Pore Lifetimes Research Articles

Shear stress induced lipid order and permeability changes of giant unilamellar vesicles

BackgroundThe permeability of a lipid bilayer is a function of its phase state and depends non-linearly on thermodynamic variables such as temperature, pressure or pH. We investigated how shear forces influence the phase state of giant unilamellar vesicles and their membrane permeability. MethodsWe determined the permeability of giant unilamellar vesicles composed of different phospholipid species under shear flow in a tube at various temperatures around and far off the melting point by analyzing the release of fluorescently labelled dextran. Furthermore, we quantified phase state changes of these vesicles under shear forces using spectral decomposition of the membrane embedded fluorescent dye Laurdan. ResultsWe observed that the membrane permeability follows a step function with increasing permeability at the transition from the gel to the fluid phase and vice versa. Second, there was an all-or-nothing permeabilization near the main phase transition temperature and a gradual dye release far off the melting transition. Third, the Laurdan phase state analysis suggests that shear forces induce a reversible melting temperature shift in giant unilamellar vesicle membranes. Major conclusionsThe observed effects can be explained best in a scenario in which shear forces directly induce membrane pores that possess relatively long pore lifetimes in proximity to the phase transition. General significanceOur study elucidates the release mechanism of thermo-responsive drug carriers as we found that liposome permeabilization is not continuous but quantized. Furthermore, the shear force induced melting temperature shift must be taken into consideration when thermo-responsive liposomes are designed.

Full-fusion and kiss-and-run in chromaffin cells controlled by irreversible vesicle size-dependent fusion pore transitions

Exocytosis operates through two distinct modes. Full-fusion leads to rapid expulsion of the entire content of a vesicle; kiss-and-run leads to slow and partial expulsion. These two modes have important biological consequences for endocrine regulation and synaptic transmission. Amperometry recordings of catecholamine release from chromaffin cells reveal single-vesicle fusion events corresponding to both of these modes, but classification is often difficult. This study introduces a new method of analyzing amperometry data to improve this classification. The ratio of the average amplitude to the peak amplitude differs between full-fusion and kiss-and-run, and the probability distribution of this ratio is well fitted by a double-Gaussian. Kiss-and-run events identified by this method have fusion pores with kinetic properties different from pores associated with full-fusion. They have slower transition rates and lifetime distributions indicative of irreversible transitions. The total-charge of an amperometric spike is expected to scale with vesicle volume during a full-fusion event. The cube root of this quantity should therefore scale with diameter, but the distribution of this quantity differs from the distribution of vesicle diameter seen in the electron microscope. Fusion pore lifetimes associated with full-fusion depend on vesicle size, and this makes the choice of mode size dependent. The fusion pore thus bifurcates after opening, and vesicle size influences this choice. The secretory vesicle protein synaptophysin influences the size dependence of fusion pore lifetime and the choice of release mode. Incorporating vesicle size into an analysis of release mode reconciled the kinetics of fusion pores, as well as the distributions of vesicle diameter and catecholamine content. Thus, the initial fusion pore emerges as a critical focus in endocrine regulation. By modulating the size-dependence of the mode of exocytosis, changes in the molecular makeup of the exocytotic apparatus can impact the shape and size of an amperometric event, and the speed and composition of secretion.

GM1 Leaflet Asymmetry Stabilizes Membrane Pores

For effective electrochemotherapy treatments and cell transfection, the cell plasma membrane needs to be transiently porated by electric fields and reseal afterwards. Successful membrane poration and cell viability depend critically on electroporation parameters such as field strength and pore lifetime. Model membranes are used to understand electroporation phenomena, but their pore lifetimes are much shorter than they are in cells. Cell membranes are highly asymmetric. Here, we investigated the influence of membrane asymmetry on electroporation of synthetic model membranes.

Chromogranin A, the major lumenal protein in chromaffin granules, controls fusion pore expansion.

Upon fusion of the secretory granule with the plasma membrane, small molecules are discharged through the immediately formed narrow fusion pore, but protein discharge awaits pore expansion. Recently, fusion pore expansion was found to be regulated by tissue plasminogen activator (tPA), a protein present within the lumen of chromaffin granules in a subpopulation of chromaffin cells. Here, we further examined the influence of other lumenal proteins on fusion pore expansion, especially chromogranin A (CgA), the major and ubiquitous lumenal protein in chromaffin granules. Polarized TIRF microscopy demonstrated that the fusion pore curvature of granules containing CgA-EGFP was long lived, with curvature lifetimes comparable to those of tPA-EGFP-containing granules. This was surprising because fusion pore curvature durations of granules containing exogenous neuropeptide Y-EGFP (NPY-EGFP) are significantly shorter (80% lasting <1 s) than those containing CgA-EGFP, despite the anticipated expression of endogenous CgA. However, quantitative immunocytochemistry revealed that transiently expressed lumenal proteins, including NPY-EGFP, caused a down-regulation of endogenously expressed proteins, including CgA. Fusion pore curvature durations in nontransfected cells were significantly longer than those of granules containing overexpressed NPY but shorter than those associated with granules containing overexpressed tPA, CgA, or chromogranin B. Introduction of CgA to NPY-EGFP granules by coexpression converted the fusion pore from being transient to being longer lived, comparable to that found in nontransfected cells. These findings demonstrate that several endogenous chromaffin granule lumenal proteins are regulators of fusion pore expansion and that alteration of chromaffin granule contents affects fusion pore lifetimes. Importantly, the results indicate a new role for CgA. In addition to functioning as a prohormone, CgA plays an important role in controlling fusion pore expansion.

Pulsatile Gating of Giant Vesicles Containing Macromolecular Crowding Agents Induced by Colligative Nonideality.

The ability of large macromolecules to exhibit nontrivial deviations in colligative properties of their aqueous solutions is well-appreciated in polymer physics. Here, we show that this colligative nonideality subjects giant lipid vesicles containing inert macromolecular crowding agents to osmotic pressure differentials when bathed in small-molecule osmolytes at comparable concentrations. The ensuing influx of water across the semipermeable membrane induces characteristic swell-burst cycles: here, cyclical and damped oscillations in size, tension, and membrane phase separation occur en route to equilibration. Mediated by synchronized formation of transient pores, these cycles orchestrate pulsewise ejection of macromolecules from the vesicular interior reducing the osmotic differential in a stepwise manner. These experimental findings are fully corroborated by a theoretical model derived by explicitly incorporating the contributions of the solution viscosity, solute diffusivity, and the colligative nonideality of the osmotic pressure in a previously reported continuum description. Simulations based on this model account for the differences in the details of the noncolligatively induced swell-burst cycles, including numbers and periods of the repeating cycles, as well as pore lifetimes. Taken together, our observations recapitulate behaviors of vesicles and red blood cells experiencing sudden osmotic shocks due to large (hundreds of osmolars) differences in the concentrations of small molecule osmolytes and link intravesicular macromolecular crowding with membrane remodeling. They further suggest that any tendency for spontaneous overcrowding in single giant vesicles is opposed by osmotic stresses and requires independent specific interactions, such as associative chemical interactions or those between the crowders and the membrane boundary.

Erlang flow of hydrophilic pore formation and closure events in a lipid bilayer during phase transition resulting from diffusion in the radius space.

The Smoluchowski equation with an energy profile of a special type and an assumed hydrophobic ("half") pore source term is used to describe the process of hydrophilic pore formation in a lipid bilayer at the gel-liquid phase transition. The source term reflects the occurrence of molecule packing defects in a lipid bilayer at phase transition. The time sequences of the pore formation and closure events are treated as non-stationary, second-order Erlang flows whose characteristics depend on the equation solution. The computed distributions of the time intervals between hydrophilic pores, and pore lifetimes agree with the previously published experimental interpulse interval and pulse duration histograms for the current fluctuations through planar bilayer membranes of DPPC immersed in a LiCl aqueous solution containing polyethylene glycol. Thus, the statistical analysis of pore formation and closure times leads us to conclude that firstly, the increased permeability of a lipid bilayer during the gel-liquid phase transition is accounted for by the emergence of additional hydrophobic defects in the heterogeneous structure of the bilayer and secondly, that the non-exponential distributions of the lipid channel closed and open times observed in experiments are evidence that the process of hydrophilic pore formation is not a one-step process but involves at least two dependent events.

Effects of Membrane Tension on Snare-Mediated Single Fusion Pores

Fast and temporally controlled release is imperative for neurotransmitter release via evoked synaptic vesicle exocytosis. This process is driven by the interaction between vesicle v-SNAREs with target t-SNAREs on the plasma membrane. The initial, nanometer-sized connection between two fusing lipid bilayers is called the fusion pore. This structure only lasts for a short time before it either dilates completely (resulting in full fusion), or reseals again (kiss-and-run fusion). Fusion pore dynamics determine the amount and size of cargo released, as well as vesicle recycling. Thus they constitute a fundamental mechanism that controls the extent of postsynaptic neuronal activity. Due to its small size and transient nature, the fusion pore has been an elusive object to study. We have recently developed an assay using biochemically well-defined components to look at current flow through single SNARE-mediated fusion pores. The components of the assay include HeLa cells with “flipped” t-SNARE proteins expressed on the outside of the cell (t-cells), and nanodiscs with v-SNARE proteins embedded within them (v-discs). When we fill a patch pipette with v-discs and establish a cell-attached patch onto a t-cell, the v- and t-SNAREs interact, leading to the formation of a fusion pore connecting the cytosol to the pipette solution. Currents through these fusion pores report single fusion pore dynamics.Membrane tension has been shown to affect fusion pore dynamics in neuroendocrine cells. In our assay, we can quickly and accurately change the pressure in the patch pipette, thereby modulate membrane tension, using a high-speed pressure clamp. We will show how an increase in membrane tension causes extended pore lifetimes, as well as increased pore conductance and charge influx through the pore.

Nanodisc-cell fusion: control of fusion pore nucleation and lifetimes by SNARE protein transmembrane domains.

The initial, nanometer-sized connection between the plasma membrane and a hormone- or neurotransmitter-filled vesicle –the fusion pore– can flicker open and closed repeatedly before dilating or resealing irreversibly. Pore dynamics determine release and vesicle recycling kinetics, but pore properties are poorly known because biochemically defined single-pore assays are lacking. We isolated single flickering pores connecting v-SNARE-reconstituted nanodiscs to cells ectopically expressing cognate, “flipped” t-SNAREs. Conductance through single, voltage-clamped fusion pores directly reported sub-millisecond pore dynamics. Pore currents fluctuated, transiently returned to baseline multiple times, and disappeared ~6 s after initial opening, as if the fusion pore fluctuated in size, flickered, and resealed. We found that interactions between v- and t-SNARE transmembrane domains (TMDs) promote, but are not essential for pore nucleation. Surprisingly, TMD modifications designed to disrupt v- and t-SNARE TMD zippering prolonged pore lifetimes dramatically. We propose that the post-fusion geometry of the proteins contribute to pore stability.

Fusion Between v-SNARE Nanodiscs and “Flipped” t-SNARE Cells: Control of Fusion Pore Nucleation and Lifetimes by SNARE Protein Transmembrane Domains

A key step during neurotransmitter or hormone release (exocytosis) is the formation of a nanometer-sized fusion pore that connects the plasma membrane to the synaptic or secretory vesicle. The pore can flicker open and closed repeatedly before dilating or resealing irreversibly. Pore dynamics affect vesicle recycling, amount and kinetics of released cargo, and activation of downstream events. The fusion pore is clearly a key intermediate; however, factors regulating its dynamics are poorly understood, in large part due to a lack of biochemically defined single-pore assays.We have developed a novel assay that probes single fusion pores formed between ∼16 nm flat lipid bilayer nanodiscs (NDs) reconstituted with neuronal/exocytotic vesicular soluble N-ethylmaleimide-sensitive factor attachment protein receptor (v-SNARE) proteins and cells engineered to express cognate “flipped” target membrane t-SNAREs on their surfaces, with the SNARE domain facing the extracellular medium. Conductance through single, voltage-clamped fusion pores directly reports sub-millisecond pore dynamics. Pore currents fluctuated, briefly returned to baseline multiple times, and disappeared ∼6 s after initial opening, as if the fusion pore fluctuated in size, flickered, and resealed. We found that putative interactions between v- and t-SNARE transmembrane domains (TMDs) promote, but are not essential for pore nucleation. Surprisingly, the same modifications that reduced pore nucleation rates dramatically prolonged pore lifetimes. Our results indicate that rod-shaped, post-fusion cis-SNARE complexes that poorly fit the highly curved fusion site rapidly vacate the pore region, leaving pore properties to be determined by lipids. In contrast, mutants deficient in TMD-zippering prevented pore resealing for >60 s, possibly due to their post-fusion Y-shape. Thus, post-fusion geometry of the proteins may determine pore stability, analogous to the well-known effects of lipid geometry on highly curved fusion intermediates.

Control of Fusion Pore Nucleation and Dynamics by SNARE Protein Transmembrane Domains

Membrane fusion is a fundamental biological process, whose initial stages have been observed in hormone-secreting cells and neurons using electrophysiological and electrochemical methods. The initial connection between the plasma membrane and a hormone- or neurotransmitter-filled vesicle -the fusion pore- can flicker open and closed repeatedly before dilating or resealing irreversibly. Pore dynamics determine events such as vesicle recycling and release kinetics, but pore properties are poorly known, because fusion pores are transient, and biochemically defined assays with single-pore sensitivity are lacking. We isolated single flickering pores connecting v-SNARE-reconstituted nanodiscs to cells ectopically expressing cognate, “flipped” t-SNAREs, voltage-clamped in the cell-attached configuration. Currents through such pores directly report sub-millisecond single-pore dynamics. We found that interactions between v- and t-SNARE transmembrane domains (TMDs) observed in a recent crystal structure promote, but are not essential for pore nucleation. Surprisingly, TMD interactions also affected pore lifetimes. Rod-shaped, post-fusion cis-SNARE complexes vacated the highly curved fusion site where they fit poorly, leaving pore properties to be determined largely by lipid bilayers. In contrast, Y-shaped mutants deficient in TMD-zippering lingered at the fusion site, preventing pore resealing for >60 s. Thus, post-fusion geometry of the proteins determines pore stability, analogous to the well-known effects of lipid geometry on highly curved fusion intermediates.

Soft perforation of cardiolipin-containing planar lipid bilayer membrane by cytochrome c and H(2)O(2).

The release of cytochrome c (cyt c) from mitochondria is responsible for initiation of cell apoptosis. Although extramitochondrial proteins are thought to initiate this release, the exact mechanism remains unclear. Cyt c binds to and penetrates lipid bilayer membranes of specific phospholipid cardiolipin (CL) contained in mitochondria. We present here the experimental results of monitoring planar BLM (pBLM) from mixtures of azolectin and of CL (4/1 by moles) by triangle voltage pulses of 100 mV in amplitude and frequency of 2 Hz. The BLM were modified by a successive addition of cyt c and of H(2)O(2) in water solution. It is shown that the addition of cyt c alone leads to a stepwise increase in the ionic conductance of the pBLM, indicating the appearance of transmembrane pores. Pore lifetimes then reached several seconds at an average pore diameter of ~2 nm. Current-voltage characteristics were then linear and passed through the origin which is characteristic for broad, nonselective ion pores. Subsequent addition of H(2)O(2) caused a dramatic increase in transmembrane current at retention of average pore size constant. Observed increase in membrane current is due to growth of a number of pores in an open state. We suggest that hydrogen peroxide in the presence of cyt c promotes a peroxidation of membrane phospholipids to form lysolipids, the embedding of which stabilizes the edge of the pore and the surface of lipid bilayer.

Effect of arginine-rich cell penetrating peptides on membrane pore formation and life-times: a molecular simulation study.

The molecular basis for the effectiveness of arginine-rich cell penetrating peptides (ARCPPs) traversing a cell membrane barrier is not well established. The fact that a threshold concentration of ARCPPs is required for efficient translocation in model membranes suggests cooperative action by ARCPPs. We used umbrella sampling simulations to calculate the free energies for membrane pore formation. Membrane-bound octaarginine (ARG8) peptides showed little cooperativity in lowering the free energy barrier to generate membrane pores by direct peptide translocation or by lipid flip-flop. Instead, high concentrations of ARG8 peptides were found to expand the surface area of the lipid bilayer due to the deep partitioning of guanidinium ions into the lipid glycerol regions. Surface-bound ARG8 peptides can also insert an arginine side chain into one existing transient membrane pore, and the lifetime of the transient membrane pore is significantly extended by arginine. This suggests a cooperative kinetic mechanism may act above a threshold adsorption concentration to facilitate the rapid uptake of these peptides.

Atomistic Simulations of Pore Formation and Closure in Lipid Bilayers

Cellular membranes separate distinct aqueous compartments, but can be breached by transient hydrophilic pores. A large energetic cost prevents pore formation, which is largely dependent on the composition and structure of the lipid bilayer. The softness of bilayers and the disordered structure of pores make their characterization difficult. We use molecular-dynamics simulations with atomistic detail to study the thermodynamics, kinetics, and mechanism of pore formation and closure in DLPC, DMPC, and DPPC bilayers, with pore formation free energies of 17, 45, and 78 kJ/mol, respectively. By using atomistic computer simulations, we are able to determine not only the free energy for pore formation, but also the enthalpy and entropy, which yields what is believed to be significant new insights in the molecular driving forces behind membrane defects. The free energy cost for pore formation is due to a large unfavorable entropic contribution and a favorable change in enthalpy. Changes in hydrogen bonding patterns occur, with increased lipid-water interactions, and fewer water-water hydrogen bonds, but the total number of overall hydrogen bonds is constant. Equilibrium pore formation is directly observed in the thin DLPC lipid bilayer. Multiple long timescale simulations of pore closure are used to predict pore lifetimes. Our results are important for biological applications, including the activity of antimicrobial peptides and a better understanding of membrane protein folding, and improve our understanding of the fundamental physicochemical nature of membranes.

F-Actin and Myosin II Accelerate Catecholamine Release from Chromaffin Granules

The roles of nonmuscle myosin II and cortical actin filaments in chromaffin granule exocytosis were studied by confocal fluorescence microscopy, amperometry, and cell-attached capacitance measurements. Fluorescence imaging indicated decreased mobility of granules near the plasma membrane following inhibition of myosin II function with blebbistatin. Slower fusion pore expansion rates and longer fusion pore lifetimes were observed after inhibition of actin polymerization using cytochalasin D. Amperometric recordings revealed increased amperometric spike half-widths without change in quantal size after either myosin II inhibition or actin disruption. These results suggest that actin and myosin II facilitate release from individual chromaffin granules by accelerating dissociation of catecholamines from the intragranular matrix possibly through generation of mechanical forces.

Myosin 2 Maintains an Open Exocytic Fusion Pore in Secretory Epithelial Cells

Many studies have implicated F-actin and myosin 2 in the control of regulated secretion. Most recently, evidence suggests a role for the microfilament network in regulating the postfusion events of vesicle dynamics. This is of potential importance as postfusion behavior can influence the loss of vesicle content and may provide a new target for drug therapy. We have investigated the role of myosin 2 in regulating exocytosis in secretory epithelial cells by using novel assays to determine the behavior of the fusion pore in individual granules. We immunolocalize myosin 2A to the apical region of pancreatic acinar cells, suggesting it is this isoform that plays a role in granule exocytosis. We further show myosin 2 phosphorylation increased on cell stimulation, consistent with a regulatory role in secretion. Importantly, in a single-cell, single-granule secretion assay, neither the myosin 2 inhibitor (-)-blebbistatin nor the myosin light chain kinase inhibitor ML-9 had any effect on the numbers of granules stimulated to fuse after cell stimulation. These data indicate that myosin 2, if it has any action on secretion, must be targeting postfusion granule behavior. This interpretation is supported by direct study of fusion pore opening in which we show that (-)-blebbistatin and ML-9 promote fusion pore closure and decrease fusion pore lifetimes. Our work now adds to a growing body of evidence showing that myosin 2 is an essential regulator of postfusion granule behavior. In particular, in the case of the secretory epithelial cells, myosin 2 activity is necessary to maintain fusion pore opening.

Active Mechanisms Are Needed to Describe Cell Responses to Submicrosecond, Megavolt-per-Meter Pulses: Cell Models for Ultrashort Pulses

Intracellular effects of submicrosecond, megavolt-per-meter pulses imply changes in a cell's plasma membrane (PM) and organelle membranes. The maximum reported PM transmembrane voltage is only 1.6 V and phosphatidylserine is translocated to the outer membrane leaflet of the PM. Passive membrane models involve only displacement currents and predict excessive PM voltages (∼25 V). Here we use a cell system model with nonconcentric circular PM and organelle membranes to demonstrate fundamental differences between active (nonlinear) and passive (linear) models. We assign active or passive interactions to local membrane regions. The resulting cell system model involves a large number of interconnected local models that individually represent the 1), passive conductive and dielectric properties of aqueous electrolytes and membranes; 2), resting potential source; and 3), asymptotic membrane electroporation model. Systems with passive interactions cannot account for key experimental observations. Our active models exhibit supra-electroporation of the PM and organelle membranes, some key features of the transmembrane voltage, high densities of small pores in the PM and organelle membranes, and a global postpulse perturbation in which cell membranes are depolarized on the timescale of pore lifetimes.

The Amplitude Distribution of Release Events through a Fusion Pore

Neurotransmitters, hormones, or dyes may be released from vesicles via a fusion pore, rather than by full fusion of the vesicle with the plasma membrane. If the lifetime of the fusion pore is comparable to the time required for the substance to exit the vesicle, only a fraction of the total vesicle content may be released during a single pore opening. Assuming 1), fusion pore lifetimes are exponentially distributed ( τ P), as expected for simple single channel openings, and 2), vesicle contents are lost through the fusion pore with an exponential time course ( τ D), we derive an analytical expression for the probability density function of the fraction of vesicle content released ( F): dP/ dF = A (1 − F) (A-1), where A = τ D/ τ P. If A > 1, the maximum of the distribution is at F = 0; if A < 1, the maximum is at F = 1; if A = 1, the distribution is perfectly flat. Thus, the distribution never has a peak in the middle (0 < F < 1). This should be considered when interpreting the distribution of miniature synaptic currents, or the fraction of FM dye molecules lost during a single fusion pore opening event.

Transient pores in vesicles

AbstractWe present our observations of transient pores in giant unilamellar vesicles, placed under tension, by optical illumination. When the membrane tension reached a certain level, transient pores appeared. Pore opening is driven by the membrane tension, σ, and its closure by the pore's line tension, 𝒯. By use of viscous mixtures of glycerol and water, we slowed down the leak out of the inner liquid in the presence of a pore. This allowed pores to reach large sizes (a few micrometres) and last at least a few seconds so that they could be visualized by fluorescence videomicroscopy. Line tension was inferred from the measurements of the closure velocity of the pores. By addition of cholesterol, which increased 𝒯 (reducing pore lifetimes), or of surfactants, which decreased 𝒯 (increasing pore lifetimes), we demonstrate how 𝒯, and consequently pore lifetimes, can be controlled over nearly two orders of magnitude. Addition of surfactants also has a dramatic effect on vesicle fusion. We discuss how our results can be extended to less viscous aqueous solutions which are more relevant for liposomal drug delivery formulations.© 2003 Society of Chemical Industry

Cascades of Transient Pores in Giant Vesicles: Line Tension and Transport

Under ordinary circumstances, the membrane tension of a giant unilamellar vesicle is essentially nil. Using visible light, we stretch the vesicles, increasing the membrane tension until the membrane responds by the sudden opening of a large pore (several micrometers in size). Only a single pore is observed at a time in a given vesicle. However, a cascade of transient pores appear, up to 30–40 in succession, in the same vesicle. These pores are transient: they reseal within a few seconds as the inner liquid leaks out. The membrane tension, which is the driving force for pore opening, is relaxed with the opening of a pore and the leakage of the inner liquid; the line tension of the pore’s edge is then able to drive the closure of a pore. We use fluorescent membrane probes and real-time videomicroscopy to study the dynamics of the pores. These can be visualized only if the vesicles are prepared in a viscous solution to slow down the leakout of the internal liquid. From measurements of the closure velocity of the pores, we are able to infer the line tension, T . We have studied the effect of the shape of inclusion molecules on T . Cholesterol, which can be modeled as an inverted cone-shaped molecule, increases the line tension when incorporated into the bilayers. Conversely, addition of cone-shaped detergents reduces T . The effect of some detergents can be dramatic, reducing T by two orders of magnitude, and increasing pore lifetimes up to several minutes. We give some examples of transport through transient pores and present a rough measurement of the leakout velocity of the inner liquid through a pore. We discuss how our results can be extended to less viscous aqueous solutions which are more relevant for biological systems and biotechnological applications.

Molecular dynamics study of water pores in a phospholipid bilayer

A molecular dynamics (MD) study of a hydrated phospholipid bilayer is reported. Therein the stability of initially formed water pores in phospholipid membranes is investigated by means of computer experiments based on a microscopic membrane model. The pore lifetimes were found to be in the range of 100 ps. The probability of finding water molecules in the hydrophobic interior of the membrane is discussed. Our results give evidence for the existence of water pores in phospholipid membranes which are sufficiently thin.