Clathrin assembly proteins: affinity purification and a model for coat assembly

Assembly protein (AP) preparations from bovine brain coated vesicles have been fractionated by clathrin-Sepharose affinity chromatography. Two distinct fractions that possess coat assembly activity were obtained and are termed AP-1 and AP-2. The AP-1, not retained on the resin, has principal components with molecular weights of 108,000, 100,000, 47,000, and 19,000. The AP-2, bound to the resin and eluted by Tris-HCl at a concentration that parallels the latter's effect on coat disassembly, corresponds to the active complex described previously (Zaremba, S., and J. H. Keen, 1983, J. Cell Biol., 97:1339-1347). Its composition is similar to that of the AP-1 in that it contains 100,000-, 50,000-, and 16,000-mol-wt polypeptides in equimolar amounts; minor amounts of 112,000- and 115,000-mol-wt polypeptides are also present. Both are distinct from a recently described assembly protein of larger subunit molecular weight that we term AP-3. These results indicate the existence of a family of assembly proteins within cells. On incubation with clathrin both AP-1 and AP-2 induce the formation of coat structures, those containing AP-1 slightly smaller (mean diameter = 72 nm) than those formed in the presence of AP-2 (mean diameter = 79 nm); both structures have been detected previously in coated vesicle preparations from brain. Coats formed in the presence of AP-2 consistently contain approximately one molecule each of the 100,000-, 50,000-, and 16,000-mol-wt polypeptides per clathrin trimer. By low angle laser light scattering the molecular weight of native AP-2 was determined to be approximately 343,000, indicating that it is a dimer of each of the three subunits, and implying that it is functionally bivalent in clathrin binding. A model for AP-mediated coat assembly is proposed in which a bivalent AP-2 molecule bridges the distal legs or terminal domains of two clathrin trimers that are destined to occupy adjacent vertices in the assembled coat. Binding of a second AP-2 molecule locks these two trimers in register for assembly and further addition of AP-2 to free trimer legs promotes completion of the clathrin lattice. Effects of AP binding on the angle and flexibility of the legs at the hub of the trimer (the "pucker") are suggested to account for the characteristic size distributions of coats formed under varied conditions and, more speculatively, to contribute to the transformation of flat clathrin lattices to curved coated vesicles that are thought to occur during endocytosis.     

Presence of an immunologically-defined class of peri-Golgi vesicles in cultured mammalian cells

Immunsera of mice injected with clathrin-depleted coated vesicle membranes, purified from rat liver, revealed a preferential labeling of some perinuclear structures by immunofluorescence microscopy on NRK cells. Subsequent production of 4 monoclonal antibodies was achieved. The antigen was strictly located in the Golgi area of the cell but was not an intrinsic element of the Golgi complex. The restricted location of the structures excluded these were lysosomes which appeared more dispersed in these cells. After nocodazole treatment the material was found dispersed in the cytoplasm. This provided a means of distinguishing the antigen from clathrin-coated structures and Golgi intrinsic elements. Immunolocalization at the electron microscope level confirmed the data obtained at the light level. Some peroxidase reaction product was rarely associated with Golgi elements, but predominantly stained small neighboring Golgi vesicles (50 nm diameter), as well as tubulo-elongated structures and some large (500 nm) irregular-shaped vesicles. A 32 kDa molecular weight antigen was characterized by immunopurification from NRK cells metabolically labeled with 35S-Met. This 32 kDa antigen appeared as part of a higher multimolecular membrane component of 300 kDa. A 170 kDa and a 70 kDa components were immunodetected in a semi-purified membrane fraction from rat liver, demonstrating that the antigen was a minor but very antigenic contaminant of the coated vesicle preparation used as immunogen. In conclusion, the labeled peri-Golgi structures may be part of the newly characterized trans-Golgi network and/or of the reticular/vesicular endosomal, prelysosomal structure recently described.     

Endocytosis in nonciliated epithelial cells of the ductuli efferentes in the rat

The nonciliated cells lining the ductuli efferentes presented three distinct cytoplasmic regions. The apical region contained, in addition to cisternae of endoplasmic reticulum and mitochondria, two distinct membranous elements. The tubulovesicular system consisted of dilated tubules connected to the apical plasma membrane and subjacent distended vesicular profiles. The apical tubules, not connected to the cell surface, consisted of numerous densely stained tubules of small size which contain a compact, finely granulated material. The supranuclear region, in addition to a Golgi apparatus and ER cisternae, contained dilated vacuoles, pale and dense multivesicular bodies, as well as numerous dense granules identified cytochemically as lysosomes. The basal region contained the nucleus and many lipid droplets. The endocytic activity of these cells was investigated using cationic ferritin (CF) and concanavalin-A-ferritin (Con-A-ferritin) as markers of adsorptive endocytosis; and native ferritin (NF), concanavalin-A-ferritin in the presence of alpha-methyl mannoside, and horseradish peroxidase or albumin bound to colloidal gold for demonstrating fluid-phase endocytosis. These tracers were injected separately into the rete testis, and animals were sacrificed at various time intervals after injection. At 1 min, CF or Con-A-ferritin were seen bound to the apical plasma membrane, to the membrane of microvilli, and to the membrane delimiting elements of the tubulovesicular system. Between 2 and 5 min, these tracers accumulated in the densely stained apical tubules and at 15 min in the dilated vacuoles. Between 30 min and 1 hr, the tracers appeared in multivesicular bodies of progressively increasing density, whereas at 2 hr and later time intervals, many dense lysosomal elements became labeled. The tracers for fluid-phase endocytosis showed a distribution similar to that for CF or Con-A-ferritin except that they did not bind to the apical plasma membrane, microvilli, or membrane delimiting the tubulovesicular system. At no time interval were any of the tracers observed in the abluminal spaces. Thus, the nonciliated epithelial cells of the ductuli efferentes are actively involved in fluid-phase and adsorptive endocytosis, both of which result in the sequestration of endocytosed material within the lysosomal apparatus of the cell.     

cAMP induces a rapid and reversible modification of the chemotactic receptor in Dictyostelium discoideum

Stimulation, within 1 min after cAMP stimulation, of aggregation-competent Dictyostelium discoideum amebae was found to cause a rapid (within 1 min) modification of the cell's surface cAMP receptor. The modified receptor migrated on SDS PAGE as a 47,000-mol-wt protein, as opposed to a 45,000-mol-wt protein labeled on unstimulated cells. The length of time this modified receptor could be detected depended upon the strength of the cAMP stimulus: 3-4 min after treatment with 10(-7) M cAMP, cells no longer possessed the 47,000-mol-wt form of the cAMP receptor. Instead, the 45,000-mol-wt form was present. Stimulation of cells with 10(-5) M cAMP, however, resulted in the persistent (over 15 min) expression of the modified receptor. The time course, concentration dependence, and specificity of stimulus for this cAMP-induced shift in the cAMP receptor were found to parallel the cAMP-stimulated phosphorylation of a 47,000-mol-wt protein. In addition, both phenomena were shown to occur in the absence of endogenous cAMP synthesis. The possibility that the cAMP receptor is phosphorylated in response to cAMP stimulation, and the role of this event in cell desensitization, are discussed.     

Identification of an endosomal antigen specific to absorptive cells of suckling rat ileum

A membrane fraction enriched in apical endosomal tubules was isolated from absorptive cells of suckling rat ileum and used as an immunogen to generate anti-endosome monoclonal antibodies. By immunofluorescence, one of these antibodies bound exclusively to the region of the apical endocytic complex in ileal absorptive cells, but not to other cell types. Immunoblot analysis showed the antigen as a diffuse 55-61-kD band which was highly enriched in the endosome fraction over whole-cell homogenate. The antigen appears to be an intramembrane glycoprotein: it partitioned primarily in the detergent phase after TX-114 extraction, and shifted to 44 kD after chemical deglycosylation. EM immunocytochemistry showed that the antibody bound to the luminal side of endosomal tubule membranes, a portion of endosomal vesicle membranes, and in endocytic pits of apical plasma membranes. However, it did not bind to multivesicular bodies, the giant lysosome, or other organelles. Immunocytochemistry after uptake with adsorbed or soluble tracer proteins showed that the antigen labeled portions of both prelysosomal pathways previously described in these cells (Gonnella, P.A., and M. R. Neutra, 1984, J. Cell Biol., 99:909-917). The function of this glycoprotein is not known, but inasmuch as it has been detected only in absorptive cells of suckling rat ileum, it may serve a function specific to these cells. Nevertheless, this endosomal antigen, designated glycoprotein (gp) 55-61, will serve as a useful marker for exploring membrane dynamics in early stages of the endocytic pathway.     

A multicompartmental model of fluid-phase endocytosis in rabbit liver parenchymal cells.

Fluid-phase endocytosis was studied in isolated rabbit liver parenchymal cells by using 125I-poly(vinylpyrrolidone) (PVP) as a marker. First, uptake of 125I-PVP by cells was determined. Also, cells were loaded with 125I-PVP for 20, 60 and 120 min, and release of marker was monitored for 120-220 min. Then we used the Simulation, Analysis and Modeling (SAAM) computer program and the technique of model-based compartmental analysis to develop a mechanistic model for fluid-phase endocytosis in these cells. To fit all data simultaneously, a model with three cellular compartments and one extracellular compartment was required. The three kinetically distinct cellular compartments are interpreted to represent (1) early endosomes, (2) a prelysosomal compartment equivalent to the compartment for uncoupling of receptor and ligand (CURL) and/or multivesicular bodies (MVB), and (3) lysosomes. The model predicts that approx. 80% of the internalized 125I-PVP was recycled to the medium from the early-endosome compartment. The apparent first-order rate constant for this recycling was 0.094 min-1, thus indicating that an average 125I-PVP molecule is recycled in 11 min. The model also predicts that recycling to the medium occurs from all three intracellular compartments. From the prelysosomal compartment, 40% of the 125I-PVP molecules are predicted to recycle to the medium and 60% are transferred to the lysosomal compartment. The average time for recycling from the prelysosomal compartment to the medium was estimated to be 66 min. For 125I-PVP in the lysosomal compartment, 0.3%/min was transferred back to the medium. These results, and the model developed to interpret the data, predict that there is extensive recycling of material endocytosed by fluid-phase endocytosis to the extracellular environment in rabbit liver parenchymal cells.     

LIM kinase 1: evidence for a role in the regulation of intracellular vesicle trafficking of lysosomes and endosomes in human breast cancer cells

LIM kinase (LIMK) plays a critical role in stimulus-induced remodeling of the actin cytoskeleton by linking signals from the Rho family GTPases to changes in cofilin activity. Recent studies have shown an important role for LIMK1 signaling in tumor cell invasion through regulating actin dynamics. In this study, we investigate the role of LIMK1 in intracellular vesicle trafficking of lysosomes/endosomes. We analyzed by confocal immunofluorescence microscopy the cellular distribution of lysosomal proteins and the endocytosis of an endocytic tracer, epidermal growth factor (EGF), in LIMK1-transfected cells. We found in these cells an abnormal dispersed translocation of lysosomes stained for LIMPII and cathepsin D throughout the cytoplasm. The small punctate structures that stained for these lysosomal proteins were redistributed to the periphery of the cell. Computational 3D-image analysis of confocal immunofluorescence micrographs further demonstrated that these vesicles did not colocalize with the transferrin receptor, an early endosomal marker. Furthermore, LIMPII-positive lysosomes did not colocalize with early endosomes labeled with endocytosed Texas red-transferrin. These results indicate that there is no mixing between dispersed lysosomes and early endosomes in the LIMK1-transfected cells. Moreover, LIMK1 overexpression resulted in a marked retardation in the receptor-mediated internalization of Texas red-labeled EGF in comparison with mock-transfected cells. At 30 min after internalization, most of the Texas red-EGF staining overlapped with LIMPII-positive late endosomes/lysosomes in mock-transfected cells, whereas in LIMK1 transfectants only a small fraction of internalized EGF colocalized with LIMPII-positive structures in the perinuclear region. Taken together, the findings presented in this paper suggest that LIMK1 has a role in regulating vesicle trafficking of lysosomes and endosomes in invasive tumor cells.     

Exchange kinetics and composition of endocytic membranes in terms of plasma membrane constituents: a morphometric study in macrophages

Intracellular membrane traffic, during endocytosis in mouse bone marrow-derived macrophages, was studied quantitatively by morphometric and kinetic analysis. Three functionally different markers were used: Horseradish peroxidase (HRP) served as a fluid-phase (FP) marker (1000 micrograms HRP/ml in the presence of mannan) or as a receptor-mediated (RM) membrane marker (25 micrograms HRP/ml) and, third, plasma membrane (PM) glycoconjugates, enzymatically labeled with [3H]galactose at the cell surface, served as a covalent membrane marker. The cell surface was labeled with [3H]galactose, followed by either FP or by RM uptake of HRP. The kinetics of the intracellular appearance of the markers were measured as the membrane area stained by HRP-reaction product and as the number of autoradiographic grains associated with these membranes. The following compartments were distinguished: PM, coated vesicles (VI), pinosomes or endosomes (VII), secondary lysosomes (VIII), and HRP-negative vesicles (EV). Tubular structures of VII became labeled with HRP only during RM uptake. The markers flowed first into VI and VII, and after 5 min into VIII. EV became labeled with the covalent membrane marker starting from 5 min. The ratio of autoradiographic grain number to HRP-stained membrane area remained constant with time although substantially different for the various compartments, viz. 100% (VI), 50% (VII and EV) and 30% (VIII) as compared to the PM (100%). This indicated that endosomes were only partially derived from internalized PM and that secondary lysosomes contained a substantial pool of PM constituents. The observed kinetics suggested that once every 30 to 40 min the entire PM was internalized, the bulk of which was recycled after 4 min from a prelysosomal compartment(s) leaving only 12 to 20% for recycling via membranes of secondary lysosomes after a residence time of 24 to 33 min.     

Cell-substratum adhesion in embryonic chick central nervous system is mediated by a 170,000-mol-wt neural-specific polypeptide

Embryonal chick neural retina cells release into the culture medium a complex of proteins and glycosaminoglycans, termed adherons, that promote cell to substratum adhesion. A monoclonal antibody (C1H3) blocks adheron-mediated cell to substratum adhesion and specifically binds to a 170,000-mol-wt protein present in retinal adherons (Cole, G.J., and L. Glaser, 1984, J. Biol. Chem., 259:4031-4034). The 170,000-mol-wt protein also can be identified in embryonic chick brain and peripheral nervous tissue. In the neural retina, C1H3 also binds to a second antigen with a molecular weight of 140,000 that is absent in the brain. Embryonic brain, therefore, provides a source for the immunopurification of the 170,000-mol-wt protein. Brain adherons also contain the 170,000-mol-wt protein, and cell to substratum adhesion mediated by these adherons is blocked by the C1H3 monoclonal antibody. The 170,000-mol-wt protein in the brain is therefore functionally identical to that in the retina. To demonstrate that adheron-mediated cell to substratum adhesion is caused by cell binding to the 170,000-mol-wt protein, we showed that (a) protease digestion, but not glycosaminoglycan hydrolase digestion of adherons, blocked their ability to bind cells to substratum; (b) the immunopurified 170,000-mol-wt protein blocks adheron-mediated cell to substratum adhesion; and (c) cells can bind to immunopurified 170,000-mol-wt protein bound to glass surfaces.     

Pre-lysosomal divergence of alpha 2-macroglobulin and transferrin: a kinetic study using a monoclonal antibody against a lysosomal membrane glycoprotein (LAMP-1)

We have used electron microscopic immunocytochemistry to compare the distribution of LAMP-1, a marker for lysosomal membranes, with the intracellular localization of alpha 2-macroglobulin (alpha 2-M) and transferrin at various time points after their endocytosis into cultured NIH 3T3 cells. The purposes of this study were (a) to determine how soon endocytic ligands reach lysosomal organelles, (b) to examine whether the intermediate endocytic vesicles gained lysosomal markers gradually or in a precipitous, discrete event, and (c) to examine the relationship, if any, between the pathway of recycling ligands and lysosomes. At early time points (0-5 min) after initiation of endocytosis, most structures containing alpha 2-M labeled with colloidal gold (receptosomes) were not labeled by anti-LAMP-1 detected using ferritin bridge or peroxidase immunocytochemistry. At late time points (greater than or equal to 15 min), the structures containing alpha 2-M (lysosomes) were strongly labeled by anti-LAMP-1. In contrast, transferrin that was directly labeled with ferritin was mostly located in LAMP-1-negative structures at all time points studied. The proportion of alpha 2-M-gold containing vesicles strongly labeled for LAMP-1 roughly paralleled the proportion of alpha 2-M-gold-containing structures positive for cytochemically detectable acid phosphatase. Our data indicate that ligands such as transferrin that are internalized through coated pits and receptosomes, but not delivered to lysosomes, do not traverse a lysosomal organelle compartment as marked by LAMP-1 content. Ligands such as alpha 2-M that are destined for lysosomal delivery reach a LAMP-1-positive organelle compartment only after they traverse LAMP-1-negative, non-lysosomal vesicles previously described as receptosomes.     

Endocytosis in spermatids during spermiogenesis of the mouse

In the course of spermiogenesis in the mouse, spermatid cytoplasm contains numerous membrane pits, vesicles and membranous tubules which are frequently anastomosed. Pale and dense multivesicular bodies (MVB) and secondary lysosome-like structures are also present in the cytoplasm. In order to study the pathway of non-specific adsorptive endocytosis in spermatids, cationic ferritin (CF) was directly microinjected into the lumen of seminiferous tubules, and added to germinal cell culture. Tissue and cultures were fixed at various time intervals after injection. Two-5 hr after microinjection of tracer, CF was found simultaneously in vesicles, tubules, MVB and in lysosome-like bodies present in spermatids at all steps of spermiogenesis. Various membranous components of the Golgi medulla, and the innermost transsaccule of the Golgi cortex were labelled simultaneously. In primary cultures of spermatids, the vesicles contained the marker 5 min after its deposition; 10 min after deposition, CF was evident in tubules; at 30 min, CF was present in pale MVB; at 1 hr, the dense MVB and lysosome-like bodies were labelled. Finally, at 2 hr 30 min, vesicles and tubules of the Golgi medulla contained CF grains. Apparently spermatids are very active cells in the process of adsorptive endocytosis throughout spermiogenesis. Endocytosis in spermatids is probably one of the mechanisms involved in the uptake of material used to build up spermatozoa components. The strong labelling of the Golgi region probably point to its role in recycling endocytosed membranes.     

A heparin-binding domain from N-CAM is involved in neural cell- substratum adhesion

Cell-substratum adhesion in the embryonic chicken nervous system has been shown to be mediated in part by a 170,000-mol-wt polypeptide that is a component of adherons. Attachment of retinal cells to the 170,000-mol-wt protein is inhibited by the C1H3 monoclonal antibody and by heparan sulfate (Cole, G. J., D. Schubert, and L. Glaser, 1985, J. Cell Biol., 100:1192-1199). In the present study we have demonstrated that the 170,000-mol-wt C1H3 polypeptide is immunologically identical to the neural cell adhesion molecule N-CAM, and that the 170,000-mol-wt component of N-CAM is preferentially secreted by cells as a component of adherons. We have identified a monoclonal antibody, designated B1A3, that inhibits heparin binding to N-CAM and cell-to-substratum adhesion. A 25,000-mol-wt heparin (heparan sulfate)-binding domain of N-CAM has been identified by limited proteolysis, and this fragment promotes cell attachment when bound to glass surfaces. The fragment also partially inhibits cell binding to adherons when bound to retinal cells, and the B1A3 monoclonal antibody inhibits retinal cell attachment to substrata composed of intact N-CAM or the heparin-binding domain. These data are the first evidence that N-CAM is a multifunctional protein that contains both cell-and heparin (heparan sulfate)-binding domains.     

Prelysosomal divergence of transferrin and epidermal growth factor during receptor-mediated endocytosis

The routes followed by epidermal growth factor and transferrin during their endocytosis by human epithelial cells were compared in double-label studies by using density gradient centrifugation of cell homogenates and fluorescence microscopy with intact cells. Gradient centrifugation studies of cells incubated with radioactively labeled epidermal growth factor and transferrin indicated that both ligands initially were associated with a class of vesicles having a density of 1.037 g/mL and then were rapidly transferred to a membrane compartment having a slightly higher density (1.039 g/mL). Subsequently, the two ligands diverged. Epidermal growth factor ultimately was transferred to a membranous compartment containing lysosomal enzymes (density (1.08 g/mL) where it was degraded. Transferrin was released intact from the cells; very little was transferred to lysosomes. Using fluorescently labeled ligands, it was observed that after cells were warmed to 37 degrees C for 5 min, transferrin and epidermal growth factor gave coincident, punctate fluorescent patterns, strongly suggesting they were localized within the same endocytic vesicles. Subsequently, the epidermal growth factor signal was observed in lysosomes whereas the transferrin signal became weaker and diffuse and did not coincide with the punctate epidermal growth factor fluorescence. The time course of the divergence of the radioactive and fluorescent ligands coupled with the previous morphologic studies on the pathway of epidermal growth factor internalization [Willingham, M. C., & Pastan, I. (1982) J. Cell Biol. 94, 207-212] suggests that the sorting process is prelysosomal and possibly Golgi associated.     

Interaction of the 70,000-mol-wt amino-terminal fragment of fibronectin with the matrix-assembly receptor of fibroblasts

Plasma fibronectin binds saturably and reversibly to substrate-attached fibroblasts and is subsequently incorporated into the extracellular matrix (McKeown-Longo, P.J., and D. F. Mosher, 1983, J. Cell Biol., 97:466-472). We examined whether fragments of fibronectin are processed in a similar way. The amino-terminal 70,000-mol-wt catheptic D fragment of fibronectin bound reversibly to cell surfaces with the same affinity as intact fibronectin but did not become incorporated into extracellular matrix. The 70,000-mol-wt fragment blocked binding of intact fibronectin to cell surfaces and incorporation of intact fibronectin into extracellular matrix. Binding of the 70,000-mol-wt fragment to cells was partially abolished by cleavage into 27,000-mol-wt heparin-binding and 40,000-mol-wt gelatin-binding fragments and more completely abolished by reduction and alkylation of disulfide bonds. Binding of the 70,000-mol-wt fragment to cells was not blocked by gelatin or heparin. When coated onto plastic, the 70,000-mol-wt fragment did not mediate attachment and spreading of suspended fibroblasts. Conversely, fibronectin fragments that had attachment and spreading activity did not block binding of exogenous fibronectin to substrate-attached cells. These results indicate that there is a cell binding site in the 70,000-mol-wt fragment that is distinct from the previously described cell attachment site and is required for assembly of exogenous fibronectin into extracellular matrix.     

Isolation and characterization of epinectin, a novel adhesion protein for epithelial cells

A 70,000-mol-wt protein was isolated from A431 carcinoma cell extracellular matrix that promotes cell substratum adhesion of these epidermoid tumor cells. Extracellular matrix was isolated by a modification of a procedure described by Hedman et al. (Hedman, K., M. Kurkinen, K. Alitalo, A. Vaheri, S. Johansson, and M. Hook, 1979 J. Cell Biol., 81:83-91) and Yamada and Weston (Yamada, K., and J. A. Weston, 1974, Proc. Natl. Acad. Sci. USA, 71:3492-3496). Cells were solubilized with 0.5% deoxycholate, 10 mM Tris, 0.9% NaCl, and 1 mM phenylmethylsulfonyl fluoride, pH 8.0. The residual matrix was then removed from the plates with 6 M urea and 1 mM phenylmethylsulfonyl fluoride and phosphate-buffered saline. SDS PAGE gels of the 6 M urea extract showed one major band at 70,000-mol-wt by Coomassie Blue staining. A 70,000-mol-wt isotopically-labeled band could also be extracted from the matrix of cells incubated with [35S]methionine. Because of the presence of this protein on squamous-derived epithelial cells we have called the 70,000-mol-wt molecule epinectin. Indirect immunofluorescence with polyclonal rabbit antibodies against epinectin stained A431 cells pericellularly in dense punctate accumulations and along the plasma membrane. Enzyme-linked immunoassays and gel-transfer immunolocalization studies showed that the extract did not cross-react with antibodies to fibronectin, laminin, serum-spreading factor, epibolin, or keratin. Additionally, antibodies to epinectin did not cross-react with these proteins. Further studies showed that epinectin does not bind to gelatin. Cell-adhesion assay, using radiolabeled A431 carcinoma cells on various adhesion-promoting substrates, showed that epinectin has similar adhesion-promoting capacity as serum-spreading factor, was somewhat less active than fibronectin, but more effective than laminin or epibolin. Epinectin appears to be a unique protein isolated from epidermoid tumor cells that is distinct from other known adhesion proteins.     

Internalization of cationized ferritin by isolated pancreatic acinar cells

The internalization of cationized ferritin (CF) was studied in isolated pancreatic acinar cells in vitro. Horseradish peroxidase (HRP) was used in conjunction with CF to compare internalization of soluble-phase and membrane-bound tracers. The mode of internalization of CF was dependent upon tracer concentration and origin of the plasma membrane (apical vs. lateral-basal). At the lower tracer concentrations (0.19 and 0.38 mg/ml), internalization from the apical cell surface occurred via small vesicles. The tracer then appeared in multivesicular bodies, in tubules, and in irregular membrane-bound structures. After 15 min, CF particles were seen in many small vesicles near the Golgi apparatus, but not in the Golgi saccules. In contrast, at the lateral-basal cell surface the CF particles tended to form clusters. These clusters were more pronounced at higher CF concentrations (0.76 and 1.5 mg/ml) and were associated with elongated cellular processes, which seemed to engulf CF accumulations in a phagocytic manner. Once internalized, CF was found primarily in large irregular structures which appeared to migrate slowly toward the nucleus, reaching a juxtanuclear position after approximately 30 min. CF was observed in lysosomes after 30-45 min and by 90 min most of the CF was confined to large vacuoles and to trimetaphosphatase-positive lysosomes. Similar routes were observed when cells were double-labeled with CF and HRP, where endocytic structures showed co-localization of both tracers. The results of this study indicate the importance of the Golgi region in the intracellular sorting of internalized apical membrane. Furthermore, this work confirms the presence of distinct endocytic pathways at the apical and lateral-basal cell surfaces.     

Receptor-mediated endocytosis in rat liver: purification and enzymic characterization of low density organelles involved in uptake of galactose-exposing proteins

Rat liver organelles involved in receptor-mediated endocytosis were labeled with a conjugate of galactosylated BSA to horseradish peroxidase [( 3H]galBSA-HRP), injected 10 min before sacrifice. These organelles were recovered at low density (1.11-1.13 g/ml) in sucrose gradients (Quintart, J., P. J. Courtoy, J. N. Limet, and P. Baudhuin, 1983, Eur. J. Biochem., 131:105-112). Upon incubation of such low density fractions in 3,3'-diaminobenzidine (DAB) and H2O2 and equilibration in a second sucrose gradient, galBSA-HRP-containing particles selectively shifted towards heavier densities (Courtoy, P. J., J. Quintart, and P. Baudhuin, 1984, J. Cell Biol., 98:870-876, companion paper), resulting in up to 250-to 300-fold purification with respect to the homogenate. The most purified preparations, wherein DAB- stained structures represented approximately 85% of the total volume of particles, contained only trace activities of enzymes usually regarded as markers for other subcellular entities. These minor activities could reflect either contamination or true enzyme association to the ligand- containing structures. Considering the latter hypothesis, at most 1.0% of alkaline phosphodiesterase I and 2.6% of 5'-nucleotidase (markers for plasma membrane), 3.6% of N-acetyl-beta-glucosaminidase (lysosomes), and 6.0% of galactosyltransferase (Golgi complex) from the homogenate would be associated with the whole population of ligand- containing organelles. After DAB cytochemistry on liver fixed 10 min after galBSA-HRP injection, ligand-containing structures accounted for 0.78-0.89% of the fractional volume of the hepatocytes and displayed a membrane area of 2,100 cm2/cm3, compared with 6,700 cm2/cm3 for the pericellular membrane. Altogether, our data support the hypothesis that these ligand-containing organelles are structurally distinct from plasma membrane, lysosomes, and Golgi complex.     

A quantitative analysis of the endocytic pathway in baby hamster kidney cells

A morphological analysis of the compartments of the endocytic pathway in baby hamster kidney (BHK) cells has been made using the fluid-phase marker horseradish peroxidase (HRP). The endocytic structures labeled after increasing times of endocytosis have been identified and their volume and surface densities measured. In the first 2 min of HRP uptake the volume density of the labeled structures increased rapidly and thereafter remained constant for the next 13-18 min. This plateau represents the volume density of endosome organelles and accounts for 0.65% of the cytoplasmic volume (or 6.8 microns 3 per cell). The labeled structures consist of tubular-cisternal elements which are frequently observed in continuity with 300-400 nm vesicles. After 15-20 min of internalization the volume density of HRP-labeled structures again increased rapidly and reached a second plateau between 30 and 60 min of labeling. This second increase corresponded to detectable levels of HRP reaching later, acid phosphatase (AcPase)-reactive compartments. These structures, comprising the prelysosomes and lysosomes, were mostly vesicular and collectively accounted for 3.5% of the cytoplasmic volume (or 37 microns 3 per cell). The absolute peripheral surface areas of the two classes of organelles (endosomes and prelysosomes/lysosomes) were estimated to be 430 and 370 microns 2 per cell, respectively. The volume of fluid internalized in the first 2 min of uptake was five- to sevenfold less than the volume of the compartment labeled in this time. To account for these results we propose that, after uptake from the cell surface, HRP is delivered to, and diluted in, endosomes that are preexisting organelles initially devoid of the marker. With increasing times of endocytosis the concentration of HRP in the early endosomes increases, as more of the marker enters this compartment. An elevation in HRP concentration in endosomes during the early time points was shown directly using anti- HRP antibodies and colloidal gold on cryosections. The stereological values given in the present study, in combination with earlier studies, provide a minimum estimate for both the total surface area of membranes and the rate of membrane synthesis in a BHK cell.     

Evidence that globular Golgi clusters in mitotic HeLa cells are clustered tubular endosomes.

By conventional electron microscopy we observed in mitotic HeLa cells the structures termed Golgi clusters by Lucocq et al. (J. Cell Biol. 104, 865-874 (1987)) and interpreted by them as clusters of vesicular remnants of the Golgi apparatus. Golgi clusters consist of tubular and vesicular profiles about 50 nm in diameter, sometimes associated with larger 250 nm vesicles. When cultures of HeLa cells were incubated for 60 min or 120 min with medium containing high specific activity horseradish peroxidase (HRP) at 10 mg/ml we found that the membrane-bound compartments in the Golgi clusters in mitotic cells contained heavy deposits of HRP reaction product. Neither interphase nor mitotic HeLa cells contain an endogenous peroxidase activity. We concluded that Golgi clusters are an endocytic compartment and confirmed this by showing that Golgi clusters could be labeled with two other endocytic tracers--bovine serum albumin conjugated to colloidal gold and transferrin conjugated to HRP. When cultures were incubated with HRP for only 15 min most of the Golgi clusters in the mitotic cells were either unlabeled or consisted of a mixture of HRP-labeled and unlabeled profiles. Since during mitosis endocytosis is inhibited this was the expected result. When interphase HeLa cells were incubated with Brefeldin A (BFA), the Golgi apparatus disassembled and immunofluorescence microscopy showed that 1,4 beta galactosyltransferase had relocated to the endoplasmic reticulum. When cells in the presence of BFA and lacking the Golgi apparatus were allowed to endocytose HRP and then entered mitosis, typical HRP-labeled Golgi clusters were seen in the mitotic cells. It is therefore highly unlikely that these structures contain membrane derived from the Golgi cisternae that are sensitive to BFA, including in HeLa cells those containing galactosyltransferase. Finally, we found that interphase HeLa cells incubated with okadaic acid contain structures that are morphologically indistinguishable from Golgi clusters but can be labeled by endocytic tracer. Taken together, this evidence indicates that most, if not all, of the membrane-bound compartments in Golgi clusters are tubular early endosomes.