Insulin binding and processing by H4IIEC3 hepatoma cells: ultrastructural and biochemical evidence for a unique route of internalization and processing

Biochemical and ultrastructural studies of insulin binding and cellular processing by cultured H4IIEC3 hepatoma cells were performed. Insulin binding and intracellular accumulation were rapid and after 30 min at 37 degrees C, 65% of the total cell-associated 125I-insulin was in an acid-stable compartment. Chloroquine had no significant effect on the amount of total cell-associated insulin or the percentage of insulin in the acid-stable compartment or cell-associated insulin degradation under those conditions, but after 60-min incubations, it slightly decreased the rate of dissociation of internalized hormone. Ultrastructural analysis revealed that monomeric ferritin-insulin (Fm-I) initially bound to single or paired receptors on microvilli. Within 5 min occupied insulin receptors microaggregated and migrated to the intervillous cell surface. During the next 5-10 min occupied receptors aggregated into large clusters on the plasma membrane. Large amounts of insulin were internalized by macropinocytosis and the majority of internalized Fm-I was found in phagosomes. Less than 10% of the membrane-bound insulin was associated with pinocytotic invaginations or coated pits and less than 5% of the total cell-associated insulin was found in lysosomes. Chloroquine had no detectable effect on the amount of Fm-I or its distribution among the intracellular organelles. These studies demonstrated that, compared to previous studies with rat adipocytes or 3T3-L1 adipocytes, insulin interalization and intracellular processing in this hepatoma cell were unique. These differences provide further evidence that insulin binding and processing may be controlled by cell-specific mechanisms and that substantial heterogeneity exists in pathways previously presumed to be similar for all cell types.     

Insulin and alpha 2-macroglobulin-methylamine undergo endocytosis by different mechanisms in rat adipocytes: I. Comparison of cell surface events

This ultrastructural study compared the endocytosis of a peptide hormone, ferritin-labeled insulin (Fm-I) or gold-labeled insulin (Au-I), and a non-hormonal ligand, gold-labeled alpha-2-macroglobulin-methylamine (Au-alpha 2MGMA), by rat adipocytes. Quantitative analysis of the cell surface showed that coated pits occupied 0.4% of the adipocyte surface. This was one fifth to one tenth of that which has been reported on fibroblasts and hepatocytes, cell types in which receptor-mediated endocytosis has been extensively studied. In contrast, uncoated micropinocytotic invaginations were quite numerous and occupied 13.1% of the adipocyte cell surface. The frequency of micropinocytotic invaginations, 13.8 per micron 2 of plasma membrane, was 7-12 times greater than has been reported on fibroblasts. Therefore, the ultrastructure of the endocytic apparatus on rat adipocytes was different from more commonly studied cell types. At 4 degrees C, Au-alpha 2MGMA concentrated within coated pits to a density that was 52 times greater than that on the uncoated plasma membrane. Au-alpha 2MGMA was excluded from micropinocytotic invaginations by more than 93%; this exclusion was unrelated to the size of the Au-alpha 2MGMA particle. In contrast, at 4 degrees C, Fm-I did not concentrate within coated pits and occupied micropinocytotic invaginations in a random manner. At 37 degrees C, coated pits accounted for all of the endocytosis of Au-alpha 2MGMA, proving that these structures were functional despite their atypically low density. In contrast, greater than 99% of the endocytosis of Fm-I or Au-I occurred through micropinocytotic invaginations. These results demonstrated for the first time by a comparative, quantitative, ultrastructural method that insulin and Au-alpha 2MGMA undergo endocytosis by dissimilar mechanisms on rat adipocytes. Dissimilarities in the endocytosis of insulin and Au-alpha 2MGMA may be related to the different biological roles of these two molecules.     

Quantitative ultrastructural analysis of receptor-mediated insulin uptake into adipocytes

Monomeric ferritin-insulin was used as an ultrastructural marker to determine by quantitative electron microscopy the time course and route of insulin uptake in rat adipocytes. To approximate steady state membrane binding conditions prior to any internalization, adipocytes were prefixed with glutaraldehyde and incubated for 30 min with 70 nM monomeric ferritin-insulin. Electron micrographs of these cells showed that the ferritin-insulin particles were predominantly in small groups of receptor sites on the plasma membrane and in pinocytotic-like invaginations of the plasma membrane. Significant amounts of ferritin-insulin were observed in cytoplasmic vesicles of unfixed cells as early as 2 min and in multivesicular bodies and lysosome-like structures within 5 to 10 min after the addition of the ligand. Ferritin-insulin accumulation reached steady state levels in the cytoplasmic vesicles in 5 to 10 min and in the lysosome-like structures in 15 min. Little ferritin-insulin was bound to coated pits, and the relative paucity of coated pits found in adipocytes suggested that these specialized endocytotic structures have a relatively insignificant role in insulin uptake in fat cells. Quantitative analysis of the uptake process suggested that a proportion of the insulin internalized by the cell may not be transported to lysosomes, but may be recycled along with the insulin receptor to the plasma membrane.     

Surface interactions and intracellular localization of cationic ferritin: similarity to 125I-insulin in 3T3-L1 adipocytes

We previously reported that in 3T3-L1 adipocytes 125I-insulin associates preferentially with microvilli and coated pits at low temperatures and early times of incubation. At higher temperatures it is internalized through a series of membrane limited intracellular compartments. In the present study, we used a high resolution probe, cationic ferritin (CF), to track adsorptive endocytosis in the 3T3-L1 adipocyte. We find that CF initially associates with coated pits at 2 min of incubation at 37 degrees C. With further incubation at 37 degrees C CF is internalized and after 2 to 10 min of incubation is predominantly localized to coated and non-coated clear vesicles. Approximately 50% of the apparent coated vesicles seen near the plasma membrane on single thin sections are shown by serial sectioning to be true vesicles (i.e., without a surface connection). At later time points CF is localized predominantly to lysosomal structures and, to a much smaller extent, Golgi-related structures. The remarkable similarity between 125I-insulin and CF with respect to post-binding processing suggests that while the membrane receptor confers the initial specificity, post-binding events are common for different types of ligands after they bind to cell surfaces and are subject to adsorptive endocytosis.     

Opposite Effects of Insulin on Focal Adhesion Proteins in 3T3-L1 Adipocytes and in Cells Overexpressing the Insulin Receptor

Insulin can regulate the abundance and organization of filamentous actin within cells in culture. Early studies using cell lines that overexpress the insulin receptor demonstrated that insulin caused a rapid reversible disassembly of actin filaments that coincided with the rapid tyrosine dephosphorylation of focal adhesion kinase. We have extended these studies by demonstrating that paxillin, another focal adhesion protein, and Src undergo tyrosine dephosphorylation in response to insulin in Chinese hamster ovary (CHO) and rat hepatoma (HTC) cells that overexpress the insulin receptor. This contrasted with the effect of insulin in parental CHO and HTC cells in which focal adhesion proteins were not dephosphorylated in response to the hormone. In addition, insulin caused a dispersion of focal adhesion proteins and disruption of actin filament bundles only in cells that overexpressed the insulin receptor. Moreover, in 3T3-L1 adipocytes, which are considered prototypic insulin-responsive cells, actin filament assembly was stimulated, and focal adhesion protein tyrosine phosphorylation was not altered. 3T3-L1 cells have more insulin receptors than either parental CHO or HTC cells but have fivefold less insulin receptors than the overexpressing cell lines. We hypothesize that a threshold may exist in which the overexpression of insulin receptors determines how insulin signaling pathways regulate the actin cytoskeleton.     

Tissue-specific sorting of the human LDL receptor in polarized epithelia of transgenic mice

The distribution of human low density lipoprotein (LDL) receptors was studied by immunofluorescence and immunoelectron microscopy in epithelial cells of transgenic mice that express high levels of receptors under control of the metallothionein-I promoter. In hepatocytes and intestinal epithelial cells, the receptors were confined to the basal and basolateral surfaces, respectively. Very few LDL receptors were present in coated pits or intracellular vesicles. In striking contrast, in the epithelium of the renal tubule the receptors were present on the apical (lumenal) surface where they appeared to be concentrated at the base of microvilli and were abundant in vesicles of the endocytic recycling pathway. Intravenously administered LDL colloidal gold conjugates bound to the receptors on hepatocyte microvilli and were slowly internalized, apparently through slow migration into coated pits. We conclude that (a) sorting of LDL receptors to the surface of different epithelial cells varies with each tissue; and (b) in addition to a signal for clustering in coated pits, the LDL receptor may contain a signal for retention in noncoated membrane that is manifest in hepatocytes and intestinal epithelial cells, but not in renal epithelial cells or cultured human fibroblasts.     

Tyrosine kinase-defective insulin receptors undergo insulin-induced microaggregation but do not concentrate in coated pits

Biologically active colloid-gold complexes were used to compare ligand-induced microaggregation, redistribution, and internalization of insulin receptors on Rat 1 fibroblasts expressing wild type (HIRc) or tyrosine kinase-defective (HIR A/K1018) human insulin receptors. Insulin-like growth factor I (IGF I) and alpha 2-macroglobulin receptors also were compared. On both cell types, all four unoccupied receptor types occurred predominantly as single receptors. Ligand binding caused receptor microaggregation. Microaggregation of wild type or kinase-defective insulin receptors or IGF I receptors was not different. alpha 2-Macroglobulin receptors formed larger microaggregates. Compared to wild type insulin or IGF I receptors, accumulation of kinase-defective insulin receptor microaggregates in endocytic structures was decreased, and the size of microaggregates in coated pits was significantly smaller. As a result, receptor-mediated internalization of gold-insulin by HIR A/K1018 cells was less than 6% of the cell-associated particles compared to approximately 60% of the particles in HIRc cells. On HIR A/K1018 cells, alpha 2-macroglobulin and IGF I were internalized via coated pits demonstrating that those structures were functional. These results suggest that: 1) ATP binding, receptor autophosphorylation, and activation of receptor kinase activity are not required for receptor microaggregation; 2) receptor microaggregation per se is not sufficient to cause ligand-induced receptor-mediated internalization or the biological effects of insulin; and 3) autophosphorylation of the beta-subunit or activation of the receptor kinase activity is required for the insulin-induced concentration of occupied receptors in coated pits.     

Localization of the insulin receptor in caveolae of adipocyte plasma membrane.

The insulin receptor is a transmembrane protein of the plasma membrane, where it recognizes extracellular insulin and transmits signals into the cellular signaling network. We report that insulin receptors are localized and signal in caveolae microdomains of adipocyte plasma membrane. Immunogold electron microscopy and immunofluorescence microscopy show that insulin receptors are restricted to caveolae and are colocalized with caveolin over the plasma membrane. Insulin receptor was enriched in a caveolae-enriched fraction of plasma membrane. By extraction with beta-cyclodextrin or destruction with cholesterol oxidase, cholesterol reduction attenuated insulin receptor signaling to protein phosphorylation or glucose transport. Insulin signaling was regained by spontaneous recovery or by exogenous replenishment of cholesterol. beta-Cyclodextrin treatment caused a nearly complete annihilation of caveolae invaginations as examined by electron microscopy. This suggests that the receptor is dependent on the caveolae environment for signaling. Insulin stimulation of cells prior to isolation of caveolae or insulin stimulation of the isolated caveolae fraction increased tyrosine phosphorylation of the insulin receptor in caveolae, demonstrating that insulin receptors in caveolae are functional. Our results indicate that insulin receptors are localized to caveolae in the plasma membrane of adipocytes, are signaling in caveolae, and are dependent on caveolae for signaling.     

Localization of transferrin receptors and insulin-like growth factor II receptors in vesicles from 3T3-L1 adipocytes that contain intracellular glucose transporters

Transferrin receptors in detergent extracts of subcellular membrane fractions prepared from 3T3-L1 adipocytes were measured by a binding assay. There was a small but significant increase (1.2-fold) in the amount of receptor in a crude plasma membrane fraction and a 40% decrease in the number of transferrin receptors in microsomal membranes prepared from insulin-treated cells, when compared with corresponding fractions from control cells. Intracellular vesicles containing insulin-responsive glucose transporters (GT) have been isolated by immunoadsorption from the microsomal fraction (Biber, J. W., and G. E. Lienhard. 1986. J. Biol. Chem. 261:16180-16184). All of the transferrin receptors in this fraction were localized in these vesicles; however, because the GT vesicles contain approximately 30-fold fewer transferrin receptors than GT, on the average only one vesicle in three contains a transferrin receptor. The binding of 125I-pentamannose 6-phosphate BSA to 3T3-L1 adipocytes at 4 degrees C was used to monitor surface insulin-like growth factor II (IGF-II)/mannose 6-phosphate receptors. Exposure of cells to insulin at 37 degrees C for 5 min resulted in a 2.5-4.5-fold increase in surface receptors. There was a corresponding 20% decrease in the amount of IGF-II receptors in the microsomal membranes prepared from insulin-treated cells, as assayed by immunoblotting. Moreover, the IGF-II receptors and GT were located in the same intracellular vesicles, since antibodies to the carboxyterminal peptide of either protein immunoadsorbed vesicles containing 70-95% of both proteins initially present in the microsomal fraction. In conjunction with other studies, these results indicate that in 3T3-L1 adipocytes, three membrane proteins (the GT, the transferrin receptor, and the IGF-II receptor) respond similarly to insulin, by redistributing to the surface from intracellular compartment(s) in which they are colocalized.     

The effects of chronic ethanol administration on the rates of internalization of various ligands during hepatic endocytosis.

The purpose of the present study was to further characterize the ethanol-induced impairments in hepatic endocytosis. Specifically, we examined the effects of ethanol treatment on receptor-ligand internalization via the coated and noncoated pit pathways. Insulin, epidermal growth factor (EGF) and asialoorosomucoid (ASOR) were used as model ligands to study internalization by isolated hepatocytes. ASOR and EGF are thought to be internalized strictly in coated pit regions of the cell membrane, while insulin may be internalized in both coated and uncoated membrane regions. Ethanol administration for 5-7 weeks decreased internalization of ASOR and EGF while internalization of insulin was unchanged during a single round of endocytosis of surface-bound ligand. Similarly, a more quantitative measure of endocytosis, the endocytic rate constant, was decreased for EGF and ASOR but not for insulin in livers of experimental rats. When endocytosis of Lucifer yellow, a fluorescent dye known to be internalized in the cell by fluid-phase endocytosis was examined, the initial rates of dye uptake were not significantly altered by alcohol administration. These results indicate that ethanol may selectively impair internalization occurring by coated pits while it has a minimal effect on initial uptake of molecules which are internalized by noncoated membrane regions.     

An analysis of the relationship between the cellular distribution and the rate of turnover for the separate classes of unoccupied, noncovalently occupied, and covalently occupied insulin receptor

To further investigate insulin's role in regulating the turnover of insulin receptor during down-regulation in 3T3-L1 adipocytes, the relationship between the cellular distribution and turnover of unoccupied, noncovalently occupied, and covalently occupied receptor was examined. At steady-state 12% of the unoccupied receptors and 46% of covalently occupied receptors are intracellular. The apparent first-order rate constant (Kapp) for turnover of the total pool of covalently occupied receptors (0.16 h-1) is 3.8-fold higher than that for unoccupied receptors (0.042 h-1). When unlabeled insulin is added, identical values for both Kapp (0.10 h-1) and distribution (26% internal) are measured for noncovalently and covalently occupied receptors. The rate constant (Kdeg), describing the relative sensitivity of internalized receptor to degradation, is identical (0.36-0.41 h-1) for unoccupied, noncovalently occupied, and permanently occupied pools of internal receptor. Mechanisms for down-regulation postulating: (a) an occupancy-dependent alteration in the conformation of internal receptor increasing receptor sensitivity to internal proteases, (b) a preferential sorting of internal occupied receptor to degradative pathways, or (c) induction of intracellular proteases by insulin, would all reflect a substantial change in Kdeg for occupied receptor and thus are unlikely mechanisms by which insulin increases the rate of receptor turnover. The turnover of insulin receptor in 3T3-L1 adipocytes is regulated primarily by its intracellular concentration and not by the state of occupancy of internalized receptor.     

Insulin elicits a redistribution of transferrin receptors in 3T3-L1 adipocytes through an increase in the rate constant for receptor externalization

Incubation of 3T3-L1 adipocytes with insulin at 37 degrees C resulted in a 2-fold increase in specific binding of transferrin to cell-surface receptors, as measured by a subsequent incubation of cells at 4 degrees C with 125I-transferrin. The insulin concentration required for half-maximal effect was 10 nM, and the half-time for insulin action was 40 s. By comparison, insulin stimulated hexose transport in 3T3-L1 adipocytes with a half-maximal effect at 8 nM and a half-time of 105 s. Scatchard analysis of 125I-transferrin binding to cells at 4 degrees C showed that the insulin-induced increase in transferrin receptor binding was due to an increase in the number of surface transferrin receptors. When cells were incubated for 2 h at 37 degrees C with 125I-transferrin to achieve steady-state binding and then exposed to insulin, there was a 1.7-fold increase in surface-bound transferrin (acid-sensitive) and a corresponding decrease in intracellularly bound transferrin (acid-insensitive). Thus, insulin elicits translocation of intracellular transferrin receptors to the plasma membrane. Concomitant with the 2-fold increase in surface receptors in response to insulin, there was a 2-fold increase in the rate of 59Fe3+ uptake from 59Fe3+-loaded transferrin. The rate of externalization of the intracellular 125I-transferrin-receptor complex at 37 degrees C was determined for basal and insulin-treated cells. Insulin increased the first-order rate constant for this process 1.7-fold. The effect of insulin on the rate of externalization is sufficient to account for the increase in surface transferrin receptors.     

Hepatic receptor for asialo-glycoproteins. Ultrastructural demonstration of ligand-induced microaggregation of receptors

Endocytosis of asialo-glycoproteins in hepatocytes is mediated by a lectin-like receptor with specificity for D-galactose. Early events of receptor-ligand interactions have been studied by ultrastructural analysis. Hepatocytes were isolated from the rat liver by collagenase perfusion and incubated with a galactosylated electron dense marker (gold-Gal-BSA, glactosylated bovine serum albumin adsorbed onto colloidal gold particles). Initial binding of gold-Gal-BSA particles occurs to receptors diffusely distributed at hepatic microvilli of the former space of Disé. No lectin activity was found in membrane areas that had formed in situ the region of hepatic cell contact or bile canaliculi. Microaggregation of receptor-ligand complexes is seen as an early consequence of particle binding. Microaggregates contain 2-5 particles and are located outside coated pits. After prolonged incubation larger clusters are formed, these are found associated with coated membrane areas. It is concluded that at least three steps precede the uptake of galactosylated proteins by hepatocytes. These are: (i) binding of ligand at diffusely distributed binding sites; (ii) local microaggregation of receptor-ligand complexes; (iii) formation of larger clusters and association with coated pits.     

Colloidal gold-labeled insulin complex. Characterization and binding to adipocytes

Biologically active insulin gold complex was used as an ultrastructural marker to study insulin binding sites, uptake, and internalization in isolated rat adipocytes. The preparation conditions for monodispersed particles, ca. 16 nm in diameter and loaded with approximately 100 insulin molecules, are reported. The complex is stable for at least six weeks. Single particles or small clusters were scattered across the cell membrane. The distribution of unbound receptors seemed to be independent of the extensive system of pre-existing surface connected vesicles in adipocytes. The uptake of particles took place predominantly via non-coated pinocytotic invaginations; clathrin-coated pits did not seem to be important for this process. Lysosome-like structures contained aggregates of 10-15 particles. These data suggest that insulin gold complex is a useful marker for the specific labeling of insulin binding sites.     

Effect of preferential insertion of LDL receptors near coated pits

Recent experiments suggest that low density lipoprotein (LDL) receptors on human fibroblasts are not inserted into the plasma membrane uniformly, as earlier experiments indicated, but are inserted into specialized regions, called plaques, where coated pits form. If the consequent reduction in the time required for LDL receptors to diffuse to coated pits were significant, this could alter conclusions drawn from previous calculations based on the assumption that LDL receptors are inserted uniformly. In particular, the conclusion could be wrong that diffusion of LDL receptors to coated pits is the rate limiting step in the interaction of cell surface LDL receptors with coated pits. Here we calculate the extent of the reduction in mean travel time of an LDL receptor to a coated pit, as a function of the plaque radius. We find that only if LDL receptor insertion is limited to a very small portion of the plasma membrane near coated pit sites is there a substantial decrease in the average time it would take an LDL receptor to diffuse to a coated pit. In order for preferential insertion of LDL receptors into plaques to cut the mean receptor travel time in half, plaques would have to take up no more than 10% of the cell surface area; to reduce the travel time by a factor of 10 plaques would have to cover only 2% of the cell surface, approximately twice the area covered by coated pits at 37°C.     

Stoichiometric translocation of adipocyte insulin receptors from the cell-surface to the cell-interior. Studies using a novel method to rapidly remove detergent and concentrate soluble receptors

A rapid one-step method was developed for harvesting and concentrating insulin receptors from solubilized adipocytes, which entails precipitating soluble receptors with polyethylene glycol and resuspending the receptor-containing pellet in a reduced volume of binding buffer. With this procedure 90-100% of receptors were recovered, while 80% of cellular protein was removed, thus resulting in a marked reduction of both ligand and receptor proteases and about a 5-fold purification of the receptor. More importantly, greater than 98% of the Triton X-100 detergent was removed during this procedure so that the reduced receptor affinity observed in solubilized extracts (due to detergent) was restored to normal. Reconstituted receptors exhibited normal binding characteristics similar to those observed for plasma membrane receptors. The general utility of our receptor precipitation-reconstitution method is highlighted by studies on insulin-induced translocation of receptors from the cell-surface to the cell-interior of adipocytes and studies on the assessment of the binding affinity of nascent intracellular receptors. The results of these studies are consistent with the following. 1) Insulin initiates endocytotic uptake of insulin receptors, which then recycle back to the cell-surface. 2) Chloroquine impairs the recycling of internalized receptors while preventing receptor degradation, resulting in the progressive trapping and accumulation of receptors within cells during insulin treatment. 3) Receptor translocation during acute insulin-induced down-regulation is stoichiometric in that receptors lost from the cell-surface can be quantitatively recovered within the cell-interior. 4) In the absence of ligand, these receptors within adipocytes are mainly newly synthesized receptors enroute to the cell-surface, and they possess an affinity similar, if not identical, to mature receptors on the plasma membrane.     

Effect of preferential insertion of LDL receptors near coated pits

Recent experiments suggest that low density lipoprotein (LDL) receptors on human fibroblasts are not inserted into the plasma membrane uniformly, as earlier experiments indicated, but are inserted into specialized regions, called plaques, where coated pits form. If the consequent reduction in the time required for LDL receptors to diffuse to coated pits were significant, this could alter conclusions drawn from previous calculations based on the assumption that LDL receptors are inserted uniformly. In particular, the conclusion could be wrong that diffusion of LDL receptors to coated pits is the rate limiting step in the interaction of cell surface LDL receptors with coated pits. Here we calculate the extent of the reduction in mean travel time of an LDL receptor to a coated pit, as a function of the plaque radius. We find that only if LDL receptor insertion is limited to a very small portion of the plasma membrane near coated pit sites is there a substantial decrease in the average time it would take an LDL receptor to diffuse to a coated pit. In order for preferential insertion of LDL receptors into plaques to cut the mean receptor travel time in half, plaques would have to take up no more than 10% of the cell surface area; to reduce the travel time by a factor of 10, plaques would have to cover only 2% of the cell surface, approximately twice the area covered by coated pits at 37 degrees C.     

The antibody-induced clustering and endocytosis of HLA antigens on cultured human fibroblasts

It has previously been shown by immunofluorescence experiments that the cross-linking of HLA antigens into patches (by antibody reagents directed to human β₂-microglobulin) on the surfaces of cultured human fibroblasts leads to the lining up of the patches over the actomyosin-containing stress fibers lying immediately under the surface membrane. These experiments have now been extended to the resolution of the electron microscope by the use of ferritin-conjugated antibody. The results show that a substantial part of the HLA surface clusters that form by 5 min after the addition of the antibody reagents is found in small uncoated surface invaginations which are subsequently endocytosed and ultimately fuse with lysosomal bodies. At no stage in this process is there any indication that coated pits or coated vesicles participate. These and other results suggest, therefore, that there are at least two distinct mechanisms for the ligand-induced endocytosis and lysosomal processing of membrane components, one involving coated pits and the other the noncoated invaginations described in this paper. Transmembrane associations of clusters with intracellular actomyosin-containing structures may have a role in the endocytosis of these noncoated invaginations.     

Relationship between insulin stimulation and endogenous regulation of 2-deoxyglucose uptake in 3T3-L1 adipocytes

The occurrence of the endogenous regulatory response to high rates of 2-deoxyglucose (2-DG) uptake, as previously described for C6 glioma cells during incubation with 2 mM 2-DG (Lange et al.: J. Cell. Physiol., 1989), was studied in 3T3-L1 preadipocytes and adipocytes, and the influence of insulin on this endogenous uptake regulation was examined. In contrast to 3T3-L1 preadipocytes, insulin-sensitive differentiated 3T3-L1 adipocytes displayed the time-dependent cyclic pattern of 2-DG uptake rates characteristic of the membrane-limited and endogenously regulated cellular state of hexose utilization. Although insulin induced a threefold stimulation of 2-DG tracer uptake in adipocytes, the hormone did not additionally stimulate the uptake rates or affect the periodic response: maximum and minimum levels of uptake remained unchanged. Scanning electron microscopy (SEM) revealed that the acquirement of the differentiated state is accompanied by a conspicuous transformation of the smooth surface of undifferentiated 3T3-L1 cells into a surface covered by numerous microvilli of uniform size and appearance. Treatment with insulin (10 mU/ml; 10 minutes) converted these microvilli into voluminous saccular membrane protrusions of the same type as had been formed during incubation of 3T3-L1 adipocytes with 2 mM 2-DG, and which have previously been shown to be involved in the endogenous uptake regulation of C6 glioma cells (Lange et al.: J. Cell. Physiol., 1989). These insulin-induced saccated membrane areas appeared to become integrated into the cell surface. Accordingly, insulin treatment caused a twofold increase of the intracellular distribution space of 3-O-methylglucose (3-OMG) in 3T3-L1 adipocytes. This insulin-induced increase of the 3-OMG distribution space exhibited the same time (t1/2 = 2-2.5 minutes) and dose dependence (EC50 = 20 nM) as the insulin-induced stimulation of 3-OMG transport. Glucose deprivation during the differentiation period inhibited the outgrowth of microvilli from the cell surface. Glucose starvation (18 hours at less than 0.5 mM) induced a conspicuous reduction of the length of microvilli on differentiated 3T3-L1 cells. In this state, the stalks of the microvilli are almost invisible and the enlarged spherical tips of the microvilli (with an average diameter of 370 nm compared to 230 nm of fed cells) appeared to protrude directly out of the cell surface. Starvation-induced shortening of microvilli was accompanied by a threefold increase of the basal 3-OMG transport rate and a greater than twofold increase of the intracellular 3-OMG distribution space as compared to fed cells (10 mM; 18 hours).(ABSTRACT TRUNCATED AT 400 WORDS)     

Intracellular receptor sorting during endocytosis: comparative immunoelectron microscopy of multiple receptors in rat liver

Using double-label quantitative immunoelectron microscopy on ultrathin cryosections of rat liver, we have compared the endocytotic pathways of the receptors for asialoglycoprotein (ASGP-R), mannose-6-phosphate ligands (MP-R), and polymeric IgA (IgA-R). All three were found within the Golgi complex, along the entire plasma membrane, in coated pits and vesicles, and within a compartment of uncoupling of receptors and ligand ( CURL ). The receptors occurred randomly at the cell surface, in coated pits and vesicles. Within CURL tubules ASGP-R and MP-R were colocalized , but IgA-R and ASGP-R displayed dramatic microheterogeneity. Thus, in addition to its role in uncoupling and sorting recycling receptor from ligand, CURL serves as a compartment to segregate recycling receptor (e.g. ASGP-R) from receptor involved in transcytosis (e.g. IgA-R).