The long-known resistance to pathogens provided by host-associated microbiota fostered the notion that adding protective bacteria could prevent or attenuate infection. However, the identification of endogenous or exogenous bacteria conferring such protection is often hindered by the complexity of host microbial communities. Here, we used zebrafish and the fish pathogen Flavobacterium columnare as a model system to study the determinants of microbiota-associated colonization resistance. We compared infection susceptibility in germ-free, conventional and reconventionalized larvae and showed that a consortium of 10 culturable bacterial species are sufficient to protect zebrafish. Whereas survival to F. columnare infection does not rely on host innate immunity, we used antibiotic dysbiosis to alter zebrafish microbiota composition, leading to the identification of two different protection strategies. We first identified that the bacterium Chryseobacterium massiliae individually protects both larvae and adult zebrafish. We also showed that an assembly of 9 endogenous zebrafish species that do not otherwise protect individually confer a community-level resistance to infection. Our study therefore provides a rational approach to identify key endogenous protecting bacteria and promising candidates to engineer resilient microbial communities. It also shows how direct experimental analysis of colonization resistance in low-complexity in vivo models can reveal unsuspected ecological strategies at play in microbiota-based protection against pathogens.

由宿主相关微生物群提供的对病原体的长期已知抗性促进了这样一种观念，即添加保护性细菌可以预防或减轻感染。然而，赋予这种保护的内源性或外源性细菌的鉴定常常受到宿主微生物群落的复杂性的阻碍。在这里，我们使用斑马鱼和鱼类病原体Flavobacterium columnare作为模型系统来研究微生物群相关定植抗性的决定因素。我们比较了无菌、常规和再常规化幼虫的感染易感性，并表明 10 种可培养细菌物种的联合体足以保护斑马鱼。而生存到F. columnare感染不依赖于宿主先天免疫，我们使用抗生素失调来改变斑马鱼微生物群的组成，从而确定了两种不同的保护策略。我们首先确定了细菌Chryseobacterium massiliae分别保护幼虫和成年斑马鱼。我们还表明，9 种内源斑马鱼物种的集合，它们不会单独保护，赋予群落水平的感染抵抗力。因此，我们的研究提供了一种合理的方法来识别关键的内源性保护细菌和有希望的候选者来设计有弹性的微生物群落。它还展示了在低复杂性体内模型中对定植抗性的直接实验分析如何揭示在基于微生物群的病原体保护中发挥作用的意想不到的生态策略。