Evolution of the Twist Subfamily Vertebrate Proteins: Discovery of a Signature Motif and Origin of the Twist1 Glycine-Rich Motifs in the Amino-Terminus Disordered Domain

Twist proteins belong to the basic helix-loop-helix (bHLH) family of multifunctional transcriptional factors. These factors are known to use domains other than the common bHLH in protein-protein interactions. There has been much work characterizing the bHLH domain and the C-terminus in protein-protein interactions but despite a few attempts more focus is needed at the N-terminus. Since the region of highest diversity in Twist proteins is the N-terminus, we analyzed the conservation of this region in different vertebrate Twist proteins and study the sequence differences between Twist1 and Twist2 with emphasis on the glycine-rich regions found in Twist1. We found a highly conserved sequence motif in all Twist1 (SSSPVSPADDSLSNSEEE) and Twist2 (SSSPVSPVDSLGTSEEE) mammalian species with unknown function. Through sequence comparison we demonstrate that the Twist protein family ancestor was “Twist2-like” and the two glycine-rich regions found in Twist1 sequences were acquired late in evolution, apparently not at the same time. The second glycine-rich region started developing first in the fish vertebrate group, while the first glycine region arose afterwards within the reptiles. Disordered domain and secondary structure predictions showed that the amino acid sequence and disorder feature found at the N-terminus is highly evolutionary conserved and could be a functional site that interacts with other proteins. Detailed examination of the glycine-rich regions in the N-terminus of Twist1 demonstrate that the first region is completely aliphatic while the second region contains some polar residues that could be subject to post-translational modification. Phylogenetic and sequence space analysis showed that the Twist1 subfamily is the result of a gene duplication during Twist2 vertebrate fish evolution, and has undergone more evolutionary drift than Twist2. We identified a new signature motif that is characteristic of each Twist paralog and identified important residues within this motif that can be used to distinguish between these two paralogs, which will help reduce Twist1 and Twist2 sequence annotation errors in public databases.