Compartmentalized signal transduction by receptor tyrosine kinases.

Signal transduction through receptor tyrosine kinases is believed to occur mainly at the plasma membrane. Ligands bind to their cognate receptors and trigger autophosphorylation events, which are detected by intracellular signalling molecules. However, ligands, such as epidermal growth factor and insulin, induce the rapid internalization of their receptors into endosomes. Although this event is traditionally thought to attenuate the ligand-induced response, in this article the authors discuss an alternative scenario in which selective and regulated signal transduction from receptor tyrosine kinases occurs within the endosome.

Localization of GTP-stimulated core glycosylation to fused microsomes.

Purified rough microsomes from liver maximally incorporated N-acetyl-[3H]glucosamine into endogenous acceptors from UDP-N-acetyl-[3H]glucosamine substrate, providing the associated ribosomes were removed and 0.5 mM GTP was added. These conditions also led to the coalescence of microsomes into large fused membranes. By measurement of membrane profiles on electron micrographs, a correlation was observed between GTP-stimulated glycosylation and microsomal membrane length (r2 = 0.92). Membrane fusion was not observed in the absence of GTP, with sugar transfer inhibited by greater than 90% for acid-resistant acceptors (protein), and approximately 50% for acid-labile acceptors (lipid-linked intermediates). When radiolabeled acceptors were localized by electron microscope radioautography, high concentrations of silver grains (83 grains/100 microns membrane length) were observed over fused membranes with lower grain densities observed over unfused membranes in the same preparation (20 grains/100 microns). These studies directly link microsomal membrane fusion to GTP-stimulated core glycosylation. The observations extend the suggestion of Godelaine et al. (1979, Eur. J. Biochem. 96:17-26) that physiological levels of GTP promote the translocation of substrate across endoplasmic reticulum membranes which, we propose, occurs via a membrane fusion phenomenon.

Passage of serum-destined proteins through the Golgi apparatus of rat liver. An examination of heavy and light Golgi fractions.

The participation of hepatic Golgi apparatus in the intracellular transport of blood-destined proteins has been analyzed using Golgi fractions enriched in cis and trans components of the Golgi apparatus. SDS-polyacrylamide gel electrophoresis of the liver Golgi fractions showed several proteins corresponding in relative proportions and mobilities with serum proteins. After a pulse injection of labeled leucine, the secretory content of the cis Golgi fraction was labeled earlier than the trans Golgi fraction. Taken together, the results show the participation of the liver Golgi apparatus in the secretion of most of the serum proteins and provide documentation for a sequential progression of secretory protein through the cis and trans components of the Golgi apparatus.

Direct in vivo demonstration by radioautography of specific binding sites for calcitonin in skeletal and renal tissues of the rat.

An in vivo binding assay using radioautography was employed to visualize calcitonin receptors in rat tissues. At 2 min after intravenous injection of biologically active 125I-salmon calcitonin, free hormone was separated from bound hormone by intracardiac perfusion with lactated Ringer's followed by fixation with 2.5% glutaraldehyde. Various tissues were removed and processed for light and electron microscope radioautography. These were compared to tissues removed from animals that received identical amounts of labeled hormone with a large excess of unlabeled calcitonin. Among the tissues investigated, kidney and bone demonstrated labeling. In kidney, most silver grains were located over vesicles below the brush border of cells of theproximal convoluted tubules. These grains were still present after simultaneous injection of excess unlabeled hormone and most likely represented binding to sites involved with ingestion and degradation of hormone from the urinary filtrate. In contrast, grains localized to the basal surfaces of distal convoluted tubule cells were significantly reduced in number in control animals and represented sites of saturable, specific hormone binding. In bone, specific binding sites were found only at the periphery of osteoclasts. These labeled cells were located at resorption sites examined in tibia, humerus, and alveolar bone. This demonstration of the localization of 124I-calcitonin in situ provides a new approach for study the interaction of calcium-regulating hormones with their target cells.

Synthesis of rat liver microsomal cytochrome b5 by free ribosomes.

Free and membrane-bound polyribosomes were separated from liver homogenates and characterized by electron microscopy. Using the wheat germ cell-free translation system, total translation products of poly A+RNA extracted from free polyribosomes (poly A+RNAf) showed some correlation to total liver cytosol proteins. In contrast, translation products of poly A+RNA from membrane-bound polyribosomes (poly A+RNAmb) showed some similarity to rat serum. Antibody to purified rat serum albumin immunoprecipitated from only the translation products of poly A+RNAmb a single polypeptide of mol wt 68,000. i.e., 3,000 greater than secreted serum albumin. In contrast, antibody to detergent-extracted cytochrome b5 immunoprecipitated from only the translation products of poly A+RNAf a single polypeptide of mol wt 17,500, identical to that of microsomal cytochrome b5. A consideration of the known properties of cytochrome b5 is consistent with an exclusive site of synthesis on free ribosomes.

Cytochemistry of Golgi fractions prepared from rat liver.

Cytochemical tests for several marker enzymes were applied to liver tissue and to the three Golgi fractions (GF(1), GF(2), GF(3)) separated by the procedure of Ehrenreich et al. from liver homogenates of alcohol-treated rats. 5'-Nucleotidase (AMPase) reaction product was found in all three fractions but in different locations: It occurred along the inside of the membrane of VLDL-filled vacuoles in GF(1) and GF(2), and along the outside of the cisternal membranes in GF(3). In the latter it was restricted to the dilated cisternal rims and was absent from the cisternal centers. The AMPase activity found in the fractions by biochemical assay is therefore indigenous to Golgi components and is not due to contamination by plasma membrane. Acid phosphatase (AcPase) reaction product was detected within lysosomal contaminants in GF(1) and within many VLDL-filled vacuoles in GF(1) and GF(2), indicating that AcPase activity is due not only to contaminating lysosomes, but also to enzyme indigenous to Golgi secretory vacuoles. G-6-Pase reaction product was present in GF(3) and within contaminating endoplasmic reticulum fragments, but not in other fractions. Thiamine pyrophosphatase (TPPase) was localized to some of the VLDL-filled vacuoles and cisternae in GF(1) and GF(2), and was not found in the cisternae in GF(3). The results demonstrate the usefulness of cytochemical methods in monitoring the fractionation procedure: They have (a) allowed a reliable identification of contaminants, (b) made possible a distinction between indigenous and contaminating activities, and (c) shown, primarily by the results of the TPPase test, that the procedure achieves a meaningful subfractionation of Golgi elements, with GF(1) and GF(3), representing primarily trans-Golgi elements from the secretory Golgi face, and GF(3) consisting largely of cis-Golgi components from the opposite face.

Reconstitution of the Golgi apparatus after microinjection of rat liver Golgi fragments into Xenopus oocytes.

We have studied the reconstitution of the Golgi apparatus in vivo using an heterologous membrane transplant system. Endogenous glycopeptides of rat hepatic Golgi fragments were radiolabeled in vitro with [3H]sialic acid using detergent-free conditions. The Golgi fragments consisting of dispersed vesicles and tubules with intraluminal lipoprotein-like particles were then microinjected into Xenopus oocytes and their fate studied by light (LM) and electron microscope (EM) radioautography. 3 h after microinjection, radiolabel was observed by LM radioautography over yolk platelet-free cytoplasmic regions near the injection site. EM radioautography revealed label over Golgi stacked saccules containing the hepatic marker of intraluminal lipoprotein-like particles. At 14 h after injection, LM radioautographs revealed label in the superficial cortex of the oocytes between the yolk platelets and at the oocyte surface. EM radioautography identified the labeled structures as the stacked saccules of the Golgi apparatus, the oocyte cortical granules, and the plasmalemma, indicating that a proportion of microinjected material was transferred to the surface via the secretion pathway of the oocyte. The efficiency of transport was low, however, as biochemical studies failed to show extensive secretion of radiolabel into the extracellular medium by 14 h with approximately half the microinjected radiolabeled constituents degraded. Vinblastine (50 microM) administered to oocytes led to the formation of tubulin paracrystals. Although microinjected Golgi fragments were able to effect the formation of stacked saccules in vinblastine-treated oocytes, negligible transfer of heterologous material to the oocyte surface could be detected by radioautography. The data demonstrate that dispersed fragments of the rat liver Golgi complex (i.e., unstacked vesicles and tubules) reconstitute into stacked saccules when microinjected into Xenopus cytoplasm. After the formation of stacked saccules, reconstituted Golgi fragments transport constituents into a portion of the exocytic pathway of the host cell by a microtubule-regulated process.

The cytoplasmic droplet of rat epididymal spermatozoa contains saccular elements with Golgi characteristics.

The cytoplasmic droplet of epididymal spermatozoa is a small localized outpouching of cytoplasm of the tail of unknown significance. EM revealed flattened saccular elements as the near exclusive membranous component of the droplet. Light and electron microscopic immunolabeling for Golgi/TGN markers showed these saccules to be reactive for antibodies to TGN38, protein affinity-purified alpha 2,6 sialyltransferase, and anti-human beta 1,4 galactosyltransferase. The saccules were isolated by subcellular fractionation and antibodies raised against this fraction immunolabeled the saccules of the droplet in situ as well as the Golgi region of somatic epithelial cells lining the epididymis. The isolated droplet fraction was enriched in galactosyltransferase and sialyltransferase activities, and endogenous glycosylation assays identified the modification of several endogenous glycopeptides. EM lectin staining in situ demonstrated galactose and N-acetyl galactosamine constituents in the saccules. Endocytic studies with cationic and anionic ferritin as well as HRP failed to identify the saccules as components of the endocytic apparatus. Epididymal spermatozoa were devoid of markers for the ER as well as the Golgi-associated coatamer protein beta-COP. It is therefore unlikely that the saccular elements of the droplet participate in vesicular protein transport. However, the identification of Golgi/TGN glycosylating activities in the saccules may be related to plasma membrane modifications which occur during epididymal sperm maturation.

Concentration of intracellular hepatic apolipoprotein E in Golgi apparatus saccular distensions and endosomes.

The intrahepatic distribution of apolipoprotein E has been assessed by immunogold labeling of cryosections as well as by Western blotting of organelles isolated from liver homogenates. Both techniques supported the prior analytical fractionation studies of Wong (1989) who concluded that intrahepatic apoE was largely endosomal. All endosomal components decorated by gold particles indicative of apoE antigenicity in cryosections appeared filled with lipoprotein-like particles thereby accounting for this prominent morphological feature of isolated liver endosomes. The distribution of gold particles about the hepatic Golgi apparatus revealed a high content of apoE in closely apposed endosomes, ca. 400 nm in diameter, double labeled for apoE and internalized HRP. Remarkably, apoE (but not internalized HRP) was also observed within saccular distensions of all saccules of stacked Golgi cisternae but absent from the flattened saccular components as was also observed for apoB. This contrasted with albumin, the major secretory protein, which was uniformly distributed throughout the hepatic Golgi apparatus. These observations support a growing body of evidence for intra-Golgi sorting of secretory material in hepatic Golgi apparatus. The lack of any immunoreactive apoE or albumin in small 70-90 nm vesicles about the Golgi cisternae suggests limits to current models of vesicle-mediated intra-Golgi transport.

Association of the tyrosine phosphorylated epidermal growth factor receptor with a 55-kD tyrosine phosphorylated protein at the cell surface and in endosomes.

After the intraportal injection of EGF, the EGF receptor (EGFR) is rapidly internalized into hepatic endosomes where it remains largely receptor bound (Lai et al., 1989. J. Cell Biol. 109:2751-2760). In the present study, we evaluated the phosphotyrosine content of EGFRs at the cell surface and in endosomes in order to assess the consequences of internalization. Quantitative estimates of specific radioactivity of the EGFR in these two compartments revealed that tyrosine phosphorylation of the EGFR was observed at the cell surface within 30 s of ligand administration. However, the EGFR was also highly phosphorylated in endosomes reaching levels of tyrosine phosphorylation significantly higher than those of the cell surface receptor at 5 and 15 min after EGF injection. A 55-kD tyrosine phosphorylated polypeptide (pyp55) was observed in association with the EGFR at the cell surface within 30 s of EGF injection. The protein was also found in association with the EGFR in endosomes as evidenced by coprecipitation studies using a mAb to the EGFR as well as by coelution with the EGR in gel permeation chromatography. Limited proteolysis of isolated endosomes indicated that the tyrosine phosphorylated domains of the EGFR and associated pyp55 were cytosolically oriented while internalized EGF was intraluminal. The identification of pyp55 in association with EGFR in both hepatic plasma membranes and endosomes may be relevant to EGFR function and/or trafficking of the EGFR.

Effect of colchicine on internalization of prolactin in female rat liver: an in vivo radioautographic study.

Binding and internalization of 125I-ovine prolactin into hepatocytes of female rats was visualized by the in vivo radioautographic method (Bergeron, J. J. M., G. Levine, R. Sikstrom, D. O'Shaughnessey, B. Kopriwa, N. J. Nadler, and B. I. Posner, 1977, Proc. Natl. Acad. Sci. USA, 745:051-5055). Receptor-mediated internalization of label was observed into lipoprotein-filled vesicles in the Golgi/bile canalicular region of the hepatocyte. Colchicine treatment had no effect on the internalization of label into the lipoprotein-filled vesicles. However, the location of the radio-labeled lipoprotein-filled vesicles was altered from the Golgi/bile canalicular region to subsinusoidal. Radioactive content of hepatocytes decreased as a function of time after injection of 125I-prolactin; however, colchicine treatment markedly retarded this loss of label. Subcellular fractionation experiments indicated that colchicine treatment led to decreased levels of 125I-prolactin accumulation in microsomes but augmented the accumulation of label in the L fraction. It is concluded that in normal female rats prolactin is internalized into lipoprotein-filled vesicles in the Golgi region before degradation of the hormone. Colchicine treatment accumulates labeled lipoprotein-containing vesicles in a subsinusoidal region and retards hormone catabolism. The labeled vesicles observed after colchicine treatment may correspond to the unique vesicles previously observed in the L fraction and found to be enriched in prolactin receptors (Khan, M. N., B. I. Posner, A. K. Verma, R. J. Khan, and J. J. M. Bergeron, 1981, Proc. Natl. Acad. Sci. USA, 78:4980-4981).

Galactose transfer to endogenous acceptors within Golgi fractions of rat liver.

The distribution of galactosyl transferase was studied using trans and cis Golgi fractions isolated by a modification of the Ehrenreich et al. procedure (1973. J. Cell Biol. 59:45-72) as well as an intact Golgi fraction isolated by a new one-step procedure. Two methods of assay were used. The first method analyzed the ability of Golgi fractions to transfer galactose (from uridine diphosphogalactose [UDP-gal] substrate) to the defined exogenous acceptor ovomucoid. The second method assessed the transfer of galactose from UDP-gal substrate to endogenous acceptors (endogenous glycosylation). The trans Golgi fraction (Golgi light) was highly active by the first method but revealed only low activity by the second method. Golgi fractions enriched in central and cis elements (the Golgi intermediate, heavy and especially the intact Golgi fraction) were highly active in both methods of assay. The endogenous glycosylation approach was validated by gel fluorography of the endogenous acceptors. For all Golgi fractions, transfer of galactose was revealed to secretory glycopeptides. It is concluded that galactosyl transferase activity in vivo occurs primarily in central and cis Golgi elements but not trans Golgi vesicles.

Binding and uptake of 125I-insulin into rat liver hepatocytes and endothelium. An in vivo radioautographic study.

Electron microscope radioautography has been used to study hormone-receptor interaction. At intervals of 3, 10, and 20 min after the injection of 125I-insulin, free hormone was separated from bound hormone by whole body perfusion with modified Ringer's solution. The localization of bound hormone, fixed in situ by perfusion with glutaraldehyde, was determined. At 3 min, 125I-insulin has been shown to be exclusively localized to the hepatocyte plasmalemma (Bergeron et al., 1977, Proc. Natl. Acad. Sci. U. S. A., 74:5051--5055). In the present study, quantitation indicated that 10(5) receptors were present per cell and distributed equally along the sinusoidal and lateral segments of the hepatocyte plasmalemma. At later times, label was found in the Golgi region. At 10 min, both secretory elements of the Golgi apparatus and lysosome-like vacuoles were labeled, and at 20 min the label was especially concentrated over the latter vacuoles. Acid phosphatase cytochemistry showed that the vacuoles did not react and therefore were presumed not to be lysosomal. These Golgi vacuoles may constitute a compartment involved in the initial degradation and/or site of action of the hormone. Control experiments were carried out at all time intervals and consisted of parallel injections of radiolabeled insulin with excess unlabeled hormone. At all times in controls, label was diminished over hepatocytes and was found primarily over endothelial cells and within the macropinocytotic vesicles and dense bodies of these cells. Kupffer cells and lipocytes were unlabeled after the injection of 125I-insulin with or without excess unlabeled insulin.

Effect of colchicine on the uptake of prolactin and insulin into Golgi fractions of rat liver.

In previous studies we have shown that 125I-labeled prolactin is taken up by a receptor-dependent process and concentrated in an intact form in Golgi elements from female rat liver (J. Biol. Chem., 1979, 254:209-214). In this study we have examined the effect of colchicine on this uptake process into Golgi elements. Colchicine [25 mumol (10 mg)/100 gm body wt] was injected intraperitoneally in adult female rats, and hepatic Golgi fractions were prepared at 1, 2, and 3 h postinjection. The enzyme recoveries and morphological appearance of fractions from colchicine-treated and control (alcohol alone) animals were similar. At times greater than 1 h after colchicine there was a marked (greater than 60%) inhibition of uptake of 125I-ovine prolactin (125I-oPRL) into Golgi light and intermediate fractions but no inhibition of uptake into Golgi heavy and plasmalemma elements. At times from 2 to 45 min postinjection, 125I-oPRL was extracted from Golgi elements and found to be largely intact as judged by rebinding to receptors. The inhibitory effect of colchicine was seen at doses ranging from 0.25 mumol to 25 mumol/100 g body wt. Vincristine also inhibited 125I-oPRL uptake into the Golgi light and intermediate fractions but lumicolchicine had no inhibitory effect. There was a smaller effect of colchicine both at early (1 h) and later (3 h) times on the extent and pattern of 125I-insulin uptake. Colchicine treatment did not produce a significant change in lactogen receptor levels in the Golgi fractions. These results demonstrate that colchicine treatment inhibited the transfer of prolactin into Golgi vesicular elements. The much smaller effect on insulin uptake suggests that there may be differences in the manner in which the two hormones are handled in the course of internalization.

Golgi fractions prepared from rat liver homogenates. I. Isolation procedure and morphological characterization.

In devising a new procedure for the isolation of Golgi fractions from rat liver homogenates, we have taken advantage of the overloading with very low density lipoprotein (VLDL) particles that occurs in the Golgi elements of hepatocytes approximately 90 min after ethanol is administered (0.6 g/100 g body weight) by stomach tube to the animals. The VLDLs act as morphological markers as well as density modifiers of these elements. The starting preparation is a total microsomal fraction prepared from liver homogenized (1:5) in 0.25 M sucrose. This fraction is resuspended in 1.15 M sucrose and loaded at the bottom of a discontinuous sucrose density gradient. Centrifugation at approximately 13 x 10(6)g.min yields by flotation three Golgi fractions of density >1.041 and <1.173. The light and intermediate fractions consist essentially of VLDL-loaded Golgi vacuoles and cisternae. Nearly empty, often collapsed, Golgi cisternae are the main component of the heavy fraction. A procedure which subjects the Golgi fractions to hypotonic shock and shearing in a French press at pH 8.5 allows the extraction of the content of the Golgi elements and the subsequent isolation of their membranes by differential centrifugation.

Golgi fractions prepared from rat liver homogenates. II. Biochemical characterization.

The three Golgi fractions isolated from rat liver homogenates by the procedure given in the companion paper account for 6-7% of the protein of the total microsomal fraction used as starting preparation. The lightest, most homogeneous Golgi fraction (GF(1)) lacks typical "microsomal" activities, e.g., glucose-6-phosphatase, NADPH-cytochrome c-reductase, and cytochrome P-450. The heaviest, most heterogeneous fraction (GF(3)) is contaminated by endoplasmic reticulum membranes to the extent of approximately 15% of its protein. The three fractions taken together account for nearly all the UDP-galactose: N-acetyl-glucosamine galactosyltransferase of the parent microsomal fraction, and for approximately 70% of the activity of the original homogenate. Omission of the ethanol treatment of the animals reduces the recovery by half. The transferase activity is associated with the membranes of the Golgi elements, not with their content. Galactose is transferred not only to N-acetyl-glucosamine but also to an unidentified lipid-soluble component.

ER/Golgi intermediates acquire Golgi enzymes by brefeldin A-sensitive retrograde transport in vitro.

Secretory proteins exit the ER in transport vesicles that fuse to form vesicular tubular clusters (VTCs) which move along microtubule tracks to the Golgi apparatus. Using the well-characterized in vitro approach to study the properties of Golgi membranes, we determined whether the Golgi enzyme NAGT I is transported to ER/Golgi intermediates. Secretory cargo was arrested at distinct steps of the secretory pathway of a glycosylation mutant cell line, and in vitro complementation of the glycosylation defect was determined. Complementation yield increased after ER exit of secretory cargo and was optimal when transport was blocked at an ER/Golgi intermediate step. The rapid drop of the complementation yield as secretory cargo progresses into the stack suggests that Golgi enzymes are preferentially targeted to ER/Golgi intermediates and not to membranes of the Golgi stack. Two mechanisms for in vitro complementation could be distinguished due to their different sensitivities to brefeldin A (BFA). Transport occurred either by direct fusion of preexisting transport intermediates with ER/Golgi intermediates, or it occurred as a BFA-sensitive and most likely COP I-mediated step. Direct fusion of ER/Golgi intermediates with cisternal membranes of the Golgi stack was not observed under these conditions.

Ligand-mediated internalization, recycling, and downregulation of the epidermal growth factor receptor in vivo.

EGF receptor internalization, recycling,a nd downregulation were evaluated in liver parenchyma as a function of increasing doses of injected EGF. The effect of ligand occupancy in vivo on the kinetics and extent of internalization was studied with changes in the receptor content of isolated plasmalemma and endosome fractions evaluated by direct binding, Scatchard analysis, and Western blotting. For all doses of injected EGF, receptor was lost from the plasmalemma and accumulated in endosomes in a time- and dose-dependent fashion. However, at doses of injected EGF equivalent to less than or equal to 50% surface receptor occupancy (i.e., less than or equal to 1 microgram/100 g body weight), receptor levels returned by 120 min to initial values. This return was resistant to cycloheximide and therefore did not represent newly synthesized receptor. Neither was the return due to replenishment by an intracellular pool of low-affinity receptors as such a pool could not be detected by Scatchard analysis or Western blotting. Therefore, receptor return was due to the recycling of previously internalized receptor. At doses of injected EGF greater than 50% receptor occupancy, net receptor loss-i.e., downregulation-was observed by evaluating the receptor content of total particulate fractions of liver homogenates. At the higher saturating doses of injected EGF (5 and 10 micrograms/100 g body weight), the majority of surface receptor content was lost by 15 min and remained low for at least an additional 105 min. As the kinetics of ligand clearance from the circulation and liver parenchyma were similar for all doses of EGF injected, then the ligand-mediated regulation of surface receptor content and downregulation were not a result of a prolonged temporal interaction of ligand with receptor. Rather, the phenomena must be a consequence of the absolute concentrations of EGF interacting with receptor at the cell surface and/or in endosomes.

Ligand-mediated autophosphorylation activity of the epidermal growth factor receptor during internalization.

The association of EGF with its receptor in endosomes isolated from rat liver homogenates was assessed biochemically by polyethylene glycol precipitation and morphologically by electron microscope radioautography. The proportion of receptor-bound ligand in endosomes at 15 min after the injection of doses of 0.1 and 1 microgram EGF/100 g body weight was 57%. This value increased to 77% for the dose of 10 micrograms EGF injected. Quantitative electron microscope radioautography carried out on endosomes isolated at 15 min after the injection of 10 micrograms 125I-EGF demonstrated that most radiolabel was over the endosomal periphery thereby indicating that ligand-receptor complexes were in the bounding membrane but not in intraluminal vesicles of the content. EGF receptor autophosphorylation activity during internalization was evaluated in plasmalemma and endosome fractions. This activity was markedly but transiently reduced on the cell surface shortly after the administration of saturating doses of EGF. The same activity, however, was augmented and prolonged in endosomes for up to 30 min after EGF injection. The transient desensitization of cell surface activity was not due to prior in vivo phosphorylation since receptor dephosphorylation in vitro failed to restore autophosphorylation activity. Transient desensitization of cell surface autophosphorylation activity coincided with a diminished capacity for endocytosis of 125I-EGF with endocytosis returning to normal after the restoration of cell surface autophosphorylation activity. The inhibition of cell surface autophosphorylation activity and the activation of endosomal autophosphorylation activity coincident with downregulation suggest that EGF receptor traffic is governed by ligand-regulated phosphorylation activity.

Membrane fusion and glycosylation in the rat hepatic Golgi apparatus.

When purified Golgi fractions were incubated with UDP-[3H]galactose in the absence of Triton-X-100, radioactivity was incorporated into an endogenous lipid and several peptide acceptors. Electron microscope analysis of Golgi fractions incubated in the endogenous galactosyl transferase assay medium revealed extensive fusion of Golgi saccules. Systematic removal of constituents in the galactosyl transferase assay medium showed enhanced (minus beta-mercaptoethanol) or reduced (minus ATP, minus sodium cacodylate buffer or minus MnCl2) fusion of Golgi membranes compared to the complete medium, Stereologic analysis revealed a correlation between membrane fusion and galactosyl transferase activity (r = 0.99, P less than 0.001). Electron microscope radioautography was carried out after incubation of Golgi fractions with UDP-[3H]galactose. Silver grains were not observed over trans elements of Golgi but were revealed mainly over large fused saccules with the number of silver grains being proportionate to membrane fusion (r = 0.92, P less than 0.001). Bilayer destabilization at points of Golgi membrane fusion may act to translocate galactose across the Golgi membrane and thereby provide a fusion regulated substrate for terminal glycosylation.