Obligate endosymbiosis enables genome expansion during eukaryogenesis

Abstract

The endosymbiosis of an alpha-proteobacterium that gave rise to mitochondria was one of the key events in eukaryogenesis. One striking outcome of eukaryogenesis was a much more complex cell with a large genome. Despite the existence of many alternative hypotheses for this and other patterns potentially related to endosymbiosis, a constructive evolutionary model in which these hypotheses can be studied is still lacking. Here, we present a theoretical approach in which we focus on the consequences rather than the causes of mitochondrial endosymbiosis. Using a constructive evolutionary model of cell-cycle regulation, we find that genome expansion and genome size asymmetry arise from emergent host-symbiont cell-cycle coordination. We also find that holobionts with large host and small symbiont genomes perform best on long timescales and mimic the outcome of eukaryogenesis. By designing and studying a constructive evolutionary model of obligate endosymbiosis, we uncovered some of the forces that may drive the patterns observed in nature. Our results provide a theoretical foundation for patterns related to mitochondrial endosymbiosis, such as genome size asymmetry, and reveal evolutionary outcomes that have not been considered so far, such as cell-cycle coordination without direct communication.

Fig. 1. Overview of the model and experimental setup.

a Host and symbiont each consist of a genome (with discrete beads representing genes and binding sites), a gene regulatory network and a cell cycle. Holobionts (disks) live in space and comprise one host (black square) and one or more symbionts (purple circles) that compete for nutrients. For both the host and symbiont, a division event (blue arrow) and a death event (blurred shape) are shown. Lattice site colors depict nutrient abundance for an influx of n_influx = 30. b Two in silico evolution experiments with endosymbiosis were performed (green and purple boxes): one starting with primitive cells on a nutrient gradient and one starting with pre-evolved prokaryotes from our previous work. Simulation time and grid sizes are shown for each experiment; colors depict nutrient abundance.

Fig. 2. Primitive holobionts adapt to nutrient gradient.

a Population size increases, coinciding with an increase in symbiont numbers. b Genomes expand early, and later genome size symmetry breaks between host and symbiont. c Regulatory repertoires expand further than free-living cells for all hosts and symbionts that survive. The maximum population size is given by the grid size: 25 × 275 = 6875. Adaptation generally occurs before t = 10⁷ (see Supplementary Fig. S2 for trajectories until t = 2 ⋅ 10⁷). In panels (b) and (c), the size of the markers is scaled by the final population size.

Fig. 3. Evolutionary dynamics of the most successful replicate with primitive FECA, P9.

Each panel shows a different variable in spacetime. The overlaid lines represent summaries across the entire gradient (for the top panel, the overlaid line shows total population size rather than local cell density, in correspondence with Fig. 2a). On the right, genomes of the host and symbiont in the ancestor at t = 9 ⋅ 10⁶ are depicted from top to bottom, along with the regulatory interactions (L denotes total genome size, R regulatory repertoire size). For an example of a different evolutionary trajectory, see Supplementary Fig. S3.

Fig. 4. Emergence of an r- and K-strategy along the ancestry of P9.

Over time, cell density (a) and symbiont number (b) have increased, but growth measured as invasion speed (c) and nutrient availability (d) have decreased. The final population resembles the most recent ancestor, but strains at high and low nutrient influxes have adapted to those specific environments. Diamonds depict the conditions in which the ancestor lived.

Fig. 5. Evolution of host–symbiont cell-cycle coordination.

Host–symbiont cell-cycle coordination evolves, first through differential cell-cycle timing, later through niche differentiation whereby the host becomes a generalist and the symbiont a specialist. Each row shows the growth rates (left panels) and cell-cycle durations (right panels) of an ancestor of the P9 replicate. In contrast to Fig. 4, the x-axis here depicts fixed nutrient conditions where host and symbiont cell-cycle dynamics are assessed independently. The gray areas in the cell-cycle duration panels mark the minimum possible duration of a successful cell cycle (with enough time for replication) given the genome size of the host (filled) and symbiont (hatch). The cell-cycle duration is only drawn in bold for the range where ρ > 0.5, which roughly coincides with r > 0.

Fig. 6. Asymmetry in genome size correlates strongly with asymmetry in cell-cycle efficiency between host and symbiont.

Along the ancestral lineage, genome size and cell-cycle efficiency are measured (see “Calculating cell-cycle efficiency” in “Methods”). Trajectories are shown for the experiments where the largest genome size asymmetry evolved (P1, P9 and C9–12); the trajectory of P9 is in bold. One technical replicate is shown for each of the replicates initialized with complex FECA. For the experiment initialized with primitive FECA, the size of the markers is scaled by the final population size.