Molecular Evolution of RAMOSA1 (RA1) in Land Plants

Abstract

RAMOSA1 (RA1) is a Cys2-His2-type (C2H2) zinc finger transcription factor that controls plant meristem fate and identity and has played an important role in maize domestication. Despite its importance, the origin of RA1 is unknown, and the evolution in plants is only partially understood. In this paper, we present a well-resolved phylogeny based on 73 amino acid sequences from 48 embryophyte species. The recovered tree topology indicates that, during grass evolution, RA1 arose from two consecutive SUPERMAN duplications, resulting in three distinct grass sequence lineages: RA1-like A, RA1-like B, and RA1; however, most of these copies have unknown functions. Our findings indicate that RA1 and RA1-like play roles in the nucleus despite lacking a traditional nuclear localization signal. Here, we report that copies diversified their coding region and, with it, their protein structure, suggesting different patterns of DNA binding and protein-protein interaction. In addition, each of the retained copies diversified regulatory elements along their promoter regions, indicating differences in their upstream regulation. Taken together, the evidence indicates that the RA1 and RA1-like gene families in grasses underwent subfunctionalization and neofunctionalization enabled by gene duplication.

Keywords: collinearity; divergent motifs; embryophyte; nuclear localization; paralogs; phylogeny; promoter; secondary structure; zinc finger.

Figure 1

SUP evolution and the origin of RA1. Majority rule consensus tree (N = 22502 trees) of RA1 transcription factor in embryophytes generated by Bayesian inference using 73 peptide sequences (Figure S1 and Table S1). Black dots indicate Bayesian posterior probability (PP) = [0.9 to 1]. Green arrows indicate gene duplication events. Black asterisk points out proteins with known function (Table S2).

Figure 2

Evolution of RA1 in grasses. (a) Majority rule consensus tree (N = 22502 trees) of RA1 and RA1-like transcription factors in grass species generated by Bayesian inference using 35 amino acid sequences (Figures S2 and S3 and Table S1). Each clade is identified by a color. Black dots indicate Bayesian posterior probability (PP) = [0.9 to 1]. Black asterisk points out proteins with known function (Table S3). (b) Motif distribution patterns on RA1 and RA1-like amino acid sequences in grass species. Colored boxes represent motif occurrence (Figures S7 and S8). (c) Sequence representation logo of Motif 1 obtained from the multiple sequence alignment (Figure S9). Black arrowhead indicates the amino acid variant (A- > G) in Motif 1.

Figure 3

Sub-cellular localization analysis of RA1 and RA1-like proteins. Sub-cellular localization of the RA1-GFP, SvRA1-GFP, SvRA1-like A-GFP, and SvRA1-like B-GFP constructs in tobacco leaf epidermal cells. Green color is GFP protein signal. Blue color represents DAPI-stained nucleus.

Figure 4

Protein secondary structure and conformational plasticity. (a) Secondary structure prediction of SUP, ZmRA1-like A, ZmRA1-like B, and RA1 proteins obtained by PsiPred server. α-helix corresponding to the zinc finger domain is represented in orange. α-helix corresponding to Motif 2, putative EAR motif, is represented in violet. Other α-helices are represented in pink. β-strands are represented in yellow. (b) Relative disorder levels of the structures measured in a range of 0 to 1.0 (Figure S10 and Table S4). Levels above 0.5 (dashed line) are considered disordered regions. Abbreviations: ZF, zinc finger domain; EAR, EAR repressor motifs.

Figure 5

Zinc finger molecular dynamics simulation. (a) Time evolution of the secondary structure of SUP and RA1 proteins as determined by the DSSP algorithm using the simulated structures in the thermodynamics equilibrium interval (0.8 to 1.0 μs). At the side of the graph, the amino acid sequences are quoted with their residue number in the proteins. In the center of the graph, the relative position of each residue in reference to the first one of the α-helix N-terminal is indicated. (b) RMSF of the backbone atoms of each amino acid sequence in the equilibrium interval (0.8 to 1.0 μs), taking as reference for fitting the structures the one with smallest RMSD in the same period.

Figure 6

Noncoding cis-elements identified in predicted promoter regions of RA1 and RA1-like genes from grass species. Colored circles represent presence of conserved motifs in the promoter. Motif consensus sequences are showed in Table S5.