Reevaluating the green contribution to diatom genomes

Abstract

Photosynthetic diatom plastids have long been suggested to have originated by the secondary endosymbiosis of a red alga. However, recent phylogenomic studies report a high number of diatom nuclear genes phylogenetically related to green algal and green plant genes. These were interpreted as endosymbiotic gene transfers (EGT) from a cryptic green algal endosymbiosis. We reanalyzed this issue using a larger set of red algal genomic data. We show that previous studies suffer from a taxonomic sampling bias and point out that a majority of gene phylogenies are either poorly resolved or do not describe EGT events. We finally show that genes having a complete descent from cyanobacteria to diatoms through primary and secondary EGTs have been mostly transferred via a red alga. We conclude that, even if some diatom genes still support a putative green algal origin, these are not sufficient to argue for a cryptic green algal secondary endosymbiosis.

Fig. 1.—

Phylogenetic reanalysis of putative EGT genes of green algal and plant origin in diatom genomes. (A) Summary of the results reported by Moustafa et al. (2009) and used as a reference for this work. (B) Classification of the protein-based phylogenetic tree topologies in five classes: Black—Trees not sufficiently resolved to allow any inference about gene evolutionary origin; Gray—Similarity searches retrieved a poor set of species, leading to inconclusive trees; Purple—Trees supporting vertical transmission in all eukaryotes. Blue—Trees compatible with a secondary EGT scenario; Green—Only Viridiplantae and diatom homologous sequences were retrieved by similarity searches. (C) Identification of the donor group (red or green alga) for the remaining putative EGT candidates.

Fig. 2.—

Examples of ML phylogenetic trees recovered in each of the five classes. Node supports lower than 70% are hidden. (A) Eukaryotic gene; (B) only Viridiplantae and diatoms; (C) poor species sampling; (D) putative secondary EGT; (E) unresolved tree. Sequence accession numbers are from GenBank or JGI.