Regions where homology between sites was doubtful were manually removed from the align ments before phylogenetic analyses. A supermatrix was built by concatenating the alignments corresponding to the 24 APC C components, adaptor co activators http://www.selleckchem.com/products/Oligomycin-A.html and targets inferred to be present in LECA. Conserved functional domain search The identification of functional domains was carried out using the HMMER package and the HMM profiles of the Pfam database. HMM profiles having e values lower than 0. 1 were considered as significant. Phylogenetic analysis Phylogenetic trees were reconstructed for each single protein dataset, except for Apc9, Apc14, Apc15, Apc16 and securin because of their very restricted number of homologues.

Maximum Likelihood phylogenetic trees were computed using Treefinder with the LG model and a gamma correction to take into account the heterogeneity of evolutionary rates across sites, as proposed by the model selection tool implemented in Treefinder. The branch robustness of ML trees was esti mated using the non parametric bootstrap procedure implemented in Treefinder with the same parameters than for ML tree inference. Bayesian trees were computed using MrBAYES 3. 1. 2 with a mixed amino acid substitu tion model and a gamma correction. The Markov chain Monte Carlo search was run with 4 chains for 1, 000, 000 generations, with trees being sampled every 100 generations. Individual Bayesian trees are provided as Additional file 3, Data S1. TreeFinder and PhyloBayes 3. 2 were used to per form ML and Bayesian analyses on the concatenated dataset.

Treefinder was run with the same parameters than for the analysis of single datasets. Phylobayes was run using the CAT model and a gamma correction with 4 rate categories. For each dataset, two independent chains were run for at least 10000 cycles, saving one tree in ten. The first 300 trees were discarded as burn in, and the remaining trees from each chain in each dataset were used to test for convergence and compute the 50% majority rule consensus tree. Inference of the origin of APC C components and main targets The phylogenetic analysis of each individual component allowed distinguishing orthologues that originated from speciation events from paralogues that resulted from gene duplications. This point is important because infer ring the evolutionary origin of a protein requires the analysis of the evolutionary history of orthologues.

To determine the origin of each component and subunit, we used a parsimony criterion, making the assumption that horizontal gene transfers between eukaryotes are rare. Accordingly, the ancestral absence of a component in the ancestor of a taxonomic group was inferred if no orthologues Entinostat are found in any present day representative of the group. By contrast, the presence of orthologues in representatives of two lineages was inter preted as the existence of the corresponding component in their last common ancestor.