The microbiome in the hindgut of wood-feeding termites comprises various species of bacteria, archaea, and protists. This gut community is indispensable for the termite, which thrives solely on recalcitrant and nitrogen-poor wood. However, the difficulty in culturing these microorganisms has hindered our understanding of the function of each species in the gut. Although protists predominate in the termite gut microbiome and play a major role in wood digestion, very few culture-independent studies have explored the contribution of each species to digestion. Here, we report single-cell transcriptomes of four protists species comprising the protist population in worldwide pest Coptotermes formosanus. Comparative transcriptomic analysis revealed that the expression patterns of the genes involved in wood digestion were different among species, reinforcing their division of roles in wood degradation. Transcriptomes, together with enzyme assays, also suggested that one of the protists, Cononympha leidyi, actively degrades chitin and assimilates it into amino acids. We propose that C. leidyi contributes to nitrogen recycling and inhibiting infection from entomopathogenic fungi through chitin degradation. Two of the genes for chitin degradation were further revealed to be acquired via lateral gene transfer (LGT) implying the importance of LGT in the evolution of symbiosis. Our single-cell-based approach successfully characterized the function of each protist in termite hindgut and explained why the gut community includes multiple species.

The libraries prepared from single-cells of protist inhabiting the hindgut of wood-feeding termites Coptotermes formosanus following Quarz-seq protocol. Three libraries were prepared for the each spieces, i.e., Pseudotrichonympha grassii, Holomastigotoides hartmanii (commonly confused with Holomastigotoides mirabile), Holomastigotoides minor, and Cononympha leidyi (previously called Spirotrichonympha leidyi).   The generated reads were concatenated by species and assembled with Trinity. The apparent contamination in the assemblies were removed as described in Manuscript. The gene model was predicted from the assembly with Transdecoder. 

Although we carefully removed the contamination, the assembly files still contain the contaminatin.