Lipid modulation of protein-induced membrane domains as a mechanism for controlling signal transduction

The reason for the enormous lipid variety present in eukaryotic membranes remains largely an enigma. We suggest that its role is to provide an on-off switch for a signaling event at the membrane level. This is achieved through lipid-lipid interactions that convert membrane protein binding and association events into very cooperative processes while maintaining reversibility. We have previously shown [Hinderliter, A., at al. (2001) Biochemistry 40, 4181-4191] that thermodynamic linkage between an intrinsic tendency for lipid demixing and a preferential interaction of a protein with a specific lipid within the mixture leads to dramatic changes in lipid and protein domain formation. Here, we tested the hypothesis that small alterations in lipid chemical structure alter the magnitude of the net interaction free energy (omega(AB)) between unlike lipids in a predictable manner, and that even very small changes in omega(AB) lead to dramatic changes in bilayer organization when coupled with protein binding. We systematically varied the chemical structure of phosphatidylcholine (PC), in mixtures with a fixed phosphatidylserine (PS), by changing the PC acyl chain length and the degree of unsaturation, and examined domain formation upon addition of a peripheral protein, the synaptotagmin I C2A motif. Experimental excimer/monomer ratios (E/M) of pyrene-substituted lipids mimicking the PS were interpreted using Monte Carlo computer simulations. E/M is larger if the PC melting temperature is lower, suggesting that domain formation is a thermodynamic consequence of weak interactions between PC and PS. Consistent with our hypothesis, only very small changes in omega(AB) were required for prediction of large changes in lipid and protein domain formation.     

Protein-promoted membrane domains

The current notion of biological membranes encompasses a very complex structure, made of dynamically changing compartments or domains where different membrane components partition. These domains have been related to important cellular functions such as membrane sorting, signal transduction, membrane fusion, neuronal maturation, and protein activation. Many reviews have dealt with membrane domains where lipid-lipid interactions direct their formation, especially in the case of raft domains, so in this review we considered domains induced by integral membrane proteins. The nature of the interactions involved and the different mechanisms through which membrane proteins segregate lipid domains are presented, in particular with regard to those induced by the nAChR. It may be concluded that coupling of favourable lipid-lipid and lipid-protein interactions is a general condition for this phenomenon to occur.     

Lipid enrichment and selectivity of integral membrane proteins in two-component lipid bilayers

A model recently used to study lipid-protein interactions in one-component lipid bilayers (Sperotto and Mouritsen, 1991 a, b) has been extended in order to include two different lipid species characterized by different acyl-chain lengths. The model, which is a statistical mechanical lattice model, assumes that hydrophobic matching between lipid-bilayer hydrophobic thickness and hydrophobic length of the integral protein is an important aspect of the interactions. By means of Monte Carlo simulation techniques, the lateral distribution of the two lipid species near the hydrophobic protein-lipid interface in the fluid phase of the bilayer has been derived. The results indicate that there is a very structured and heterogeneous distribution of the two lipid species near the protein and that the protein-lipid interface is enriched in one of the lipid species. Out of equilibrium, the concentration profiles of the two lipid species away from the protein interface are found to develop a long-range oscillatory behavior. Such dynamic membrane heterogeneity may be of relevance for determining the physical factors involved in lipid specificity of protein function.     

A Monte Carlo simulation study of protein-induced heat capacity changes and lipid-induced protein clustering.

Monte Carlo simulations were used to describe the interaction of peripheral and integral proteins with lipids in terms of heat capacity profiles and protein distribution. The simulations were based on a two-state model for the lipid, representing the lipid state as being either gel or fluid. The interaction between neighboring lipids has been taken into account through an unlike nearest neighbor free energy term delta omega, which is a measure of the cooperativity of the lipid transition. Lipid/protein interaction was considered using the experimental observation that the transition midpoints of lipid membranes are shifted upon protein binding, a thermodynamic consequence of different binding constants of protein with fluid or gel lipids. The difference of the binding free energies was used as an additional parameter to describe lipid-protein interaction. The heat capacity profiles of lipid/protein complexes could be well described for both peripheral and integral proteins. Binding of proteins results in a shift and an asymmetric broadening of the melting profile. The model results in a coexistence of gel and fluid lipid domains in the proximity of the thermotropic transition. As a consequence, bound peripheral proteins aggregate in the temperature range of the lipid transition. Integral proteins induce calorimetric melting curves that are qualitatively different from that of peripheral proteins and aggregate in either gel or liquid crystalline lipid phase. The results presented here are in good agreement with calorimetric experiments on lipid-protein complexes and have implementations for the functional control of proteins.     

Monte Carlo simulation of protein-induced lipid demixing in a membrane with interactions derived from experiment

Lipid domain formation induced by annexin was investigated in mixtures of phosphatidylcholine (PC), phosphatidylserine (PS), and cholesterol (Chol), which were selected to mimic the inner leaflet of a eukaryotic plasma membrane. Annexins are ubiquitous and abundant cytoplasmic, peripheral proteins, which bind to membranes containing PS in the presence of calcium ions (Ca²⁺), but whose function is unknown. Prompted by indications of interplay between the presence of cholesterol in PS/PC mixtures and the binding of annexins, we used Monte Carlo simulations to investigate protein and lipid domain formation in these mixtures. The set of interaction parameters between lipids and proteins was assigned by matching experimental observables to corresponding variables in the calculations. In the case of monounsaturated phospholipids, the PS-PC and PC-Chol interactions are weakly repulsive. The interaction between protein and PS was determined based on experiments of annexin binding to PC/PS mixtures in the presence of Ca²⁺. Based on the proposal that PS and cholesterol form a complex in model membranes, a favorable PS-Chol interaction was postulated. Finally, protein-protein favorable interactions were also included, which are consistent with observations of large, two-dimensional, regular arrays of annexins on membranes. Those net interactions between pairs of lipids, proteins and lipids, and between proteins are all small, of the order of the average kinetic energy. We found that annexin a5 can induce formation of large PS domains, coincident with protein domains, but only if cholesterol is present.     

NMR-Based Simulation Studies of Pf1 Coat Protein in Explicit Membranes

As time- and ensemble-averaged measures, NMR observables contain information about both protein structure and dynamics. This work represents a computational study to extract such information for membrane proteins from orientation-dependent NMR observables: solid-state NMR chemical shift anisotropy and dipolar coupling, and solution NMR residual dipolar coupling. We have performed NMR-restrained molecular dynamics simulations to refine the structure of the membrane-bound form of Pf1 coat protein in explicit lipid bilayers using the recently measured chemical shift anisotropy, dipolar coupling, and residual dipolar coupling data. From the simulations, we have characterized detailed protein-lipid interactions and explored the dynamics. All simulations are stable and the NMR restraints are well satisfied. The C-terminal transmembrane (TM) domain of Pf1 finds its optimal position in the membrane quickly (within 6 ns), illustrating efficient solvation of TM domains in explicit bilayer environments. Such rapid convergence also leads to well-converged interaction patterns between the TM helix and the membrane, which clearly show the interactions of interfacial membrane-anchoring residues with the lipids. For the N-terminal periplasmic helix of Pf1, we identify a stable, albeit dynamic, helix orientation parallel to the membrane surface that satisfies the amphiphatic nature of the helix in an explicit lipid bilayer. Such detailed information cannot be obtained solely from NMR observables. Therefore, the present simulations illustrate the usefulness of NMR-restrained MD refinement of membrane protein structure in explicit membranes.     

NMR-Based Simulation Studies of Pf1 Coat Protein in Explicit Membranes

As time- and ensemble-averaged measures, NMR observables contain information about both protein structure and dynamics. This work represents a computational study to extract such information for membrane proteins from orientation-dependent NMR observables: solid-state NMR chemical shift anisotropy and dipolar coupling, and solution NMR residual dipolar coupling. We have performed NMR-restrained molecular dynamics simulations to refine the structure of the membrane-bound form of Pf1 coat protein in explicit lipid bilayers using the recently measured chemical shift anisotropy, dipolar coupling, and residual dipolar coupling data. From the simulations, we have characterized detailed protein-lipid interactions and explored the dynamics. All simulations are stable and the NMR restraints are well satisfied. The C-terminal transmembrane (TM) domain of Pf1 finds its optimal position in the membrane quickly (within 6 ns), illustrating efficient solvation of TM domains in explicit bilayer environments. Such rapid convergence also leads to well-converged interaction patterns between the TM helix and the membrane, which clearly show the interactions of interfacial membrane-anchoring residues with the lipids. For the N-terminal periplasmic helix of Pf1, we identify a stable, albeit dynamic, helix orientation parallel to the membrane surface that satisfies the amphiphatic nature of the helix in an explicit lipid bilayer. Such detailed information cannot be obtained solely from NMR observables. Therefore, the present simulations illustrate the usefulness of NMR-restrained MD refinement of membrane protein structure in explicit membranes.     

Quantitative characterization of the lateral distribution of membrane proteins within the lipid bilayer.

The dependence of the lateral distribution of membrane proteins on the size, protein/lipoid molar ratio, and the magnitude of the interaction potentials has been investigated by computer modeling protein-lipid distributions with Monte Carlo calculations. These results have allowed us to develop a quantitative characterization of the distribution of membrane proteins and to correlate these distributions with experimental observables. The topological arrangement of protein domains, protein plus annular lipid domains, and free lipid domains is described in terms of radial distribution, pair connectedness, and cluster distribution functions. The radial distribution functions are used to measure the distribution of intermolecular distances between protein molecules, whereas the pair connectedness functions are used to estimate the physical extension of compositional domains. It is shown that, at characteristic protein/lipid molar ratios, previously isolated domains become connected, forming domain networks that extend over the entire membrane surface. These changes in the lateral connectivity of compositional domains are paralleled by changes in the calculated lateral diffusion coefficients and might have important implications for the regulation of diffusion controlled processes within the membrane.     

A theoretical study of lipid-protein interactions in bilayers.

We present a theoretical study of the effect of different types of lipid-protein interactions on the thermodynamic properties of protein-containing lipid bilayers. The basis of this work is a theoretical model for pure lipid bilayer phase transitions developed earlier by Scott. Simple assumptions on the nature of the lipid conformations near a protein strongly affect the predicted properties of the model. Here we consider (a) random protein-lipid contacts, (b) enhanced contact between protein and lipid with a number of gauche bonds, and (c) enhanced contact between protein and all-trans but tilted lipid chains. Comparison of predicted results with experimental data seems to favor point c above but, by itself point c does not work well at larger protein concentrations. The results are discussed in the light of spectroscopic data, lipid-protein (plus annular lipid) miscibility, and interprotein forces.     

Finding a Needle in a Haystack: The Role of Electrostatics in Target Lipid Recognition by PH Domains

Interactions between protein domains and lipid molecules play key roles in controlling cell membrane signalling and trafficking. The pleckstrin homology (PH) domain is one of the most widespread, binding specifically to phosphatidylinositol phosphates (PIPs) in cell membranes. PH domains must locate specific PIPs in the presence of a background of approximately 20% anionic lipids within the cytoplasmic leaflet of the plasma membrane. We investigate the mechanism of such recognition via a multiscale procedure combining Brownian dynamics (BD) and molecular dynamics (MD) simulations of the GRP1 PH domain interacting with phosphatidylinositol (3,4,5)-trisphosphate (PI(3,4,5)P(3)). The interaction of GRP1-PH with PI(3,4,5)P(3) in a zwitterionic bilayer is compared with the interaction in bilayers containing different levels of anionic 'decoy' lipids. BD simulations reveal both translational and orientational electrostatic steering of the PH domain towards the PI(3,4,5)P(3)-containing anionic bilayer surface. There is a payoff between non-PIP anionic lipids attracting the PH domain to the bilayer surface in a favourable orientation and their role as 'decoys', disrupting the interaction of GRP1-PH with the PI(3,4,5)P(3) molecule. Significantly, approximately 20% anionic lipid in the cytoplasmic leaflet of the bilayer is nearly optimal to both enhance orientational steering and to localise GRP1-PH proximal to the surface of the membrane without sacrificing its ability to locate PI(3,4,5)P(3) within the bilayer plane. Subsequent MD simulations reveal binding to PI(3,4,5)P(3), forming protein-phosphate contacts comparable to those in X-ray structures. These studies demonstrate a computational framework which addresses lipid recognition within a cell membrane environment, offering a link between structural and cell biological characterisation.     

Targeting membrane proteins to liquid-ordered phases: molecular self-organization explored by fluorescence correlation spectroscopy

The complex and dynamic architecture of biological membranes comprises of various heterogeneities, some of which may include lipid-based and/or protein-based microdomains called "rafts". Due to interactions among membrane components, several types of domains can form with different characteristics and mechanisms of formation. Model membranes, such as giant unilamellar vesicles (GUVs), provide a key system to study lipid-lipid and lipid-protein interactions, which are potentially relevant to raft formation, by (single-molecule) optical microscopy. Here, we review studies of combined confocal imaging and fluorescence correlation spectroscopy (FCS) on lipid dynamics and organization in domains assembled in GUVs, prepared from various lipid mixtures, which are relevant to the problem of raft formation. Finally, we summarize the results on lipid-protein interactions, which govern the targeting of several putative raft- and non-raft-associated membrane proteins to domain-exhibiting GUVs.     

Lateral diffusion of gramicidin S, M-13 coat protein and glycophorin in bilayers of saturated phospholipids. Mean field and Monte Carlo studies

We have developed a general model that relates the lateral diffusion coefficient of one isolated large intrinsic molecule (mol. wt. greater than or approximately 1000) in a phosphatidylcholine bilayer to the static lipid hydrocarbon chain order. We have studied how protein lateral diffusion can depend upon protein-lipid interactions but have not investigated possible non-specific contributions from gel-state lattice defects. The model has been used in Monte Carlo simulations or in mean-field approximations to study the lateral diffusion coefficients of Gramicidin S, the M-13 coat protein and glycophorin in dimyristoyl- and dipalmitoylphosphatidylcholine (DMPC and DPPC) bilayers as functions of temperature. Our calculated lateral diffusion coefficients for Gramicidin S and the M-13 coat protein are in good agreement with what has been observed and suggest that Gramicidin S is in a dimeric form in DMPC bilayers. In the case of glycophorin we find that the 'ice breaker' effect can be understood as a consequence of perturbation of the lipid polar region around the protein. In order to understand this effect is necessary that the protein hydrophilic section perturb the polar regions of at least approx. 24 lipid molecules, in good agreement with the numbers of 29-30 measured using 31P-NMR. Because of lipid-lipid interactions this effect extends itself out to four or five lipid layers away from the protein so that the hydrocarbon chains of between approx. 74 and approx. 108 lipid molecules are more disordered in the gel phase, so contributing less to the transition enthalpy, in agreement with the numbers of 80-100 deduced from differential scanning calorimetry (DSC). An understanding of the abrupt change in the diffusion coefficient at a temperature below the main bilayer transition temperature requires an additional mechanism. We propose that this change may be a consequence of a 'coupling-uncoupling' transition involving the protein hydrophilic section and the lipid polar regions, which may be triggered by the lipid bilayer pretransition. Our calculation of the average number of gauche bonds per lipid chain as a function of temperature and distance away from an isolated polypeptide or integral protein shows the extent of statically disordered lipid around such molecules. The range of this disorder depends upon temperature, particularly near the main transition.     

Synaptotagmin perturbs the structure of phospholipid bilayers

Synaptotagmin I (syt), an integral protein of the synaptic vesicle membrane, is believed to act as a Ca2+ sensor for neuronal exocytosis. Syt's cytoplasmic domain consists largely of two C2 domains, C2A and C2B. In response to Ca2+ binding, the C2 domains interact with membranes, becoming partially embedded in the lipid bilayer. We have imaged syt C2AB in association with lipid bilayers under fluid, using AFM. As expected, binding of C2AB to bilayers required both an anionic phospholipid [phosphatidylserine (PS)] and Ca2+. C2AB associated with bilayers in the form of aggregates of varying stoichiometries, and aggregate size increased with an increase in PS content. Repeated scanning of bilayers revealed that as C2AB dissociated it left behind residual indentations in the bilayer. The mean depth of these identations was 1.81 nm, indicating that they did not span the bilayer. Individual C2 domains (C2A and C2B) also formed aggregates and produced bilayer indentations. Binding of C2AB to bilayers and the formation of indentations were significantly compromised by mutations that interfere with binding of Ca2+ to syt or reduce the positive charge on the surface of C2B. We propose that bilayer perturbation by syt might be significant with respect to its ability to promote membrane fusion.     

Lateral diffusion of gramicidin S,M-13 coat protein and glycophorin in bilayers of saturated phospholipids: Mean field and Monte Carlo studies

We have developed a general model that relates the lateral diffusion coefficient of one isolated large intrinsic molecule (mol. wt. ?1000) in a phosphatidylcholine bilayer to the static lipid hydrocarbon chain order. We have studied how protein lateral diffusion can depend upon protein-lipid interactions but have not investigated possible non-specific contributions from gel-state lattice defects. The model has been used in Monte Carlo simulations or in mean-field approximations to study the lateral diffusion coefficients of Gramicidin S, the M-13 coat protein and glycophorin in dimyristoyl- and dipalmitoylphosphatidylcholine (DMPC and DPPC) bilayers as functions of temperature. Our calculated lateral diffusion coefficients for Gramicidin S and the M-13 coat protein are in good agreement with what has been observed and suggest that Gramicidin S is in a dimeric form in DMPC bilayers. In the case of glycophorin we find that the ‘ice breaker’ effect can be understood as a consequence of perturbation of the lipid polar region around the protein. In order to understand this effect is is necessary that the protein hydrophilic section perturb the polar regions of at least approx. 24 lipid molecules, in good agreement with the numbers of 29–30 measured using ³¹P-NMR. Because of lipid-lipid interactions this effect extends itself out to four or five lipid layers away from the protein so that the hydrocarbon chains of between approx. 74 and approx. 108 lipid molecules are more disordered in the gel phase, so contributing less to the transition enthalpy, in agreement with the numbers of 80–100 deduced from differential scanning calorimetry (DSC). An understanding of the abrupt change in the diffusion coefficient at a temperature below the main bilayer transition temperature requires an additional mechanism. We propose that this change may be a consequence of a ‘coupling-uncoupling’ transition involving the protein hydrophilic section and the lipid polar regions, which may be triggered by the lipid bilayer pretransition. Our calculation of the average number of gauche bonds per lipid chain as a function of temperature and distance away from an isolated polypeptide or integral protein shows the extent of statically disordered lipid around such molecules. The range of this disorder depends upon temperature, particularly near the main transition.     

Nonideal mixing of phosphatidylserine and phosphatidylcholine in the fluid lamellar phase.

The mixing of phosphatidylserine (PS) and phosphatidylcholine (PC) in fluid bilayer model membranes was studied by measuring binding of aqueous Ca2+ ions. The measured [Ca2+]aq was used to derive the activity coefficient for PS, gamma PS, in the lipid mixture. For (16:0, 18:1) PS in binary mixtures with either (16:0, 18:1)PC, (14:1, 14:1)PC, or (18:1, 18:1)PC, gamma PS > 1; i.e., mixing is nonideal, with PS and PC clustered rather than randomly distributed, despite the electrostatic repulsion between PS headgroups. To understand better this mixing behavior, Monte Carlo simulations of the PS/PC distributions were performed, using Kawasaki relaxation. The excess energy was divided into an electrostatic term Uel and one adjustable term including all other nonideal energy contributions, delta Em. Uel was calculated using a discrete charge theory. Kirkwood's coupling parameter method was used to calculate the excess free energy of mixing, delta GEmix, hence In gamma PS,calc. The values of In gamma PS,calc were equalized by adjusting delta Em in order to find the simulated PS/PC distribution that corresponded to the experimental results. We were thus able to compare the smeared charge calculation of [Ca2+]surf with a calculation ("masked evaluation method") that recognized clustering of the negatively charged PS: clustering was found to have a modest effect on [Ca2+]surf, relative to the smeared charge model. Even though both PS and PC tend to cluster, the long-range nature of the electrostatic repulsion reduces the extent of PS clustering at low PS mole fraction compared to PC clustering at an equivalent low PC mole fraction.     

Multiscale modeling of lipids and lipid bilayers

A multiscale modeling approach is applied for simulations of lipids and lipid assemblies on mesoscale. First, molecular dynamics simulation of initially disordered system of lipid molecules in water within all-atomic model was carried out. On the next stage, structural data obtained from the molecular dynamics (MD) simulation were used to build a coarse-grained (ten sites) lipid model, with effective interaction potentials computed by the inverse Monte Carlo method. Finally, several simulations of the coarse-grained model on longer length- and time-scale were performed, both within Monte Carlo and molecular dynamics simulations: a periodical sample of lipid molecules ordered in bilayer, a free sheet of such bilayer without periodic boundary conditions, formation of vesicle from a plain membrane, process of self-assembly of lipids randomly dispersed in volume. It was shown that the coarse-grained model, developed exclusively from all-atomic simulation data, reproduces well all the basic features of lipids in water solution.     

Real-time imaging of lipid domains and distinct coexisting membrane protein clusters

A detailed understanding of biomembrane architecture is still a challenging task. Many in vitro studies have shown lipid domains but much less information is known about the lateral organization of membrane proteins because their hydrophobic nature limits the use of many experimental methods. We examined lipid domain formation in biomimetic Escherichia coli membranes composed of phosphatidylethanolamine and phosphatidylglycerol in the absence and presence of 1% and 5% (mol/mol) membrane multidrug resistance protein, EmrE. Monolayer isotherms demonstrated protein insertion into the lipid monolayer. Subsequently, Brewster angle microscopy was applied to image domains in lipid matrices and lipid-protein mixtures. The images showed a concentration dependent impact of the protein on lipid domain size and shape and more interestingly distinct coexisting protein clusters. Whereas lipid domains varied in size (14-47μm), protein clusters exhibited a narrow size distribution (2.6-4.8μm) suggesting a non-random process of cluster formation. A 3-D display clearly indicates that these proteins clusters protrude from the membrane plane. These data demonstrate distinct co-existing lipid domains and membrane protein clusters as the monofilm is being compressed and illustrate the significant mutual impact of lipid-protein interactions on lateral membrane architecture.