Phosphoinositide‐mimicking peptide sequences are binding targets for PH domains

Pleckstrin homology (PH) domains are the 11 th most common domain in the human proteome and are best known as phosphoinositide (PI) targeting modules. However, less than 10% of PH domains recognize PIs with high affinity and specificity raising the question of what is the target of the other 90%. We have identified peptides enriched in acidic or phosphoserine residues as potential protein targets for PH domains. Nearly 40% of all S. cerevisiae PH domains were able to recognize the highly acidic protein Nsr1p in pulldown experiments, making this a more common property than high affinity PI binding. We also showed binding of GST‐fusions of PLC‐δ 1 or yeast Boi2p PH domains to a multiphosphorylated peptide of IRBIT with estimated K D values of 0.3μM and 8.4μM respectively. PH domain mutations that disrupted PI binding also abrogated IRBIT binding. We hypothesize that stretches of acidic residues and/or phosphoserines structurally mimic inositol phosphates/phosphoinositide headgroups, and thus bind to at least a subclass of PH domains. Such ' phosphoinositide mimicry ' could represent a previously undescribed molecular recognition mechanism for driving specific PH domain/protein interactions.