A novel functional role of collagen glycosylation: interaction with the endocytic collagen receptor uparap/ENDO180

Abstract

Collagens make up the most abundant component of interstitial extracellular matrices and basement membranes. Collagen remodeling is a crucial process in many normal physiological events and in several pathological conditions. Some collagen subtypes contain specific carbohydrate side chains, the function of which is poorly known. The endocytic collagen receptor urokinase plasminogen activator receptor-associated protein (uPARAP)/Endo180 plays an important role in matrix remodeling through its ability to internalize collagen for lysosomal degradation. uPARAP/Endo180 is a member of the mannose receptor protein family. These proteins all include a fibronectin type II domain and a series of C-type lectin-like domains, of which only a minor part possess carbohydrate recognition activity. At least two of the family members, uPARAP/Endo180 and the mannose receptor, interact with collagens. The molecular basis for this interaction is known to involve the fibronectin type II domain but nothing is known about the function of the lectin domains in this respect. In this study, we have investigated a possible role of the single active lectin domain of uPARAP/Endo180 in the interaction with collagens. By expressing truncated recombinant uPARAP/Endo180 proteins and analyzing their interaction with collagens with high and low levels of glycosylation we demonstrated that this lectin domain interacts directly with glycosylated collagens. This interaction is functionally important because it was found to modulate the endocytic efficiency of the receptor toward highly glycosylated collagens such as basement membrane collagen IV. Surprisingly, this property was not shared by the mannose receptor, which internalized glycosylated collagens independently of its lectin function. This role of modulating its uptake efficiency by a specific receptor is a previously unrecognized function of collagen glycosylation.

FIGURE 1.

uPARAP recombinant proteins. A, domain composition of uPARAP and recombinant uPARAP constructs. The constructs D1–3, D1–4, and D1–10 comprise the first 3, 4, and 10 N-terminal domains, respectively, of the complete sequence of mature uPARAP. CysR, cysteine-rich domain; CTLD, C-type lectin-like domain with the indicated number. Note that only CTLD-2 contains an intact Ca²⁺-binding site and possesses carbohydrate binding activity. His tag and uPAR-DIII tag designate purification tags on the recombinant proteins; see “Experimental Procedures.” TM and Cyto, transmembrane and cytoplasmic regions of native uPARAP. B, SDS-PAGE analysis of purified recombinant uPARAP proteins. The calculated theoretical molecular masses are 38 (D1–3), 55 (D1–4), and 179 kDa (D1–10), respectively. The electrophoretic mobilities of molecular mass marker proteins are indicated to the left.

FIGURE 2.

SPR analysis of uPARAP interactions with solubilized collagens. Approximately equal amounts of D1–4 and D1–3 were immobilized in parallel flow channels of a BIAcore sensor chip, using near irreversible capture on mAb 2h9. Solubilized collagens (10 μg/ml) were then injected into parallel flow channels with captured D1–4 or D1–3. Collagen IV (A and C) was injected in the native state, whereas collagen I (B and D) was injected both in the form of native and heat-enatured protein. Blue and red curves depict the binding of native collagens to D1–4 and D1–3, respectively. Dark and light green curves represent the binding of heat-denatured collagen I to the same two uPARAP constructs. After the indicated 600-s phase of collagen injection, dissociation was allowed to proceed in running buffer for 240 s, followed by a 600-s injection phase with the indicated inhibitory reagents. Finally, the flow was shifted back to running buffer. The sensorgrams are shown after subtraction of buffer bulk effect (parallel flow cell without uPARAP) and a blank run with injection of running buffer instead of collagen material.

FIGURE 3.

ELISA analysis of uPARAP interactions with immobilized collagens. Heat-denatured collagen types IV and I were immobilized in ELISA wells as indicated below each panel. D1–4 (27 nm, A), D1–3 (27 nm, B), or D1–10 (55 nm, C) was added to each well. The binding was analyzed in assay buffer with 1 mm CaCl₂ and no further additions (gray columns), buffer with 3 mm EDTA (hatched gray columns), buffer with 1 mm CaCl₂ and 50 mm mannose (white columns), and buffer with 1 mm CaCl₂ and 50 mm α-Me-Gal (hatched white columns). Binding of the recombinant uPARAP proteins was detected with mAb 2h9 against uPARAP, followed by a secondary HRP-coupled rabbit anti-mouse antibody. Controls comprised wells without immobilized collagen (coating control, black columns) and wells without recombinant uPARAP (negative in all cases). Error bars represent S.D. of duplicate samples.

FIGURE 4.

Collagen deglycosylation affects the interaction with uPARAP. Collagen types IV (A, C, and E) and I (B, D, and F) were mock treated or deglycosylated with TFMS under anhydrous conditions. Following treatment, collagens were immobilized on ELISA plates, after which D1–4 (A and B), D1–3 (C and D), or D1–10 (E and F) were added. Analysis of binding and representation of results were performed as described in the legend to Fig. 3.

FIGURE 5.

Carbohydrate dependence of uPARAP-mediated collagen endocytosis. ¹²⁵I-Labeled collagens types IV (A and B) or I (C and D), holotransferrin (E and F),or mAb 2h9 (G and H) were added to uPARAP^+/+ and uPARAP^−/− fibroblasts, as indicated (A, C, E, and G). In a parallel experiment, the same ligands were added to MG63 cells (B, D, F, and H). The labeled proteins (133 ng/ml) were added in endocytosis assay buffer alone (no competitor), in buffer including 50 mm of the indicated monosaccharides, or in buffer including 10 μg/ml of mAb 5f4. After 4 h at 37 °C the cells were harvested, the intracellular fraction of cell samples was isolated, and the amount of internalized labeled protein was determined. Data are presented as radioactivity in the intracellular fraction of cells, relative to the total amount of radioactivity added. Error bars represent the S.D. of triplicate samples.

FIGURE 6.

Influence of carbohydrate on lysosomal collagen accumulation. uPARAP^+/+ and uPARAP^−/− fibroblasts were incubated with Alexa 488-labeled collagen IV (A) or Oregon Green-labeled gelatin (B) in the absence (no competitor) or presence of 50 mm mannose. Endocytosis was allowed to proceed during incubation overnight. Subsequently, cells were released, re-seeded on coverslips, fixed, and stained with a cell surface marker (anti-uPAR antibody followed by Cy3-conjugated secondary antibody; red fluorescence). Cell nuclei were stained with DAPI (blue). Cells were then examined by confocal microscopy. The left panels show the channel with green fluorescence alone (fluorescent collagen IV/gelatin), whereas the right panels show the merged images with green, red, and blue fluorescence. Note the vesicular accumulation of internalized collagen/gelatin.

FIGURE 7.

Collagen internalization in macrophages. Cultured macrophages were incubated with ¹²⁵I-labeled collagens type IV (A) or type I (B), mannose-BSA (C), or holotransferrin (D). Incubation was performed in the absence (no competitor) or presence of the following reagents: mannose (50 mm), a-MR pAb (10 μg/ml), or mAb 5f4 against uPARAP (5f4; 10 μg/ml). Following incubation for 4 h at 37 °C, the internalized fraction of labeled protein was determined. Data are presented as described in the legend to Fig. 5.