Although the transport properties of human erythrocyte water channels have been well characterized, the identity of the protein(s) mediating water flow remains unclear. Recent evidence that glucose carriers can conduct water raised the possibility that the glucose carrier, which is abundant in human erythrocytes, is the water channel. To test this possibility, water permeabilities and glucose fluxes were measured in large unilamellar vesicles (LUV) containing human erythrocyte lipid alone (lipid LUV), reconstituted purified human erythrocyte glucose carrier (Glut1 LUV), or reconstituted glucose carrier in the presence of other human erythrocyte ghost proteins (ghost LUV). In glucose and ghost LUV, glucose carriers were present at 25% of the density of native erythrocytes, were oriented randomly in the bilayer, and exhibited characteristic inhibition of glucose flux when exposed to cytochalasin B. Osmotic water permeability (Pf, in centimeters per second; n = 4) averaged 0.0012 +/- 0.00033 in lipid LUV, 0.0032 +/- 0.0015 in Glut1 LUV, and 0.006 +/- 0.0014 in ghost LUV. Activation energies of water flow for the three preparations ranged between 10 and 13 kcal/mol; p-(chloromercuri)benzenesulfonate (pCMBS), an organic mercurial inhibitor of erythrocyte water channels, and cytochalasin B did not alter Pf. These results indicate that reconstitution of glucose carriers at high density increases water permeability but does not result in water channel activity. However, because the turnover number of reconstituted carriers is reduced from that of native carriers, experiments were also performed on erythrocyte ghosts with intact water channel function. In ghosts, Pf averaged 0.038 +/- 0.013 (n = 9), while the activation energy for water flow averaged 3.0 +/- 0.3 kcal/mol.(ABSTRACT TRUNCATED AT 250 WORDS)

Bile salts have been hypothesized to mediate cytotoxicity by increasing membrane permeability to aqueous solutes. We examined whether submicellar bile salt concentrations affect model and native membrane permeability to small uncharged molecules such as water, urea, and ammonia. Osmotic water permeability (Pf) and urea permeability were measured in large unilamellar vesicles composed with egg yolk phosphatidylcholine (EYPC) +/- cholesterol (Ch) or rat liver microsomal membranes by monitoring self-quenching of entrapped carboxyfluorescein (CF). Ammonia permeability was determined utilizing the pH dependence of CF fluorescence. Submicellar bile salt concentrations did not significantly alter Pf of EYPC +/- Ch or rat liver microsomal membranes. At taurodeoxycholate (TDC) or tauroursodeoxycholate concentrations approaching those that solubilized membrane lipids, CF leakage occurred from vesicles, but Pf remained unchanged. Higher bile salt concentrations (0.5-2 mM TDC) did not alter Pf of equimolar EYPC/Ch membranes. The activation energy for transmembrane water flux was unchanged (12.1 +/- 1.2 kcal/mol for EYPC) despite the presence of bile salts in one or both membrane hemileaflets, suggesting strongly that bile salts do not form transmembrane pores that facilitate water flux. Furthermore, submicellar bile salt concentrations did not increase membrane permeability to urea or ammonia. We conclude that at submicellar concentrations, bile salts do not form nonselective convective channels that facilitate transmembrane transport of small uncharged molecules. These results suggest that bile salt-mediated transport of specific substrates, rather than nonselective enhancement of membrane permeability, underlies bile salt cytotoxicity for enterocytes and hepatocytes.

No abstract