Lunch-to-lunch LINXS Workshop in Lund
Recent years have seen strong research efforts on the lipid component of biological membranes. While many studies have been focused on the membrane structure, the dynamics of such systems are crucial for the function of the membrane including membrane bound proteins. With this Dynamics workshop, our goal is to bring together leading experts in the field of membrane dynamics, with a particular focus on neutrons and X-rays, but also complementary techniques including NMR, single molecule spectroscopy and computer modeling.
Prof. Erik Lindahl, Stockholm University, Sweden
Prof. Dr. Claudia Steinem, University of Goettingen, Germany
Ass. Prof. Marta Bally, Umeå University, Sweden
Prof. Dr. Motumu Tanaka, Heidelberg University, Germany and Kyoto University, Japan
Dr. John Katsaras (Oak Ridge National Laboratory, TN, USA
Prof. Tim Salditt, University of Göttingen, Germany
Dr. Olaf Holderer, Julich Center for Neutron Scattering at MLZ, Garching, Germany
Prof. Mei Hong, MIT, Boston, MA, USA
Prof. John Seddon, Imperial College, London, UK
Dr. Michihiro Nagao, NIST Center for neutron research, Gaithersburg MD, USA
Dr. Victoria Garcia Sakai, ISIS, STFC, Rutherford Appleton Laboratory, Harwell, Oxfordshire, UK
Dr. Lukasz Cwiklik, Heyrovsky Institute of Physical Chemistry, Czech Academy of Sciences, Prague, Czech Republic
Dr. Lionel Porcar, Institut Laue-Langevin, Grenoble, France
Mounting evidence suggests that the genetic disorders/mutation and diseases change not only the protein expression patterns but also membranes themselves. In my talk, I will show you some examples how such biological cues influence the dynamic properties of membranes. For this purpose we employ the combination of experimental techniques in real and reciprocal spaces, including grazing incidence X-ray scattering and fluorescence, off-specular neutron scattering, flicker spectroscopy, etc.
Native membrane derived polymer-supported lipid bilayers (nSLBs) are poised to bridge the gap between live cell experiments and traditional model membrane architectures that by offering a combination of accessibility by surface sensitive analytical instrumentation and a composition which more closely resembles cellular membranes by displaying a diversity of endogenous membrane proteins, lipids, and carbohydrates. With the increasing popularity of nSLB systems there is a growing need for a standardized workflow to accurately characterize their quality, composition, and structure to aid in their development and expand their utility.
We recently developed a generic method for producing polymer-supported lipid bilayers directly from cell-derived native membrane vesicles (NMVs). Due to the lack of detergent solubilization and reconstitution steps, the nSLBs created using this approach contain essentially all of the native lipids, as well as the membrane-associated proteins and carbohydrates from the donor membrane. This new approach has been shown to preserve mobility and enzymatic activity of transmembrane proteins in the resulting nSLB. As cell membranes are both dynamic and compositionally complex, replicating these aspects in a model membrane are essential.
A combination of fluorescence microscopy, neutron reflectometry, and time-of-flight secondary-ion mass spectrometry data will be presented which characterizes the structure and composition of this new supported lipid bilayer category. The methodology presented allows the amount of native membrane material in the nSLB to be precisely controlled and display a uniform lateral distribution. Insights into the nSLBs z-dimensional structure are also discussed. The methodology presented is meant to guide future researchers in producing nSLBs from their cellular membrane of choice, as well as how to investigate their quality and composition.
In atherosclerosis lipids and fibrous elements accumulate in the blood vessels forming plaques that eventually can lead to myocardial infarction or stroke. HDL and LDL particles have been shown to play a role in the development and the progression of the plaque build-up and are currently used as biological markers in addition to measurements of total lipid and cholesterol. Yet many people still develop the disease even when their blood lipid values fall within the healthy range. The function and the lipid exchange of the HDL and LDL particles seems to be more important than the actual HDL and LDL levels. Understanding the dynamic/structure relationship of different lipoproteins and especially the mode of action with which they release or accept their lipid cargo is therefore a prerequisite for the development of better standards and methods for diagnostics of atherosclerosis to aid in the development of targeted therapies in the fight against CVD. Here, small-angle neutron scattering in combination with selective deuteration was used to follow the molecular lipid exchange between native lipoprotein particles and complex cell-membrane mimics. Focusing on the lipid transport kinetics between both native HDL and LDL and large unilamellar vesicles made of “neutron invisible” natural, monounsaturated, phosphocholine mixtures we show that the the lipid exchange is assisted by collision and particle tethering to the membrane. The data also show that the two particles exhibit different kinetic regimes, suggesting that the apolipoprotein plays a key role in enhancing lipid exchange. The method developed here allows for molecular exchange events between complex biological protein-lipid systems to be followed in an elegant, systematic and controlled fashion. For the case of lipoproteins and their function in atherosclerosis, this approach can be used to provide unique information on the role that each lipid species exerts on lipid uptake and exchange, as well as the mode of action of specific apolipoprotein types and subtypes.
Virus entry is a complex dynamic multistep process requiring a series of fine-tuned events mediating virus diffusion through the glycocalyx, its attachment to the cell membrane and lateral diffusion to the point of entry. This is followed by entry, involving membrane fusion or membrane deformation into an endocytic vesicle.
The aim of our research is elucidate the mechanisms by which viral pathogens interact with the cell’membrane to crosses it and penetrate into the cell. We focus on the interplay between the membrane’s physico-chemical properties and the virus attachment process and study how cellular and viral molecules act in concert to modulate the processes through multivalency. Our research strategy widely relies of the use of artificial lipid bilayers to mimic in vitro the basic molecular architecture of the cell membrane. We further, take advantage of total internal fluorescence microscopy and single particle tracking to study, on a single particle level, virus attachment to and diffusion on the cell surface.
To illustrate the potential of such a biophysical approach, I will, in my presentation, first focus on the interaction between norovirus and glycolipid-containing membranes and investigate the role of ligands mobility and ligand clustering in modulating the affinity of the virus particle to the membrane. In a second example, I concentrate on the role of Influenza’s matrix protein in virus budding and search for mechanisms by which the protein can induce membrane deformations.  In a last example, we use model membranes carrying glycosaminoglycans, to elucidate the molecular mechanisms modulating attachment and release of the herpes simplex virus. [3, 4]
Taken together, these examples illustrate the potential of artificial cell membrane mimics in the study of processes occurring at the surface of a cell and demonstrate how such biophysical data can complement more classical cell-biology experiments.
Lipid membranes, which are the most biologically relevant lipid aggregates, have a bilayer structure. Still, in a human body, there are two important lipid assemblies, Lung Surfactant (LS) and Tear Film Lipid Layer (TFLL), which have not bi- but rather a multilayer character. Lung surfactant lines the gas-exchange interface in the lung alveoli and reduces the surface tension at the air-liquid boundary to minimize the work of breathing. It consists mainly of phospholipids with a small number of proteins. At the very interface, it forms a lipid monolayer connected to complex multilayer lipid reservoirs in the subphase. Tear Film Lipid Layer is a multilayer lipid assembly covering the aqueous tear film at the cornea surface. It is a highly dynamic and non-equilibrium structure forming the very first environment-eye barrier. It contains polar and nonpolar lipids, forming a complex and dynamic interface.
We investigate models of Lung Surfactant and TFLL employing combined theoretical and experimental approaches. We use molecular dynamics computer simulations to obtain a molecular-level picture of the systems. The simulations are complemented by experiments using Langmuir trough combined with fluorescent microscopy to address macroscopic-level phenomena. In our studies, special emphasis is given to interactions of LS and TFLL with topical drug molecules.
Amphotericin B (AmB) is a naturally occurring antibiotic with a broad spectrum against systemic fungal and parasitic infections. AmB interacts preferentially with ergosterol in cell membranes, giving rise to the fungal specificity, but it also interacts with a lower affinity with cholesterol, giving rise to toxic side effects that are often dose-limiting. The classical AmB mechanism is based on aqueous pores formed by AmB-sterol complexes, but it has recently been shown that ergosterol-binding alone , or extraction of ergosterol  form the basis of antifungal activity. We have used neutron reflection (NR) and deuterium labeling to study ergosterol and cholesterol extraction by AmB in yeast and model membranes, coupled to analysis of the lipid composition of both pathogenic and non-pathogenic yeast cells to elucidate the effect of membrane composition on AmB activity.
The structure of yeast membranes and their response to AmB [3,4] differ considerably from typical model lipid bilayers and depends on the degree of lipid polyunsaturation. AmB inserts in yeast membranes both in the absence and presence of ergosterol, but forms no aqueous pores. We have also observed in-situ a highly hydrated extramembraneous AmB layer, which does not form on simple POPC-sterol membranes. While AmB inserts to a much higher degree in cholesterol containing membranes, the amount of cholesterol extracted is very limited.
In membranes from the pathogenic yeast Candida glabrata, genetically manipulated to show either increased or decreased AmB resistance, resistance is related to changes in both AmB insertion and ergosterol extraction.
Cyanobacteria are relative simple unicellular photosynthetic prokaryotes, considered ancestors of higher plant chloroplasts. We have investigated thylakoid stacking and dynamics of three cyanobacterial species with various thylakoid arrangements in vivo by means of small-angle scattering and have correlated it to the results of transmission electron microscopy.
We have formulated a mathematical model describing thylakoid membrane ultrastructure and stacking using the designated form factor and the lamellar structure factor. This mathematical model has been implemented in the scattering curve-fitting framework 'WillItFit?', what now enables the fitting of experimental scattering data from photosynthetic organisms and obtaining thylakoid ultrastructural parameters: thylakoid membrane thickness, lumen width, thylakoid repeat distance and related uncertainty parameters.
The investigation of cyanobacterial thylakoid membranes by small-angle scattering can provide information on photosynthetic organism adaptation and thylakoid dynamics in relation to environmental factors and stimuli: e.g. ion concentration, illumination or temperature changes.
Biological membranes are intrinsically out-of-thermal equilibrium, driven by a vast range of external forces, excerted for example by motor proteins, ion channels, or interactions with other membranes such as in membrane fusion. We woul like to gain a quantitative understanding in the generic differenced of membrane dynamics under such out-of-equilibrium conditions. To this end, we have used different diffraction experiments probing the collective dynamics of well controlled model membranes, in particular aligned lipid membrane stacks, both in thermal equilibrium as well as in an active state, subjected to external driving forces (optical or acoustoelectric). While the equilibrium collective membrane dynamics (in plane density waves, undulations) can be well captured by ineleastic neutron scattering in terms of characterisitc dispersion relations , we turned to time-resolved x-ray diffraction in order to probe the relaxation pathway of membranes which are driven out of thermal equilibrium [2,3]. Beyond dynamics of multilamellar phases, I will then present ongoing work on membrane shape transformation and in particular membrane fusion in vesicle suspensions, using advanced x-ray optics and microfluidic devices. I will finish with an outlook on novel oppertunities by the x-ray free electron lasers as well as the European Spallation Source (ESS).
Pore-spanning membranes (PSMs) are well-suited to investigate single lipid diffusion as well as lipid domain diffusion. Recent findings have highlighted the dynamic nature of such domains in the plasma membrane and the key role of the underlying cytoskeleton meshwork in stabilizing them. We used porous substrates with different pore radii serving as a static meshwork to modulate the size of lipid domains in liquid ordered (lo)/liquid disordered (ld) phase-separated continuous PSMs composed of DOPC, sphingomyelin, cholesterol and the globoside Gb3. We analyzed domain formation and domain dynamics by fluorescence video microscopy. Analysis of the diffusion of mobile lo-domains entrapped in the freestanding parts of the PSMs showed that the domains’ diffusion constants are slowed down by orders of magnitude due to the confinement of the PSM, where the drag force is governed by both the friction in the bilayer and the coupling to the aqueous phase compared to the unrestricted case. The globoside Gb3 served as a receptor for the bacterial protein Shiga toxin, which is known to reorganize phase-separated lipid membranes to a great extent eventually leading to invaginations in the plasma membrane that result in the internalization of the protein into the cell.
The phospholipid bilayer is the basic structural motif of most biological membranes. As such, many biological processes occur within or in the proximity of the cell membrane, and therefore, interest in the properties and behavior of lipids in membranes is considerable. For example, it is found that in nature the lipid distribution across the inner and outer leaflet of cell membranes is asymmetric and this asymmetry plays a prominent role in processes like cell fusion, activation of the coagulation processes and the recognition and removal of apoptotic cells by macrophages. Therefore, there is great interest in studying the factors determining lipid movement across membranes as well as the resulting lipid mapping in the membrane, both of which are far from being understood and characterized.
In the literature, big discrepancies in the timescale of the occurrence of lipid flip-flop in model bilayer systems are found, partly due to the fact that these measurements were based on the indirect observation of the process and hampered by artifacts emerging from these different methodologies. Combining time resolved small angle neutron scattering and neutron reflectivity, we show that it is possible to capture inter and intra vesicular exchange as well as lipid composition differences in the leaflets of a model bilayer with the sub-nanometer spatial resolution and for times scales as short as a few minutes. Starting from these results we extensively studied, temperature and time dependence of the structure of lipid bilayers looking for traces of structural asymmetry and consequent relaxation towards an equilibrated symmetric bilayer.
By in situ monitoring the structure of a) a solid supported lipid bilayer exposed to a solution of isotopically labeled vesicles and b) a bulk mixture of hydrogenated and deuterated vesicles, we can provide new insight on the characteristics of inter- and intra-bilayer rearrangement processes. I will report on the rates and energetics of pure lipids as well as cholesterol transfer in different lipids environments. Particularly, we markedly found that cholesterol moves very slowly (tens to hundreds of minutes) across a single bilayer (flip-flop), over a large energy barrier which was not significantly different from the time or energy that it took cholesterol to move between vesicles (exchange). These results will be discussed with respect to MD predictions and to the presence of defects that can render the lipid intra-bilayer movement (flip-flop) the time limiting process or not.
It is today clear that lipids are involved in many physiological functions that go beyond the traditional view of compartmentalisation of the cell and its organelles. The metabolism of lipids including cholesterol involves the production, in the liver, of lipid carrying particles known as lipoproteins. Lipoproteins are nano emulsion like particles composed of fats and proteins (named apolipoproteins). The complexity of lipoproteins is great, with different compositions not only in terms of the amounts of the fat and proteic components, but also on the specific protein type and isoform. Specific apolipoproteins are known to mark an increased risk for developing atherosclerosis where fat accumulation to form plaques occurs at the initial stages of this terrible disease. In this talk, I will present the efforts of my group to explore the role of lipid dynamics in their transport throughout the body by lipoproteins. We show the unique power of using small angle neutron scattering and neutron reflection as two complementary techniques to map the structural and compositional changes taken place at both lipoproteins and model membranes over time. Contrast matching and deuteration are key to highlight the different aspects of these complex natural nanoparticles.
The highly dynamic nature of lipid membranes is crucial to accommodate the numerous types of biological molecules embedded within the cell membrane and help facilitate their functions. It is therefore essential to understand how lipid membrane dynamics are affected by different types of inclusions from a molecular level. One of the most important dynamic is the bending elastic modulus that controls both the nanoscale membrane fluctuations as well as the cell shape and deformability. Among other microscopic, spectroscopic, and scattering techniques, neutron spin echo (NSE) spectroscopy accesses collective membrane undulation fluctuations on the nanosecond time scales and can be used to quantify the bending modulus by comparing the measured relaxation time with theoretical derivations for the height-height fluctuations. More recently, collective thickness fluctuations with 100s nanosecond time scales have also been measured with NSE and an asymmetric bilayer model predicts that the membrane thickness fluctuations are driven by membrane compressibility and dissipated through the solvent and membrane viscosities. Accordingly, a measurement of the bilayer thickness fluctuations by NSE allows us to estimate the membrane viscosity. On one hand, we apply this technique to quantify the effects of different inclusions on the membrane elastic and viscous properties. On the other hand, this research opens new possibilities to directly compare experimental membrane viscosity with molecular diffusion behavior measured by neutron backscattering spectroscopy (BS) or other spectroscopic techniques. In this presentation, I will discuss the methodology we apply to analyze NSE data from model membrane systems as well as present a direct comparison between the membrane viscosity extracted from measurements of the thickness fluctuations and the lipid diffusion measured with BS technique.
We report on diffusion of hydration between phospholipid membranes using incoherent quasi-elastic neutron scattering (QENS) and computer simulations . The combination of a well-aligned stack of DMPC membranes with the large, 2-dimensional detectors available at the neutron spectrometer Let (ISIS, UK) allows for simultaneous access to water motions lateral and perpendicular to the membranes. The resulting 2-dimensional maps of relaxation time and stretching exponent evidence anomalous (sub-diffusive) and anisotropic diffusion of membrane hydration water varying on nanometer distances. By combining molecular dynamics and coarse-grained Brownian dynamics simulations, the overall behavior is reproduced, and the apparent features can be linked back to an intrinsic sub-diffusivity of water at picosecond time scales, and the anisotropy of confinement and local dynamical environments.
Despite the enormous potential for pharmaceutical applications, the molecular details of the changes in the stratum corneum (SC) associated with high permeability in the presence of different added chemicals are still not fully understood. Solvents in formulations or compounds that may facilitate transdermal drug delivery, called “penetration enhancers”, are among these added chemicals. These different molecules likely influence SC molecular components in very different ways. The aim of the present study is to characterize the molecular effect of different classes of molecules on SC lipid and protein components.
At normal relative humidity and ambient temperature, the main fraction of SC lipid and protein components are solid and highly ordered, while there is a very minor co-existing fraction that is fluid/mobile. Changes in this minor fluid fraction is inherently difficult to detect in experimental studies, however, it is considered crucial to SC barrier and mechanical properties. Through recent developments, Polarization Transfer Solid-State NMR (PT ssNMR) method together with almost complete peak assignment of SC components permits the detection of small changes in the molecular dynamics of the minor fluid lipid and protein components upon the added chemicals. Simultaneously we are able to monitor changes in molecular dynamics of the added molecules inside SC, enabling us to draw conclusions on interactions and partitioning of these molecules in SC. By correlating the effects on SC molecular components and SC barrier function, we aim at depended understanding of diffusional transport in SC and how this is related to the fluidity of the SC molecular components.
Model membrane studies involving X-rays and neutrons have greatly informed our understanding of the structure and dynamics of biological membranes. Recently, membrane lateral organization – thought to play a vital part in the function of membranes in vivo – has been emulated in model membrane systems, and neutron scattering approaches have been developed to study this organization on the nanoscale. At ORNL we have developed neutron scattering approaches to study membrane models, which have then been applied to in vivo systems (i.e., B. subtilis). During the seminar I will describe static and dynamic studies of laterally organized membranes and elaborate on future developments to the neutron spin echo (NSE) technique for accurately obtaining the bilayer bending modulus (Kc) – one of the most important physical constants characterizing lipid membranes – and membrane viscosity.
Modified messenger RNA constitues a new intresting approach for transient protein expression in different therapies. Based on learnings from DNA delivery the intracellular delivery of such macromolecules remains a challenge. In our work we have prepared so called lipid nanoparticles (LNPs), containing several copies of mRNA in each, using microfluidic mixing. The size of such LNPs can be easily manipulated using different amounts of PEG-ylated stabilizers resulting in rather monodisperse nanoparticles having number-averaged diameters between 50 and 150 nm. In these studies we have been using a erythropoietin mRNA (5-methylcytidine, pseudouridine). The focus of this presentation will be on structural studies using cryo-transmission electron microscopy, small-angle x-ray and neutron scattering for LNPs of different sizes. For the case of neutron scattering using contrast variation is especially informative. We use this information to enable more detailed studies of the LNP internal structure and to tailor the LNP surface composition. Furthermore, we have also performed in vitro cell (human hepatocytes and adipocytes) and in vivo (intravenous and subcutaneous administration) studies studies measuring both LNP uptake and the concurrent protein expression with the ambition to correlate LNP stucture to their biological function.
The high energy resolution of Neutron Spin Echo (NSE) Spectroscopy is perfectly suited for studying thermally driven membrane fluctuations in soft matter systems. Measuring membrane fluctuations allows determining the membrane bare bending rigidity of a membrane patch and is complementary to phase diagram measurements and SANS . Adding co-surfactants such as diblock copolymers or homopolymers to a microemulsion modifies the bending rigidity and has been followed with NSE. A recent topic in membrane dynamics was the influence of a rigid interface to a surfactant or phospholipid membrane. Microscopic flat interfaces provided by clay particles  as well as macroscopic interfaces were investigated . The latter has been studied in grazing incidence geometry, a new development in NSE Spectroscopy where in analogy to GISANS only the part of the sample at a solid/liquid interface is illuminated with an evanescent neutron wave. Despite the low intensity, a carefully designed low background sample cell together with new neutron optical components such as neutron resonator structures  allowed to gather interface specific information and to observe modifications of the undulation dispersion relations in soft membranes such as microemulsions  and also phospholipid membranes [6,7].
Surfactant/lipid vesicles are closed bilayer aggregates that are interesting to understand because of their importance in several biological processes. They are often surprisingly stable, partly because of an intriguing stability against Ostwald ripening. If, in addition, fusion events are rare, a vesicle dispersion may retain its size distribution for weeks and months. For this reason, kinetic stability of vesicles is sometimes misinterpreted as thermodynamic stability. Here we will focus on vesicle (membrane) fusion and how its kinetics depends on the surfactant monolayer spontaneous curvature, H0. As model system we have studied the binary water-C10E3 (CH3(CH2)9(OCH2CH2)3OH) system, where H0 of the non-ionic surfactant monolayer can be conveniently tuned by varying the temperature.(We use the convention that curvature away from water is counted as positive, thus, H0 decreases with increasing temperature, H0≈10-3(T0-T) where T0 is the “balanced temperature” where H0=0). In the vicinity of H0=0 (here, T≈26 °C), the surfactant may form two different bilayer phases. A lamellar phase, when H0>0 (T<26 °C) and a sponge phase when H0<0 (T>26 °C). Interestingly, it is found that the lamellar phase can in excess water be fragmented into kinetically stable uni-lamellar vesicles, while the sponge phase can not. Above 26 °C vesicles spontaneously fuse and the rate increases with increasing temperature. The fact that vesicle fusion typically requires H0<0 is consistent with membrane fusion models involving the so-called stalk intermediate structure. Vesicle fusion was also studied with giant uni-lamellar vesicles using rapid confocal laser scanning microscopy.
We have studied the effects of pressure on the gel-fluid transition in sphingomyelin bilayer membranes, and have found that the ordering of the chains and the development of the ripples on forming the gel phase occur on different timescales.
We have demonstrated by x-ray diffraction that fluid-fluid phase separation in ternary DOPC / DPPC / cholesterol mixtures can be induced in bulk phases by hydrostatic pressure. We have been able to image directly this pressure-induced phase separation in lipid vesicles by using high pressure optical microscopy.
By incorporation of charged phospholipids, we have been able to swell inverse bicontinuous cubic phases to approx. 500 Å, with water channels of approx. 220 Å diameter, potentially expanding the range of usefulness of such phases for applications].
We studied the effect of hydrostatic pressure on the structure and stability of the inverse micellar cubic phase Fd3m, and have discovered a number of novel effects. We have also studied the structure of this phase by contrast variation neutron scattering, and showed that the more weakly amphiphilic diacylglycerol component is preferentially located in the smaller, more highly curved inverse micelles. We discovered a lyotropic phase of space group P63/mmc, whose structure is based upon an hcp packing of quasi-spherical inverse micelles, in a hydrated mixture of DOPC, dioleoyl glycerol, and cholesterol. This phase is expected to have a greater chain packing frustration than the Fd3m cubic phase, and it appears that the cholesterol is able to relieve the chain packing frustration within the hydrophobic region of this phase, allowing the P63/mmc phase to form.
We also discovered a novel inverse ribbon phase in the branched-chain polyoxyethylene surfactant system C14C16EO4 in excess water. This phase is stabilised by the application of hydrostatic pressure, with the structure becoming increasingly distorted away from 2-D hexagonal symmetry with increasing pressure.
The properties of self-assembled amphiphilic molecules are of key relevance to understanding the complex processes that take place in biological membranes. Albeit structural characterization being the initial step to understanding such systems, ultimately the dynamics over a wider range of timescales are key to understanding function and the interactions tha take place between the multitude of components that can be found in a cell membrane. In this talk I will show how neutron spectroscopy, in particular quasi-elastic scattering, offers attractive possibilities to learn more about the heterogeneous dynamical landscape that exists in membranes. I will use example from simple micelles, simple model membranes and more complex lipidic assemblies.
The stratum corneum is the outermost layer of human skin and primary barrier towards the environment. The main component is stacked layers of saturated long-chain ceramides, free fatty acids and cholesterol, but we do not yet know the molecular structure or formation details. Here, I will present our work on new methods to fit models to low-resolution cryo-EM microscopy vitreous section (CEMOVIS) data, in particular by generating molecular models and using cryo-EM simulation to generate electron diffraction micrographs that can be compared directly to experimental data, and iteratively use these to improve the models. This has enabled us to create a number of alternative models, compare how they fit existing experimental data, and also use coarse-grained simulations to understand the formation process where cubic phases turn into bilayers depending on the lipid composition. These types of models can be highly useful tools for understanding the barrier properties, and we are currently combining it with free energy calculations to explore rapid prediction of permeation properties from CEMOVIS-derived models, which could have important applications in developing new generations of skin-permeating drugs.
Arginine-rich cell-penetrating peptides (RRPs) spontaneously traverse cell mem-
branes and are promising candidates for drug delivery. The internalization mechanism has been suggested to involve a cooperative process initiated by the self-aggregation of RRPs adsorbed onto the plasma membrane1. Likewise, formation of aggregates in solution has been shown to be related to the efficiency of cellular uptake2. Using small-angle X-ray scattering experiments and all-atom simulations, we study the solution behavior of arginine and lysine decapeptides. Despite its large positive charge, we ?find that deca-arginine self-associates in aqueous solution3. Simulations elucidate the molecular origin of the attraction, whereas inspection of the Protein Data Bank reveals that the mode of deca-arginine dimerization commonly occurs in protein crystal structures. To investigate the concerted interaction between multiple RRPs and a lipid bilayer over large time and length scales, we develop a computationally efficient coarse-grained model which can be readily parametrized against all-atom simulations as well as experimental data. The force ?eld is based on a granular representation of the mismatch between the dielectric constant of lipids and aqueous medium, combined with an accurate description of membrane elastic properties. Via constant-pH Monte Carlo simulations, the model is used to study the inuence of acid-base equilibria and chain length on the energetics of RRPs membrane permeation.
Membrane curvature generation and membrane remodeling underlie many biological processes such as virus entry into cells and virus budding. How proteins mediate this curvature generation is a fundamental question that is still poorly understood. I will present our recent structural studies, using solid-state NMR spectroscopy, of three membrane proteins that give insights into the mechanisms of virus-cell membrane fusion and virus budding. 1) We have investigated the structure of the membrane-interacting domains of the fusion proteins of the parainfluenza virus 5 (PIV5) and human immunodeficiency virus (HIV). For the PIV5 fusion protein, the fusion peptide (FP) and the transmembrane domain (TMD) show striking membrane-dependent conformations. The β-sheet conformation causes negative Gaussian curvature and membrane dehydration, which are required for membrane merger, while the α-helical conformation resides in low-curvature lamellar membranes and forms three-helix bundles. Therefore, the local lipid composition of the membrane is a key regulator of the site of virus-cell fusion. 2) For the HIV fusion protein gp41, we have determined the oligomeric structure of the membrane-proximal external region (MPER) and the TMD. We find that this domain is trimerized in the lipid membrane with a helix-turn-helix conformation, suggesting that this domain stabilizes the trimer structure of gp41 and promotes membrane curvature during the fusion process. 3) The influenza virus buds from host cells in a cholesterol-dependent manner using the matrix protein M2. To understand how cholesterol interacts with M2 to generate membrane curvature, we have determined the cholesterol-binding site of M2 by measuring protein-cholesterol distances and cholesterol orientation in the membrane. The data represent the first direct determination of the cholesterol-binding structure of a membrane protein in lipid bilayers, and moreover indicate a specific mechanism by which cholesterol concentration gradients in the membrane drive the M2 protein to the neck of the budding virus to conduct membrane scission.
Challenges, tools and opportunities to reveal dynamics of membranes and their constituents to further our understanding of their function. Debate leader Prof. John Katsaras, SNS, Oak Ridge, USA and panel members Prof. Peter Schurtenberger, LINXS and Lund University; Prof. Dr. Motomu Tanaka, University of Heidelberg and Kyoto University; Dr. Marité Cardenas, Malmö University; Dr Lennart Lindfors, AstraZeneca.