Morphological Stasis and Proteome Innovation in Cephalochordates

Abstract

Lancelets, extant representatives of basal chordates, are prototypic examples of evolutionary stasis; they preserved a morphology and body-plan most similar to the fossil chordates from the early Cambrian. Such a low level of morphological evolution is in harmony with a low rate of amino acid substitution; cephalochordate proteins were shown to evolve slower than those of the slowest evolving vertebrate, the elephant shark. Surprisingly, a study comparing the predicted proteomes of Chinese amphioxus, Branchiostoma belcheri and the Florida amphioxus, Branchiostoma floridae has led to the conclusion that the rate of creation of novel domain combinations is orders of magnitude greater in lancelets than in any other Metazoa, a finding that contradicts the notion that high rates of protein innovation are usually associated with major evolutionary innovations. Our earlier studies on a representative sample of proteins have provided evidence suggesting that the differences in the domain architectures of predicted proteins of these two lancelet species reflect annotation errors, rather than true innovations. In the present work, we have extended these studies to include a larger sample of genes and two additional lancelet species, Asymmetron lucayanum and Branchiostoma lanceolatum. These analyses have confirmed that the domain architecture differences of orthologous proteins of the four lancelet species are because of errors of gene prediction, the error rate in the given species being inversely related to the quality of the transcriptome dataset that was used to aid gene prediction.

Keywords: Asymmetron; Branchiostoma; amphioxus; domain architecture; gene prediction; genome annotation; lancelet; proteome; stasis; transcriptome.

Figure 1

Ryk receptor tyrosine kinase genes are present in lancelet species. (A) Although the Branchiostoma belcheri genome encodes a full-length Ryk receptor tyrosine kinase (247370_PRF0), no ortholog is found in the predicted proteome of Branchiostoma floridae. The transcriptome of Asymmetron lucayanum contains transcripts encoding full-length Ryk receptor tyrosine kinase (GETC01129892.1), but in the case of B. lanceolatum, only fragments of the RYK gene are represented in its transcriptome; (B) Analysis of the B. floridae genome and transcriptome with FixPred provided evidence for a RYK gene that encodes a protein (RYK_BRAFL) with the same domain architecture as those of B. belcheri (RYK_BRABE) and A. lucayanum (RYK_ASYLU). The figure illustrates the Pfam domain architectures of the proteins and the position of protein fragments of the Ryk receptor tyrosine kinases. Color code for Pfam-A domains: Wnt-inhinitory factor (WIF)—green; Pkinase_Tyr—red.

Figure 2

WAP, Follistatin, Immunoglobulin, Kunitz and NTR Domain-Containing (WFIKKN) genes encoding proteins with domain architectures typical of vertebrate WFIKKN proteins are present in the genomes of B. belcheri, B. floridae and A. lucayanum. (A) Proteins 328450_PRF0 and 328460_PRF0 of B. belcheri correspond to different regions of the WFIKKN protein of B. belcheri and the WFIKKN protein of B. floridae (XP_002607185.1). The transcriptome of A. lucayanum contains three transcripts (GESY01012772.1, GESY01032162.1, and GESY01032163.1) of the WFIKKN gene and the genome assembly of A. lucayanum contains a single scaffold encoding the 5′ part half of the WFIKKN gene (LZCU01164105.1); (B) Analysis of the transcriptome and genome of A. lucayanum with FixPred has provided evidence that its WFIKKN gene encodes a protein (WFIKKN_ASYLU) with the same domain architecture as those of B. belcheri (WFIKKN_BRABE) and B. floridae (WFIKKN_BRAFL). The figure illustrates the Pfam domain organizations of the proteins and the position of protein fragments of WFIKKN proteins. Color code for Pfam-A domains: WAP—green; Kazal—red; Immunoglobulin—blue; Kunitz—yellow; netrin (NTR)—purple.

Figure 3

Genes encoding proteins with domain architectures typical of vertebrate calcium-binding mitochondrial carrier proteins are present in the genomes of B. belcheri, B. floridae, B. lanceolatum, and A. lucayanum. (A) The B. floridae protein XP_002602990.1 is an ortholog of the calcium-binding mitochondrial carrier protein of B. belcheri (330210_PRF0), but it contains a C-terminal extension with a Peptidase_M2 domain. The transcriptome of A. lucayanum contains full-length transcripts of the gene (GESY01078193.1) encoding a protein with a domain architecture characteristic of calcium-binding mitochondrial carrier proteins. The transcriptome of B. lanceolatum contains several transcripts of the calcium-binding mitochondrial carrier protein gene and one of them, matching the C-terminal part of the protein (JT887996.1), supports the conclusion that it does not have a C-terminal Peptidase_M2 domain; (B) The corrected sequence (CMC1_BRAFL) of the calcium-binding mitochondrial carrier protein of B. floridae [21] has the same domain architecture as its orthologs of B. belcheri (CMC1_BRABE) and A. lucayanum (CMC1_ASYLU). The figure illustrates the Pfam domain architectures of the proteins and the position of protein fragments of calcium-binding mitochondrial carrier proteins. Color code for Pfam-A domains: EF-hand—green; Mito_carr—red; Peptidase_M2—blue.

Figure 4

Myb Like, SWIRM And MPN Domains 1 (MYSM1) genes encoding proteins with domain architectures typical of histone H2A deubiquitinases are present in the genomes of B. belcheri, B. floridae, and A. lucayanum. (A) The architectures of the predicted histone H2A deubiquitinases of B. belcheri and B. floridae deviate from the conserved domain architecture of these proteins in that one of the fragments of the B. belcheri protein (256530_PRF0) lacks the JAB domain, another fragment of the B. belcheri ortholog (256520_PRF0) lacks the Myb_DNA-binding and SWIRM domains, whereas the B. floridae protein (B6MUN4.1) lacks the Myb_DNA-binding domain. The transcriptome of A. lucayanum contains full-length transcripts of the MYSM1 gene, but the transcriptome of B. lanceolatum had only one transcript fragment derived from its MYSM1 gene; (B) Re-annotation of the MYSM1 genes of B. belcheri and B. floridae with the FixPred protocol has shown that these deviations are due errors in gene prediction, permitting the correction of these sequences (MYSM1_BRABE, MYSM1_BRAFL). The figure illustrates the Pfam domain architectures of the proteins and position of protein fragments of MYSM1 proteins. Color code for Pfam-A domains: Myb_DNA-binding—green; SWIRM—red; JAB—blue.

Figure 5

Chordin (CHRD) genes encoding proteins with domain architectures typical of vertebrate chordins are present in the genomes of B. belcheri, B. floridae, and A. lucayanum. (A) The characteristic domain architecture of vertebrate chordins is conserved in B. floridae (Q0Q581), but the structure of the B. belcheri ortholog (Bb_056190F) is markedly different in that it lacks some of the domains and has a long C-terminal extension containing a polycystic kidney disease (PKD)_channel domain. The transcriptome of A. lucayanum contains full-length transcripts of the CHRD gene (GETC01124733.1). The transcriptome of B. lanceolatum had several transcripts of the CHRD gene and one of them matching the C-terminal part of the protein (JT869941.1) supports the conclusion that the protein does not have a PKD_channel domain; (B) Re-annotation of the B. belcheri CHRD gene with the MisPred/FixPred protocol has shown that these deviations are due to errors in gene prediction, permitting the correction of this sequence (CHRD_BRABE). The figure illustrates the Pfam domain architectures of the proteins and position of protein fragments of chordins. Color code for Pfam-A domains: von Willebrand factor type C domain (VWC)—green; chordin (CHRD)—red; PKD_channel—blue.