The Unicellular Ancestry of Groucho-Mediated Repression and the Origins of Metazoan Transcription Factors

Abstract

Groucho is a co-repressor that interacts with many transcription factors playing a crucial role in animal development. The evolutionary origins of Groucho are not clear. It is generally regarded as being a distinct animal-specific protein, although with similarities to the yeast Tup-like proteins. Here, it is shown that Groucho has true orthologs in unicellular relatives of animals. Based on their phylogenetic distribution, and an analysis of ligand-binding residues, these genes are unlikely to be orthologs of the fungal Tup-like genes. By identifying conserved candidate Groucho interaction motifs (GIMs) in nonmetazoan transcription factors, it is demonstrated that the details of molecular interactions between Groucho and transcription factors are likely to have been established prior to the origin of animals, but that the association of GIMs with many transcription factor types can be regarded as a metazoan innovation.

Keywords: animal evolution; transcription factor evolution; transcriptional repression.

Fig. 1.—

The N-terminal domains of Groucho/TLE (PDB: 4om2) and TUP1 (PDB: 3vp8) do not share statistically significant sequence similarity: (a) they adopt different quaternary structures—chains colored from blue at N-terminus to red at C-terminus; (b) both contain coiled-coil regions but with differing degrees of curvature and Groucho/TLE contains additional C-terminal helices.

Fig. 2.—

(a) Phylogenetic tree of aligned WD40 sequences from Groucho and Tup, with representative WDR5 proteins as an outgroup. The Tup and Groucho clades are boxed and labeled. Sequences within the Tup group typically include a Tup_N N-terminal motif, and those within the Groucho group a TLE_N motif. The Naegleria sequences are indicated with arrows. ‘Teretosporea’ is a clade of Icthyosporea and Corallochytrium, defined in Torruella et al. (2015). Black circles on nodes represent complete boostrap support, with numbers giving values for other nodes central to the discrimination of Tup and Groucho. Sequences that uniquely define the leaves are available in the Supplementary Material online. (b) The distribution of Groucho and Tup orthologs identified in this study with respect to the three major eukaryotic groups (eukaryotic tree adapted from He et al. 2014).

Fig. 3.—

Side view of the ligand-binding pocket residues of TLE1 (three structures, with bound EH1 (PDB: 2ce8), WRPW (PDB: 2ce9), and nonbound forms (PDB: 1gxr)), with equivalent TUP1 (PDB: 1erj) residues superimposed. The F and W of the TLE1 bound ligands are shown (Sprague et al. 2000; Jennings et al. 2006).

Fig. 4.—

Sequence conservation of the ligand-binding pocket of the Groucho and Tup proteins (outgroup WDR5 members are also shown). The region is extracted from the full multiple sequence alignment, columns that are >80% identical within a class are colored by amino acid type (Taylor 1997). The dichotomous F,E (in Groucho) and Y,L (in Tup) residues, likely to contribute to ligand recognition, are marked with arrows.

Fig. 5.—

Example associations from nonbilaterian metazoans of EH1 motifs with (a) Homeobox, (b) Forkhead, and (c) T-box domains. Sc = Sycon ciliatum, a sponge; ML = Mnemiopsis leidyi, a ctenophore; Ta = Trichoplax adhaerens. Sequence accessions correspond to the databases described in the Methods. Domain diagrams represent the Sycon sequences.