RNA consensus structures for inferring green algal phylogeny: A three-taxon analysis for Golenkinia/Jenufa, Sphaeropleales and Volvocales (Chlorophyta, Chlorophyceae)

Background and Aims: "Order please!" was a recently published, very nice and apt title for a persistent problem in green algae phylogenetics. The green algal class Chlorophyceae comprises two clades (SV and OCC) made up of five orders (Sphaeropleales, Volvocales, Oedogoniales, Chaetopeltidales and Chaetophorales). A variety of data and methods have shown that one further group (Golenkinia+Jemfa) cannot be unambiguously placed among the five orders. In addition, concerning Sphaeropleaceae and/or Treubarinia the monophyly of Sphaeropleales and Volvocales remain unresolved. Methods: IUPAC (International Union of Pure and Applied Chemistry) encoded consensus sequences for different 18S rDNA data sets as well as corresponding 18S rDNA consensus structures were assembled for Sphaeropleales, Volvocales (with or without Treubarinia) and a Golenkinia+Jenufa assemblage to independently infer phylogenetic relationships in a three-taxon analysis. Using a known template structure, individual 18S rDNA secondary structures were predicted by homology modelling. Sequences and their individual secondary' structures, automatically encoded by a 12-letter alphabet (each nucleotide with its three structural states, paired left, paired right, or unpaired), were simultaneously aligned; consensus structures and IUPAC encoded consensus sequences were read out from the different sequence-structure alignments. Key Results: In contrast to previous studies using 18S rDNA data, results of this study corroborate chloroplast data and strongly support a sister group relationship between Golenkinia+Jemtfa and the Sphaeropleales. The Golenkinia+Jemfa assemblage shows 330 matches to the Sphaeropleales (sequence-structure consensus) but only 214 matches to the Volvocales (sequence-structure consensus). Phylogenetieally informative nucleotides are highlighted and visualized in their structural context taking into account structural domains (I-IV) and hypervariable regions (V2-V9). Rooting the three-taxon scenario remains difficult because the extremely long branches of Golenkinia and Jenufa are attracted to a chosen outgroup, reducing bootstrap support in any conceivable four-taxon tree obtained by parsimony or profile-neighbor-joining analysis. Conclusions: This algal case study of 18S rDNA consensus data (IUPAC encoded sequences and consensus structures) coupled with profile distances between groups of sequences, demonstrated that a phylogenetic problem can be reduced to a three-taxon analysis.