Search for a command to run...
Abstract Chlorophyll is one of the most abundant pigments on Earth. Although its degradation is well understood in plants, the role of prokaryotes in this process – despite their vast metabolic capabilities – remains unknown. Recent developments in the field of AI-predicted protein structures have opened new avenues for investigating functional homologies between evolutionary-distant organisms previously inaccessible through traditional sequence- or profile-based methods. Here, we present the first evidence of Chlorophyll a (Chl a ) degradation by prokaryotes, discovered through a novel bioinformatic framework which bridges the gap across the domains of life via structural alignments of functionally characterised plant proteins, followed by structure similarity graph-based clustering. Metagenomic sequencing data was assembled and binned, yielding over 70,000 medium- to high-quality genomes in total, furthermore publicly available datasets containing genomes from prokaryotic isolates, metagenome-assembled genomes, as well as single-cell genomes were then mined for prokaryotic homologues of Chl a degradation genes. Our analysis revealed over 400 genomes from diverse taxonomic groups and habitats that possess a complete pathway, more than 50% stemming from isolates. Additionally, many other genomes harbour partial pathways, suggesting that Chl a degradation capabilities are globally widespread across diverse ecosystems. We then validated our in silico findings using the model organism Shewanella acanthi and confirmed its Chl a degradation capability via growth experiments, fluorescence spectroscopy and HPLC analyses. Our findings reveal a previously unrecognised pathway in prokaryotes, highlighting the power of structure-based remote homology detection for uncovering metabolic capabilities and evolutionary relationships.