The Modular Adaptor Protein ARH Is Required for Low Density Lipoprotein (LDL) Binding and Internalization but Not for LDL Receptor Clustering in Coated Pits*

ARH is an adaptor protein required for efficient endocytosis of low density lipoprotein (LDL) receptors (LDLRs) in selected tissues. Individuals lacking ARH ( ARH (cid:1) / (cid:1) ) have severe hypercholesterolemia due to impaired hepatic clearance of LDL. Immortalized lymphocytes, but not fibroblasts, from ARH-deficient subjects fail to internalize LDL. To further define the role of ARH in LDLR function, we compared the subcellular distribution of the LDLR in lymphocytes from normal and ARH (cid:1) / (cid:1) subjects. In normal lymphocytes LDLRs were predominantly located in intracellular compartments, whereas in ARH (cid:1) / (cid:1) cells the receptors were almost ex-clusively on the plasma membrane. Biochemical assays and quantification of LDLR by electron microscopy indicated that ARH (cid:1) / (cid:1) lymphocytes had > 20-fold more LDLR on the cell surface and a (cid:1) 27-fold excess of LDLR outside of coated pits. The accumulation of LDLR on the cell surface was not due to failure of receptors to local-ize

the cytoplasmic tail of the receptor (1). The molecular machinery that sorts the LDLR to coated pits and promotes its rapid internalization has not been fully defined. Recently a crucial component of the machinery that internalizes LDLR was identified by elucidation of the molecular basis of autosomal recessive hypercholesterolemia (ARH), a rare form of severe hypercholesterolemia. Patients with ARH have normal LDLR but markedly reduced clearance of circulating LDL by the liver (2,3). The disorder is caused by mutations that inactivate a 308amino acid adaptor protein named ARH (4).
ARH contains four highly conserved domains (5,6). The first is a ϳ40-amino acid N-terminal domain of unknown function that is followed by a phosphotyrosine binding (PTB) domain (4). PTB domains are found in a variety of adaptor proteins involved in receptor trafficking and signaling (7). Typically PTB domains of adaptor proteins bind receptors via a consensus sequence, NPXY, in the cytoplasmic tail. In vitro studies with purified recombinant proteins indicate that the PTB domain of ARH binds to the unphosphorylated FDNPVY internalization sequence in LDLR in a sequence-specific manner (8,9). PTB domains share structural similarity with phosphoinositidebinding pleckstrin homology domains (10), and some PTB domains, including the domain in ARH, bind phosphoinositides (8,11,12).
Downstream of the PTB domain in ARH is a canonical clathrin box sequence (LLDLE), a conserved motif that mediates binding of several adaptor proteins to the terminal domain of the clathrin heavy chain (13,14). The clathrin box sequence in ARH can mediate high affinity binding to the heavy chain of clathrin (8,9). Finally ARH contains a highly conserved sequence at its C terminus that binds the ␤ 2 -subunit of AP2, a second structural protein in coated pits (9). The combination of these three functional regions in one protein coupled with the requirement of ARH for LDL endocytosis led to the proposition that ARH is required either to chaperone LDLR to the coated pit or to promote the internalization of the receptor (8,9).
Elucidation of the specific role of ARH in LDLR endocytosis has been hampered by the fact that cultured skin fibroblasts, which provided critical insights into other aspects of LDLR function, do not recapitulate the defective LDLR internalization observed in ARH subjects (2). Whereas fibroblasts from LDLR Ϫ/Ϫ patients (homozygous familial hypercholesterolemia) take up radiolabeled LDL at less than 10% of the rate observed in fibroblasts from normal individuals, fibroblasts from most ARH subjects take up and degrade LDL at 50 -100% of the normal rate (15). In contrast to fibroblasts, LDLR function is significantly impaired in immortalized lymphocytes from ARH patients (16,17). To further define the role of ARH in LDLR function we examined the cellular distribution and ligand binding characteristics of the LDLR in ARH Ϫ/Ϫ lymphocytes.

EXPERIMENTAL PROCEDURES
Materials-All cell culture reagents were from Invitrogen. Rabbit anti-LDLR IgG used for immunofluorescence and immunoelectron microscopy was from Maine Biotechnology Services, Inc. (Portland, ME). Rabbit anti-LDLR (4548) used in biochemistry experiments was a gift from Joachim Herz. The mouse monoclonal antibody to LDL receptor (C7) was from Santa Cruz Biotechnology Inc. (Santa Cruz, CA), mouse monoclonal anti-human transferrin receptor antibody was from Zymed Laboratories Inc. (San Francisco, CA), and mouse monoclonal anti-EEA was from BD Transduction Laboratories. Alexa-Fluor 488-conjugated goat anti-rabbit IgG and Alexa-Fluor 568-conjugated goat anti-mouse IgG were from Molecular Probes (Eugene, OR), and 10-nm gold-labeled goat anti-rabbit IgG and PD-10 columns were from Amersham Biosciences. Formaldehyde was from Fluka (Buchs, Switzerland), and sulfosuccinimidyl-6-biotinamido hexanoate and neutravidin-agarose were from Pierce. All other chemicals were from Sigma. LDL was prepared from freshly drawn human plasma as described previously (18).
Cell Culture-Lymphocytes, isolated from circulating blood of normal subjects, subjects with ARH deficiency, and subjects with familial hypercholesterolemia, were immortalized using the Epstein-Barr virus (19). Lymphocytes were maintained in Medium A (RPMI 1640 medium supplemented with 15% (v/v) fetal bovine serum, 20 mM HEPES, pH 7.5, 100 units/ml penicillin G, and 100 g/ml streptomycin). To upregulate expression of the LDLR, lymphocytes were grown in Medium B (RPMI 1640 medium supplemented with 10% (v/v) human lipoprotein-poor serum, 20 mM HEPES, pH 7.5, 100 units/ml penicillin G, and 100 g/ml streptomycin).
Immunofluorescence Microscopy-Immortalized normal, ARH Ϫ/Ϫ , and LDLR Ϫ/Ϫ lymphocytes were grown in Medium B for 48 h prior to mounting on poly-L-lysine-coated glass coverslips. The coverslips were rinsed briefly with Hanks' balanced salt solution and Buffer A (10 mm HEPES, pH 7.3) before fixation in freshly prepared 3% (v/v) formaldehyde in Buffer A for 30 min at room temperature. The coverslips were rinsed twice in PBS (0.15 M NaCl, 10 mM phosphate buffer, pH 7.3) and then incubated in 50 mM NH 4 Cl in PBS for 30 min. After rinsing twice in PBS, the fixed cells were permeabilized in PBS plus 0.1% (v/v) Triton X-100 for 7 min at 4°C and then conditioned with 1% (w/v) bovine serum albumin in PBS for 30 min. For double immunofluorescence staining, rabbit anti-LDLR IgG (0.5 g/ml) with either mouse monoclonal anti-human transferrin receptor (10 g/ml) or mouse monoclonal anti-EEA (2.5 g/ml) antibodies were applied to the coverslips overnight at 4°C. The coverslips were rinsed with PBS and then incubated for 2 h with Alexa-Fluor 488-conjugated goat anti-rabbit IgG (10 g/ml) and Alexa-Fluor 568-conjugated goat anti-mouse IgG (10 g/ml). The coverslips were rinsed in PBS prior to mounting and imaged using a Leica TCS SP confocal microscope.
Electron Microscopy-Cultured lymphocytes were immersion-fixed in 3% (w/v) paraformaldehyde in Buffer A for 60 min. The lymphocytes were suspended in 2% (w/v) agar and then infused in 2 M sucrose containing 15% (w/v) polyvinylpyrrolidone (10 kDa). Frozen ultrathin sections were prepared using a Leica Ultracut UCT ultramicrotome equipped with a Leica EMFCS cryochamber. The sections were lifted onto nickel grids in 1.8% (v/v) methylcellulose and 2.3 M sucrose (1:1) and stored overnight on gelatin at 4°C (20 -22). Before immunolabeling, the gelatin was liquefied at 37°C, and the nickel grids were removed. The sections were washed by floating the sections on droplets of PBS.
For immunogold localization, the grids with the attached thin sections were conditioned on droplets containing 1% (w/v) bovine serum albumin, 0.01% (v/v) Triton X-100, and 0.01% (v/v) Tween 20 in PBS (Buffer B) for 10 min at room temperature. The grids were incubated for 2 h in the presence of rabbit anti-LDLR IgG antibody and diluted in Buffer B to a final concentration of 1 g/ml. The sections were rinsed on droplets of PBS and then incubated with 10-nm gold-labeled goat antirabbit IgG (diluted 1:40) in Buffer B containing 10% normal goat serum. Finally the grids with the attached thin sections were rinsed in PBS, fixed with 2% (v/v) glutaraldehyde in PBS for 10 min, embedded, and stained with methylcellulose and uranyl acetate (20,21).
For cell surface immunolabeling, the lymphocytes were formaldehyde-fixed as described above, rinsed in Buffer B, and incubated with rabbit anti-LDLR IgG antibody (1 g/ml) in Buffer B for 17 h. The primary antibodies were localized by incubating the cells for 2 h with 10-nm gold-labeled goat anti-rabbit IgG (diluted 1:40) in Buffer B. After extensive washings, the cells were fixed with 2% glutaraldehyde and postfixed in 1% (w/v) osmium tetroxide. Subsequently the specimens were rinsed in distilled water, dehydrated with graded ethanol, and then Epon-embedded according to the manufacturer's protocol (Electron Microscope Science). Ultrathin sections (ϳ80 nm) were cut with a diamond knife using a Leica Ultracut R ultramicrotome and placed on Formvar/carbon-coated nickel grids. Ultrathin sections were stained with 3% aqueous uranyl acetate (15 min) and lead acetate (5 min). Electron micrographs were taken using a JEOL 1200 electron microscope operating at 80 kV.
Colloidal gold-conjugated LDL was produced as described previously (23). Briefly 100 ml of 10% (w/v) gold chloride was added to 100 ml of boiling H 2 O. After 10 s, 2 ml of 1% trisodium citrate was added, and the mixture was maintained at 100°C for 5 min. The reaction was cooled to room temperature, adjusted to a pH of 6 with dipotassium phosphate, and then centrifuged at 800 ϫ g for 30 min at 4°C to remove aggregates. The colloidal gold was collected by centrifugation at 17,500 ϫ g for 40 min at 4°C. The centrifuged material was aspirated to a volume of 1 ml. The colloidal gold pellet was resuspended and added to an equal volume of human LDL (1 mg/ml), which had been dialyzed overnight in 50 mM EDTA, pH 6.0. After allowing the mixture to equilibrate to room temperature (60 min), the mixture was overlaid on a 35% sucrose cushion and centrifuged at 17,500 ϫ g for 60 min at 12°C. The colloidal gold-LDL in the pellet was resuspended in PBS (2.5 ml), solvent-exchanged over a PD-10 column equilibrated in PBS, and dialyzed twice against 3 liters of PBS at 4°C. The colloidal gold-labeled LDL was used within 5 days of synthesis.
Immortalized lymphocytes were seeded at 5 ϫ 10 5 cells/ml in Medium B and cultured for 2 days. Cells were then counted and resuspended at 5 ϫ 10 6 cells in 1 ml of Medium B at either 4 or 37°C prior to addition of 10 g/ml colloidal gold-labeled LDL. The cells were maintained at either 4 or 37°C for 90 min prior to centrifugation at 900 ϫ g for 5 min at 4°C and then washed three times using Buffer C (50 mM Tris-HCl, 150 mM NaCl, 2 mg/ml bovine serum albumin, pH 7.4). Cells were fixed, embedded, sectioned, and placed on nickel grids as described above.
Quantification of Gold Labeling-Electron microscope images were obtained by randomly taking 100 photographs of each cell type (normal, ARH Ϫ/Ϫ , and LDLR Ϫ/Ϫ ) that had been labeled with the LDLR-specific antibody and 25 photographs of each cell type labeled with the colloidal gold-LDL. The length of the non-coated pit membranes, the diameter of the coated pits, and the number of gold particles associated with each of these two regions were determined. The labeling intensity was expressed as the number of gold particles per micrometer length of the different regions of the plasma membrane. Clusters of tightly associated gold particles in the experiments using LDL were counted as one since these likely represented aggregates of LDL. Gold particles separated by a gap greater than twice their diameter were counted as single particles.
Quantification of Cell Surface LDLR-A total of 5 ϫ 10 6 lymphocytes were washed with ice-cold Medium C (1ϫ Eagle's modified minimal essential medium supplemented with 20 mM HEPES, pH 7.4) and incubated on a rotator with 125 I-labeled monoclonal antibody to human LDLR (C7) or a rabbit polyclonal antibody to the human LDLR (4548) (7 g/ml) for 90 min at 4°C. Cells incubated with 4548 were washed five times in Buffer C and then resuspended in 1 ml of Buffer C with 1 Ci of 125 I-Protein A prior to rotating for 60 min at 4°C. After incubation with either 125 I-C7 or 125 I-Protein A, cells were washed three times with Buffer C and twice with Buffer D (50 mM Tris-HCl, 150 mM NaCl, pH 7.4) and then resuspended in 1 ml of Buffer D. Cellular protein was determined by the method of Bradford (24), and the 125 I counts were determined using half of the sample. The counts were normalized to the cellular protein. In the case of C7, counts were converted to fmol of C7. All experiments were performed in triplicate.
An additional method used to quantify cell surface LDLR was by biotinylating the cell surface proteins (25). Briefly 2.5 ϫ 10 6 cells were washed twice with cold PBS and resuspended in 1 ml of PBS plus sulfosuccinimidyl-6-biotinamido hexanoate (1 mg/ml) for 30 min at 4°C with end-over-end mixing. Cells were then washed with Buffer D and incubated with Buffer D for 30 min on ice to quench any residual biotin reagent. Cells were washed twice with cold PBS and lysed with 60 l of lysis buffer (1% Triton X-100, 4 mM EGTA, 10 mM Tris-HCl, pH 8) at room temperature for 30 min with end-over-end mixing. Cells were subjected to centrifugation at 15,000 rpm for 5 min in a microcentrifuge to remove insoluble debris. A total of 15 l of lysate was retained (Lysate). A total of 40 l of lysate was added to 100 l of a 50% slurry of neutravidin-agarose and 660 l of lysis buffer. The mixture was rotated for 60 min at 4°C and then pelleted. The supernatants consisted of unbiotinylated material, which contained internal proteins (Internal). The beads were washed three times with 15 mM Tris-HCl, 4 mM EGTA, 500 mM NaCl, and 0.5% Triton X-100, pH 8.0 and then in the same buffer without NaCl. Proteins that remained associated with the beads were biotinylated and represented proteins exposed on the cell surface. This material was eluted by incubating the beads in SDS sample buffer at 90°C for 5 min (Surface). Samples were then sizefractionated by 5-17% gradient SDS-PAGE, transferred to polyvinyli-dene difluoride membranes (Millipore), and processed for immunoblot analysis using a polyclonal anti-LDLR antibody (4548).
Endocytosis of LDLR-Endocytosis of the LDLR was monitored using a modification of the protocol described previously (26). Briefly 5 ϫ 10 6 lymphocytes were aliquoted into 5 ml of Medium B. Monensin was added to a final concentration of 25 mM for 0, 5, 20, or 60 min such that all incubations ended at the same time. Cells were plunged into ice-cold  PBS to inhibit further endocytosis, washed with cold PBS, resuspended in 1 ml of PBS plus sulfosuccinimidyl-6-biotinamido hexanoate (1 mg/ ml), and incubated for 30 min at 4°C with end-over-end mixing. Samples were then processed for surface expression of LDLR using neutravidin-agarose as described above.
Lipoprotein Binding Assays-Radiolabeled human 125 I-LDL and rabbit 125 I-labeled ␤-migrating very low density lipoprotein ( 125 I-␤-VLDL) binding assays were performed in triplicate as described previously (18). Acid and EDTA washes were conducted at 4°C using Buffer E containing 50 mM Tris maleate, 150 mM NaCl, and 2 mM CaCl 2 , pH 7.6 (control), Buffer E at pH 5.0 (acid wash), or Buffer E with 10 mM EDTA (EDTA wash) for 1 h with end-over-end mixing prior to addition of 125 I-LDL. The experiments used to compile the data for the Scatchard plots included the following modifications. In each experiment, 2.5 ϫ 10 6 lymphocytes were suspended in 100 l of ice-cold Medium C. The cells were then incubated with 5 g/ml 125 I-LDL together with increasing concentrations of cold LDL (up to 5 mg/ml). The specific activity of the labeled LDL in each assay was recalculated based upon the total LDL concentration. Bound and free counts were separated by centrifugation of the cells through a cushion of Medium C plus 10% sucrose at 10,000 ϫ g for 5 min. Tubes were then frozen in liquid nitrogen and cut to separate the bound and free LDL. The amount of free and bound LDL in each assay was calculated using the specific activity of LDL in each assay and assembled into a Scatchard plot. Ligand blot assays were performed as described previously (27).

RESULTS
In normal cells, LDL binds to the LDLR on the cell surface, and the LDL⅐LDLR complex is internalized and delivered to endosomes. In the low pH environment of the endosome, the receptor dissociates from the lipoprotein and recycles to the cell surface. LDL that is released from the receptor in the endosome is degraded, delivering cholesterol to the cell. The degradation of 125 I-LDL is severely compromised in ARH Ϫ/Ϫ lymphocytes despite an increased amount of LDL binding to the surfaces of these cells (16,17). To determine whether the decreased LDL degradation in ARH Ϫ/Ϫ cells is due to a failure of LDLR internalization, ARH Ϫ/Ϫ and normal lymphocytes were treated with monensin to block LDLR recycling, and the rates of disappearance of LDLR from the cell surface were compared in the two cell lines (Fig. 1). The half-life of LDLR at the plasma membrane was ϳ8 min in normal lymphocytes, a value similar to that seen in fibroblasts (26). In contrast, no loss of surface LDLR was detected in ARH Ϫ/Ϫ cells after 60 min. These results are consistent with a failure of LDLR to be internalized in ARH Ϫ/Ϫ lymphocytes.
A potential consequence of the failure to internalize LDLR is an accumulation of receptors on the cell surface. We examined the localization of the LDLR in normal and ARH Ϫ/Ϫ cells using immunofluorescence. In normal lymphocytes LDLR staining was observed predominantly in internal compartments (Fig. 2). The internal LDLR staining co-localized with the early endosome marker EEA1 (Fig. 2) and partially co-localized with the lysosomal marker lysosome-associated membrane protein-1 (data not shown). In contrast, LDLR was predominantly on the cell surface of ARH Ϫ/Ϫ lymphocytes. The distribution of LDLR in the ARH Ϫ/Ϫ lymphocytes mirrors the distribution observed previously in hepatocytes from ARH Ϫ/Ϫ mice (28). The relative abundance of surface LDLR on normal and ARH Ϫ/Ϫ cells was quantified by biotinylating cell surface proteins, precipitating with neutravidin-agarose, and immunoblotting for the LDLR. ARH Ϫ/Ϫ cells had ϳ20-fold more LDLR on the cell surface than normal cells (Fig. 3A). A similar increase in LDLR number was

FIG. 3. Loss of ARH dramatically increases the amount of immunodetectable LDLR on the cell surface. In
A, LDLR Ϫ/Ϫ , normal, and ARH Ϫ/Ϫ cells were treated with a non-cell-permeable biotinylation reagent, washed, and lysed with Triton X-100. After removal of insoluble debris, one-fourth of the lysate was set aside, while two-thirds of the lysate was incubated with neutravidin-agarose to precipitate biotinylated proteins. Shown in the upper portion of A is an immunoblot of lysates (lanes 1-3), biotinfree material (lanes 4 -6), and biotinylated material (lanes 7-9). Lanes 1, 4, and 7 were from LDLR Ϫ/Ϫ lymphocytes; lanes 2, 5, and 8 were from normal cells; and lanes 3, 6, and 9 were from ARH Ϫ/Ϫ cells.
The lower portion of A shows a quantification of the immunoblot normalized to sample size. In B, the relative number of LDLRs on the cell surface was determined by incubating LDLR Ϫ/Ϫ , ARH Ϫ/Ϫ , and normal lymphocytes with an 125 I-labeled mouse monoclonal antibody against the LDLR (C7) at 4°C for 90 min. Cells were washed, lysed, and counted on a ␥ counter. The cellular protein content was determined and used to normalize each sample. In C, the relative number of LD-LRs on the cell surface was determined by incubating LDLR Ϫ/Ϫ , normal, and ARH Ϫ/Ϫ lymphocytes and a no-cell control with a rabbit polyclonal antibody against the LDLR at 4°C. Cells were then washed, incubated with 125 I-Protein A, washed again, lysed, and counted on a ␥ counter. The cellular protein content was used to normalize each sample. seen when cell surface LDLRs were detected with 125 I-labeled monoclonal anti-LDLR antibody or with polyclonal anti-LDLR followed by 125 I-Protein A (Fig. 3, B and C).
To determine whether the accumulation of LDLR on the plasma membranes of ARH Ϫ/Ϫ lymphocytes reflected a failure to cluster the receptors in coated pits, the surface distribution of LDLR was examined by immunoelectron microscopy. In agreement with the immunofluorescence data (Fig. 2), LDLR was located predominantly on the cell surfaces of ARH Ϫ/Ϫ lymphocytes, while in normal cells the LDLR was associated with internal vesicular structures resembling endosomes and lysosomes (Fig. 4). Quantification of immunogold particles on the cell surface revealed a ϳ23-fold increase in the number of LDLRs/m of membrane in the ARH Ϫ/Ϫ cells (Table I). Surprisingly there were at least as many LDLRs/m of membrane in the coated pits of ARH Ϫ/Ϫ cells as in normal cells (Table I).
The large increase (Ͼ27-fold) in the number of LDLRs in the non-coated pit membrane of the ARH Ϫ/Ϫ cells resulted in a significant reduction in the proportion of surface LDLRs in coated pits of the ARH Ϫ/Ϫ cells as compared with normal cells (1 versus 31%). Thus, the loss of ARH resulted in a redistribution of LDLR from endosomes to the non-coated pit portion of the plasma membrane without significantly affecting the level of LDLR in coated pits.
The relative increase in the number of cell surface LDLRs in ARH Ϫ/Ϫ lymphocytes observed in this study (ϳ15-20-fold) is substantially higher than the relative increase in LDL binding (2-fold) observed previously (16,17,29). Therefore, we tested the capacity of ARH Ϫ/Ϫ cells to bind two well characterized LDLR ligands: ␤-VLDL particles, which bind to LDLR through apoE, and LDL, which binds to LDLR through apoB-100. The surface binding capacity of 125 I-␤-VLDL to ARH Ϫ/Ϫ cells was markedly greater (14-fold) than the binding capacity of the normal cells (Fig. 5A), which is consistent with the large increase in cell surface LDLR in the ARH Ϫ/Ϫ cells. By contrast 125 I-LDL binding was only 2-fold higher in ARH Ϫ/Ϫ cells than in the normal cells (Fig. 5B) in agreement with prior estimates of the relative LDL binding capacity of ARH Ϫ/Ϫ lymphocytes (16,17,29).
To determine whether the decrease in LDL binding was due to a reduction in the affinity of the LDLR for LDL in the ARH Ϫ/Ϫ cells, we performed more detailed binding studies. Scatchard plot analysis indicated that the 2-fold increase in LDL binding to the ARH Ϫ/Ϫ cells was due to a ϳ2-fold increase in the number of high affinity LDL binding sites rather than to a large increase in low affinity binding sites (Fig. 5C). Together these binding experiments indicated that the majority (Ͼ90%) of LDLR on the cell surface of ARH Ϫ/Ϫ cells was unable to bind lipoproteins via apoB-100.
To determine which LDLRs on the cell surface bind LDL, we  incubated the ARH Ϫ/Ϫ and normal cells with colloidal goldlabeled LDL (LDL-gold) for 1 h at either 4 or 37°C and performed electron microscopy (Fig. 6). At 4°C the ARH Ϫ/Ϫ cells bound ϳ2-fold more LDL-gold than did normal cells, which is consistent with the 125 I-LDL binding assays (Table II and Fig.  5). At 37°C, ARH Ϫ/Ϫ cells also had more surface-associated LDL-gold than did normal cells, although the increase was somewhat less than 2-fold. Within coated pits, the number of LDL-gold particles/m of membrane was proportional to the number of LDLR-immunogold particles/m of membrane both in normal and in ARH Ϫ/Ϫ cells (Tables I and II). In contrast, LDL-gold binding to the surface of non-coated pit membranes was much less efficient in ARH Ϫ/Ϫ cells than in normal cells. Thus, in ARH Ϫ/Ϫ lymphocytes the LDLR binds LDL in coated pits but largely fails to bind LDL outside coated pits (Ͼ95% failure).
How does ARH promote LDL binding to the LDLR outside the coated pit? One possibility is that the LDLR in ARH Ϫ/Ϫ cells undergoes a post-translational modification that interferes with LDL binding. We explored this possibility by comparing the electrophoretic mobility on denaturing gels of the LDLR from the ARH Ϫ/Ϫ and normal cells; no differences in mobility were detected (data not shown). Moreover 125 I-LDL ligand blots performed using Triton X-100 extracts from the two cell lines showed no difference in LDL binding activity (Fig.  7). Thus, the defect in LDL binding in ARH Ϫ/Ϫ cells appears not to be intrinsic to the LDLR but rather is specific to the cellular context of the receptor.
Next we tested whether the inability of most LDLRs on ARH Ϫ/Ϫ cells to bind LDL was due to non-covalent interactions between the ectodomain of LDLR and another protein, such as receptor-associated protein or Boca (30,31). Cells were washed with acidic or EDTA-containing buffers to release any attached ligands prior to assaying for LDL binding. No increases in the LDL binding capacity were seen in the ARH Ϫ/Ϫ cells after these treatments (Fig. 8). We also did not see any difference in the electrophoretic mobility of LDLR from ARH Ϫ/Ϫ cells on nondenaturing blue native gels nor did we detect any difference in chemical cross-linking behavior of LDLRs (32) from the two cell types (data not shown). Thus, the failure of LDLR to bind LDL did not appear to be due to association of alternative ligands to the receptor ectodomain.

DISCUSSION
The results of this study indicate that the modular adaptor protein ARH plays a crucial role in both ligand binding and internalization of the LDLR. Loss of ARH resulted in failure of LDLR endocytosis and a dramatic redistribution of LDLR such that the vast majority of receptors were on the cell surface rather than in internal compartments. Electron microscopy revealed that most of the LDLRs on the surfaces of ARH Ϫ/Ϫ cells were outside coated pits and were unable to bind LDL. Biochemical studies confirmed that despite a Ͼ20-fold increase in immunoreactive LDLR on the cell surface, LDL cell surface binding was only modestly (2-fold) increased in ARH Ϫ/Ϫ cells. Within coated pits, ARH Ϫ/Ϫ cells had at least as many LDLR and LDL binding sites as did normal cells. Remarkably, however, the LDLR and LDL were not efficiently endocytosed from these pits. Thus, ARH was not required for localization of the LDLR to coated pits but rather appeared to be required for endocytosis of the LDLR from pits. In addition the failure of Ͼ90% of LDLR in ARH Ϫ/Ϫ cells to bind LDL suggested that ARH potentiated the ability of LDLR to bind LDL in vivo.
We have shown previously that ARH is required for normal uptake and degradation of LDL in human lymphocytes and in mouse liver (17,28), but the specific role of ARH in LDLRmediated endocytosis has not been defined. The observation that ARH can bind to LDLR, clathrin, and AP2 suggested that ARH couples the LDLR to the endocytic machinery and may direct LDLR to coated pits or anchor the receptor in the pit during endocytosis. Our present results indicated that the number of LDLRs in coated pits was similar in normal and ARH Ϫ/Ϫ lymphocytes and that the number of LDL particles bound in coated pits was normal in ARH Ϫ/Ϫ cells. Unlike normal cells, ARH Ϫ/Ϫ cells did not accumulate LDL-gold in multivesicular bodies characteristic of late endosomes and lysosomes when maintained at 37°C (Fig. 6). We were also unable to detect LDLR internalization after monensin treatment in ARH Ϫ/Ϫ cells. Thus, despite the presence of both LDLR and LDL in coated pits of ARH Ϫ/Ϫ cells, neither was internalized. The failure to internalize LDLR is consistent with the reduced uptake and degradation of 125 I-LDL observed previously in ARH Ϫ/Ϫ lymphocytes (16,17,29). These observations suggested that the loss of ARH arrested the LDLR trafficking cycle at the point of endocytosis of LDLR from coated pits. How might ARH be involved in the endocytosis of LDLR from coated pits?
ARH may facilitate the endocytosis of LDLR and LDLR⅐LDL complexes from coated pits by stabilizing the interaction between the receptor and the structural components of the pit.
Efficient trapping of receptors in coated pits can be mediated by a so called "rugged energy landscape" in which multiple interactions of relatively low but varying dissociation constants act in concert to reduce the dissociation rate of receptors from coated pits (33). In accordance with this mechanism, coated pits contain many potential binding partners for LDLR including AP2 (34), PTB domain-containing adaptor proteins such as Dab2 (12,28) and the clathrin heavy chain (35). The associations of these molecules individually or collectively may enable ARH Ϫ/Ϫ cells to accumulate LDLR in coated pits but may be insufficient to hold the receptor in the invaginating portion of the budding pit. In fibroblasts, many LDLRs appear to be expelled from the invaginating portion of coated pits during pit budding (36), suggesting that poorly anchored receptors are not readily captured by the budding portion of the coated pit. ARH may stabilize the association of the receptor with the invaginating portion of the budding pit, thereby increasing the efficiency of LDLR internalization.
The second major finding of this study is that the absence of ARH abolished the ability of more than 90% of cell surface LDLR to bind LDL. In coated pits, the ratio of LDL to LDLR was similar in normal and ARH Ϫ/Ϫ cells. Comparison of the ratio of LDL-gold to LDLR-gold in the non-coated pit membranes of normal and ARH Ϫ/Ϫ cells indicated that the receptors that failed to bind LDL on ARH Ϫ/Ϫ cells were predominantly located outside coated pits. The ratio of LDL to LDLR also indicated that in normal cells LDL bound to the LDLR with lower efficiency inside coated pits than outside. This lower efficiency can be explained by receptor occlusion resulting from binding of LDL to adjacent LDLR when the receptors are at high density as they are in coated pits (37). In contrast, ARH Ϫ/Ϫ cells exhibited less efficient binding of LDL to receptors outside of coated pits than inside pits. These observations strongly suggest that ARH plays a role in the LDL binding activity of receptors outside coated pits.
It is currently unclear how ARH participates in LDL binding by the LDLR. Scatchard plots of LDL binding to ARH Ϫ/Ϫ cells indicated that the small proportion of LDLR that bound to LDL did so with normal affinity (Fig. 5C). The receptors from ARH Ϫ/Ϫ and normal cells did not appear to have different covalent modifications since the LDLR had equivalent electrophoretic mobility on denaturing SDS-polyacrylamide gels and bound equivalent amounts of 125 I-LDL in ligand blots. Noncovalent associations of the LDLR with alternative ligands also did not appear to be the cause of the LDL binding defect since washing ARH Ϫ/Ϫ cells with acidic or EDTA-containing buffers did not improve their ability to bind LDL (Fig. 8). We also did not see any difference in the electrophoretic mobility of LDLR from ARH Ϫ/Ϫ cells on non-denaturing blue native gels nor did we detect any difference in chemical cross-linking behavior of LDLR (32) from the two cell types. Thus, the failure of most LDLRs in ARH Ϫ/Ϫ cells to bind LDL was not an intrinsic property of the LDLR nor were the LDLRs blocked by an alternative ligand. How then might ARH affect LDL binding?
We propose the following model. In the absence of ARH, the LDLR located outside the pits associates with the cell surface in such a manner that the ligand binding domains are partially occluded. The apoB-100 binding site on LDLR encompasses a large surface (38), requiring cysteine-rich repeats 3-7 and the first epidermal growth factor repeat. In contrast, the apoE binding site only requires cysteine-rich repeat 5 (39,40). Thus, partial occlusion may be sufficient to prevent binding to apoB-100 but not apoE binding. ARH binding to the cytoplasmic tail of LDLR, perhaps in combination with ARH binding to phosphatidylinositol 4,5-bisphosphate, may relieve the partial occlusion, allowing the full exposure of the ligand binding domains of LDLR and thereby making them accessible to LDL. Since LDL is able to bind to the LDLR in the coated pits of the ARH-deficient cells, other LDLR binding components can promote the availability of the ligand binding domains. The finding that ARH is required for both ligand binding and internalization of LDLR suggests that ARH plays a more complex role in LDL metabolism than originally envisioned.