Research Article: Disulfide Bonds within the C2 Domain of RAGE Play Key Roles in Its Dimerization and Biogenesis

Date Published: December 17, 2012

Publisher: Public Library of Science

Author(s): Wen Wei, Leonie Lampe, Sungha Park, Bhavana S. Vangara, Geoffrey S. Waldo, Stephanie Cabantous, Sarah S. Subaran, Dongmei Yang, Edward G. Lakatta, Li Lin, Michael P. Bachmann.


The receptor for advanced glycation end products (RAGE) on the cell surface transmits inflammatory signals. A member of the immunoglobulin superfamily, RAGE possesses the V, C1, and C2 ectodomains that collectively constitute the receptor’s extracellular structure. However, the molecular mechanism of RAGE biogenesis remains unclear, impeding efforts to control RAGE signaling through cellular regulation.

We used co-immunoprecipitation and crossing-linking to study RAGE oligomerization and found that RAGE forms dimer-based oligomers. Via non-reducing SDS-polyacrylamide gel electrophoresis and mutagenesis, we found that cysteines 259 and 301 within the C2 domain form intermolecular disulfide bonds. Using a modified tripartite split GFP complementation strategy and confocal microscopy, we also found that RAGE dimerization occurs in the endoplasmic reticulum (ER), and that RAGE mutant molecules without the double disulfide bridges are unstable, and are subjected to the ER-associated degradation.

Disulfide bond-mediated RAGE dimerization in the ER is the critical step of RAGE biogenesis. Without formation of intermolecular disulfide bonds in the C2 region, RAGE fails to reach cell surface.

This is the first report of RAGE intermolecular disulfide bond.

Partial Text

RAGE was initially identified as a receptor for advanced glycation end products (AGE) [1], which are generated via non-enzymatic crosslinking of carbohydrates to proteins and other biological molecules [2]. Since then, other ligands for RAGE have been discovered including chromatin binding protein HMGB1 (high-mobility group box 1), s100 family of small calcium binding peptides, amyloid β protein, and phosphatidylserine [3]–[6], making RAGE one of the pattern recognizing receptors that participate in innate immunity [7], [8]. In addition, RAGE also functions as an adhesion molecule on the cell surface of neutrophiles, enhancing its recruitment to vascular endothelial cells during inflammation [9]. Similar to the Toll-like receptors (TLRs), engagement of RAGE by its ligands triggers several intracellular signaling programs including NF-κB and Erk1/2 transcription pathways, leading to inflammation [10]–[12]. However, unlike ligands for TLRs, which are mainly derived from exogenous pathogens, RAGE ligands are generated either endogenously, or are derived from the diet [13], [14]. RAGE-associated signaling, therefore, appears to participate mainly in pathophysiological events such as inflammation-related tissue remodeling and maladaptation, and has been implicated in atherogenesis, diabetes, and Alzheimer’s disease [11], [15], [16]. Recent reports have also shown that RAGE may be involved in immune defense mechanisms [17]–[19]. Despite its significant roles in pathogenesis and inflammation, the signaling mechanism of RAGE remains elusive, and cytosolic factors that relay the cell surface signals to specific cellular programs have not been elucidated [8], [12], [16].

Oligomerization of receptors may occur via two basic mechanisms: ligand-induced or constitutive. It has been unclear whether RAGE dimerization needs ligand, and our results showed that RAGE dimerization belongs to the second group (Figures 1 and 7). Using multilateral biochemical approaches, we demonstrated that RAGE in mammalian cells forms dimer-based oligomers (Figure 1). In crosslinking studies, sequential RAGE multimers were detected on SDS-PAGE (Figure 1B), whereas in native PAGE, sRAGE was found to form dimer-based oligomers (Figure 1C). This is due to random crosslinking by the crosslinkers within the oligomeric RAGE complex. Although it is still possible that the observed RAGE oligomers on the cell surface in our studies are formed under an overexpression condition in transfected cells, the relatively short spacer arm (11.4 Å, equal to 8 atoms) of the chosen crosslinker, BS3, renders crosslinking among non-complexed RAGE molecules in the crosslinking experiment unlikely. In addition, the expression of RAGE cysteine-to-serine mutants is about 2–4 times lower than the wild-type counterpart (Figures 2D and 8A), whereas the formation of dimers in RAGE carrying single cysteine-to-serine mutation is not affected, suggesting that there is no correlation between the formation of dimers/oligomers and the level of RAGE expression.