Flexible pivoting of dynamin pleckstrin homology domain catalyzes fission: insights into molecular degrees of freedom

Making the cut: Multiscale simulation of membrane remodeling

Biological membranes are dynamic heterogeneous materials, and their shape and organization are tightly coupled to the properties of the proteins in and around them. However, the length scales of lipid and protein dynamics are far below the size of membrane-bound organelles, much less an entire cell. Therefore, multiscale modeling approaches are often necessary to build a comprehensive picture of the interplay of these factors, and have provided critical insights into our understanding of membrane dynamics. Here, we review computational methods for studying membrane remodeling, as well as passive and active examples of protein-driven membrane remodeling. As the field advances towards the modeling of key aspects of organelles and whole cells - an increasingly accessible regime of study - we summarize here recent successes and offer comments on future trends.

Mast cell secretory granule fusion with amphisomes coordinates their homotypic fusion and release of exosomes

Secretory granule (SG) fusion is an intermediate step in SG biogenesis. However, the precise mechanism of this process is not completely understood. We show that Golgi-derived mast cell (MC) SGs enlarge through a mechanism that is dependent on phosphoinositide (PI) remodeling and fusion with LC3+ late endosomes (amphisomes), which serve as hubs for the fusion of multiple individual SGs. Amphisome formation is regulated by the tyrosine phosphatase PTPN9, while the subsequent SG fusion event is additionally regulated by the tetraspanin protein CD63 and by PI4K. We also demonstrate that fusion with amphisomes imparts to SGs their capacity of regulated release of exosomes. Finally, we show that conversion of PI(3,4,5)P3 to PI(4,5)P2 and the subsequent recruitment of dynamin stimulate SG fission. Our data unveil a key role for lipid-regulated interactions with the endocytic and autophagic systems in controlling the size and number of SGs and their capacity to release exosomes.

Cryo-EM structures of membrane-bound dynamin in a post-hydrolysis state primed for membrane fission

Dynamin assembles as a helical polymer at the neck of budding endocytic vesicles, constricting the underlying membrane as it progresses through the GTPase cycle to sever vesicles from the plasma membrane. Although atomic models of the dynamin helical polymer bound to guanosine triphosphate (GTP) analogs define earlier stages of membrane constriction, there are no atomic models of the assembled state post-GTP hydrolysis. Here, we used cryo-EM methods to determine atomic structures of the dynamin helical polymer assembled on lipid tubules, akin to necks of budding endocytic vesicles, in a guanosine diphosphate (GDP)-bound, super-constricted state. In this state, dynamin is assembled as a 2-start helix with an inner lumen of 3.4 nm, primed for spontaneous fission. Additionally, by cryo-electron tomography, we trapped dynamin helical assemblies within HeLa cells using the GTPase-defective dynamin K44A mutant and observed diverse dynamin helices, demonstrating that dynamin can accommodate a range of assembled complexes in cells that likely precede membrane fission.

Reconstituting membrane fission using a high content and throughput assay

Protein-mediated membrane fission has been analyzed both in bulk and at the single event resolution. Studies on membrane fission in vitro using tethers have provided fundamental insights into the process but are low in throughput. In recent years, supported membrane template (SMrT) have emerged as a facile and convenient assay system for membrane fission. SMrTs provide useful information on intermediates in the pathway to fission and are therefore high in content. They are also high in throughput because numerous fission events can be monitored in a single experiment. This review discusses the utility of SMrTs in providing insights into fission pathways and its adaptation to annotate membrane fission functions in proteins.

Non-affine deformation analysis and 3D packing defects: A new way to probe membrane heterogeneity in molecular simulations

Here, we discuss a new framework developed over the last 5 years in our group to probe nanoscale membrane heterogeneity. The framework is based on the idea of characterizing lateral heterogeneity through non-affine deformation (NAD) measurements, transverse heterogeneity through three dimensional (3D) lipid packing defects, and using these approaches to formalize the seemingly trivial correlation between lateral organization and lipid packing in biological membranes. We find that measurements from NAD analysis, a prescription which is borrowed from Physics of glasses and granular material, can faithfully distinguish between liquid-ordered and disordered phases in membranes at molecular length scales and, can also be used to identify phase boundaries with high precision. Concomitantly, 3D-packing defects can not only distinguish between the two co-existing fluid phases based on their molecular scale packing (or membrane free volume), but also provide a route to connect the membrane domains to their functionality, such as exploring the molecular origins of inter-leaflet domain registration and peptide partitioning. The correlation between lateral membrane order and transverse packing presents novel molecular design-level features that can explain functions such as protein/peptide partitioning and small-molecule permeation dynamics in complex and heterogeneous membranes with high-fidelity. The framework allows us to explore the nature of lateral organization and molecular packing as a manifestation of intricate molecular interactions among a chemically rich variety of lipids and other molecules in a membrane with complex membrane composition and asymmetry across leaflets.

"Gearing" up for dynamin-catalyzed membrane fission

Endocytic dynamins self-assemble into helical scaffolds and utilize energy from GTP hydrolysis to constrict and sever tubular membranous necks of budded endocytic intermediates. They bind the membrane using a pleckstrin-homology domain (PHD). The PHD is characterized by four unstructured loops, two of which partially insert into the membrane. Recent studies reveal that loop insertion lowers the bending rigidity of the membrane and that mutations in these two loops produce separable and opposite effects on the efficiency of dynamin-catalyzed membrane fission. Here, we review the current understanding of dynamin-catalyzed membrane fission and attempt to reconcile contrasting notions that have emerged from biochemical and cellular studies evaluating the role of the PHD in this process. We propose that two membrane-inserting loops act as "gears" that define the catalytic efficiency of the dynamin helical scaffold in membrane fission.

Computational modeling of membrane trafficking processes: From large molecular assemblies to chemical specificity

In the last decade, molecular dynamics (MD) simulations have become an essential tool to investigate the molecular properties of membrane trafficking processes, often in conjunction with experimental approaches. The combination of MD simulations with recent developments in structural biology, such as cryo-electron microscopy and artificial intelligence-based structure determination, opens new, exciting possibilities for future investigations. However, the full potential of MD simulations to provide a molecular view of the complex and dynamic processes involving membrane trafficking can only be realized if certain limitations are addressed, and especially those concerning the quality of coarse-grain models, which, despite recent successes in describing large-scale systems, still suffer from far-from-ideal chemical accuracy. In this review, we will highlight recent success stories of MD simulations in the investigation of membrane trafficking processes, their implications for future research, and the challenges that lie ahead in this specific research domain.

Mechanistic analysis of a novel membrane-interacting variable loop in the pleckstrin-homology domain critical for dynamin function

Classical dynamins are best understood for their ability to generate vesicles by membrane fission. During clathrin-mediated endocytosis (CME), dynamin is recruited to the membrane through multivalent protein and lipid interactions between its proline-rich domain (PRD) with SRC Homology 3 (SH3) domains in endocytic proteins and its pleckstrin-homology domain (PHD) with membrane lipids. Variable loops (VL) in the PHD bind lipids and partially insert into the membrane thereby anchoring the PHD to the membrane. Recent molecular dynamics (MD) simulations reveal a novel VL4 that interacts with the membrane. Importantly, a missense mutation that reduces VL4 hydrophobicity is linked to an autosomal dominant form of Charcot-Marie-Tooth (CMT) neuropathy. We analyzed the orientation and function of the VL4 to mechanistically link data from simulations with the CMT neuropathy. Structural modeling of PHDs in the cryo-electron microscopy (cryo-EM) cryoEM map of the membrane-bound dynamin polymer confirms VL4 as a membrane-interacting loop. In assays that rely solely on lipid-based membrane recruitment, VL4 mutants with reduced hydrophobicity showed an acute membrane curvature-dependent binding and a catalytic defect in fission. Remarkably, in assays that mimic a physiological multivalent lipid- and protein-based recruitment, VL4 mutants were completely defective in fission across a range of membrane curvatures. Importantly, expression of these mutants in cells inhibited CME, consistent with the autosomal dominant phenotype associated with the CMT neuropathy. Together, our results emphasize the significance of finely tuned lipid and protein interactions for efficient dynamin function.

Cofilin-Membrane Interactions: Electrostatic Effects in Phosphoinositide Lipid Binding

The actin cytoskeleton interacts with the cell membrane primarily through the indirect interactions of actin-binding proteins such as cofilin-1. The molecular mechanisms underlying the specific interactions of cofilin-1 with membrane lipids are still unclear. Here, we performed coarse-grain molecular dynamics simulations of cofilin-1 with complex lipid bilayers to analyze the specificity of protein-lipid interactions. We observed the maximal interactions with phosphoinositide (PIP) lipids, especially PIP₂ and PIP₃ lipids. A good match was observed between the residues predicted to interact and previous experimental studies. The clustering of PIP lipids around the membrane bound protein leads to an overall lipid demixing and gives rise to persistent membrane curvature. Further, through a series of control simulations, we observe that both electrostatics and geometry are critical for specificity of lipid binding. Our current study is a step towards understanding the physico-chemical basis of cofilin-PIP lipid interactions.

A Review of Mechanics-Based Mesoscopic Membrane Remodeling Methods: Capturing Both the Physics and the Chemical Diversity

Specialized classes of proteins, working together in a tightly orchestrated manner, induce and maintain highly curved cellular and organelles membrane morphology. Due to the various experimental constraints, including the resolution limits of imaging techniques, it is non-trivial to accurately elucidate interactions among the various components involved in membrane deformation. The spatial and temporal scales of the systems also make it formidable to investigate them using simulations with molecular details. Interestingly, mechanics-based mesoscopic models have been used with great success in recapitulating the membrane deformations observed in experiments. In this review, we collate together and discuss the various mechanics-based mesoscopic models for protein-mediated membrane deformation studies. In particular, we provide an elaborate description of a mesoscopic model where the membrane is modeled as a triangulated sheet and proteins are represented as either nematics or filaments. This representation allows us to explore the various aspects of protein-protein and protein-membrane interactions as well as examine the underlying mechanistic pathways for emergent behavior such as curvature-mediated protein localization and membrane deformation. We also put forward current efforts in the field towards back-mapping these mesoscopic models to finer-grained particle-based models-a framework that could be used to explore how molecular interactions propagate to physical scales and vice-versa. We end the review with an integrative-modeling-based road map where experimental imaging micrograph and biochemical data are combined with mesoscopic and molecular simulations methods in a theoretically consistent manner to faithfully recapitulate the multiple length and time scales in the membrane remodeling processes.

Membrane Remodeling Due to a Mixture of Multiple Types of Curvature Proteins

We present an extension of the Monte Carlo based mesoscopic membrane model, where the membrane is represented as a dynamically triangulated surface and the proteins are modeled as anisotropic inclusions formulated as in-plane nematic field variables adhering to the deformable elastic sheet. In the extended model, we have augmented the Hamiltonian to study membrane deformation due to a mixture of multiple types of curvature generating proteins. This feature opens the door for understanding how multiple kinds of curvature-generating proteins may be working in a coordinated manner to induce desired membrane morphologies. For example, among other things, we study membrane deformations and tubulation due to a mixture of positive and negative curvature proteins as mimics of various proteins from BAR domain family. We also study the effect of membrane anisotropy that manifests as differential binding affinity and organization of curvature proteins, leading to insights into the tightly regulated cargo sorting and transport processes. Our simulation results show different morphology of deformed vesicles that depend on membrane tension, the curvatures and number of the participating proteins as well as on protein-protein and membrane-protein interactions.

Systematic simulation of the interactions of pleckstrin homology domains with membranes

Pleckstrin homology (PH) domains can recruit proteins to membranes by recognition of phosphatidylinositol phosphate (PIP) lipids. Several family members are linked to diseases including cancer. We report the systematic simulation of the interactions of 100 mammalian PH domains with PIP-containing membranes. The observed PIP interaction hotspots recapitulate crystallographic binding sites and reveal a number of insights: (i) The β1 and β2 strands and their connecting loop constitute the primary PIP interaction site but are typically supplemented by interactions at the β3-β4 and β5-β6 loops; (ii) we reveal exceptional cases such as the Exoc8 PH domain; (iii) PH domains adopt different membrane-bound orientations and induce clustering of anionic lipids; and (iv) beyond family-level insights, our dataset sheds new light on individual PH domains, e.g., by providing molecular detail of secondary PIP binding sites. This work provides a global view of PH domain/membrane association involving multivalent association with anionic lipids.

Simulations of Kindlin-2 PIP binding domains reveal protonation-dependent membrane binding modes

Kindlin-2, a member of the Kindlin family of peripheral membrane proteins, is important for integrin activation and stabilization of epidermal growth factor receptor. It associates with the cytoplasmic face of the plasma membrane via dedicated phosphatidylinositol phosphate binding domains located in the N-terminal F0 and Pleckstrin Homology domains. These domains have binding affinity for phosphatidylinositol 4,5-bisphosphate and, to a greater degree, phosphatidylinositol 3,4,5-trisphosphate. The biological significance of the differential binding of these phosphatidylinositol phosphates to Kindlin-2 and the mechanism by which they activate Kindlin-2 are not well understood. Recently, ssNMR identified the predominant protonation states of phosphatidylinositol 4,5-bisphosphate and phosphatidylinositol 3,4,5-trisphosphate near physiological pH in the presence of anionic lipids. Here, we perform atomistic simulation of the bound state of the Pleckstrin Homology and F0 domains of Kindlin-2 at membranes containing phosphatidylinositol 4,5-bisphosphate/phosphatidylinositol 3,4,5-trisphosphate with differing protonation states. This computational approach demonstrates that these two phosphatidylinositol phosphates differently modulate Kindlin-2 subdomain binding in a protonation-state-dependent manner. We speculate these variations in binding mode provide a mechanism for intracellular pH and Ca²⁺ influx to control the membrane binding behavior and activity of Kindlin-2.