Humans, and many other species, are host to diverse symbionts. It is often suggested that the mutual benefits of host-microbe relationships can alone explain cooperative evolution. Here, we evaluate this hypothesis with evolutionary modelling. Our model predicts that mutual benefits are insufficient to drive cooperation in systems like the human microbiome, because of competition between symbionts. However, cooperation can emerge if hosts can exert control over symbionts, so long as there are constraints that limit symbiont counter evolution. We test our model with genomic data of two bacterial traits monitored by animal immune systems. In both cases, bacteria have evolved as predicted under host control, tending to lose flagella and maintain butyrate production when host-associated. Moreover, an analysis of bacteria that retain flagella supports the evolution of host control, via toll-like receptor 5, which limits symbiont counter evolution. Our work puts host control mechanisms, including the immune system, at the centre of microbiome evolution.

人类和许多其他物种都是不同共生体的宿主。人们经常认为，宿主-微生物关系的互惠互利可以单独解释合作进化。在这里，我们用进化模型评估这个假设。我们的模型预测，由于共生体之间的竞争，互惠互利不足以推动人类微生物组等系统的合作。然而，如果宿主能够对共生体施加控制，只要存在限制共生体逆向进化的约束，合作就可能出现。我们使用动物免疫系统监测的两种细菌性状的基因组数据来测试我们的模型。在这两种情况下，细菌都在宿主控制下按照预期进化，在与宿主相关时往往会失去鞭毛并维持丁酸盐的产生。此外，对保留鞭毛的细菌的分析支持了通过Toll样受体5进行宿主控制的进化，这限制了共生体的逆进化。我们的工作将包括免疫系统在内的宿主控制机制置于微生物组进化的中心。