The evolution of cooperation is thought to be promoted by pleiotropy, whereby cooperative traits are co-regulated with traits that are important for personal fitness. However, this hypothesis faces a key challenge: what happens if mutation targets a cooperative trait specifically rather than the pleiotropic regulator? Here we explore this question with the bacterium Pseudomonas aeruginosa, which cooperatively digests complex proteins using elastase. We empirically measure and theoretically model the fate of two mutants – one missing the whole regulatory circuit behind elastase production and the other with only the elastase gene mutated – relative to the wild-type. We first show that, when elastase is needed, neither of the mutants can grow if the wild-type is absent. And, consistent with previous findings, we show that regulatory gene mutants can grow faster than the wild-type when there are no pleiotropic costs. However, we find that mutants only lacking elastase production do not outcompete the wild-type, because the individual cooperative trait has a low cost. We argue that the intrinsic architecture of molecular networks makes pleiotropy an effective way to stabilize cooperative evolution. While individual cooperative traits experience loss-of-function mutations, these mutations may result in weak benefits, and need not undermine the protection from pleiotropy.
CFUs of mixes of strains in CAA and BSA at 0 and 48 hours
Population sizes of each of the strains (wildtype WT, lasB mutants DLB and lasR mutants DLR) were quantified in monoculture and in coculture at different ratios of WT to mutant (9:1, 1:1 or 1:9). Experiments were conducted in two growth media: CAA (casamino acids + M9), which all strains could take up and BSA (casamino acids + M9 + BSA, which could only be digested using elastase). To distinguish the strains in cocultures, fluorescently-labelled strains (YFP or DsRed) were used and mixed in all possible combinations to test for labelling effects. We quantified population sizes by plating out dilution series and counting colony-forming units. These are listed in the file for t=0h and t=48h. This shows how the ratio of the strains and the total population size change throughout the experiment. These data were used to generate Fig. 1B, 2B and 3.
CFUs_CAA_BSA_0h_48h.xlsx
Growth curves (OD_600) over 48 hours in CAA medium
Population sizes of each of all experimental treatments (monocultures of wildtype WT, lasB mutants DLB and lasR mutants DLR) and cocultures of WT and one of the mutants at different ratios (9:1, 1:1 or 1:9) growing in CAA medium (casamino acids, which all strains should be able to take up). We quantified population sizes using a Tecan plate reader measuring OD_600. These are listed in the file every hour for 48 hours. These data were used to generate Fig. 1A.
growth_curves_CAA.xlsx
Growth curves (OD_600) over 48 hours in BSA medium
Population sizes of each of all experimental treatments (monocultures of wildtype WT, lasB mutants DLB and lasR mutants DLR) and cocultures of WT and one of the mutants at different ratios (9:1, 1:1 or 1:9) growing in BSA medium (casamino acids with BSA, which can only be degraded by elastase). We quantified population sizes using a Tecan plate reader measuring OD_600. These are listed in the file every hour for 48 hours. These data were used to generate Fig. 2A.
growth_curves_BSA.xlsx