Systemic Glycosaminoglycan Clearance by HARE/Stabilin-2 Activates Intracellular Signaling

Abstract

Scavenger receptors perform essential functions, critical to maintaining mammalian physiologic homeostasis by continuously clearing vast numbers of biomolecules from blood, interstitial fluid and lymph. Stabilin-2 (Stab2) and the Hyaluronic Acid Receptor for Endocytosis (HARE), a proteolytic isoform of Stab2, are important scavenger receptors responsible for the specific binding and internalization (leading to degradation) of 22 discrete molecules, macromolecular complexes and cell types. One-third of these ligands are glycosaminoglycans (GAGs). Full-length Stab2, but not HARE, mediates efficient phagocytosis of apoptotic cells and bacteria via binding to target surface ligands. HARE, the C-terminal half of Stab2, mediates endocytosis of all the known soluble ligands. HA was the first ligand identified, in 1981, prior to receptor purification or cloning. Seven other GAG ligands were subsequently identified: heparin, dermatan sulfate, chondroitin and chondroitin sulfates A, C, D and E. Synthetic dextran sulfate is also a GAG mimic and ligand. HARE signaling during HA endocytosis was first discovered in 2008, and we now know that activation of HARE/Stab2 signaling is stimulated by receptor-mediated endocytosis or phagocytosis of many, but not all, of its ligands. This review focuses on the HARE-mediated GAG activation of intracellular signaling, particularly the Extracellular Signal-Regulated Kinase 1/2 pathway.

Keywords: ERK1/2; clathrin-coated pit; phagocytosis; receptor isoform; receptor recycling; receptor-mediated endocytosis; scavenger receptor; systemic clearance.

Figure 1

Glycosaminoglycan diversity. The schematic representation of glycosaminoglycans (GAG) structures shows that heparin (Hep), dermatan sulfate (DS), chondroitin (CDR), chondroitin sulfates A (CS-A), C (CS-C), D (CS-D), and E (CS-E), but not HA, are synthesized onto core proteins (dashed line), which can be membrane-bound or soluble, via a common linker sequence (Galactose-Galactose-Xylose; two circles-star) attached to Ser. All GAGs are made as linear polymers with repeating disaccharide units containing either N-acetylglucosamine (GlcNAc, blue square) or N-acetylgalactosamine (GalNAc, yellow square) and either glucuronic acid (GlcU, blue diamond) or iduronic acid (IdU, brown diamond). Arrows denote the polymer growing end, to which new sugars are added. Sulfation occurs during biosynthesis and different cell types can create different negative charge patterns (unique molecular signatures) that are specifically recognized by other soluble or membrane proteins (e.g., Hyaluronic Acid Receptor for Endocytosis (HARE)). GAGs built on a core protein are attached at their reducing end, and new sugars are added at the nonreducing end. In contrast, HA is not attached to or built onto a core protein. Rather, HA synthesis is initiated de novo and extended by HA synthase until the UDP-HA chain is released, at which point it cannot be rebound and extended further [2]. Additionally different is that HA synthesis occurs at the reducing end, which is still attached to UDP. Thus, the first sugars assembled remain at the nonreducing end (NR end) and the linkages in this unique region are chitin [GlcNAc(β1,4)_n] rather than HA. HA synthase initiates HA disaccharide synthesis only after first assembling a chitin-UDP (UDP-GlcNAc_3-4) primer for starting HA synthesis [2]. The primer then remains as a chitin cap at the NR end [3].

Figure 2

Human Stab2 domain organization. Full-length Stab2 is a type I membrane protein [17] of 2551 aa with a large N-terminal extracellular domain, a 92 aa HA-binding Link module (red), a 22 aa transmembrane domain (TMD, black), and a 72 aa C-terminal cytoplasmic domain (CD, gray). The ectodomain contains 7 Fasciclin-1 domains (F1–F7; green) and 4 larger EGF-like domains (blue; E1–E4) spanning the ectodomain length. Multiple binding sites for phosphatidylserine (PS) and for αMβ2 and αvβ5 integrins are within the E1-E4 and F1–F7 domains, respectively. PS binding sites enable receptor recognition of dying apoptotic cells [18] and integrin-binding sites enable receptor recognition of lymphocytes [19] in coordination with PS interactions [20]. A proteolytic cleavage site (arrow) that generates the 190 kDa HARE isoform at Ser¹¹³⁶ [21] is within Fasciclin-1 domain 4 (F4). This constitutive cleavage process gives two half-receptors: a soluble N-terminal half, whose fate is unknown, and the membrane-bound C-terminal half that is HARE (Section 5).

Figure 3

All GAG and other soluble ligand-binding domains are within the HARE portion of Stab2. The half-receptor HARE isoform (1416 aa) is the C-terminal half of Stab2 and contains the CD (gray), membrane (black) and Link (red) domains, as well as Fasciclin-1 domains F5–F57 and EGF-like domain E4 of the full-length protein (Figure 2). The N-terminal region of HARE is a truncated portion of the Stab2 E3 domain (pE3). All known GAG binding sites (yellow circles and gray ellipses) are within HARE [24,25]. The Link module enables binding to HA, CDR, CS-A, CS-C and CS-D (right yellow circle); these four CDR GAGs inhibit HA binding and Link module deletion eliminates all five binding activities [26]. A separate Hep binding site is on the N-terminal side of the Link domain. A third independent binding site for acetylated low-density lipoprotein (AcLDL; white circle), is assigned to the N-terminal side of the Hep site. Based on competition studies [26], DS binds to or occludes the Hep binding site and partially blocks both HA and AcLDL binding. DxS competes with both Hep and AcLDL binding but does not block HA binding. CS-E competes at a low level with both HA and Hep but does not affect AcLDL binding.

Figure 4

Cycles of HARE expression occur during rat embryonic liver development. HARE/Stab2 protein was detected between day 13 (13d) and day18 (18d) using monoclonal antibodies 30 and 154 raised against purified rat HARE (Magnification 100×). Commercial paraffin tissue blocks (Zyagen) were processed for immunocytochemical staining as described by McGary and co-workers [54,55]. Carl T. McGary, MD PhD is at University of Minnesota Medical School of Medicine-Duluth Campus, Department of Biomedical Sciences, SMED 227, 1035 University Drive, Duluth, MN 55812.

Figure 5

The Stab2/HARE CD has four potential endocytic targeting motifs. The 72 aa CD (Y²⁴⁸⁰–L²⁵⁵¹) next to the TMD contains four different AP-2/clathrin-mediated targeting motifs (boldface: M1-M4) that enable coated pit mediated endocytosis by both receptor isoforms, free or ligand-bound. Motifs in the top row (red) target HARE•HA complexes for endocytosis; M4 (black) is the only motif that does not mediate HA uptake. Motifs in the bottom row (blue) target HARE•Hep complexes for endocytosis; M2 (black) is the only motif that does not mediate Hep uptake. A tryptic peptide (underlined) contains phospho-Ser²⁴⁹⁷ (asterisk). Dot symbols (•) denote noncovalent bonding.

Figure 6

Ser²⁴⁹⁷ is phosphorylated in the HARE CD. (A) Immunopurified WT HARE cut from an SDS-PAGE gel was trypsin digested and samples were untreated (top) or treated with alkaline phosphatase (bottom). Mass spectrometry analyses using a Voyager Elite MALDI-TOF instrument in reflector mode revealed signals at m/z 2284.98 for a phospho-(Thr²³⁸⁹-R²⁵⁰⁸) peptide, which was reduced by phosphatase treatment, resulting in strong signals for the predicted free peptide (m/z 2205.02). (B) HARE was immunopurified from stable Flp-in 293 cells expressing WT (left) or S2497A (right) HARE, run on SDS-PAGE, electroblotted, stained with Pro-Q for phosphoryl groups [57,58] and HARE (arrows) was visualized by fluorescence microscopy (top) and then, after washing and staining, with Coomassie blue (bottom).

Figure 7

Cell signaling in response to HARE-mediated HA endocytosis is highly dependent on HA size. Specially prepared narrow-size range HAs of different mass (14–967 kDa; 0 is no HA), as indicated in each bar, were tested for the ability to activate signaling in cells expressing HARE, using a dual LUC reporter gene assay [81]. Identically treated empty-vector cells were at a ratio ≤1.0 (dotted line), indicating no NF-κB promoter-specific signaling (*, p < 0.05).

Figure 8

A model for HA-size dependent activation of HARE and ERK signaling. The scheme [80] leads to HARE dimer formation and activation of signal transduction (bottom middle). Symbols denote; membrane (horizontal black lines), CD (black rectangles), HA binding Link module (light gray indented ovals), ectodomain (N-terminal to Link module; long dark gray ovals) and HA size ranges of <40 kDa (small black dots), 40–400 kDa (small black ovals) and >400 kDa (large black ovals).

Figure 9

A HARE/Stab2 Tissue-Stress Sensor System. The scheme [84] shows parallel systemic loops in which intracellular signaling is stimulated by HARE-mediated endocytosis (left) of four key soluble ligands (HA, Hep, DS and AcLDL) or by Stab2-mediated phagocytosis of particulate debris or cells (right). These signaling cascades activate genes regulated by NF-κB promoters (red arrows) stimulating synthesis and secretion of pro-inflammatory cytokines such as Transforming Growth Factor-β (TGFβ) [18,50] that can direct new tissue responses (blue arrows) to the physiologic or homeostatic threats [85,86,87]. Soluble ligands shed from body tissues during infection, necrosis or biomatrix turnover are endocytosed by fixed SECs in the liver, lymph nodes and spleen (left). Particle phagocytosis by circulating macrophages (right) removes and destroys apoptotic red or white blood cells (red ovals), wound debris and bacterial cells (black ovals).