Human and Murine Interleukin 23 Receptors Are Novel Substrates for A Disintegrin and Metalloproteases ADAM10 and ADAM17*

IL-23 (interleukin 23) regulates immune responses against pathogens and plays a major role in the differentiation and maintenance of TH17 cells and the development of autoimmune diseases and cancer. The IL-23 receptor (IL-23R) complex consists of the unique IL-23R and the common IL-12 receptor β1 (IL-12Rβ1). Differential splicing generates antagonistic soluble IL-23R (sIL-23R) variants, which might limit IL-23-mediated immune responses. Here, ectodomain shedding of human and murine IL-23R was identified as an alternative pathway for the generation of sIL-23R. Importantly, proteolytically released sIL-23R has IL-23 binding activity. Shedding of IL-23R was induced by stimulation with the phorbol ester phorbol 12-myristate 13-acetate (PMA), but not by ionomycin. PMA-induced shedding was abrogated by an ADAM (A disintegrin and metalloprotease) 10 and 17 selective inhibitor, but not by an ADAM10 selective inhibitor. ADAM17-deficient but not ADAM10-deficient HEK293 cells failed to shed IL-23R after PMA stimulation, demonstrating that ADAM17 but not ADAM10 cleaves the IL-23R. Constitutive shedding was, however, inhibited by an ADAM10 selective inhibitor. Using deletions and specific amino acid residue exchanges, we identified critical determinants of ectodomain shedding within the stalk region of the IL-23R. Finally, interaction studies identified domains 1 and 3 of the IL-23R as the main ADAM17 binding sites. In summary, we describe human and murine IL-23R as novel targets for protein ectodomain shedding by ADAM10 and ADAM17.

The pro-inflammatory cytokine IL-23 is composed of the IL-12p40 subunit and p19 (1). IL-23 is a key factor for the development of T H 17 cells (2), which control antimicrobial and antifungal responses, but is also critically involved in the pathogenesis of chronic inflammatory disorders (3). The receptor complex is composed of the common IL-12 receptor ␤1 (IL-12 ␤1), 2 shared with IL-12, and the unique IL-23 receptor (4,5).
Single nucleotide polymorphisms within the IL23R gene were associated with various autoimmune diseases and the risk to develop cancer (1). Upon recruitment of the receptors by IL-23, which results in a noncanonical receptor complex formation (6), signaling is initiated by activation of associated Tyk2 (tyrosine kinase 2) and Jak2 (Janus kinase 2), which phosphorylate predominantly STAT3, and to a lesser extent STAT1, STAT4, and STAT5 (5). Recently, a noncanonical tyrosine-independent STAT3 activation site within the IL-23R was identified (7). In addition to STAT proteins PI3K, MAPK and NF-B signaling pathways were activated (7,28).
Ectodomain shedding of membrane-bound proteins leads to receptor protein down-regulation on the cell surface and the generation of soluble protein ectodomains with agonistic or antagonistic properties. Members of the ADAM (A disintegrin and metalloprotease) gene family are major ectodomain shedding proteinases. ADAM17 and its close relative ADAM10 are the major sheddases of this family, (8), with extensive overlap and compensation for several substrates, including EGF receptor ligands, TNF, TNF receptor, and IL-6R (9,10). Activation of ADAM proteases is achieved by different stimuli including phorbol ester (phorbol-12-myristate-13-acetate (PMA)), ionomycin, ligands of G-protein-coupled receptors, ATP, bacterial toxins, bacterial metalloproteinases, and apoptosis (8). For some ADAM target proteins such as Notch, induction of intracellular signaling by the remaining intracellular domain cleavage product has been described (11). Previously, it was shown that alternative splicing of IL-23R result in a series of truncated soluble IL-23R proteins (12)(13)(14).
Here, we discovered murine and human IL-23R as novel substrates of ADAM10 and ADAM17, resulting in the release of soluble IL-23R proteins, which retained their ability to bind to IL-23. Distinct areas within the murine and human IL-23R, which are important for ectodomain shedding, were identified in murine and human IL-23R. Immunoprecipitation analysis revealed domains 1 and 3 of IL-23R as critical ADAM17 interaction sites. Thus, we propose that ectodomain shedding is a second mechanism that contributes to the generation of soluble IL-23R variants.
Construction of Expression Plasmids-Plasmids containing the coding sequences for hIL-23R and hIL-12R␤1 were obtained from Esther van de Vosse (Leiden University Medical Center, Leiden, The Netherlands) (17,18) and cloned into p409 (19) for transient transfection of HEK293 cells and into the retroviral plasmids pMOWS-puro or pMOWS-hygro (20) for generation of stable Ba/F3-gp130 cell lines. The eukaryotic expression plasmid for mIL-23R was described (7) and used as template for cloning of stalk deletion variants. Human and murin IL-23R deletion cDNA variants and the IL-23R-(stalk hIL-6R)-chimeric cDNA were constructed using standard cloning procedures via splicing by overlapping extension PCRs and subcloned into p409. As mentioned above, cDNAs for hIL-23R variants were additionally subcloned into the pMOWSpuro plasmid and hIL-12R␤1 into the pMOWS-hygro plasmid for retroviral transduction of Ba/F3-gp130 cells (7). The resulting cDNAs for mIL-23R variants were inserted into the pMOWS-puro vector and used for the retroviral transduction of Ba/F3-gp130-mIL-12R␤1 cells. Mutations of valine within the putative murine ADAM17 cutting region of mIL-23R (V342D, V339D, and V339D/V342D) were introduced by PCR using Phusion high fidelity DNA polymerase (FINNZYMES, Thermo Fisher Scientific) followed by DpnI digestion of methylated template DNA. The coding sequence of Hyper-IL-23, a fusion protein of murine p40 followed by a synthetic linker (RGGGGSGGGGSVE) and murine p19 (21) was C-terminally fused with human Fc tag to generate Hyper-IL-23Fc (HIL-23Fc). Furthermore, human Fc was fused to the C terminus of murine IL-23R containing amino acids 24 -315 using standard cloning procedures to obtain the mIL23R D1D2D3-Fc fusion protein (6). Accordingly, mIL-23R D1D2-Fc (aa 24 -217) and IL-23R D1-Fc (aa 24 -123) have been generated. Splicing by overlapping extension PCRs was used to generate expression plasmids for mIL-23R D2D3-Fc (aa 122-315), mIL-23R D2-Fc (aa 122-217), and mIL-23R D3-Fc (aa 218 -315), respectively. ADAM17-GFP was cloned in our group by PCR amplification of murine ADAM17 and C-terminally fusing the PCR product in the open reading frame of GFP.
Shedding Assays-To analyze IL-23R ectodomain shedding in stably transduced Ba/F3-gp130-IL-12R␤1-IL-23R cells, the cells were washed once with PBS and diluted to a concentration of 1.2 ϫ 10 6 /ml. The cells were incubated with the inhibitors GI254023X (GI) or GW280264X (GW) for 30 min and afterward for 2 h with PMA (100 nM) or 1 h with ionomycin (1 M). For constitutive shedding, the cells were incubated for 24 h with GI (3 M), GW (3 M), Marimastat (10 M), or DMSO, which was added after 3 and 6 h. Cell supernatants were centrifuged to remove cellular debris. A volume of 1 ml of each supernatant was then incubated overnight at 4°C under constant agitation with 1 ml of supernatant of COS-7 cells and transiently transfected with the cDNA coding for HIL-23Fc to bind soluble IL-23R (sIL-23R). Afterward the complex was incubated with protein A-agarose for 4 h at 4°C under constant agitation. The agarose beads were then collected by centrifugation, and the supernatant was removed. The beads were washed two times with ice-cold IP buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 2.5 mM sodium pyrophosphate, 1 mM ␤-glycerophosphate, 1 mM Na 3 VO 4 , and a complete protease inhibitor mixture tablet/50 ml buffer (Roche Diagnostics)) and directly boiled in Laemmli buffer, which was subsequently analyzed by Western blot. To investigate ectodomain shedding of IL-23R in transiently transfected HEK293 cells, the cells were seeded out at 1 ϫ 10 6 /ml 24 h after transfection. Shedding assays were performed as described for Ba/F3 cells.
Flow Cytometry-For detection of cell surface expression of the different murine IL-23R variants, 1 ϫ 10 6 Ba/F3-gp130-IL-12R␤1-IL-23R cell lines were washed once with FACS buffer (PBS, 0.5% BSA) and incubated with 2.5 g of monoclonal mouse IL-23R antibody for 1 h on ice. After three subsequent washing steps with FACS buffer, the cells were incubated in 100 l of FACS buffer containing a 1:100 dilution of Alexa Fluor 647-conjugated Fab goat anti-rat IgG for 1 h at 4°C. The cells were washed with FACS buffer, resuspended in 500 l of FACS buffer, and analyzed by flow cytometry on a BD Bioscience FACS Canto II. The resulting data were evaluated using the FCS Express software (De Novo Software, Los Angeles, CA). Human IL-23R surface expression was detected with 0.5 g of human IL-23R-biotinylated monoclonal antibody/100 l of FACS buffer followed by incubation with 0.2 g of APC-streptavidin/100 l of FACS buffer.
Expression and Purification of GFP Nanobodies-The GFP nanobody (V H H) used for co-immunoprecipitation was designed according to (22). A C-terminal myc and His 6 tag was fused to the GFP-binding V H H domain, cloned into pET23a-PelB vector, and transformed in component Escherichia coli BL21 cells. 2 liters of culture medium was inoculated with 50 ml of E. coli preculture and incubated at 37°C and 140 rpm to an A 600 of 0.8. Bacterial expression was induced with 1 mM isopropyl ␤-D-1-thiogalactopyranoside, and the cells were incubated overnight at 30°C and 140 rpm. The cells were removed by centrifugation (1 h, 4°C, 3000 ϫ g), and supernatant was subjected to purification. Bacterial supernatant was filtered by a 0.45-m SCFA NALGENE Rapid-Flow TM bottle top filter (Thermo Fisher Scientific) to remove cell debris, mixed 1:1 with PP buffer (500 mM NaCl, 50 mM Na 2 HPO 4 /NaH 2 PO 4 , pH 7.4), and loaded on a pre-equilibrated nickel-nitrilotriacetic acid column (His-Trap FF; GE Healthcare). His-tagged GFP nanobodies were eluted by PP buffer containing 500 mM imidazole. After removal of imidazole the purified protein was concentrated using Amicon Ultra-15 10K centrifugal filter devices (Merck Millipore).
Coupling of GFP Nanobody to Sepharose Beads-For coupling of the purified GFP nanobodies to Sepharose, 1 mg of purified GFP nanobodies were covalently linked to 2 ml of NHS-activated Sepharose 4 Fast Flow (Ge Healthcare), according to the manufacturer's instructions. Therefore, beads were washed three times with ice-cold 1 mM HCl (2 min, 3000 ϫ g, 4°C). Purified protein was added to the beads and incubated for 2 h at room temperature under constant agitation. By adding blocking buffer (0.5 ethanolamine, 0.5 NaCl, pH 8.3) for 2 h under constant agitation, remaining free binding sites were blocked. Finally, beads were washed twice with 0.1 M Tris-HCl, pH 8, and stored in 20% ethanol at 4°C.
Co-immunoprecipitation of IL-23R and ADAM17-Interaction of ADAM17 and IL-23R was analyzed by co-immunoprecipitation. Therefore, HEK293 cells were co-transfected with murine ADAM17-GFP and the corresponding IL-23R variant. 48 h after transfection, the cells were lysed in lysis buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 0.5% Nonidet P-40 (Tergitol-type Nonidet P-40) and a complete protease inhibitor mixture tablet/50 ml of buffer) for 1 h at 4°C under constant agitation and centrifuged at 20000 ϫ g for 20 min. The resulting cell lysate was mixed with 300 l of dilution buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.5 mM EDTA, and a complete protease inhibitor mixture tablet/50 ml of buffer). For input control (I), 50 l of this solution was mixed with 50 l of 5ϫ Laemmli buffer and incubated for 10 min at 95°C. Sepharose beads that were coupled with GFP nanobodies were washed twice with dilution buffer to remove the ethanol and subjected to the cell lysate mixture mentioned before. Beads and lysate were incubated at room temperature for 2 h under constant agitation to bind mADAM17-GFP to the GFP nanobody Sepharose beads. Afterward Sepharose beads were collected by centrifugation. 50 l of resulting supernatant were mixed with Laemmli buffer and incubated for 10 min at 95°C (referred to as nonbound). The beads were washed three times with ice-cold dilution buffer and finally directly boiled in Laemmli buffer. After centrifugation, the resulting supernatant was subsequently used for Western blotting referred to as bound fraction.
Western Blotting-25 l of proteins from precipitated supernatant or cell lysate were separated by SDS-PAGE under reducing conditions and transferred to a PVDF membrane (GE Healthcare). The membranes were blocked with 5% skim milk powder in TBST-T (10 mM Tris-HCl, pH 7.6, 150 mM NaCl, and 0.05% Tween 20) and probed with the primary antibody in 5% skim milk powder in TBS-T (anti-human Fc, anti-GFP) or 5% BSA in TBS-T (hIL-23R biotinylated mAb, mIL-23R biotinylated mAb) at 4°C overnight. After several washing steps, membranes were probed with the secondary peroxidase-conjugated antibody or streptavidin-HRP for 1 h at room temperature before applying the ECL Prime Western blotting detection reagent (GE Healthcare) according to the manufacturer´s instructions. To remove the antibodies, membranes were stripped in 62.5 mM Tris-HCl (pH 6.8), 2% SDS, and 0.1% ␤-mercaptoethanol for 20 min at 60°C.
Internalization-2 ϫ 10 6 Ba/F3-gp130-mIL-12R␤1-mIL-23R cells were stimulated with Marimastat (10 M), GI (3 M), GW (3 M), or DMSO for 30 min at 37°C; washed with FACS buffer (0.5% BSA/PBS); and incubated with primary antibody (␣-mIL-23R, ␣-hIL-23R) for 1 h on ice. Afterward the cells were resuspended in prewarmed medium supplemented with HIL-23 (0.2% of conditioned cell culture medium from a stable CHO-K1 cell line secreting HIL-23 (7)), Marimastat (10 M), GI (3 M), GW (3 M), or DMSO and incubated for the indicated time points at 37°C and 5% CO 2 . Subsequently cells were washed and stored on ice. After the last time point, cells were incubated for 1 h on ice with secondary antibody (Alexa Fluor 647 conjugate). After removing the nonbound antibodies by washing, the cells were analyzed by flow cytometry on a BD Bioscience FACS Canto II, and the data were visualized using the FCS Express software.
Programs and Statistical Analysis-Western blots were analyzed by ImageJ. Statistical analysis was performed by using Student's t test. A p value of Ͻ 0.05 (*) was defined as statistically significant; p Ͻ 0.01 (**), p Ͻ 0.001 (***).

Inhibition of Ectodomain Shedding Prolonged the Half-life of IL-23R on the Cell Surface-
We analyzed the cell surface halflife of the murine IL-23R. Murine IL-23R cDNA was stably introduced into Ba/F3-gp130-mIL-12R␤1 cells. Ba/F3 cells are ideal tools for determining the cell surface half-life of membrane proteins by flow cytometry (23). Cell surface expressed mIL-23R was labeled with IL-23R antibodies at 4°C, and nonbound antibodies were washed away. Cells were shifted back to 37°C for 0, 10, 20, 30, 60, and 90 min to allow IL-23R internalization and ectodomain shedding in the presence and absence of HIL-23. After the indicated time points, the cells were stained with secondary Alexa Fluor 647-conjugated Fab goat anti-rat IgG at 4°C to quantify the remaining cell surface expression of IL-23R by flow cytometry. The time point T0 is the same in all experiments shown in Fig. 1A because the experiment for all conditions were done in parallel. During the flow cytometry pulse-chase experiment, the time-dependent reduction of cell surface IL-23R expression was analyzed. Cell surface halflife of mIL-23R in the presence and absence of HIL-23 was comparable with ϳ14 min without and ϳ18 min with cytokine stimulation (Fig. 1, A and B). Interestingly, inhibition of ectodomain shedding by GI, GW, or Marimastat resulted in prolonged cell surface exposition of IL-23R of ϳ28, 35, and 37 min, suggesting that disappearance of membrane-bound IL-23R was conducted by a combination of internalization and ectodomain shedding.
Induced Ectodomain Shedding of the IL-23R Is Mediated by ADAM17-Human and murine IL-23R cDNAs were stably introduced into Ba/F3-gp130-IL-12R␤1 cells. Cell surface expression of human IL-23R (hIL-23R) was verified by flow cytometry (Fig. 2A). Proliferation of Ba/F3-gp130-IL-12R␤1-IL-23R cells was dependent on IL-23, confirming biological activity of both IL-23 receptors (data not shown). Ba/F3-gp130-IL-12R␤1-IL-23R cells were treated with PMA and ionomycin. Soluble IL-23R (sIL-23R) within the cellular supernatant was co-precipitated with HIL-23Fc fusion protein and protein A-agarose. Co-precipitated sIL-23R was detected by Western blotting. As depicted in Fig. 2B, human and murine sIL-23Rs were detected in the cell culture supernatant after PMA but almost not after ionomycin treatment. As control for the co-precipitation, HIL-23Fc was detected by Western blotting using an Fc-specific antibody. Typically PMA induces ADAM17, whereas ADAM10 is stimulated by ionomycin (9). Our results clearly demonstrate that sIL-23R was generated after PMA treatment, suggesting ectodomain shedding of the receptor. Next, Ba/F3-gp130-IL-12R␤1 cells expressing human or murine IL-23R were treated with PMA in the presence of the ADAM10-selective inhibitor GI or the ADAM10/ADAM17selective inhibitor GW (9). Again, PMA treatment resulted in release of sIL-23R, which was inhibited by GW but not by GI (Fig. 2C). These results suggest that ADAM17 is activated after PMA treatment and releases IL-23R by ectodomain shedding.
We have shown previously that the IL-6R stalk region is cleaved by ADAM17 after PMA stimulation and by ADAM10 after ionomycin stimulation (9). Because we did not observe ionomycin-induced shedding of the IL-23R, we asked whether transfer of the IL-6R stalk region results in ionomycin-induced shedding. We generated the synthetic hybrid receptor, consisting of human IL-23R extracellular domains (D1-D3), transmembrane domain, and the intracellular domain, with an interspersed exchange of the IL-23R stalk region by the IL-6R stalk domain. In transiently transfected HEK293 cells, the IL-23R-(stalk-IL-6R) chimeric receptor was cleaved after PMA stimulation with ADAM17, because GW completely suppressed the generation of sIL-23R, and interestingly, restored ionomycininduced IL-23R shedding mediated by ADAM10 was completely suppressed by GI (Fig. 2, D and E), suggesting that the composition of the stalk region was sufficient to enable induction of ADAM10-induced shedding.
Even though we did not detect ionomycin-induced IL-23R shedding, we analyzed whether IL-23R is a substrate of constitutive shedding commonly mediated by ADAM10 (11). Equal numbers of Ba/F3-gp130-IL-12R␤1 cells expressing human or murine IL-23R were incubated for 24 h without further stimulation, and sIL-23R was detected by HIL-23Fc co-precipitation and Western blotting. After 24 h, human and murine sIL-23R were detected in the supernatants of the cells (Fig. 2F). Interestingly, the broad spectrum inhibitor Marimastat, the ADAM10-selective inhibitor GI, and the ADAM10/ADAM17selective inhibitor GW completely suppressed the generation of sIL-23R (Fig. 2F), suggesting that ADAM10 is the constitutive sheddase of IL-23R.
We also made use of HEK293T cells, transiently transfected with murine or human IL-23R cDNA, in which ADAM10, ADAM17, or both proteases were genetically inactivated by CRISPR/Cas9 targeting. Stimulation of PMA resulted in the formation of murine and human sIL-23R, which was completely absent in ADAM17-and ADAM10/ADAM17-deficient HEK293T cells but not in ADAM10 deficient cells. As a control, sIL-23R was detected in ADAM17-deficient HEK293T cells, which were reconstituted by transient transfection with a cDNA coding for murine ADAM17 (mADAM17-GFP fusion protein) (Fig. 2, G and H). Our results demonstrate that murine and human IL-23R were novel substrates of ADAM10 and ADAM17 and that the soluble receptors retained binding capacity toward IL-23 as approved by HIL-23Fc co-precipitation.
Homology alignment of predicted human and murine IL-23R stalk region revealed an overall low homology, except of the region 318 -325 of the human IL-23R with 338 -345 of the murine IL-23R (Fig. 4A, boxed amino acid residues). Although ADAM17 has no defined cleavage side, some prediction of preferred amino acids in the P1Ј position can be made. ADAM17 prefers valine as amino acid in the P1Ј position (24), and we identified two conserved valine residues at positions 339 and 342 in the murine IL-23R and at positions 320 and 323 in the human IL-23R. Next, we introduced point mutations in the mIL-23R and replaced valine 339 and valine 342 to aspartic acid, which was the most unfavorable amino acid residue at P1Ј for ADAM17. Cell surface expression of mIL-23R V342D , mIL-23R V339D , and the double mutant mIL-23R V339D/V342D was comparable in Ba/F3-gp130-mIL-12R␤1 cells as verified by Western blotting and flow cytometry (Fig. 4, B and C). Although some residual shedding of mIL-23R V342D , mIL-23R V339D , and the double mutant mIL-23R V339D/V342D was detected, it was greatly reduced as compared with wild-type murine IL-23R (Fig. 4D). Using ImageJ for quantification of Western blots from three independent experiments, our data indicate that shedding of all variants was significantly reduced by ϳ70 -80% as compared with wild-type IL-23R (Fig. 4E). Interestingly, not only PMA-induced shedding but also constitutive shedding was significantly reduced (Fig. 4, F and G), suggesting that these amino acid residue exchanges also compromised constitutive shedding of the mIL-23R by ADAM10.
Interaction of IL-23R with ADAM17 Is Mainly Mediated by Domains 1 and 3 of IL-23R-Interaction of mADAM17 and mIL-23R was analyzed using a C-terminal fusion protein of mADAM17 with GFP and membrane-bound mIL-23R. GFPtagged mADAM17 was precipitated with a GFP nanobody coupled to Sepharose (22). Transiently transfected HEK293 cells were co-transfected with cDNAs coding for mADAM17-GFP and mIL-23R. After lysis of the cells, mIL-23R and the noncleavable mIL-23R⌬333-372 variant were co-precipitated with ADAM17-GFP (Fig. 5A), demonstrating that both proteins formed a complex, which was not depending on the cleaving capability of the substrate. As a control, IL-23R and the mIL-23R⌬333-372 variant were not precipitated by the GFP nanobody in the absence of ADAM17-GFP (Fig. 5A). Also human IL-23R was co-precipitated by mADAM17-GFP and GFP nanobody-Sepharose (Fig. 5B). Next, soluble variants of mIL-23R fused to an Fc domain from an IgG1 antibody were co-precipitated by mADAM17-GFP to identify the responsible domain within the mIL-23R for mADAM17 complex formation. Here, we generated soluble mIL-23R-Fc fusion proteins composed of all extracellular domains 1-3 (aa 24 -315) lacking the stalk region of the IL-23R named smIL-23R-D1D2D3, of domains D1 and D2 (smIL-23R-D1D2, aa 24 -217), of domains D2 and D3 (smIL-23R-D2D3, aa 122-315), and of only one of the domains (smIL-23R-D1, aa 24 -123; smIL-23R-D2, aa 122-217; an smIL-23R-D3, aa 220 -315). After co-transfection of cDNAs coding for sIL-23-Fc variants with mADAM17-GFP, cells were lysed and incubated with GFP nanobody Sepharose. Interestingly, smIL-23R-D1D2D3 was co-precipitated by mADAM17-GFP, demonstrating that the interaction of IL-23R and ADAM17 is mediated by the extracellular domains FIGURE 5. Interaction of IL-23R with ADAM17 is mainly mediated by domain 1 and 3 of IL-23R. A, left panel, cDNAs coding for mIL-23R and mADAM17-GFP were co-transfected in HEK293 cells. 48 h after transfection, the cells were lysed and precipitated by GFP nanobody Sepharose. mIL-23R and mADAM17-GFP were detected by Western blotting using mIL-23R and GFP antibodies. Right panel, schematic illustration of IL-23R domain structure. B, cDNAs coding for hIL-23R and mADAM17-GFP were co-transfected in HEK293 cells and interaction of IL-23R and mADAM17-GFP was analyzed as in A. C, cDNAs coding for smIL-23R and variants thereof with mADAM17-GFP were co-transfected in HEK293 cells, and interaction of smIL-23R and variants thereof with mADAM17-GFP was analyzed as in A. I, input; NB, nonbound after precipitation; B, bound after precipitation; I, immunoprecipitation. of both proteins. Interestingly, binding was also shown for smIL-23R-D1D2 and, although to a lesser extent, by smIL-23R-D2D3. Confusingly, ADAM17 efficiently precipitated mIL-23R-D3, to a lesser extent mIL-23R-D1 but not IL-23R-D2 (Fig.  5C). Western blotting against the GFP tag of mADAM17 verified efficient precipitation of mADAM17-GFP. Our results demonstrate that the interaction of smIL-23R and mADAM17 is mainly mediated by domains 1 and 3 of the mIL-23R.

Discussion
Here, we describe that the murine and human IL-23R were substrates of the ADAM proteases ADAM10 and ADAM17. Initially, we found that the cell surface half-life of murine IL-23R was prolonged after broad spectrum inhibition of metalloproteases by Marimastat, suggesting ectodomain shedding of IL-23R. Because commercially available ELISAs were not sensitive enough to detect sIL-23R in cell culture supernatants, we made use of recombinant Fc-tagged Hyper-IL-23 to co-immunoprecipitate sIL-23R, followed by Western blotting of sIL-23R. Detailed analysis showed that ADAM17 cleaved IL-23R after stimulation with the phorbol ester PMA, whereas ADAM10 was responsible for constitutive shedding of the IL-23R but surprisingly not able to shed IL-23R after ionomycin treatment. To determine the responsible ADAM proteases, we made use of selective chemical inhibitors of ADAM10 und ADAM17 and HEK293 cells in which ADAM10/ADAM17 were genetically inactivated by CRISPR/Cas9. Exchange of the stalk region of the IL-23R by the IL-6R renders the synthetic IL-23R-(stalk-IL-6R) variant also sensitive for ionomycin-inducible ADAM10-mediated shedding. Co-immunoprecipitation of sIL-23R by IL-23 also showed that sIL-23R was biologically active in terms of IL-23 binding. We also tested whether the released sIL-23R was able to inhibit IL-23 signaling; however, the amount of sIL-23R in the cell culture supernatant was too low to inhibit IL-23 signaling (data not shown). Cleavage of ADAM proteases mostly occurs in the membrane-proximal region (11); stepwise deletion of the stalk region of the IL-23R, however, failed to determine a membrane-proximal target cleavage sequence within the IL-23R. Instead deletion of the complete stalk region suppressed constitutive and inducible shedding of IL-23R. Homology alignment of murine and human IL-23R revealed an overall low homology of the stalk region. Interestingly, a short amino acid stretch close to the domain 3 of IL-23R was highly conserved. Here, we identified two possible cleavage sites within the IL-23R. Mutation of two valines, which are preferred P1Ј amino acids of ADAM17 cleavage (24), resulted in IL-23R variants, in which constitutive and induced shedding was dramatically reduced, suggesting that this was the cleavage site for ADAM10 and ADAM17 or that this region was critically involved in substrate recognition.
Human IL-23R is mainly expressed on activated memory T cells, natural killer cells, and innate lymphoid cells and at lower levels on monocytes, macrophages, and dendritic cells (25). Interestingly, no study directly approved cell surface expression of endogenous IL-23R by flow cytometry or Western blotting. We were, however, also not able to detect endogenous IL-23R on these cells, and therefore, our study lacks the direct proof that endogenous IL-23R is also a target of ADAM proteases.
Binding of ADAM proteases to substrates is considered to be mediated not via the cleavage site. First, we showed by co-immunoprecipitation that ADAM17 and murine and human IL-23R physically interact and that the putative cleavage site of the murine IL-23R was not needed for this interaction. Interestingly, smIL-23-D1D2D3 only consisting of the extracellular domains 1-3 was also efficiently co-immunoprecipitated by ADAM17, demonstrating also that the intracellular domain, the transmembrane domain, and the complete stalk region of IL-23R were not needed for substrate binding. Expression of single and combined extracellular domains of IL-23R showed, that ADAM17 interacts, although to a lesser extent with smIL-23-D1D2 and smIL-23-D2D3 and the single domains smIL-23R-D1 and smIL-23R-D3, but not with smIL-23-D2.
Alternative splicing of IL-23R result in a series of truncated soluble IL-23R proteins (12)(13)(14). Soluble variants of the IL-23R, lacking the transmembrane and intracellular region, represent potential therapeutic agents for the selective blockade of IL-23 signal transduction. Recombinant soluble IL-23R variants have been shown to block IL-23 and inhibit the generation of T H 17 cells (26,27). Furthermore administration of a human soluble IL-23R protein protected mice against the development of EAE (26).
Here, we provide first evidence that IL-23R is a substrate of ectodomain shedding. Shedding of IL-23R might contribute to negative regulation of IL-23 signaling with two major steps. First, ectodomain shedding of IL-23R result in down-regulation of cell surface expressed IL-23R molecules and thereby might reduce the cellular responsiveness toward IL-23. Second, sIL-23R is able to bind to IL-23 and thereby might act as competitive antagonist of IL-23 signaling.