Analysis of the source of heterogeneity in the osmotic response of plant membrane vesicles

Plasma membrane vesicles have been widely employed to understand the biophysics of water movements, especially when active aquaporins are present. In general, water permeability coefficients in these preparations outcome from the analysis of the osmotic response of the vesicles by means of light scattering. As from now, this is possible by following a theoretical approach that assumes that scattered light follows a single exponential function and that this behavior is the consequence of vesicle volume changes due to an osmotic challenge. However, some experimental data do not necessarily fit to single exponentials but to double ones. It is argued that the observed double exponential behavior has two possible causes: different vesicle population in terms of permeability or in terms of size distribution. As classical models cannot identify this source of heterogeneity, a mathematical modeling approach was developed based on phenomenological equations of water transport. In the three comparative models presented here, it was assumed that water moves according to an osmotic mechanism across the vesicles, and there is no solute movement across them. Interestingly, when tested in a well described plasma membrane vesicle preparation, the application of these models indicates that the source of heterogeneity in the osmotic response is vesicles having different permeability, clearly discarding the variable size effect. In conclusion, the mathematical approach presented here allows to identify the source of heterogeneity; this information being of particular interest, especially when studying gating mechanisms triggered in water channel activity.     

Determination of solute reflection coefficients in kidney brush-border membrane vesicles by light scattering: influence of the refractive index

Solute reflection coefficients sigma of cell membrane vesicles or liposomes are commonly determined by comparison of the water flow induced by a gradient of the studied solute and that of a reference molecule using light scattering techniques. However, variations in scattered light which are mainly related to change in vesicle volume are also influenced by the refractive index of the surrounding medium. Therefore comparing kinetics of vesicle shrinkage induced by hyperosmotic solutions which have different refractive indexes might lead to an under or over estimation of sigma. We determined sigma NaCl in rat kidney brush-border membrane vesicles by two different approaches using mannitol, a poorly permeant molecule, as reference. (1) The refractive index of the hyperosmotic NaCl solution was adjusted to that of mannitol by addition of polyvinyl pyrrolidone (Mr 40,000), without a significant increase in osmolality. Thereby the change in scattered light intensity induced by osmotic vesicle shrinkage due to gradients of NaCl and mannitol were comparable and led to a sigma NaCl value close to one instead of the previously published value of 0.53. (2) The reflection coefficient was calculated from the lifetime of vesicle shrinkage which is not refractive index-dependent. Again sigma NaCl was not different from one. These results suggest that the water proteic pathways found in the luminal membrane of proximal tubules are not shared by salts.     

Evidence for proteic water pathways in the luminal membrane of kidney proximal tubule

The osmotic permeability of the apical membrane of proximal tubule cells was studied on rat brush-border membrane vesicles by following their rate of shrinkage with a stopped-flow device coupled to light transmission recording. The mercuric sulfhydryl reagent para-chloromercuribenzenesulfonic acid (PCMBS) reduced the water permeability of the membrane, in a time- and dose-dependent manner, to 35% of the control value. Mercuric chloride was a more potent inhibitor and decreased the osmotic water permeability of the brush-border membrane to 15% of the control. This inhibition was reversed by an excess of cysteine, while cysteine per se did not modify the rate of vesicle shrinkage. These results suggest that most of the osmotic water movements across kidney brush-border membranes are through polar pathways which involve the integrity of the membrane proteins.     

Proton decay kinetics for vesicles containing buffers—an analytical solution

The problem of predicting the kinetics of proton efflux and the decay of the internal proton concentration for vesicles containing one or more buffers for which the internal proton concentration is initially higher than that of the surrounding medium is examined. An analytical solution is derived that describes the time course of the proton efflux from vesicles and the decay of the internal proton concentration under conditions of zero transmembrane electric potential. The effect of the internal buffers is to increase the time required for the proton concentration gradient to equilibrate across the membrane. To simplify the analysis we assume that the equilibration of the internal and external proton activity is due primarily to proton diffusion through the membrane, and not to hydroxyl ion flux. For a vesicle containing a single buffer the solution requires six independent physical parameters: the initial internal proton concentration, the external proton concentration, the ratio of the vesicle surface area to the internal volume, the permeability coefficient of the membrane for protons, the total concentration of the internal buffer, and the equilibrium constant for the dissociation of the internal buffer. Determination of these physical values is sufficient to predict the time dependence of the internal proton concentration and of the proton efflux. Over a pH range that is below or near the pK of the internal buffer the solution is complex. However, if the initial pH is one unit or more higher than the pK of the internal buffer the kinetics of the internal proton concentration and proton efflux can be described by a pseudo first order reaction. In this case the apparent rate constant depends linearly on the permeability coefficient and is dominated by the total internal buffer concentration and its pK. For example, increasing the internal buffer concentration inside a vesicle by 10-fold results in an approximately 10-fold increase in the half-time of the proton efflux kinetics. The theoretical analysis is applied to thylakiod vesicles using experimentally determined values for the physical parameters. The predictions of the analysis are compared to experimentally observed kinetics.     

PIP1 aquaporins,sterols, and osmotic water permeability of plasma membranes from etiolated pea seedlings

Plasma membrane isolated from microsomal membranes of pea seedling root and shoot cells by means of aqueous two-phase polymer system was separated by flotation in discontinuous OptiPrep gradient into “light” (≤1.146 g/cm³) and “heavy” (≥1.146 g/cm³) fractions. Osmotic water permeability of plasma membrane and its two fractions was investigated by inducing transmembrane osmotic gradient on the vesicle membrane and recording the kinetics of vesicle osmotic shrinkage by the stopped-flow method. Rate constants of osmotic shrinkage and coefficients of osmotic water permeability of the membranes were estimated on the basis of the kinetic curve approximation by exponential dependencies and using electron microscope data on vesicles sizes. In plasma membrane and its fractions the content of sterols and PIP1 aquaporins was determined. It was found that in “light” PM fractions from both roots and shoots the content of PIP1 aquaporins and sterols was higher and the osmotic water permeability coefficient was lower than in “heavy” fractions of plasma membrane. The results indicate that plasma membrane of roots and shoots is heterogeneous in osmotic water permeability. This heterogeneity may be related with the presence of microdomains with different content of aquaporins and sterols in the membrane.     

Membrane water and solute permeability determined quantitatively by self-quenching of an entrapped fluorophore

Quantitative determination of rapid water and solute transport and solute reflection coefficients by light-scattering methods is complicated by dependence of vesicle or cell light scattering on nonvolume factors including solution refractive index, cell motion, and membrane aggregation. To overcome these difficulties, a fluorescence technique has been developed to measure accurately (1) osmotic water permeability (Pf), (2) solute permeability (Ps), and (3) solute reflection coefficient (sigma). The time course of vesicle volume is determined by the self-quenching of entrapped fluorescein sulfonate (FS), the best of a series of dyes screened for self-quenching, brightness, and vesicle loading/trapping. To validate the method, rabbit renal brush border vesicles (BBV) were loaded with 1-10 mM FS for 12 h at 4 degrees C and washed to remove extravesicular FS. FS leakage occurred over greater than 6 h at 4 degrees C and greater than 30 min at 23 degrees C. FS fluorescence vs vesicle volume was calibrated from the time course of fluorescence decrease (excitation 470 nm, emission greater than 515 nm) in response to a series of inward osmotic gradients in a stopped-flow apparatus. At 23 degrees C Pf was 0.005 +/- 0.001 cm/s, independent of osmotic gradient size, and inhibited 67% by 0.5 mM HgCl2. Urea Ps was 2 x 10(-6) cm/s with sigma 0.95-1.00 on the basis of the fluorescence time course analysis and the extravesicular [urea] required to obtain zero initial volume flow (null method) when vesicles were loaded with sucrose.(ABSTRACT TRUNCATED AT 250 WORDS)     

Inhibition of anion permeability of sarcoplasmic reticulum vesicles by 4-acetoamido-4'-isothiocyanostilbene-2,2'-disulfonate

The permeabilities of sarcoplasmic reticulum vesicle membrane for various ions and neutral molecules were measured by following the change in light scattering intensity due to the osmotic volume change of the vesicles. 4-Acetoamido-4'-isothiocyanostilbene-2,2'-disulfonate (SITS), which is a potent inhibitor for the anion permeability of red blood cells membrane, inhibited the permeability of sarcoplasmic reticulum for anions such as Cl-, Pi and methanesulfonate, while it slightly increased that for cations and neutral molecules such as Na+, K+, choline and glycerol. Binding of 5 mumol SITS/g protein was necessary for the inhibition of anion permeability. These results suggest the existence of a similar anion transport system in sarcoplasmic reticulum membrane as revealed in red blood cell membrane.     

Influence of the growth at high osmolality on the lipid composition, water permeability and osmotic response of Lactobacillus bulgaricus

Changes in water permeability and membrane packing were measured in cells of Lactobacillus bulgaricus and in vesicles prepared with lipids extracted from them. The osmotic response of whole cells and vesicles is compared with the one of bacteria grown in a high osmolal medium. Both bacteria and vesicles, behave as osmometers. This means that the volume decrease is promoted by the outflow of water, driven by the NaCl concentration difference, arguing that neither Na+ nor Cl- permeates the cell or the lipid membrane in these conditions. Therefore, the volume changes can be correlated with the rate of water permeation across the cell or the vesicle membranes. The permeation of water was analyzed as a function of the lipid species by measuring the volume changes and the saturation ratio of the lipids. To put into relevance the membrane processes, the permeation properties of lipid vesicles prepared with lipids extracted from bacteria grown in normal and high osmolality conditions were also analyzed. The permeation response was correlated with the physical properties of the membrane of whole cells and vesicles, by means of fluorescence anisotropy of diphenyl hexatriene (DPH). The modifications in membrane properties are related with the changes in the membrane composition triggered by the growth in a high osmolal medium. The changes appear related to an increase in the sugar content of the whole pool of lipids and in the saturated fatty acid residues.     

Semipermeability of the nuclear membrane in the intact cell

The osmotic properties of nuclei in intact cells were studied by injecting solutions into the cytoplasm of amphibian oocytes. Subsequent changes in nuclear volume were recorded photographically. The injection of solutions containing polyvinylpyrrolidone or bovine serum albumin caused changes in nuclear volume which were related to the colloid osmotic pressure of the solution injected. The concentration in which no significant nuclear volume change occurred (the isotonic range) was 1.0 to 1.5 per cent polyvinylpyrrolidone (2.0 to 3.75 x 10(-4)M). 2 per cent bovine serum albumin had no significant effect on nuclear volume, whereas 4 per cent caused a significant decrease. The significance of these findings is discussed in terms of the permeability characteristics of the nuclear membrane.     

Water permeabilities and salt reflection coefficients of luminal, basolateral and intracellular membrane vesicles isolated from rabbit kidney proximal tubule

The mechanisms of water transport across the rabbit renal proximal convoluted tubule were approached by measuring osmotic permeabilities and solute reflection coefficients of the brush-border and the basolateral membranes. Plasma and intracellular membrane vesicles were isolated from rabbit renal cortex by centrifugation on a Percoll gradient. Three major turbidity bands were obtained: a fraction of purified basolateral membranes (BLMV), the two others being brush-border (BBMV) and endoplasmic reticulum (ERMV) membrane vesicles. The osmotic permeability (Pf) of the three types of vesicle was measured using stop-flow techniques and their geometry was determined by quasi-elastic light scattering. Pf was equal to 123 +/- 8 microns/s (n = 10) for BBMV, 166 +/- 10 microns/s (n = 10) for BLMV and 156 +/- 9 microns/s (n = 4) for ERMV (T = 26 degrees C). A transcellular water permeability, per unit of apical surface area, of 71 microns/s was calculated considering that the luminal and the basolateral membranes act as two conductances in series. This value is in close agreement, after appropriate normalizations, with previously reported transepithelial water permeabilities obtained using in vitro microperfusion techniques thus supporting the hypothesis of a predominantly transcellular route for water flow across rabbit proximal convoluted tubule. The addition of 0.4 mM HgCl2, a sulfhydryl reagent, decreased Pf about 60% in three types of membrane providing evidence for the existence of proteic pathways. NaCl and KCl reflection coefficients were measured and found to be close to one for plasma and intracellular membranes suggesting that the water channels are not shared by salts.     

Effects of divalent cations,temperature, osmotic pressure gradient,and vesicle curvature on phosphatidylserine vesicle fusion

Summary Fusion of phosphatidylserine vesicles induced by divalent cations, temperature and osmotic pressure gradients across the membrane was studied with respect to variations in vesicle size. Vesicle fusion was followed by two different methods: 1) the Tb/DPA fusion assay, whereby the fluorescent intensity upon mixing of the internal aqueous contents of fused lipid vesicles was monitored, and 2) measurement of the changes in turbidity of the vesicle suspension due to vesicle fusion. It was found that the threshold concentration of divalent cations necessary to induce vesicle fusion depended on the size of vesicles; as the diameter of the vesicle increased, the threshold value increased and the extent of fusion became less. For the osmotic pressure-induced vesicle fusion, the larger the diameter of vesicles, the smaller was the osmotic pressure gradient required to induce membrane fusion. Divalent cations, temperature increase and vesicle membrane expansion by osmotic pressure gradient all resulted in increase in surface energy (tension) of the membrane. The degree of membrane fusion correlated with the corresponding surface energy changes of vesicle membranes due to the above fusion-inducing agents. The increase in surface energy of 9.5 dyn/cm from the reference state corresponded to the threshold point of phosphatidylserine membrane fusion. An attempt was made to explain the factors influencing fusion phenomena on the basis of a single unifying theory.     

Inhibition of anion permeability of sarcoplasmic reticulum vesicles by 4-acetoamido-4′-isothiocyanostilbene-2,2′-disulfonate

The permeabilities of sarcoplasmic reticulum vesicle membrane for various ions and neutral molecules were measured by following the change in light scattering intensity due to the osmotic volume change of the vesicles. 4-Acetoamido-4′-isothiocyanostilbene-2,2′-disulfonate (SITS), which is a potent inhibitor for the anion permeability of red blood cells membrane, inhibited the permeability of sarcoplasmic reticulum for anions such as Cl^?, P_i and methanesulfonate, while it slightly increased that for cations and neutral molecules such as Na⁺, K⁺, choline and glycerol. Binding of 5μmol SITS/g protein was necessary for the inhibition of anion permeability. These results suggest the existence of a similar anion transport system in sarcoplasmic reticulum membrane as revealed in red blood cell membrane.     

Macromolecule entrapment in human placental microvillous membrane vesicles

An osmotic lysis technique was developed to induce transient permeability in human placental microvillous membrane vesicles. The degree of vesicle opening and resealing was quantitated using the fluorescent markers, 6-carboxyfluorescein and fluorescein dextran. Compared to freeze-thaw and sonication methods, hypotonic lysis was significantly more efficient, causing greater than 90% lysis with greater than 90% subsequent resealing under optimal conditions. The transient increase in vesicle permeability permitted the unrestricted entry of macromolecules with molecular masses up to 70,000 kDa. Passive transport of water, protons, and erythritol and carrier-mediated transport of L-valine and sodium-proton exchange were unaltered by the lysis/resealing procedure. Bovine tracheal vesicles were lysed to an extent similar to placental microvillous vesicles, but rabbit renal cortical brush border and basolateral membranes were lysed to a lesser extent (approximately 60%). These results show that hypotonic lysis is a suitable method for the loading and trapping of macromolecules in isolated membrane vesicles for studies of intracellular regulation of transport.     

Fusion of phospholipid vesicles with a planar membrane depends on the membrane permeability of the solute used to create the osmotic pressure

Phospholipid vesicles fuse with a planar membrane when they are osmotically swollen. Channels in the vesicle membrane are required for swelling to occur when the vesicle-containing compartment is made hyperosmotic by adding a solute (termed an osmoticant). We have studied fusion using two different channels, porin, a highly permeable channel, and nystatin, a much less permeable channel. We report that an osmoticant's ability to support fusion (defined as the magnitude of osmotic gradient necessary to obtain sustained fusion) depends on both its permeability through lipid bilayer as well as its permeability through the channel by which it enters the vesicle interior. With porin as the channel, formamide requires an osmotic gradient about ten times that required with urea, which is approximately 1/40th as permeant as formamide through bare lipid membrane. When nystatin is the channel, however, fusion rates sustained by osmotic gradients of formamide are within a factor of two of those obtained with urea. Vesicles containing a porin-impermeant solute can be induced to swell and fuse with a planar membrane when the impermeant bathing the vesicles is replaced by an isosmotic quantity of a porin-permeant solute. With this method of swelling, formamide is as effective as urea in obtaining fusion. In addition, we report that binding of vesicles to the planar membrane does not make the contact region more permeable to the osmoticant than is bare lipid bilayer. In the companion paper, we quantitatively account for the observation that the ability of a solute to promote fusion depends on its permeability properties and the method of swelling. We show that the intravesicular pressure developed drives fusion.     

Membrane stress causes inhibition of water channels in brush border membrane vesicles from kidney proximal tubule

Brush border membrane vesicles (BBMV) from rabbit kidney proximal tubule cells, prepared with different internal solute concentrations (cellobiose buffer 13, 18 or 85 mosM) developed an hydrostatic pressure difference across the membrane of 18.7 mosM, that causes a membrane tension close to 5 × 10⁻⁵ N cm⁻¹. When subjected to several hypertonic osmotic shocks an initial delay of osmotic shrinkage (a lag time), corresponding to a very small change in initial volume was apparent. This initial osmotic response, which is significantly retarded, was correlated with the initial period of elevated membrane tension, suggesting that the water permeability coefficient is inhibited by membrane stress. We speculate that this inhibition may serve to regulate cell volume in the proximal tubule.     

Permeability to ions of bovine retinal disk membrane vesicles in the bleached state

The permeability of the bleached disk membrane of retinal rod outer segments to univalent and divalent ions is studied by light scattering. The membranes are isolated from frozen dark-adapted bovine retinae, swollen into spherical vesicles in a hypotonic medium and bleached in dilute suspension and their size is determined by elastic and quasi-elastic light scatterings. Various electrolytes are then added to the suspending medium in order to examine their osmotic activity relative to the vesicles deformation characteristics. By following the deformation behavior of the membrane vesicles by elastic light scattering in terms of the oblate ellipsoidal shell model, the osmotic activity of a given electrolyte is qualitatively deduced and thereby the permeability of the membrane to the electrolyte is ranked in reference to a chosen standard, i.e., sucrose. By this method, we show that the permeabilities to Na⁺, K⁺, Mg²⁺ and Ca²⁺ are all alike, and those to halides (F^?, Cl^?, Br^?, I^?), nitrate and phosphates (HPO₄^2?/H₂PO₄^?) are similar. Acetate, however, is about 3-times more permeative, while sulfate is less permeative than the other anions by about the same factor. The viability of our method is checked with use of an ionophore, lasolocid (X-537A), by establishing partial recovery from the osmotic deformation through the suppression of the cation osmotic effect. Ion-induced aggregation and pH-dependent size and shape changes are both found to be insignificant.     

Cell volume and plasma membrane osmotic water permeability in epithelial cell layers measured by interferometry.

The development of strategies to measure plasma membrane osmotic water permeability (Pf) in epithelial cells has been motivated by the identification of a family of molecular water channels. A general approach utilizing interferometry to measure cell shape and volume was developed and applied to measure Pf in cell layers. The method is based on the cell volume dependence of optical path length (OPL) for a light beam passing through the cell. The small changes in OPL were measured by interferometry. A mathematical model was developed to relate the interference signal to cell volume changes for cells of arbitrary shape and size. To validate the model, a Mach-Zehnder interference microscope was used to image OPL in an Madin Darby Canine Kidney (MDCK) cell layer and to reconstruct the three-dimensional cell shape (OPL resolution < lambda/25). As predicted by the model, a doubling of cell volume resulted in a change in OPL that was proportional to the difference in refractive indices between water and the extracellular medium. The time course of relative cell volume in response to an osmotic gradient was computed from serial interference images. To measure cell volume without microscopy and image analysis, a Mach-Zehnder interferometer was constructed in which one of two interfering laser beams passed through a flow chamber containing the cell layer. The interference signal in response to an osmotic gradient was analyzed to quantify the time course of relative cell volume. The calculated MDCK cell plasma membrane Pf of 6.1 x 10(-4) cm/s at 24 degrees C agreed with that obtained by interference microscopy and by a total internal reflection fluorescence method. Interferometry was also applied to measure the apical plasma membrane water permeability of intact toad urinary bladder; Pf increased fivefold after forskolin stimulation to 0.04 cm/s at 23 degrees C. These results establish and validate the application of interferometry to quantify cell volume and osmotic water permeability in cell layers.     

Rapid determination of internal volumes of membrane vesicles with electron spin resonance-stopped flow technique

We have developed an electron spin resonance (ESR)-stopped flow technique and employed it for the simple and rapid determination of internal volumes of biomembrane vesicles and liposomes. A vesicle suspension containing a neutral and membrane-permeable spin label, 2,2,6,6-tetramethyl-4-oxopiperidine-1-oxyl (TEMPONE), was mixed in the stopped-flow apparatus with an isotonic solution of relatively impermeable line broadening agents, potassium tris(oxalato)chromate(III) or potassium ferricyanide, and an ESR spectrum was recorded. From the relative intensity of the sharp triplet signal due to TEMPONE in the aqueous space within vesicles, the determination of the internal aqueous volume was straightforward. Using this technique, it is possible to measure intravesicular volumes in 0.1 s. The internal volume of sonicated phospholipid vesicles was approximately 0.3 microliter/mg lipid. The light fraction of sarcoplasmic reticulum membrane vesicles isolated from rabbit skeletal muscle was estimated to have an internal volume of 2.2-2.6 microliter/mg protein in its resting state. Activation of Ca2+ pumps in the membrane upon addition of ATP and Ca2+ ions decreased the internal volume by about 10%. This finding supports the hypothesis that the Ca2+ pump is electrogenic and that the efflux of potassium ions compensates for the influx of positive charges. The present technique is widely applicable to the simple and rapid determination of the internal volumes of membrane vesicles.     

A theory of osmotic lysis of lipid vesicles

Osmotic lysis of vesicles is shown to begin when the membrane expansion due to osmotic pressure exceeds its critical value, delta S, at which a membrane ruptures to form a pore. The dependence of delta S on the vesicle radius and respective osmotic pressures are obtained. It is found that osmotic pressure necessary for small (100 A) vesicles to rupture should exceed 30 atm, for large (10 000 A) vesicles it being as small as 10(-3) atm. In the case of large (greater than or approximately 1000 A) vesicles the value of relative expansion of the membrane at which its rupture occurs in a reasonable time only depends slightly on the vesicle radius. For instance, for 10 000 A vesicles it amounts to 3%. The tension of membrane rupture is about 8 dyn/cm for large vesicles. Membrane tension, although it decreases considerably as a result of rupture and pore formation, does not vanish completely. It supports the residual intravesicular pressure causing the efflux of vesicle (cell) contents. Simultaneously, osmotic influx of water through the membrane occurs that results in either complete rupture of the membrane with the efflux of the whole of the contents, or its gradual washout in either of two, quasi-steady or pulse-wise regimes. In the first case a pore is steadily open, whereas in the second case it alternately opens and closes, ejecting about 5% of internal solution each time. Lysis kinetics is analyzed. Pulse-wise regime of lysis is shown to be the most likely one.     

Transmembrane flux and receptor desensitization measured with membrane vesicles. Homogeneity of vesicles investigated by computer simulation.

The use of membrane vesicles to make quantitative studies of transmembrane transport and exchange processes involves an assumption of homogeneity of the membrane vesicles. In studies of 86Rb+ exchange mediated by acetylcholine receptor from the electric organ of Electrophorus electricus and of 36Cl- exchange mediated by GABA receptor from rat brain, measurements of ion exchange and receptor desensitization precisely followed first order kinetics in support of this assumption. In other measurements a biphasic decay of receptor activity was seen. To elucidate the molecular properties of receptors from such measurements it is important to appreciate what the requirements of vesicle monodispersity are for meaningful results and what the effect of vesicle heterogeneity would be. The experiments were simulated with single vesicle populations with variable defined size distributions as well as with mixtures of different populations of vesicles. The properties of the receptors and their density in the membrane could be varied. Different receptors could be present on the same or different membrane vesicles. The simulated measurements were not very sensitive to size dispersity. A very broad size distribution of a single vesicle population was necessary to give rise to detectable deviations from first order kinetics or errors in the determined kinetic constants. Errors could become significant with mixtures of different vesicle populations, where the dispersity in initial ion exchange rate constant, proportional to the receptor concentration per internal volume, became large. In this case the apparent rate of receptor desensitization would diverge in opposite directions from the input value when measured by two different methods, suggesting an experimental test for such kinetic heterogeneity. A biphasic decrease of receptor activity could not be attributed to vesicle heterogeneity and must be due to desensitization processes with different rates. Significant errors would not arise from the size dispersity apparent in subpopulations of vesicles seen by imaging techniques in membrane preparations.