Obligate mutualistic cooperation limits evolvability

Abstract

Cooperative mutualisms are widespread and play fundamental roles in many ecosystems. Given that these interactions are often obligate, the Darwinian fitness of the participating individuals is not only determined by the information encoded in their own genomes, but also the traits and capabilities of their corresponding interaction partners. Thus, a major outstanding question is how obligate cooperative mutualisms affect the ability of organisms to adapt evolutionarily to changing environmental conditions. Here we address this issue using a mutualistic cooperation between two auxotrophic genotypes of Escherichia coli that reciprocally exchanged costly amino acids. Amino acid-supplemented monocultures and unsupplemented cocultures were exposed to stepwise increasing concentrations of different antibiotics. This selection experiment reveals that metabolically interdependent bacteria are generally less able to adapt to environmental stress than autonomously growing strains. Moreover, obligate cooperative mutualists frequently regain metabolic autonomy, resulting in a collapse of the mutualistic interaction. Together, our results identify a limited evolvability as a significant evolutionary cost that individuals have to pay when entering into an obligate mutualistic cooperation.

Fig. 1. Design of the evolution experiment.

In a previous study, serial coevolution of two bacterial genotypes of Escherichia coli, which were auxotrophic for the two amino acids tryptophan (ΔtrpB, red cells) or tyrosine (ΔtyrA, blue cells), had resulted in the evolution of obligate mutualistic cooperation between both cell types. One of these mutualistic consortia was used in the current study. The two strains were either grown together in coculture (minimal medium) or in individual monocultures (minimal medium + required amino acid, 100 µM each, red circle, blue triangle). Initial populations, which were sensitive to the four different antibiotics ampicillin, kanamycin, chloramphenicol and tetracycline, were serially propagated for 15 transfers, during which the concentration of these four antibiotics was gradually increased.

Fig. 2. Mutualistic cooperation limits the ability of strains to adapt to environmental stress.

a, b Mean growth (±95% confidence interval, n = 80 per point) quantified as OD_600nm and c, d proportion of surviving replicates in percent (n = 80 per strain) of auxotrophic monocultures (TRP, TYR) and mutualistic cocultures (CO, purple circles) of the tryptophan (TRP, red triangles) and tyrosine (TYR, blue star) auxotrophic strains throughout the evolution experiment. Antibiotic concentrations were increased in a stepwise manner after each transfer (i.e. every 72 h) (Supplementary Fig. 5). Grey-shaded areas indicate periods without antibiotic treatment. The green triangles above represent the increasing antibiotic concentrations in the evolution experiment (a–d). a, c, e kanamycin treatment, b, d, f tetracycline treatment. Dashed lines mark the point when antibiotic concentrations exceeded sub-MIC values. Monocultures were supplemented (⁺) with both amino acids (100 µM each), while cocultured bacteria depended on the amino acids provided by their respective cross-feeding partner (⁻). e, f Clustering trees of cell density profiles of experimental cultures across transfers indicate differences in the evolutionary trajectories taken by the different populations. Each leaf within a given tree represents a replicate (n = 80 per strain). A radial embedding layout was used to display trees. For exact P-values, see Supplementary Table 3. Source data are provided as a Source Data file.

Fig. 3. Strain-level differences cause increased susceptibility of cooperative consortia to environmental stress.

Minimum inhibitory concentration (MIC) values of ancestral (Supplementary Fig. 3) and derived strains and consortia that had evolved in the presence of a chloramphenicol and b tetracycline were assessed. ΔMIC is the difference between both values, thus indicating the increase in resistance over the course of the evolution experiment. Cocultures (CO, purple boxes) and monocultures of the tryptophan (TRP, red boxes) and tyrosine auxotrophs (TYR, blue boxes) of coevolved and monoevolved populations were analysed with (⁺) or without (⁻) amino acid supplementation (100 µM each). Box plots show median values (horizontal line in boxes) and the upper and lower quartiles (i.e. 25–75% of data, boxes). Whiskers indicate the 1.5x interquartile range. Different letters above boxes indicate significant differences between groups (two-sided Mann–Whitney U-test followed by Benjamini–Hochberg correction: P < 0.05, n = 8). For exact P-values, see Supplementary Table 4. Source data are provided as a Source Data file.

Fig. 4. Both coevolutionary history and mutualistic dependence affect the ability of cooperative mutualists to adapt to environmental change.

Shown is the growth of coevolved auxotrophs (CO_EVO, purple circles and boxes) and cocultured monoevolved auxotrophs (CO_AUX, green circles and boxes) determined as population density (OD_600nm) with (⁺) or without (⁻) supplementation of tryptophan and tyrosine (100 µM each). The antibiotic, to which the respective consortia have been exposed in the evolution experiment, is indicated in each panel. Cultures were grown with a, b chloramphenicol, d, e tetracycline or c, f not treated with any antibiotic. The dashed lines mark the typical working concentration of the respective antibiotic. Data is shown as (a, b, d, e) mean (±95% confidence interval) or as c, f box plots with median values (horizontal line in boxes) and the upper and lower quartiles (i.e. 25–75% of data, boxes). Whiskers indicate the 1.5x interquartile range. Different letters above boxes indicate significant differences between groups (two-sided Mann–Whitney U-test followed by Benjamini–Hochberg correction: P < 0.05, n = 16). For exact P-values, see Supplementary Table 6. Source data are provided as a Source Data file.

Fig. 5. Environmental stress favours reversion to metabolic autonomy.

Shown is the population-level proportion of initially auxotrophic genotypes that evolved in the presence of chloramphenicol (left) and tetracycline (right), which remained auxotrophic (filled bar) or reverted to prototrophy (hatched bars). Tryptophan auxotrophic strains (TRP) are depicted in red and tyrosine auxotrophic strains (TYR) in blue. Populations of monoevolved genotypes (TRP_M, TYR_M) and genotypes isolated from coevolved cultures (TRP_CO, TYR_CO) are compared. The plotted colony counts (%) are the number of colonies analysed per strain relative to the total number of colonies tested in the respective cultures. In both treatments, the reversion rates of coevolved tyrosine auxotrophs were significantly higher than those of their monoevolved counterpart (chloramphenicol-treated TYR two-sided Pearson χ² test: P = 5.4 × 10⁻⁷⁰, n = 240–285; tetracycline-treated TYR two-sided Pearson χ² test: P = 4.8 × 10⁻⁹², n = 240–342). Source data are provided as a Source Data file.