Fredric S. Cohen - US grants

DESCRIPTION (provided by applicant): Fusion between membranes is a ubiquitous yet poorly understood cellular process. Infection of cells by enveloped viruses is initiated by formation of a fusion pore between the virus and a host cell membrane. The process is induced by fusion proteins, and for many unrelated viruses these proteins share the common feature of three monomers that fold into six-helix bundles in their final structure. Among these proteins are the hemagglutinin (HA) of influenza virus and Env of HIV-1. Agents that inhibit bundle formation in HIV-1 Env prevent fusion and have been shown to have the potential to control the progression of AIDS. Vaccines against influenza are usually directed against HA. The formation of the six-helix bundle causes the transmembrane domains (TMDs) in the viral membrane and fusion peptides of the protein inserted in the target membrane to come into close proximity. A central question of the field is whether the movement of fusion peptides and TMDs toward each other is the direct cause of fusion pore formation. This will be tested for both HA and HI-1 Env using several approaches. For HA, the N-cap region and the region proximal to the TMD interact to cause proximity of the TMD and fusion peptide, so these regions will be mutated and the prediction that hemifusion, but not full fusion, occurs will be tested. The TMDs of HA must separate from each other to create proximity with the fusion peptides; preventing this separation by cross-linking monomers will determine which steps of fusion can occur without proximity. For Env, residues will be mutated to reduce bundle formation and the effects on fusion will be measured. Inhibitory peptides and mutation experiments will determine whether bundle formation drives fusion pore enlargement for Env. Recombinants of the bundle-forming portion of Env inhibit fusion; determining how they do so will either provide the free energy released by bundle formation or yield a means to explore associations between Env's. Extent of association between proteins will also be investigated by measuring kinetics of fusion as a function of HA density. The accumulation of HA-GFP into the region of contact and/or into large complexes, as fusion proceeds, will be explored. Protein associations will also be probed through the use of a simplified model system for Env receptor and coreceptor. These aims will be investigated by video fluorescence dye spread measurements, electrical admittance techniques, and laser scanning confocal microscopy, as appropriate.

Fusion between membranes occurs in a wide variety of cellular processes. Infection of cells by enveloped virus is initiated when viral and cell membranes fuse. Influenza virus enters cells by endocytosis. Low PH within endosomes causes a massive conformational change of the envelope glycoprotein hemagglutinin (HA) which leads to merger between endosomal membrane and viral envelope. This permits the viral nucleocapsid to pass through the fusion pore and into cytosol. HA is a model protein for the study of membrane fusion. Flu vaccines that must be developed each year are directed against HA. Thus, determining how HA mediates fusion is important not only for understanding the molecular mechanism of a critical cellular process but also for preventing disease. Functions of each of the three domains of HA -- the ectodomain, transmembrane (TM) domain, and cytosolic tail (CT) -- will be identified by fusing HA-expressing cells to red blood cells and planar bilayer membranes. Electrical admittance measurements will measure fusion pores and video fluorescence microscopy will monitor lipid continuity. The hypotheses that the ectodomain is responsible for hemifusion and that the TM domain causes the transition from hemifusion to full fusion will be tested. Truncation of the TM domain (including elimination of the CT) will isolate the role of the ectodomain and determine whether the full length of the TM domain is required for pore formation. The question of whether any TM domain will support fusion or particular primary sequences are necessary will be answered by using a chimeric protein that contains the ectodomain of HA but the TM domain of an integral membrane protein not involved in fusion. In the region connecting the ectodomain and TM domain, two conserved glycines will be eliminated to determine if decreasing flexibility in this area hinders fusion. Controlled chimera and mutation experiments will establish whether the general determining factor in pore flickering is acylation of a CT. New methods to detect hemifusion by fluorescence resonance energy transfer (FRET) will be developed. Preventing small fusion pores from further growth by manipulating growth conditions will show if lipid continuity between membranes is established before or after the pore forms. Agents that increase the positive spontaneous curvature, a fundamental characteristic of lipids, will be incorporated into inner membrane leaflets to investigate the extent of lipid control of growth, and possibly formation, of fusion pores. Experimental control of both lipid properties and protein structure is an integrated approach to the study of membrane fusion.

Colicins of the E1 family are bactericidal proteins which exert their lethal action by forming voltage-gated ion channels in the cytoplasmic bacterial membrane. The crystallographic structure in the water soluble state of one colicin is known. The formation and gating of colicin E1 channels are being studied in voltage-clamped, solvent-free, planar bilayer phospholipid membranes. These studies will lead to knowledge of the movements, on an atomic scale, of the protein during refolding from its water soluble to membrane-bound form and the conformational changes responsible for opening and closing of the channels. The properties of colicin E1 channels in planar bilayers thus provide a well-defined system to delineate the physico-chemical principles that underly the physiological processes of channel gating and protein translocation through bilayers. The enumeration of these principles would facilitate strategies for coupling protein toxins (e.g. ricin, abrin, diptheria) to carrier proteins so that the toxins retain their activity and cross targeted (e.g. transformed) cell membranes. This is, for example, the goal when designing immunotoxins -- toxins coupled to antibodies. This proposal is specifically directed toward determining the regions and residues of colicin that are translocated when the channel is gated by voltage and resolving the folding pattern of these regions in the bilayer. The voltage-dependence of deactivation of site-directed mutants that have charges added or deleted at defined residues will be measured. This dependence will give the fraction of the applied voltage sensed by each of the altered residues. Because this fraction sets the location of the mutated residues within the bilayer, the folding pattern of the channel will be obtained. Direct confirmation of proposed folding patterns and translocated regions will be sought by complexing membrane-impermeant avidin, added to the trans-aqueous compartment, with biotinylated colicin to lock biotinylated residues to the trans side. If the residues are translocated, deactivation will be inhibited. The deactivation kinetics are strongly dependent on the pH of the trans-aqueous compartment. Acidic residues, facing the trans compartment, that are neutralized by protonation at low pH (<4 - 5) are thought to be responsible for this pH dependence. Acidic residues that are candidates to face the trans compartment will be mutated to neutral ones to determine if the rates of deactivation are increased at high pH.

DESCRIPTION The research proposed desires to understand the events occurring within the lipid bilayer when the HA protein of influenza virus mediates membrane fusion. Fluorescence resonance energy transfer (FRET), polarization anisotropy and patch clamp measurements will be used to monitor changes in both viral and target membranes during the fusion process. Changes will be monitored using a set of fluorescent probes that have been synthesized and developed by the foreign collaborators. The specific aims are: 1. Ganglioside clustering during virus binding and membrane fusion that is triggered by lowering pH will be studied. 2. The distances between virus and liposome membranes in the prefusion state will be evaluated. 3. The role of ganglioside-bound HA and unbound HAs in causing fusion will be determined. 4. The kinetics of mixing of the inner and outer lipid monolayers during virus-liposome fusion will be compared. 5. The influence of the lipid composition in liposomes on the kinetics of lipid mixing will be studied. 6. The kinetics of pore growth between HA-expressing cells and large unilamellar vesicles will be measured.

Membrane fusion is a ubiquitous cellular process, but the proteins responsible for fusion have been unambiguously identified only in the case of enveloped virus. For enveloped virus, infection of cells is initiated by membrane fusion. A fusion pore forms and enlarges and the viral genome passes through the pore and into cytosol. The biophysical mechanism of fusion has been more extensively studied for hemagglutinin (HA) of influenza virus than for any other fusion protein. As HA and many other viral fusion proteins, including that of HIV-1, have the same core structure, and all viral fusion proteins initiate their action by insertion of fusion peptides into membranes, the overall mechanism by which HA induces fusion is probably similar for many, if not all, viral fusion proteins. Fusion, including pore behavior, has been most extensively studied by expressing HA on cell surfaces and fusing these cells to target membranes. However, cellular proteins could alter the fusion process and affect the pores. Individual influenza virions will therefore be fused to phospholipid bilayer membranes, which are free of protein, and the steps leading up to the formation of the fusion pore, the pore itself, and its subsequent enlargement will be characterized by electrical capacitance measurements. Whether full lipid continuity between membranes is established immediately upon fusion of a virus will be assessed by determining if fluorescent lipid dye can pass through the small fusion pore that initially forms. The density of HA in the viral envelope will be systematically reduced by proteolytically removing it and kinetics will be measured so that the number of HA molecules that associate in the creation of a pore can be estimated. Intermediate states of protein conformation and lipid monolayer arrangement from the bound state to fusion have been inferred for cellular systems that express HA. Whether these intermediate states do in fact precede fusion in the viral system will be established. An experimentally testable theoretical model will be constructed that relates the structural changes HA is known to undergo when fusion is triggered to the configurations through which membrane monolayers are thought to proceed. Whether a given change in HA can cause a corresponding change in monolayer configuration will be established by explicit calculation.

[unreadable] DESCRIPTION (provided by applicant): Rafts are specialized regions of membranes that consist of phase-separated domains of cholesterol and sphingolipids enriched in particular proteins. A large number of cellular processes - such as signal transduction and intracellular trafficking - are thought to be controlled by raft behavior. The wide-ranging importance of rafts has also linked them to many diseases, and some viruses even appear to fuse and/or bud at raft sites. But specific structures and dynamics of raft formation, growth, and composition are as yet unknown. The planar lipid bilayer model system has many advantages for discovering the physical chemical principles that govern these aspects of rafts; lipid phase separation, partitioning of proteins into cholesterol/sphingolipid domains, and control of formation of these domains by proteins can all be investigated in bilayer membranes. By including cholesterol, sphingomyelin, and fluorescent probes in bilayers, kinetic aspects of phase-separated cholesterol/sphingomyelin domains will be studied by fluorescence microscopy with selectivity and sensitivity not possible using cells. Fluorescent and non-fluorescent (quencher) probes will be constructed to partition into selected domains and placed in raft forming bilayers at high concentrations. These techniques will allow small lipid-microdomain rafts to be detected, their growth and dissolution to be quantified, and their stability to be characterized. Rafts within a single monolayer leaflet will be studied and any coupling to a liquid-ordered domain in the opposite monolayer will be investigated. The extent to which contact between acyl chains of lipids in opposite monolayers controls coupling will be determined. Among the proteins thought to partition into rafts, GPI anchored proteins are prominent; GPI-GFP will be used as a model protein to assess relationships between proteins and rafts under varying conditions. The hypothesis that cholesterol-binding protein can serve as a center for nucleation of rafts will be tested. The results of experimental aims will be used to adapt theory developed for phase creation and growth in other systems, so that an integrated understanding of rafts can be based on fundamental physical principles.

DESCRIPTION (provided by applicant): All viruses that contain class II or class III fusion proteins (and some with class I) fuse from within endosomes. For these viruses, the endosome is the initial site of infection. In response to the low-pH endosomal environment, a fusion protein undergoes conformational changes that cause merger of the viral and endosomal membranes, releasing the viral genetic material into cytosol. If regions on the fusion protein critical for infection are located, they will provide targets for new anti-viral drugs and vaccines. Properties of the endosomal membrane itself and its interior such as membrane voltage, acidic lipids, and redox potentials could also exert profound effects on fusion; if regulatory properties are identified and could be modified, new methods of halting infection could result. Voltage across endosomal membranes has been shown by our laboratory to control fusion of a number of types of virions that have class II or class III fusion proteins: the naturally occurring negative voltag across a membrane promotes fusion; positive voltage inhibits it. The universality of voltage dependence of class II and III fusion proteins is now reasonably certain. Chimera experiments strongly suggest the transmembrane domain (TMD) as the region of the fusion protein that confers voltage sensitivity. Voltage dependence could arise either because a TMD directly responds to voltage, or because acidic (negatively charged) lipids in outer membrane leaflets bind to TMDs. The concentration of acidic lipids in outer leaflets varies with voltage-dependent flip-flop between leaflets: enriching the concentration of acidic lipids in outer leaflets by experimentally incorporating them and measuring the consequences to voltage-dependent fusion will determine if acidic lipid binding causes voltage-dependent fusion. If it does, the binding region on the fusion protein will be identified by altering the protein. If the TMD is the voltage sensor, measuring displacement currents of synthetic TMDs will determine whether a large dipole moment is the key for sensing voltage. The energy required to transfer electrons (redox potentials) may also have an important role in regulating viral fusion: The redox potential of an endosome depends on its level of NADPH oxidase (NOX). Inhibition of NOX activity within endosomes indicates that the number of virions that fuse varies with the oxidation state in the same manner it does in cell-cell fusion. NOX activity will be altered, and fusion within endosomes monitored by confocal microscopy, to determine the relevance of redox potentials in infection. Methods to monitor membrane insertion of segments of viral fusion proteins will be developed through coupling lipophilic, charged probes to a fusion protein and electrophysiologically determining if voltage dependence of fusion is altered. This method will have far greater sensitivity than current methods. Class II and III proteins share some structural features; insertion studies could thus yield fundamental principles that unify the mechanisms of action of fusion proteins in these two classes. Clinically, identifying the mechanisms for endosomal control of viral fusion will reveal which processes could be interrupted to reduce or prevent infection.

Abstract Understanding mechanisms cells use to maintain cholesterol homeostasis are critical in cell biology and many diseases. To achieve this, the chemical activity of cholesterol in cell plasma membranes must be measured because activity controls cholesterol?s effects on cellular processes. To date, plasma membrane cholesterol concentration has been used to quantify cholesterol activity. But the activity of cholesterol is determined by its chemical potential; concentration contributes to, but does not accurately reflect membrane activity. Because a method to measure cholesterol chemical potential had not been available, it was not possible to properly evaluate many of cholesterol?s effects, including those on cellular signaling. We have now developed methods to do so. These methods and a new perfusion fluorimetry apparatus we have devised allow us to follow the chemical potential of cholesterol of plasma membranes in real time. We have discovered that cells quickly respond to changes in extracellular cholesterol by adjusting the cholesterol chemical potential of their plasma membranes without changing the total content of cellular cholesterol. This finding reveals a previously unknown mechanism to maintain cholesterol homeostasis: quick adjustment of plasma membrane chemical potentials to control cholesterol influx and efflux. We have identified protein scaffolded domains, as typified by caveolae, as sites at which cells sense and rapidly respond to external cholesterol. The abundance and total amount of cholesterol that resides in caveolae are determined by the extent of phosphorylation at position Ser80 of caveolin-1, the foundational protein of the domain. The shuttling of cholesterol between scaffolded domains and the surround which must result upon Ser80 phosphorylation alters cholesterol chemical potential. We therefore hypothesize that signaling cascades initiated within scaffolded domains are responsible for maintaining cholesterol homeostasis when cells are subjected to changes in external cholesterol and to growth factors. We further posit that these activated signaling cascades feed back to the plasma membrane to maintain chemical potentials. Cells will be stimulated with growth factors and relevant signaling cascades will be identified. The abundance of caveolae will be assessed by measuring the FRET (fluorescence resonance energy transfer) signals between caveolins. Our preliminary evidence strongly implicates that growth factors and/or changes in the level of external cholesterol stimulate the PI3K/Akt/mTOR signaling pathway that feeds back to achieve cholesterol homeostasis. Optogenetic techniques will be used to determine whether it and/or others are indeed responsible for control of cholesterol. Parallel experiments using the same strategies will determine if flotillins, analogous to caveolin, also serve as sensors/regulators of cholesterol chemical potentials.