Evolutionary limits to cooperation in microbial communities

Microbes produce many compounds that are costly to a focal cell but promote the survival and reproduction of neighboring cells. This observation has led to the suggestion that microbial strains and species will commonly cooperate by exchanging compounds. Here, we examine this idea with an ecoevolutionary model where microbes make multiple secretions, which can be exchanged among genotypes. We show that cooperation between genotypes only evolves under specific demographic regimes characterized by intermediate genetic mixing. The key constraint on cooperative exchanges is a loss of autonomy: strains become reliant on complementary genotypes that may not be reliably encountered. Moreover, the form of cooperation that we observe arises through mutual exploitation that is related to cheating and “Black Queen” evolution for a single secretion. A major corollary is that the evolution of cooperative exchanges reduces community productivity relative to an autonomous strain that makes everything it needs. This prediction finds support in recent work from synthetic communities. Overall, our work suggests that natural selection will often limit cooperative exchanges in microbial communities and that, when exchanges do occur, they can be an inefficient solution to group living.

‘Benefit-of-the-species’ arguments … provide for the reader an escape from inner conflict, exacting nothing emotionally beyond what most of us learn to accept in childhood, that most forms of life exploit and prey on one another.

Hamilton, 1975 (1)

Microbes typically live in dense communities containing many strains and species. These genetically diverse societies are widespread and central to how microbes affect us, including examples such as the gut microbiome, polymicrobial infections, and communities vital to bioremediation and nutrient cycling (2, 3). In these collectives, ecological interactions are thought to be both common and strong given that cell density is typically high and that microbes possess many phenotypes that influence the reproduction and survival of surrounding cells (4, 5). Such social traits include many secretions, such as extracellular enzymes and scavenging molecules (4–6), and other beneficial “leaky” traits, such as detoxification agents (7) or amino acids (8, 9).A central explanation for cooperative phenotypes in microbes is that they function to help cells of the same genotype (10, 11), which is backed up by a growing body of theory and experiments (12–18). However, it is also clear that, in nature, microbes commonly interact with many different genotypes (both different strains and species) in complex ecological networks (19–21). Do these different microbial genotypes cooperate with one another? Understanding this question is central to building models of microbial communities and how they will respond to both environmental and anthropogenic perturbations (22).Studies involving genetically engineered (8, 9, 23, 24) and/or artificially selected communities (23, 25, 26) emphasize how easily cooperation between genotypes can be achieved in the laboratory. Additionally, there are a growing number of suggestions that cooperation should commonly evolve between microbial strains and species (27–30). This view contrasts with empirical surveys of natural bacterial communities, which suggest that competitive interactions predominate over cooperative interactions (31). However, it has also been suggested that cooperation between different genotypes may explain the unculturability of many species in the laboratory when in monoculture (32–34). If correct, studies with culturable species could underestimate cooperativity in microbial communities.The potential for cooperation between different microbial genotypes then remains unclear. Indeed, we even lack clear predictions of what to expect. There is a need for general theory on cooperation between microbial genotypes. One microbial interaction that has been explored theoretically is syntrophy, where one species produces a toxic waste product that another species consumes (35–38). Syntrophy is likely to be ecologically important and under some conditions (36), can benefit both species. However, syntrophic species need not pay energetic costs to interact: one species is producing waste, and the other species is feeding. Such byproduct cooperation can, therefore, readily evolve but is fundamentally different to the exchange of costly secretions (39, 40). Other models have analyzed when costly cooperation between species is expected in microbes and other organisms (35, 39, 41). However, these models assume that there is no opportunity for one partner to express the beneficial trait of the other. Although this constraint will sometimes occur, there is considerable functional overlap in the cooperative traits of microbial species (7). In addition, the potential for horizontal gene transfer in microbes means that there is a broad scope for a focal strain to pick up the phenotypes of co-occurring strains and species (42–44).Here, we examine the potential for microbial cooperation between different strains and species. We base our work on the well-established models of within-genotype microbial cooperation for a single public good (12, 18, 45–47) so that the relationship to previous work is clear (SI Materials and Methods). We add one key feature to these models: we allow cells to invest in multiple distinct cooperative secretions, such that there is the potential for different genotypes to exchange secretions with one another. Our analysis shows that the degree of genetic mixing defines the potential for cooperation both within and between genotypes. Low mixing favors genotypes that produce all secretions, whereas high mixing favors genotypes that do not produce any at all. Only for intermediate levels of genetic mixing do we find between-genotype cooperation, where strains produce a subset of secretions and rely on other genotypes for the complementary traits. Moreover, the form of cooperation that emerges is inefficient and results in a loss of productivity relative to one genotype making all secretions. Natural selection limits both the occurrence and effectiveness of cooperation within microbial communities.