|Home | About | Journals | Submit | Contact Us | Français|
Profiling of the four known galactose-binding receptors in the C-type lectin family has been undertaken in parallel on a glycan array. The results are generally consistent with previous assays using various different formats, but they provide a direct comparison of the properties of the four receptors, revealing that they fall into two distinct groups. The major subunit of the rat asialoglycoprotein receptor and the rat Kupffer cell receptor show similar broad preference for GalNAc-terminated glycans, while the rat macrophage galactose lectin and the human scavenger receptor C-type lectin bind more restricted sets of glycans. Both of these receptors bind to Lewisx-type structures, but the macrophage galactose lectin also interacts strongly with bi-antennary galactose- and GalNAc-terminated glycans. Although the similar glycan-binding profiles for the asialoglycoprotein receptor and the Kupffer cell receptor might suggest that these receptors are functionally redundant, analysis of fibroblasts transfected with full-length Kupffer cell receptor reveals that they fail to endocytose glycosylated ligand.
The asialoglycoprotein receptor on hepatocytes is the prototype for the C-type lectin family. The receptor removes glycoproteins with terminal galactose or GalNAc residues from circulation (Ashwell and Harford, 1982). Three additional receptors in the C-type lectin family bind galactose- or GalNAc-containing ligands: the Kupffer cell receptor, the macrophage galactose lectin and the scavenger receptor C-type lectin (SRCL) (Hoyle and Hill, 1988; Ii et al., 1990; Nakamura et al., 2001). The potentially overlapping functions of these receptors remain poorly understood. It has been suggested that the Kupffer cell receptor might function in parallel to the asialoglycoprotein receptor, explaining failure of asialoglycoprotein receptor-deficient mice to accumulate uncleared asialoglycoproteins (Fadden et al., 2003; Ishibashi et al., 1994). Because the human macrophage galactose lectin binds strongly to the carcinoma-associated Tn antigen, GalNAc-Ser/Thr (Yamamoto et al., 1994), this receptor may have a role in recognition of malignant cells by tumoricidal macrophages. Knockout studies suggest that one of the two mouse macrophage galactose lectins is involved in clearance of apoptotic cells (Yuita et al., 2005).
The monosaccharide binding profiles of the asialoglycoprotein receptor and Kupffer cell receptor are very similar: both show a marked preference for GalNAc over galactose. In contrast, the rat macrophage galactose lectin exhibits no preference for GalNAc over galactose (Iobst and Drickamer, 1996). Binding of the asialoglycoprotein receptor (Lee et al., 1983; Park et al., 2003), the Kupffer cell receptor (Tiemeyer et al., 1992) and the macrophage galactose lectin (Yamamoto et al., 1994; Tsuiji et al., 2002; van Vliet et al., 2005) to distinct sets of oligosaccharide ligands has also been demonstrated. However, these results were obtained in a variety of different assay formats, so the results are not always directly comparable.
When the proteins are used in parallel to probe a recently developed glycan array, the results reveal that the receptors fall into two distinct subclasses that show selectivity for distinct classes of glycans. Further characterization of the intracellular trafficking capabilities of the Kupffer cell receptor indicates that it is unlikely to function in parallel with the asialoglycoprotein receptor in glycoprotein clearance.
In order to gain a broad view of the oligosaccharide-binding specificities of the galactose-binding receptors, fluorescently-labeled trimeric extracellular domain fragments from each protein were used to probe a glycan array consisting of biotinylated oligosaccharides presented on immobilized streptavidin (Figure 1). Data from a previous screening of the array with SRCL (Coombs et al., 2005) are shown for comparison. The results show that there is substantial overlap in binding specificities between the macrophage galactose lectin and SRCL. In contrast, the major subunit of the asialoglycoprotein receptor, rat hepatic lectin-1 (RHL-1), and the Kupffer cell receptor both bind to a much broader range of oligosaccharides. Thus, the galactose-binding receptors can be divided into two specificity groups, the first group containing human SRCL and the rat macrophage galactose lectin, and the second containing RHL-1 and the rat Kupffer cell receptor.
The glycan array profile for the macrophage galactose lectin reveals that this receptor binds two specific subsets of glycans with terminal galactose or GalNAc residues: branched glycans with two terminal galactose or GalNAc residues or smaller glycans in which fucose is adjacent to a terminal galactose or GalNAc residue. Amongst the fucose-containing ligands, Lewisx and Lewisa, as well as Lewisx in which the galactose is replaced with GalNAc, are all bound. The remaining highest affinity ligands are all biantennary structures containing terminal galactose or GalNAc residues, except for glycan 69, which contains adjacent terminal galactose and NeuAc residues. Based on the known structures of C-type CRDs with bound ligands, these results suggest that a galactose or GalNAc residue in the high affinity ligands makes primary interactions with the macrophage galactose lectin, but binding is enhanced by secondary interactions with adjacent terminal galactose, GalNAc, fucose or NeuAc residues. Binding of the macrophage galactose lectin is also observed to a few other galactose-terminated ligands including glycan 127 (Galβ1-3GlcNAcβ1-3Galβ1-4Glcβ). The extended structure of glycan 127 raises the possibility that binding may be due to the way the terminal galactose is presented on this glycan. Evidence that presentation can affect the binding of the receptor to glycans on the array is provided by the fact that Lewisa on a longer spacer arm (glycan 77) is bound more tightly than Lewisa on a short spacer arm (glycan 29).
SRCL exhibits an even narrower specificity than the macrophage galactose lectin, as it binds highly selectively to Lewisx, with weaker binding to Lewisa and Lewisx in which the galactose is replaced with GalNAc. SRCL has an absolute requirement for adjacent terminal galactose and fucose residues and does not bind galactose- or GalNAc-terminating glycans unless they form part of a Lewisx-type epitope.
Although the glycan array reveals that RHL-1 binds to a broad range of galactose- and GalNAc-terminated oligosaccharides over a range of affinities, some patterns in the type of ligand bound are evident:
First, the results for RHL-1 reflect the previously observed preferential binding of this receptor to GalNAc over galactose. Comparison of glycans 4 and 5 (α-/β-linked galactose) versus glycans 8 and 9 (α-/β-linked GalNAc), confirms the strong preference for GalNAc. RHL-1 also shows strong binding to blood group A antigen (GalNAcα1-3(Fucα1-2)Galβ), but almost no binding to the similar blood group B antigen (Galα1-3(Fucα1-2)Galβ) (glycans 23, 77 and 144), where the only difference is the presence of terminal galactose instead of GalNAc. However, preferential binding of GalNAc-terminating ligands by RHL-1 does not correlate with the ability of the receptor to bind a few monoantennary galactose-terminated ligands on the array, such as glycans 119, 120 and 145.
Second, modification of GalNAc by substitution on the 6 position is tolerated. These substituents include sulfate (glycan 63), Galβ1-4GlcNAc (glycan 78), GlcNAc (glycan 113), and sialic acid (glycans 126 and 127). The orientation of the GalNAc in the CRD, where the sugar co-ordinates Ca2+ via its 3- and 4-hydroxyls, would project any substituent at the 6 position away from the primary binding site. In some cases, 6-substitution enhances binding, suggesting that 6 substituents make secondary contacts with the receptor. Binding to similarly substituted galactose residues is not observed. The ability to accommodate 6-substitution of galactose with GlcNAc or 6-substitution of GalNAc with sialic acid has been previously observed (Lee, 1982; Park et al., 2003).
Third, binding to galactose- and GalNAc- terminated glycans is enhanced when there are multiple terminal residues. Previous studies on the asialoglycoprotein receptor show dramatic increases in binding in going from mono- to bi- to triantennary galactose-terminated structures (Rice et al., 1990). However, RHL-1 shows much weaker binding to biantennary glycans terminating in galactose (glycans 109 and 141) rather than GalNAc (glycans 101 and 161). The highest affinity ligand on the array for RHL-1 is a complex biantennary glycan, in which both antennae terminate in a GalNAcβ1-4GlcNAc structure. This glycan also contains outer arm α1-2 linked fucose on the sub-terminal GlcNAc and α1-2 linked fucose on the asparagine-linked GlcNAc (glycan 161), but the non-fucosylated form (glycan 101) also binds well to RHL-1. Unfortunately, more branched tri- and tetra-antennary glycans that have been shown to compete particularly effectively for binding to the asialoglycoprotein receptor (Lee et al., 1983) are not present on the array.
Finally, projection of glycans on a polypeptide core in the form of O-linked glycopeptides enhances binding. For example, binding of glycan 50, presented as an O-linked glycopeptide, is enhanced compared to the binding of glycan 119, the same oligosaccharide presented on a short spacer arm. The O-linked glycopeptides consist of oligosaccharides linked to an 18-amino acid peptide. The strong binding of galactose-and GalNAc-terminated O-linked glycopeptides suggests that O-linked glycans may be among the ligands bound by the receptor in vivo. Binding to a more limited set of O-linked glycans has been observed previously (Baenziger and Maynard, 1980).
The glycan array results indicate that the Kupffer cell receptor binds to essentially all of the same glycans as RHL-1, but the relative affinities of the two receptors differ in several respects. For example, compared to RHL-1, the Kupffer cell receptor shows less selectivity for GalNAc over galactose. Overall, the Kupffer cell receptor binds to very few glycans on the array substantially better than it binds to GalNAc. The two best ligands are biantennary GalNAc-terminated glycans, suggesting that for the Kupffer cell receptor the primary determinant of enhanced ligand binding is the formation of branched structures. In addition, the Kupffer cell receptor has enhanced affinity for the O-linked glycans on the array.
Like RHL-1, the Kupffer cell receptor binds to several GalNAc-oligosaccharides modified at the 6 position with sulfate, sialic acid or GlcNAc. However, the Kupffer cell receptor exhibits approximately equal or lower affinity for these glycans compared to those with terminal GalNAc presented alone, such as glycans 8, 9, and 22, implying the additional substituents are accommodated, but do not contribute to binding affinity. In contrast to RHL-1, the Kupffer cell receptor does show weak binding to some 6-substituted galactose-containing glycans (glycans 83 and 155).
The similarity between the ligand binding profiles of the Kupffer cell receptor and RHL-1 on the glycan array is consistent with the possibility that the Kupffer cell receptor might function in parallel with the asialoglycoprotein receptor as a second clearance receptor for desialylated ligands. The Kupffer cell receptor releases ligand at endosomal pH (Lehrman et al., 1986a; unpublished observations), as required for receptor-mediated endocytosis, but the ability of the receptor to mediate endocytosis has not been demonstrated. To examine the endocytic capabilities of the Kupffer cell receptor, the full-length protein was expressed in rat fibroblasts. These cells show no internalization and degradation of 125I-Gal-BSA, indicating that the receptor does not endocytose glycosylated ligands (Figure 2A). In a parallel assay, fibroblasts expressing SRCL exhibit 125I-Gal-BSA uptake and degradation. Expression was confirmed by blotting of cell extracts with antibodies raised against the CRD of the receptor (Figure 2B). Two broad protein bands at ~70 and ~62 kDa represent the full-length receptor and the receptor after proteolytic cleavage, both with varying extents of glycosylation. Quantification of blots revealed that up to 250,000 molecules were expressed per cell, which is at least as high as the levels of other C-type lectins that reveal endocytic activity when expressed in these cells (Stambach and Taylor, 2003; Coombs et al. 2005). When membrane-impermeable biotin was used to label cell surface proteins, blotting of biotinylated proteins isolated from the cells on avidin confirmed the presence of Kupffer cell receptor at the cell surface.
It is possible that the Kupffer cell receptor requires the presence of accessory proteins not present in the fibroblasts to enable it to endocytose. However, this would be atypical for the C-type lectin family of receptors, many of which function as endocytic receptors when transfected into the cell line used in these experiments (Coombs et al., 2005; Guo et al., 2004; Stambach and Taylor, 2003). Unlike the other receptors investigated here, all of which are endocytic, the cytoplasmic tail of the Kupffer cell receptor lacks a tyrosine-based internalization motif, although it does contain a di-leucine motif (Figure 3).
The results presented here demonstrate that the glycan array provides a powerful tool for distinguishing the binding characteristics of superficially similar receptors. While previous studies have provided evidence that the galactose-binding C-type lectins differ in their relative affinities for monosaccharide ligands such as galactose and GalNAc, the full extent of their ligand-binding selectivity is only evident when larger oligosaccharide ligands are tested. The array provides an efficient way to make such comparisons. The major outcome of these studies is the demonstration that the galactose-binding receptors can be divided into two groups based on the distinct classes of ligands that they bind.
The binding profile of the Kupffer cell receptor presented here confirms that the receptor is primarily a galactose- and GalNAc-binding receptor. Although the receptor was originally referred to as the ‘fucose receptor’, it has low affinity for the a-linked fucose commonly found in naturally occurring glycans (Lehrman et al., 1986a; Lehrman et al., 1986b). The absence of fucose-binding activity, combined with the failure of the receptor to mediate endocytosis, makes it unlikely that it has a role in clearing fucose-containing glycoproteins from circulation. Results from the present studies combined with earlier work reveal that there are multiple C-type lectins that can endocytose fucose-containing ligands. Glycans bearing fucose in conjunction with another terminal sugar such as sialic acid or galactose, creating epitopes such as Lewisx, sialyl-Lewisx or Lewisa, may be cleared by the asialoglycoprotein receptor and SRCL. Fucose is also bound by several receptors with mannose-type CRDs, including the mannose receptor, DC-SIGN and langerin (Taylor, 2001; Guo et al., 2004; Stambach and Taylor, 2003). This selection of C-type lectins that might be responsible for endocytosis of fucose-containing glycoproteins may explain earlier studies which suggested the presence of receptors for fucosylated ligands in addition to the asialoglycoprotein receptor and mannose receptor (Lehrman et al., 1986b).
It has been suggested that the Kupffer cell receptor may function as a particle receptor, capable of internalizing particles greater than 12 nm in diameter coated with galactose or GalNAc (Biessen et al., 1994). Such particles are much larger than Gal-BSA (~4 nm diameter). This type of phagocytic activity might require additional specialized cellular machinery found in Kupffer cells but not in fibroblasts. Alternatively, the Kupffer cell receptor might function as an adhesion receptor. In any case, the function is not essential in mammalian biology, because no human ortholog of the Kupffer cell receptor has been identified (Fadden et al., 2003).
Recent results for the human macrophage galactose lectin suggest that it differs significantly from the rat receptor studied here. An Fc-human macrophage galactose lectin chimera binds preferentially to terminal GalNAc (van Vliet et al., 2005) and shows no enhanced affinity for fucosylated glycans such as GalNAcβ1-4(Fucα1-3)GlcNAc or for biantennary over monoantennary glycans. Other studies suggest a preference of mouse macrophage galactose lectin-1 for Lewisx and of mouse macrophage galactose lectin-2 for GalNAc-terminated sugars (Tsuiji et al., 2002). It appears that rat macrophage galactose lectin is more closely related in ligand-binding characteristics to mouse macrophage galactose lectin-1 and the human receptor is more like mouse macrophage galactose lectin-2. Thus, rat and mouse both express macrophage galactose lectins capable of binding Lewisx but this binding activity is absent in humans. Similarly, mouse and rat hepatic lectin-1 have been found to differ in their selectivity for Siaa2-6GalNAcβ-terminated glycans compared to GalNAcβ-terminated glycans, even though they share very high amino acid sequence identity (Park and Baenziger, 2004). Thus, the results reported here, combined with other recent studies of ligand-binding specificity of galactose-binding receptors, suggest the need for caution in making assumptions about the characteristics of receptors from one species based on results from another.
Both RHL-1 and the Kupffer cell receptor bind strongly to the GalNAcβ1-4GlcNAc epitope on the glycan array. This epitope is common on glycoproteins of invertebrates, such as helminth parasites (Cummings and Nyame, 1999), but is also found on some mammalian glycoproteins. For example, GalNAcβ1-4GlcNAc is present on glycoprotein hormones, where it is capped by 4-O-sulphate. Recognition of GalNAcβ1-4GlcNAc by RHL-1 raises the possibility that this receptor may be involved in interactions with parasites. Alternatively, RHL-1 might serve as a backup for the mannose receptor in clearance of glycoprotein hormones (Fiete et al., 1991) on which GalNAc has become exposed due to loss of the 4-O-sulphate.
Trimeric extracellular domain fragments of the Kupffer cell receptor (KCR-B), RHL-1 and the macrophage galactose lectin (Fadden et al., 2003; Iobst and Drickamer, 1996) were expressed and purified by affinity chromatography on galactose-Sepharose as previously described. The major subunit of the asialoglycoprotein receptor, RHL-1, is capable of forming homotrimers and binding glycosylated ligands (Braiterman et al., 1989). The Kupffer cell receptor and the macrophage galactose lectin were dialyzed into 150 mM NaCl, 100 mM Na-HEPES, pH 7.8, 10 mM CaCl2, and RHL-1 was dialyzed into the same buffer but containing 500 mM NaCl. Aliquots (~2 ml) containing 200 - 400 μg of each protein were mixed with 5 aliquots of 10 μl of fluorescein isothiocyanate dissolved at 1 mg/ml in dimethyl sulfoxide and left to react overnight at 4 °C. Addition of an equal volume of loading buffer (1.25 M NaCl, 25 mM Tris-Cl, pH 7.8, 25 mM CaCl2) was followed by repurification on a 1-ml column of galactose-Sepharose. Columns were washed with loading buffer and proteins were eluted with 1.25 M NaCl, 25 mM Tris-Cl, pH 7.8, 2.5 mM EDTA. Fractions containing the purified fluorescein-labeled extracellular domain fragments were used to probe the glycan array developed by the Consortium for Functional Glycomics (www.functionalglycomics.org) following their standard protocol.
cDNA encoding the cytoplasmic tail and transmembrane region of the rat Kupffer cell receptor was amplified from rat liver cDNA (Marathon Ready cDNA, Clontech) using the primers 5′-aggacagaccttagaatcgtggggcaagag-3′ (forward) and 5′-gtgcagttgcttgggggccatctggaggaa-3′ (reverse). Following denaturation at 95 °C for 1 min, 40 cycles of 95 °C for 30 s, and 65 °C for 1 min were carried out. PCR products were cloned into the vector pCRII-TOPO (Invitrogen) and sequenced on an ABI prism 310 Genetic Analyzer. A fragment encoding the cytoplasmic tail and transmembrane domain was combined with the extracellular domain cDNA (Fadden et al., 2003) to create the full-length receptor cDNA which was inserted into the retroviral expression vector pVcos along with the neomycin resistance gene (Stambach and Taylor, 2003). The plasmid was transfected into Ψ-nCre packagingcells, which produce a pseudovirus that was used to infect Rat-6 fibroblasts. Lines stably expressing Kupffer cell receptor were selected using G418. Analysis of the uptake and degradation of 125I-Gal-BSA (Gal-BSA from E-Y Laboratories) by Kupffer cell receptor and SRCL expressing fibroblasts was performed as previously described (Stambach and Taylor, 2003).
For blotting experiments, cells were harvested by scraping, suspended in 10 volumes of 125 mM NaCl, 10 mM Tris-Cl, pH 7.5, 0.5% Triton X-100, sonicated briefly, incubated 1 h at 4 °C, and spun for 2 min at 18,000 x g. The supernatants were analyzed by SDS-polyacrylamide gel electrophoresis on a 15% gel, which was blotted onto nitrocellulose. The membrane was blocked in 5% BSA in Tris-buffered saline and incubated with affinity-purified antibody at 1:100 dilution followed by protein A-alkaline phosphatase conjugate at 1:5000 dilution. For surface labeling (Davis et al., 1998), cells at approximately 70% confluence were rinsed twice at 37 °C with phosphate-buffered saline containing 0.1 mM CaCl2 and 1 mM MgCl2 followed by incubation with 2 ml of a 1 mg/ml solution of sulfo-N-hydroxysuccinimdyl-biotin (Perbio) in the same buffer at 4 °C for 20 min, with gentle shaking. The biotin solution was removed and the plates were washed three times with phosphate-buffered saline containing 100 mM glycine and incubated for 45 min at 4 °C in the final wash. Plates were rinsed twice more and the cells were lysed by brief sonication in 1 ml of 100 mM Tris-Cl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100. Following incubation at 4 °C for 10 min, lysates were centrifuged at 18,000 x g for 5 min and incubated with 100 μl of a 50% suspension of avidin-conjugated beads (Perbio) for 1 h at room temperature. Beads were recovered by centrifugation at 18,000 x g for 5 min and washed five times with 1 ml of lysis buffer. Proteins were eluted in 50 μl of 2x sample buffer containing 2% β-mercaptoethanol at 100 °C for 5 min. Beads were removed by centrifugation and aliquots of the supernatant were analyzed on a 15% SDS-polyacrylamide gel that was blotted and stained as above.
We thank Richard Alvarez and Angela Lee of the Consortium for Functional Glycomics for screening the glycan array. This work was supported by grant 075565 from the Wellcome Trust, studentship 10845 from the Biotechnology and Biological Research Council and National Institutes of Health grant GM62116 to the Consortium for Functional Glycomics.