Understanding the energetics of peripheral protein-membrane interactions is important to many areas of biophysical chemistry and cell biology. Estimating free-energy landscapes by molecular dynamics (MD) simulation is challenging for such systems, especially when membrane recognition involves complex lipids, e.g., phosphatidylinositol phosphates (PIPs). We combined coarse-grained MD simulations with umbrella sampling to quantify the binding of the well-explored GRP1 pleckstrin homology (PH) domain to model membranes containing PIP molecules. The experimentally observed preference of GRP1-PH for PIP3 over PIP2 was reproduced. Mutation of a key residue (K273A) within the canonical PIP-binding site significantly reduced the free energy of PIP binding. The presence of a noncanonical PIP-interaction site, observed experimentally in other PH domains but not previously in GRP1-PH, was also revealed. These studies demonstrate how combining coarse-grained simulations and umbrella sampling can unmask the molecular basis of the energetics of interactions between peripheral membrane proteins and complex cellular membranes.

Pleckstrin homology (PH) domains mediate protein-membrane interactions by binding to phosphatidylinositol phosphate (PIP) molecules. The structural and energetic basis of selective PH-PIP interactions is central to understanding many cellular processes, yet the molecular complexities of the PH-PIP interactions are largely unknown. Molecular dynamics simulations using a coarse-grained model enables estimation of free-energy landscapes for the interactions of 12 different PH domains with membranes containing PIP or PIP, allowing us to obtain a detailed molecular energetic understanding of the complexities of the interactions of the PH domains with PIP molecules in membranes. Distinct binding modes, corresponding to different distributions of cationic residues on the PH domain, were observed, involving PIP interactions at either the "canonical" (C) and/or "alternate" (A) sites. PH domains can be grouped by the relative strength of their C- and A-site interactions, revealing that a higher affinity correlates with increased C-site interactions. These simulations demonstrate that simultaneous binding of multiple PIP molecules by PH domains contributes to high-affinity membrane interactions, informing our understanding of membrane recognition by PH domains in vivo.

Association of peripheral proteins with lipid bilayers regulates membrane signaling and dynamics. Pleckstrin homology (PH) domains bind to phosphatidylinositol phosphate (PIP) molecules in membranes. The effects of local PIP enrichment on the interaction of PH domains with membranes is unclear. Molecular dynamics simulations allow estimation of the binding energy of GRP1 PH domain to PIP3-containing membranes. The free energy of interaction of the PH domain with more than two PIP3 molecules is comparable to experimental values, suggesting that PH domain binding involves local clustering of PIP molecules within membranes. We describe a mechanism of PH binding proceeding via an encounter state to two bound states which differ in the orientation of the protein relative to the membrane, these orientations depending on the local PIP concentration. These results suggest that nanoscale clustering of PIP molecules can control the strength and orientation of PH domain interaction in a concentration-dependent manner.

Kindlins co-activate integrins alongside talin. They possess, like talin, a FERM domain (4.1-erythrin–radixin–moiesin domain) comprising F0–F3 subdomains, but with a pleckstrin homology (PH) domain inserted in the F2 subdomain that enables membrane association. We present the crystal structure of murine kindlin-3 PH domain determined at a resolution of 2.23 Å and characterise its lipid binding using biophysical and computational approaches. Molecular dynamics simulations suggest flexibility in the PH domain loops connecting β-strands forming the putative phosphatidylinositol phosphate (PtdInsP)-binding site. Simulations with PtdInsP-containing bilayers reveal that the PH domain associates with PtdInsP molecules mainly via the positively charged surface presented by the β1–β2 loop and that it binds with somewhat higher affinity to PtdIns(3,4,5)P3 compared with PtdIns(4,5)P2. Surface plasmon resonance (SPR) with lipid headgroups immobilised and the PH domain as an analyte indicate affinities of 300 µM for PtdIns(3,4,5)P3 and 1 mM for PtdIns(4,5)P2. In contrast, SPR studies with an immobilised PH domain and lipid nanodiscs as the analyte show affinities of 0.40 µM for PtdIns(3,4,5)P3 and no affinity for PtdIns(4,5)P2 when the inositol phosphate constitutes 5% of the total lipids (∼5 molecules per nanodisc). Reducing the PtdIns(3,4,5)P3 composition to 1% abolishes nanodisc binding to the PH domain, as does site-directed mutagenesis of two lysines within the β1–β2 loop. Binding of PtdIns(3,4,5)P3 by a canonical PH domain, Grp1, is not similarly influenced by SPR experimental design. These data suggest a role for PtdIns(3,4,5)P3 clustering in the binding of some PH domains and not others, highlighting the importance of lipid mobility and clustering for the biophysical assessment of protein–membrane interactions.