2011; respectively, clades I and II in Hagino et al. 2011). The diversity within each of these clades differed according to the

marker: for example clade β was not well-defined in the rpl16 phylogeny, while cox3 showed the highest inter- and intra-clade diversity. G. oceanica and E. huxleyi strains were separated and mitochondrial clades Metformin in vivo α and β retrieved in the 26 strain tree (Fig. 2) inferred from concatenated sequences of three genes representative of each genomic compartment (28S rDNA, tufA and cox1). Despite the fact that the two morpho-species were genetically delineated in this analysis, no relationship was found between the genetic grouping and morphotypes within E. huxleyi. The comparison of multiple genes in the search for genetic barcodes for accurate species delineation is relatively common for multicellular eukaryotes (plants, animals, and fungi), but has rarely been undertaken for the older and highly diverse protistan lineages (Pawlowski et al. 1997), where nuclear ribosomal DNA markers are still by far the most commonly Sirolimus cost used barcodes. However, ribosomal genes, occurring in numerous copies in the nuclear genome and interacting with numerous partners during protein synthesis, are under strong purifying selection pressure and are best suited to resolve high-rank relationships due to their slow evolutionary rate and very high level of conservation

(Sogin et al. 1986). For marine protists, a particularly high level of conservation of rDNA genes

is theoretically expected due to their potentially very high medchemexpress effective population size (Piganeau et al. 2011). Our multigene analysis confirms that rDNAs evolve too slowly to discriminate morpho-species within the Emiliania/Gephyrocapsa species complex, which diversified relatively recently during the Quaternary. Likewise, the 16S rDNA from the plastid genome, also involved in protein synthesis, is highly conserved, as is the rbcL gene that codes for the large subunit of RuBisCO and thus plays a central role in carbon fixation by photosynthesis. Neither of these conserved plastid markers are suited for either identification or evolutionary studies of E. huxleyi/G. oceanica. However, all other gene markers tested in this study exhibited higher nucleotide substitution rates, with the partial sequences of plastidial tufA (long) and mitochondrial dam displaying the highest degrees of variability for the relatively large set of strains analysed, with mean overall substitution rates of, respectively, 6.4% and 6.0% (Table 1). The general pattern that emerges from our data set is that the plastidial markers do not produce consistent groupings both between and within morpho-species, while the mitochondrial markers delineate a coherent set of genetic clades at both inter- and intra-morpho-species levels.