Laurie Graham

Bacterial adhesins attach their hosts to surfaces that the bacteria will colonize. This surface adhesion occurs through specific ligand‐binding domains located towards the distal end of the long adhesin molecules. However, recognizing which of the many adhesin domains are structural and which are ligand binding has been difficult up to now. Here we have used the protein structure modeling program AlphaFold2 to predict structures for these giant 0.2‐ to 1.5‐megadalton proteins. Crystal structures previously solved for several adhesin regions are in good agreement with the models. Whereas most adhesin domains are linked in a linear fashion through their N‐ and C‐terminal ends, ligand‐binding domains can be recognized by budding out from a companion core domain so that their ligand‐binding sites are projected away from the axis of the adhesin for maximal exposure to their targets. These companion domains are “split” in their continuity by projecting the ligand‐binding domain outwards. The “split domains” are mostly β‐sandwich extender modules, but other domains like a β‐solenoid can serve the same function. Bioinformatic analyses of Gram‐negative bacterial sequences revealed wide variety ligand‐binding domains are used in their Repeats‐in‐Toxin adhesins. The ligands for many of these domains have yet to be identified but known ligands include various cell‐surface glycans, proteins, and even ice. Recognizing the ligands to which the adhesins bind could lead to ways of blocking colonization by bacterial pathogens. Engineering different ligand‐binding domains into an adhesin has the potential to change the surfaces to which bacteria bind.