557th MEETING, LIVERPOOL
747
enzymes is due to a more specificprocess than generalizedcelldamage or loss (Holdsworth & Coleman, 1975). We have now attempted to make an initial separation of these enzymes from other biliary proteins, before investigating their relationship with the bile-canaliculus region of the hepatocyte plasma membrane. Fresh sheep bile was treated with solid protamine sulphate and centrifuged to remove the majority of pigments and lipids. The supernatant was concentrated tenfold with a Diaflo membrane and then applied to columns (1.2cm x 8.0cm) of concanavalin A-Sepharose 4B, equilibrated with 0.1 % deoxycholate in 0.1 M-Tris-HC1, pH 7.7. Almost all the biliary protein was removed by washing the columns with the equilibrating buffer. A small amount of protein containing the enzymes was eluted from the column with a-methyl D-glucoside, suggesting the presence of glycoproteins. The appearance of bands, stainable with periodic acid-Schiff reagent, after polyacrylamidegel electrophoresis in the presence of sodium dodecyl sulphate, provided further evidence for the presence of glycoproteins. When a discontinuous gradient of a-methyl Dglucoside was used to elute the affinity column, 5’-nucleotidase and alkaline phosphatase were eluted before alkaline phosphodiesterase I and L-leucine 8-naphthylamidase. All four enzymes have been identified as glycoproteins in preparations containing liver plasma membranes (Evans, 1974; Evans & Gurd, 1973; Kaplan, 1972; Sternes & Behal, 1974). The glycoprotein nature of these same enzymes in bile not only gives a useful preliminary stage in their purification, but also indicates a possible relationship between the membrane-bound and biliary enzymes. Isolation and purification of some of these enzymes from liver plasma membranes involves an initial solubilization of the enzymes with detergents (Evans & Gurd, 1973; Evans, 1974; Widnell & Unkeless, 1968). It is possible that the release of enzymes into the bile is also related to a solubilization with detergents, since an important stage in bile formation is the output of bile salts from the liver cell and their accumulation in the bile canaliculus. We thank the Medical Research Council for financial support. Evans, W. H. (1974) Nature (London) New Biol. 250,391-394 Evans, W. H. & Gurd, J. W. (1973) Biochem.J. 133,189-199 Holdsworth, G. & Coleman, R. (1975) Biochim. Biophys. Acta 389,47-50 Kaplan, M. M. (1972) Gastreoenterology62,452-468 Sternes, W. L. & Behal(l974) Biochemistry 13,3221-3227 Widnell, C . & Unkeless, J. C. (1968) Proc. Natl. Acad. Sci. U.S.A.61, 1050-1057
Effects of Detergents on Erythrocyte Membranes: Different Patterns of Solubilizationof the Membrane Proteins by Dihydroxy and Trihydroxy Bile Salts ROGER COLEMAN and GEORGE HOLDSWORTH Department of Biochemistry, University of Birmingham, P.O. Box 363, Birmingham B15 2TT, U.K. Human erythrocyte ‘ghosts’ were incubated at 37°C in an iso-osmotic medium with a range of concentrations of several detergents. Solubilized (supernatant) and residue (pellet) materials were obtained by centrifugation for 1h at 150000g, and the lipid and protein patterns studied by t.1.c. and sodium dodecyl sulphate-polyacrylamide-gel electrophoresis respectively. Triton X-100,a synthetic detergent with a low hydrophilic/lipophilic balance, solubilized up to 95 % of the phospholipid and about 70 % of the protein. The solubilized proteins were preferentially enriched in polypeptide bands 3, 4.1,4.2, 4.5 (as described by Steck, 1975),and also contained some of bands 5,6 and 7. Many of these polypeptides have been identified as intrinsic proteins of this membrane. Yu et al. (1973) and Vol. 3
748
BIOCHEMICAL SOCIETY TRANSACTIONS
Kirkpatrick et al. (1974) have observed similar effects of Triton X-100 on erythrocyte membranes. Cholate, a trihydroxy bile salt, solubilizes essentially all the phospholipids of the membrane, but only about 40% of the protein. The polypeptide profile of the solubilized proteins was quite distinct from that of the proteins solubilized by Triton X-100; it contained essentially bands 1 and 2, with some of bands 5 and 6. These polypeptide bands largely represent the extrinsic proteins of this membrane and contain ‘spectrin’. Dehydrocholate, a synthetic keto analogue of cholate, which is even more hydrophilic than cholate, and does not readily form micelles, did not solubilize either protein or phospholipid from the membrane. Deoxycholate, which is less hydrophilic than cholate owing to the possession of only two hydroxyl groups, caused complete solubilization of all components of the membrane. When cholate was added to sub-maximal amounts of deoxycholate, it increased the solubilization of phospholipid, but had no effect on the extent of protein solubilization. The residue remaining after Triton was shown to be a filamentous material (Yu et al., 1973), largely containing spectrin, together with some sphingomyelin and cholesterol. On the other hand, the residue remaining after cholate treatment resembled a small ‘ghost cell’ when examined by phase-contrast microscopy, and gave the familiar trilamellar feature of intact membranes when thin-sectioned for electron microscopy. The proteins of these delipidated ‘membranes’ were essentially those thought normally to be inserted into the lipid layer in the intact membrane. These solubilization studies show that dihydroxy bile salts are potentially more damaging to membranes than trihydroxy bile salts in that, in addition to releasing phospholipids, they are able to release proteins normally embedded in the lipid layer. Normal bile contains both trihydroxy and dihydroxy bile salts, usually with the former in considerable excess. Several of the morphological and metabolic effects of liver damage seen during cholestasis (from a variety of causes) appear to be better correlated with the accumulation of dihydroxy rather than trihydroxy bile salts (see Moritz & Snodgrass, 1972; Heaton, 1972; Schaffner, 1973). In addition, when dihydroxy bile salts are present in abnormal amounts in the large intestine, they cause increased permeability and fluid loss. Such effects are not observed with trihydroxy bile salts (Mekhjian et al., 1971 ;Sladen & Harries, 1972; Feldman et al., 1973). It may be that some aspects of the toxicity of dihydroxy bile salts is related to their ability to interact with and release intrinsic proteins from membranes. We thank the Medical Research Council for financial support. Feldrnan, S., Reinhard, M. & Willson, C. (1973) J. Pharm. Sci. 62,1961-1964 Heaton, K. W. (1972) Bile Salts in Health and Disease, Churchill, Edinburgh Kirkpatrick, F. H., Gordesky, S. E. & Marinetti, G. V. (1974) Biochim. Biophys. Acfu 345, 154161
Mekhjian, H. S., Phillips, S. F. &Hoffman (1971) J. Clin. Znoest. 50, 1569-1584 Moritz, M. & Snodgrass, P. J. (1972) Gastroenterology 62,93-100 Schaffner, F. (1972) Helu. Med. Acta37, 183-192 Sladen, G . E. & Harries, J. L. (1972) Biochim. Biophys. Acta 288,443-456 Steck, T. L. (1974) J. Cell Biol. 62, 1-19 Yu, J., Fischrnan, D. A. & Steck, T. L. (1973) J. Supramol. Struct. 1,233-248
1975
747
enzymes is due to a more specificprocess than generalizedcelldamage or loss (Holdsworth & Coleman, 1975). We have now attempted to make an initial separation of these enzymes from other biliary proteins, before investigating their relationship with the bile-canaliculus region of the hepatocyte plasma membrane. Fresh sheep bile was treated with solid protamine sulphate and centrifuged to remove the majority of pigments and lipids. The supernatant was concentrated tenfold with a Diaflo membrane and then applied to columns (1.2cm x 8.0cm) of concanavalin A-Sepharose 4B, equilibrated with 0.1 % deoxycholate in 0.1 M-Tris-HC1, pH 7.7. Almost all the biliary protein was removed by washing the columns with the equilibrating buffer. A small amount of protein containing the enzymes was eluted from the column with a-methyl D-glucoside, suggesting the presence of glycoproteins. The appearance of bands, stainable with periodic acid-Schiff reagent, after polyacrylamidegel electrophoresis in the presence of sodium dodecyl sulphate, provided further evidence for the presence of glycoproteins. When a discontinuous gradient of a-methyl Dglucoside was used to elute the affinity column, 5’-nucleotidase and alkaline phosphatase were eluted before alkaline phosphodiesterase I and L-leucine 8-naphthylamidase. All four enzymes have been identified as glycoproteins in preparations containing liver plasma membranes (Evans, 1974; Evans & Gurd, 1973; Kaplan, 1972; Sternes & Behal, 1974). The glycoprotein nature of these same enzymes in bile not only gives a useful preliminary stage in their purification, but also indicates a possible relationship between the membrane-bound and biliary enzymes. Isolation and purification of some of these enzymes from liver plasma membranes involves an initial solubilization of the enzymes with detergents (Evans & Gurd, 1973; Evans, 1974; Widnell & Unkeless, 1968). It is possible that the release of enzymes into the bile is also related to a solubilization with detergents, since an important stage in bile formation is the output of bile salts from the liver cell and their accumulation in the bile canaliculus. We thank the Medical Research Council for financial support. Evans, W. H. (1974) Nature (London) New Biol. 250,391-394 Evans, W. H. & Gurd, J. W. (1973) Biochem.J. 133,189-199 Holdsworth, G. & Coleman, R. (1975) Biochim. Biophys. Acta 389,47-50 Kaplan, M. M. (1972) Gastreoenterology62,452-468 Sternes, W. L. & Behal(l974) Biochemistry 13,3221-3227 Widnell, C . & Unkeless, J. C. (1968) Proc. Natl. Acad. Sci. U.S.A.61, 1050-1057
Effects of Detergents on Erythrocyte Membranes: Different Patterns of Solubilizationof the Membrane Proteins by Dihydroxy and Trihydroxy Bile Salts ROGER COLEMAN and GEORGE HOLDSWORTH Department of Biochemistry, University of Birmingham, P.O. Box 363, Birmingham B15 2TT, U.K. Human erythrocyte ‘ghosts’ were incubated at 37°C in an iso-osmotic medium with a range of concentrations of several detergents. Solubilized (supernatant) and residue (pellet) materials were obtained by centrifugation for 1h at 150000g, and the lipid and protein patterns studied by t.1.c. and sodium dodecyl sulphate-polyacrylamide-gel electrophoresis respectively. Triton X-100,a synthetic detergent with a low hydrophilic/lipophilic balance, solubilized up to 95 % of the phospholipid and about 70 % of the protein. The solubilized proteins were preferentially enriched in polypeptide bands 3, 4.1,4.2, 4.5 (as described by Steck, 1975),and also contained some of bands 5,6 and 7. Many of these polypeptides have been identified as intrinsic proteins of this membrane. Yu et al. (1973) and Vol. 3
748
BIOCHEMICAL SOCIETY TRANSACTIONS
Kirkpatrick et al. (1974) have observed similar effects of Triton X-100 on erythrocyte membranes. Cholate, a trihydroxy bile salt, solubilizes essentially all the phospholipids of the membrane, but only about 40% of the protein. The polypeptide profile of the solubilized proteins was quite distinct from that of the proteins solubilized by Triton X-100; it contained essentially bands 1 and 2, with some of bands 5 and 6. These polypeptide bands largely represent the extrinsic proteins of this membrane and contain ‘spectrin’. Dehydrocholate, a synthetic keto analogue of cholate, which is even more hydrophilic than cholate, and does not readily form micelles, did not solubilize either protein or phospholipid from the membrane. Deoxycholate, which is less hydrophilic than cholate owing to the possession of only two hydroxyl groups, caused complete solubilization of all components of the membrane. When cholate was added to sub-maximal amounts of deoxycholate, it increased the solubilization of phospholipid, but had no effect on the extent of protein solubilization. The residue remaining after Triton was shown to be a filamentous material (Yu et al., 1973), largely containing spectrin, together with some sphingomyelin and cholesterol. On the other hand, the residue remaining after cholate treatment resembled a small ‘ghost cell’ when examined by phase-contrast microscopy, and gave the familiar trilamellar feature of intact membranes when thin-sectioned for electron microscopy. The proteins of these delipidated ‘membranes’ were essentially those thought normally to be inserted into the lipid layer in the intact membrane. These solubilization studies show that dihydroxy bile salts are potentially more damaging to membranes than trihydroxy bile salts in that, in addition to releasing phospholipids, they are able to release proteins normally embedded in the lipid layer. Normal bile contains both trihydroxy and dihydroxy bile salts, usually with the former in considerable excess. Several of the morphological and metabolic effects of liver damage seen during cholestasis (from a variety of causes) appear to be better correlated with the accumulation of dihydroxy rather than trihydroxy bile salts (see Moritz & Snodgrass, 1972; Heaton, 1972; Schaffner, 1973). In addition, when dihydroxy bile salts are present in abnormal amounts in the large intestine, they cause increased permeability and fluid loss. Such effects are not observed with trihydroxy bile salts (Mekhjian et al., 1971 ;Sladen & Harries, 1972; Feldman et al., 1973). It may be that some aspects of the toxicity of dihydroxy bile salts is related to their ability to interact with and release intrinsic proteins from membranes. We thank the Medical Research Council for financial support. Feldrnan, S., Reinhard, M. & Willson, C. (1973) J. Pharm. Sci. 62,1961-1964 Heaton, K. W. (1972) Bile Salts in Health and Disease, Churchill, Edinburgh Kirkpatrick, F. H., Gordesky, S. E. & Marinetti, G. V. (1974) Biochim. Biophys. Acfu 345, 154161
Mekhjian, H. S., Phillips, S. F. &Hoffman (1971) J. Clin. Znoest. 50, 1569-1584 Moritz, M. & Snodgrass, P. J. (1972) Gastroenterology 62,93-100 Schaffner, F. (1972) Helu. Med. Acta37, 183-192 Sladen, G . E. & Harries, J. L. (1972) Biochim. Biophys. Acta 288,443-456 Steck, T. L. (1974) J. Cell Biol. 62, 1-19 Yu, J., Fischrnan, D. A. & Steck, T. L. (1973) J. Supramol. Struct. 1,233-248
1975
