A molecular timescale for eukaryote evolution with implications for the origin of red algal-derived plastids

Abstract

In modern oceans, eukaryotic phytoplankton is dominated by lineages with red algal-derived plastids such as diatoms, dinoflagellates, and coccolithophores. Despite the ecological importance of these groups and many others representing a huge diversity of forms and lifestyles, we still lack a comprehensive understanding of their evolution and how they obtained their plastids. New hypotheses have emerged to explain the acquisition of red algal-derived plastids by serial endosymbiosis, but the chronology of these putative independent plastid acquisitions remains untested. Here, we establish a timeframe for the origin of red algal-derived plastids under scenarios of serial endosymbiosis, using Bayesian molecular clock analyses applied on a phylogenomic dataset with broad sampling of eukaryote diversity. We find that the hypotheses of serial endosymbiosis are chronologically possible, as the stem lineages of all red plastid-containing groups overlap in time. This period in the Meso- and Neoproterozoic Eras set the stage for the later expansion to dominance of red algal-derived primary production in the contemporary oceans, which profoundly altered the global geochemical and ecological conditions of the Earth.

Fig. 1. Serial plastid endosymbioses models as proposed by Stiller et al. (left) and Bodył et al. (right).

The tree topology shown here is based on the results obtained in our study. Further models have been suggested but are not compatible with this topology^,,. Numbers denote the level of endosymbiosis events. Note, Myzozoa were not included by Stiller et al. and the dashed line indicates engulfment of an ochrophyte by the common ancestor of Myzozoa as suggested by Sevcikova et al..

Fig. 2. Test for endosymbiotic gene transfers in red plastid-containing lineages based on the analysis of 320 single-gene ML trees.

For each clade in the ML tree to which a single taxonomic label could be assigned, relative frequencies with which all other clades in the tree were recovered as the closest sister group are given (for details, see ‘Methods’).

Fig. 3. Time-calibrated phylogeny of extant eukaryotes.

Divergence times were inferred with MCMCTree under an autocorrelated relaxed clock model and 33 fossil calibration points as soft-bound uniform priors (Table 1). The tree topology was reconstructed using IQ-TREE under the LG + C60 + G + F model and a constrained tree search following the OTU-reduced Bayesian CAT + GTR + G topology (Supplementary Fig. 4). Approximate likelihood calculations on the 320 gene concatenation under LG + G and a birth–death tree prior were used. Bars at nodes are 95% HPD. Bars corresponding to the first and last common ancestors of extant red plastid-donating and -containing lineages are highlighted in red and their stems are shaded as indicated. Crowns denote the common ancestors of the extant members of these groups. An absolute time scale in Ma and a geological time scale are shown. The tree depicted here was rooted on Amorphea. An equivalent time-calibrated tree rooted on Excavata is shown in Supplementary Fig. 7. Cry Cryptophyta, Hap Haptophyta, Myz Myzozoa, Och Ochrophyta, Rho Rhodophytina.

Fig. 4. Summary of inferred timeframes for the spread of complex red plastids using different software and models.

Vertical lines correspond to the lower and upper 95% HPD intervals from the nodes defining the branch of interest and dots indicate their posterior mean divergences. Faded boxes represent the temporal windows for the secondary endosymbioses, constrained by the 95% HPDs and tree topology under the two proposed symbiotic scenarios. Numbers at arrows denote the level of endosymbiosis events (compare with Fig. 1). AC autocorrelated clock model, UC uncorrelated clock model.