The hemifusion intermediate and its conversion to complete fusion: regulation by membrane composition.

To fuse, membranes must bend. The energy of each lipid monolayer with respect to bending is minimized at the spontaneous curvature of the monolayer. Two lipids known to promote opposite spontaneous curvatures, lysophosphatidylcholine and arachidonic acid, were added to different sides of planar phospholipid membranes. Lysophosphatidylcholine added to the contacting monolayers of fusing membranes inhibited the hemifusion we observed between lipid vesicles and planar membranes. In contrast, fusion pore formation depended upon the distal monolayer of the planar membrane; lysophosphatidylcholine promoted and arachidonic acid inhibited. Thus, the intermediates of hemifusion and fusion pores in phospholipid membranes involve different membrane monolayers and may have opposite net curvatures, Biological fusion may proceed through similar intermediates.     

Liposome composition effects on lipid mixing between cells expressing influenza virus hemagglutinin and bound liposomes

The involvement of contacting and distal lipid monolayers in different stages of protein-mediated fusion was studied for fusion mediated by influenza virus hemagglutinin. Inclusion of non-bilayer lipids in the composition of the liposomes bound to hemagglutinin-expressing cells affects fusion triggered by low pH. Lysophosphatidylcholine added to the outer membrane monolayers inhibits fusion. The same lipid added to the inner monolayer of the liposomes promotes both lipid and content mixing. In contrast to the inverted cone-shaped lysophosphatidylcholine, lipids of the opposite effective shape, oleic acid or cardiolipin with calcium, present in the inner monolayers inhibit fusion. These results along with fusion inhibition by a bipolar lipid that does not support peeling of one monolayer of the liposomal membrane from the other substantiate the hypothesis that fusion proceeds through a local hemifusion intermediate. The transition from hemifusion to the opening of an expanding fusion pore allows content mixing and greatly facilitates lipid mixing between liposomes and cells.     

Structural intermediates in influenza haemagglutinin-mediated fusion

Fusion pore formation in the haemagglutinin (HA)-mediated fusion is a culmination of a multistep process, which involves low-pH triggered refolding of HA and rearrangement of membrane lipid bilayers. This rearrangement was arrested or slowed down by either altering lipid composition of the membranes, or lowering the density of HA, and/ or temperature. The results suggest that fusion starts with the lateral assembly of activated HA into multimeric complexes surrounding future fusion sites. The next fusion stage involves hemifusion, i.e. merger of only contacting membrane monolayers. Lysophosphatidylcholine reversibly arrests fusion prior to this hemifusion stage. In the normal fusion pathway, hemifusion is transient and is not accompanied by any measurable transfer of lipid probes between the membranes. A temperature of 4degreeC stabilizes this `restricted hemifusion' intermediate. The restriction of lipid flow through the restricted hemifusion site is HA-dependent and can be released by partial cleaving of low pH-forms of HA with mild proteinase K treatment. Lipid effects indicate that fusion proceeds through two different lipid-involving intermediates, which are characterized by two opposite curvatures of the lipid monolayer. Hemifusion involves formation of a stalk, a local bent connection between the outer membrane monolayers. Fusion pore formation apparently involves bending of the inner membrane monolayers, which come together in hemifusion. To couple low pH-induced refolding of HA with lipid rearrangements, it is proposed that the extension of the alpha -helical coiled coil of HA pulls fusion peptides inserted into the HA-expressing membrane and locally bends the membrane into a saddle-like shape. Elastic energy drives self-assembly of these HA-containing membrane elements into a ring-like complex and causes the bulging of the host membrane into a dimple growing towards the target membrane. Bending stresses in the lipidic top of the dimple facilitate membrane fusion.     

Structural intermediates in influenza haemagglutinin-mediated fusion.

Fusion pore formation in the haemagglutinin (HA)-mediated fusion is a culmination of a multistep process, which involves low-pH triggered refolding of HA and rearrangement of membrane lipid bilayers. This rearrangement was arrested or slowed down by either altering lipid composition of the membranes, or lowering the density of HA, and/or temperature. The results suggest that fusion starts with the lateral assembly of activated HA into multimeric complexes surrounding future fusion sites. The next fusion stage involves hemifusion, i.e. merger of only contacting membrane monolayers. Lysophosphatidylcholine reversibly arrests fusion prior to this hemifusion stage. In the normal fusion pathway, hemifusion is transient and is not accompanied by any measurable transfer of lipid probes between the membranes. A temperature of 4 degrees C stabilizes this 'restricted hemifusion' intermediate. The restriction of lipid flow through the restricted hemifusion site is HA-dependent and can be released by partial cleaving of low pH-forms of HA with mild proteinase K treatment. Lipid effects indicate that fusion proceeds through two different lipid-involving intermediates, which are characterized by two opposite curvatures of the lipid monolayer. Hemifusion involves formation of a stalk, a local bent connection between the outer membrane monolayers. Fusion pore formation apparently involves bending of the inner membrane monolayers, which come together in hemifusion. To couple low pH-induced refolding of HA with lipid rearrangements, it is proposed that the extension of the alpha-helical coiled coil of HA pulls fusion peptides inserted into the HA-expressing membrane and locally bends the membrane into a saddle-like shape. Elastic energy drives self-assembly of these HA-containing membrane elements into a ring-like complex and causes the bulging of the host membrane into a dimple growing towards the target membrane. Bending stresses in the lipidic top of the dimple facilitate membrane fusion.     

The Pathway of Membrane Fusion Catalyzed by Influenza Hemagglutinin: Restriction of Lipids, Hemifusion, and Lipidic Fusion Pore Formation

The mechanism of bilayer unification in biological fusion is unclear. We reversibly arrested hemagglutinin (HA)-mediated cell–cell fusion right before fusion pore opening. A low-pH conformation of HA was required to form this intermediate and to ensure fusion beyond it. We present evidence indicating that outer monolayers of the fusing membranes were merged and continuous in this intermediate, but HA restricted lipid mixing. Depending on the surface density of HA and the membrane lipid composition, this restricted hemifusion intermediate either transformed into a fusion pore or expanded into an unrestricted hemifusion, without pores but with unrestricted lipid mixing. Our results suggest that restriction of lipid flux by a ring of activated HA is necessary for successful fusion, during which a lipidic fusion pore develops in a local and transient hemifusion diaphragm.     

Reversible merger of membranes at the early stage of influenza hemagglutinin-mediated fusion

Fusion mediated by influenza hemagglutinin (HA), a prototype fusion protein, is commonly detected as lipid and content mixing between fusing cells. Decreasing the surface density of fusion-competent HA inhibited these advanced fusion phenotypes and allowed us to identify an early stage of fusion at physiological temperature. Although lipid flow between membranes was restricted, the contacting membrane monolayers were apparently transiently connected, as detected by the transformation of this fusion intermediate into complete fusion after treatments known to destabilize hemifusion diaphragms. These reversible connections disappeared within 10-20 min after application of low pH, indicating that after the energy released by HA refolding dissipated, the final low pH conformation of HA did not support membrane merger. Although the dynamic character and the lack of lipid mixing at 37 degrees C distinguish the newly identified fusion intermediate from the intermediate arrested at 4 degrees C described previously, both intermediates apparently belong to the same family of restricted hemifusion (RH) structures. Because the formation of transient RH structures at physiological temperatures was as fast as fusion pore opening and required less HA, we hypothesize that fusion starts with the formation of multiple RH sites, only a few of which then evolve to become expanding fusion pores.     

Model of membrane fusion: Continuous transition to fusion pore with regard of hydrophobic and hydration interactions

We consider the process of fusion of lipid membranes from the stage of stalk with minimal radius to the stage of fusion pore. We assume that stalk directly developed into the fusion pore, omitting the stage of hemifusion diaphragm. Energy of intermediate stages is calculated on the basis of the classical elasticity theory of liquid crystals adapted for lipid membranes. The trajectory of transition from stalk to pore is obtained with regard to hydrophobic and hydration interactions. Continuous change of orientation of lipids in distal monolayers occurs along the trajectory. The orientation changes from the direction along rotational axis of the system specific to stalk to the direction corresponding to the fusion pore. Dependence of energy of intermediate stages on the value of spontaneous curvature of distal monolayers of the fusing membranes is obtained. We demonstrate that the energy barrier of the stalk-to-pore transition decreases when distal monolayers have positive spontaneous curvature, which is in accordance with available experimental data.     

Meta-stability of the hemifusion intermediate induced by glycosylphosphatidylinositol-anchored influenza hemagglutinin.

Fusion between influenza virus and target membranes is mediated by the viral glycoprotein hemagglutinin (HA). Replacement of the transmembrane domain of HA with a glycosylphosphatidylinositol (GPI) membrane anchor allows lipid mixing but not the establishment of cytoplasmic continuity. This observation led to the proposal that the fusion mechanism passes through an intermediate stage corresponding to hemifusion between outer monolayers. We have used confocal fluorescence microscopy to study the movement of probes for specific bilayer leaflets of erythrocytes fusing with HA-expressing cells. N-Rh-PE and NBD-PC were used for specific labeling of the outer and inner membrane leaflet, respectively. In the case of GPI-HA-induced fusion, different behaviors of lipid transfer were observed, which include 1) exclusive movement of N-Rh-PE (hemifusion), 2) preferential movement of N-Rh-PE relative to NBD-PC, and 3) equal movement of both lipid analogs. The relative population of these intermediate states was dependent on the time after application of a low pH trigger for fusion. At early time points, hemifusion was more common and full redistribution of both bilayers was rare, whereas later full redistribution of both probes was frequently observed. In contrast to wild-type HA, the latter was not accompanied by mixing of the cytoplasmic marker Lucifer Yellow. We conclude that 1) the GPI-HA-mediated hemifusion intermediate is meta-stable and 2) expansion of an aqueous fusion pore requires the transmembrane and/or cytoplasmic domain of HA.     

Probing the mechanism of fusion in a two-dimensional computer simulation

A two-dimensional (2D) model of lipid bilayers was developed and used to investigate a possible role of membrane lateral tension in membrane fusion. We found that an increase of lateral tension in contacting monolayers of 2D analogs of liposomes and planar membranes could cause not only hemifusion, but also complete fusion when internal pressure is introduced in the model. With a certain set of model parameters it was possible to induce hemifusion-like structural changes by a tension increase in only one of the two contacting bilayers. The effect of lysolipids was modeled as an insertion of a small number of extra molecules into the cis or trans side of the interacting bilayers at different stages of simulation. It was found that cis insertion arrests fusion and trans insertion has no inhibitory effect on fusion. The possibility of protein participation in tension-driven fusion was tested in simulation, with one of two model liposomes containing a number of structures capable of reducing the area occupied by them in the outer monolayer. It was found that condensation of these structures was sufficient to produce membrane reorganization similar to that observed in simulations with "protein-free" bilayers. These data support the hypothesis that changes in membrane lateral tension may be responsible for fusion in both model phospholipid membranes and in biological protein-mediated fusion.     

A point mutation in the transmembrane domain of the hemagglutinin of influenza virus stabilizes a hemifusion intermediate that can transit to fusion

A hemagglutinin (HA) of influenza virus having a single semiconserved Gly residue within the transmembrane domain mutated to Leu (G520L) was expressed on cells; these cells were bound to red blood cells. By decreasing pH at 23 degrees C rather than 37 degrees C, an intermediate with properties expected of hemifusion just as the membranes are about to transit to full fusion was captured. As evidence: 1) increasing temperature to 37 degrees C at neutral pH allowed fusion to proceed; 2) after achieving the intermediate, the two membranes did not separate from each other after proteolytic cleavage of G520L because cells treated with proteinase K could not fuse upon temperature increase but could fuse upon the addition of chlorpromazine; and 3) at the point of the intermediate, adding exogenous lipids known to promote or inhibit the creation of hemifusion did not significantly alter the lipid dye spread that occurred upon increasing temperature, implying that at the intermediate, contacting membrane leaflets had already merged. A stable intermediate of hemifusion that could transit to fusion was also generated for wild-type HA, but pH had to be reduced at the significantly lower temperature of 4 degrees C. The fusion pores generated by G520L did not enlarge, whereas those induced by wild-type HA did. The finding that a state of transitional hemifusion can be readily obtained via a point mutation without the need for unusually low temperature supports the hypothesis that hemifusion occurs before pore formation.     

Effect of membrane curvature-modifying lipids on membrane fusion by tick-borne encephalitis virus

Enveloped viruses enter cells by fusion of their own membrane with a cellular membrane. Incorporation of inverted-cone-shaped lipids such as lysophosphatidylcholine (LPC) into the outer leaflet of target membranes has been shown previously to impair fusion mediated by class I viral fusion proteins, e.g., the influenza virus hemagglutinin. It has been suggested that these results provide evidence for the stalk-pore model of fusion, which involves a hemifusion intermediate (stalk) with highly bent outer membrane leaflets. Here, we investigated the effect of inverted-cone-shaped LPCs and the cone-shaped oleic acid (OA) on the membrane fusion activity of a virus with a class II fusion protein, the flavivirus tick-borne encephalitis virus (TBEV). This study included an analysis of lipid mixing, as well as of the steps preceding or accompanying fusion, i.e., binding to the target membrane and lipid-induced conformational changes in the fusion protein E. We show that the presence of LPC in the outer leaflet of target liposomes strongly inhibited TBEV-mediated fusion, whereas OA caused a very slight enhancement, consistent with a fusion mechanism involving a lipid stalk. However, LPC also impaired the low-pH-induced binding of a soluble form of the E protein to liposomes and its conversion into a trimeric postfusion structure that requires membrane binding at low pH. Because inhibition is already observed before the lipid-mixing step, it cannot be determined whether impairment of stalk formation is a contributing factor in the inhibition of fusion by LPC. These data emphasize, however, the importance of the composition of the target membrane in its interactions with the fusion peptide that are crucial for the initiation of fusion.     

Hemifusion between cells expressing hemagglutinin of influenza virus and planar membranes can precede the formation of fusion pores that subsequently fully enlarge

The chronological relation between the establishment of lipid continuity and fusion pore formation has been investigated for fusion of cells expressing hemagglutinin (HA) of influenza virus to planar bilayer membranes. Self-quenching concentrations of lipid dye were placed in the planar membrane to monitor lipid mixing, and time-resolved admittance measurements were used to measure fusion pores. For rhodamine-PE, fusion pores always occurred before a detectable amount of dye moved into an HA-expressing cell. However, with DiI in the planar membrane, the relationship was reversed: the spread of dye preceded formation of small pores. In other words, by using DiI as probe, hemifusion was clearly observed to occur before pore formation. For hemifused cells, a small pore could form and subsequently fully enlarge. In contrast, for cells that express a glycosylphosphatidylinositol-anchored ectodomain of HA, hemifusion occurred, but no fully enlarged pores were observed. Therefore, the transmembrane domain of HA is required for the formation of fully enlarging pores. Thus, with the planar bilayer membranes as target, hemifusion can precede pore formation, and the occurrence of lipid dye spread does not preclude formation of pores that can enlarge fully.     

Calculating Transition Energy Barriers and Characterizing Activation States for Steps of Fusion

We use continuum mechanics to calculate an entire least energy pathway of membrane fusion, from stalk formation, to pore creation, and through fusion pore enlargement. The model assumes that each structure in the pathway is axially symmetric. The static continuum stalk structure agrees quantitatively with experimental stalk architecture. Calculations show that in a stalk, the distal monolayer is stretched and the stored stretching energy is significantly less than the tilt energy of an unstretched distal monolayer. The string method is used to determine the energy of the transition barriers that separate intermediate states and the dynamics of two bilayers as they pass through them. Hemifusion requires a small amount of energy independently of lipid composition, while direct transition from a stalk to a fusion pore without a hemifusion intermediate is highly improbable. Hemifusion diaphragm expansion is spontaneous for distal monolayers containing at least two lipid components, given sufficiently negative diaphragm spontaneous curvature. Conversely, diaphragms formed from single-component distal monolayers do not expand without the continual injection of energy. We identify a diaphragm radius, below which central pore expansion is spontaneous. For larger diaphragms, prior studies have shown that pore expansion is not axisymmetric, and here our calculations supply an upper bound for the energy of the barrier against pore formation. The major energy-requiring deformations in the steps of fusion are: widening of a hydrophobic fissure in bilayers for stalk formation, splay within the expanding hemifusion diaphragm, and fissure widening initiating pore formation in a hemifusion diaphragm.     

Molecular Dynamics Simulation Analysis of Membrane Defects and Pore Propensity of Hemifusion Diaphragms

Membrane fusion often exhibits slow dynamics in electrophysiological experiments, involving prespike foot and fusion pore-flickering, but the structural basis of such phenomena remains unclear. Hemifusion intermediates have been implicated in the early phase of membrane fusion. To elucidate the dynamics of formation of membrane defects and pores within the hemifusion diaphragm (HD), atomistic and coarse-grained models of hemifusion intermediates were constructed using dipalmitoylphosphatidylcholine or dioleoylphosphatidylcholine membranes. The work necessary to displace a lipid molecule to the hydrophobic core of the bilayer was measured. For a lipid within the HD with radius of 4 nm, the work was ∼80 kJ/mol, similar to that in a planar bilayer. The work was much less (∼40 kJ/mol) when the HD was surrounded by a steep stalk, i.e., stalk wings forming a large angle at the junction of three bilayers. In the latter case, the lipid displacement engendered formation of a pore contacting the HD rim. The work was similarly small (40 kJ/mol) for a small HD of 1.5 nm radius, where a pore formed and grew rapidly, quickly generating a toroidal structure (<40 ns). Combining the steep stalk and the small HD decreased the work further, although quantitative analysis was difficult because the latter system was not in a stable equilibrium state. Results suggest that fine tuning of fusion dynamics requires strict control of the HD size and the angle between the expanded stalk and HD. In additional free simulations, the steep stalk facilitated widening of a preformed pore contacting the HD rim.     

Lipid intermediates in membrane fusion: formation,structure, and decay of hemifusion diaphragm

Lipid bilayer fusion is thought to involve formation of a local hemifusion connection, referred to as a fusion stalk. The subsequent fusion stages leading to the opening of a fusion pore remain unknown. The earliest fusion pore could represent a bilayer connection between the membranes and could be formed directly from the stalk. Alternatively, fusion pore can form in a single bilayer, referred to as hemifusion diaphragm (HD), generated by stalk expansion. To analyze the plausibility of stalk expansion, we studied the pathway of hemifusion theoretically, using a recently developed elastic model. We show that the stalk has a tendency to expand into an HD for lipids with sufficiently negative spontaneous splay, (~)J(s)< 0. For different experimentally relevant membrane configurations we find two characteristic values of the spontaneous splay. (~)J*(s) and (~)J**(s), determining HD dimension. The HD is predicted to have a finite equilibrium radius provided that the spontaneous splay is in the range (~)J**(s)< (~)J(s)<(~)J*(s), and to expand infinitely for (~)J(s)<(~)J**(s). In the case of common lipids, which do not fuse spontaneously, an HD forms only under action of an external force pulling the diaphragm rim apart. We calculate the dependence of the HD radius on this force. To address the mechanism of fusion pore formation, we analyze the distribution of the lateral tension emerging in the HD due to the establishment of lateral equilibrium between the deformed and relaxed portions of lipid monolayers. We show that this tension concentrates along the HD rim and reaches high values sufficient to rupture the bilayer and form the fusion pore. Our analysis supports the hypothesis that transition from a hemifusion to a fusion pore involves radial expansion of the stalk.     

The 1-127 HA2 construct of influenza virus hemagglutinin induces cell-cell hemifusion.

Conformational changes in the HA2 subunit of influenza hemagglutinin (HA) are coupled to membrane fusion. We investigated the fusogenic activity of the polypeptide FHA2 representing 127 amino-terminal residues of the ectodomain of HA2. While the conformation of FHA2 both at neutral and at low pH is nearly identical to the final low-pH conformation of HA2, FHA2 still induces lipid mixing between liposomes in a low-pH-dependent manner. Here, we found that FHA2 induces lipid mixing between bound cells, indicating that the "spring-loaded" energy is not required for FHA2-mediated membrane merger. Although, unlike HA, FHA2 did not form an expanding fusion pore, both acidic pH and membrane concentrations of FHA2, required for lipid mixing, have been close to those required for HA-mediated fusion. Similar to what is observed for HA, FHA2-induced lipid mixing was reversibly blocked by lysophosphatidylcholine and low temperature, 4 degrees C. The same genetic modification of the fusion peptide inhibits both HA- and FHA2-fusogenic activities. The kink region of FHA2, critical for FHA2-mediated lipid mixing, was exposed in the low-pH conformation of the whole HA prior to fusion. The ability of FHA2 to mediate lipid mixing very similar to HA-mediated lipid mixing is consistent with the hypothesis that hemifusion requires just a portion of the energy released in the conformational change of HA at acidic pH.     

Inner but Not Outer Membrane Leaflets Control the Transition from Glycosylphosphatidylinositol-anchored Influenza Hemagglutinin-induced Hemifusion to Full Fusion

Cells that express wild-type influenza hemagglutinin (HA) fully fuse to RBCs, while cells that express the HA-ectodomain anchored to membranes by glycosylphosphatidylinositol, rather than by a transmembrane domain, only hemifuse to RBCs. Amphipaths were inserted into inner and outer membrane leaflets to determine the contribution of each leaflet in the transition from hemifusion to fusion. When inserted into outer leaflets, amphipaths did not promote the transition, independent of whether the agent induces monolayers to bend outward (conferring positive spontaneous monolayer curvature) or inward (negative curvature). In contrast, when incorporated into inner leaflets, positive curvature agents led to full fusion. This suggests that fusion is completed when a lipidic fusion pore with net positive curvature is formed by the inner leaflets that compose a hemifusion diaphragm. Suboptimal fusion conditions were established for RBCs bound to cells expressing wild-type HA so that lipid but not aqueous dye spread was observed. While this is the same pattern of dye spread as in stable hemifusion, for this “stunted” fusion, lower concentrations of amphipaths in inner leaflets were required to promote transfer of aqueous dyes. Also, these amphipaths induced larger pores for stunted fusion than they generated within a stable hemifusion diaphragm. Therefore, spontaneous curvature of inner leaflets can affect formation and enlargement of fusion pores induced by HA. We propose that after the HA-ectodomain induces hemifusion, the transmembrane domain causes pore formation by conferring positive spontaneous curvature to leaflets of the hemifusion diaphragm.     

Point-like protrusion as a prestalk intermediate in membrane fusion pathway

The widely accepted pathway of membrane fusion begins with the fusion stalk representing the initial intermediate of hemifusion. The lipid structures preceding hemifusion and their possible influence on fusion kinetics were not addressed. Here, we suggest the point-like protrusion as a prestalk fusion intermediate, which has energy lower than that of stalk and, therefore, does not limit the fusion rate. We demonstrate that by calculating the energy of the point-like protrusion, which depends on the lipid monolayer elastic parameters and the strength of the intermembrane hydration repulsion. The point-like protrusion completes the fusion-through-hemifusion model of membrane merger.     

A hemifused complex is the hub in a network of pathways to membrane fusion

Membrane fusion is a critical step for many essential processes, from neurotransmission to fertilization. For over 40 years, protein-free fusion driven by calcium or other cationic species has provided a simplified model of biological fusion, but the mechanisms remain poorly understood. Cation-mediated membrane fusion and permeation are essential in their own right to drug delivery strategies based on cell-penetrating peptides or cation-bearing lipid nanoparticles. Experimental studies suggest calcium drives anionic membranes to a hemifused intermediate that constitutes a hub in a network of pathways, but the pathway selection mechanism is unknown. Here we develop a mathematical model that identifies the network hub as a highly dynamic hemifusion complex. Multivalent cations drive expansion of this high-tension hemifusion interface between interacting vesicles during a brief transient. The fate of this interface determines the outcome, either fusion, dead-end hemifusion, or vesicle lysis. The model reproduces the unexplained finding that calcium-driven fusion of vesicles with planar membranes typically stalls at hemifusion, and we show the equilibrated hemifused state is a novel lens-shaped complex. Thus, membrane fusion kinetics follow a stochastic trajectory within a network of pathways, with outcome weightings set by a hemifused complex intermediate.     

Stalk model of membrane fusion: solution of energy crisis.

Membrane fusion proceeds via formation of intermediate nonbilayer structures. The stalk model of fusion intermediate is commonly recognized to account for the major phenomenology of the fusion process. However, in its current form, the stalk model poses a challenge. On one hand, it is able to describe qualitatively the modulation of the fusion reaction by the lipid composition of the membranes. On the other, it predicts very large values of the stalk energy, so that the related energy barrier for fusion cannot be overcome by membranes within a biologically reasonable span of time. We suggest a new structure for the fusion stalk, which resolves the energy crisis of the model. Our approach is based on a combined deformation of the stalk membrane including bending of the membrane surface and tilt of the hydrocarbon chains of lipid molecules. We demonstrate that the energy of the fusion stalk is a few times smaller than those predicted previously and the stalks are feasible in real systems. We account quantitatively for the experimental results on dependence of the fusion reaction on the lipid composition of different membrane monolayers. We analyze the dependence of the stalk energy on the distance between the fusing membranes and provide the experimentally testable predictions for the structural features of the stalk intermediates.