Effects of thermal adaptation at 40 degrees C on membrane viscosity and the sodium-potassium pump in Chinese hamster ovary cells.

We have demonstrated increased heat resistance in Chinese hamster ovary cells grown to confluence at 37 degrees C and thermally adapted at the nonlethal temperature of 40 degrees C for 24 h. Membrane viscosity, estimated by fluorescence anisotropy, was inversely related to temperature, from 5 degrees C to 45 degrees C. For a given temperature viscosity was consistently higher in thermally adapted cells than in native cells. Having demonstrated a change in membrane structure associated with thermal adaptation, we carried out a study of the Na+-K+ pump in native and thermally adapted cells as an example of a vital active transport process known to be sensitive to membrane viscosity. 86Rb uptake measured from 31 degrees C to 50 degrees C increased steadily to 46 degrees C and then decreased rapidly in both native and thermally adapted cells. Detailed measurement of ouabain-sensitive 86Rb influx demonstrated an increase in both Km and Vmax between 37 degrees C and 45 degrees C, but there was no difference between native and thermally adapted cells. We have thus demonstrated an adaptive structural change in the cell membrane of mammalian cells which may be related to the induction of thermal resistance at 40 degrees C but which is not associated with any change in this active transport system.

Phospholipid asymmetry in renal brush-border membranes.

The topological distribution of phospholipids between the inside and the outside of rabbit kidney brush-border membranes has been investigated by incubating membrane vesicles with sphingomyelinase, phospholipases A2 from bee venom and hog pancreas, phospholipases C and D, and trinitrobenzene sulfonate. Orientation and integrity of vesicles upon phospholipase treatment was determined by using two monoclonal antibodies recognizing an extracytoplasmic and a cytoplasmic domain, respectively, of the neutral endopeptidase (EC 3.4.24.11). It is shown that the transbilayer distribution of phospholipids is highly asymmetrical in kidney brush-border membranes: sphingomyelin accounted for 75% of the phospholipids present in the external leaflet, whereas phosphatidylethanolamine and phosphatidylserine plus phosphatidylinositol were found to comprise the majority of the inner layer of the membrane.

The involvement of cytoskeletal proteins in the maintenance of phospholipid topology in renal brush-border membranes.

When incubated for 14 h at 37 degrees C in the absence of energy supply, brush-border membrane vesicles from rabbit kidney cortex maintain, as judged by the use of sphingomyelinase and trinitrobenzene sulfonate as membrane probes, their highly asymmetrical phospholipid distribution. In particular, sphingomyelin still accounts for 75% of the phospholipids present on the outer membrane leaflet. Pretreatment of the vesicles with 5 mM diamide resulted in extensive crosslinking of membranous and cytoskeletal proteins. Although it had no immediate effect on the topology of phospholipids, this crosslinking resulted in a limited but significant increase in the amount of aminophospholipids present on the outer membrane leaflet after 14-h incubations. Degradation of aminophospholipids, upon incubation with hog pancreas and bee venom phospholipases A2, was also enhanced by diamide. However, this enhanced hydrolysis was observed immediately after the diamide treatment. A similar increase in degradation of aminophospholipids was obtained when vesicles were incubated with dihydrocytochalasin B. Our results strongly suggest that cytoskeletal proteins, via interactions with aminophospholipids, stabilize the lipid bilayer of the brush-border membrane. It is also suggested that, due to a low transbilayer migration rate, sphingomyelin may play an important role in the maintenance of the lipid asymmetry in these membranes.

Effects of benzyl alcohol on enzyme activities and D-glucose transport in kidney brush-border membranes.

Addition of increasing amounts of benzyl alcohol progressively reduced the steady-state anisotropies of diphenylhexatriene and trimethylammoniumdiphenylhexatriene in brush-border membranes from rat kidney. The decrease in order of membrane lipids, equivalent for 50 mM benzyl alcohol to that produced by a rise in temperature of approx. 6 degrees C, had no effect on the activities of alkaline phosphatase or gamma-glutamyltranspeptidase. On the other hand, benzyl alcohol markedly inhibited the D-glucose uptakes measured in the presence of a 100 mM sodium gradient. For concentrations less than 30 mM, benzyl alcohol reduced the Jmax without significant effects on Km, 22Na+ uptake or the vesicular volume of brush-border preparations. Comparable results were obtained substituting octanol for benzyl alcohol. Our data strongly suggest that, at constant temperature, the D-glucose carrier present in renal brush-border membranes is extremely sensitive to variations in membrane physical state.

Water permeation in Madin-Darby canine kidney cells is modulated by membrane fluidity.

Simultaneous determinations of water and antipyrine permeations in monolayers of Madin-Darby canine kidney (MDCK) cells grown on a permeant support were done to study the relationships between water transport and membrane fluidity in these epithelial cells. The changes in permeation of the lipophilic non-electrolyte antipyrine were used to probe the modifications in membrane fluidity. In controls, the apparent diffusional permeability coefficient for water (PDw) was three times higher than the antipyrine's one, PDAp (4.2.10(-5) vs. 1.4.10(-5) cm s-1). Addition of vasopressin or dibutyryl cyclic AMP to the monolayers induced a biphasic increase in antipyrine permeation with peak values at t = 2 min, 3-4-fold that of controls. Variations in water permeation were of similar amplitude and obeyed the same time course, leaving the water to antipyrine permeation ratios unchanged. Compound H7, an inhibitor of protein kinases, blunted the increase in permeation for both antipyrine and water. Finally, addition of the fluidizing agent benzyl alcohol to the monolayers resulted in a parallel increase in PDAp and PDw. These results suggest that the physical state of membrane lipids may control water permeation in MDCK cells.

Temperature-dependent relationship between K+ influx, Mg2+-ATPase activity, transmembrane potential and membrane lipid composition in mycoplasma.

The temperature-dependent relationship between K+ active influx, Mg2+-ATPase activity, transmembrane potential (delta psi) and the membrane lipid composition has been investigated in mycoplasma PG3. Native organisms were grown in a medium containing 10 microgram/ml cholesterol and either oleic plus palmitic (chol (+), O + P) or elaidic (chol (+), E) acids. Adapted cells were grown in a medium free of exogenous cholesterol and supplemented with elaidic acid (chol (-), E). Arrhenius plots of 42K+ active influx gave a linear relationship for (chol (+), O + P) cells (EA = -9 kcal). On the other hand, when oleic plus palmitic acids are replaced by elaidic acid, an upward discontinuity appears between 28 and 30 degrees C, which is associated with a large increase in the apparent activation energy of the process (t > 30 degrees C, EA = -24 kcal; t < 30 degrees C, EA = -40 kcal). Finally, a biphasic response with a break at approx. 23 degrees C (EA = -7 kcal, t > 23 degrees C; EA = -44 kcal, t < 23 degrees C) is observed for (chol(-), E) organisms. From the lack of correspondence between these effects on the K+ influx and the temperature dependence of both the Mg2+-ATPase activity and delta psi, it is suggested that changes in the membrane lipid composition affect the K+ transport at the level of the K+ carrier itself. Differential scanning calorimetry, steady-state fluorescence polarization of diphenylhexatriene and freeze-fracture electron microscopy experiments further suggest that the effect is largely due to modifications of the membrane microviscosity and that the K+ carrier is associated with the most fluid lipid species present in the membrane.

Active K+ transport in Mycoplasms mycoides var. Capri. Relationships between K+ distribution, electrical potential and ATPase activity.

The addition of 5 . 10(-5) M or less of dicyclohexylcarbodiimide to Mycoplasma mycoides var. Capri preferentially influences K+ influx rather than efflux and reduces by 30--40% the activity of the membrane-bound Mg2+- ATPase. Adding valinomycin to metabolizing cells does not markedly affect K+ distribution but induces a rapid and complete loss of intracellular K+ in non-metabolizing cells. Uncoupling agents such as dinitrophenol, carbonylcyanide p-trifluoromethoxyphenylhydrazone, dissipate the K+ concentration gradient only when combined with valinomycin. Variations in the merocyanine fluorescence intensity indicate that a transmembrane electrical potential (delta psi) is generated on cell energization. This delta psi, not affected by valinomycin or uncouplers when used alone, is collapsed by a mixture of both. No change in fluorescence intensity can be detected when glucose is added to dicyclohexylcarbodiimide treated organisms. These experiments suggest that the membrane-bound Mg-ATPase activity control K+ distribution in these organisms through the generation of a transmembrane electrical potential difference.

Effect of membrane cholesterol on potassium transport in Mycoplasma mycoides var. Capri (PG3).

Relationships between membrane lipid composition and physiological properties, particularly intracellular potassium levels, have been studied at 37 degrees C in Mycoplasma mycoides var. Capri (PG3). Native organisms grown on medium supplemented with either oleic acid plus palmitic acid or elaidic acid have identical growth characteristics, acidification properties and intracellular K content. On the other hand, when the cholesterol normally present in the membrane (20--25% of total lipids) is reduced to less than 2%, we observe: (1) the intracellular K content decreases (20 microgram K/mg cell protein instead of 40) and is independent of the phase of growth; (2) K passive permeability is drastically increased but K distribution remains in equilibrium with the transmembrane potential (delta psi); (3) organisms stop growing at pH 6.5 (instead of 5.2) and acidification is reduced by 40%, suggesting a large increase in H+ permeability, and (4) intracellular Na contents rise from 3 to 9 microgram Na/mg cell protein. Replenishing cholesterol in membranes of depleted cells results in a recovery of the high intracellular K level (35--40 microgram K/mg cell protein) and acidification properties. It is suggested that cholesterol affects the cation content via the increase in proton permeability which in turn controls the value of the delta psi responsible for the value of intracellular K equilibrium. Changes in K passive permeability, although related to the amount of cholesterol present in the plasma membrane, are probably not involved in the control of the intracellular K level.

Active K+ transport in Mycoplasma mycoides var. Capri. Net and unidirectional K+ movements.

Analysis of the cation composition of growing Mycoplasma mycoides var. Capri indicates that these organisms have a high intracellular K+ concentration (Ki: 200--300 mM) which greatly exceeds that of the growth medium, and a low Na+ concentration (Na+i: 20 mM). Unlike Na+i,K+i varies with cell aging. The K+ transport properties studied in washed organisms resuspended in buffered saline solution show that cells maintain a steady and large K+ concentration gradient across their membrane at the expense of metabolic energy mainly derived from glycolysis. In starved cells, K+i decreases and is partially compensated by a gain in Na+. This substitution completely reverses when metabolic substrate is added (K+ reaccumulation process). Kinetic analysis of K+ movement in cells with steady K+ level shows that most of K+ influx is mediated by an autologous K+-K+ exchange mechanism. On the other hand, during K+ reaccumulation by K+-depleted cells, a different mechanism (a K+ uptake mechanism) with higher transport capacity and affinity drives the net K+ influx. Both mechanisms are energy-dependent. Ouabain and anoxia have no effect on K+ transport mechanisms; in contrast, both processes are completely blocked by dicyclohexylcarbodiimide, an inhibitor of the Mg2+ -dependent ATPase activity.

Analysis of membrane fractions from Mycoplasma gallisepticum.

Membrane fractions have been isolated from Mycoplasma gallisepticum following a procedure derived from that described by Maniloff, J. and Quinlan, D.C. (J. Bacteriol. (1974) 120, 495-501). A light fraction F1 was obtained which contained structures resembling the bleb-infrableb apparatus characteristic of M. gallisepticum. It was enriched in DNA and had an electrophoretic profile different from that of unfractionated membranes. Cholesterol-to-phospholipid ratios higher than two and elevated values of the ratio of saturated to unsaturated fatty acids were other characteristics of this fraction. The two other fractions isolated (FII and FIV) also differed from intact membranes by their cholesterol and phospholipid content as well as by their saturation ratios. The membrane fluidity of FII and FIV, estimated by fluorescence polarization, was similar to that of unfractionated membranes while a slight but significant difference was recorded for the light fraction. Possible relationships between the lateral heterogeneity of the M. gallisepticum membrane and the obtainment of fractions are discussed.

Effects of energization on membrane organization in mycoplasma.

Fluorescence polarization and ESR experiments using various probes demonstrated that addition of glucose to resting Mycoplasma capricolum and Mycoplasma mycoides subs capri had, if any, a very limited effect on the physical state of their membrane lipids. Under the same conditions the degree of exposure of primary amino groups of membrane proteins to the aqueous surrounding, estimated from fluorescence labeling by fluorescamine and the cycloheptaamylose-fluorescamine complex was significantly increased. This energy dependent increase was blocked by dicyclohexylcarbodiimide (DCCD), an inhibitor of the membrane bound Mg2+ stimulated ATPase of mycoplasma and by carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP) which, in mycoplasma, only affects the chemical component of the proton-motive force. Variations in the proton activity gradient across the membrane induced by changing the pH of the labeling medium resulted in parallel variations in the ratio of relative intensities of labeling of energized to resting cells. The values taken by this ratio were up to two for a maximal proton gradient of 0.9 pH unit and tended to unity when the intracellular and extracellular pH tended to equalize. It is concluded that, upon mycoplasma cell energization, membrane proteins undergo a conformational change resulting in the exposure of new free amino groups. This conformational change is primarily dependent on the existence of a delta ph across the membrane and occurs in the absence of important modifications in the physical state of membrane lipids.

Benzyl alcohol differently affects fluid phase endocytosis and exocytosis in renal epithelial cells.

The effects of benzyl alcohol, a local anaesthetic commonly used for modification of membrane fluidity, on fluid phase endocytosis and on exocytosis have been investigated in MDCK cells. Fluid phase endocytosis in confluent cells monolayer grown on solid support was determined, at 37 degrees C, by the uptake of the fluorescent dye Lucifer Yellow (LY). Exocytosis was estimated from the release of LY by cells preloaded with the dye. Addition of benzyl alcohol resulted in a concentration dependent inhibition of fluid phase endocytosis. For 30 mM benzyl alcohol, the inhibition obtained (83%) compared with that produced by preincubating the cells in a solution made hypertonic with 0.25 M sucrose. The inhibitory effect of benzyl alcohol was reversed within 30 min by washing. Endocytosis inhibition by benzyl alcohol was also observed in LLC-PK1 cells and OK cells, two renal epithelial cell lines of proximal tubule origin. In contrast, benzyl alcohol had no effect on exocytosis in LLC-PK1 cells, a limited but significant (15% at 30 mM) stimulatory effect on exocytosis in MDCK cells and a marked stimulatory effect (75% at 30 mM) in OK cells. These data demonstrate that benzyl alcohol affects endocytosis and exocytosis processes in renal epithelial cells. They suggest that membrane fluidity may alter membrane trafficking in living renal epithelial cells.

Increase in membrane fluidity modulates sodium-coupled uptakes and cyclic AMP synthesis by renal proximal tubular cells in primary culture.

In order to evaluate the influence of membrane fluidization on three apical transport systems and on a basolateral enzyme, and to analyse the mechanisms involved, we studied, in cultured rabbit proximal tubular cells, the effect of increasing concentrations of the local anesthetic drug benzyl alcohol on Na(+)-dependent uptakes of phosphate (Pi), methyl alpha-D-glucopyranoside (MGP), and L-alanine, as well as on basal and stimulated cyclic AMP content. At 10 mM, benzyl alcohol increased the Vmax of Pi uptake by 31%, decreased that of MGP uptake by 24%, and did not affect alanine uptake. Km values were not affected. Benzyl alcohol, up to 40 mM, increased in a concentration-dependent manner basal, PTH-stimulated, and cholera toxin-stimulated, but not forskolin-stimulated cyclic AMP accumulation. In the presence of 40 mM benzyl alcohol, the magnitude of PTH-induced inhibition of Pi uptake was enhanced from 11% to 24%. It is concluded that: (i) fluidization of apical membranes affected differently Na+/Pi, Na+/MGP, and Na+/alanine cotransports, reflecting differences in the lipidic environments of these transport system; (ii) fluidization of basolateral membranes enhanced PTH-stimulated cyclic AMP generation through improved coupling between the receptor-GS complex and the catalytic subunit of adenylate cyclase; (iii) these variations may result in physiological and pathophysiological modulation of the renal handling of solutes and of the phosphaturic effect of PTH.

Benzyl alcohol increases membrane fluidity and modulates cyclic AMP synthesis in intact renal epithelial cells.

To evaluate a possible modulation by membrane fluidity of hormonal, cAMP-mediated effects on renal epithelial cells, we studied the effect of the neutral local anesthetic, benzyl alcohol, on membrane fluidity and on basal and stimulated intracellular cAMP content in intact MDCK cells. Benzyl alcohol induced a dose-dependent decrease of lipid order which was measured by steady-state fluorescence anisotropy using trimethylammonium-diphenylhexatriene and propionyl-diphenylhexatriene as fluorescent probes. Benzyl alcohol induced a 2-fold increase in basal cAMP content, likely as a consequence of increased prostaglandin synthesis since this effect was abolished by indomethacin. The effect of benzyl alcohol on stimulated cAMP synthesis depended on the nature of the ligand: 10 mM benzyl alcohol increased significantly the stimulatory effect of prostaglandin E2, glucagon and forskolin but not of vasopressin. At higher concentrations (40 mM), benzyl alcohol did not affect significantly the glucagon-stimulated cAMP content, while it inhibited significantly the prostaglandin E2-, forskolin- and vasopressin-stimulated cAMP synthesis. The 40 mM benzyl alcohol-induced inhibition was reversed by 1 mM Mn2+, which is known to block the inhibitory GTP-binding protein Ni. These results suggest that: (i) the various components of the adenylate cyclase-cAMP system and their coupling are affected differently by changes in membrane fluidity, which might reflect differences in their lipid environment, (ii) changes in membrane fluidity can modulate responses of renal tubular cells to hormones, and thus tubular functions.

Transport of choline by Madin-Darby canine kidney cells.

Choline is an essential precursor for the synthesis of phosphatidylcholine, the most abundant phospholipid classes in renal cells, as well as for the synthesis of the osmolyte glycerophosphorylcholine. The characteristics of choline uptake in the renal epithelial cell line MDCK were investigated. In the range of physiological concentrations, choline entered MDCK cells, grown as a monolayer on solid support, via a specific sodium-independent transport system (apparent Km = 43 microM, apparent Vmax = 284 pmol/mg protein per 5 min). Cell ATP depletion, addition of KCl to the medium to reduce the cell membrane potential, and hemicholinium-3 (HC-3) inhibited choline uptake. Specific binding of [3H]HC-3 was detected on the apical membrane of cells grown on plastic dishes, whereas it occurred only on the basolateral domain of cells grown on permeant support. When growing cells on filter, choline uptake from the basolateral side was 10-times the apical uptake. This suggests that the choline carrier present at the apical domain of cells grown on solid support is either inactivated or no longer targeted to the apical but to the basolateral membrane of MDCK cells grown on filter.

Composition and physical properties of lipids from plasma membranes of dog kidney.

Lipid composition, physical state of major phospholipid classes and transbilayer migration of phosphatidylcholine have been determined in plasma membranes of the dog kidney. The lipid composition of brush-border membranes markedly differs from that of antiluminal membranes with respect to: (a) the total phospholipid content; (b) the cholesterol to phospholipid ratio (C/P); (c) the distribution of the major phospholipid classes. Sphingomyelin present in large amounts in both luminal and antiluminal membranes extracts exhibits a transition of phase between 20 and 44 degrees C approximately. In the range of temperature studied (5-55 degrees C) no phase transitions were detected for the other phospholipid species. Our data suggest that: (1) at physiological temperature the higher C/P ratio of brush-border membranes is in large part responsible for their lower fluidity; (2) both the relatively low cholesterol and high sphingomyelin contents contribute to the thermotropic transitions observed in intact membranes. Finally transbilayer migration of phosphatidylcholine in brush-border membranes is a very slow process with a half time of 6.5 h at 37 degrees C which compares with that of other biological membranes.

Temperature dependence of endocytosis in renal epithelial cells in culture.

Temperature dependence of fluid-phase endocytosis was determined in two renal epithelial cell lines, MDCK cells and LLC-PK1 cells, using Lucifer Yellow or horseradish peroxidase as markers. For both cell lines, grown on solid support as a confluent monolayer, biphasic curves of marker uptake vs. temperature were obtained. The changes in slope occurred around 27 degrees C, a critical temperature at which the lipids of the plasma membrane of MDCK cells enter in the gel state. Activation energies were significantly higher above 27 degrees C (15-22 kcal/mol) than below that critical temperature (9-12 kcal/mol). These data indicate that changes in membrane physical state have marked effects on endocytic processes. They suggest that two mechanisms, with different activation energies are involved in the fluid phase endocytosis by renal epithelial cells in culture.

Incorporation of exogenous phosphatidylcholine in the plasma membrane of MDCK cells by a specific transfer protein.

The possibility to introduce exogenous phosphatidylcholine (PC) in the plasma membrane of Madin-Darby canine kidney (MDCK) cells other than by fusion of liposomes with virus-infected cells (Van Meer, G. and Simons, K. (1983) J. Cell Biol. 97, 1365-1374) was studied. Monolayers of confluent MDCK cells grown on a permeable support were exposed to unilamellar vesicles of dipalmitoylphosphatidylcholine (DPPC), a phospholipid that does not exchange spontaneously, and were incubated with or without the PC-specific transfer protein (PC-TP), at 4 and 37 degrees C. Added either on the apical or basolateral side of monolayers grown in the presence of [14C]choline, PC-TP stimulated the transfer of 14C-labeled PC from the cell membrane to the liposomes, even at 4 degrees C. Conversely, PC-TP promoted the transfer, by a temperature-dependent process, of [3H]DPPC from liposomes to the cell plasma membrane. The amount of DPPC imported at 37 degrees C was higher than 100 pmol/well for apical incubations. The data demonstrate that, in MDCK cells: (a) PC-TP can modify the PC species present in the plasma membrane; (b) PC accounts for a significant amount of the polar lipids present in the external leaflet of the apical membrane domain.

Fate of a fluorescent inhibitor of endopeptidase-24.11 using enzyme-expressing MDCK cells. Modification of its cellular processing with a monoclonal antibody.

Neutral endopeptidase-24.11 (NEP) is a membrane-bound zinc metallopeptidase which cleaves biologically active peptides such as the enkephalins and atrial natriuretic peptide. Using the specific and fluorescent thiol inhibitor of the enzyme, N-[fluoresceinyl]-N'-[1-(6-(3-mercapto-2-benzyl-1-oxopropyl)-amino-1- hexyl]-thiocarbamide (FTI), the fate of the inhibitor-enzyme complex was investigated by videomicrofluorimetry using MDCK epithelial cells expressing the rabbit peptidase thanks to a retroviral expression vector. N-[3-(R,S)-[(hydroxyamino) carbonyl]-2-benzyl-1-oxopropyl]- glycine (HACBOGly) and the corresponding tritiated molecule were also used to measure the cellular pathway of inhibitor-NEP complexes. In the present paper, we demonstrate that, for short incubation times, the fluorescent probe preferentially labeled brush border membranes of the apical side of the MDCK cells. After more than 1 h incubation, a honeycomb pattern of fluorescence was observed in videomicrofluorimetry suggesting that part of the inhibitor was bound or localized close to the basolateral plasma membrane. Confocal experiments confirmed the transcytosis of FTI/NEP complex, from the apical to the basolateral domain. Using [3H]HACBOGly on filter-grown cells, after 2 and 4 h incubation at 37 degrees C, the percentage of basolateral membrane-bound molecules was estimated to be about 12 and 23%, respectively. The coincubation of the cells with FTI and 2B12, a monoclonal antibody raised against the rabbit enzyme, greatly modified the fluorescence pattern. A patchy fluorescence was observed for short incubation times, corresponding to cluster formation induced by antigen-antibody binding. For longer incubation times (> 1 h), in addition to the basolateral labeling, some intracellular fluorescent vesicles were observed essentially localized in the vicinity of the nucleus. The colocalization of FTI with Texas Red isothiocyanate-labeled Concanavalin A (TRITC-Con A) strongly suggests an endosomal/lysosomal internalization pathway when FTI was incubated in the presence of 2B12 mAb.

The use of atomic force microscopy for the observation of corneal epithelium surface.

PURPOSE: To evaluate the feasibility of imaging normal corneal epithelium by means of atomic force microscopy (AFM). METHODS: Twelve normal corneas from six albino rabbits were examined using a commercial atomic force microscope. Six corneas were examined in balanced salt solution after fixation in glutaraldehyde 2.5% and six without any fixation. Rectangular silicon nitride cantilevers with a spring constant of 10 to 20 mN/m were used. The measured forces after imaging were less than 100 pN. All reported images were made with 512x512-pixel definition with typical scan rates ranging from 1 to 5 Hz. RESULTS: High-quality images of corneal epithelium surface were obtained from fixed and unfixed specimens in magnifications ranging from x2000 to x2,000,000. Imaging of fixed specimens was always easier. In unfixed specimens fuzzy images were very common, probably because of the presence of the cell glycocalyx. AFM revealed the typical polygonal corneal epithelial cells. The cell surface was covered by microprojections; at cell borders the microprojections were arranged in two characteristic parallel rows. Craterlike formations were revealed in several specimens. The microprojections' morphology and their surface details were revealed using magnifications up to x2,000,000. Three-dimensional representation of the images facilitated better understanding of the surface topography. Measurements in horizontal and vertical plane were made using the section analysis tool. CONCLUSIONS: In this work the AFM parameters appropriate for corneal epithelium imaging in physiological medium were defined. AFM represents a new powerful tool for corneal epithelium imaging, and its application in this field warrants further investigation.