Metapangenomics of wild and cultivated banana microbiome reveals a plethora of host-associated protective functions

BACKGROUND: Microbiomes are critical to plants, promoting growth, elevating stress tolerance, and expanding the plant’s metabolic repertoire with novel defense pathways. However, generally microbiomes within plant tissues, which intimately interact with their hosts, remain poorly characterized. These endospheres have become a focus in banana (Musa spp.)—an important plant for study of microbiome-based disease protection. Banana is important to global food security, while also being critically threatened by pandemic diseases. Domestication and clonal propagation are thought to have depleted protective microbiomes, whereas wild relatives may hold promise for new microbiome-based biological controls. The goal was to compare metapangenomes enriched from 7 Musa genotypes, including wild and cultivated varieties grown in sympatry, to assess the host associations with root and leaf endosphere functional profiles. RESULTS: Density gradients successfully generated culture-free microbial enrichment, dominated by bacteria, with all together 24,325 species or strains distinguished, and 1.7 million metagenomic scaffolds harboring 559,108 predicted gene clusters. About 20% of sequence reads did not match any taxon databases and ~ 62% of gene clusters could not be annotated to function. Most taxa and gene clusters were unshared between Musa genotypes. Root and corm tissues had significantly richer endosphere communities that were significantly different from leaf communities. Agrobacterium and Rhizobium were the most abundant in all samples while Chitinophagia and Actinomycetia were more abundant in roots and Flavobacteria in leaves. At the bacterial strain level, there were > 2000 taxa unique to each of M. acuminata (AAA genotype) and M. balbisiana (B-genotype), with the latter ‘wild’ relatives having richer taxa and functions. Gene ontology functional enrichment showed core beneficial functions aligned with those of other plants but also many specialized prospective beneficial functions not reported previously. Some gene clusters with plant-protective functions showed signatures of phylosymbiosis, suggesting long-standing associations or heritable microbiomes in Musa. CONCLUSIONS: Metapangenomics revealed key taxa and protective functions that appeared to be driven by genotype, perhaps contributing to host resistance differences. The recovery of rich novel taxa and gene clusters provides a baseline dataset for future experiments in planta or in vivo bacterization or engineering of wild host endophytes. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s40793-023-00493-x.