Three classes of spiny-(Clathrin-) coated vesicles in rodent liver based on size distributions

Summary Clathrin-coated vesicles in rodent (rat and mouse) liver distribute into three distinct populations based on measurements of vesicle diameter. The first population consists of 60–80 nm vesicles found almost exclusively within the Golgi apparatus region. The second population is of 100–160 nm coated vesicles located within 100–500 nm of the cell surface. A third population of coated vesicles of intermediate diameter (ca. 90 nm) is present both at the Golgi apparatus and at the cell surface. We speculate that clathrin and clathrin-coated vesicles within the region of the Golgi apparatus and of the cell surface exist in two recycling populations. The third population of vesicles of intermediate diameter could represent a shuttle to link the two major compartments.     

Calmodulin binding and protein phosphorylation in adrenal medulla coated vesicles

Coated vesicles from bovine adrenal medulla contained clathrin and major detergent-insoluble polypeptides of 120-100, 51 and 49 kDa. Intact coated vesicles and vesicles lacking clathrin light chains were bound by immobilized calmodulin in the presence of Ca2+. Clathrin in the form of 700 A cages was not bound. The calmodulin binding components in intact coated vesicles are therefore contributed by the enclosed vesicle or by the 120-100, 50 or 49 kDa polypeptides. The 51 kDa component incorporated 32Pi from labelled ATP by an endogenous kinase activity; no other coat or vesicle membrane protein was phosphorylated in vitro, either by intrinsic or exogenous kinases.     

Isolation of a vesicular intermediate in the cell-free transfer of membrane from transitional elements of the endoplasmic reticulum to Golgi apparatus cisternae of rat liver

Preparations enriched in part-smooth (lacking ribosomes), part-rough (with ribosomes) transitional elements of the endoplasmic reticulum when incubated with ATP plus a cytosol fraction responded by the formation of blebbing profiles and approximately 60-nm vesicles. The 60-nm vesicles formed resembled closely transition vesicles in situ considered to function in the transfer of membrane materials between the endoplasmic reticulum and the Golgi apparatus. The transition elements following incubation with ATP and cytosol were resolved by preparative free-flow electrophoresis into fractions of differing electronegativity. The main fraction contained the larger vesicles of the transitional membrane elements, while a less electronegative minor shoulder fraction was enriched in the 60-nm vesicles. If the vesicles concentrated by preparative free-flow electrophoresis were from material previously radiolabeled with [3H]leucine and then added to Golgi apparatus immobilized to nitrocellulose, radioactivity was transferred to the Golgi apparatus membranes. The transfer was rapid (T1/2 of about 5 min), efficient (10-30% of the total radioactivity of the transition vesicle preparations was transferred to Golgi apparatus), and independent of added ATP but facilitated by cytosol. Transfer was specific and apparently unidirectional in that Golgi apparatus membranes were ineffective as donor membranes and endoplasmic reticulum vesicles were ineffective as recipient membranes. Using a heterologous system with transition vesicles from rat liver and Golgi apparatus isolated from guinea pig liver, coalescence of the small endoplasmic reticulum-derived vesicles with Golgi apparatus membranes was demonstrated using immunocytochemistry. Employed were polyclonal antibodies directed against the isolated rat transition vesicle preparations. When localized by immunogold procedures at the electron microscope level, regions of rat-derived vesicles were found fused with cisternae of guinea pig Golgi apparatus immobilized to nitrocellulose strips. Membrane transfer was demonstrated from experiments where transition vesicle membrane proteins were radioiodinated by the Bolton-Hunter procedure. Additionally, radiolabeled peptide bands not present initially in endoplasmic reticulum appeared following coalescence of the derived vesicles with Golgi apparatus. These bands, indicative of processing, required that both Golgi apparatus and transition vesicles be present and did not occur in incubated endoplasmic reticulum preparations or on nitrocellulose strips to which no Golgi apparatus were added.(ABSTRACT TRUNCATED AT 400 WORDS)     

Antibodies to hepatic endosomes. Identification of two endosome antigens

Endosome fractions were prepared from rat liver homogenates, and antibodies were raised in rabbits against the integral membrane proteins. Immunofluorescent studies showed that these antibodies identified primarily intracellular structures in liver sections, isolated hepatocytes and HepG-2 cells. Immunoelectron microscopy using protein A-gold confirmed that endocytic multivesicular structures, especially those located at the biliary pole of the hepatocyte, were labeled. Biochemical analysis showed that approximately 12 endosome antigens were present. A major 43 kDa glycosylated antigen corresponded to the asialoglycoprotein receptor subunit. A further antigen identified in endosomes was a 115 kDa polypeptide pI 4.3 previously identified as a major calmodulin-binding protein. The antigens identified in rat liver endosomes were different to those previously shown by other studies to be present in the Golgi apparatus and lysosomes.     

Monoclonal antibodies as markers of the endocytic and secretory pathways

A galactosyltransferase-rich subcellular fraction and wheat germ agglutinin(WGA)-binding microsomal proteins from rat myeloma cells have been used to immunize BALB/c mice. Fusion of the corresponding spleen cells with the Sp2/0 mouse myeloma has lead to the production of hybridomas secreting monoclonal antibodies directed against four proteins of the Golgi complex (GC) and other smooth membranes (SM). Subcellular fractionation of myeloma cells and rat liver, Triton X-114 partitioning, protease treatment and lectin binding studies have permitted us to identify--by immunoblotting--the molecular weight of the proteins involved, their topology and their mode of association with membranes. Morphological analysis has been performed by immunocytochemistry at the light and electron microscopic level. Judging by these criteria, the GCII antigen is a protein of 44 kDa which is loosely associated with the endodomain of Golgi cisternae. GCIII is a detergent-binding glycoprotein of 130 kDa whose epitope is on the endodomain of Golgi cisternae. SMI is a detergent-binding glycoprotein of 58 to 90 kDa found at several stations along the endocytic path: in coated pits, coated vesicles, endocytic vesicles, but not in lysosomes. The epitope recognized by the corresponding antibody faces the ectodomain. When this antibody is added to living cells in culture, it is rapidly internalized. SMII is a detergent-binding glycoprotein of 140 kDa. The epitope recognized is restricted to membranes of Golgi complex cisternae and multivesicular bodies. These reagents should be useful for dissection and perturbation of vesicular traffic.     

Antibodies to the Golgi complex and the rough endoplasmic reticulum

Rabbits were immunized with membrane fractions from either the Golgi complex or the rough endoplasmic reticulum (RER) by injection into the popliteal lymph nodes. The antisera were then tested by indirect immunofluorescence on tissue culture cells or frozen, thin sections of tissue. There were may unwanted antibodies to cell components other than the RER or the Golgi complex, and these were removed by suitable absorption steps. These steps were carried out until the pattern of fluorescent labeling was that expected for the Golgi complex or RER. Electron microscopic studies, using immunoperoxidase labeling of normal rat kidney (NRK) cells, showed that the anti-Golgi antibodies labeled the stacks of flattened cisternae that comprise the central feature of the Golgi complex, many of the smooth vesicles around the stacks, and a few coated vesicles. These antibodies were directed, almost entirely, against a single polypeptide with an apparent molecular weight of 135,000. The endoplasmic reticulum (ER) in NRK cells is an extensive, reticular network that pervades the entire cell cytoplasm and includes the nuclear membrane. The anit-RER antibodies labeled this structure alone at the light and electron microscopic levels. They were largely directed against four polypeptides with apparent molecular weights of 29,000, 58,000, 66,000, and 91,000. Some examples are presented, using immunofluorescence microscopy, where these antibodies have been used to study the Golgi complex and RER under a variety of physiological and experimental condition . For biochemical studies, these antibodies should prove useful in identifying the origin of isolated membranes, particularly those from organelles such as the Golgi complex, which tend to lose their characteristic morphology during isolation.     

Differential cell surface expression of four lysosomal integral membrane proteins (LIMPs) in normal rat kidney cells

Four monoclonal antibodies generated against rat, hepatocyte lysosomal integral membrane protein (LIMPs) (Barriocanal et al., 1986a, b) were used as probes to ascertain the distribution of similar proteins in normal rat kidney (NRK) cells. Comparison of immunoprecipitations of LIMPs 1-4 from hepatocytes and NRK cells revealed a marked similarity in the proteins, in both cell types, as determined by SDS-PAGE. Further, the LIMP epitopes recognized by the antibodies are situated intravesicularly. Ultrastructural immunocytochemistry, using pre-embedding peroxidase, revealed that primary and secondary lysosomes in NRK cells are readily stained with all four antibodies, as well as vesicles in the Golgi region. Immunofluorescence microscopy of non-permeabilized NRK cells with antibodies recognizing LIMPs 1 and 4 illustrated a limited but significant punctate staining pattern of the cell surface. Ultrastructural immunoperoxidase indicated these sites to be cell surface localized coated pits and vesicles. However, it is known that all LIMPs are expressed on the cell surface, albeit at different concentrations, although the total number of each LIMP per cell, respectively, is approximately the same (Barriocanal et al., 1987). Treatment of NRK cells with the acidotropic agent NH4Cl decreased the cell surface expression of LIMPs 1, 3 and 4, but had no effect on LIMP 2. Further, the relative diminution of the cell surface expression varied among the four LIMPs. These results are interpreted to suggest that not all lysosomes contain the same integral membrane proteins in their vesicle container.     

A role for clathrin in reassembly of the Golgi apparatus

The Golgi apparatus is a highly dynamic organelle whose organization is maintained by a proteinaceous matrix, cytoskeletal components, and inositol phospholipids. In mammalian cells, disassembly of the organelle occurs reversibly at the onset of mitosis and irreversibly during apoptosis. Several pharmacological agents including nocodazole, brefeldin A (BFA), and primary alcohols (1-butanol) induce reversible fragmentation of the Golgi apparatus. To dissect the mechanism of Golgi reassembly, rat NRK and GH3 cells were treated with 1-butanol, BFA, or nocodazole. During washout of 1-butanol, clathrin, a ubiquitous coat protein implicated in vesicle traffic at the trans-Golgi network and plasma membrane, and abundant clathrin coated vesicles were recruited to the region of nascent Golgi cisternae. Knockdown of endogenous clathrin heavy chain showed that the Golgi apparatus failed to reform efficiently after BFA or 1-butanol removal. Instead, upon 1-butanol washout, it maintained a compact, tight morphology. Our results suggest that clathrin is required to reassemble fragmented Golgi elements. In addition, we show that after butanol treatment the Golgi apparatus reforms via an initial compact intermediate structure that is subsequently remodeled into the characteristic interphase lace-like morphology and that reassembly requires clathrin.     

HVEM tomography of the trans-Golgi network: structural insights and identification of a lace-like vesicle coat

High voltage electron microscopy and computer axial tomography have been used to study the 3-D structure of trans-Golgi cisternae and trans- Golgi networks (TGNs) in NRK cells. Both structures were specifically labeled by photoconversion of a fluorescent analogue of ceramide using a modification of the techique of Pagano et al. (J. Cell Biol. 1991. 113: 1267-1279). Regions of the Golgi ribbon in fixed, stained cells were cut in 250-nm sections and analyzed by tilt series microscopy and subsequent tomographic reconstruction. Resolution of the reconstructions ranged from 6 to 10 nm. The size and structure of the TGN varied considerably throughout the Golgi ribbon; all reconstructions were made from regions with pronounced TGN. Most regions analyzed contained multiple (2-4) Golgi cisternae that stain with ceramide. These "peel off" from the closely stacked cisternae and are continuous at their ends with tubules that contribute to the TGN. Most vesicular profiles visualized in the TGN are connected to TGN tubules. The budding of vesicles appears to occur synchronously along the length of a TGN tubule. Two distinct coats were visualized on budding vesicles: clathrin cages and a novel, lace-like structure. Individual TGN tubules produce vesicles of only one coat type. These observations lead to the following predictions: (a) sorting of molecules must occur prior to the formation of TGN tubules; (b) vesicle formation takes place almost synchronously along a given TGN tubule; and (c) lace-like coats form an exocytic vesicles.     

Clathrin beta-light chain of rat liver coated vesicles is phosphorylated in vitro and in vivo

Clathrin beta-light chain of rat liver coated vesicles is phosphorylated in vitro in the presence of poly(L-lysine) by an endogenous protein kinase which appears to be similar to casein kinase II. Clathrin beta-light chain is also phosphorylated in vivo. After injection of [32P]phosphate into rats and preparation of purified coated vesicles in the presence of phosphatase inhibitors, electrophoretic analysis showed the presence of several labeled polypeptides including clathrin beta-light chain. A polypeptide of 50 kDa, which may correspond to the major polypeptide phosphorylated in vitro of coated vesicles, is also labeled in vivo.     

Cell-free analysis of Golgi apparatus membrane traffic in rat liver

 Cell-free systems for the analysis of Golgi apparatus membrane traffic rely either on highly purified cell fractions or analysis by specific trafficking markers or both. Our work has employed a cell-free transfer system from rat liver based on purified fractions. Transfer of any constituent present in the donor fraction that can be labeled (protein, phospholipid, neutral lipid, sterol, or glycoconjugate) may be investigated in a manner not requiring a processing assay. Transition vesicles were purified and Golgi apparatus cisternae were subfractionated by means of preparative free-flow electrophoresis. Using these transition vesicles and Golgi apparatus subfractions, transfer between transitional endoplasmic reticulum and cis Golgi apparatus was investigated and the process subdivided into vesicle formation and vesicle fusion steps. In liver, vesicle formation exhibited both ATP-independent and ATP-dependent components whereas vesicle fusion was ATP-independent. The ATP-dependent component of transfer was donor and acceptor specific and appeared to be largely unidirectional, i.e., ATP-dependent retrograde (cis Golgi apparatus to transitional endoplasmic reticulum) traffic was not observed. ATP-dependent transfer in the liver system and coatomer-driven ATP-independent transfer in more refined yeast and cultured cell systems are compared and discussed in regard to the liver system. A model mechanism developed for ATP-dependent budding is proposed where a retinol-stimulated and brefeldin A-inhibited NADH protein disulfide oxidoreductase (NADH oxidase) with protein disulfide-thiol interchange activity and an ATP-requiring protein capable of driving physical membrane displacement are involved. It has been suggested that this mechanism drives both the cell enlargement and the vesicle budding that may be associated with the dynamic flow of membranes along the endoplasmic reticulum-vesicle-Golgi apparatus-plasma membrane pathway. Accepted: 26 January 1998     

Following the Fate In Vivo of COPI Vesicles Generated In Vitro

COPI vesicles are a class of transport carriers that function in the early secretory pathway. Their fate and function are still controversial. This includes their contribution to bidirectional transport within the Golgi apparatus and their role during cell division. Here we describe a method that should address several open questions about the fate and function of COPI vesicles in vivo . To this end, fluorescently labeled COPI vesicles were generated in vitro from isolated rat liver Golgi membranes, labeled with the fluorescent dyes Alexa-488 or Alexa-568. These vesicles appeared to be active and colocalized with endogenous Golgi membranes within 30 min after microinjection into mammalian cells. The COPI vesicle-derived labeled membrane proteins could be classified into two types that behaved like endogenous proteins after Brefeldin A treatment.     

Phosphoproteins and protein kinases of the Golgi apparatus membrane

Incubation of a highly purified fraction derived from rat liver Golgi apparatus with [gamma-32P]ATP results in phosphorylation of several endogenous phosphoproteins. One phosphoprotein with an apparent Mr of 48,300 is radiolabeled to an apparent extent at least 5-fold higher than any other phosphoprotein as part of either the Golgi apparatus or highly purified rat liver fractions derived from the rough endoplasmic reticulum, mitochondria, plasma membrane, coated vesicles, cytosol, and total homogenate. Approximately 70% of the 48.3-kDa phosphoprotein appears to be a specific extrinsic Golgi membrane protein with the phosphorylated amino acid being threonine. The protein kinase which phosphorylates the 48.3-kDa protein is an intrinsic Golgi membrane protein and is dependent on Mg2+, independent of Ca2+, calmodulin, and cAMP, and is inhibited by N-ethylmaleimide. Preliminary evidence suggests that there are also intrinsic membrane protein kinases in the Golgi apparatus which are dependent on Ca2+ and cAMP. The physiological role of the above phosphoproteins and protein kinases is not known.     

Nucleocytoplasmic traffic of proteins.

We have used the synchronized formation of a mixed cytoplasm upon heterokaryon formation as a model for investigating the cisternal-specific transport of resident proteins between neighboring Golgi apparatus. Rat NRK and hamster 15B cells were fused by UV-inactivated Sindbis virus and then incubated for various time periods in the presence of cycloheximide. The resident Golgi apparatus proteins, rat GIMPc and Golgp 125, were localized with species-specific monoclonal antibodies. Immunofluorescent colocalization of rat and hamster Golgi membrane proteins was observed with a t1/2 of 1.75 h at 37 degrees C. Colocalization of resident, but not transient, Golgi membrane protein was concomitant with formation of a large extended Golgi complex and was accompanied by the acquisition of endoglycosidase H resistance by preexisting Golgp 125. Dispersal of the extended Golgi complex by nocodazole revealed that colocalization of resident Golgi proteins was due to intermixing of proteins in the same Golgi element rather than overlapping of closely apposed Golgi structures. Incubation of the polykaryons at 20 degrees C inhibited both the colocalization of GIMPc and Golgp 125 and the formation of an extended Golgi complex. Little change in the number of cisternae/stack in cross sections of the Golgi apparatus was observed upon cell fusion, and in the extended Golgi complex the hamster resident protein remained localized to one side of the Golgi stack. Surprisingly, the morphological identity of the rat and hamster Golgi units appeared to be maintained in the heterokaryons. These results suggest that the intermixing of resident Golgi membrane proteins requires direct physical continuity between Golgi elements and that resident Golgi membrane proteins are preferentially excluded from the non-clathrin-coated transport vesicles budding from Golgi cisternae.     

The organelles of the trans domain of the cell. Ultrastructural localization of sialoglycoconjugates using Limax flavus agglutinin

The subcellular distribution of sialic acid was determined at the ultrastructural level using Limax flavus agglutinin (LFA). This lectin, which is specific for N-acetylneuraminic acid and N-glycolylneuraminic acid, was covalently conjugated to horseradish peroxidase (HRP). The conjugates (LFA-HRP) were applied to aldehyde-fixed, saponin-permeabilized 3T3 cells in pre-embedding labeling electron microscopy. Peroxidase label was detected in a patchy distribution at the cell surface, and in plasma-membrane-coated pits, endocytic vesicles (receptosomes), multivesicular bodies, and lysosomes. Smooth-surfaced tubular and vesicular structures, similar to those that participate in membrane recycling, were labeled. In the Golgi complex, more than half of the cisternae contained label--typically only one cisterna on the cis side was unlabeled. Heavily labeled structures of the trans Golgi included a reticular membranous system with coated regions--50-80 nm diameter vesicular or pit-like profiles and larger coated vacuoles. Smooth 200-300 nm vacuoles were labeled on the trans side of the Golgi stack. Similar structures have been previously shown to participate in the exocytosis of plasma membrane and secretory glycoproteins from the Golgi stacks. These findings identify those intracellular organelles that are functionally at the level of, or distal to, the sialyltransferase-containing membranes of the Golgi, and distinguish them from the pre-Golgi membranous structures. The LFA-HRP conjugate is an indicator for this functional trans domain of the cell, and should be applicable for ultrastructural double-label experiments as a cis versus trans marker of the exocytic pathway.     

Intracellular vesicles involved in the transport of Semliki Forest virus membrane proteins to the cell surface.

The route of transport of Semliki Forest virus (SFV) membrane glycoproteins to the plasma membrane was studied using immunoperoxidase electron microscopy. SFV glycoproteins were localized in cultured BHK-21 fibroblasts infected with a temperature-sensitive mutant ts-1 of SFV, which shows a temperature-dependent, reversible defect in the transport of membrane glycoproteins to the cell surface. At 39 degrees C (restrictive temperature) the viral proteins were retained in the endoplasmic reticulum and the nuclear membrane. After shift of the infected cultures to 28 degrees C (permissive temperature) the proteins were synchronously transported to the Golgi complex. In the Golgi complex the labeled proteins were first (at 2.5 min) detected in large Golgi-associated vacuoles (GAV). Subsequently, i.e., at 5-30 min, the viral glycoproteins appeared in the cisternal stack: at 5 min the label was found in one or two of the proximal cisternae whereas at 15 or 30 min also the more distal cisternae were partially or uniformly labeled. At all time points examined after the temperature-shift, peroxidase label was found in 50 nm vesicles which were frequently coated. At 30 min, in addition to the 50 nm vesicles, larger 80 nm vesicles, which often had a cytoplasmic coat were labeled in the Golgi region. These results identify two major size classes of both coated and smooth vesicles which appear to function in the transport of the viral membrane proteins from the endoplasmic reticulum via distinct GAV and the stacked Golgi cisternae to the plasma membrane.     

Transitional endoplasmic reticulum membranes and vesicles isolated from animals and plants

Summary The process of formation from endoplasmic reticulum and transfer to Golgi apparatus of small 50–70 nm transition vesicles has been reconstituted in a cell-free system. Fractions enriched in transition elements derived from part-rough, part-smooth transitional regions of the endoplasmic reticulum were prepared from elongation zones of hypocotyls of etiolated seedlings of soybean and coleoptiles of maize and were compared with those from rat liver. When activated with nucleoside triphosphate, cytosol and an ATP regenerating system, time- and temperature-dependent transfer of membranes to Golgi apparatus acceptor was demonstrated. The fractions enriched in transition elements were radioiodinated with¹²⁵I by the Bolton-Hunter procedure. Acceptor Golgi apparatus stacks were immobilized to nitrocellulose strips to facilitate analysis. In heterologous transfer experiments, the plant and animal acceptors and donors could be interchanged. The transfer was limited primarily by the donor (rat liver > soybean hypocotyl > maize coleoptiles) and determined secondarily by the source of the acceptor. The acceptor fractions were most efficacious when prepared from the same source as the donor. Thus, 50–70 nm vesicles bud from transitional endoplasmic reticulum elements of plants function in a manner similar to those of animal cells to transfer membrane materials to the Golgi apparatus. The recognition signals that determine vesicle fusion appear to be conserved both among species and between the plant and animal kingdoms to the extent that donor and acceptor sources may be interchanged with only small reductions in overall efficiency of transfer.Abbrevations HEPES N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid - EDTA ethylenediaminetetraacetic acid     

Apolipoprotein E localization in rat hepatocytes by immunogold labeling of cryothin sections

The distribution of apolipoprotein (apo) E in rat hepatocytes was investigated with an affinity-purified polyclonal antibody raised against apoE isolated from hepatogeneous very low density lipoproteins (VLDL). The distribution of this antibody was visualized with colloidal gold complexed to anti-rabbit IgG. By epipolarization microscopy, apoE was found uniformly along the basolateral surfaces of all hepatic parenchymal cells, showing a striking intensity along the sinusoidal front. Punctate deposits of colloidal gold appeared to be randomly distributed within all hepatocytes. Widely scattered Kupffer cells also stained for apoE. Electron microscopic examination of immunogold-labeled cryothin sections showed that hepatocytic microvilli projecting into the space of Disse consistently contained clusters of immunogold. The gold particles were variably associated with evident lipoprotein particles, raising the possibility that apoE alone may bind to receptors or other macromolecules at the surface of hepatocytes. Endosomes near the sinusoidal front and multivesicular bodies in the Golgi/biliary area labeled intensely for apoE, consistent with a high content of apoE associated with triglyceride-rich lipoprotein remnants contained within these organelles. Some but not all nascent VLDL particles within putative forming Golgi secretory vesicles were labeled, but many other Golgi vesicles and cisternae that lacked evident VLDL particles were also labeled. These results suggest that at least some apoE associates with nascent VLDL in forming Golgi secretory vesicles. Unexpectedly, the matrix of all hepatocytic peroxisomes was heavily labeled. Immunoblots with the affinity-purified anti-rat apoE IgG against proteins from highly purified peroxisomes isolated from rat hepatocytes revealed a protein with an apparent molecular mass of 34.5 kDa, similar to that of rat apoE in rat blood plasma. In addition, gold was sometimes found in the area either adjacent to peroxisomes or between multivesicular bodies and the bile canaliculus not evidently associated with a membranous compartment. These observations suggest that apoE may participate in interorganellar cholesterol transport within hepatocytes.     

Reconstitution of vesiculated Golgi membranes into stacks of cisternae: requirement of NSF in stack formation

We have developed an in vitro system to study the biochemical events in the fusion of ilimaquinone (IQ) induced vesiculated Golgi membranes (VGMs) into stacks of cisternae. The Golgi complex in intact normal rat kidney cells (NRK) is vesiculated by treatment with IQ. The cells are washed to remove the drug and then permeabilized by a rapid freeze-thaw procedure. VGMs of 60 nm average diameter assemble into stacks of Golgi cisternae by a process that is temperature dependent, requires ATP and a high speed supernatant from cell extract (cytosol), as revealed by immunofluorescence and electron microscopy. The newly assembled stacks are functionally active in vesicular protein transport and contain processing enzymes that carry out Golgi specific modifications of glycoproteins. The fusion of VGMs requires NSF, a protein known to promote fusion of transport vesicles with the target membrane in the exocytic and endocytic pathways. Immunoelectron microscopy using Golgi specific anti-mannosidase II antibody reveals that VGMs undergo sequential changes in their morphology, whereby they first fuse to form larger vesicles of 200-300-nm average diameter which subsequently extend into tubular elements and finally assemble into stacks of cisternae.     

ATP-dependent fusion of liposomes with the Golgi apparatus of perforated cells

We examined the interactions of lipid vesicles (liposomes) labeled with various fluorescent markers with the intracellular membranes of semi-intact ("perforated") fibroblasts. When incubations were performed in the presence of an ATP-regenerating system, both vesicle lipids and entrapped water soluble markers were transferred to the Golgi apparatus of treated cells, indicative of membrane fusion. Fusion occurred using unilamellar vesicles 30-80 nm in diameter and composed of phosphatidylcholine alone, but was inhibited when equimolar amounts of either phosphatidylserine or sphingomyelin were present in the vesicles. Lipid vesicle-Golgi membrane fusion was also inhibited by pretreatment of the perforated cells with N-ethylmaleimide. These findings suggest that lipid vesicles may be useful for delivery of labeled lipids, macromolecules, and dyes to the Golgi apparatus, and for modeling the interactions of transport vesicles with the Golgi apparatus.