COB: a comprehensive database of chloroplast outer envelope beta-barrel proteins

Despite their central role in metabolite exchange, lipid trafficking, and protein import, chloroplast outer envelope beta-barrel proteins lack a dedicated comprehensive sequence database spanning many plant proteomes. Here we present COB (chloroplast outer-envelope beta-barrel), comprising 16,586 sequences organized across ten protein categories and an uncharacterized group, constructed using a machine learning classifier that identifies chloroplast beta-barrels from features of evolutionary protein contact maps. COB reveals that Streptophyta have more barrels overall and a greater variety of solute transporters than Chlorophyta. COB shows considerable structural diversity across categories, including variable beta-strand counts and a high prevalence of open barrel conformations not seen in bacterial outer membrane proteins. Structure predictions for Arabidopsis thaliana outer envelope proteins identify candidate hybrid barrel assemblies, with TOC159 family members as universal interaction partners, and single-chain multi-barrel architectures previously described only in Gram-negative bacteria. Chloroplast barrels also have more open topologies and shorter strands than bacterial barrels. COB provides a comprehensive sequence resource for chloroplast outer envelope beta-barrels, establishing a foundation for investigating the evolution, structure, and functions of this essential protein-fold.