The polar head group of a novel insulin-sensitive glycophospholipid mimics insulin action on phospholipid methyltransferase

A phospholipid has been purified from rat liver membranes which copurified with an insulin-sensitive glycophospholipid isolated from H35 hepatoma cells. The polar head group of this phospholipid was generated by treatment with a phosphatidylinositol-specific phospholipase C from Staphylococcus aureus and purified through a C18 extraction column. Like insulin, the addition of this polar head group to isolated rat adipocytes inhibited the stimulatory effect of isoproterenol on phospholipid methyltransferase. The polar head group was also active on a subcellular fraction. The addition of the polar head group to microsomes isolated from isoproterenol-treated adipocytes produced a time-dependent inactivation of phospholipid methyltransferase, approaching basal activity. It is proposed that the effects of insulin on phospholipid methyltransferase may be mediated by this polar head group.     

Inhibition of cyclic AMP-dependent protein kinase by the polar head group of an insulin-sensitive glycophospholipid

A glycophospholipid has been purified from rat liver membranes and shown to copurify with an insulin-sensitive glycophospholipid isolated from H35 hepatoma cells. The polar head group of this glycophospholipid is a phospho-oligosaccharide generated by treatment with phosphatidylinositol-specific phospholipase C from Staphylococcus aureus. It has been proposed that this phospho-oligosaccharide, which is also generated in response to insulin, may play a role in insulin action. Incubation of the catalytic subunit of cyclic AMP-dependent protein kinase with this phospho-oligosaccharide inhibited the activity of the kinase to phosphorylate histone IIA, a purified preparation of phospholipid methyltransferase and kemptide, a phosphate-accepting peptide. Inhibition of kinase activity was dose-dependent and 50% inhibition of histone phosphorylation was demonstrated with a concentration of phospho-oligosaccharide of around 2 microM. This effect was demonstrated in the presence of ATP at concentrations up to 1 mM, indicating that the phospho-oligosaccharide acts at physiological concentrations of ATP and that it does not compete with this nucleotide for the same binding site in the kinase. Inhibition by the phospho-oligosaccharide of kinase activity could be reversed by dilution or dialysis and was not reproduced by up to 50 microM myo-inositol, glucosamine, galactose, myo-inositol 1-phosphate, glucosamine 1-phosphate, galactose 1-phosphate or phosphorylcholine. The inhibitory activity was resistant to mild acid treatment but was labile to treatment with alkali, exposure to nitrous acid or incubation with sodium periodate. The phospho-oligosaccharide had no effect on the phosphorylation of lysine-rich histone by rat brain protein kinase C and on the binding of cyclic AMP to a cyclic AMP-dependent protein kinase. In conclusion, the data in this study suggested that a phospho-oligosaccharide generated from an insulin-sensitive glycophospholipid may play a role in insulin action by modulating cyclic AMP-dependent protein kinase activity.     

Localization of phosphoprotein C23 in nucleoli by immunological methods

Antiserum to a major phosphorylated nucleolar protein. C23 (MW 103000, pI 5.2) from Novikoff hepatoma was produced in rabbits. By immunodiffusion analysis, the antiserum produced precipitin bands and with various crude extracts of nucleoli, but not with extranucleolar or cytosol fractions. The specificity of the antibody was assessed using acid-urea polyacrylamide gel electropherograms of acid-soluble nucleolar proteins in which the separated proteins were transferred to nitrocellulose sheets. The purified antibody reacted predominantly with protein C23 as visualized by the immunoperoxidase procedure. By the indirect immunofluorescence technique, protein C23 was localized predominantly to nucleoli of Novikoff hepatoma or normal rat liver cells. In Novikoff hepatoma cells, traces of fluorescence were seen near the inner layer of the nuclear envelope. Additional narrow regions of fluorescence extended from the nucleoli into the extranucleolar areas of some Novikoff cells. The nucleolar areas of fluorescence were smaller but brighter in the normal liver than in Novikoff hepatoma, consistent with the small size of rat liver nucleoli. These data indicate that the major location of protein C23 is the nucleolus.     

Comparison of cytosolic and nuclear poly(A) polymerases from rat liver and a hepatoma: structural and immunological properties and response to NI-type protein kinases

Poly(A) polymerases were purified from the cytosol fraction of rat liver and Morris hepatoma 3924A and compared to previously purified nuclear poly(A) polymerases. Chromatographic fractionation of the hepatoma cytosol on a DEAE-Sephadex column yielded approximately 5 times as much poly(A) polymerase as was obtained from fractionation of the liver cytosol. Hepatoma cytosol contained a single poly(A) polymerase species [48 kilodaltons (kDa)] which was indistinguishable from the hepatoma nuclear enzyme (48 kDa) on the basis of CNBr cleavage maps. Liver cytosol contained two poly(A) polymerase species (40 and 48 kDa). The CNBr cleavage patterns of these two enzymes were distinct from each other. However, the cleavage pattern of the 40-kDa enzyme was similar to that of the major liver nuclear poly(A) polymerase (36 kDa), and approximately three-fourths of the peptide fragments derived from the 48-kDa species were identical with those from the hepatoma enzymes (48 kDa). NI-type protein kinases from liver or hepatoma stimulated hepatoma nuclear and cytosolic poly(A) polymerases 4-6-fold. In contrast, the liver cytosolic 40- and 48-kDa poly(A) polymerases were stimulated only slightly or inhibited by similar units of the protein kinases. Antibodies produced in rabbits against purified hepatoma nuclear poly(A) polymerase reacted equally well with hepatoma nuclear and cytosolic enzyme but only 80% as well with the liver cytosolic 48-kDa poly(A) polymerase and not at all with liver cytosolic 40-kDa or nuclear 36-kDa enzymes. Anti-poly(A) polymerase antibodies present in the serum of a hepatoma-bearing rat reacted with hepatoma nuclear and cytosolic poly(A) polymerases to the same extent but only 40% as well with the liver cytosolic 48-kDa enzyme.(ABSTRACT TRUNCATED AT 250 WORDS)     

Purification of detergent-solubilized form and membrane-binding domain of rat γ-glutamyltransferase by immuno-affinity and hydrophobic chromatography

A new method to purify papain- or detergent-solubilized form (papain or detergent form) of γ-glutamyltransferase from rat hepatomas as well as from rat kidney by immuno-affinity column chromatography is presented. The antibody-column was prepared by coupling the anti-kidney papain form antibody, which had been purified by using a kidney papain form-Sepharose column, to CNBr-activated Sepharose 4B. The enzyme bound to the antibody-column was eluted with 0.04 M NH₄OH. By this method, detergent forms were purified 300 and 1600-fold in approx. 50% yields from rat kidney and rat ascites hepatoma AH 13, respectively, and the papain form was also purified 16 000-fold in a similar yield from primary hepatoma which has a very low activity of this enzyme. Preparations thus obtained apparently did not contain any peptide other than heavy and light subunit peptides of this enzyme on SDS-polyacrylamide gel electrophoresis. The detergent form of kidney enzyme was preferentially adsorbed to a hydrophobic column of aminooctyl-Sepharose, while the papain form was not, suggesting that the detergent form might be adsorbed to the column through hydrophobic interaction of the membrane-binding domain. The domain peptide was also purified by the hydropholic column after release from the detergent form by papain treatment. The molecular weight of the peptide was estimated to be about 16 000 on SDS-polyacrylamide gel electrophoresis. On double immunodiffusion, the domain peptide reacted with anti-detergent form antibody but not with anti-papain form antibody. The domain-specific antibody was also purified from the anti-detergent form antibody.     

Detection of the principal protein target of a hepatic carcinogen

Antiserum was prepared against the principal liver protein conjugate of the hepatic carcinogen, 3′-methyl-4-dimethylaminoazobenzene. One precipitin band was obtained when the antiserum reacted with the purified conjugate in double immunodiffusion gel analysis. The same anti-serum detected two proteins in rat liver cytosol. Of these two proteins, one was immunoreactively identical to the purified antigen; in contrast, the other protein was only partly identical to it. Absorption of the antiserum with rat kidney cytosol yielded specific antiserum that reacted only with the protein that was immunologically identical to the purified conjugate. That protein, detected in normal rat liver cytosol, is apparently the principal protein target of the azocarcinogens in liver carcinogenesis.     

Purification of detergent-solubilized form and membrane-binding domain of rat gamma-glutamyltransferase by immuno-affinity and hydrophobic chromatography

A new method to purify papain- or detergent-solubilized form (papain or detergent form) of gamma-glutamyltransferase from rat hepatomas as well as from rat kidney by immuno-affinity column chromatography is presented. The antibody-column was prepared by coupling the anti-kidney papain form antibody, which had been purified by using a kidney papain form-Sepharose column, to CNBr-activated Sepharose 4B. The enzyme bound to the antibody-column was eluted with 0.04 M NH4OH. By this method, detergent forms were purified 300 and 1600-fold in approx. 50% yields from rat kidney and rat ascites hepatoma AH 13, respectively, and the papain form was also purified 16 000-fold in a similar yield from primary hepatoma which has a very low activity of this enzyme. Preparations thus obtained apparently did not contain any peptide other than heavy and light subunit peptides of this enzyme on SDS-polyacrylamide gel electrophoresis. The detergent form of kidney enzyme was preferentially absorbed to a hydrophobic column of aminooctyl-Sepharose, while the papain form was not, suggesting that the detergent form might be adsorbed to the column through hydrophobic interaction of the membrane-binding domain. The domain peptide was also purified by the hydrophobic column after release from the detergent form by papain treatment. The molecular weight of the peptide was estimated to be about 16 000 on SDS-polyacrylamide gel electrophoresis. On double immunodiffusion, the domain peptide reacted with anti-detergent form antibody but not with anti-papain form antibody. The domain-specific antibody was also purified from the anti-detergent form antibody.     

Comparison of ornithine decarboxylase from rat liver, rat hepatoma and mouse kidney.

Comparisons were made of ornithine decarboxylase isolated from Morris hepatoma 7777, thioacetamide-treated rat liver and androgen-stimulated mouse kidney. The enzymes from each source were purified in parallel and their size, isoelectric point, interaction with a monoclonal antibody or a monospecific rabbit antiserum to ornithine decarboxylase, and rates of inactivation in vitro, were studied. Mouse kidney, which is a particularly rich source of ornithine decarboxylase after androgen induction, contained two distinct forms of the enzyme which differed slightly in isoelectric point, but not in Mr. Both forms had a rapid rate of turnover, and virtually all immunoreactive ornithine decarboxylase protein was lost within 4h after protein synthesis was inhibited. Only one form of ornithine decarboxylase was found in thioacetamide-treated rat liver and Morris hepatoma 7777. No differences between the rat liver and hepatoma ornithine decarboxylase protein were found, but the rat ornithine decarboxylase could be separated from the mouse kidney ornithine decarboxylase by two-dimensional gel electrophoresis. The rat protein was slightly smaller and had a slightly more acid isoelectric point. Studies of the inactivation of ornithine decarboxylase in vitro in a microsomal system [Zuretti & Gravela (1983) Biochim. Biophys. Acta 742, 269-277] showed that the enzymes from rat liver and hepatoma 7777 and mouse kidney were inactivated at the same rate. This inactivation was not due to degradation of the enzyme protein, but was probably related to the formation of inactive forms owing to the absence of thiol-reducing agents. Treatment with 1,3-diaminopropane, which is known to cause an increase in the rate of degradation of ornithine decarboxylase in vivo [Seely & Pegg (1983) Biochem. J. 216, 701-717] did not stimulate inactivation by microsomal extracts, indicating that this system does not correspond to the rate-limiting step of enzyme breakdown in vivo.     

Gamma-glutamyltransferase from azo dye induced hepatoma and fetal rat liver. Similarities in their kinetic and immunological properties.

Some properties of gamma-glutamyltransferase ((gamma-glutamyl)-peptide: amino-acid gamma-glutamyltransferase EC 2.3.2.2) from azo dye induced hepatoma and fetal rat liver were studied using kinetic and immunological criteria. There was no significant difference between the hepatoma enzyme and fetal rat liver enzyme in some of their catalytic properties. Antisera against the purified hepatoma enzyme also reacted to the fetal rat liver enzyme in the inhibition test and the precipitin reaction. A structural similarity between the hepatoma enzyme and fetal rat liver enzyme was observed and the acquirement of fetal characteristics in hepatoma was discussed.     

Cathepsin S. The cysteine proteinase from bovine lymphoid tissue is distinct from cathepsin L (EC 3.4.22.15).

Cathepsin S was purified from bovine spleen by acid autolysis, (NH4)2SO4 fractionation and chromatography on CM-Sephadex C-50, CM-cellulose and activated-thiol-Sepharose. Cathepsin L was isolated from lysosomal fractions of rat liver, rat kidney and bovine liver. Generally, cathepsin L was bound tightly to CM-Sephadex C-50. Preparations of cathepsin L from rat liver, rat kidney and bovine liver were shown to have kinetic constants for the substrate benzyloxycarbonyl-Phe-Arg-7-(4-methyl)coumarylamide in the same range (Km 2-3 microM). Benzyloxycarbonyl-Phe-Phe-diazomethane proved to be a sensitive irreversible inhibitor of cathepsin L from different species. Cathepsin S differed in all these characteristics from cathepsin L. A polyclonal antibody to cathepsin L from rat reacted with bovine cathepsin L but not with bovine cathepsin S.     

Structurally and immunologically distinct poly(A) polymerases in rat liver. Occurrence of a tumor-type enzyme in normal liver

Poly(A) polymerases purified from rat liver nuclei consisted of two distinct species, a predominant enzyme of Mr = 38,000 and a minor one of Mr = 48,000. Prior to extensive purification, the minor enzyme constituted approximately 1% of the total liver poly(A) polymerase. Poly(A) polymerase purified from a rat tumor, Morris hepatoma 3924A, was comprised of a single species of Mr = 48,000 which was identical to the minor liver enzyme with respect to chromatographic and immunological characteristics. Gel filtration on Sephacryl S-200 using 0.3 M NaCl for elution showed that the major liver poly(A) polymerase had a molecular weight of 156,000, which corresponded to a tetramer of the 38-kDa polypeptide, whereas the hepatoma and minor liver 48-kDa species existed as dimers with a molecular weight of 96,000. Fractionation by Sephacryl S-200 resulted in complete loss of both liver poly(A) polymerase activities which could be restored by exogenous N1-type protein kinase. Following CNBr cleavage, the 48-kDa poly(A) polymerase from liver and hepatoma exhibited nearly identical peptide maps which were distinct from that of the major liver enzyme (38 kDa). Antibodies raised against tumor poly(A) polymerase reacted with the 48-kDa polypeptide but not with the 38-kDa liver enzyme. Immune complex formation was observed between seven of the eight CNBr cleavage products derived from the 48-kDa polypeptide of both liver and hepatoma. It is concluded that distinct genes in rat liver code for two structurally and immunologically unique nuclear poly(A) polymerases, one of which is identical to the enzyme from the hepatoma.     

Structure and properties of an under-sulfated heparan sulfate proteoglycan synthesized by a rat hepatoma cell line

A rat hepatoma cell line was shown to synthesize heparan sulfate and chondroitin sulfate proteoglycans. Unlike cultured hepatocytes, the hepatoma cells did not deposit these proteoglycans into an extracellular matrix, and most of the newly synthesized heparan sulfate proteoglycans were secreted into the culture medium. Heparan sulfate proteoglycans were also found associated with the cell surface. These proteoglycans could be solubilized by mild trypsin or detergent treatment of the cells but could not be displaced from the cells by incubation with heparin. The detergent-solubilized heparan sulfate proteoglycan had a hydrophobic segment that enabled it to bind to octyl- Sepharose. This segment could conceivably anchor the molecule in the lipid interior of the plasma membrane. The size of the hepatoma heparan sulfate proteoglycans was similar to that of proteoglycans isolated from rat liver microsomes or from primary cultures of rat hepatocytes. Ion-exchange chromatography on DEAE-Sephacel indicated that the hepatoma heparan sulfate proteoglycans had a lower average charge density than the rat liver heparan sulfate proteoglycans. The lower charge density of the hepatoma heparan sulfate can be largely attributed to a reduced number of N-sulfated glucosamine units in the polysaccharide chain compared with that of rat liver heparan sulfate. Hepatoma heparan sulfate proteoglycans purified from the culture medium had a considerably lower affinity for fibronectin-Sepharose compared with that of rat liver heparan sulfate proteoglycans. Furthermore, the hepatoma proteoglycan did not bind to the neoplastic cells, whereas heparan sulfate from normal rat liver bound to the hepatoma cells in a time-dependent reaction. The possible consequences of the reduced sulfation of the heparan sulfate proteoglycan produced by the hepatoma cells are discussed in terms of the postulated roles of heparan sulfate in the regulation of cell growth and extracellular matrix formation.     

Identification of a novel insulin-sensitive glycophospholipid from H35 hepatoma cells

This study identifies and partially characterizes an insulin-sensitive glycophospholipid in H35 hepatoma cells. The incorporation of [3H]glucosamine into cell lipids was investigated. A major labeled lipid was purified by sequential thin layer chromatography using first an acid followed by a basic solvent system. After hydrochloric acid hydrolysis and sugar analysis by thin layer chromatography, 80% of the radioactivity in the purified lipid was found to comigrate with glucosamine. H35 cells were prelabeled with [3H]glucosamine for either 4 or 24 h and treated with insulin causing a dose-dependent stimulation of turnover of the glycophospholipid which was detected within 1 min. The purified glycolipid was cleaved by nitrous acid deamination indicating that the glucosamine C-1 was linked to the lipid moiety through a glycosidic bond. [14C]Ethanolamine, [3H]inositol, and [3H]sorbitol were not incorporated into the purified glycolipid. The incorporation of various fatty acids into this glycolipid was also studied. [3H]Palmitate was found to be preferentially incorporated while myristic acid, stearic acid, oleic acid, linoleic acid, linolenic acid, and arachidonic acid were either not incorporated or incorporated less than 10% of palmitate. The purified glycolipid labeled with [3H]palmitate was cleaved by treatment with phospholipase A2 but was resistant to mild alkali hydrolysis suggesting the presence of a 1-hexadecyl,2-palmitoyl-glyceryl moiety in the purified lipid. Treatment of labeled glycophospholipid with phosphatidylinositol-specific phospholipase C from Staphylococcus aureus generated a compound migrating as 1-alkyl,2-acyl-glycerol and a polar head group with a size in the range from 800 to 3500. These findings coupled with the nitrous acid deamination demonstrate that glucosamine was covalently linked through a phosphodiester bond to the glyceryl moiety of the purified glycolipid. These findings suggest that insulin acts on this glycophospholipid by stimulating an insulin-sensitive phospholipase C. This unique glycophospholipid may play an important role in insulin action by serving as precursor of insulin-generated mediators.     

Monoclonal antibodies recognizing different epitopes of the 27-kDa gap junction protein from rat liver

Monoclonal antibodies (2-3E2, 6-3G11, and 7-3H6) against gap junction plaques purified from rat liver were prepared and characterized. Immunoblot analysis of liver gap junctions revealed that all three antibodies reacted with the 27-kDa protein, but not with the 22-kDa one. The 2-3E2 and 6-3G11 antibodies both reacted with the 27-kDa protein in gap junctions purified from livers of the rat, mouse, rabbit, and guinea pig; the 7-3H6 antibody, however, failed to react with the 27-kDa protein from guinea pig liver. The 7-3H6 antibody reacted strongly with the 24- to 26-kDa degradation products of the 27-kDa protein. Indirect immunofluorescence showed that the 6-3G11 and 7-3H6 antibodies both gave the same specific fluorescence labeling on rat liver cryosections, suggesting that these two antibodies recognized the cytoplasmic sites of the 27-kDa protein. Immunoblot analysis of protease-digested fragments from the 27-kDa protein revealed that the 7-3H6 antibody reacted with the 24- and 17-kDa fragments (including portions of the carboxyl-terminal domain of the 27-kDa protein) produced with endoproteinases Arg-C and Lys-C, respectively. Immunoblot analysis of CNBr fragments of the 27-kDa protein revealed that all three antibodies reacted with the 10-kDa fragment, which is thought to be the carboxyl-terminal domain of the 27-kDa protein. These results demonstrate that three monoclonal antibodies recognize different epitopes of the cytoplasmic sites (probably the carboxyl-terminal domain) of the 27-kDa liver gap junction protein.     

Multiple nuclear protein kinase activities in rat liver and hepatoma 3924A.

Multiple protein kinase activities were isolated from nuclei of rat and hepatoma 3924A, and purified 40- to 140-fold, respectively. Hepatic protein kinase-I exhibited high activity with casein as substrate, but was relatively inactive with either liver and hepatoma chromatin or mixed histone. In contrast, hepatoma protein kinase-I showed equivalent activity with casein and liver chromatin. Protein kinase-IIA, -IIB and-IIC from both tissues were more active with liver chromatin in comparison to casein and hepatoma chromatin, and exhibited similar electrophoretic profiles of 32P-chromatin.     

Isolation of two cholic acid-binding proteins from rat liver cytosol using affinity chromatography

Using an affinity matrix coupled with cholic acid, two proteins that recognise bile acids were isolated from rat liver cytosol. One protein of molecular weight 68 000 was immunologically identical to rat albumin. The other protein was of molecular weight 46 000. On discontinuous sodium dodecyl sulphate-polyacrylamide gel electrophoresis the 46 000 molecular weight protein dissociated to a single band with an RF value identical to the Yb subunit of the bromosulphophthalein-binding fraction (Y-fraction) of whole liver cytosol. The monomers of purified ligandin under these conditions resolved into two bands which corresponded to the Ya and Yc subunits of liver cytosol Y-fraction. Anti-serum to the purified ligandin reacted monospecifically with purified ligandin and whole liver cytosol, but did not cross-react with the Yb dimer eluted from the affinity column. The Yb dimer was shown to possess glutathione-S-transferase activity with a substrate specificity distinct from ligandin but similar to glutathione-S-transferase C. Cholic acid inhibited the catalytic activity of the transferase.     

A universal lipid exchange protein from rat hepatoma.

The postmicrosomal protein fraction from rat hepatoma 27 adjusted to pH 5.1 stimulates phospholipid exchange between rat liver microsomes and mitochondria with higher rates and in a less specific way than the corresponding fraction from rat liver. A phospholipid exchange protein has been purified to homogeneity from the hepatoma pH-5.1 supernatant by gel filtration on Sephadex G-75 and ion-exchange chromatography on carboxymethylcellulose. The isolated protein had a molecular weight of 11200 as determined by electrophoresis on polyacrylamide in the presence of dodecyl sulfate and of 11168 as calculated from the amino acid composition. Isoelectric focusing showed a single band at pH 5.2. in the assay system rat liver microsomes leads to mitochondria the protein exhibits a complete lack of substrate specificity transferring all the major microsomal phospholipids to about the same extent. The possible role of the isolated phospholipid exchange protein in the chemical dedifferentiation of hepatoma cell membranes is discussed.     

Gangliosides of plasma membranes from normal rat liver and Morris hepatoma.

Plasma membranes were isolated from normal rat liver and Morris hepatoma 5123tc by discontinuous sucrose gradient centrifugation. There was an average two and one-halffold enrichment of gangliosides in plasma membranes from normal liver and hepatoma as compared with their respective whole cells. The amount of total gangliosides in plasma membranes from hepatoma was eight times greater than that found in normal liver. This increase resulted from a fivefold increase in hematosides, an eightfold increase in monosialogangliosides and a twenty-twofold increase in disialogangliosides. Trisialogangliosides were present in normal liver but not in hepatoma.     

Identification of an amino acid-regulated mRNA from rat liver as the mammalian equivalent of bacterial ribosomal protein L22

Amino acid deprivation of rat hepatoma cells induced the levels of a 612-base pair mRNA termed ASI (Shay, N. F., Nick, H. S., and Kilberg, M. S. (1990) J. Biol. Chem. 265, 17844-17848). The ASI mRNA was present at levels equal to or greater than actin in every rat tissue tested. The corresponding full-length cDNA was cloned, and the present report demonstrates that the deduced 184-residue amino acid sequence shares greater than 30% identity to a number of bacterial and chloroplast L22 ribosomal proteins, including those from Escherichia coli and Halobacterium halobium. A monospecific anti-peptide antibody was produced that upon immunochemical analysis of subcellular fractions of rat liver recognized a band in the microsomal fraction and, more specifically, reacted with a single polypeptide in the ribosomal large subunit fraction. The antibody did not react with any proteins of the mitochondrial large subunit, but did recognize a protein in human liver homogenate at the same relative mobility (23 kDa) as that observed for rat liver.     

Identification of a transformation-sensitive 110-kDa plasma membrane glycoprotein of rat hepatocytes

Monoclonal antibodies were used to define cell surface antigens which are present on rat hepatocytes but are absent from hepatoma cells. One monoclonal antibody, referred to as Be 9.2, recognizes a major component of purified rat liver plasma membranes with a Mr of 110 000. This antigen (gp110) was not found in the transplantable Morris hepatoma 9121 and 7777 nor on two cultured hepatoma cell lines. Isoelectric focussing showed that gp110 is a very acidic membrane component with an isoelectric point of 3.6 to 3.8. Treatment with neuraminidase reduced the Mr to 95 000. Gp110 while bound to the membrane was resistant to trypsin, but sensitive to papain. The tissue distribution of gp110 was examined by indirect immunofluorescence in frozen sections. The antigen was found on the bile canalicular domain of hepatocytes, the microvillous zone of enterocytes of the small intestinal villi, the luminal plasma membrane of acinar cells in the submaxillary and extraorbital gland and of epithelial cells of the vesicular gland. Gp110 could not be detected in the stomach, pancreas, large intestine, kidney, thymus, spleen, heart, lung, muscle cells and fibers and in the brain. Identical results were obtained by the use of an antiserum raised against purified gp110. They confirm the transformation-sensitive character of this glycoprotein. A possible identity with dipeptidyl peptidase IV and aminopeptidase M, which have similar molecular weights and are also present in rat liver on the bile canalicular domains, could be excluded. The results suggest that the loss of gp110 might be regarded as a marker for transformation or dedifferentiation of hepatocytes.