast to mammals, Drosophila has only a single Myh gene, but different muscles produce functionally different isoforms through alternative splicing. The Drosophila indirect flight muscle expresses a Myh isoform with an extremely fast cross-bridge detachment rate that exclusively includes exon 9a (blue font). In contrast to exons 9b (A, Ala) and 9c (T, Thr), exon 9a exhibits a negatively charged residue (D, Asp, red arrow) that resembles the parallel substitution T512E (Thr to Glu, black arrow) in microbat/dolphin Myh2 (blue font). A negatively charged residue (Asp or Glu) at this position likely modulates the actin sliding velocity by dynamically forming a salt bridge with the positively charged Arg759 in the converter domain during the post-power stroke conformation that precedes cross-bridge detachment (39). Note that mammalian Myh4 myosin, which confers the highest muscle fiber shortening velocity, also exhibits a negatively charged residue (Glu) at position 512 (fig. S4).

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/4/9/eaat9660/DC1

    Fig. S1. Computational approach to identify parallel amino acid substitutions between pairs of independent lineages.

    Fig. S2. Phylogeny of the 30 placental mammals.

    Fig. S3. Radical parallel substitutions in conserved positions between microbat and dolphin in hearing-related genes.

    Fig. S4. Convergence of microbat and dolphin Myh2 toward Myh4.

    Fig. S5. Myh4 is deleted in the cetacean lineage.

    Fig. S6. Unassembled genomic and RNA reads confirm the absence of Myh4 in the cetacean lineage.

    Fig. S7. The amount of aligning intronic sequence distinguishes Myh orthologs from paralogs.

    Table S1. Classification of all parallel amino acid substitutions detected by reconstructing ancestral sequences with a maximum likelihood and a Bayesian approach.

    Table S2. Functional enrichments of proteins with parallel substitutions between the microbat (little brown bat) and the bottlenose dolphin.

    Table S3. Pairs of independent branches and their number of parallel substitutions in fast-twitch versus slow-twitch fiber proteins.

    Table S4. GC-biased gene conversion alone can potentially explain only one of the seven parallel substitutions in the fast-twitch muscle proteins.

    Table S5. Top 30 genes with a significantly higher expression level in the anterior cricothyroid muscle compared to breast muscle of P. parnellii.

    Table S6. Expression of calsequestrin, Ca2+ ATPase, myosin heavy chain, and myosin light chain genes in the anterior cricothyroid muscle and the breast muscle of the echolocating P. parnellii bat.

    References (4955)

  • Supplementary Materials

    This PDF file includes:

    • Fig. S1. Computational approach to identify parallel amino acid substitutions between pairs of independent lineages.
    • Fig. S2. Phylogeny of the 30 placental mammals.
    • Fig. S3. Radical parallel substitutions in conserved positions between microbat and dolphin in hearing-related genes.
    • Fig. S4. Convergence of microbat and dolphin Myh2 toward Myh4.
    • Fig. S5. Myh4 is deleted in the cetacean lineage.
    • Fig. S6. Unassembled genomic and RNA reads confirm the absence of Myh4 in the cetacean lineage.
    • Fig. S7. The amount of aligning intronic sequence distinguishes Myh orthologs from paralogs.
    • Table S1. Classification of all parallel amino acid substitutions detected by reconstructing ancestral sequences with a maximum likelihood and a Bayesian approach.
    • Table S2. Functional enrichments of proteins with parallel substitutions between the microbat (little brown bat) and the bottlenose dolphin.
    • Table S3. Pairs of independent branches and their number of parallel substitutions in fast-twitch versus slow-twitch fiber proteins.
    • Table S4. GC-biased gene conversion alone can potentially explain only one of the seven parallel substitutions in the fast-twitch muscle proteins.
    • Table S5. Top 30 genes with a significantly higher expression level in the anterior cricothyroid muscle compared to breast muscle of P. parnellii.
    • Table S6. Expression of calsequestrin, Ca2+ ATPase, myosin heavy chain, and myosin light chain genes in the anterior cricothyroid muscle and the breast muscle of the echolocating P. parnellii bat.
    • References (4955)

    Download PDF

    Files in this Data Supplement:

Navigate This Article