Fig. 2

Maximum likelihood (ML) phylogeny of the TALE superclass homeodomain in Archaeplastida supports ancient division between KNOX- and non-KNOX TALE groups. The ML trees were generated from the homeodomain alignment with 70 amino acid positions. The consensus tree out of 1000 bootstrap trees is shown. The three numbers at critical nodes show %bootstrap, %SH, and Bayesian posterior probability in support of clades. The tree contains two outgroup clades marked by black squares at nodes, and two Archaeplastida clades, one combining most KNOX sequences marked by the red square and the other combining all non-KNOX sequences marked by the blue square. Vertical bars on the right depict the distribution of outgroup in black, KNOX in red, and non-KNOX sequences in blue. Red dots by the sequence names indicate the presence of KN-A or KN-B domains, and blue dots indicate the presence of a PBC-homology domain. Truncated sequences not available for homology domain analysis are marked with open black boxes. Filled black boxes indicate the absence of a KN-A/B or PBC-homology domain. Proposed classification is indicated by the vertical lines. Dotted vertical lines indicate suggested class members placed outside the main clade for the class in the phylogeny. PBX-Red sequences are found in four separate clades, marked by purple shades on the blue section of the vertical bars. Two PBX-Red sequences marked by the purple square are exceptionally found in the KNOX-Red1 clade, having divergent amino acids at highly conserved positions at Trp19, His23, and Lys31 in their homeodomain, suggesting their false association with the KNOX-Red1. Colors of the sequence names indicate their phylogenetic group: Blue for Glaucophyta, purple for Rhodophyta, green for prasinophytes, light blue for the chlorophytes, orange for Streptophyta, and black for outgroups. The ruler shows genetic distance. Details of the sequences analyzed by this phylogeny are provided in Additional file 1: Table S2. *Gloeochaete_wittrockiana_014496 is considered as a sequence from a bannelid-type amoeba that contaminated the original culture (SAG46.84) for the MMETSP1089 transcriptome. **Association of KNOX-Red2 class sequences to Amorphea PBC sequences is attributed to a shared WFGN motif determining DNA-binding specificity of the homeodomain via convergent evolution