Nonpalmitoylated Human Asialoglycoprotein Receptors Recycle Constitutively but Are Defective in Coated Pit-mediated Endocytosis, Dissociation, and Delivery of Ligand to Lysosomes*

The hepatic asialoglycoprotein receptor (ASGP-R) internalizes desialylated glycoproteins via the clathrin-coated pit pathway and mediates their delivery to lysosomes for degradation. The human ASGP-R contains two subunits, H1 and H2. Cytoplasmic residues Cys36 in H1, as well as Cys54 and Cys58 in H2 are palmitoylated (Zeng, F.-Y., and Weigel, P. H. (1996) J. Biol. Chem. 271, 32454). In order to study the function(s) of ASGP-R palmitoylation, we mutated these Cys residues to Ser and generated stably transfected SK-Hep-1 cell lines expressing either wild-type or nonpalmitoylated ASGP-Rs. Compared with wild-type ASGP-Rs, palmitoylation-defective ASGP-Rs showed normal ligand binding, intracellular distribution and trafficking patterns, and pH-induced dissociation profiles in vitro. However, continuous ASOR uptake, and the uptake of prebound cell surface ASOR were slower in cells expressing palmitoylation-defective ASGP-Rs than in cells expressing wild-type ASGP-Rs. Unlike native ASGP-Rs in hepatocytes or hepatoma cells, which mediate endocytosis via the clathrin-coated pit pathway and are almost completely inhibited by hypertonic medium, only ∼40% of the ASOR uptake in SK-Hep-1 cells expressing wild-type ASGP-Rs was inhibited by hyperosmolarity. This result suggests the existence of an alternate nonclathrin-mediated internalization pathway, such as transcytosis, for the entry of ASGP-R·ASOR complexes into these cells. In contrast, ASOR uptake mediated by cells expressing palmitoylation-defective ASGP-Rs showed only a marginal difference under hypertonic conditions, indicating that most of the nonpalmitoylated ASGP-Rs were not internalized and processed normally through the clathrin-coated pit pathway. Furthermore, cells expressing wild-type ASGP-Rs were able to degrade the internalized ASOR, whereas ASOR dissociation was impaired and degradation was barely detectable in cells expressing nonpalmitoylated ASGP-Rs. We conclude that palmitoylation of the ASGP-R is required for its efficient endocytosis of ligand by the clathrin-dependent endocytic pathway and, in particular, for the proper dissociation and delivery of ligand to lysosomes.

Numerous studies have demonstrated that the post-translational modification of integral membrane proteins by palmitoy-lation plays a role in regulating specific protein functions (1)(2)(3)(4)(5). For example, CHO cells expressing nonpalmitoylated mutant transferrin receptors show a significant increase in the rate of receptor internalization when compared with the wild-type receptor, indicating that palmitoylation inhibits the rate of transferrin receptor endocytosis (6). Tanaka et al. (7) demonstrated that abolition of palmitoylation reduced the rate of intracellular trafficking of the human thyrotropin receptor, which resulted in delayed surface expression of newly synthesized receptor. Moreover, Blanpain et al. (8) found that failure of the CC-chemokine receptor CCR5 to be palmitoylated hindered its trafficking and thus led to the sequestration of CCR5 in intracellular biosynthetic compartments. Thus, one of the potential mechanisms for regulating intracellular trafficking may be through the covalent attachment of palmitic acid to a protein.
The ASGP-R, 1 found in liver parenchymal cells, is an endocytic receptor that can mediate the clearance of injected asialoglycoproteins from the circulation of mammals (reviewed in Refs. 9 -19). Native ASGP-Rs in isolated hepatocytes or hepatoma-derived cell lines endocytose asialoglycoproteins via the clathrin-dependent pathway. The internalized ligands are ultimately delivered to lysosomes for degradation, while the receptors recycle to the plasma membrane.
Based on a series of studies on ASGP-R-mediated endocytosis, we have previously shown that ASGP-Rs operate in two distinct pathways, called the State 1 and State 2 pathways (reviewed in Refs. 16,17,19). Receptors functioning in the State 2 pathway can be distinguished from those in the State 1 pathway based on their sensitivity to modulation by a variety of agents, such as monensin, chloroquine, microtubule drugs, or by ATP depletion (20 -22). During the process of endocytosis, only the receptors in the State 2 pathway undergo an I/R cycle that may be regulated by changes in the palmitoylation status of the receptors (23)(24)(25)(26). This two-pathway endocytosis system has also been genetically validated by Stockert et al. (27) with the isolation of a trafficking defective mutant cell line, Trf1, derived from HuH-7 cells. Although a functional State 1 pathway is present in the Trf1 cells, they are defective in endocytosis mediated by the State 2 pathway.
The human ASGP-R is composed of two subunits, H1 and H2 (28). The cytoplasmic Cys 36 in H1, as well as the analogous Cys 54 and the juxtamembrane Cys 58 in H2 are palmitoylated (29,30). 2 In order to elucidate the effects of palmitoylation on * This study was supported by National Institutes of Health Grant GM49695 from the National Institute for General Medical Sciences. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed. Tel.: 405-271-1288; Fax: 405-271-3092; E-mail: paul-weigel@ouhsc.edu. ASGP-R mediated endocytosis, the Cys residues that are palmitoylated in the H1 and H2 subunits were replaced by Ser using site-directed mutagenesis. The wild-type or mutant subunits were then stably expressed in SK-Hep-1 cells. Here we report that stable cell lines expressing either wild-type or nonpalmitoylated ASGP-Rs have a similar cellular distribution pattern of the receptors. However, cells expressing nonpalmitoylated ASGP-Rs are defective in ligand uptake, dissociation, and degradation.
Buffers and Media-Buffer 1 contains 150 mM NaCl, 6.7 mM KCl, and 10 mM HEPES, pH 7.4. Hanks balanced salt solution and PBS were made following standard recipes. DMEM and MEM were from Invitrogen. Medium 1 contains MEM supplemented with 2.4 g/liter HEPES, pH 7.4, 100 mg/ml sodium succinate, 75 mg/ml succinic acid, and 0.22 g/liter NaHCO 3 . Complete medium is DMEM supplemented with 10% fetal bovine serum (Summit Biotechnology), 2 mM L-glutamine, and 100 units/ml each of penicillin and streptomycin (Invitrogen).
Stable Transfectants-cDNAs encoding wild-type H1 and the H2 splice variant H2b (28), were generously provided by Dr. Martin Spiess (University of Basel, Switzerland). The generation of H1(C36S) by site-directed mutagenesis has been described (29). The generation of H2(C54S/C58S) will be described elsewhere. 2 For the generation of single transfectants expressing either wild-type H1 or H1(C36S), the corresponding cDNAs were subcloned into pIRES/neo (Clontech) and stably transfected SK-Hep-1 cell lines were selected and cloned. 2 Wild-type H2 and H2(C54S/C58S) cDNAs were inserted into pcDNA3.1(ϩ)/zeo (Invitrogen) using the BamHI and EcoRI sites. Each H2 cDNA was transfected into cell lines expressing H1 or H1(C36S), and stable double transfectants were selected and cloned. This resulted in the generation of four types of cell lines expressing either wild-type H1 and H2, wild-type H1 and mutant H2, mutant H1 and wild-type H2, or mutant H1 and H2 subunits. All transfections were performed using the Calcium Phosphate Transfection Kit (Invitrogen) according to the manufacturer's instructions. Cell lines were selected, cloned, and maintained in complete medium supplemented with 200 g/ml G418 (Mediatech) and 200 g/ml Zeocin (Invitrogen).
Total Cellular and Cell Surface 125 I-ASOR or 125 I-IgG Binding-Stable cell lines were grown in 12-or 24-well plates to confluence, washed, and incubated in DMEM without serum for 30 -60 min. Cells were then chilled on ice and incubated with 1.5 g/ml 125 I-ASOR, or 0.5 g/ml affinity-purified anti-H1 or anti-H2 125 I-IgG in Medium 1, in the presence (to assess total cellular ASGP-Rs) or absence (to assess cell surface ASGP-Rs) of 0.055% (w/v) digitonin for 1 h (33). Unbound ASOR or IgG was removed by extensive washing with ice-cold Medium 1. Cells were then solubilized with 0.3 M NaOH, and the radioactivity and protein content of the cell lysates were measured.
pH Sensitivity of ASGP-R⅐ASOR Complexes-Various stable cell lines were allowed to internalize 125 I-ASOR for 30 min at 37°C and chilled rapidly on ice, and unbound ligand was removed by washing three times with ice-cold Buffer 1. The rest of the procedure was performed on ice. Cells were permeabilized with 0.055% (w/v) digitonin in Buffer 1 for 15 min. Cells were washed again and then subjected to two 20-min incubations in Buffer 1 with different pH values, ranging from pH 5.0 to 7.4. The cells were washed once with Buffer 1, and the remaining cell-associated radioactivity was then determined.
Recycling of Receptor Subunits-Stable cell lines were grown to confluence in 35 mm dishes and washed with PBS. One ml of prewarmed medium containing 1 mg/ml proteinase-K (Roche Diagnostics Corp.) was added, and the cells were incubated at 37°C. At the desired time, the cells were chilled on ice, scraped, and collected in a microcentrifuge tube, pelleted at 4°C, and washed three times with ice-cold PBS containing 100 M phenylmethylsulfonyl fluoride. Cells were then solubilized in the latter buffer containing 1% Triton-X 100, 0.1% SDS, and 5 mM EDTA, and the lysates were subjected to reducing SDS-PAGE. The proteins were electrotransferred to nitrocellulose and the H1 and H2 subunits were detected by Western analysis using subunit-specific polyclonal antibodies (32). Images of the Western blots were captured digitally, and the density of the protein bands was quantified using a Fluorchem™ 8000 Imaging System from Alpha Innotech Corp.
Continuous 125 I-ASOR Uptake-The stable cell lines were grown in 24-well plates to confluence and then serum-starved for 30 -60 min, washed, and incubated with DMEM containing 1.5 g/ml 125 I-ASOR for various times at 37°C. In some experiments hyperosmolarity was used to block clathrin recycling by including 0.4 M sucrose in the medium. The cells were chilled rapidly on ice, washed three times with ice-cold Hanks, and lysed with 0.3 M NaOH. Radioactivity and protein content in the cell lysates were quantified.

125
I-ASOR Degradation-The degradation of 125 I-ASOR was determined by measuring the appearance of acid-soluble radioactivity in the medium (31). The cells were incubated with 125 I-ASOR as described above and at each time point, the medium was removed and mixed with an equal volume of 10% (w/v) phosphotungstic acid in 2 N HCl. After at least 15 min incubation on ice, the precipitates were pelleted by centrifugation, and the radioactivity remaining in the supernatant was quantified.
Internalization of Surface Bound 125 I-ASOR-Stable cell lines were grown in 12-well plates to confluence and serum-starved for 30 -60 min. Cells were then incubated on ice for 1 h in Hanks containing 1.5 g/ml 125 I-ASOR. Unbound 125 I-ASOR was removed by washing three times with Hanks. The cells were then resuspended in Hanks and incubated at 37°C. In some experiments hyperosmolarity was used to block clathrin recycling by including 0.4 M sucrose in the medium. After the desired internalization period, the medium was aspirated, and the cells were chilled rapidly on ice, followed by a 10-min incubation with Hanks containing 20 mM EGTA to remove the remaining surface-bound 125 I-ASOR. The medium was collected, and the cells were washed again and lysed. The radioactivity recovered in the EGTA wash (cell-surface associated 125 I-ASOR) and the cell lysates (internalized 125 I-ASOR) were then quantified.
Co-localization of Internalized ASOR and Lysosomes-Stable cells were grown in Lab-Tek II chamber slides (Nalge Nunc International) to ϳ70% confluence and then serum-starved for 30 min. Cells were washed, incubated with 1.5 g/ml fl-ASOR in DMEM, in the presence or absence of 0.4 M sucrose, for 1 h at 37°C. Cells were washed again and then incubated in DMEM containing 150 g/ml unlabeled ASOR for 30 min. The lysosomal marker Lysotracker Red DND-99 was added at a concentration of 50 nM, and cells were incubated at 37°C for an additional 30 min. After rinsing with PBS, the live cells were examined at room temperature using a Leica TCS confocal microscope equipped with a krypton/argon laser. Digital images of samples that had dual fluorescence staining were captured with Leica TCS NT Version 1.6.587 software. The extent of co-localization of green pixels (fl-ASOR) with red pixels (lysosome) was determined using Adobe Photoshop v5.5 and the image analysis protocol of Setiadi et al. (34).
Dissociation of Internalized 125 I-ASOR-Cells grown in 12-well plates to confluence were pulse-labeled with 1.5 g/ml 125 I-ASOR in DMEM for 3 min at 37°C and then incubated with a 100-fold excess of unlabeled ASOR for various times, after which the cells were chilled rapidly on ice. Remaining surface-bound 125 I-ASOR was removed by washing in Hanks containing 20 mM EGTA. The cells were then incubated in 0.5 ml of Hanks, with or without 0.055% (w/v) digitonin, to measure the amount of receptor-bound ligand and the total amount of intracellular ligand (both free and receptor-bound), respectively. After 15 min on ice, the cells were washed and solubilized in 0.3 M NaOH. Cell-associated radioactivity and protein content were determined. The difference in the radioactivity associated with cells in the presence or absence of digitonin represents ligand that was lost due to the permeabilization of cells and was therefore not bound to receptor (33).
Exocytosis of Internalized 125 I-ASOR-Cells grown in 12-well plates to confluence were preincubated with DMEM with or without 0.4 M sucrose for 10 min. The medium was aspirated, and the cells were pulse-labeled with 1.5 g/ml 125 I-ASOR in DMEM, in the presence or absence of 0.4 M sucrose, for 10 min at 37°C and then chilled rapidly on ice. Unbound ligand was removed by washing with ice-cold Hanks. Cell surface-bound 125 I-ASOR was then removed by a 15-min wash with Hanks containing 20 mM EGTA. The cells were then incubated at 37°C with 0.5 ml of Hanks in the presence or absence of 0.4 M sucrose. At the desired time, intact or degraded 125 I-ASOR released into the medium, and remaining cell-associated radioactivity were measured.
General-125 I radioactivity was measured using a Packard COBRA 5002 Gamma Counting System. The protein content in the cell lysates was determined by the method of Bradford (35) using BSA as the standard. SDS-PAGE was performed by the method of Laemmli (36). Specific binding was determined by subtracting the value of nonspecific binding from the total binding. Nonspecific binding was determined by incubating SK-Hep-1 cells, transfected with either the empty plasmids or plasmids containing H1 and H2 cDNAs, with 125 I-ASOR or 125 I-IgG by including a 100-fold excess of unlabeled ASOR or IgG in the mixes. Nonspecific binding was routinely ϳ20 -30% of the total bound radioactivity.

Generation of Stable Cell Lines Expressing Palmitoylationdefective ASGP-Rs-
The sequences of the cytoplasmic domains of H1 and H2, depicting the location of the palmitoylated Cys residues, are shown in Fig. 1A. In order to study the effects of Cys palmitoylation on receptor activity and function, stable cell lines expressing either wild-type or partially or completely palmitoylation-defective ASGP-Rs were created in SK-Hep-1 cells, a human hepatoma-derived cell line lacking endogenous ASGP-Rs (37). The expression of wild-type or mutant H1 and H2 subunits in all stable cell lines was confirmed by Western blot analysis using subunit-specific antibodies (not shown). All four ASGP-R types containing combinations of wild-type or Cys-variant H1 or H2 subunits have similar ligand-binding affinities in the range of 1-3 nM, as determined by equilibrium binding and Scatchard analysis. 3 To verify that ASGP-Rs from a cell line expressing [H1(C36S) and H2(C54S/C58S)] subunits were indeed unable to be palmitoylated, two different clonal cell lines that had been sequentially transfected with either [H1 and H2], or with [H1(C36S) and H2(C54S/C58S)] cDNAs were metabolically labeled with [ 3 H]palmitate (Fig. 1B). The results confirmed that ASGP-Rs containing wild-type H1 and H2 were radiolabeled with [ 3 H]palmitate (lanes 1 and 2), whereas ASGP-Rs containing H1(C36S) and H2(C54S/C58S) were not labeled (lanes 3 and 4).
Intracellular and Cell Surface Distribution of Wild-type and Palmitoylation-defective ASGP-Rs-We next determined the effect of Cys-to-Ser mutations in the H1 or H2 cytoplasmic domains on the distribution of active intracellular and cell surface receptors (Table I). All cell lines expressed comparable amounts of total cellular ASOR binding sites, assessed after digitonin permeabilization. Based on Scatchard analyses, there is a difference of only ϳ25% between the cell lines with the highest [H1(C36S)ϩH2] and the lowest [H1ϩH2(C54S/C58S)] number of ASOR binding. 3 The cell surface ASOR-binding capacities of nonpermeabilized cells were also similar among all cell lines (Table I). Roughly 75% of the total ASOR binding was intracellular. All three palmitoylation-defective ASGP-R cell lines expressed levels of H1 and H2 subunits similar to cells expressing wild-type ASGP-Rs, as determined by the amount of subunit-specific 125 I-IgG binding (Table I). We conclude that the Cys-to-Ser mutations in the H1 and H2 cytoplasmic domains have no effect on ASGP-R ligand binding activity or the distribution of active intracellular and cell surface receptors.
The Effect of Decreasing pH on the Dissociation of ASOR⅐ASGP-R Complexes-The rapid continuous uptake of ligand by a recycling receptor system requires an efficient process for dissociation of receptor-ligand complexes. Therefore, we examined the pH sensitivity of ASOR dissociation from the Cys-variant ASGP-Rs. The wild-type, partially palmitoylated and the palmitoylation-deficient ASGP-Rs showed similar profiles of ASOR dissociation as a function of decreasing pH (Fig. 2). Thus, the lack of palmitoylation did not significantly alter the sensitivity of the ASOR⅐ASGP-R complexes to pH.
Endogenous Trafficking Rates of Wild-type and Palmitoylation-defective ASGP-R Subunits-Using a proteinase K digestion method, we compared the rates of receptor subunit trafficking to the cell surface in cells expressing either the wildtype H1 and H2 subunits or their Cys-to-Ser variants (Fig. 3). The abundance of total cellular H1 after proteinase K treatment was estimated by quantifying the ϳ46-kDa mature H1 protein bands (solid arrows in Fig 3). The intracellular ϳ41-kDa precursor polypeptides of H1 (38) remained unaffected by the protease digestion (open arrows), indicating that cells were intact during the experiment. Thus the decrease in the amount of the mature H1 observed at various times was due to degradation by the exogeneous protease of subunits recycling to the cell surface. When the amount of H1 remaining was quantified and compared with the amount recovered at time 0 (Table II), the rates of H1 and H1(C36S) degradation, and thus trafficking from the cell interior to the cell surface, were almost identical with an apparent half-life of ϳ7 min. The apparent recycling rate of the wild-type H2 was also similar to that of H2(C54S/ C58S), with a half-life of ϳ7 min, when analyzed using the same protocol (not shown). These results indicate that the palmitoylation defect in the receptor subunits does not affect the rate at which H1 or H2 recycles between intracellular compartments and the cell surface.
Continuous Uptake and Degradation of 125 I-ASOR Are Slower in Cells Expressing Palmitoylation-defective ASGP-Rs-We next determined the rates of 125 I-ASOR uptake in cell lines expressing wild-type or palmitoylation-defective ASGP-Rs. Cells expressing palmitoylation-defective ASGP-Rs internalized ASOR at a reduced rate compared with cells expressing the wild-type ASGP-R (Fig. 4A). This result suggests that abolishing palmitoylation of the ASGP-R inhibited the rate of receptor internalization and ligand accumulation. In control experiments, specific uptake of ASOR was undetectable in SK-Hep-1 cells transfected with the backbone plasmids or in cells transfected with H1 cDNA alone (not shown). The rates of ASOR uptake in both of these control cell lines were comparable to the nonspecific uptake in cells transfected with both ASGP-R subunits.
We also determined the rates of 125 I-ASOR degradation in these stable cell lines (Fig. 4B). Wild-type ASGP-Rs mediated the specific degradation of ASOR at a rate at least 7-fold faster than that of the nonpalmitoylated ASGP-Rs, which was only marginally above the nonspecific rate of ASOR degradation. The overall specific ligand processing ability of cells expressing  3 and 4) were metabolically labeled with [ 3 H]palmitate. Active ASGP-Rs from these cell lines were purified and analyzed in parallel by Western blotting using anti-H1 antibody (WB) and fluorography (FL) as described under "Experimental Procedures." palmitoylation-defective ASGP-Rs was ϳ50% compared with cells expressing wild-type ASGP-Rs (Fig. 4C). The above results indicate that although cells expressing nonpalmitoylated ASGP-R are able to endocytose ASOR, they are defective in their ability to degrade the internalized ligands. In addition, cells expressing partially palmitoylation-defective ASGP-Rs, in which only H1 or H2 is not palmitoylated, showed essentially the same rates of ASOR uptake and degradation as cells expressing the completely nonpalmitoylated ASGP-Rs. 3 Transferrin Uptake Is Similar in Cells Expressing Wild-type or Palmitoylation-defective ASGP-Rs-The differences in the rates of ASOR uptake and degradation in these stable cell lines could be explained by a general defect in the coated pit endocytic pathway in an individual cell line. To test this possibility, we measured the endogenous rate of transferrin uptake and its sensitivity to hyperosmolarity as a control (Fig. 5). The rates of transferrin uptake were essentially identical in both cell lines and the accumulation of transferrin plateaued in both cases after ϳ1 h (not shown). The percent inhibition of transferrin uptake induced by hypertonic sucrose was also similar in both cell lines. However, transferrin uptake was not completely blocked by hypertonic sucrose treatment, which disrupts clathrin-coated pit mediated endocytosis and virtually eliminates ASOR uptake in rat hepatocytes (39,40). Since transferrin uptake also occurs through the clathrin-coated pit pathway (41), the partial inhibitory effect of hyperosmolarity indicates the existence of an alternate nonclathrin-mediated pathway for endocytosis of transferrin in these SK-Hep-1-derived cells. Nevertheless, these data show that the general endocytic pathways were the same in cell lines expressing either wild-type or palmitoylation-defective ASGP-Rs; they were not altered during the cell selection and cloning process.
Uptake of Surface-bound 125 I-ASOR-We also determined the initial rates of internalization of one synchronized round of surface-bound ASOR and the sensitivity of the uptake to hyperosmolarity. The binding at 4°C of ASOR to ASGP-R in transfected SK-Hep-1 cells (not shown) or in hepatocytes (39) was not affected by 0.4 M sucrose. Cells expressing wild-type ASGP-Rs were able to internalize increasing amounts of ASOR with time after the cells were warmed to 37°C (Fig. 6). Roughly 60% of the surface-bound ligand was internalized at 12 min. In addition, the majority of this ASOR uptake was inhibited by hypertonic sucrose treatment, showing that most of the initial ASOR uptake mediated by wild-type ASGP-Rs occurred via the clathrin-coated pit pathway. In contrast, internalization of prebound ASOR was not detected in cells expressing nonpalmitoylated ASGP-Rs, showing that there was a defect in the rapid endocytosis of surface-bound ASOR. These results show that  2. pH sensitivity of ASOR dissociation from palmitoylation-defective ASGP-R complexes. The indicated stable cell lines were allowed to internalize 125 I-ASOR for 30 min and were then washed, permeabilized, and washed in various buffers at the indicated pH as described under "Experimental Procedures." The amount of remaining cell-associated radioactivity was then determined and plotted relative to the amount of cell-associated radioactivity at pH 7.4 (100%).    Fig. 5 suggested that there is also a nonclathrin-mediated pathway for the endocytosis of transferrin into SK-Hep-1 cells. This prompted us to examine the effect of hyperosmolarity on continuous ASOR uptake in the two stable cell lines. Cells expressing wild-type ASGP-Rs were sensitive to hyperosmolarity and internalized ϳ40% less ASOR in the presence of 0.4 M sucrose at 3 h compared with the control (Fig. 7A). In contrast, cells expressing palmitoylationdefective ASGP-Rs were resistant to sucrose treatment, since only a small percent, if any, of ASOR uptake was inhibited (Fig.  7B). These results indicate that in cells expressing palmitoylation-defective ASGP-Rs, almost none of the ASOR uptake is through clathrin-coated pits. The data also suggest that in SK-Hep-1 cells a nonclathrin-mediated pathway is utilized for the entry of both wild-type and nonpalmitoylated ASGP-R⅐ASOR complexes.

Effect of Hyperosmolarity on Degradation of 125 I-ASOR-We
next examined the effect of hyperosmolarity on the degradation of internalized ASOR by the SK-Hep-1 stable cell lines. Cells expressing wild-type ASGP-Rs were able to degrade ASOR, and this degradation was inhibited by the presence of hypertonic medium (Fig. 8A), even though the cells continued to endocytose ASOR (see Fig. 7A). Cells expressing nonpalmitoylated ASGP-Rs were barely able to degrade the internalized ASOR either in the presence or absence of sucrose (Fig. 8B). These results suggest that the hyperosmolarity-sensitive pathway is responsible for the ASOR degradation seen in cells expressing wild-type ASGP-Rs.
Co-localization of Internalized ASOR with Lysosomes-In order to assess whether the ligand degradation deficiency observed in cells expressing palmitoylation-defective ASGP-Rs was due to the failure to deliver ligand to lysosomes, we determined the subcellular localization of the internalized ASOR (Fig. 9). Under isotonic conditions, co-localization of internalized fl-ASOR with lysosomes was apparent in both cell lines (Fig. 9, A and C), although less co-localization was observed in cells expressing nonpalmitoylated ASGP-Rs. However, in cells treated with hypertonic sucrose, the internalized fl-ASOR was generally not present in lysosomes in either cell line (Fig. 9, E  and G). These apparent differences were more evident when we I-ASOR, and endocytosis was induced by warming the cells to 37°C for 0, 5, or 12 min. Internalized (black bar) and cell surface (gray bar) ligand contents were determined as described under "Experimental Procedures." Radioactivity recovered in each fraction was expressed as a percentage of the total combined. The experiment was performed three times with similar results.

FIG. 4. Continuous uptake and degradation of 125 I-ASOR by cell lines expressing wild-type or nonpalmitoylated ASGP-Rs.
Cells expressing wild-type (q, E) or nonpalmitoylated ASGP-Rs (OE, ‚) were incubated with 1.5 g/ml 125 I-ASOR at 37°C. Nonspecific uptake (E, ‚) was determined as described under "Experimental Procedures." A, continuous uptake of 125 I-ASOR. At the indicated times, cells were chilled on ice, washed, and lysed, and cell-associated radioactivity was quantified. B, degradation of 125 I-ASOR. Medium samples were removed at the indicated times. Intact 125 I-ASOR in the medium was precipitated and the radioactivity associated with degraded ASOR remaining in the supernatant was quantified as described under "Experimental Procedures." C, the overall specific ligand processing ability of cells expressing wild-type (q) or palmitoylation-defective (OE) ASGP-Rs was calculated as the total amount of specific ASOR uptake and degradation determined in A and B, respectively. quantified the extent of co-localization of the green (fl-ASOR) with red (lysosomes) pixels from similar confocal images. Under isotonic conditions, 73 Ϯ 10% and 26 Ϯ 10% of vesicles containing internalized fl-ASOR co-localized with lysosomes, respectively, in cells expressing wild-type or nonpalmitoylated ASGP-Rs (means Ϯ standard deviations of 15 cells analyzed from different fields; p Ͻ 0.001, Student's t test). In contrast, under hypertonic conditions, almost no co-localization of fl-ASOR with lysosomes was detected (Ͻ2%) in either cell line. These results show that the cells expressing nonpalmitoylated ASGP-Rs were very inefficient in the delivery of ligand to lysosomes, compared with cells expressing the wild-type ASGP-Rs. In addition, fl-ASOR internalized by either type of ASGP-R through the nonclathrin-mediated pathway was not delivered to lysosomes.
Rate of Dissociation of Internalized 125 I-ASGP-R⅐ASOR Complexes-Cells were pulse-labeled with 125 I-ASOR and incubated with unlabeled ASOR for various times. After removing surface-bound ASOR, the total intracellular ASOR content (both free and ASGP-R-bound) and the fraction of ASOR that was still bound to receptors were then determined, respectively, by treating cells in the absence or presence of digitonin. More than 90% of the internalized ASOR was dissociated from the wild-type ASGP-Rs after 12 min, whereas only ϳ60% was dissociated from nonpalmitoylated ASGP-Rs (Fig. 10). The results indicate that the in vivo dissociation rate of ASGP-R⅐ASOR complexes was faster for wild-type ASGP-Rs compared with palmitoylation-defective ASGP-Rs.
Exocytosis of Internalized 125 I-ASOR-Although ASOR was internalized under hypertonic conditions in transfected SK-Hep-1 cells expressing either wild-type or nonpalmitylated ASGP-Rs, it was not delivered to lysosomes, and no degradation products were detected in the medium. Therefore, we examined the possible presence of a transcytosis pathway, by which ligands are internalized and transported across normally polarized cells without being degraded. The release of intact internalized ASOR into the medium was monitored in transfected SK-Hep-1 cells that had been pulse-labeled with 125 I-ASOR. Under hypertonic conditions, cells expressing either wild-type or nonpalmitylated ASGP-Rs released the same amounts (ϳ40 -50%) of internalized ASOR, as intact molecules, into the medium within 1 h (Fig. 11B). Comparatively less intact ASOR was externalized (Ͻ15%) from either cell line under isotonic conditions (Fig. 11A), but over the first 30 min the rate of intact ASOR release was 2-fold greater from cells expressing nonpalmitylated ASGP-Rs compared with wild-type ASGP-Rs. These results show that a significantly higher percentage of ASOR, internalized via the clathrin-independent pathway, is subsequently externalized. Importantly, a higher level of ASOR externalization occurred with nonpalmitylated ASGP-Rs, supporting the conclusion that their ligand processing and trafficking abilities are abnormal. DISCUSSION We have previously reported that both human ASGP-R subunits, H1 and H2, are post-translationally modified with pal- mitic acid (29,30,42). In order to elucidate the role of palmitoylation in ASGP-R function, we constructed stable cell lines expressing either the wild-type, or mutant ASGP-Rs incapable of palmitate attachment to either one or both receptor subunits. The partial or complete lack of ASGP-R palmitoylation did not affect the steady-state distribution of cell surface and intracellular receptors among all the stable cell lines tested. Approximately 75% of the total cellular receptors were localized intracellularly, which is similar to the distribution of native ASGP-Rs in isolated rat hepatocytes (33,43) and in HepG2 cells (44,45). The wild-type and palmitoylation-defective ASGP-Rs also have comparable ligand-binding affinities, 3 as well as similar profiles in vitro of dissociation of bound ligand as a function of pH.
In a series of previous studies (20 -24), we demonstrated that the State 2 ASGP-Rs are reversibly inactivated by treating cells with a variety of drugs that can also inactivate other recycling, endocytic receptors in various cell types (reviewed in Refs. 16,17,19). We also discovered that all ASGP-R subunits are palmitoylated (29,30,42) and that removal of the fatty acids from affinity-purified rat ASGP-Rs caused inactivation (loss of ligand binding activity) of the State 2, but not State 1, receptor population (26). Furthermore, State 2 ASGP-R inactivation in permeable hepatocytes (46) could be reversed by subsequent treatment with specific fatty acyl Co-A derivatives, which resulted in the reactivation of receptors (23,24). Based on these and other similar results (20,47), we proposed that during receptor-mediated endocytosis through the coated pit pathway, State 2 ASGP-Rs, as well as other recycling receptors, are transiently inactivated and then later reactivated before they recycle back to the cell surface (i.e. they undergo an I/R cycle). We suggested that this I/R cycle may be necessary for the efficient segregation and different intracellular routing of dissociated ligand and receptors, because it would eliminate the ability of ligand, at the predicted high concentration within endosomes, to rebind to receptors. Finally, we hypothesized that this I/R cycle was coupled to the fatty acylation status of State 2 ASGP-Rs. Modification of the cytoplasmic domains of the multimeric ASGP-Rs might regulate their activity, because the intercalation of fatty acyl chains into the membrane could alter the arrangement, and therefore ligand binding ability, of the extracellular carbohydrate binding domains (42). Clearly, our present results do not support the latter hypothesis, because fatty acylation is not necessary for the ligand binding activity of the ASGP-R. The lack of palmitoylation also did not affect the intracellular distribution or routing of ASGP-R subunits. Intracellular nonpalmitoylated ASGP-R subunits apparently still traffick normally within the cell. For example, the appearance of nonpalmitoylated subunits at the cell surface occurred at the same rate as the wild-type receptor subunits, indicating that receptor palmitoylation does not affect the rate at which oligomeric complexes or individual subunits recycle to and from the cell surface.
Despite the extensive similarities between wild-type and nonpalmitoylated ASGP-Rs mentioned above, SK-Hep-1 cells expressing nonpalmitoylated ASGP-Rs displayed a slower rate of continuous ASOR uptake, dissociation and degradation, compared with cells expressing wild-type ASGP-Rs. We also demonstrated that SK-Hep-1 cells expressing ASGP-Rs could internalize ASOR not only through the clathrin-dependent pathway, but also via an alternate, clathrin-independent route. ASOR internalization through noncoated pit pathways was not mediated by nonspecific adsorptive or fluid-phase endocytosis, but rather was specifically mediated by the ASGP-R, since SK-Hep-1 cells transfected with the plasmids alone did not accumulate ASOR. Hypertonic sucrose treatment reduced the rate of ASOR internalization by wild-type ASGP-R to a level similar to that of the nonpalmitoylated ASGP-R, which itself was resistant to such treatment.
Our results show that the greater ASOR uptake capacity in cells expressing wild-type, compared with palmitoylation-defective, ASGP-Rs was due to receptor-mediated uptake through clathrin-coated pits, and in turn suggest that palmitoylation of ASGP-Rs regulates their endocytosis and cellular routing via the clathrin-mediated pathway. Rates of ASOR uptake, dissociation, and degradation by cells expressing partially palmitoylation-defective ASGP-Rs, in which only one of the two ASGP-R subunits is not palmitoylated, were very similar to cells expressing ASGP-Rs with both subunits nonpalmitoylated. 3 The residual ligand uptake and processing by all three types of palmitoylation-defective ASGP-Rs were equally resistant to FIG. 10. Dissociation of internalized 125 I-ASOR. Cells expressing wild-type (q) or nonpalmitoylated (OE) ASGP-Rs were pulse-labeled with 125 I-ASOR, washed to remove surface-bound ligand, and then incubated with Hanks in the presence or absence of digitonin to measure, respectively, the amount of receptor-bound ligand and the total amount of cell-associated ligand as described under "Experimental Procedures." Results were calculated as the percentage of total intracellular ASOR that was free (dissociated) and are shown as the average Ϯ S.E. of three independent experiments, each performed in duplicate. The 7 and 12 min differences were both significant (p Ͻ 0.002, Student's t test). hyperosmolarity. Therefore, palmitoylation of both the H1 and