α/β Hydrolases: Toward Unraveling Entangled Classification.

α/β Hydrolase-like enzymes form a large and functionally diverse superfamily of proteins. Despite retaining a conserved structural core consisting of an eight-stranded, central β-sheet flanked with six α-helices, they display a modular architecture allowing them to perform a variety of functions, like esterases, lipases, peptidases, epoxidases, lyases, and others. At the same time, many α/β hydrolase-like families, even enzymatically distinct, share a high degree of sequence similarity. This imposes several problems for their annotation and classification, because available definitions of particular α/β hydrolase-like families overlap significantly, so the unambiguous functional assignment of these superfamily members remains a challenging task. For instance, two large and important peptidase families, namely S9 and S33, blend with lipases, epoxidases, esterases, and other enzymes unrelated to proteolysis, which hinders automatic annotations in high-throughput projects. With the use of thorough sequence and structure analyses, we newly annotate three protein families as α/β hydrolase-like and revise current classifications of the realm of α/β hydrolase-like superfamily. Based on manually curated structural superimpositions and multiple sequence and structure alignments, we comprehensively demonstrate structural conservation and diversity across the whole superfamily. Eventually, after detailed pairwise sequence similarity assessments, we develop a new clustering of the α/β hydrolases and provide a set of family profiles allowing for detailed, reliable, and automatic functional annotations of the superfamily members.