HARE-Mediated Endocytosis of Hyaluronan and Heparin Is Targeted by Different Subsets of Three Endocytic Motifs

The hyaluronan (HA) receptor for endocytosis (HARE) is a multifunctional recycling clearance receptor for 14 different ligands, including HA and heparin (Hep), which bind to discrete nonoverlapping sites. Four different functional endocytic motifs (M) in the cytoplasmic domain (CD) target coated pit mediated uptake: (YSYFRI2485 (M1), FQHF2495 (M2), NPLY2519 (M3), and DPF2534 (M4)). We previously found (Pandey et al. J. Biol. Chem. 283, 21453, 2008) that M1, M2, and M3 mediate endocytosis of HA. Here we assessed the ability of HARE variants with a single-motif deletion or containing only a single motif to endocytose HA or Hep. Single-motif deletion variants lacking M1, M3, or M4 (a different subset than involved in HA uptake) showed decreased Hep endocytosis, although M3 was the most active; the remaining redundant motifs did not compensate for loss of other motifs. Surprisingly, a HARE CD variant with only M3 internalized both HA and Hep, whereas variants with either M2 or M4 alone did not endocytose either ligand. Internalization of HA and Hep by HARE CD mutants was dynamin-dependent and was inhibited by hyperosmolarity, confirming clathrin-mediated endocytosis. The results indicate a complicated relationship among multiple CD motifs that target coated pit uptake and a more fundamental role for motif M3.


Introduction
Stabilin-2 (Stab2) and HARE (half-length Stab2) function as primary scavenger receptors for the systemic clearance from lymph and blood of hyaluronan (HA) [1], heparin (Hep), and 12 other functionally and structurally distinct ligands [2][3][4][5][6]. HA turnover and catabolism by HARE have been studied extensively for decades [1,7,8] and the responsible receptor was molecularly identified >15 years ago after it was purified and cloned [5,6,9,10]. HARE endocytosis of HA occurs over a broad range of sizes from ∼2.5 kDa to >MDa [11]. Hep is cleared from the body by two different mechanisms: larger Hep is rapidly cleared from blood in a high-affinity saturable binding mechanism by HARE/Stab2 in liver sinusoidal endothelial cells [12,13], whereas low mass Hep is primarily cleared by kidney [14] in a nonsaturable renal excretion mechanism [15].
HA and Hep have distinct binding sites within the HARE ectodomain and neither ligand competes for the binding and endocytosis of the other [2]. Several articles in this special issue summarize the many functions of HA. As with HA, the biological and clinical activities of Hep have been studied for decades, and Hep is the most highly prescribed drug in the USA (e.g., for preventing or treating thromboembolic diseases and postsurgery clotting) [16]. HA is synthesized by many cell types and is the longest (up to 5 × 10 4 sugars) and only unsulfated glycosaminoglycan. In contrast, Hep is synthesized by mast cells as a serglycin proteoglycan with much shorter polysaccharide chains, <50 sugars [17,18]. Hep is the most anionic glycosaminoglycan, due to extensively sulfated disaccharide isomers and binds to many different matrix, soluble, and cell surface proteins, including growth factors [19]. Hep also functions as a coreceptor or anticoagulant agent [20,21].
Many receptors require bound ligand in order to interact with adapter proteins and then be targeted to coated pits [22]. In contrast, HARE and other constitutively recycling 2 International Journal of Cell Biology clearance receptors (e.g., asialoglycoprotein and LDL receptors) are continuously targeted to coated pits, internalized, and recycled back to the cell surface whether bound to ligand or not. Endocytic receptors often contain a tyrosine-based motif (e.g., YXX or NPXY, where X is any amino acid and is a hydrophobic residue) or a dileucine motif (D/EXXXLL/I) involved in clathrin-mediated endocytosis [23][24][25][26]. YXX and dileucine motifs interact with AP-2 adaptor complexes, whereas NPXY motifs interact with AP-2 and other adaptor proteins such as Dab2 and ARH [25]. The adaptor protein GULP is required for Stab2-mediated phagocytosis of aged (apoptotic) red blood cells, by recognition of a phospho-Tyr in the HARE CD [27].
The 72-amino acid C-terminal tail of HARE (Y 2480 -L 2551 in full-length Stab2) contains at least four endocytic motifs, an unusually high number: YSYFRI 2485 (M1), FQHF 2495 (M2), NPLY 2519 (M3), and DPF 2534 (M4). Surprisingly, three of these four motifs (M1, M2, and M3) mediate endocytosis of HA, with M3 being the most active motif [26]. All three motifs participate in total coated pit targeting of HARE-HA complexes, and no single motif is required for uptake if the other functional motifs are present. The HARE CD motif network responsible for targeting the receptor to coated pits reflect either a very high level of redundancy or the presence of multiple distinct endocytic and signaling pathways. Our objective here was to identify the endocytic motifs responsible for HARE-Hep endocytosis. The results indicate that M1, M3, and M4 are utilized for Hep endocytosis, which is a different subset of three motifs compared to that used for HA, and that in the absence of other motifs, only M3 is able to mediate the endocytosis of both HA and Hep; HARE containing M2 or M4 alone did not mediate HA or Hep endocytosis.

Reagents, Buffers, Stable Cell Lines, and Normalization.
Flp-In 293 cells, FBS, DMEM, hygromycin B, Zeocin, Lipofectamine 2000, glutamate, plasmid expression vectors, and super-competent TOP10 Escherichia coli were from In-Vitrogen (Carlsbad, CA). Stable cells expressing HARE and HARE-mutants were generated as described previously [26,28]. Hep was from Celsus (Cincinnati, OH) or Sigma-Aldrich (St. Louis, MO). Low endotoxin HA, made by bacterial fermentation, was from Genzyme Corp. (Cambridge, MA). Dynasore was from Sigma-Aldrich. Sodium 125 I-iodide (100 mCi/mL; specific activity of >0.6 TBq/mg) in NaOH and PD-10 columns were from GE/Amersham Biosciences (Piscataway, NJ). Streptavidin (SA) was from Pierce (Rockford, IL). Preparation and quantification of biotinylated and iodinated ligands and the compositions of other buffers were described previously [13,29,30]. Other materials, reagents, and kits were obtained as described [26] or were from Sigma-Aldrich. HARE cDNA constructs and vectors for creation of stably transfected Flp-In 293 cell lines expressing wildtype (WT) HARE or HARE mutants with single or multiple endocytic motif deletions or site-specific substitutions were described previously [26,28]. All recombinant HARE proteins contain C-terminal V5 and His 6 epitope tags. Endocytosis Medium is DMEM with 0.05% BSA. In all experiments, the results among different HARE-expressing cell lines were normalized for HARE expression based on Western blot quantification of equal lysate protein samples [26]. Binding or endocytosis result values are expressed as the mean ± SE fmol/10 6 cells/HARE.

125 I-SA•b-Hep Binding and Endocytosis
Assays. Cells expressing WT HARE, HARE-mutants, or EV were grown in DMEM with 8% FBS and 100 g/mL hygromycin B (complete medium) in 12-well tissue culture plates for at least 2 days prior to experiments. They were processed for binding or endocytosis assays at 90-95% confluence. Radiolabeled 125 I-SA•b-Hep or 125 I-SA•b-HA complexes were prepared [13] using a 2 : 1 molar ratio of b-GAG : 125 I-SA and were incubated in 0.5 mL of Endocytosis Medium for 1 h on a rotary mixer at 22 ∘ C just prior to the experiment. For nonspecific binding controls, the same amounts of 125 I-SA and free biotin were used. 125 I-Complexes were diluted in Endocytosis Medium to the final concentrations indicated. Cells were washed with Hanks' balanced salts solution and incubated at 37 ∘ C for 1 h with Endocytosis Medium (no serum) to allow HARE-mediated internalization of any bound serum glycosaminoglycans. The medium was aspirated and replaced with Endocytosis Medium containing 50 nM preformed complexes of 125 I-SA with b-Hep or b-HA with or without a 50-fold excess of unlabeled ligand as competitor. The cells were then incubated either at 37 ∘ C for 1, 2, or 4 h to assess the rate of endocytosis or at 4 ∘ C for 2 h with or without 0.055% digitonin to assess total cellular or surface binding, respectively [31].
Nonspecific binding of 125 I-SA was also assessed in parallel samples by incubating cells with 125 I-SA•biotin complexes. The medium was removed by aspiration, and cells were washed three times (2 mL each) with cold Hanks' balanced salts solution to remove unbound ligand and solubilized in 1 mL 0.3 N NaOH. Radioactivity was measured using a Packard Cobra II gamma counter and lysate protein content was determined by the method of Bradford [32] using bovine serum albumin as standard. For each cell line, including EV, the binding of 125 I-SA•biotin was subtracted from the binding of 125 I-SA•b-ligand to correct for nonspecific binding of SA. (2 mL each) with cold Hanks' balanced salts solution to remove unbound 125 I-ligand and processed as noted above.

Statistical Analysis.
At least 2-4 independent experiments were performed in triplicate ( = 6-12) and combined data are presented as the mean ± SE. All regression lines had correlation coefficients ≥0.97 and experimental and control results were compared by unpaired Student's -tests using SigmaPlot v10 (Systat Software, Inc., Point Richmond, CA). Values of < 0.05 were considered statistically significant.

Results
HARE and Stab2 are scavenger receptors that bind and clear 14 different ligands, including seven glycosaminoglycans, from lymph and blood. We designate the full-length 315 kDa protein as Stab2 and HARE as the 190 kDa isoform generated by an unknown proteolytic mechanism [33]. Both HARE/Stab2 are the main systemic clearance receptors for HA and presumably Hep, in all mammals studied [34][35][36][37]. HARE is the predominant Stab2-related protein expressed in sinusoidal endothelial cells of lymph node and liver, the main systemic clearance tissues [10,38]. Although both HA and Hep are anionic glycosaminoglycans, they bind to discrete and nonoverlapping sites in the HARE ectodomain [2]. HA binding requires the Link domain, which it likely binds to directly, whereas Hep binds to an uncharacterized site and binding is unaffected by deletion of the Link domain [2]. Since, HA and Hep bind to different sites, we wanted to determine if HARE utilizes the same subset of three redundantly functional endocytic motifs for Hep endocytosis as found previously for HA endocytosis [26]. Most of the CD mutants used here had been characterized previously for their HARE-mediated HA binding and uptake ability. Two additional single-motif containing CD mutants were created for the present study (+M2 and +M4) to obtain a set of HARE CD variants expressing only one of the four motifs (e.g., ΔM1M2M4 = +M3); the panel of CD mutants used is shown schematically in Figure 1. We were not successful in creating cell lines expressing only motif M1.

Cell Surface and Total Hep Binding Are Similar among Multiple HARE CD Mutants.
To understand further the importance of human HARE having the ability to internalize both HA and Hep, we wanted to determine which of the four CD endocytic motifs were functional for each ligand. We previously found that HARE expression levels, as well as HA binding to surface and intracellular HARE, were similar to WT in a panel of stable Flp-In 293 cell lines expressing various CD-mutants [26]. Here we used a set of variant cell lines, expressing HARE mutants that were either single-motif deletions or containing a single-motif (i.e., three motifs deleted). To determine whether the cellular HARE distribution of Hep binding was affected in any of the variants, we compared 125 I-SA•b-Hep binding at 4 ∘ C to cell surface or total cellular HARE (cell surface and intracellular receptors) in the various HARE CD-mutant cells. Total and surface binding were monitored in the presence or absence of digitonin, respectively, under conditions that selectively permeabilize endocytic, but not nuclear, mitochondrial or lysosomal compartments [31,39]. Since Hep nonspecifically binds to many cell surface and intracellular proteins, the binding of Hep by EV cells is higher relative to WT cells than the nonspecific binding of HA [13,26]. Only small amounts of 125 I-SA•biotin (e.g., <1% of 125 I-SA•b-Hep values) bound to cells and this did not increase with time [13]. In contrast, 125 I-SA•b-Hep uptake was time-dependent and linear over 4 h, as in Figure 3. Cell surface (Figures 2(a), 2(c), and 2(e)) and total (Figures 2(b), 2(d), and 2(f)) 125 I-SA•b-Hep binding to WT or HARE CD-mutant cells were 2-3 times greater than to EV cells.As expected, the distribution of 125 I-SA•b-Hep binding sites between surface and internal was similar to that for HA binding in WT and the HARE CD mutants [26]. HARE is a constitutively active receptor involved in continuous and repeated cycles of ligand internalization and the HARE   recycling time of 7-9 min [28,40] is similar to that of other constitutively active recycling receptors [13,41,42]. The majority of recycling receptors, including HARE [28,33], are localized in intracellular endocytic and recycling compartments. Thus, Hep total binding (surface and internal) by WT or CD-mutant cells was much greater than surface binding, as expected. Among the group of nine CD-mutants examined, there were no significant differences in Hep surface binding (Figure 2 top panels), confirming that deletion of one or more endocytic motifs did not alter the dynamic ongoing movement of HARE to and from the cell surface; the steadystate surface receptor pool was similar among a set of HARE variants. Total Hep binding was identical to WT among the set of nine HARE mutants except for ΔM1 and ΔM3 (Figure 2(b)), which were significantly higher ( < 0.05).

Internalization of 125 I-SA•b-Hep by HARE Single-Motif Deletion Mutants.
To assess the contributions of the various endocytic motifs to the kinetics of Hep endocytosis, cells expressing WT, HARE-mutants, or EV were incubated at 37 ∘ C with 125 I-SA•b-Hep for different times (Figure 3). Partial impairment of 125 I-SA•b-Hep endocytosis relative to WT cells occurred in ΔM1 or ΔM4 cells; HARE-specific uptake (WT uptake minus EV uptake) was 65% and 68% of WT rates, respectively, for ΔM1 or ΔM4 cells (Figure 3(a)). Cells expressing the ΔM3 mutant showed even greater impairment of HARE-specific Hep endocytosis (35% of WT), indicating that M3 is responsible for more targeting to coated pits than M1 or M4. Surprisingly, ΔM2 cells did not show a defect in HARE-specific Hep endocytosis, but rather a 35% increase in HARE-specific uptake compared to WT as though M2 itself International Journal of Cell Biology

The Role of Y 2519 in HARE-Mediated Internalization of 125 I-SA•b-Hep.
Since it is well known that phosphorylated Tyr residues in NPXY motifs are important in signaling pathways [43,44], we wanted to identify further the importance of Y 2519 in NPLY 2519 for targeting HARE-ligand complexes to coated pits. We used two CD-mutant HARE cell lines, one with only a Y-to-A substitution, WT (Y2519A), and the other with the same substitution in the ΔM1M2M4 background, +M3 (Y2519A). WT (Y2519A) cells showed no significant defect in 125 I-SA•b-Hep endocytosis (95% of specific WT uptake), whereas +M3 (Y2519A) cells were identical to EV cells, showing complete impairment of Hep endocytosis (Figure 3(b)). The results show that Tyr in the HARE NPLY motif is critical for targeting to coated pits by +M3 cells, but it is not required if HARE has functional M2 and M4 motifs; these motifs compensate for a potential defect in M3. The data are consistent with either the ability of NPLA 2519 to retain targeting function in the presence of, but not the absence of, the two other Hep•HARE targeting motifs or the ability of M1 and M4 to compensate for NPLA 2519 and perform the targeting function.  (Figure 3(a)), we expected +M2 cells to be defective in Hep endocytosis. However, the inability of +M4 cells to take up Hep was unexpected, since this motif is functional in WT cells. The unexpected functional differences among the single-motif containing HARE variants are not ligand specific, as the same pattern was observed when HA endocytosis was examined (Figure 4(b)). Again, +M2 cells (expected to be active; Figure 4(a)) or +M4 cells (expected to be inactive) were identical to EV cells; they were both unable to internalize HA, indicating the lack of coated pit targeting and uptake. In contrast +M3 cells showed ∼60% of the endocytic capability of WT cells, a very similar result to that for Hep uptake (Figure 4(a)). Thus, although both HA (data not shown) and Hep (Figures 2(c) and 2(d)) bind equally well to HARE variants with only a single M2, M3 or M4 motif and these variants show similar surfaceinternal distributions (Figure 2), only M3 by itself is able to target HARE-ligand complexes to coated pits and mediate efficient uptake. The quantitative and relative rates of 125 I-SA•b-Hep endocytosis of the various HARE CD mutants are summarized and compared to the values for HA uptake [26] in Table 1.

Internalization of Hep or HA by HARE CD Mutants Is
Inhibited by Hyperosmolarity. The unexpected behavior of HARE single-motif containing variants prompted us to verify that the various HARE CD variants mediate endocytosis using a clathrin coated pit pathway, as shown previously for native and recombinant WT HARE [28,40,45]. Hyperosmolar conditions inhibit clathrin assembly into coated pits and, thus, clathrin-dependent internalization of many plasma membrane receptors [46][47][48]. To verify further that ligand uptake by the various HARE CD-mutant cells is clathrin-dependent, we assessed the effects of hyperosmolarity on endocytosis using medium containing 0.45 M sucrose. Hyperosmolar sucrose treatment blocked internalization of HA by ∼77% in WT cells compared to control (untreated) cells and by 40-70% in the single-motif deletion HARE CD mutants ( Figure 5(a)). Similar results were obtained for the effects of hyperosmolarity on Hep uptake ( Figure 5(b)). Overall, the results confirm that Hep and HA internalization by the various HARE CD mutants occurs via clathrin-coated pit pathways.

Inhibition of Dynamin Activity Blocks HA or Hep Endocytosis by HARE.
Many endocytic pathways in mammalian cells, including those involving clathrin-coated pits, phagocytosis, and caveolae, require the molecular motor protein dynamin for vesicle formation [49]. Dynasore is a small cell-permeable chemical that specifically inhibits the GTPase activity of dynamin and interferes with dynamin-dependent endocytic pathways [50]. Although not absolutely specific for clathrin-mediated uptake, dynasore should inhibit HAREmediated uptake that occurs via coated pits. Dynasore treatment significantly inhibited HA uptake by HARE CD variants with alterations in the motif subset involved in HA uptake, compared to DMSO-alone controls (Figure 6(a)).
Similarly, dynasore inhibited Hep endocytosis by ∼70-85% in several HARE CD mutants of the motif subset involved in Hep uptake (Figure 6(b)). As expected, the above dynasore and hyperosmolar sensitivity results indicate that the various HARE CD mutants mediate HARE-Hep and HARE-HA endocytosis by dynamin-dependent clathrin-coated pit pathways.

Discussion
Full-length Stab2 and 190 kDa HARE (the C-terminal half of Stab2) are the primary scavenger receptors for systemic clearance of multiple structurally distinct ligands (most of which are derived from tissue biomatrix degradation or cell debris) including HA, Hep, chondroitin sulfates (types A, C, D, and E), dermatan sulfate, advanced glycation end products, acetylated or oxidized LDL, collagen N-terminal propeptides, and M 2 and 5 5 integrins [2,4,13,28,45,[51][52][53][54]. In addition, apoptotic cells and debris are rapidly cleared from blood and lymph by macrophages and sinusoidal endothelial cells via HARE-mediated binding to phosphatidylserine and then phagocytosis [3,55]. Both functional receptor isoforms are expressed in sinusoidal endothelial cells of liver, lymph node, spleen, and bone marrow with the 190 kDa HARE being the predominant species [10]. Both are also expressed in some specialized tissues, such as corneal and lens epithelium, heart valve mesenchymal cells, epithelial cells in renal papillae, and oviduct [5]. HARE-mediated endocytosis of HA [56] or Hep [57] activates intracellular signaling leading to activation of ERK1/2 and NF-B stimulation of gene expression. Uptake of the HARE ligands dermatan sulfate and acetylated LDL [57] and phosphatidylserine [55] also activated NF-B mediated gene expression, whereas chondroitin sulfates types A, C, D, and E did not. Although all 9 of these ligands are effectively International Journal of Cell Biology 7  [58] that responds to the amounts and ratios of multiple biomatrix ligands via a signal transduction network that leads to the secretion of TGF- [59] and other, yet to be identified, factors such as pro-or anti-inflammatory cytokines. The physiological importance of HARE/Stab2 for HA homeostasis was verified in Stab2 knockout mice, which have impaired systemic clearance of HA resulting in abnormally high circulating HA levels [60]. Cytokine profiles have not yet been determined in these animals. HARE may also act as a homing receptor for human prostate tumor cells, allowing metastasis to lymph nodes [61] and likely other HARE-expressing tissues, such as liver and bone marrow. Metastasis was >95% blocked by treating mice with a specific anti-HARE HA-blocking antibody.
It is well established that HARE-HA uptake is clathrin coated pit-mediated [40,45], and this was confirmed for Hep uptake in various HARE CD mutants based on the inhibition of ligand uptake in cells treated with either the dynamin inhibitor dynasore or sucrose, under hyperosmolar conditions (Figures 5 and 6). Many endocytic receptors utilize a single CD motif for endocytosis, such as YXX (e.g., transferrin and asialoglycoprotein receptors [62,63]) or NPXY (e.g., LDL, insulin, and EGF receptors [64,65]). To our knowledge few other, if any, receptors contain multiple different endocytic motifs that are cooperatively utilized for endocytosis. For example, LDL receptor-related protein contains five possible endocytic motifs (1, YXX ; 2, NPXY; and 2, LL), but only YXX is utilized as the dominant endocytic signal [66]. HARE is unusual and possibly unique in having four different functional endocytic motifs and in utilizing subsets of three motifs for the uptake of HA and Hep.
An unexpected finding in this study was that HARE utilizes a different subset of three motifs for the endocytosis of Hep compared to HA (Figure 7). Three of the four endocytic motifs in the HARE CD (M1 (YSYFRI 2485 ), M3 (NPLY 2519 ), and M4 (DPF 2534 )) are utilized for Hep internalization. In contrast, a different subset of three motifs (M1, M2 (FQHF 2595 ), and M3) is utilized for HA endocytosis [26]. This result and the previous finding that Hep and HA bind to independent nonoverlapping sites in the HARE ectodomain [2] indicate that the binding of HA or Hep may create distinct conformational states within the intracellular CD that promoted differential recognition of endocytic motifs M2 and M4 by the relevant adaptor proteins. Different conformational or multimeric states of the intracellular CD could favor efficient binding of particular adaptor proteins to specific motifs. The CD conformation of HARE-HA complexes may allow M2 recognition by an appropriate adaptor protein, but not M4 recognition, whereas the CD conformation of HARE-Hep complexes may allow M4 recognition by an appropriate adaptor protein, but not M2 recognition. Consistent with the idea that binding in the ectodomain may influence intracellular signaling, Hep does not bind within the HA-binding HARE Link domain, whereas both HA and Hep bind to the Link domain of TSG6 [67].
The consequences of this differential mechanism of Hep versus HA endocytosis are unknown but might include different downstream signaling events or trafficking outcomes for a portion of the internalized pool of Hep or HA. The impairment of HA or Hep endocytosis due to a single-motif deletion was not compensated by the other two functional motifs, indicating that each motif mediates targeting and endocytosis by a distinct independent and saturable pathway, M3 is highlighted in red to indicate its special significance as the only motif of the three we were able to test (M2, M3, and M4) that enabled HARE to target coated pit mediated endocytosis of both HA and Hep.
perhaps through a subset of coated pits. If true, this has significant implications for possible independent concurrent signaling pathways mediated by different HARE-ligand complexes. One difference between the signaling stimulated by HA uptake and the signaling stimulated by Hep uptake is that HA signaling is very size-dependent. Only HA sizes between 40 kDa and 400 kDa are able to activate HARE-mediated ERK1/2 and NF-B signaling pathways; smaller or larger HA is endocytosed but does not activate signaling [11]. HARE-Hep activation of both signaling pathways is independent of Hep size [57]. Perhaps the use of different motif subsets for HA and Hep uptake is related to the mechanism by which HA size dependence is achieved during internalization of HARE-HA complexes.
The results indicate that each subset of three motifs participates in the total uptake of HA or Hep, but that the nature of their cooperation is unequal and complicated. Although the loss of only M3 (in ΔM3 cells) impaired Hep or HA endocytosis by ∼40%, indicating that M3 shares onethird of the Hep uptake burden, the loss of the other two Hep uptake motifs M1 and M4 (in +M3 cells) only decreased endocytosis by the same amount, 35%. This was a surprising functional difference among the three motifs, since they appear to function together when all are present, but only one can function if alone. Hep and HA endocytosis were completely eliminated in +M3 (Y2519A) cells, showing that Tyr 2519 is important for the endocytic process mediated by M3 alone. In WT (Y2519A) cells there was essentially no effect on uptake of either ligand. However, ongoing studies show that WT (Y2519A) cells are completely unable to activate NF-B during uptake of HA, Hep, dermatan sulfate, or acetylated LDL [57]. Thus, Tyr 2519 is critical for signaling to downstream effectors, when the receptor is endocytosing loaded cargo, but it is not needed for just cargo endocytosis alone.
Further studies are required to define the adaptor proteins (e.g., Gulp or AP-2) that interact with the four endocytic motifs in the HARE CD and to understand the biological relevance of the complex coated pit targeting network and how it is coupled to signal transduction for a subset of internalized ligands.