Early evolution of the T-box transcription factor family

Abstract

Developmental transcription factors are key players in animal multicellularity, being members of the T-box family that are among the most important. Until recently, T-box transcription factors were thought to be exclusively present in metazoans. Here, we report the presence of T-box genes in several nonmetazoan lineages, including ichthyosporeans, filastereans, and fungi. Our data confirm that Brachyury is the most ancient member of the T-box family and establish that the T-box family diversified at the onset of Metazoa. Moreover, we demonstrate functional conservation of a homolog of Brachyury of the protist Capsaspora owczarzaki in Xenopus laevis. By comparing the molecular phenotype of C. owczarzaki Brachyury with that of homologs of early branching metazoans, we define a clear difference between unicellular holozoan and metazoan Brachyury homologs, suggesting that the specificity of Brachyury emerged at the origin of Metazoa. Experimental determination of the binding preferences of the C. owczarzaki Brachyury results in a similar motif to that of metazoan Brachyury and other T-box classes. This finding suggests that functional specificity between different T-box classes is likely achieved by interaction with alternative cofactors, as opposed to differences in binding specificity.

Keywords: Holozoa; Porifera; origin multicellularity; premetazoan evolution; subfunctionalization.

Fig. 1.

Phylogenetic distribution of different T-box classes among opisthokonts. The first column indicates the minimum and maximum number of T-box genes found in each lineage. Consensus phylogenetic relationships are shown (, –24). See also Fig. S1 and Dataset S1.

Fig. 2.

C. owczarzaki Brachyury (CoBra) and Tbox-3 (CoTbx3) mRNAs rescue XBra_En injections in Xenopus assays. (A) All panels show MyoD expression in stage-30 embryos injected with 500 pg of the mRNA or the combinations of mRNAs indicated in the Upper Left. The phenotypes were classified in three categories based on the amount of trunk structures observed in the injected embryos as determined by muscle MyoD expression. Wild-type embryos showed complete trunk and full MyoD expression. Mild affected embryos showed partial reduction of the trunk with reduced MyoD expression domain. Severe affected embryos lacked almost all trunk tissue, and the expression of MyoD was hardly or not detected. Lower Right shows bar plots summarizing the different phenotypes observed in each case. (B) Quantitative RT-PCR experiments showing two muscle (MyoD and Muscle actin) and one notochord (Shh) genes in the different injected embryos. Controls are noninjected embryos. Bar plots represent relative expression, normalized with endogenous Histone 4 expression levels. Error bars represent SD from at least two different biological replicates.

Fig. 3.

Molecular phenotype of Xenopus animal caps or embryos injected with different Brachyury orthologs mRNAs. (A) Quantitative RT-PCR experiments showing several mesendoderm markers in animal caps injected with different mRNAs. Controls are noninjected embryos. Bar plots show relative expression, normalized with endogenous Histone 4. Error bars represent SD from at least two different biological replicates. (B) In situ hybridization to detect chordin and Sox17 expression in animal caps injected with the Brachyury mRNAs indicated above each panel. (C) Whole-mount in situ hybridization of Sox17, chordin, and Wnt11 genes in stage 11–12 Xenopus embryos injected with the Brachyury orthologs indicated above each panel. All embryos are shown in the same orientation. Dotted lines represent the closing blastopore; the black arrowhead indicates the dorsal side, and the white arrowhead highlights ectopic expression.

Fig. 4.

CoBra-binding motifs derived from PBM data (SI Methods). For comparison, different mouse T-box classes binding motifs also derived from PBM data [except mouse T, based on SELEX (Systematic Evolution of Ligands by Exponential Enrichment) (34)]. See also Dataset S2.