Structure of isolated plant Golgi apparatus revealed by negative staining.

Sucrose-gradient-purified dictyosomes of plant Golgi apparatus appear, after glutaraldehyde stabilization, as stacks of highly fenestrate and tubate cisternae when negatively stained with phosphotungstic acid, shadowed with heavy metal, or OsO(4)-stained in thin section. The tubular proliferations (diameter 200 to 400 A) extend for several microns from the central region and are united at intervals into an anastomosing network. Associated with the tubules are two kinds of vesicles which are distinguishable on the basis of texture, size, shape, and staining characteristics. One vesicle type is rough-surfaced, nearly spherical, and of uniform dimensions (diameter approximately 600 A). Metal shadowing shows that these vesicles remain spherical after drying. The other vesicle type is smooth-surfaced and varies in both size and shape. Intercisternal elements are revealed, by negative staining, on the surface of internal cisternae after fragmentation of the dictyosome. The progressive differentiation of cisternae from the forming face to the maturing face is observed in thin sections of these isolated preparations. The morphological characteristics observed in negatively stained dictyosomes indicate regions of functional specialization within the dictyosome cisternae and reveal a dictyosome structure more extensive than that envisioned from sections.

Isolation and characterization of the principal ATPase associated with transitional endoplasmic reticulum of rat liver.

The transfer of membranes from the endoplasmic reticulum to the Golgi apparatus occurs via 50-70 nm transition vesicles which derive from part-rough, part-smooth transitional elements of the endoplasmic reticulum (TER). Vesicle budding from the TER is an ATP-dependent process both in vivo and in vitro. An ATPase with a monomer molecular weight of 100 kD by SDS-PAGE has been isolated from TER and designated as TER ATPase. The native TER ATPase has been characterized as a hexamer of six 100-kD subunits by gel filtration. The protein catalyzes the hydrolysis of [gamma 32-P]ATP and is phosphorylated in the presence of Mg2+. It is distinct from the classical transport ATPases based on pH optima, ion effects, and inhibitor specificity. Electron microscopy of negatively stained preparations revealed the TER ATPase to be a ring-shaped structure with six-fold rotational symmetry. A 19-amino acid sequence of TER ATPase having 84% identity with valosin-containing protein and 64% identity with a yeast cell-cycle control protein CDC48p was obtained. Anti-synthetic peptide antisera to a 15-amino acid portion of the sequence of TER ATPase recognized a 100-kD protein from TER. These antisera reduced the ATP-dependent cell-free formation of transition vesicles from isolated TER of rat liver. In a reconstituted membrane transfer system, TER ATPase antisera inhibited transfer of radiolabeled material from endoplasmic reticulum to Golgi apparatus, while preimmune sera did not. The results suggest that the TER ATPase is obligatorily involved in the ATP requirements for budding of transition vesicles from the TER. cDNA clones encoding TER ATPase were isolated by immunoscreening a rat liver cDNA library with the affinity-purified TER ATPase antibody. A computer search of deduced amino acid sequences revealed the cloned TER ATPase to be the rat equivalent of porcine valosin-containing protein, a member of a novel family of ATP binding, homo-oligomeric proteins including the N-ethylmaleimide-sensitive fusion protein.

Isolation of a Golgi apparatus-rich fraction from rat liver. IV. Thiamine pyrophosphatase.

The thiamine pyrophosphatase (the enzyme [s] catalyzing the release of inorganic phosphate with thiamine pyrophosphate as the substrate) activities of Golgi apparatus-, plasma membrane-, endoplasmic reticulum-, and mitochondria-rich fractions from rat liver were compared at pH 8. Activity was concentrated in the Golgi apparatus fractions, which, on a protein basis, had a specific activity six to eight times that of the total homogenates or purified endoplasmic reticulum fractions. However, only 1-3% of the total activity was recovered in the Golgi apparatus fractions under conditions where 30-50% of the UDPgalactose:N-acetylglucosamine-galactosyl transferase activity was recovered. Considering both recovery of galactosyl transferase and fraction purity, we estimate that approximately 10% of the total thiamine pyrophosphatase activity of the liver was localized within the Golgi apparatus, with a specific activity of about ten times that of the total homogenate. Cytochemically, reaction product was found in the cisternae of the endoplasmic reticulum as well as in the Golgi apparatus. This is in contrast to results obtained in most other tissues, where reaction product was restricted to the Golgi apparatus. Thus, enzymes of rat liver catalyzing the hydrolysis of thiamine pyrophosphate, although concentrated in the Golgi apparatus, are widely distributed among other cell components in this tissue.

Biochemical and morphological comparison of plasma membrane and milk fat globule membrane from bovine mammary gland.

Purified plasma membrane fractions from lactating bovine mammary glands and membranes of milk fat globules from the same source were similar in distribution and fatty acid composition of phospholipids. The sphingomyelin content of the phospholipid fraction of both membranes was higher than in these fractions from other cell components, beta-carotene, a constituent characteristic of milk fat, was present in the lipid fraction of the plasma membrane. Cholesterol esters of plasma membrane were similar in fatty acid composition to those of milk fat globule membranes. Disc electrophoresis of either membrane preparation on polyacrylamide gels revealed a single major protein component characteristic of plasma membrane from other sources. Distinct morphological differences between plasma membrane and milk fat globule membranes were observed in both thin sections and in negatively stained material. Plasma membrane was vesicular in appearance while milk fat globule membranes had a platelike aspect. These observations are consistent with derivation of fat globule membrane from plasma membrane accompanied by structural rearrangement of membrane constituents.

B-type cytochromes in plasma membranes isolated from rat liver, in comparison with those of endomembranes.

Fractions of plasma membranes, Golgi apparatus, endoplasmic reticulum (ER), and nuclear envelope were isolated from rat liver and were characterized by electron microsocpe and biochemical methods. The purity of the fractions was controlled by morphometry and by marker enzyme activities. Amounts of cytochromes b5, P-450, and P-420 were measured, as well as the NADPH- and NADPH-cytochrome c reductase activities. The pigments of the microsomal electron transport system were found in all membrane fractions in relatively high amounts, thus excluding an origin by microsomal contamination. Purified preparations of plasma membrane and Golgi apparatus contained approximately 30% of the cytochrome b5 and cytochrome P-450 + P-420 found in ER membranes. Plasma membranes were also characterized by a high ratio of P-420/450. Degradation of cytochromes P-450 and P-420 was relatively rapid in all fractions, except in the ER. Cytochrome b5 extracted from plasma membranes was spectrophotometrically and enzymatically indistinguishable from ER cytochrome b5. However, immunnlogical characterization with rabbit antibodies against the trypsin-resistant core of microsomal cytochrome b5 showed the presence of at least two types of cytochrome b5 in ER membranes, in contrast to the plasma membranes in which only one of these components was detected. This immunological differentiation also demonstrates that the plasma membrane-bound cytochrome b5 is endogenous to this membrane and does not reflect contamination by ER elements. We conclude that cytochromes b5, P-450, and P-420 are not confined only to ER and nuclear membranes but also occur in signficant amounts in Golgi apparatus and plasma membranes. The findings are discussed in relation to observations of similar redox components in Golgi apparatus, secretory vesicles, and plasma membranes of other cells.

Isolation of a Golgi apparatus-rich fraction from rat liver. I. Method and morphology.

Golgi apparatus were released without fixatives from rat hepatocytes by gentle homogenization, concentrated by differential centrifugation, and purified by sucrose gradient centrifugation. Examination of sections of purified fractions by electron microscopy showed fields of morphologically intact units of Golgi apparatus consisting of stacks of parallel flattened cisternae, secretory vesicles, and small vesicular profiles. Negative staining of unfixed pellets revealed a complex network of anastomotic tubules continuous with platelike structures and secretory vesicles. These structures corresponded, respectively, to the small vesicular profiles and parallel flattened cisternae with attached secretory vesicles of sectioned material. Small fragments of granular endoplasmic reticulum were often closely associated with the peripheral tubules, suggesting sites of continuity in intact hepatocytes.

Mammary carcinoma: enzymatic block in disialoganglioside biosynthesis.

The sialyl transferase of disialoganglioside formation is depressed in mammary tumors induced in the rat by 7,12-dimethylbenz[a]anthracene. Specific activities of other glycosyltransferases of the pathway ceramide to monosialo-ganglioside are unchanged or elevated so that the ganglioside GM(1) accumulates and higher gangliosides are depressed. These findings with a solid tumor are critical to an involvement of gangliosides in the cell-surface changes of tumorigenesis.

Polyribosomes associated with forming acrosome membranes in guinea pig spermatids.

Ribosomes, some of which are arranged in polyribosomal configurations, are attached to specialized regions of the acrosomal membrane in guinea pig spermatids. This finding indicates a new functional dimension for the acrosomal membrane, that of protein synthesis, and suggests that during acrosome formation, proteins of the acrosomal membrane or acrosomal contents need not be synthesized before or during passage through the Golgi apparatus.

Cytomembrane differentiation in the endoplasmic reticulum-Golgi apparatus-vesicle complex.

Diversity of cytomembrane types is confirmed in hyphae of the fungus Pythium ultimum by electron microscopy. A transition in membrane morphology across stacks of dictyosome cisternae (from endoplasmic reticulumlike at one pole to plasma membrane-like at the opposite pole) suggests that dictyosomes of the Golgi apparatus are sites of membrane interconversion.

Vitamin A: concentration in the rat liver Golgi apparatus.

Vitamin A compounds (principally as retinyl esters) are concentrated in Golgi apparatus fractions from rat liver. The amounts vary with the vitamin A status of the liver and show an inverse relation to the concentration of ubiubinone. The results suggest a specific role of the Golgi apparatus in the mobilization or action, or both, of vitamin A compounds.

Regular oscillatory behavior of aqueous solutions of CuII salts related to effects on equilibrium dynamics of ortho/para hydrogen spin isomers of water.

Cell surface and growth-related NADH oxidases with protein disulfide-thiol interchange activity, ECTO-NOX, exhibit copper-dependent, clock-related, temperature-independent and entrainable patterns of regular oscillations in the rate of oxidation of NAD(P)H as do aqueous solutions of copper salts. Because of time scale similarities, a basis for the oscillatory patterns in nuclear spin orientations of the hydrogen atoms of the copper-associated water was sought. Extended X-ray absorption fine structure (EXAFS) measurements at 9302 eV on pure water were periodic with a ca. 3.5 min peak to peak separation. Decomposition fits revealed 5 unequally spaced maxima similar to those observed previously for Cu(II)Cl(2) to generate a period length of about 18 min. With D(2)O, the period length was proportionately increased by 30% to 24 min. The redox potential of water and of D(2)O also oscillated with 18 and 24 min period lengths, respectively. Measurements in the middle infrared spectral region above a water sample surface revealed apparent oscillations in the two alternative orientations of the nuclear spins (ortho and para) of the hydrogen atoms of the water or D(2)O with 5 unequally spaced maxima and respective period lengths of 18 and 24 min. Thus, the time keeping oscillations of ECTO-NOX proteins appear to reflect the equilibrium dynamics of ortho-para hydrogen atom spin ratios of water where the presence of metal cations such as Cu(II) in solution determine period length.

Structural observations of time dependent oscillatory behavior of CuIICl2 solutions measured via extended X-ray absorption fine structure.

Cell surface ECTO-NOX proteins exhibit a clock-related, temperature-independent entrainable pattern of periodic (24 min) oscillations in the rate of oxidation of NAD(P)H. Aqueous solutions of copper salts also oxidize NAD(P)H with a similar temperature-independent pattern. For both, five maxima are observed, two of which are separated by 6 min and the remaining three are separated by 4.5 min. In D2O, the pattern is retained but the period length is proportionately increased to 30 min in direct relationship to the 30 h circadian day observed with D2O-grown organisms. With copper solutions, periodic changes in redox potential correlate precisely with the periodic changes in the rates of NAD(P)H oxidation. Consequently, the local environment of the Cu2+ ion in copper chloride solutions was investigated by X-ray absorption spectroscopy. Detailed extended X-ray absorption fine structure (EXAFS) analyses revealed a pattern of oscillations closely resembling those of the copper-catalyzed oxidation of NADH. With CuCl2 in D2O, a pattern with a period length of 30 min was observed. The findings suggest a regular pattern of distortion in the axial and/or equatorial oxygen atoms of the coordinated water molecules which correlate with redox potential changes sufficient to oxidize NADH. A metastable equilibrium condition in the ratio of ortho to para nuclear spin orientation of the water associated hydrogen atoms would be kinetically consistent with a 24-30 min timeframe. The temperature independence of the biological clock can thus be understood as the consequence of a physical rather than a chemical basis for the timing events.

Role of p97 and syntaxin 5 in the assembly of transitional endoplasmic reticulum.

Transitional endoplasmic reticulum (tER) consists of confluent rough and smooth endoplasmic reticulum (ER) domains. In a cell-free incubation system, low-density microsomes (1.17 g cc(-1)) isolated from rat liver homogenates reconstitute tER by Mg(2+)GTP- and Mg(2+)ATP-hydrolysis-dependent membrane fusion. The ATPases associated with different cellular activities protein p97 has been identified as the relevant ATPase. The ATP depletion by hexokinase or treatment with either N-ethylmaleimide or anti-p97 prevented assembly of the smooth ER domain of tER. High-salt washing of low-density microsomes inhibited assembly of the smooth ER domain of tER, whereas the readdition of purified p97 with associated p47 promoted reconstitution. The t-SNARE syntaxin 5 was observed within the smooth ER domain of tER, and antisyntaxin 5 abrogated formation of this same membrane compartment. Thus, p97 and syntaxin 5 regulate assembly of the smooth ER domain of tER and hence one of the earliest membrane differentiated components of the secretory pathway.

Characteristics of transplantable tumors induced in the rat by N-2-fluorenylacetamide: elevations in tissue and serum sialic acid.

The correlation between levels of sialic acid and sialic acid-containing glycolipids (gangliosides) in tumors and serum with the growth characteristics of the tumors was investigated in transplantable hepatomas and squamous cell carcinomas initiated with the carcinogen N-2-fluorenylacetamide and propagated in vivo and in tissue culture. Tumor lines varied in histologic classification, growth rate, and ability to form pulmonary metastases. There was neither a correlation between growth rate and histologic classification nor between either of these two parameters and the ability to metastasize. Total and ganglioside sialic acid levels were elevated in carcinogen-treated liver and in transplantable hepatomas when contrasted with normal liver. Levels of sialic acid showed a weak correlation with the growth rate of hepatomas. Gangliosides from nonmetastatic hepatoma lines exhibited less N-acetylneuraminic acid--galactose--glucose-N--acylsphingosine (GM3) and an increased ratio of total monosialogangliosides to disialogangliosides than did metastatic lines. Ganglioside patterns of metastatic hepatoma lines more closely resembled the ganglioside patterns of normal liver than did those of the nonmetastatic lines. Concomitant elevations of total and ganglioside sialic acid levels were observed in sera of animals bearing subcutaneous implants. Serum levels of total sialic acid did correlate with total sialic acid levels found in the tumor tissues. The levels of serum sialic acid were not correlated directly with levels of serum sialyltransferase activity. Elevations of both tissue and serum ganglioside sialic acid were consistent features of liver tumorigenesis in the rat after N-2-fluorenylacetamide administration. They appeared, furthermore, to be early events not directly related to tumor cell differentiation or metastasis.

Gangliosides of liver tumors induced by N-2-fluorenylacetamide. I. Ganglioside alterations in liver tumorigenesis and normal development.

Hyperplastic nodules and hepatocellular carcinomas were induced in livers of rats by a low-protein diet containing 0.05% of the carcinogen N-2-fluorenylacetamide. Ganglioside amounts and composition were determined for histologically different hepatocellular carcinomas and compared with those for control livers, hyperplastic nodules, and liver tissue surrounding hepatomas and nodules as well as those for livers of fetal, newborn, 1-week-old, weanling, and adult Sprague-Dawley rats. Ganglioside sialic acid levels were elevated above those of normal adult liver in all liver tissues following the carcinogen treatment regimen. Livers of fetal and newborn rats contained nearly twice the amount of ganglioside sialic acid on a protein or DNA basis as did livers of adult rats. Analyses of individual nodules and hepatomas revealed two populations of tumors in which the levels of ganglioside sialic acid were 2.3 and 3.8 times normal. Ganglioside sialic acid content was at hepatoma levels in small nodules. Individual gangliosides were evenly distributed between products of the monosialoganglioside and disialoganglioside pathways in normal liver with a ratio of [N-acetylneuraminic acid (sialic acid)] (NAN)-galactose (Gal)-N-acetylgalactosamine (GalNAc)-(NAN)-Gal-glucose (Glc)-ceramide (Cer) (GD1a) to Gal-GalNAc-(NAN)2-Gal-Glc-Cer (GD1b) of about one. In contrast, the monosialogangliosides predominated in liver tissues following administration of the carcinogen. Increased levels of specific monosialogangliosides were present in nodules, in liver of carcinogen-treated animals prior to the appearance of tumors, and in the liver tissues surrounding nodules and hepatomas. In single hepatomas, ganglioside patterns correlated with tumorigenicity. A well-differentiated hepatoma had a normal complement of most gangliosides but was deficient in trisialogangliosides. In a poorly diferentiated but well-circumscribed hepatoma, the relative levels of all higher gangliosides were reduced. The monosialoganglioside Gal-GalNAc-(NAN)-Gal-Glc-Cer (GM1) accounted for 80% of the total ganglioside in a poorly circumscribed and poorly differentiated hepatoma. The ganglioside pattern of fetal livers most closely resembled that of a poorly differentiated hepatoma. During the first week post natum, levels of all higher monosialogangliosides and disialogangliosides declined, but the decline was most pronounced for gangliosides GM1 and GD1a. The ratio of GM1 + GD1a to GD1b + NAN-Gal-GalNAc-(NAN)2-Gal-Glc-Cer or (NAN)3-Gal-Glc-Cer (GT), used as an index of the relative predominance of the monoslaloganglioside and disialoganglioside pathways, fell from 2.7 for fetal liver to 0.4 for adult liver. Pools of precursor gangliosides increased during development, transiently for GalNAc-(NAN)-Gal-Glc-Cer and for more than 3 weeks for NAN-Gal-Glc-Cer. When hyperplastic nodules and hepatocellular carcinomas were compared, a reverse pattern was observed. The ratio of GM1 + GD1a to GD1b + GT rose steadily to values of 2.7 and 11...

Isolation of purified plasma membranes from cultured cells and hepatomas by two-phase partition and preparative free-flow electrophoresis.

This report describes and documents the isolation of plasma membranes from hepatomas and tissue culture cells by aqueous two-phase partition. The method used previously for normal liver was effective, rapid, and reproducible. Preparations from both cells and hepatomas were more than 90% plasma membrane-derived based on electron microscope morphometry and assays of marker enzyme activities. Relative enrichments over starting homogenates were near theoretical as determined by electron microscope morphometry of starting cells and tissues. Recoveries were about 10% or greater. Briefly, the membranes to be separated were mixed with a combination of two different polymers that themselves separated into two phases. For tissue culture cells and hepatomas, a mixture of 6.6% (w/w) dextran and 6.6% (w/w) polyethylene glycol containing 0.25 M sucrose and 5 mM potassium phosphate (pH 7.2) was used. A potassium-stimulated, ouabain-inhibited, p-nitrophenylphosphatase was employed as a plasma membrane marker to monitor yield and recovery. In combination with free-flow electrophoresis, the preparations of plasma membranes from cultured cells were resolved further into fractions enriched in vesicles of right side-out or inside-out orientations. Unlike centrifugation methods, the same type of two-phase separations provided useful plasma membrane fractions when applied to different types of cultured cells as well as to solid tumors and normal tissues. The lower phase membranes, after two-phase partition, provided a plasma membrane-depleted source of membranes other than plasma membrane for use as reference fractions. The procedures should find wide application to problems of cancer research where facile and decisive separations of surface and internal membranes may be required.

Decreased NADH-oxidoreductase activities as an early response in rat liver to the carcinogen 2-acetylaminofluorene.

Reduced nicotinamide adenine dinucleotide (NADH):ferricyanide reductase and DT-diaphorase specific activity in total homogenates of rat liver are markedly decreased as a very early biochemical event of hepatocarcinogenesis induced by the carcinogen 2-acetylaminofluorene (AAF). A 50 to 75% decrease in NADH:ferricyanide reductase was observed after 1 day of AAF (0.025% in the diet) feeding and persisted throughout a 7-week continuum of AAF administration. Carcinogen added directly to cell extracts had no effect. Similar results were obtained with single injections of either AAF or diethylnitrosamine. Xanthine dehydrogenase was also reduced in liver following AAF administration to nearly the same extent as NADH:ferricyanide reductase and DT-diaphorase. Total NADH-cytochrome c reductase and mitochondrial activity as estimated from succinic dehydrogenase were not affected by carcinogen administration relative to basal dietary controls. The reduced nicotinamide adenine dinucleotide phosphate:cytochrome c reductase that functions in drug detoxification was elevated. With livers of animals fed 4-acetamidophenol, a hepatotoxin chemically related to AAF, small decreases were noted in NADH:ferricyanide reductase, but not in xanthine dehydrogenase nor in DT-diaphorase. Initial lowering of these activities in the livers of the carcinogen-treated animals is preceded by or concomitant with a reduction in the levels of extramitochondrial pyridine nucleotides known from other studies to result from DNA damage.

NADH oxidase activity of HeLa plasma membranes inhibited by the antitumor sulfonylurea N-(4-methylphenylsulfonyl)-N'-(4-chlorophenyl) urea (LY181984) at an external site.

NADH oxidase activity from HeLa plasma membranes was inhibited by the antitumor sulfonylurea N-(4-methylphenylsulfonyl)-N'-(4-chlorophenyl)urea (LY181984). With sealed right side-out vesicles, the drug inhibited half maximally at about 30 nM and the inhibition was nearly complete. A closely related but growth-inactive sulfonylurea, N-(4-methylphenylsulfonyl)-N'-(phenyl)urea (LY181985), did not inhibit the activity. With plasma membranes first solubilized with 2% Triton X-100, activity also was inhibited by LY181984 and not by LY181985 but the maximum inhibition at 10 microM LY181984 was only 50%. When sealed right side-out plasma membrane vesicles were frozen and thawed repeatedly to evert some of the vesicles into an inside-out configuration, the NADH oxidase activity again was only about 50% inhibited by 1 microM LY181984. In such preparations, the right side-out vesicles exhibited an electrophoretic mobility greater than that of the inside-out vesicles. Sidedness was confiremd by measurements of ATPase latency and binding of immunogold-labeled concanavalin A. When the two vesicle populations were resolved by preparative free-flow electrophoresis, the active antitumor sulfonylurea LY181984 inhibited only the NADH oxidase activity of the right side-out vesicles. These findings suggested two NADH sites or activity isoforms for the plasma membrane NADH oxidase. One activity, inhibited by LY181984, appeared to be accessible to external NADH only with sealed right side-out vesicles. The other, not inhibited by LY181984, was accessible to NADH only with inside-out vesicles or after membrane disruption by Triton X-100. The findings demonstrate that the NADH oxidation site inhibited as a result of binding the active antitumor sulfonylurea LY181984 is at the external cell surface. Plasma membrane vesicles from HeLa cells are able to oxidize NADH supplied to either membrane surface but only with inside-out vesicles is NADH oxidation sensitive to inhibition by the antitumor sulfonylurea.

Glycolipids modulate glycosyl transfer to endogenous protein acceptors.

Regulation by gangliosides of glycosylation of endogenous membrane glycoproteins is indicated from in vitro studies in which incorporation of radioactive sugars into endogenous protein acceptors was measured and from in vitro studies where transferase activities of membranes were correlated with ganglioside content during hepatic tumorigenesis. Galactosyl transfer from UDP galactose exhibited a complex response pattern and was stimulated by lactosyl ceramide and the ganglioside N-acetylgalactosaminyl-(N-acetylneuraminyl)-galactosylglucosylceramide (GM2) but was inhibited by higher gangliosides. Except for N-acetylneuraminylgalactosylglucosylceramide (GM3), which had no effect, inhibition was proportional to ganglioside complexity. Inhibition of glycosylation of the exogenous acceptor, ovomucoid, by ganglioside was slight by comparison. While marked structure-linked latency was observed with the high molecular weight exogenous acceptor, no latency was observed for incorporation into endogenous acceptors suggesting that the membranes were permeable to sugar nucleotides. Membrane disruption with detergents lessened rather than enhanced inhibition by gangliosides. Sialyl transfer from CMPsialic acid, on the other hand, was unaffected or stimulated by gangliosides. Stimulation by galactosyl-N-acetylgalactosaminyl-(N-acetylneuraminyl)-galactosylglucosylceramide (GM1) was proportional to concentration and reached 2-fold at 240 micrograms/mg protein. The results suggest that the ganglioside content of membrane may affect glycosylation of membrane glycoproteins.

Translocation of cytidine 5'-monophosphosialic acid across Golgi apparatus membranes.

Golgi apparatus, isolated from rat liver, incorporate [14C]sialic acid from CMP[14C]sialic acid into endogenous glycolipid and glycoprotein acceptors. Incorporation of [14C]sialic acid into endogenous glycoprotein acceptors was stimulated an average of 3-fold by Triton X-100 at an optimal concentration of 0.05% and was inhibited at higher concentrations. Incorporation of [14C]sialic acid into endogenous glycolipid acceptors was not stimulated by detergent. The major glycolipid product was identified by thin-layer chromatography as the ganglioside GD3. SDS-polyacrylamide gel electrophoresis on the glycoprotein products demonstrated incorporation of [14C]sialic acid into 6--7 major bands. Neuraminidase studies determined that approximately 60% of the [14C]sialic acid incorporated into endogenous acceptors in the absence of detergent had a luminal orientation. Furthermore, electron microscopy studies showed that the isolated Golgi apparatus fraction consisted of intact membrane cisternae. Our results demonstrate that sialylation of cisternal acceptors located on the inside of the membrane occurs in the absence of detergent. They are consistent with carrier-mediated transport as a mechanism to allow CMPsialic acid to traverse the Golgi apparatus membrane and to be used to glycosylate endogenous glycoprotein and glycolipid acceptors.