Tracing the taxonomic distribution of plant cell wall degrading enzymes across the tree of life using feature architecture aware orthology assignments

Abstract

The decomposition of plant material is a key driver of the global carbon cycle, traditionally attributed to fungi and bacteria. However, some invertebrates also possess orthologs to bacterial or fungal cellulolytic enzymes, likely acquired via horizontal gene transfer. This reticulated mode of evolution necessitates ortholog searches in large taxon sets to comprehensively map the repertoire of plant cell wall degrading enzymes (PCDs) across the tree of life, a task surpassing capacities of current software. Here, we use fDOG, a novel profile-based ortholog search tool to trace 235 potential PCDs across more than 18,000 taxa. fDOG allows to start the ortholog search from a single protein sequence as a seed, it performs on par with state-of-the-art software that require the comparison of entire proteomes, and it is unique in routinely scoring protein feature architecture differences between the seed protein and its orthologs. Visualizing the presence-absence patterns of PCD orthologs using a UMAP highlights taxa where recent changes in the enzyme repertoire indicate a change in lifestyle. Three invertebrates have a particularly rich set of PCD orthologs encoded in their genome. Only few of the orthologs show differing protein feature architectures relative to the seed that suggest functional modifications. Thus, the corresponding species represent lineages within the invertebrates that may contribute to the global carbon cycle. This study shows how fDOG can be used to create a multi-scale view on the taxonomic distribution of a metabolic capacity that ranges from tree of life-wide surveys to individual feature architecture changes within a species.