Fig 2: Crystal structures of the mouse IL-27 in complex with IL-27Rα and human IL-27 in complex with SRF388Fab(A) Overview of the IL-12 family of cytokines and their receptors. Immunoglobin (Ig) domains are shown as ovals with thick black outline. The cytokine-binding homology region (CHR) consisting of tandem fibronectin type III domains (FNIII) is shown with a double line in the upper domain and a single line in the lower domain. For IL-35 signaling complexes comprising gp130/gp130, IL-12Rβ2/IL-12Rβ2, and IL-27Rα/IL-12Rβ2 have also been described (Pylayeva-Gupta, 2016).(B) Comparison of the signaling activity of the recombinant mIL-27 (used for crystallization), recombinant single chain mouse IL-27, mIL-27sc (used for BLI), and commercially produced recombinant single chain mouse IL-27 from BioLegend, mIL-27scBL. STAT1 and STAT3 activity was measured in CD8+ T cells by flow cytometry upon stimulation with increasing concentrations of mIL-27. n = 3 technical replicates/data points, with error bars indicating SEM. These data are representative of two independent experiments with splenocytes from two distinct mice.(C) Chromatogram of the mIL-27:mIL-27Rα:Nb5 complex from the Superdex200 increase column. Coomassie-stained TGX gel run under denaturing conditions depicts the complex used for the crystallographic trials. Western blot of the same sample with antibodies specific to the respective subunits run in parallel is depicted on the right. Full blot is available in Figure S1B.(D) Kinetic binding profiles of hIL-27-SRF388 interaction characterized by the BLI. Anti-human IgG Fc (AHC) biosensors were used to immobilize SRF388 followed by binding measurements in different concentrations of hIL-27. One biosensor was used as a reference channel for background subtraction. Data traces (black) were fitted using a 1:1 interaction model (red) to quantify the kinetics (ka, kd) and binding affinity (KD) of the interactions. Data were analyzed with Octet Data Analysis software v10.0.1.7 (ForteBio). Data presented represent one experiment performed using optimized assay conditions. Additional BLI experiments were conducted; however, the kd exceeded the limit for dissociation. MSD solution-phase studies were used to confirm the BLI measurements (see STAR methods for additional detail).(E) Cartoon representation of the mIL-27:mIL-27Rα:Nb5 crystal structure. Nanobody 5 used as a crystallization adjuvant is shown in gray surface representation.(F) Cartoon representation of the hIL-27:SRF388Fab crystal structure. Inhibiting antibody SRF388Fab is shown in surface representation with light chain (LC) in magenta and heavy chain (HC) in light pink. Related to Figures S1–S3.

Fig 3: High-affinity binding of gp130 to site 3 occurs independently of site 2(A) Predicted location of the gp130 receptor binding consistent with the site 3 mode of interaction. Top panel: crystal structure of the mIL-27:mIL-27Rα:Nb5 complex (Nb5 not shown), bottom panel: crystal structure of the hIL-27:SRF388Fab complex (SRF388Fab not shown).(B–E) (B) Kinetic profiles of the receptor complex mediated by mIL-27 characterized by BLI. Biotinylated mIL-27sc comprising the EBI3 subunit fused to the p28 subunit via a (GGGS)4 linker (see method details), was coupled to the surface of streptavidin-coated BLI sensors, followed by binding measurements in different concentrations of mIL-27RαCHR, (C) mgp130IgCHR in the presence of 100 nM of mIL-27RαCHR, (D) mgp130IgCHR, and (E) mgp130CHR. Data traces (black) were fitted using a 1:1 interaction model (red) to quantify the kinetics (ka, kd) and binding affinity (KD) of the interactions using the Octet Analysis Studio 12.2.1.24 software. For each experiment three technical replicates were performed. The reported KD, ka, and kd values represent average values from three technical replicate experiments.(F) The anti-gp130Ig antibody B-T2 blocks IL-27 signaling in U937 cells and (G) PBMCs. U937 cells or Ficoll-isolated human PBMCs were cultured in RPMI with various concentrations of anti-gp130 antibodies and rhIL-27 for 20 min at 37°C. Cells were fixed and stained for pSTAT1 (pY701). Samples were washed with FACS buffer, read on an LSR Fortessa (BD Biosciences), and analyzed using the FlowJo Software analysis program (TreeStar). Cytokine stimulated conditions represents 0% inhibition and unstimulated conditions represents 100% inhibition. The data presented represent the mean of two technical replicates per data point with error bars indicating SD. These data are representative of two independent experiments with U937 cell line or PBMCs from healthy donors.

Fig 4: The IL-27 receptor assembly is distinct from IL-12/IL-23 complexes and is structurally intermediate between IL-6 and IL-12/23 mediated receptor assemblies(A) For IL-23, IL-23p19 binds IL-12p40 via site 1; however, IL-23R binds IL-23p19 exclusively with its N-terminal immunoglobulin domain via a site 3 instead of using the elbow region between D2 and D3 as in other type I cytokine receptors. IL-12Rβ1 receptor, which is shared with the IL-12 cytokine, does not engage the classical site 2 but binds p40 at a site opposite to p19, here labeled 2′. Structure of the minimal IL-23 receptor complex (PDB: 6wdq) was supplemented with an AlphaFold model of D2-D5 of IL-12Rβ1.(B) IL-27 uses the canonical site 1 for the p28:EBI3 interaction as well as site 2 for the p28:IL27Rα binding akin to IL-6:IL-6Rα and IL-6:gp130 binding respectively. gp130 is predicted to bind the conserved W195 in mouse and W197 in human p28. Here the mouse minimal complex presented in this manuscript was combined with the AlphaFold models of D3-D5 of IL-27Rα and a full-length gp130.(C) IL-6 receptor complex is composed of an IL-6:IL-6Rα:gp130 complex that forms a hexamer. gp130 engages both sites 2 and 3 on IL-6. First it binds site 2 of IL-6 with its elbow region between D2D3 forming an intermediate nonsignaling complex. The resulting complex dimerizes via D1 of gp130 and site 3 of the cytokine on another IL-6:IL-6Rα:gp130 complex. Surface representations of models of the extracellular signaling complexes. Models were assembled from published crystal structures supplemented by AlphaFold2 predictions.

Fig 5: Comparison of binding interfaces of mouse and human IL-27(A) Close-up of the site 1 interface in mouse (top) and human (bottom) IL-27.(B) Close-up of the site 2 interface between mp28 and mIL-27Rα (top) and hp28 and SRF388Fab.(C) Sequence of the SRF388Fab with CDRs visualized with magenta box (light chain) and light pink box (heavy chain). Residues taking part in site II interface are marked with magenta circles. Sequences were shortened for clarity.(D) Crystal structure of the mIL-27:mIL-27Rα:Nb5 binary complex with mp28 in cartoon representation (light blue) and mEBI3 and mIL-27Rα in surface electrostatic representation. Structure of the mEBI3 (left) reveals a large positively charged patch on domain 2 at the site of mIL-27Rα binding, whereas mIL-27Rα possesses a complementary patch that is in turn negatively charged (right).(E) Close-up view of the mEBI3:mIL-27Rα interaction. Binding occurs via a large ∼660 Å2 interface.(F) Arginine 51 of the mp28 is interacting simultaneously with E182 of the mEBI3 as well as with Q154 of the mIL-27Rα.(G) Residue R55 from hp28 is observed forming hydrogen bonds with E184 of hEBI3 similarly to the mouse orthologue. Related to Figures S4 and S5, and Tables S1 and S2.