Intracellular transport and degradation of chylomicron remnants in rat liver cells after in vivo endocytosis.

The intracellular transport and degradation of in vivo endocytosed chylomicron remnants labelled with 125I in the protein moiety was studied in rat liver cells by means of subcellular fractionation in Nycodenz and sucrose density gradients. Initially, the radioactivity was located in low-density endosomes and was sequentially transferred to light and dense lysosomes. Data from gel filtration of the light and dense lysosomal fractions showed radioactive material with a molecular weight of about 1000-2000, representing short peptide fragments or amino acids which remain attached to iodinated tyramine cellobiose. In addition, undegraded apoproteins accumulated in both types of lysosome. Our data suggest that endocytosed chylomicron remnant apoproteins are first located in low-density endosomes and are sequentially transferred to light and dense lysosomes. Furthermore, the degradation process starts in the light lysosomes.

Endocytosis mediated by the mannose receptor in liver endothelial cells. An immunocytochemical study.

Immunocytochemical labeling of ultrathin cryosections from rat liver showed that mannose-terminated glycoproteins are removed rapidly from the blood stream mainly by the sinusoidal endothelial cells. The mannose-terminated glycoprotein ovalbumin was injected intravenously into rats 1 min, 6 min, and 24 min before perfusion fixation of the liver. Several minor and at least three major subcellular compartments were shown to be involved in the endocytic process. One minute after injection, ovalbumin was found at the cell surface, in coated pits, in coated vesicles, in tubular structures, and bound to the membrane of large early endosomes of which some showed a cisternal structure. After 6 min, ovalbumin was found in the lumen of large electron-lucent late endosomes and after 24 min in electron-dense structures, presumably lysosomes. The early endosomes have an ultrastructure which, together with the labeling pattern, indicates that this compartment has the same function as the CURL identified in parenchymal liver cells. The results are in accordance with recent biochemical findings indicating that ovalbumin endocytosed by endothelial cells is found sequentially in three different subcellular fractions depending on the time between injection and cooling for fractionation (G. M. Kindberg, T. Berg: Intracellular transport of endocytosed mannose terminated glycoproteins in rat liver endothelial cells. In: E. Wisse, D. L. Knook, K. Decker (eds.): Cells of the Hepatic Sinusoid. Vol. 2. pp. 120-124. Kupffer Cell Foundation. Rijswijk The Netherlands 1989).

Distribution of liposome-encapsulated iodixanol in rat liver cells.

Distribution of liposome-encapsulated [(125)I]iodixanol in different types of liver cells following intravenous injection was studied in rats. The data showed that liposome-encapsulated [(125)I]iodixanol was rapidly taken up by the liver; after 15 min, radioactivity corresponding to nearly 25% of the injected radioactivity could be recovered therein. After 4 hr, approximately 60% of the injected radioactivity was in the liver. One week after injection, nearly 30% of the encapsulated radioactivity could still be recovered in the liver. Liposome-encapsulated [(125)I]iodixanol was taken up both by hepatocytes and the Kupffer cells. On a per cell basis, the uptake of liposome-encapsulated [(125)I]iodixanol in Kupffer cells was more than 10-fold greater than that in hepatocytes, while the contribution of liver endothelial cells to uptake was negligible. Osmotic protection studies showed that iodixanol does not readily diffuse across lysosomal membranes, indicating that loss of iodixanol from the liver probably occurred by recycling rather than by diffusion across phagolysosomal and plasma membranes.

Clearance of purified human liver gamma-glutamyltransferase after intravenous injection in the rat.

The clearance of gamma-glutamyltransferase was studied by injecting the purified human liver enzyme intravenously in the rat. The results show a biphasic clearance, with a rapid initial rate of removal. The initial uptake is more rapid for neuraminidase-treated GT. Liver accounts for the bulk organ uptake and the enzyme is almost exclusively taken up into the parenchymal cells. We suggest that the uptake of circulating GT is receptor mediated, most likely by the galactose receptor of the parenchymal cells.

Mangafodipir trisodium injection, a new contrast medium for magnetic resonance imaging: in vitro metabolism and protein binding studies of the active component MnDPDP in human blood.

The binding to human serum proteins of MnDPDP (manganese(II) dipyridoxyl diphosphate), the active component of the magnetic resonance imaging contrast medium mangafodipir trisodium injection (Teslascan) was studied in ultrafiltration experiments. Sera from three males and three females were incubated with 86 microM [14C]MnDPDP for 60 min at room temperature (20-23 degrees C), followed by centrifugation through filters with a cut-off of 30 kDa. Analysis of the filtrates and the initial incubation mixtures for manganese, by ICP-AES, and for DPDP and its dephosphorylated metabolites DPMP (dipyridoxyl monophosphate) and PLED (dipyridoxyl ethylenediamine diacetate) by liquid scintillation counting, showed a clear difference in protein binding of manganese and the ligands under these conditions. Only 2.2 +/- 1.8% (mean +/- S.E.; n = 6) of DPDP, DPMP and PLED were bound to protein, whereas 26.9 +/- 2.9% (mean +/- S.E.; n = 6) of manganese was bound to protein. No binding of DPDP, DPMP or PLED to blood cells was observed when whole blood, containing either heparin or EDTA as anticoagulant, was spiked with [14C]MnDPDP and the cell-free fraction and the lysed cell fraction analysed by liquid scintillation counting. The extent of protein binding of manganese corresponded well with results from an in vitro metabolism study, in which MnDPDP was added to heparinized human whole blood, showing that approximately 25% of DPDP, DPMP or PLED were not bound to manganese. The in vitro metabolism study revealed that transmetallation with zinc was nearly complete within 1 min, and that dephosphorylation is a sequential process going from DPDP to the monophosphate DPMP, and then to the fully dephosphorylated compound PLED.

Uptake and degradation of radioactively labelled albumin microspheres as markers for Kupffer cell phagocytosis.

Rats were injected with liposomes containing iodixanol (CTP10 Injection; 100 mg iodine per kg body weight) followed by a second injection of 125I-tyramine-cellobiose-albumin microspheres. The amounts of phagocytosed and degraded labelled albumin in liver were measured. A reduced uptake and degradation of albumin microspheres was observed when the labelled microspheres were injected 2 h or 24 h after the liposomes compared with that obtained in control animals receiving saline. No effect on the uptake and degradation of labelled microspheres was observed when the time lag between the injection of liposomes and labelled microspheres was 1 week. The data show that the uptake and degradation of 125I-tyramine-cellobiose-albumin microspheres can be used as indicators of Kupffer cell phagocytotic function following drug uptake by these cells.

Lysosomal and endosomal heterogeneity in the liver: a comparison of the intracellular pathways of endocytosis in rat liver cells.

Air-filled albumin microspheres, asialoorosomucoid and formaldehyde-treated serum albumin are selectively taken up by endocytosis in rat liver Kupffer cells, parenchymal cells and endothelial cells, respectively. Intracellular transport and degradation of endocytosed material were studied by subcellular fractionation in sucrose and Nycodenz gradients after intravenous injection of the ligand. By using ligands labeled with 125I-tyramine-cellobiose, the subcellular distribution of labeled degradation products can be studied because they are trapped at the site of formation. The results show that the kinetics of intracellular transport are different in hepatic parenchymal, endothelial and Kupffer cells. In endothelial cells, the ligand is associated with two types of endosomes during the first minutes after internalization and then is transferred rapidly to the lysosomes. In parenchymal cells, 125I-tyramine-cellobiose-asialoorosomucoid was located in a relatively slowly sedimenting vesicle during the first minute after internalization and subsequently in denser endosomes. Degradation of 125I-tyramine-cellobiose-asialoorosomucoid in parenchymal cells started later than that of 125I-tyramine-cellobiose-formaldehyde-treated serum albumin in endothelial cells. Furthermore, the ligand seemed to be transferred relatively slowly from endosomes to lysosomes, and most of the undegraded ligand was in the endosomes. The rate-limiting step of proteolysis in parenchymal cells is probably the transport from endosomes to lysosomes. In Kupffer cells, most 125I-tyramine-cellobiose-microspheres are found as undegraded material in very dense endosomes up to 3 hr after injection. After 20 hr, most of the ligand is degraded in lysosomes distributed at a lower density than the endosomes in Nycodenz and sucrose gradients.

Renal uptake and degradation of trapped-label insulin.

In order to study the kinetics of insulin degradation in the kidneys and liver, insulin was labelled by a trapped-label procedure and injected into rats. In contrast to conventional 125I-insulin, the trapped-label preparation allows quantitative measurements of the extent of degradation in vivo because the final degradation products do not leave the cells. One hour after injection, the amount of radioactivity in the kidneys from a trace dose of trapped-label insulin was 10 times higher that from conventionally labelled insulin; over 80% of the increase was due to low molecular weight degradation products which were retained in the kidneys. The amount of acid-precipitable radioactivity in the blood was the same for both labelled preparations, indicating that their rates of clearance were similar. In the kidney, we detected no degradation products of molecular weight intermediate between intact insulin and the end products of proteolysis. After 2 h, 33% of the injected dose remained in the kidneys and only 13% in the liver. Over 80% of the renal radioactivity was sedimentable in an isotonic density gradient, indicating that intact insulin, as well as degradation products in the cells, were enclosed within membrane-bound vesicles.

Receptor-mediated endocytosis of ovalbumin by two carbohydrate-specific receptors in rat liver cells. The intracellular transport of ovalbumin to lysosomes is faster in liver endothelial cells than in parenchymal cells.

1. The uptake of ovalbumin (OVA) in rat liver parenchymal cells (PC) and non-parenchymal cells was studied in vivo and in vitro in order to compare the cellular expression of glycoprotein receptors and the kinetics of intracellular transport of ligand endocytosed by these receptors. 2. Ovalbumin was labelled with 125I or with 125I-tyramine-cellobiose (125I-TC). By using 125I-TC-OVA the labelled degradation products were trapped in the cells. 3. 125I-TC-OVA was rapidly cleared from blood mainly by receptor-mediated uptake in the liver. At 30 min after injection, 50% of the ligand was recovered in the liver. The endothelial cells (EC) and the PC were the predominant cell types responsible for uptake. 4. The uptake in PC was strongly inhibited by asialo-orosomucoid (AOM), but not by mannan, indicating that the uptake in these cells was mediated by the galactose receptor and not by the mannose receptor. This finding is compatible with the observation that a proportion of the OVA contains terminal galactose residues in the carbohydrate moiety. 5. In vitro uptake of OVA in cultured EC was saturable and inhibited by mannan, mannose, fructose, N-acetylglucosamine, EDTA or monensin, but not by galactose or AOM. The uptake of OVA in these cells was therefore mediated by the mannose receptor. 6. To label the organelles involved in endocytosis in PC and EC, 125I-TC-OVA was injected intravenously together with an excess of either AOM or mannan. In this way the labelled ligand could be directed selectively to EC or PC respectively. Subcellular fractionation of total liver in sucrose and Nycodenz gradients revealed that in EC the intracellular transport of OVA is so fast that endocytosed ligand accumulates and thus increases the density of the lysosomes. Conversely, in PC transfer of ligand is slower, with the result that accumulation of undegraded ligand in the lysosomes does not occur. These findings are interpreted to mean that in EC the rate-limiting step of handling of endocytosed ligand is intralysosomal degradation, whereas in PC the rate-limiting step is transport of ligand to the lysosomes. 7. Altogether, these findings suggest that endocytosis of OVA by the liver EC and PC is mediated by mannose and galactose receptors respectively, and that the kinetics of intracellular transport of OVA differ in the two cell types.

The effect of vanadate on receptor-mediated endocytosis of asialoorosomucoid in rat liver parenchymal cells.

Vanadate is a phosphate analogue that inhibits enzymes involved in phosphate release and transfer reactions (Simons, T. J. B. (1979) Nature 281, 337-338). Since such reactions may play important roles in endocytosis, we studied the effects of vanadate on various steps in receptor-mediated endocytosis of asialoorosomucoid labeled with 125I-tyramine-cellobiose (125I-TC-AOM). The labeled degradation products formed from 125I-TC-AOM are trapped in the lysosomes and may therefore serve as lysosomal markers in subcellular fractionation studies. Vanadate reduced the amount of active surface asialoglycoprotein receptors approximately 70%, but had no effect on the rate of internalization and retroendocytosis of ligand. The amount of surface asialoglycoprotein receptors can be reduced by lowering the incubation temperature gradually from 37 to 15 degrees C (Weigel, P. H., and Oka, J. A. (1983) J. Biol. Chem. 258, 5089-5094); vanadate affected only the temperature--sensitive receptors. Vanadate inhibited degradation of 125I-TC-AOM 70-80%. Degradation was much more sensitive to vanadate than binding; half-maximal effects were seen at approximately 1 mM vanadate for binding and approximately 0.1 mM vanadate for degradation. By subcellular fractionation in sucrose and Nycodenz gradients, it was shown that vanadate completely prevented the transfer of 125I-TC-AOM from endosomes to lysosomes. Therefore, the inhibition of degradation by vanadate was indirect; in the presence of vanadate, ligand did not gain access to the lysosomes. The limited degradation in the presence of vanadate took place in a prelysosomal compartment. Vanadate did not affect cell viability and ATP content.

Intracellular transport of asialoglycoproteins in rat hepatocytes. Evidence for two subpopulations of lysosomes.

The intracellular transport and degradation of asialoorosomucoid (AOM) in isolated rat hepatocytes was studied by means of subcellular fractionation in Nycodenz gradients. The asialoglycoprotein was labelled by covalent attachment of a radioiodinated tyramine-cellobiose adduct ( [125I]TC) which leads to labelled degradation products being trapped intracellularly and thus serving as markers for the degradative organelles. The ligand was initially (1 min) in a slowly sedimenting (small) vesicle and subsequently in larger endosomes. Acid-soluble, radioactive degradation products were first found in a relatively light lysosome whose distribution coincided in the gradient with that of the larger endosome. Later (30 min) degradation products were found in denser lysosomes which banded in the same region of the gradient as the lysosomal enzyme, beta-acetylglucosaminidase. Colchicine, monensin and leupeptin all inhibited degradation of [125I]tyramine-cellobiose asialoorosomucoid ( [125I]TC-AOM) and reduced the formation of degradation products in both the light and the dense lysosomes. In presence of monensin and colchicine no undegraded ligand was seen in the dense lysosome, suggesting that uptake in these vesicles was inhibited. Leupeptin allowed accumulation of undegraded ligand in the dense lysosome. Therefore, transfer from light to dense lysosomes is not dependent on degradation as such. In the presence of monensin two peaks of undegraded ligand were found in the gradients. It seems possible that in the monensin-sensitive endosomes, dissociation of the ligand-receptor complex is inhibited, allowing ligand to recycle with the receptors in small vesicles.

Intracellular transport of ovalbumin afterin vivo endocytosis in rainbow trout liver.

The intracellular handing of a mannose-terminated glycoprotein taken up in rainbow trout liver cells by receptor-mediated endocytosis has been studied. The intracellular transport and degradation of ovalbumin (OA) were studied by means of subcellular fractionation in Nycodenz gradients and by differential centrifugation following intravenous injection of the ligand. By using OA labelled with(125)I-tyramine cellobiose ((125)I-TC), the subcellular distribution of labelled degradation products could be studied, since they are trapped intracellularly in the organelle where the degradation takes place. (125)I-TC-OA was shortly after injection (45 min) localized in a homogenous population of endosomes. Labelled degradation products firs appeared in an organelle with the same density distribution as the endosomes. In livers homogenized 2h after injection the degradation products appeared in organelles with increasing size and density. After 24h, the degradation products were recovered in at least two populations of lysosomes with a distribution profile which coincided with that of the lysosomal enzyme ß-acetylglucosaminidase.The heterogeneous distribution of the late degradation products seemed not to be due to uptake of ligand in different liver cell types as only the parenchymal liver cells took up labelled OA after intravenous injection of the ligand.

Receptor-mediated endocytosis of retinol-binding protein by liver parenchymal cells: interference by radioactive iodination.

Retinol-binding protein (RBP) was iodinated directly by radio-iodine substitution on the tyrosyl residues by the sodium hypochlorite (NaOCl) or the Enzymobead (EB) methods, or indirectly by linkage of 125I-tyramine-cellobiose (TC) or 125I-N-succinimidyl-3-(4- hydroxyphenyl)propionic acid ester (SHPP) adduct on to free amino residues of RBP. Binding, uptake and degradation of iodinated RBP were studied in isolated rat and rabbit liver parenchymal cells. The amount of ligand bound to cells at 4 degrees C was dependent on the type of labelling, in that the 125I-TC ligand was bound to a lesser extent than NaClO-labelled 125I-RBP, EB-labelled 125I-RBP and 125I-SHPP-RBP. At 37 degrees C, the 125I-SHPP-RBP and the EB-labelled 125I-RBP became cell-associated more rapidly than the other two ligands. The higher cell association at 37 degrees C than at 4 degrees C suggests that internalization of the ligand occurred at the higher temperature. The degradation of the ligands was also different. The EB-labelled 125I-RBP, the 125I-TC-RBP and the 125I-SHPP-RBP showed an apparent lag phase before a steady increase in acid-soluble radioactivity was observed. Much less of EB-labelled 125I-RBP and 125I-TC-RBP were degraded (about 6%) than of the other two ligands (about 16%) after 120 min. About 50% of the acid-soluble radioactivity in these experiments could be accounted for by degradation in the medium, suggesting that about half of the degradation observed was intracellular. The present study therefore shows that the different labelling techniques yield varying estimates of the cellular handling of RBP. In addition, a rapid release of RBP was observed in experiments where cells were pulsed with radioactive RBP at 4 degrees C, washed and incubated further at 37 degrees C. Between 50% and 70% was released after 5 min of incubation. By increasing the temperature during the pulse to 37 degrees C, or by lowering the temperature during the chase to 4 degrees C, much less RBP was released from the cells. These data suggest that the release process represents recycling of internalized ligand from an early endosome.

Heterogeneity of degradation of B-cell endocytosed monoclonal antibodies reacting with different sIgM epitopes.

The degradation of a panel of monoclonal antibodies (MoAb) bound to surface IgM (sIgM) was studied in three human Burkitt's lymphoma cell lines. The panel included MoAb that recognize several distinct epitopes associated with the F(c mu)5 domain, the c mu 2 domain and kappa or lambda light chains. The amount of degraded MoAb and the rate of their degradation varied considerably between the various antibodies. Properties of MoAb such as avidity or ability to cross-link sIgM did not significantly influence their degradation. The most consistent correlation between rate of degradation and MoAb used was the location of the epitope recognized by the individual MoAb. Thus, 7 out of 8 anti-light chain MoAb were degraded at a higher rate than 5 out of 5 anti-F(c mu)5 MoAb. One anti-c mu 2 MoAb was degraded at a rate similar to the majority of anti-light chain MoAb. The intracellular transport of an anti-kappa light chain MoAb and an anti-F(c mu)5 MoAb was studied in detail by subcellular fractionation in sucrose gradients. We found that the anti-kappa light chain MoAb was transported more rapidly to lysosomes than the anti-F(c mu)5 MoAb, showing that they were sorted differently intracellularly.

Rat hepatocyte hyaluronan/glycosaminoglycan binding proteins: evidence for distinct divalent cation-independent and divalent cation-dependent activities.

We have previously shown (Biochemistry, 29, 10425, 1990) that hepatocytes contain intracellular specific binding sites for hyaluronan (HA). Although HA-binding activity is not dependent on divalent cations, it is increased in the presence of Ca+2. Here we report that a novel photoaffinity HA derivative (ASD-HA) crosslinks specifically to different proteins in permeable cells in the presence or absence of Ca+2. With Ca+2 present, two proteins of approximately 24 kD and 43 kD were labeled. Additionally, a broad zone of specific crosslinking was observed in the region of 40-100 kD. However, in the presence of the chelator EGTA this zone was absent and the 24 and 43 kD proteins were also not cross-linked to the HA photoaffinity derivative. In the absence of Ca+2, only a 54 kD protein was specifically labeled. The results indicate that different intracellular hepatocyte proteins are responsible for the Ca+2-independent and the Ca+2-dependent binding of HA.

Degradation of a monoclonal anti-mu chain antibody in a human surface IgM-positive B cell line starts in prelysosomal vesicle.

The surface IgM-mediated endocytosis and intracellular transport of an anti-F(c mu)5 mAb was studied by using subcellular fractionation in sucrose gradients. The results of such experiments showed that antibody was initially endocytosed in vesicles of low density, and later transferred to a presumably lysosomal compartment of higher density. SDS-PAGE analysis of gradient fractions showed that high Mr degradation fragments of the endocytosed antibody were formed in the low density vesicles before terminal degradation could be recorded. The partial degradation of the antibody was not blocked by low temperature or enzyme inhibitors, such as leupeptin and benzyloxycarbonyl-phenylalanylalanine-diazomyethyl-ketone, all of which severely retarded terminal degradation. The data also suggested that the recycling of partially degraded antibody to the cell surface employed a pool of such low density prelysosomal vesicles.

The relationship between autophagy and the intracellular degradation of asialoglycoproteins in cultured rat hepatocytes.

The relationship between autophagy and the intracellular distribution of endocytosed asialoorosomucoid was studied in cultured rat hepatocytes. Overt autophagy was induced by shifting the cells to a minimal salt medium. Incubation in minimal salt medium led to the formation of buoyant lysosomes at the expense of denser lysosomes manifested as a dual distribution of these organelles in Nycodenz gradients. Asialoorosomucoid was labeled with 125I-tyramine cellobiose. The labeled degradation products formed from this ligand are trapped at the site of degradation and may therefore serve as markers for the subgroup of lysosomes involved in the degradation. In control cells the degradation of the ligand was initiated in a light prelysosomal compartment and continued in denser lysosomes. In cells with high autophagic activity, the degradation of labeled asialoorosomucoid took place exclusively in a buoyant group of lysosomes. These results suggest that degradation of endocytosed ligand takes place in the same secondary lysosomes as substrate sequestered by autophagic mechanisms. These light lysosomes represent a subgroup of active lysosomes which are gradually recruited from dense bodies. Data are also presented that indicate that insulin may prevent the change in buoyant density brought about by incubation in deficient medium.

Separation of endocytic vesicles in Nycodenz gradients.

The endocytosis of 125I-labeled asialofetuin by rat hepatocytes was studied using Nycodenz/sucrose gradients. It was shown in pulse chase experiments that the ligand endocytosed initially (after 1/2 to 1 min) was in small, slow-sedimenting vesicles of similar sizes. The vesicles containing the ligand increased in size, and after about 2.5 min 20-30% of the ligand was recovered in larger, faster-sedimenting vesicles. After 15 min almost all internalized ligand was recovered in the fast-sedimenting vesicles. The initial, small endocytic vesicles and the later, larger endocytic vesicles have similar buoyant densities; the maturation of the endosomes can only be revealed by rate sedimentation, not by isopycnic centrifugation. Dissociation of ligand from receptor was found to occur in the larger, faster-sedimenting vesicles. The presence of ammonia inhibited the increase in size of the ligand-containing endosomes. The methods employed here offer the possibility of obtaining endocytic vesicles at various stage of their development for further studies.

Intracellular transport of endocytosed chylomicron [3H]retinyl ester in rat liver parenchymal cells. Evidence for translocation of a [3H]retinoid from endosomes to endoplasmic reticulum.

The intracellular transport of chylomicron remnants labeled with [3H]retinyl ester was studied in rat liver parenchymal cells by means of subcellular fractionation in Nycodenz and sucrose density gradients. The data presented indicate that endocytosed chylomicron remnant [3H]retinyl ester initially is located in low density endosomes. Radioactivity is subsequently transferred to a denser vesicle. Equilibrium as well as rate zonal centrifugation suggest that this denser [3H] retinoid-containing vesicle may represent endoplasmic reticulum. We have compared the intracellular transport of chylomicron remnant [3H]retinyl ester and 125I-asialofetuin. The receptor-mediated endocytosis of asialoglycoproteins in rat liver parenchymal cells is a thoroughly studied system. Our results suggest that the [3H] retinoid and 125I-asialofetuin follow the same path initially to the endosomes. After transit in endosomes, the intracellular transport differs. While asialofetuin is transported to the lysosomes, the retinoid is probably transferred to the endoplasmic reticulum.

Intracellular transport of endocytosed proteins in rat liver endothelial cells.

1. Receptor-mediated endocytosis of mannose-terminated glycoproteins in rat liver endothelial cells has been followed by means of subcellular fractionation and by immunocytochemical labelling of ultrathin cryosections after intravenous injection of ovalbumin. For subcellular-fractionation studies the ligand was labelled with 125-tyramine-cellobiose adduct, which leads to labelled degradation products being trapped intracellularly in the organelle where the degradation takes place. 2. Isopycnic centrifugation in sucrose gradients of a whole liver homogenate showed that the ligand is sequentially associated with three organelles with increasing buoyant densities. The ligand was, 1 min after injection, recovered in a light, slowly sedimenting vesicle and subsequently (6 min) in larger endosomes. After 24 min the ligand was recovered in dense organelles, where also acid-soluble degradation products accumulated. 3. Immunocytochemical labelling of ultrathin cryosections showed that the ligand appeared rapidly after internalization in coated vesicles and subsequently in two larger types of endosomes. In the 'early' endosomes (1 min after injection) the labelling was seen closely associated with the membrane of the vesicle; after 6 min the ligand was evenly distributed in the lumen. At 24 min after injection the ligand was found in the lysosomes. 4. A bimodal distribution of endothelial cell lysosomes with different buoyant densities was revealed by centrifugation in iso-osmotic Nycodenz gradients, suggesting that two types of lysosomes are involved in the degradation of mannose-terminated glycoproteins in liver endothelial cells. Two populations of lysosomes were also revealed by sucrose-density-gradient centrifugation after injection of large amounts of yeast invertase. 5. In conclusion, ovalbumin is transferred rapidly through three endosomal compartments before delivering to the lysosomes. The degradation seems to take place in two populations of lysosomes.