The distribution of anionic sites on the surfaces of mitochondrial membranes. Visual probing with polycationic ferritin

Polycationic ferritin, a multivalent ligand, was used as a visual probe to determine the distribution and density of anionic sites on the surfaces of rat liver mitochondrial membranes. Both the distribution of bound polycationic ferritin and the topography of the outer surface of the inner mitochondrial membrane were studied in depth by utilizing thin sections and critical-point dried, whole mount preparations for transmission electron microscopy and by scanning electron microscopy. Based on its relative affinity for polycationic ferritin, the surface of the inner membrane contains discrete regions of high density and low density anionic sites. Whereas the surface of the cristal membrane contains a low density of anionic sites, the surface of the inner boundary membrane contains patches of high density anionic sites. The high density anionic sites on the inner boundary membrane were found to persist as stable patches and did not dissociate or randomize freely when the membrane was converted osmotically to a spherical configuration. The observations suggest that the inner mitochondrial membrane is composed of two major regions of anionic macromolecular distinction. It is well-known that an intermembrane space exists between the two membranes of the intact mitochondrion; however, a number of contact sites occur between the two membranes. We determined that the outer membrane, partially disrupted by treatment with digitonin, remains attached to the inner membrane at these contact sites as inverted vesicles. Such attached vesicles show that the inner surface of the outer membrane contains anionic sites, but of decreased density, surrounding the contact sites. Thus, the intermembrane space in the intact mitochondrion may be maintained by electronegative surfaces of the two mitochondrial membranes. The distribution of anionic sites on the outer surface of the outer membrane is random. The nature and function of fixed anionic surface charges and membrane contact sites are discussed with regard to recent reports relating to calcium transport, protein assembly into mitochondrial membranes, and membrane fluidity.     

Involvement of contact sites in phosphatidylserine import into liver mitochondria

The synthesis, translocation, and decarboxylation of phosphatidylserine can occur in a cell-free system (Voelker, D. R. (1989) J. Biol. Chem. 264, 8019-8025). We made use of the spatial separation of the site of biosynthesis and the site of decarboxylation of phosphatidylserine to demonstrate that mitochondrial contact sites are intimately involved in the translocation of phosphatidylserine prior to decarboxylation. In that sense, the inhibition of phosphatidylserine decarboxylase leads to an accumulation of this phospholipid in the contact site-enriched fractions without mixing the inner membrane phospholipid pool. On the other hand, newly synthesized phosphatidylethanolamine can be exported very rapidly to the mitochondrial surface in the same way, i.e. via contact sites. These data provide further evidence for the existence of a structural and functional microcompartmentation at the inner mitochondrial membrane surface.     

Identification of a GTP-binding protein in the contact sites between inner and outer mitochondrial membranes

Whilst investigating whether GTP hydrolysis may be required for the import of preproteins into mitochondria we have found that a GTP-binding protein is located at the contact sites between mitochondrial inner and outer membranes. When mitochondrial outer membranes purified from rat liver were UV-irradiated in the presence of [alpha-32P]GTP, a 52 kDa protein was radiolabelled, whereas [alpha-32P]ATP did not label this protein. GTP-binding proteins were also labelled in the cytosolic and microsomal fractions, but the 52 kDa protein was concentrated in mitochondrial membranes and was the only protein specifically labelled by GTP in these membranes. Fractionation of mitochondrial membrane vesicles into outer membranes, inner membranes and contact sites between outer and inner membranes showed that the GTP-binding activity was highly enriched in contact sites, the location at which preprotein import is believed to occur. A protein of almost identical size was also found to be labelled in mitochondria from yeast.     

Mg-nucleotides induced dissociation of liposome-bound creatine kinase: reversible changes in its secondary structure and in the fluidity of the bilayer

MgADP binding to mitochondrial creatine kinase (mtCK) adsorbed on liposomes was induced by the photorelease of caged ADP. The nucleotide binding produced two types of structural changes. One was related to the well-established release of mtCK from the liposomes. The other corresponded to reversible structural changes induced by nucleotide binding to mtCK as demonstrated here. Infrared spectroscopy data show that the MgADP-induced desorption of mtCK from vesicles led to a slight increase in alpha-helix structures in mtCK at the expense of a small decrease in beta-sheet structures and a concomitant increase in the fluidity of the membranes. The desorption of mtCK induced by MgADP and MgATP was almost complete, as shown by centrifugation and enzymatic activity measurements. The photorelease of MgADP in a reactive medium containing phosphocreatine and mtCK associated with liposomes led to nucleotide binding and to the formation of MgATP and creatine. Addition of phosphocreatine also desorbed mtCK from liposomes, while addition of creatine did not. Interpretation of these results would suggest that ADP, ATP or phosphocreatine induce the release of mtCK from membranes, increase the phospholipid bilayer fluidity, and may also decrease the number of contact sites between inner and outer mitochondrial membranes, thus affecting the activity of other mitochondrial enzymes. It is tempting to propose that membrane mtCK binding regulation by nucleotide and PCr concentrations may serve as a physiological adaptation for energy supply.     

Fluorescence polarization and spin-label studies of the fluidity of stromal and granal chloroplast membranes

The lipid fluidity of thylakoid membrane regions separated by Yeda press and sonication methods has been investigated using diphenylhexatriene fluorescence polarization measurements and rotational correlation times derived from the ESR spectra of the spin-labels 5-doxyldecane and 12-doxylstearate. According to both techniques, stromal lamellae vesicles with essentially only Photosystem I activity were more fluid than the granal membranes. The differences in lipid fluidity between the two fractions were interpreted in terms of the ratio of the amounts of protein compared to lipid in the membranes. Stromal lamellae fractions contained lower protein/lipid ratios compared with the granal membranes.     

Plasma membranes from Candida tropicalis grown on glucose or hexadecane. I. Isolation, identification and purification.

Plasma membranes from Candida tropicalis grown on glucose or hexadecane were isolated using a method based on the difference in surface charge of mitochondria and plasma membranes. After mechanical disruption of the cells, a fraction consisting of mitochondrial and plasma membrane vesicles was obtained by differential centrifugation. Subsequently the mitochondria were separated from the plasma membrane vesicles by aggregation of the mitochondria at a pH corresponding to their isoelectric point. Additional purification of the isolated plasma membrane vesicles was achieved by osmolysis. Surface charge densities of mitochondria and plasma membranes were determined and showed substrate-dependent differences. The isolated plasma membranes were morphologically characterized by electron microscopy and, as a marker enzyme, the activity of Mg2+-dependent ATPase was determine. By checking for three mitochondrial marker enzymes the plasma membrane fractions were estimated to be 94% pure with regard to mitochondrial contamination.     

Mg-nucleotides induced dissociation of liposome-bound creatine kinase: reversible changes in its secondary structure and in the fluidity of the bilayer

MgADP binding to mitochondrial creatine kinase (mtCK) adsorbed on liposomes was induced by the photorelease of caged ADP. The nucleotide binding produced two types of structural changes. One was related to the well-established release of mtCK from the liposomes. The other corresponded to reversible structural changes induced by nucleotide binding to mtCK as demonstrated here. Infrared spectroscopy data show that the MgADP-induced desorption of mtCK from vesicles led to a slight increase in &#102 -helix structures in mtCK at the expense of a small decrease in &#103 -sheet structures and a concomitant increase in the fluidity of the membranes. The desorption of mtCK induced by MgADP and MgATP was almost complete, as shown by centrifugation and enzymatic activity measurements. The photorelease of MgADP in a reactive medium containing phosphocreatine and mtCK associated with liposomes led to nucleotide binding and to the formation of MgATP and creatine. Addition of phosphocreatine also desorbed mtCK from liposomes, while addition of creatine did not. Interpretation of these results would suggest that ADP, ATP or phosphocreatine induce the release of mtCK from membranes, increase the phospholipid bilayer fluidity, and may also decrease the number of contact sites between inner and outer mitochondrial membranes, thus affecting the activity of other mitochondrial enzymes. It is tempting to propose that membrane mtCK binding regulation by nucleotide and PCr concentrations may serve as a physiological adaptation for energy supply.     

Protein import into mitochondria: ATP-dependent protein translocation activity in a submitochondrial fraction enriched in membrane contact sites and specific proteins

To identify the membrane regions through which yeast mitochondria import proteins from the cytoplasm, we have tagged these regions with two different partly translocated precursor proteins. One of these was bound to the mitochondrial surface of ATP-depleted mitochondria and could subsequently be chased into mitochondria upon addition of ATP. The other intermediate was irreversibly stuck across both mitochondrial membranes at protein import sites. Upon subfraction of the mitochondria, both intermediates cofractionated with membrane vesicles whose buoyant density was between that of inner and outer membranes. When these vesicles were prepared from mitochondria containing the chaseable intermediate, they internalized it upon addition of ATP. A non-hydrolyzable ATP analogue was inactive. This vesicle fraction contained closed, right-side-out inner membrane vesicles attached to leaky outer membrane vesicles. The vesicles contained the mitochondrial binding sites for cytoplasmic ribosomes and contained several mitochondrial proteins that were enriched relative to markers of inner or outer membranes. By immunoelectron microscopy, two of these proteins were concentrated at sites where mitochondrial inner and outer membranes are closely apposed. We conclude that these vesicles contain contact sites between the two mitochondrial membranes, that these sites are the entry point for proteins into mitochondria, and that the isolated vesicles are still translocation competent.     

Plasma membranes from candida tropicalis grown on glucose or hexadecane. I. Isolation,identification and purification

Plasma membranes from Candida tropicalis grown on glucose or hexadecane were isolated using a method based on the difference in surface charge of mitochondria and plasma membranes.After mechanical disruption of the cells, a fraction consisting of mitochondrial and plasma membrane vesicles was obtained by differential centrifugation.Subsequently the mitochondria were separated from the plasma membrane vesicles by aggregation of the mitochondria at a pH corresponding to their isoelectric point. Additional purification of the isolated plasma membrane vesicles was achieved by osmolysis. Surface charge densities of mitochondria and plasma membranes were determined and showed substrate-dependent differences.The isolated plasma membranes were morphologically characterized by electron microscopy and, as a marker enzyme, the activity of Mg²⁺-dependant ATPase was determined.By checking for three mitochondrial marker enzymes the plasma membrane fractions were estimated to be 94% pure with regard to mitochondrial contamination.     

Membrane properties modulate the activity of a phosphatidylinositol transfer protein from the yeast, Saccharomyces cerevisiae

A phospholipid transfer protein from yeast (Daum, G. and Paltauf, F. (1984) Biochim. Biophys. Acta 794, 385-391) was 2800-fold enriched by an improved procedure. The specificity of this transfer protein and the influence of membrane properties of acceptor vesicles (lipid composition, charge, fluidity) on the transfer activity were determined in vitro using pyrene-labeled phospholipids. The yeast transfer protein forms a complex with phosphatidylinositol or phosphatidylcholine, respectively, and transfers these two phospholipids between biological and/or artificial membranes. The transfer rate for phosphatidylinositol is 19-fold higher than for phosphatidylcholine as determined with 1:8 mixtures of phosphatidylinositol and phosphatidylcholine in donor and acceptor membrane vesicles. If acceptor membranes consist only of non-transferable phospholipids, e.g., phosphatidylethanolamine, a moderate but significant net transfer of phosphatidylcholine occurs. Phosphatidylcholine transfer is inhibited to a variable extent by negatively charged phospholipids and by fatty acids. Differences in the accessibility of the charged groups of lipids to the transfer protein might account for the different inhibitory effects, which occur in the order phosphatidylserine which is greater than phosphatidylglycerol which is greater than phosphatidylinositol which is greater than cardiolipin which is greater than phosphatidic acid which is greater than fatty acids. Although mitochondrial membranes contain high amounts of negatively charged phospholipids, they serve effectively as acceptor membranes, whereas transfer to vesicles prepared from total mitochondrial lipids is essentially zero. Ergosterol reduces the transfer rate, probably by decreasing membrane fluidity. This notion is supported by data obtained with dipalmitoyl phosphatidylcholine as acceptor vesicle component; in this case the transfer rate is significantly reduced below the phase transition temperature of the phospholipid.     

Lipid topology and physical properties of the outer mitochondrial membrane of the yeast, Saccharomyces cerevisiae

The outer membrane of yeast mitochondria was studied with respect to its lipid composition, phospholipid topology and membrane fluidity. This membrane is characterized by a high phospholipid to protein ratio (1.20). Like other yeast cellular membranes the outer mitochondrial membrane contains predominantly phosphatidylcholine (44% of total phospholipids), phosphatidylethanolamine (34%) and phosphatidylinositol (14%). Cardiolipin, the characteristic phospholipid of the inner mitochondrial membrane (13% of total phospholipids) is present in the outer membrane only to a moderate extent (5%). The ergosterol to phospholipid ratio is higher in the inner (7.0 wt%) as compared to the outer membrane (2.1 wt.%). Attempts to study phospholipid asymmetry by selective degradation of phospholipids of the outer leaflet of the outer mitochondrial membrane failed, because isolated right-side-out vesicles of this membrane became leaky upon treatment with phospholipases. Selective removal of phospholipids of the outer leaflet with the aid of phospholipid transfer proteins and chemical modification with trinitrobenzenesulfonic acid on the other hand, gave satisfactory results. Phosphatidylcholine and phosphatidylinositol are more or less evenly distributed between the two sides of the outer mitochondrial membrane, whereas the majority of phosphatidylethanolamine is oriented towards the intermembrane space. The fluidity of mitochondrial membranes was determined by measuring fluorescence anisotropy using diphenylhexatriene (DPH) as a probe. The lower anisotropy of DPH in the outer as compared to the inner membrane, which is an indication for an increased lipid mobility in the outer membrane, was attributed to the higher phospholipid to protein and the lower ergosterol to phospholipid ratio. The data presented here show, that the outer mitochondrial membrane, in spite of its close contact to the inner membrane, is distinct not only with respect to its protein pattern, but also with respect to its lipid composition and physical membrane properties.     

Distribution of the integral membrane protein NADH-cytochrome b5 reductase in rat liver cells, studied with a quantitative radioimmunoblotting assay.

The intracellular localization of the post-translationally inserted integral membrane protein, NADH-cytochrome b5 reductase, was investigated, using a quantitative radioimmunoblotting method to determine its concentration in rat liver subcellular fractions. Subcellular fractions enriched in rough or smooth microsomes, Golgi, lysosomes, plasma membrane and mitochondrial inner or outer membranes were characterized by marker enzyme analysis and electron microscopy. Reductase levels were determined both with the NADH-cytochrome c reductase activity assay, and by radioimmunoblotting, and the results of the two methods were compared. When measured as antigen, the reductase was relatively less concentrated in microsomal subfractions, and more concentrated in fractions containing outer mitochondrial membranes, lysosomes and plasma membrane than when measured as enzyme activity. Rough and smooth microsomes had 4-5-fold lower concentrations, on a phospholipid basis than did mitochondrial outer membranes. Fractions containing Golgi, lysosomes and plasma membrane had approximately 14-, approximately 16, and approximately 9-fold lower concentrations of antigen than did mitochondrial outer membranes, respectively, and much of the antigen in these fractions could be accounted for by cross-contamination. No enzyme activity or antigen was detected in mitochondrial inner membranes. Our results indicate that the enzyme activity data do not precisely reflect the true enzyme localization, and show an extremely uneven distribution of reductase among different cellular membranes.     

Localization of phosphatases in Trypanosoma rhodesiense

Phosphatase activity in Trypanosoma rhodesiense has been examined histochemically by light and electron microscopy and by enzymatic assay in homogenate fractions. Using a method with lead as capture ion, acid phosphatase was found in lysosome-like vesicles and in the flagellar pocket. No alkaline adenosine triphosphatase (ATPase) was detectable by this method. Direct assay of p-nitrophenylphosphatase activity in homogenate fractions showed that acid phosphatase activity was strongly membrane-bound, but that activity at pH 9 was minimal in both soluble and particulate fractions. "Endogenous" ATPase activity was localized specifically and reproducibly in the mitochondrial membranes and under the plasma membrane of he flagellum. This nonenzymic reaction product could not be eradicated by glycerol extraction or glucose depletion. Unlike the membrane staining, which was manifest only after lead treatment, heat-resistant electron-dense material was found in the matrix of lysosomal vesicles in trypanosomes fixed in glutaraldehyde only and not subjected to further treatment with heavy metal reagents. X-ray emission analysis showed the presence of calcium and phosphorus, indicating that the matrix might have a phosphate storage function.     

Synthesis of plasmalemmal glycoproteins in intestinal epithelial cells. Separation of Golgi membranes from villus and crypt cell surface membranes; glycosyltransferase activity of surface membrane

The relationship between Golgi and cell surface membranes of intestinal cells was studied. These membranes were isolated from intestinal crypt cells and villus cells. The villus cell membranes consisted of microvillus membrane, a Golgi-rich fraction, and two membrane fractions interpreted as representing lateral-basal membranes. The villus cell microvillus membrane was purified by previously published techniques while the other membranes were obtained from isolated cells by differential centrifugation and density gradient velocity sedimentation. The two membrane fractions obtained from villus cells and considered to be lateral-basal membranes were enriched for Na+,K+-ATPase activity, but one also showed enrichment in glycosyltransferase activity. The Golgi membrane fraction was enriched for glycosyltransferase activity and had low to absent Na+,K+-ATPase activity. Adenylate cyclase activity was present in all membrane fractions except the microvillus membrane but co-purified with Golgi rather than lateral-basal membranes. Electron microscopy showed that the Golgi fraction consisted of variably sized vesicles and cisternalike structures. The two lateral-basal membrane fractions showed only vesicles of smaller, more uniform size. After 125I labeling of isolated intact cells, radioactivity was found associated with the lateral-basal and microvillus membrane fractions and not with the Golgi fraction. Antibody prepared against lateral-basal membrane fractions reacted with the surface membrane of isolated villus cells. The membrane fractions from isolated crypt cells demonstrated that all had high glycosyltransferase activity. The data show that glycosyltransferase activity, in addition to its Golgi location, may be a significant property of the lateral-basal portion of the intestinal villus cell plasma membrane. Data obtained with crypt cells support earlier data and show that the crypt cell surface membrane possesses glycosyltransferase activity.     

Release of the phosphodiesterase activator by cyclic AMP-dependent ATP:protein phosphotransferase from subcellular fractions of rat brain.

The subcellular distribution of the endogenous phosphodiesterase activator and its release from membranes by a cyclic AMP-dependent ATP:protein phosphotransferase was studied in fractions and subfractions of rat brain homogenate. These fractions were obtained by differential centrifugation and sucrose density gradient; their identity was ascertained by electron microscopy and specific enzyme markers. In the subcellular particulate fractions, the concentration of activator is highest in the microsomal fraction, followed by the mitochondrial and nuclear fractions. Gradient centrifugation of the main mitochondrial subfraction revealed that activator was concentrated in those fractions containing mainly synaptic membranes. Activator was releasted from membranes by a cyclic AMP-dependent phosphorylation of membrane protein. The release of activator occurred mainly from the mitochondrial subfractions containing synaptic membranes and synaptic vesicles. The data support the view that a release of activator from membranes may be important in normalizing the elevated concentration of cyclic AMP following persistent transsynaptic activation of adenylate cyclase.     

Mitochondrial creatine kinase mediates contact formation between mitochondrial membranes

Purified mitochondrial creatine kinase (Mi-CK) (EC 2.7.3.2) from chicken heart was shown to interact simultaneously with purified inner and outer mitochondrial membranes, thereby creating an intermembrane chondrial membranes, thereby creating an intermembrane were purified from rat liver and thus were fully devoid of Mi-CK. Intermembrane contact formation was demonstrated by measuring the binding of inner membrane vesicles to outer membranes spread at the air-water interface. Mi-CK also mediated intermembrane adhesion when membranes formed with total lipid extracts of both membranes were used, pointing to the role of lipids as potential membrane anchors of Mi-CK in the mitochondrial intermembrane space. Other enzymes of the intermembrane space that (like Mi-CK) are also cationic, as well as cytosolic isoenzymes of creatine kinase, failed to induce contact formation. Thus, of the proteins tested, membrane contact formation was specific for Mi-CK. The two oligomeric forms of Mi-CK (octamer and dimer) differed in their ability to mediate intermembrane adhesion, the octamer being more potent. Highly basic peptides, i.e. poly-L-lysines, were shown to strongly interact with membranes formed with lipid extracts of mitochondrial membranes: they both induced intermembrane binding and fusion. Interestingly, the extent of contact formation mediated by poly-L-lysines was lower than that of octameric Mi-CK. The implications of these findings on the function and localization of Mi-CK and on the structure of the mitochondrial intermembrane compartment are discussed.     

Cytochemical studies on the localization and functional properties of calcium in anterior pituitary cells

The localization of calcium and its functional properties in anterior pituitary cells were studied using a potassium pyroantimonate technique. In all kinds of secretory cells, the precipitates of the calcium-pyroantimonate complex were distributed on the limiting membrane of the secretory granule. They were present also in the cytoplasmic matrix, the mitochondrial matrix, small smooth vesicles, coated vesicles, and in the nuclear euchromatin area. The precipitates were usually seen at the contact region between the limiting membranes of two adjacent secretory granules, or between the granule limiting membrane and the plasma membrane. When the tissues were incubated in the medium containing A23187 (10 microM) for 5 min, the deposits on the granule limiting membrane were increased in number and those on the mitochondrial matrix were decreased; the reaction products almost disappeared on the limiting membranes of the secretory granules after membrane fusion following single or multigranular exocytosis induced by A23187-treatment. In addition, small vesicles in the capillary endothelium contained reaction precipitates. Based on these results we propose a hypothetical model for the relationship between the localization of calcium and secretory activity.     

Contact sites from human placental mitochondria: characterization and role in progesterone synthesis

To understand the functional compartmentalization of human placental mitochondria, we analyzed the composition and steroidogenic activity of contact sites. Several fractions containing contact sites were isolated using osmotic shock treatment and sucrose gradient centrifugation. These fractions contained various proteins and marker enzymes associated with mitochondrial membranes. The fractions containing the cytochrome P450 side chain cleavage system, cholesterol, nicotinamide adenine dinucleotide phosphate-isocitrate dehydrogenase, porin, and adenosine 5(')-triphosphate-diphosphohydrolase activity showed the capacity to synthesize progesterone. Our observations indicate that all necessary elements and enzymes for steroidogenesis are present and functional in placental mitochondrial contact sites. This organization may facilitate the metabolism of cholesterol delivered to the outer mitochondrial membrane into steroid hormones by the inner mitochondrial membrane cholesterol side chain cleavage system.     

Localization of Phosphatases in Trypanosoma rhodesiense1

ABSTRACT. Phosphatase activity in Trypanosoma rhodesiense has been examined histochemically by light and electron microscopy and by enzymatic assay in homogenate fractions. Using a method with lead as capture ion, acid phosphatase was found in lysosome-like vesicles and in the flagellar pocket. No alkaline adenosine triphosphatase (ATPase) was detectable by this method. Direct assay of p-nitrophenylphosphatase activity in homogenate fractions showed that acid phosphatase activity was strongly membrane-bound, but that activity at pH 9 was minimal in both soluble and particulate fractions. “Endogenous” ATPase activity was localized specifically and reproducibly in the mitochondrial membranes and under the plasma membrane of the flagellum. This nonenzymic reaction product could not be eradicated by glycerol extraction or glucose depletion. Unlike the membrane staining, which was manifest only after lead treatment, heat-resistant electron-dense material was found in the matrix of lysosomal vesicles in trypanosomes fixed in glutaraldehyde only and not subjected to further treatment with heavy metal reagents. X-ray emission analysis showed the presence of calcium and phosphorus, indicating that the matrix might have a phosphate storage function.     

Relationship between lipid fluidity and water permeability of bovine tracheal epithelial cell apical membranes

Apical membrane vesicles were prepared from bovine tracheal epithelial cells. These membranes were enriched in alkaline phosphatase specific activity 35-fold compared to cellular homogenates. Steady-state fluorescence polarization studies of these membranes, using three fluorophores, demonstrated that they possessed a relatively low fluidity. Studies using the probe 1,6-diphenyl-1,3,5-hexatriene detected thermotropic transitions at 25.7 +/- 0.4 and 26.8 +/- 0.6 degrees C in these membranes and their liposomes, respectively. Analysis of the composition of these membranes revealed a fatty acyl saturation index of 0.59 +/- 0.02, a protein/lipid ratio (w/w) of 0.60 +/- 0.06, a cholesterol/phospholipid ratio (mol/mol) of 0.83 +/- 0.11, and a sphingomyelin/lecithin ratio (mol/mol) of 0.64 +/- 0.10. Membrane vesicles were osmotically active when studied by a stopped-flow nephelometric technique. Arrhenius plots of rates of osmotic water efflux demonstrated break points at approximately 28 and 18 degrees C, with activation energies of 16.7 +/- 0.2 kcal mol-1 from 35 to 28 degrees C, 8.3 +/- 0.5 kcal mol-1 from 28 to 18 degrees C, and approximately 3.0 kcal mol-1 below 18 degrees C. Treatment of membrane vesicles with benzyl alcohol, a known fluidizer, decreased lipid order (increased fluidity) and increased the rate of osmotic water efflux. The present results suggest that water crosses tracheal epithelial cell apical membranes by solubility-diffusion across the lipid domain and that increases in fluidity correlate with increases in the water permeability of these membranes.