Min Of Internalization Research Articles

Constitutive Internalization of G Protein-coupled Receptors and G Proteins via Clathrin-independent Endocytosis

Although agonist-dependent endocytosis of G protein-coupled receptors (GPCRs) as a means to modulate receptor signaling has been widely studied, the constitutive endocytosis of GPCRs has received little attention. Here we show that two prototypical class I GPCRs, the beta2 adrenergic and M3 muscarinic receptors, enter cells constitutively by clathrin-independent endocytosis and colocalize with markers of this endosomal pathway on recycling tubular endosomes, indicating that these receptors can subsequently recycle back to the plasma membrane (PM). This constitutive endocytosis of these receptors was not blocked by antagonists, indicating that receptor signaling was not required. Interestingly, the G proteins that these receptors couple to, Galpha(s) and Galpha(q), localized together with their receptors at the plasma membrane and on tubular recycling endosomes. Upon agonist stimulation, Galpha(s) and Galpha(q) remained associated with the PM and these endosomal membranes, whereas beta2 and M3 receptors now entered cells via clathrin-dependent endocytosis. Deletion of the third intracellular loop (i3 loop), which is thought to play a role in agonist-dependent endocytosis of the M3 receptor, had no effect on the constitutive internalization of the receptor. Surprisingly, with agonist, the mutated M3 receptor still internalized and accumulated in cells but through clathrin-independent and not clathrin-dependent endocytosis. These findings demonstrate that GPCRs are versatile PM proteins that can utilize different mechanisms of internalization depending upon ligand activation.

Topology of the signal transduction of the G protein-coupled somatostatin receptor sst2 in human glioma cells.

By a dual approach, using electron microscopy and biochemical techniques, we investigated the topology of the somatostatin receptor sst2 with its inhibitory G protein Gialpha after ligand-induced stimulation and internalization in human glioma cells. On intact cells, the sst2 was labeled at 8 degrees C by an antibody directed to its extracellular sequence followed by a 15-nm gold-labeled secondary antibody. In the presence of the ligand, internalization was induced by exposure to 37 degrees C for 5-10 min. Then, cells were either fixed for immunoelectron-microscopic analysis or homogenized for density gradient separation. After post-embedding staining of the sst2-labeled sections with anti-Gialpha1- 3 or anti-caveolin, a co-localization of sst2, Gialpha and caveolin was detected in endosomal vesicles after 5 min of internalization, but not after 10 min. Furthermore, the gold-labeled organelles containing the internalised receptor were separated from the non-labeled ones on sucrose gradients (density shift separation) and analyzed by Western blotting. Also here, in fractions with higher densities, sst2 could be costained with Gialpha and caveolin after 5 min. From these congruent results from both methods, it can be concluded that, in human glioma cells, the receptor sst2 (1) is internalised in caveolin-positive vesicles and (2) is neighboured to its Gialpha proteins at the plasma membrane and early endosomes.

Acidification-dependent dissociation of endocytosed insulin precedes that of endocytosed proteins bearing the mannose 6-phosphate recognition marker.

A key step in the sorting of endocytosed ligands from their receptors is dissociation, which is triggered by the acidic pH of endosomes. To determine whether dissociation occurs synchronously for all ligands, we compared in Chinese hamster ovary cells the intracellular dissociation of insulin, which dissociates between pH 6.3 and 7.0, with that of lysosomal hydrolases bearing the mannose 6-phosphate recognition marker (Man-6-P proteins), which dissociate around pH 5.8. Chinese hamster ovary cells were pulsed for 2 min with 125I-insulin, acid-washed to remove surface binding, and chased. During a 40-min period, about 50% of the internalized 125I-insulin was released intact via a retrocytotic pathway. Retrocytosis was not inhibited by monensin, suggesting that the release was not dependent on acidic endosomes. The remaining insulin dissociated from its receptor in an acidification-sensitive manner and was eventually degraded. Dissociation was 70% complete within 5 min of internalization. When cells were similarly incubated with 125I-Man-6-P proteins, about 35% of the internalized radioactivity was released during a 1-h chase, reflecting proteolytic maturation of the Man-6-P proteins. Dissociation of Man-6-P proteins was acidification-dependent (i.e. inhibited by monensin), and was 50% complete after about 11 min. The results indicate that acidification-dependent dissociation of ligands does not occur in a single step and suggest that multiple endocytic compartments are involved in receptor/ligand sorting.

Endocytosis in filter-grown Madin-Darby canine kidney cells.

In this paper, we have characterized the apical and basolateral endocytic pathways of epithelial MDCK cells grown on filters. The three-dimensional organization of the endocytic compartments was analyzed by confocal microscopy after internalization of a fluorescent fluid-phase marker from either side of the cell layer. After 5 min of internalization, distinct sets of apical and basolateral early endosomes were observed lining the plasma membrane domain from which internalization had occurred. At later time points, the apical and the basolateral endocytic pathways were shown to converge in the perinuclear region. Mixing of two different fluorescent markers could be detected after their simultaneous internalization from opposite sides of the cell layer. The extent of the meeting was quantitated by measuring the amount of complex formed intracellularly between avidin internalized from the apical side and biotinylated horseradish peroxidase (HRP) from the basolateral side. After 15 min, 14% of the avidin marker was complexed with the biotinylated HRP and this value increased to 50% during a subsequent chase of 60 min in avidin-free medium. We also determined the kinetics of fluid internalization, recycling, transcytosis, and intracellular retention using HRP as a marker. Fluid was internalized with the same rates from either surface domain (1.2 x 10(-4) microns 3/min per microns 2 of surface area). However, significant differences were observed for each pathway in the amounts and kinetics of marker recycled and transcytosed. The content of apical early endosomes was primarily recycled and transcytosed (45% along Bach route after 1 h internalization), whereas delivery to late endocytic compartments was favored from the basolateral early endosome (77% after 1 h). Our results demonstrate that early apical and basolateral endosomes are functionally and topologically distinct, but that the endocytic pathways converge at later stages in the perinuclear region of the cell.

Reconstitution of Endosomal Proteolysis in a Cell-free System: Transfer of immune complexes internalized via Fc receptors to an endosomal proteolytic compartment

The presence of acid proteases in the endosomal compartment of macrophages has been recently demonstrated (Diment, S., Leech, M. S., and Stahl, P. D. (1988) J. Biol. Chem. 263, 6901-6907). This proteolytic activity allows the early degradation of ligands internalized by receptor-mediated endocytosis. To study the early steps that initiate the proteolytic processing of ligands, immune complexes formed with anti-dinitrophenol monoclonal IgG and radiolabeled dinitrophenol-derivatized bovine serum albumin were bound at 4 degrees C to Fc receptors of J774 macrophages. Cells were allowed to internalize immune complexes bound to the plasma membrane for different periods of time at 37 degrees C. Vesicle preparations generated from these cells were incubated in vitro at acidic pH to allow the hydrolysis of ligands located in protease-positive compartments. Ligand hydrolysis was observed after about 5 min of internalization, suggesting that at earlier times immune complexes were located in protease-free vesicles. Upon incubation of cell lysates under conditions that support in vitro endosome-endosome fusion, early protease-free endosomes containing ligand acquire proteolytic activity. Reconstitution of fusion-dependent proteolysis required energy, ions, membrane-associated factors, and cytosol. Cytosol was inactivated by incubation with N-ethylmaleimide. The proteolytic compartment formed upon in vitro incubation colocalized with endosomes in the light region of a Percoll gradient. Reconstitution was also achieved using an endosomal preparation separated from lysosomes in a Percoll gradient. Our results indicate that a fusion step between newly formed endocytic vesicles and a light density, protease-positive compartment triggers the proteolytic processing of ligands internalized by receptor-mediated endocytosis.

Fractionation of endocytic vesicles and glucose-transporter-containing vesicles in rat adipocytes.

We subfractionated intracellular vesicles from rat adipocytes in order to examine the subcellular distribution of endocytic vesicles or endosomes with respect to insulin-regulatable glucose-transporter (GT)-containing vesicles [James, Lederman & Pilch (1987) J. Biol. Chem. 262, 11817-11824]. Vesicles mediating fluid-phase endocytosis sedimented as a single major peak of greater density than the single distinct peak of GT-containing vesicles. This difference was also apparent during cellular insulin exposure and after insulin removal. Endocytosis of insulin and IGF (insulin-like growth factor) II was also examined. In sucrose gradients, IGF II-containing vesicles were less dense than those containing internalized insulin. Receptor-mediated endocytic vesicles were distinct from fluid-phase endocytic vesicles, but overlapped with the GT-containing vesicles. Vesicles containing internalized ligand were further fractionated by agarose-gel electrophoresis after various times of internalization. At least three different vesicle subpopulations containing the iodinated ligands were resolved after 5 min of internalization. Endocytic vesicles containing rapidly internalized insulin (1.5 min at 37 degrees C) consistently co-migrated with GT-containing vesicles. These data indicate that fluid-phase and receptor-mediated endocytosis occur via different pathways in adipocytes. Furthermore, whereas the intracellular GT-containing vesicles are distinct from fluid-phase vesicles, a rapidly labelled pool of insulin-containing vesicles consistently co-fractionated with GT-containing vesicles when separation techniques based on size, density and charge were used. This suggests that the insulin receptor may directly interact with the intracellular GT-containing vesicles after insulin-induced endocytosis.

Resolution of multiple endosomal compartments associated with the internalization of epidermal growth factor and transferrin.

Morphological studies have indicated divergent pathways for the endocytosis of epidermal growth factor (EGF) and transferrin (Tf). In order to obtain biochemical evidence for the pathways associated with the endocytosis of EGF and Tf, a series of Percoll density gradients were employed to separate individual cellular components. Subcellular fractionation of murine fibroblasts exposed to a 2-min pulse of either 125I-Tf or 125I-EGF results in the detection of a total of six cellular compartments related to the internalization process of these ligands. The results of kinetic analysis of the entry of EGF into five membranous fractions is consistent with a model in which ligand is transferred sequentially from the plasma membrane through three distinct prelysosomal environments prior to reaching secondary lysosomes. Each prelysosomal compartment exhibits distinct density and temporal properties in a Percoll density gradient and may represent preexisting endocytic vesicles and/or specific domains of a continuous tubular structure, vesicularized during the process of cell disruption. In addition, the observed differential migration on Percoll density gradients of Tf and EGF containing compartments indicates that the majority of cell bound Tf segregates from EGF and enters a compartment lacking EGF within 5 min of internalization.

Macrophage endosomes contain proteases which degrade endocytosed protein ligands.

Rabbit alveolar macrophages rapidly internalize and degrade mannosylated bovine serum albumin (125I-mannose-BSA). Trichloroacetic acid-soluble degradation products appear in the cells as early as 6 min after uptake at 37 degrees C, and in the extracellular medium after 10 min. Incubation of endocytic vesicles containing this ligand in isotonic buffers at pH 7.4 + ATP resulted in intravesicular proteolysis, which was inhibited by monensin, nigericin, or ammonium chloride. At pH 5.0, degradation proceeded rapidly and was abolished by lysis of the vesicles with 0.1% Triton X-100. Readdition of lysosomes to the incubation mixture did not increase the rate of prelysosomal degradation. Proteolysis of 125I-mannose-BSA was optimal at pH 4.5, and inhibited by low concentrations of the cathepsin D inhibitor pepstatin A. After subcellular fractionation of the macrophages on Percoll gradients, 125I-mannose-BSA sedimented with prelysosomal vesicles and was not transported to secondary lysosomes. Addition of pepstatin A to extracellular medium during internalization of prebound 125I-mannose-BSA partially inhibited degradation of ligand, and resulted in transfer of undegraded 125I-mannose-BSA to lysosomes after 20 min. Using 125I-bovine serum albumin as a substrate for the protease in the presence of 0.1% Triton X-100, we have shown that as much as 36% of the total pepstatin A-sensitive activity sediments with nonlysosomal membranes. After intraendosomal iodination using lactoperoxidase, a labeled protease was isolated by affinity chromatography on pepstatin-agarose. The labeled protease, which had a subunit size of 46 kDa, was detected in endocytic vesicles after 5 min of internalization. These results suggest that a cathepsin D-like protease is responsible for the degradation of 125I-mannose-BSA in macrophages, and that this ligand is degraded in a prelysosomal vesicle.

Complete inhibition of transferrin recycling by monensin in K562 cells.

Monensin blocks human transferrin recycling in a dose-dependent and reversible manner in K562 cells, reaching 100% inhibition at a noncytocidal dose of 10(-5) M, whereas transferrin recycling is virtually unaffected by noncytocidal doses of chloroquine. The intracellular pathway of human transferrin in K562 cells, both in the presence and absence of 10(-5) M monensin, was localized by indirect immunofluorescence. Monensin blocks transferrin recycling by causing internalized ligand to accumulate in the perinuclear region of the cell. The effect of 10(-5) M monensin on human transferrin kinetics was quantitatively measured by radioimmunoassay and showed a positive correlation with immunofluorescent studies. Immunoelectron microscopic localization of human transferrin as it cycles through K562 cells reveals the appearance of perinuclear transferrin-positive multivesicular bodies within 3 min of internalization, with subsequent exocytic delivery of the ligand to the cell surface via transferrin-staining vesicles arising from these perinuclear structures within 5 min of internalization. Inhibition of ligand recycling with 10(-5) M monensin causes dilated transferrin-positive multivesicular bodies to accumulate within the cell with no evidence of recycling vesicles. A coordinated interaction between multivesicular bodies and the Golgi apparatus appears to be involved in the recycling of transferrin in K562 cells. Cell-surface-binding sites for transferrin were reduced by 50% with 10(-5) M monensin treatment; however, this effect was not attenuated by 80% protein synthesis inhibition with cycloheximide, supporting the idea that the transferrin receptor is also recycled through the Golgi.