Shifts from cooperative to individual-based predation defense determine microbial predator-prey dynamics

Predation defense is an important feature of predator-prey interactions adding complexity to ecosystem dynamics. Prey organisms have developed various strategies to escape predation which differ in mode (elude vs. attack), reversibility (inducible vs. permanent), and scope (individual vs. cooperative defenses). While the mechanisms and controls of many singular defenses are well understood, important ecological and evolutionary facets impacting long-term predator-prey dynamics remain underexplored. This pertains especially to trade-offs and interactions between alternative defenses occurring in prey populations evolving under predation pressure. Here, we explored the dynamics of a microbial predator-prey system consisting of bacterivorous flagellates (Poteriospumella lacustris) feeding on Pseudomonas putida. Within five weeks of co-cultivation corresponding to about 35 predator generations, we observed a consistent succession of bacterial defenses in all replicates (n = 16). Initially, bacteria expressed a highly effective cooperative defense based on toxic metabolites, which brought predators close to extinction. This initial strategy, however, was consistently superseded by a second mechanism of predation defense emerging via de novo mutations. Combining experiments with mathematical modeling, we demonstrate how this succession of defenses is driven by the maximization of individual rather than population benefits, highlighting the role of rapid evolution in the breakdown of social cooperation.


Supplementary material
Mutations identified in filamentous P. putida Table S1: Single nucleotide variations in cell division-related genes found in filamentous isolates of P. putida (n=9) with reference to a single-celled control strain.

Isolate ID Position in Genome
Base Length distribution of bacterial filaments

Additional simulation results
To elucidate why filamentous bacteria become dominant over single-celled toxin-producing bacteria, and to determine how sensitive this result is to the costliness of filamentation compared to the costliness of toxin production, we ran additional simulations. In those, we varied the cost of filamentation and performed a more detailed analysis on the results.
These simulations show that the eventual dominance of filamentous bacteria is highly robust: the simulated dynamics are almost identical even if the growth rate of filamentous bacteria was reduced by 40% (Fig. S5). For comparison, toxin production carries a growth cost of 11% (Table 5). The reason behind this result can be found in the relative costs (measured by the growth rates) and benefits (measured by the predation rates) of the two strategies. Despite being much less efficient on a population level, filamentation is a highly effective form of defense on an individual level, as the grazing risk of filamentous bacteria is strongly reduced compared to that of toxin-producing bacteria ( Fig. S6 A-C, dashed lines). This benefit is so strong that it can outweigh even a severe growth cost ( Fig. S6 A-C, solid lines), and allows the filamentous bacteria to quickly rise to dominance unless the cost is very high (as seen in Fig. S5 D).
Interestingly, if filamentation comes at a cost, it is always inferior to the toxin-producing strategy when toxin concentration is high enough to inhibit flagellate growth (days 10-15 in all panels of Fig.  S6). This is because, while all bacteria benefit from the toxin-induced grazing inhibition, it actually removes the advantage that filamentous bacteria have over toxin-producing bacteria. From a competitive viewpoint, filamentous bacteria benefit most strongly from high grazing pressure combined with a lack of toxin production (days 30-40 in Fig. S6 A-C), as this combination leaves the non-filamentous bacteria completely vulnerable while the filamentous bacteria benefit from their grazing protection. This generates a self-stabilizing effect of filamentation: because the majority of the available resources accumulates in bacterial filaments, the density of single-celled bacteria remains too low to trigger toxin production; this, in turn, keeps predator density high, which causes the filamentous bacteria to maintain their advantage. Figure S6: Per capita growth rates and predation rates belonging to the simulated dynamics in Fig. S5. Growth and predation rates of the "toxin-producing" strategy (Bo + Bx) are shown in light blue, those of the "filamentation" strategy (Bfs + Bff) in orange. Solid lines and filled symbols represent growth rates; dashed lines and crosses represent predation rates. Values represent the average for each day for clarity of presentation; note that days 1-6 do not contain values for the filamentation strategy, as this only emerges on day 7.
Finally, we modeled a hypothetical scenario where a "cheater" mutant arises: a genotype where upregulation of toxin production is suppressed, which allows them to benefit from the toxins produced by others without having to carry the cost themselves. It is well known that, without additional factors such as spatial structure, such cheaters will always have a higher fitness, and this is indeed reflected in the simulated dynamics (Fig. S7). However, it is notable that while the cheaters clearly have an advantage over the original toxin-producing strategy, their advantage is not nearly as strong as that of the filamentous bacteria. While filaments typically rise to dominance before day 30 (Fig. S5), cheaters remain outnumbered by non-cheaters after 100 days (Fig. S7). Clearly, losing the costs of defense is not nearly as advantageous as gaining an additional form of grazing protection, even when this extra protection comes at a substantial cost.