On, BLAST searches were conducted based on annotated sequences. If multiple splice variants of the protein were reported, the canonical sequence was used.Sequence Alignment and Phylogenetic AnalysesThe amino acid sequence alignment was performed using ClustalW [26] with a gap opening penalty of 10 and gap extension penalty of 0.2. The highly variable 39 terminal ends of the sequences were trimmed to avoid ambiguity. No other variable regions were excluded from the alignment. Phylogenetic analyses were performed using neighbour-joining (NJ) and Bayesian inference (BI) approaches, in MEGA [27] and MrBayes [28] respectively. Members of other P-type II ATPases sub-families were used as outgroups (Table S1) in all phylogenetic reconstructions. Outgroup sequences included the Daporinad Plasma Membrane Ca2+ ATPase (ATP2B1), Secretory Pathway Ca2+ ATPase (ATP2C1 and ATP2C2), Na+/K+-transporting ATPase (ATP4A1), K+-transporting ATPase (ATP4A), as well as the fungi-specific Na+/K+ ATPase (ACU1) and Na+ transport ATPase (ENA1) [29]. We used a Bayesian inference method for tree construction with the WAG model of amino acid substitution. This model was selected prior to the final analysis using a model jumping algorithm implemented in MrBayes [28,30]. This algorithm regularly swaps between 9 different fixed-rate amino acid models throughout the analysis and selects the model with the highest contribution to posterior probability density [28,30]. The WAG model can handle a large number of sequences and is applicable to a wide range of protein families, but retains the advantages of a maximum-likelihood approach and accounts for multiple substitutions at the same site [31]. Three independent runs of 4 Markov chains were conducted for 1,000,000 generations with a sampling frequency of 10 and the first 25 of sampled trees discarded as burn-in. We confirmed the topology of the Bayesian tree with a cluster based neighbour-joining tree using pairwise deletion and the JonesTaylor-Thornton (JTT) amino acid model [32]. Node support was analyzed using 1000 Fingolimod (hydrochloride) bootstrap replicates.Results and Discussion Overall Phylogenetic PatternThe SERCA alignment consisted of 81 sequences (61 unique taxa) spanning 1575 amino acids and contained 220 conserved and 818 parsimony-informative sites. Both the BI and NJ analyses returned highly congruent tree topologies. There were two major monophyletic clades. The first group contains clades A, B and C, which consist of metazoan, fungal, and plant sequences, respectively (Fig. 1). The second group contains clade D that encompasses plant and protist sequences (Fig. 1). Within clade A, the chordates are monophyletic and contain twoThe Evolution of Sarco(endo)plasmic Calcium ATPaseThe Evolution of Sarco(endo)plasmic Calcium ATPaseFigure 1. Bayesian phylogenetic reconstruction 1527786 of SERCA amino acid sequences from 57 taxa. The numbers at the nodes indicate posterior probabilities/bootstrap supports. Nodes highlighted with gray circles represent consensus neighbouring-joining (NJ) and Bayesian Inference (BI) analyses with bootstrap support higher than 70 . doi:10.1371/journal.pone.0052617.greciprocally monophyletic clades corresponding to vertebrates and tunicates.SERCA Gene Duplication and EvolutionWithin metazoans, the SERCA sequences of the chordates form a well supported monophyletic group that includes two sister clades, corresponding to the vertebrates and tunicates. In vertebrates, each of the three SERCA isoforms (i.e. SERCA1-3; coded by AT.On, BLAST searches were conducted based on annotated sequences. If multiple splice variants of the protein were reported, the canonical sequence was used.Sequence Alignment and Phylogenetic AnalysesThe amino acid sequence alignment was performed using ClustalW [26] with a gap opening penalty of 10 and gap extension penalty of 0.2. The highly variable 39 terminal ends of the sequences were trimmed to avoid ambiguity. No other variable regions were excluded from the alignment. Phylogenetic analyses were performed using neighbour-joining (NJ) and Bayesian inference (BI) approaches, in MEGA [27] and MrBayes [28] respectively. Members of other P-type II ATPases sub-families were used as outgroups (Table S1) in all phylogenetic reconstructions. Outgroup sequences included the Plasma Membrane Ca2+ ATPase (ATP2B1), Secretory Pathway Ca2+ ATPase (ATP2C1 and ATP2C2), Na+/K+-transporting ATPase (ATP4A1), K+-transporting ATPase (ATP4A), as well as the fungi-specific Na+/K+ ATPase (ACU1) and Na+ transport ATPase (ENA1) [29]. We used a Bayesian inference method for tree construction with the WAG model of amino acid substitution. This model was selected prior to the final analysis using a model jumping algorithm implemented in MrBayes [28,30]. This algorithm regularly swaps between 9 different fixed-rate amino acid models throughout the analysis and selects the model with the highest contribution to posterior probability density [28,30]. The WAG model can handle a large number of sequences and is applicable to a wide range of protein families, but retains the advantages of a maximum-likelihood approach and accounts for multiple substitutions at the same site [31]. Three independent runs of 4 Markov chains were conducted for 1,000,000 generations with a sampling frequency of 10 and the first 25 of sampled trees discarded as burn-in. We confirmed the topology of the Bayesian tree with a cluster based neighbour-joining tree using pairwise deletion and the JonesTaylor-Thornton (JTT) amino acid model [32]. Node support was analyzed using 1000 bootstrap replicates.Results and Discussion Overall Phylogenetic PatternThe SERCA alignment consisted of 81 sequences (61 unique taxa) spanning 1575 amino acids and contained 220 conserved and 818 parsimony-informative sites. Both the BI and NJ analyses returned highly congruent tree topologies. There were two major monophyletic clades. The first group contains clades A, B and C, which consist of metazoan, fungal, and plant sequences, respectively (Fig. 1). The second group contains clade D that encompasses plant and protist sequences (Fig. 1). Within clade A, the chordates are monophyletic and contain twoThe Evolution of Sarco(endo)plasmic Calcium ATPaseThe Evolution of Sarco(endo)plasmic Calcium ATPaseFigure 1. Bayesian phylogenetic reconstruction 1527786 of SERCA amino acid sequences from 57 taxa. The numbers at the nodes indicate posterior probabilities/bootstrap supports. Nodes highlighted with gray circles represent consensus neighbouring-joining (NJ) and Bayesian Inference (BI) analyses with bootstrap support higher than 70 . doi:10.1371/journal.pone.0052617.greciprocally monophyletic clades corresponding to vertebrates and tunicates.SERCA Gene Duplication and EvolutionWithin metazoans, the SERCA sequences of the chordates form a well supported monophyletic group that includes two sister clades, corresponding to the vertebrates and tunicates. In vertebrates, each of the three SERCA isoforms (i.e. SERCA1-3; coded by AT.
