Functional and evolutionary implications of enhanced genomic analysis of rhomboid intramembrane proteases

Rhomboids are a recently discovered family of widely distributed intramembrane serine proteases. They have diverse biological functions, including the regulation of growth factor signaling, mitochondrial fusion, and parasite invasion. Despite their existence in all branches of life, the sequence identity between rhomboids is low. We have combined BLAST-based database mining with functional and structural data to generate a comprehensive genomic analysis of eukaryotic rhomboid-like proteins. We show that robust membrane topology models are necessary to classify active rhomboid proteases unambiguously, and we define rules for distinguishing predicted active proteases from the larger evolutionary group of rhomboid-like proteins. This leads to a revision of estimates of numbers of proteolytically active rhomboids. We identify three groups Of eukaryotic rhomboid-like proteins: true active rhomboids, a tightly clustered group of novel inactive rhomboids that we name the iRhoms, and a small number of other inactive rhomboid-like proteins. The active proteases are themselves subdivided into secretase and PARL-type (mitochondrial) subfamilies; these have distinct transmembrane topologies. This enhanced genomic analysis leads to conclusions about rhomboid enzyme function. It suggests that a given rhomboid can only cleave a single orientation of substrate, and that both products of rhomboid catalyzed intramembrane cleavage can be released from the membrane. Our phylogeny predictions also have evolutionary implications: Despite the complex classification of rhomboids, our data suggest that a rhomboid-type intramembrane protease may have been present in the last eukaryotic common ancestor.