Throughout the lengthy process of evolution, various animals have formed symbiotic associations with diverse microbes. Symbiotic associations between insects and microorganisms enable insects to adapt to a variety of ecological niches and exploit diverse food resources. In most cases, the host insects have evolved specialized cells and organs, called bacteriocytes and bacteriomes, in particular of the plant sap-feeding auchenorrhynchan insects of Hemiptera, to accommodate their bacterial symbionts. Recruitment of novel symbiont is a key strategy for hosts to escape dependence on outdated, inefficient symbiont. However, symbiont replacement is a dynamic process, and its occurrence – from emergence to eventual disappearance – is extremely difficult to ascertain. Prior work has indicated that, in many cicada lineages, the obligate symbiont Hodgkinia has been replaced by yeast-like fungal symbionts (YLS), and YLS replacement by another fungus or YLS has also occurred repeatedly. Considerable diversity has been observed in the genome of Hodgkinia among different cicada species, which possibly suggest that Hodgkinia independently infected the ancestors of different cicada lineages on multiple occasions.

In the present work, phylogenetic analyses reveal that Hodgkinia infected the common ancestor of cicadas and has co-diversified with its host cicadas before replaced by YLS. We demonstrate events of Hodgkinia loss in Cicadidae occurred at least seven times and YLS replaced by another Ophiocordyceps fungus or YLS occurred repeatedly at least 18 times. The most extraordinary replacements and supplementation of symbionts in Cicadidae could be closely related to that cicadas live underground usually for many years but aboveground only several days. Most importantly, we present direct evidence that the extremely genome-degraded Hodgkinia established a nested symbiosis with Karelsulcia – immigrated to the cytoplasm of Karelsulcia to seek refuge when being replaced by YLS. This answers the enduring puzzles in the life sciences of how host insects can break away from their dependence on an ancient and potentially inefficient symbiont and what adaptative or nonadaptive evolution may evolve before a symbiont is going to be replaced by another symbiont. The findings offer fresh insights into the endosymbiotic origins of cellular organelles and intensify our curiosity about the origins of life.