Item

Abstract

Trimeric autotransporter adhesins (TAAs) are a family of highly repetitive, homotrimeric surface proteins found in many Gram-negative bacteria. In pathogenic bacteria, TAAs play multiple roles during host infection, like adhesion, biofilm formation, and immune evasion. Structurally, TAAs exhibit a modular architecture, sharing a common anchor domain and possessing a variable number of stalk and head domains. The modularity of TAAs has been used for the bioinformatic and experimental characterization of their constituent domains and full-length homotrimeric complexes. However, many TAAs remain poorly annotated because of their extreme sequence diversity and the challenges underlying their experimental characterization. To produce a comprehensive annotation of TAAs, we identified all TAAs within the RefSeq protein sequence database and predicted the homotrimeric structures of ~9000 representative TAAs using AlphaFold-multimer. As TAAs vary from under 200 amino acids to over 3000, we used their modular architecture for refined predictions based on sequence and structure analysis. We identified many uncharacterized stalk motifs and domains, variations of previously known stalk domains, and many structurally diverse head domains.