GM-CSF (granulocyte/macrophage colony-stimulating factor) is an important mediator of inducible haemopoiesis and inflammation, and has a critical role in the function of alveolar macrophages. Its clinical applications include the mobilization of haemopoietic progenitors, and a role as an immune stimulant and vaccine adjuvant in cancer patients. GM-CSF signals via a specific α receptor (GM-CSFRα) and the shared hβc (human common β-subunit). The present study has investigated the role of the Ig-like domain of GM-CSFRα in GM-CSF binding and signalling. Deletion of the Ig-like domain abolished direct GM-CSF binding and decreased growth signalling in the presence of hβc. To locate the specific residues in the Ig-like domain of GM-CSFRα involved in GM-CSF binding, a structural alignment was made with a related receptor, IL-13Rα1 (interleukin-13 receptor α1), whose structure and mode of interaction with its ligand has recently been elucidated. Mutagenesis of candidate residues in the predicted region of interaction identified Val51 and Cys60 as having critical roles in binding to the α receptor, with Arg54 and Leu55 also being important. High-affinity binding in the presence of hβc was strongly affected by mutation of Cys60 and was also reduced by mutation of Val51, Arg54 and Leu55. Of the four key residues, growth signalling was most severely affected by mutation of Cys60. The results indicate a previously unrecognized role for the Ig-like domain, and in particular Cys60, of GM-CSFRα in the binding of GM-CSF and subsequent activation of cellular signalling.

INTRODUCTION

GM-CSF (granulocyte/macrophage colony-stimulating factor), interleukin (IL)-5 and IL-3 are three cytokines produced by activated T-cells during immune responses that are important mediators of inducible haemopoiesis and inflammation. They signal through a shared receptor {hβc [human βc (common β-subunit)]} and play an important role in the pathogenesis of allergic disorders and inflammatory diseases of the lung, such as asthma. Eosinophilia is controlled primarily by IL-5 [1,2] and, to a lesser extent, by IL-3 and GM-CSF. IL-3 plays a critical role in basophil and mast cell responses in parasite infections [3]. GM-CSF is believed to be centrally involved in chronic inflammatory diseases, such as arthritis and multiple sclerosis [4]. Of the three cytokines, GM-CSF has received the most attention in terms of clinical applications. GM-CSF has been applied to the supportive care of cancer patients and, more recently, to the mobilization of progenitors. The ability of GM-CSF to stimulate cytotoxic immune responses gives it great potential as an immune stimulant and vaccine adjuvant in cancer patients [4]. GM-CSF is also an essential factor in maintaining alveolar homoeostasis through its action on alveolar macrophages, and mutations affecting GM-CSFRα (GM-CSF receptor α subunit) have recently been detected in patients with pulmonary alveolar proteinosis [5]. A critical first step in GM-CSF signalling involves its binding to the specific GM-CSFRα. In the present study, we have demonstrated an important role of the first domain of GM-CSFRα in binding GM-CSF, an aspect that has not been studied previously.

The receptors for IL-5, IL-3 and GM-CSF consist of cytokine-specific α receptors essential to the activation of hβc, which is believed to be the main signalling entity [69]. Human GM-CSF and IL-3 bind to their cognate α receptors with low affinities (1–50 nM) but, in the presence of hβc, high-affinity complexes are formed (30–200 pM). By convention, the terms ‘low’ and ‘high’ affinity are used to distinguish the two receptor-binding modes in studies of cytokine–receptor interactions, even though in many areas of biochemistry the nanomolar affinity associated with ‘low’-affinity binding would be considered very high. The formation of a complex involving ligand, α and βc receptors is necessary for receptor activation and signalling. The cytoplasmic portions of α and βc subunits possess no intrinsic tyrosine kinase activity [10], but the activated receptor complexes formed with all three ligands interact with and activate JAK2 (Janus kinase 2) [11], leading to the phosphorylation of eight tyrosine residues located in the βc cytoplasmic domain [12,13]. Subsequently, several signalling pathways are induced [14].

Structurally, the α and βc subunits of hGM-CSFR (human GM-CSFR), hIL-3R (human IL-3 receptor) and hIL-5R (human IL-5 receptor) belong to the cytokine class I receptor superfamily (or haemopoietin receptor family), which includes GHR [GH (growth hormone) receptor], the EPO (erythropoietin) receptor, gp130 and IL-4R (IL-4 receptor)/IL-13R (IL-13 receptor). The characteristic feature of this family is the extracellular cytokine-receptor homology module (CRM or CRH), composed of two fibronectin III domains that contain a number of conserved sequence elements (reviewed in [15]). Crystal structural analyses of the ligand–receptor complexes of GH, EPO, IL-4, IL-6, IL-12, G-CSF (granulocyte colony-stimulating factor) and, very recently, IL-13 have demonstrated that the extracellular domains of these receptors possess CRMs consisting of two approximately orthogonal fibronectin III domains that each contributes key residues for ligand binding from loops at the receptor ‘elbow’ region [1621].

In 2001, we reported the crystal structure of the ectodomain of βc of the hGM-CSFR which showed that it was a radical departure from the GHR paradigm [22]: being a pre-formed intertwined dimer with novel chain swapping between domains. Subsequent mutagenesis elucidated the functional epitope and showed that the βc possessed an unusual ligand-binding elbow region formed by the two protein chains [23,24]. GM-CSFRα together with IL-3Rα and IL-5Rα is part of a sub-group of the class I cytokine receptor superfamily which possess an additional ‘Ig-like’ domain (D1) at the N-termini of their ectodomains in addition to the CRM. In other cytokine receptor systems, structural studies have clearly established the pivotal roles that Ig-like domains play in mediating cytokine binding and receptor activation. For example, the crystal structure of the IL-6–IL-6Rα (IL-6 receptor α subunit)–gp130 complex demonstrates that the Ig-like domain of gp130 has a bridging function, interacting with the binding epitope on IL-6 to recruit another trimer, to generate the hexameric signalling complex [25]. Very recently, IL-13Rα1 D1 was shown to play a key role in ligand-binding by engaging the dorsal surfaces of IL-4 and IL-13 in their respective crystal structures [26]. In the case of IL-5, alanine scanning mutagenesis of the IL-5Rα subunit identified the residues Asp55, Asp56 and Tyr57 in IL-5Rα D1 as essential determinants of IL-5 binding [27,28]. Additionally, deletion of D1 of the hIL-5Rα (human IL-5Rα) blocks IL-5 signalling [27]. A different situation exists for IL-3Rα. We have demonstrated the existence of an isoform lacking D1 which, although it does not bind IL-3 as tightly as the full-length receptor, is competent for signalling [29].

In the case of hGM-CSF (human GM-CSF), whether D1 of the GM-CSFRα subunit has a role in ligand binding and receptor activation has not been established. Several studies have been performed on the CRM of hGM-CSFRα to identify the functional determinants involved in the GM-CSF–GM-CSFRα interaction, the first step in receptor activation. Guided by homology modelling with the GHR complex, Arg280, from the F′–G′ loop of GM-CSFRα, was identified as being involved in ligand binding [30]. A recent crystal structure of hGM-CSF in complex with the ectodomains of the GM-CSFRα and βc subunits has been an interesting new development in structural studies of the GM-CSFR, and the authors have suggested that the active complex is a dodecamer [31]. The new structure has provided valuable verification of the previously determined βc structure [22] and its functional epitope [23], and showed that the loop residues 241–251 and 299–305 of GM-CSFRα interact with residues 11–23 (helix A) and residues 112–118 (helix D) of the cytokine [31]. Unfortunately, the electron density for D2 was relatively poor and no electron density was present for D1 of GM-CSFRα. The structure therefore has provided no insights into (i) whether domain D1 plays a role in GM-CSF recognition; and (ii) whether this Ig-like domain participates in the signalling complex.

In the present study, we demonstrate the critical importance of hGM-CSFRα D1 for GM-CSF binding and signal transduction. To aid in the investigation of the role of the Ig-like domain we prepared a structural homology model between hGM-CSFRα and IL-13Rα1. Twelve residues in the probable region of interaction with GM-CSF were mutated using alanine substitution mutagenesis. This enabled us to identify the critical determinants, Cys60, Val51, Arg54 and Leu55, for which their respective alanine substitution mutants exhibited defective GM-CSF recognition and/or receptor activation, despite normal levels of cell-surface expression. These findings clearly implicate the Ig-like domain of hGM-CSFRα as a crucial determinant of GM-CSF binding and receptor activation, and highlight the Ig-like domain as a novel target for the development of therapeutics to ameliorate chronic inflammatory diseases and multiple sclerosis.

EXPERIMENTAL

Preparation of hGM-CSFRα ΔD1

hGM-CSFRα ΔD1, lacking exons 2 and 3, was constructed by deletion mutagenesis using the QuikChange® method (Stratagene) with Pfu Turbo DNA polymerase and the following primers: hGMa del25-113 fwd: cctcctgatcccagagaaatcAGGAAGGGAGGGTACCGC; and hGMa del25-113 rev: GCGGTACCCTCCCTTCCTgatttctctgggatcaggagg (the uppercase and lowercase letters correspond to the respective 5′ and 3′ sequences flanking the deleted region). The deletion of exons 2 and 3 eliminates D1 comprising the residues 25–113 (between the amino acid sequences IPEK24 and S114GREG), whilst leaving the signal peptide encoded by exon 1 intact. A modified form of hGM-CSFRα ΔD1 was similarly constructed by insertion mutagenesis to introduce an N-terminal FLAG epitope to allow better detection of protein cell-surface expression by flow cytometry.

Site-directed mutagenesis of hGM-CSFRα

Site-directed mutagenesis was carried out using the QuikChange® method with Pfu Turbo Ultra DNA polymerase. The complete sequences of mutant GM-CSFRα cDNAs were verified by Big Dye Terminator cycle sequencing (Applied Biosystems).

Expression constructs

For expression in COS7 cells, the cDNAs encoding hβc and wild-type or mutant GM-CSFRα subunits were subcloned into pCEX-V3-Xba, a vector derived from pCEX-V3 [32]. This expression vector contains the SV40 (simian virus 40) origin of replication, enhancer and promoter regions, so that, after transfection into COS7 cells, the plasmid replicates episomally and transcribes efficiently across the insert. For expression in CTLL2 cells, cDNAs encoding wild-type or mutant hGM-CSFRα subunits were subcloned into pEF-IRES-N, and βc were subcloned into pEF-IRES-P [33] (a gift from Dr Steve Hobbs, Institute of Cancer Research, London, U.K.).

Cytokines and radiolabelling

hGM-CSF and mIL-2 (murine IL-2) were produced using the baculovirus expression system, as described previously [34]. Purified GM-CSF was radiolabelled using the Iodogen reagent, as described previously for murine IL-3 [24,35]. Radiolabelled GM-CSF was stored at 4 °C and used for up to 10 days. 125I was purchased from PerkinElmer Life Sciences and stored at room temperature (22 °C).

Cells and DNA transfections

COS7 cells were maintained in DMEM (Dulbecco's modified Eagle's medium) containing 10% (v/v) FBS (fetal bovine serum). CTLL2 cells were maintained in RPMI 1640 containing 10% (v/v) FBS, 50 μM 2-mercaptoethanol and 10 units/ml mIL-2. Briefly, COS7 cells were harvested using trypsin and electroporated with 10 μg of wild-type or mutant GM-CSFRα either alone or together with 25 μg of wild-type βc at 200 V and 960 μF. Binding studies and antibody staining were performed on cells 64–68 h after transfection. CTLL2 cell lines stably expressing wild-type βc and wild-type or mutant GM-CSFRα subunits were generated in two steps with DNA introduced by electroporation, essentially as described previously [34].

Equilibrium binding analysis for GM-CSF

‘Hot’ saturation binding experiments were performed on COS7 cells transfected with DNA constructs encoding the relevant receptors, as described previously [23]. In order to assess the capacity of hGM-CSFRα subunits to bind hGM-CSF directly, ‘cold’ saturation assays were performed on COS7 cells transfected with cDNAs encoding wild-type or mutant hGM-CSFRα subunits in the absence of hβc cDNA. 125I-Labelled hGM-CSF (1 nM) was added to 106 cells in 200 μl of binding medium (RPMI 1640 supplemented with 0.5% BSA and 10 mM Hepes, pH7.5) with a serial dilution from a 200-fold excess of unlabelled hGM-CSF. Non-specific binding was determined by performing the assay in the presence of 1 μM unlabelled hGM-CSF. After incubation for 2.5–3 h at 4 °C with intermittent agitation, the assay was terminated by centrifugation through 2:1 (v/v) dibutyl phthalate/dinonyl phthalate at 12000 g for 4 min. The tips of tubes and visible cell pellets were counted using a Packard 5780 Auto-gamma counter. Kd (dissociation constant) values were determined from specific binding data by using the programs EBDA and LIGAND [36,37], as described previously [23,24,35].

Flow cytometry

COS7 or CTLL2 cells transfected with cDNAs encoding FLAG-tagged hGM-CSFRα ΔD1 were incubated in the presence of a FITC-conjugated anti-FLAG M2 monoclonal antibody (Sigma) and then washed before analysis of cell-surface expression was performed on a FACScan flow cytometer (BD Biosciences). COS7 or CTLL2 cells transfected with cDNAs encoding wild-type or mutant hGM-CSFRα subunits were incubated in the presence of purified anti-(human CD116) (10 μg/ml; clone 4H1, mouse IgG1; Biolegend) for 30 min at 4 °C. The cells were washed and incubated with a FITC-conjugated rat anti-(mouse IgG1) secondary antibody (5 μg/ml; BD Pharmingen) for 25 min at 4 °C and then washed before analysis of cell-surface expression was performed on a FACScan flow cytometer (BD Biosciences). For analysis of hβc expression by FACS, cells were harvested and washed with ice-cold PBS, and then cells were incubated with a 1:200 dilution of primary antibody [biotin-conjugated mouse anti-(human CDw131); 0.5mg/ml; clone 3D7, cat. no. 554535; BD Biosciences] in FACS incubation buffer [0.2% BSA and 5% (v/v) FBS in ice-cold PBS] for 1 h. After three washes with FACS washing buffer (0.2% BSA in ice-cold PBS), the cells were then incubated with secondary antibody [1:1000 dilution of streptavadin–PE (phycoerythrin); 0.5 mg/ml; cat. no. 554061; BD Pharmingen) for 30 min. Cells were washed three times with FACS washing buffer before being subjected to FACS.

Stably transfected CTLL2 cells expressing wild-type or mutant hGM-CSFRα subunits and hβc were also sorted on the basis of hGM-CSFRα expression. To separate two distinct cell populations, FITC-positive and FITC-negative cells were sorted using the FACSsorter (BD Bioscience). Dead cells were identified and eliminated by propidium iodide staining and forward side-scatter gating respectively. Gates for each population were set so that the two subsets sorted based on FITC staining would not overlap when re-analysed. Sorted FITC-positive cells were immediately suspended in growth medium containing antibiotics and cultured to obtain sufficient numbers for further characterization.

Proliferation assays

[3H]Thymidine incorporation assays were performed, as described previously [34], to determine the capacity of the wild-type or mutant hGM-CSFRα subunits to deliver a proliferative signal in CTLL2 stable cell lines co-transfected with hβc in response to hGM-CSF.

RESULTS

Critical role of hGM-CSFRα D1 in binding GM-CSF

To analyse the functional role of hGM-CSFRα D1 we generated a truncated version of hGM-CSFRα lacking the first domain (hGM-CSFRα ΔD1; Figure 1A). To determine the capacity of this mutant hGM-CSFRα to bind GM-CSF, ‘cold’ saturation binding assays were performed on transiently transfected COS7 cells using 125I-labelled human GM-CSF and the Kd values were determined (Table 1). COS7 cells expressing the wild-type GM-CSFRα subunit bound 125I-labelled GM-CSF with low affinity (Table 1 and Figure 1B), consistent with the original description by Gearing et al. [38]. In contrast, when we expressed the truncated GM-CSFRα subunit hGM-CSFRα ΔD1 we were unable to detect 125I-labelled GM-CSF binding (Table 1). To ensure that the loss of GM-CSF binding did not arise from loss of receptor expression, we performed flow cytometry to check that the truncated receptor was expressed on the cell surface. As it is possible that the anti-(GM-CSFRα) antibody epitope is located in D1, we incorporated a FLAG epitope between the signal peptide and mature hGM-CSFRα ΔD1-coding sequence and expressed this variant in COS7 cells. Flow cytometry following anti-FLAG antibody staining showed that the D1-deleted hGM-CSFRα is expressed on the cell surface at the same level as the wild-type (see Figure 4). The bulk of the cells expressed detectable levels of receptor and a small subset expressed higher levels. The expression data indicate that deletion of the Ig-like domain does not compromise folding, trafficking or cell-surface expression. These observations suggest an obligatory role for the Ig-like domain in low-affinity GM-CSF binding, analogous to the role of D1 in the hIL-5Rα subunit in mediating binding to its cognate ligand. In contrast, IL-3Rα D1 contributes to high-affinity binding, but is not obligatory [29,39].

Table 1
hGM-CSF binding to COS7 cells expressing wild-type or mutant hGM-CSFRα subunits

Binding was determined using the ‘cold’ saturation binding assay, and the Kd values±S.E. were determined from co-analysis of data obtained from separate binding experiments using LIGAND [37]. −, no binding was detected above the background non-specific binding; n, number of experiments.

GM-CSFRα Kd (nM) n 
Wild-type 4.8±0.33 
ΔD1 − 
R49A 12.5±1.4 
V50A 7.5±0.6 
V51A − 
E52A 9.2±0.6 
R54A 34.2±2.6 
L55A 31±1.8 
S56A 7.4±1.3 
N57A 5.7±0.6 
N58A 12.5±1.3 
E59A 7.78±1 
C60A − 
S61A 3.98±0.3 
GM-CSFRα Kd (nM) n 
Wild-type 4.8±0.33 
ΔD1 − 
R49A 12.5±1.4 
V50A 7.5±0.6 
V51A − 
E52A 9.2±0.6 
R54A 34.2±2.6 
L55A 31±1.8 
S56A 7.4±1.3 
N57A 5.7±0.6 
N58A 12.5±1.3 
E59A 7.78±1 
C60A − 
S61A 3.98±0.3 

Deletion of hGM-CSFRα D1 disrupts GM-CSF binding

Figure 1
Deletion of hGM-CSFRα D1 disrupts GM-CSF binding

(A) Schematic models showing component domains of hGM-CSFRα and hGM-CSFRα ΔD1 (deletion of exons 2 and 3, 280 bp, which corresponds to amino acid residues 25–113, between IPEK24 and S114GREG). (B) Scatchard representation of hGM-CSF direct binding assays, depicting a representative ‘cold’ saturation binding assay performed on COS7 cells expressing the wild-type hGM-CSFRα alone.

Figure 1
Deletion of hGM-CSFRα D1 disrupts GM-CSF binding

(A) Schematic models showing component domains of hGM-CSFRα and hGM-CSFRα ΔD1 (deletion of exons 2 and 3, 280 bp, which corresponds to amino acid residues 25–113, between IPEK24 and S114GREG). (B) Scatchard representation of hGM-CSF direct binding assays, depicting a representative ‘cold’ saturation binding assay performed on COS7 cells expressing the wild-type hGM-CSFRα alone.

Binding properties of wild-type and mutant hGM-CSFRα in the presence of hβc

To study the role of GM-CSFRα D1 in the formation of the high-affinity GM-CSF–GM-CSFRα–hβc complex, we co-transfected COS7 cells with expression vectors encoding hβc and wild-type or hGM-CSFRα ΔD1 subunits and performed ‘hot’ saturation binding assays using 125I-labelled hGM-CSF. These binding studies were performed using hGM-CSFRα subunits that were not tagged with an N-terminal FLAG epitope. Cells expressing both wild-type hGM-CSFRα and hβc exhibited two GM-CSF-binding sites, as illustrated by the curvilinear Scatchard plot (Figure 2). The Kd values of 66 pM (high affinity) and 4.8 nM (low affinity) (Table 2) are consistent with those reported previously for the GM-CSFR system [6,38]. When COS7 cells were co-transfected with hGM-CSFRα ΔD1 and hβc, we observed a single GM-CSF-binding site of intermediate affinity (481 pM), whereas a low-affinity site corresponding to GM-CSF binding by GM-CSFRα alone could not be detected (Table 2 and Figure 2). The presence of the intermediate-affinity site, rather than a high-affinity GM-CSF-binding site, observed with hGM-CSFRα ΔD1 suggests that, although neither the hGM-CSFRα ΔD1 mutant nor wild-type hβc bind GM-CSF detectably alone, their co-expression enables ligand binding. These data suggest that the binding contributions of hβc partially compensate for the deficiency in binding of hGM-CSFRα ΔD1 to GM-CSF.

Scatchard representation of hGM-CSF ‘hot’ saturation binding assays

Figure 2
Scatchard representation of hGM-CSF ‘hot’ saturation binding assays

Scatchard plots depicting 125I-labelled GM-CSF ‘hot’ saturation binding data for COS7 cells expressing hβc plus wild-type or mutant GM-CSFRα ΔD1, V51A, R54A, L55A or C60A, as labelled on the Figure. Data from a representative ‘hot’ saturation binding experiment are shown in each plot with the line-of-best-fit determined by co-analysis of data from several binding experiments using LIGAND [37]. The derived Kd values are shown in Table 2.

Figure 2
Scatchard representation of hGM-CSF ‘hot’ saturation binding assays

Scatchard plots depicting 125I-labelled GM-CSF ‘hot’ saturation binding data for COS7 cells expressing hβc plus wild-type or mutant GM-CSFRα ΔD1, V51A, R54A, L55A or C60A, as labelled on the Figure. Data from a representative ‘hot’ saturation binding experiment are shown in each plot with the line-of-best-fit determined by co-analysis of data from several binding experiments using LIGAND [37]. The derived Kd values are shown in Table 2.

Table 2
hGM-CSF binding to COS7 cells expressing hβc and wild-type or mutant hGM-CSFRα subunits

Binding was determined using the ‘hot’ saturation binding assay and Kd values±S.E. were determined using LIGAND [37]. *To improve the accuracy of determination of the high-affinity Kd, the low-affinity Kd was fixed according to the values from Table 1. †High-affinity binding data sets for L55A gave both one-site and two-site models, consistent with residual though reduced binding for hGM-CSFRα L55A alone (Table 1). −, a one site binding model was statistically significant; n, number of experiments.

 Kd  
hGM-CSFRα High-affinity (pM) Low-affinity (nM) n 
Wild-type 65.7±8.8 4.8* 
ΔD1 481±39.5 − 
V51A 190±13 − 
R54A 211±8.4 − 
L55A 179±18 31* 
 148±7.6† − 
C60A 646±59.4 − 
 Kd  
hGM-CSFRα High-affinity (pM) Low-affinity (nM) n 
Wild-type 65.7±8.8 4.8* 
ΔD1 481±39.5 − 
V51A 190±13 − 
R54A 211±8.4 − 
L55A 179±18 31* 
 148±7.6† − 
C60A 646±59.4 − 

Identification of critical residues in hGM-CSFRα D1 for binding to hGM-CSF

The recent publication of the structure of the IL-13Rα1 complex (Figure 3D) [26] provided, for the first time, a good model for the α receptors interacting with their ligands and illustrated a clear role for the C′ strand within D1 in mediating binding to its ligands IL-13 and IL-4. As a previous study identified key residues in hIL-5Rα D1 involved in IL-5 binding [27], we used these data to verify that IL-13Rα1 was a good model for hIL-5Rα and its sister receptor hGM-CSFRα. As shown in Figures 3(A)–3(D), and Supplementary Figure S1(A) (available at http://www.BiochemJ.org/bj/426/bj4260307add.htm), the structural alignment of IL-13Rα1 and hIL-5Rα gave a precise positioning of the key IL-5-binding residues close to the relevant binding residues identified from the IL-13Rα1–IL-13 crystal structure [26]. Other expected structural features also aligned well. The lone cysteine residue in the Ig-like domain of hIL-5Rα was shown previously to be solvent-exposed by virtue of its covalent modification by isothiazolone compounds [42]. These data provide further validation for the suitability of hIL-13Rα1 (human IL-13Rα1) as a template for hIL-5Rα structural alignment, as this cysteine residue is predicted to lie in the E–F loop of hIL-5Rα D1, a solvent-exposed region of the structure (Supplementary Figure S1A). The disulfides between strands 2A and 2B and strands 2D and 2E aligned with conserved disulfides in IL-13Rα1. Similarly, there was alignment of the WSXWS motif and its associated RXR motif, both of which are involved in the Trp-Arg zipper. It was therefore likely that a structural alignment of GM-CSFRα with IL-13Rα1 would also be informative and this was carried out (Supplementary Figure S1B). It also showed conservation of expected structural features. A further guide to the position of the residues of hGM-CSFRα D1 likely to be involved in cytokine binding was the conserved exon structures of the genes, as an exon boundary was close to the critical residues in IL-13Rα1 and hIL-5Rα (Supplementary Figure S1A). The structural alignment of hGM-CSFRα also showed that the corresponding exon boundary was aligned with the same region (Supplementary Figure S1B).

Structural homology models of hGM-CSFRα and hIL-5Rα

Figure 3
Structural homology models of hGM-CSFRα and hIL-5Rα

(A) The structural model of the extracellular domain of GM-CSFRα (left) was generated using FUGUE [40] by homology with the IL-13Rα1 crystal structure (from PDB 3BPO). GM-CSF (PDB 1CSG) and GM-CSFRα were superimposed on to IL-13 and IL-13Rα1 respectively, from PDB 3BPO using the program Coot [41]. GM-CSF is drawn in grey and GM-CSFRα is in red. (B) Expansion of the boxed region from (A). Side chains of residues from D1 of hGM-CSFRα that were mutated in the present study due to their potential involvement in the interaction with the ligand are drawn as blue sticks and are labelled. (C) Homology model of the hIL-5Rα ectodomain, based on the crystal structure of IL-13Rα1, generated using FUGUE [40]. The side chains of residues from D1 that are critically involved in IL-5 interactions are drawn as blue sticks and are labelled. (D) The IL-13–IL-13Rα1 complex drawn from PDB 3BPO. The side chains of residues located at the Ig-like domain of IL-13Rα1 involved in an anti-parallel β-strand interaction with ligand are drawn as blue sticks and are labelled. Cartoon structural models were drawn using PyMOL (DeLano Scientific; http://www.pymol.org). (E) Multi-species alignment of D1 and a portion of D2 of GM-CSFRα. Conserved cysteine residues are highlighted in green. Potential binding residues around the predicted C′ strand of D1 of hGM-CSFRα are highlighted in turquoise, and the residues identified as involved in ligand binding in the present study are highlighted in red text and are reproduced below the multi-species alignment in red text. The predicted secondary structure features are marked above the sequence; green arrows represent β-strands.

Figure 3
Structural homology models of hGM-CSFRα and hIL-5Rα

(A) The structural model of the extracellular domain of GM-CSFRα (left) was generated using FUGUE [40] by homology with the IL-13Rα1 crystal structure (from PDB 3BPO). GM-CSF (PDB 1CSG) and GM-CSFRα were superimposed on to IL-13 and IL-13Rα1 respectively, from PDB 3BPO using the program Coot [41]. GM-CSF is drawn in grey and GM-CSFRα is in red. (B) Expansion of the boxed region from (A). Side chains of residues from D1 of hGM-CSFRα that were mutated in the present study due to their potential involvement in the interaction with the ligand are drawn as blue sticks and are labelled. (C) Homology model of the hIL-5Rα ectodomain, based on the crystal structure of IL-13Rα1, generated using FUGUE [40]. The side chains of residues from D1 that are critically involved in IL-5 interactions are drawn as blue sticks and are labelled. (D) The IL-13–IL-13Rα1 complex drawn from PDB 3BPO. The side chains of residues located at the Ig-like domain of IL-13Rα1 involved in an anti-parallel β-strand interaction with ligand are drawn as blue sticks and are labelled. Cartoon structural models were drawn using PyMOL (DeLano Scientific; http://www.pymol.org). (E) Multi-species alignment of D1 and a portion of D2 of GM-CSFRα. Conserved cysteine residues are highlighted in green. Potential binding residues around the predicted C′ strand of D1 of hGM-CSFRα are highlighted in turquoise, and the residues identified as involved in ligand binding in the present study are highlighted in red text and are reproduced below the multi-species alignment in red text. The predicted secondary structure features are marked above the sequence; green arrows represent β-strands.

The IL-13Rα1 complex showed contacts between residues Lys76 to Glu81 including the region of the C′ strand (Figures 3A and 3B) [26]. From the structural alignment model of GM-CSFRα, we identified 12 candidate GM-CSF-binding residues in the region of the predicted C′ strand of D1 and, subsequently, performed site-directed mutagenesis to generate individual alanine mutants (Table 1). In studies on cytokine class I receptors, alanine substitution is commonly used for the determination of residues whose side chains are involved in ligand binding [15], since alanine possesses a small hydrophobic side chain which is often found in solvent-exposed regions of proteins. Furthermore, alanine substitution does not disrupt the geometry of the main chain, and the methyl side chain does not cause electrostatic or steric interference.

Binding properties of hGM-CSFRα mutants containing mutations in the region of the putative C′ strand

Each of the 12 alanine mutants of residues in the region of the predicted C′ strand were transiently expressed in COS7 cells and subjected to the ‘cold’ saturation assay to ascertain the effect of the mutations on low-affinity hGM-CSF binding. Pro53 was not mutated in the present study, as this substitution is likely to result in major structural perturbation of the C–C′ loop, rendering interpretation of any loss of binding ambiguous. None of the GM-CSFRα subunits used in these experiments were tagged with the FLAG epitope at their N-termini. COS7 cells transfected with the cDNA encoding wild-type GM-CSFRα were found to bind GM-CSF with low affinity (Kd=4.8 nM), as described above, and the Kd values of R49A, V50A, E52A, S56A, N57A, N58A, E59A and S61A mutant hGM-CSFRα subunits were not greatly affected (Table 1). In contrast, the R54A and L55A hGM-CSFRα subunits showed a reduction in affinity for hGM-CSF binding of 7- and 6-fold respectively (Table 1), and complete abolition of hGM-CSF binding was observed with the V51A and C60A hGM-CSFRα subunits. These results implicate Val51 and Cys60 as key hGM-CSF-binding determinants and suggest that Arg54 and Leu55 also play an important additional role in hGM-CSF binding. To determine whether the residues critical for direct (low-affinity) hGM-CSF binding were also involved in high-affinity binding, COS7 cells were transiently co-transfected with expression vectors encoding wild-type hβc and wild-type or mutant hGM-CSFRα subunits, and ‘hot’ saturation binding assays were performed in which increasing amounts of 125I-labelled hGM-CSF were added (Table 2 and Figure 2). Interestingly, all of these mutants were capable of binding hGM-CSF with intermediate-affinity (Kd=148–646 pM compared with wild-type affinity of 66 pM). Mutation of Cys60 gave the strongest effect, reducing high-affinity binding by a similar amount to D1 deletion.

Cell-surface expression of mutant hGM-CSFRα receptors

To verify that the loss of ligand binding with the V51A, R54A, L55A and C60A mutant hGM-CSFRα receptors did not result from a perturbation of the receptor translation, folding or presence on the cell surface, flow cytometry was used to detect mutant hGM-CSFRα cell-surface expression. COS7 cells transiently expressing wild-type or mutant hGM-CSFRα subunits were stained with purified anti-(human CD116) (clone 4H1) and a FITC-conjugated secondary antibody and were examined by indirect flow cytometry (Figure 4). The level of cell-surface expression of the V51A, R54A, L55A and C60A mutant hGM-CSFRα was observed to be equivalent to wild-type hGM-CSFRα expression with only small variations observed because of differences in transfection efficiency. COS7 cells transfected with the empty vector, pcEX-V3-Xba, exhibited no increase in fluorescence intensity above background when stained with secondary antibody alone (see Figure 4). Additionally, hGM-CSFRα mutants were capable of participating in high-affinity hGM-CSF-binding complexes (Table 2), providing further evidence for their presence on the cell surface. However, it was also important to examine whether the Ig-like domain may play a critical role in reorientating receptor subunits to initiate activation of intracellular signalling by the high-affinity receptor complex, as investigated below.

Cell-surface expression of wild-type and mutant hGM-CSFRα subunits as detected by flow cytometry

Figure 4
Cell-surface expression of wild-type and mutant hGM-CSFRα subunits as detected by flow cytometry

COS7 cells transfected with cDNAs encoding wild-type or mutant GM-CSFRα after 64–68 h were stained with purified anti-(human CD116) antibody and FITC-conjugated secondary antibody [FITC-conjugated rat anti-(mouse IgG1) antibody]. Cells stained with only the FITC-conjugated antibody were used as a control for each assay (depicted by the grey-shaded area). hGM-CSFRα ΔD1 cell-surface expression was examined using a modified form of the receptor bearing an N-terminal FLAG-tag, followed by staining with an anti-FLAG antibody. COS7 cells transfected with empty vector gave the same results as the control. FACS analyses of wild-type and mutant GM-CSFRα subunit expression in CTLL2 stable cell lines (both unsorted and sorted for GM-CSFRα levels) are shown in Supplementary Figure S2(B) (at http://www.BiochemJ.org/bj/426/bj4260307add.htm).

Figure 4
Cell-surface expression of wild-type and mutant hGM-CSFRα subunits as detected by flow cytometry

COS7 cells transfected with cDNAs encoding wild-type or mutant GM-CSFRα after 64–68 h were stained with purified anti-(human CD116) antibody and FITC-conjugated secondary antibody [FITC-conjugated rat anti-(mouse IgG1) antibody]. Cells stained with only the FITC-conjugated antibody were used as a control for each assay (depicted by the grey-shaded area). hGM-CSFRα ΔD1 cell-surface expression was examined using a modified form of the receptor bearing an N-terminal FLAG-tag, followed by staining with an anti-FLAG antibody. COS7 cells transfected with empty vector gave the same results as the control. FACS analyses of wild-type and mutant GM-CSFRα subunit expression in CTLL2 stable cell lines (both unsorted and sorted for GM-CSFRα levels) are shown in Supplementary Figure S2(B) (at http://www.BiochemJ.org/bj/426/bj4260307add.htm).

Critical role of D1 and Cys60 of hGM-CSFRα in receptor activation

Although the formation of a high-affinity complex is a prerequisite for receptor activation, it is not necessarily sufficient for signalling. We therefore considered it important to correlate the binding data with an unequivocal signalling response, in this case proliferation. hβc DNA was stably transfected into the mIL-2-dependent lymphoid cell line CTLL2, via the pEF-IRES-P vector, which allows for selection by puromycin resistance. The pEF-IRES expression vector utilizes the strong EF-1α (elongation factor-1α) promoter, which is highly active in CTLL2 cells and the expression of the puromycin-resistance gene is indicative of the expression of the cloned receptor. Verification of hβc expression was carried out by FACS (Supplementary Figure S2A, available at http://www.BiochemJ.org/bj/426/bj4260307add.htm). Puromycin-resistant cells were subsequently separately transfected with wild-type hGM-CSFRα or the mutant hGM-CSFRα subunits (ΔD1, V51A, R54A, L55A or C60A) in the vector pEF-IRES-N, which encodes G418 resistance. Antibiotic-resistant cells were selected in the presence of mIL-2, and proliferation assays carried out to determine responsiveness to hGM-CSF after mIL-2 was removed. hGM-CSFRα ΔD1 and C60A exhibited a greatly reduced response to GM-CSF compared with cells expressing the wild-type receptor (Figures 5A and 5B). They not only required markedly higher concentrations of ligand for growth stimulation, but the maximum stimulation of the cells was also much lower than that of the wild-type. These mutants were therefore studied in detail.

Proliferation of CTLL2 cell lines stably expressing wild-type or mutant hGM-CSFRα plus hβc

Figure 5
Proliferation of CTLL2 cell lines stably expressing wild-type or mutant hGM-CSFRα plus hβc

Plots depict the percentage of maximal growth response to a serial dilution of hGM-CSF for CTLL2 cells stably co-expressing wild-type hβc with: (A) wild-type or ΔD1 GM-CSFRα; (B) wild-type, V51A, R54A, L55A or C60A GM-CSFRα; and (C) wild-type, ΔD1 or C60A GM-CSFRα using cells sorted by FACS for GM-CSFRα expression. The data shown in (AC) are representative of triplicate experiments. (D) Scatchard plots of 125I-labelled GM-CSF ‘hot’ saturation binding data for CTLL2 cells (unsorted or sorted) expressing hβc plus wild-type, ΔD1 or C60A mutant GM-CSFRα as labelled in the Figure. Data from a representative ‘hot’ saturation binding experiment are shown in each plot with the line of best fit determined from co-analysis of multiple experiments using LIGAND [37]. The derived Kd values are shown in Table 3.

Figure 5
Proliferation of CTLL2 cell lines stably expressing wild-type or mutant hGM-CSFRα plus hβc

Plots depict the percentage of maximal growth response to a serial dilution of hGM-CSF for CTLL2 cells stably co-expressing wild-type hβc with: (A) wild-type or ΔD1 GM-CSFRα; (B) wild-type, V51A, R54A, L55A or C60A GM-CSFRα; and (C) wild-type, ΔD1 or C60A GM-CSFRα using cells sorted by FACS for GM-CSFRα expression. The data shown in (AC) are representative of triplicate experiments. (D) Scatchard plots of 125I-labelled GM-CSF ‘hot’ saturation binding data for CTLL2 cells (unsorted or sorted) expressing hβc plus wild-type, ΔD1 or C60A mutant GM-CSFRα as labelled in the Figure. Data from a representative ‘hot’ saturation binding experiment are shown in each plot with the line of best fit determined from co-analysis of multiple experiments using LIGAND [37]. The derived Kd values are shown in Table 3.

Duplicate CTLL2 cell lines expressing wild-type hGM-CSFRα, hGM-CSFRα ΔD1 and hGM-CSFRα C60A, together with hβc, were generated from separate transfections. Following selection of antibiotic-resistant cells in the presence of mIL-2, the stable cell lines were sorted by FACS to enrich for high hGM-CSFRα expression, and both the sorted and unsorted stable cell lines were compared. Expression of wild-type and mutant hGM-CSFRα was verified by FACS (Supplementary Figure S2B). In each case the sorting further enhanced the level of hGM-CSFRα expression in the cell populations. The hGM-CSFRα ΔD1 expressed in these experiments possessed an N-terminal FLAG epitope to allow sorting using an anti-FLAG antibody, whereas wild-type- and C60A hGM-CSFRα-expressing cells were sorted using an anti-(hGM-CSFRα) antibody. Continued high expression of hβc in each of the stable cell lines was also verified (results not shown).

‘Hot’ saturation binding assays were used to measure high-affinity hGM-CSF binding (Table 3 and Figure 5D). The unsorted wild-type hGM-CSFRα/wild-type hβc-expressing cells gave a Kd value of 91pM for high-affinity binding, which decreased to 50 pM in the FACS-sorted cells. The number of low-affinity-binding sites (attributable to hGM-CSFRα) was significantly higher in the sorted cells, as expected, but the number of high-affinity-binding sites was only slightly increased (Table 3). In the case of hGM-CSFRα ΔD1, the Kd value for high-affinity binding in unsorted cells was much higher than that for their wild-type counterparts (314 pM), and there was a 25-fold reduction in the number of binding sites. This was despite similar receptor expression as measured by FACS (Supplementary Figure S2A), suggesting that the mutant hGM-CSFRα complex with hβc may be unstable. Although the hGM-CSFRα subunit expressed in this CTLL2 cell line bears an N-terminal FLAG-tag, a comparable Kd value for high-affinity binding (481 pM) was observed in ‘hot’ saturation hGM-CSF-binding assays when hGM-CSFRα ΔD1 lacking a FLAG-tag was co-expressed with wild-type hβc in COS7 cells (Figure 2 and Table 2). With the sorted hGM-CSFRα ΔD1/wild-type hβc-expressing cells, the number of high-affinity-binding sites was greatly increased, together with a lowering of the Kd value to 135 pM. This suggests that some stabilization of the complex with hβc is occurring in the cells expressing higher levels of hGM-CSFRα ΔD1. hGM-CSFRα C60A/wild-type hβc-expressing cells also showed a much higher Kd value for high-affinity binding (451 pM) and an 8-fold reduction in the number of binding sites in unsorted cells. The number of binding sites increased 4-fold in the sorted cells, but the Kd value for high-affinity binding did not change (Table 3).

Table 3
hGM-CSF binding to CTLL2 cells expressing hβc and wild-type or mutant hGM-CSFRα subunits

*Each stable CTLL2 transfectant population was sorted by FACS for elevated expression of hGM-CSFRα, and the binding characteristics of sorted (+) and unsorted (−) cells were compared. †Binding was determined using the ‘hot’ saturation binding assay and Kd values±S.E. were determined using LIGAND [37]. ‡To improve the accuracy of determination of the high-affinity Kd, the low-affinity Kd was fixed at 4.8 nM from Table 1. n, number of experiments.

  High-affinity Low-affinity  
FACS-sorted (hGM-CSFRα)* hGM-CSFRα Kd (pM)† Sites/cell (nKd (nM) Sites/cell (nn 
− Wild-type 91±17.5 3035±315 4.8‡ 13610±1497 
Wild-type 50±20 3613±568 4.8‡ 34085±3215 
− ΔD1 314±19 120±60 − − 
ΔD1 135±10.6 6949±390 − − 
− C60A 451±54 361±32 − − 
C60A 465±55.8 1433±144 − − 
  High-affinity Low-affinity  
FACS-sorted (hGM-CSFRα)* hGM-CSFRα Kd (pM)† Sites/cell (nKd (nM) Sites/cell (nn 
− Wild-type 91±17.5 3035±315 4.8‡ 13610±1497 
Wild-type 50±20 3613±568 4.8‡ 34085±3215 
− ΔD1 314±19 120±60 − − 
ΔD1 135±10.6 6949±390 − − 
− C60A 451±54 361±32 − − 
C60A 465±55.8 1433±144 − − 

The sorted stable transfectants were also tested in proliferation assays to assess their responsiveness to hGM-CSF after mIL-2 was removed. Sorting for higher levels of expression of hGM-CSFRα ΔD1 improved the growth signalling compared with unsorted cells, but it was still noticeably abnormal in terms of both responsiveness and the stimulation plateau that was reached (Figure 5C). Sorting did not improve the abnormal growth signalling shown by hGM-CSFRα C60A (Figure 5C). The abnormal growth signalling shown by cells co-expressing C60A or ΔD1 hGM-CSFRα and wild-type hβc were similar, both showing a dramatic reduction in the stimulation plateau as well as reduced responsiveness. The data indicate defective growth signalling by the two mutants, and that hGM-CSFRα D1 is critical for both normal binding and for receptor activation.

An alignment of extracellular domains of GM-CSFRα from different species was carried out to examine the conservation of important residues including the ones highlighted in the present study. Part of this alignment is shown in Figure 3(E). The cysteine residues in D2 comprising the two disulfides between strands A and B and strands D and E were found to be absolutely conserved. There was no similar evidence of disulfides in D1, where cysteine residues vary in number and position across species. The key residue Cys60 was highly conserved and Val51 was conserved as either valine or isoleucine. Arg54 and Leu55 also some showed conservation between species, with Arg54 conserved largely as arginine or lysine and Leu55 as leucine, valine or isoleucine.

DISCUSSION

To date, the definition of ligand-binding interfaces in the cytokine class I receptor family has focused predominantly on epitopes formed by CRMs, as typified by the GHR. The contributions of other domains within cytokine receptors, such as the Ig-like domain, have until recently remained poorly understood in the absence of structural information. The recently reported X-ray crystal structures of IL-4 and IL-13 in complex with the IL-13Rα1 ectodomain have provided key insights into the contribution of IL-13Rα1 D1, in combination with its canonical adjacent CRM, in binding its cytokine ligands [26]. These crystal structures provide a template for extending our knowledge of how cytokine receptors with analogous topologies, such as the hβc co-receptors, the IL-3Rα, IL-5Rα and GM-CSFRα, may utilize their D1 in ligand recognition and receptor activation. The importance of the Ig-like domains of IL-3Rα and IL-5Rα in binding their cognate ligands is now well-established [2729,39], but oddly, the role of the GM-CSFRα D1 has escaped scrutiny. Furthermore, no electron density was observed in the recently described GM-CSF–GMCSFRα–βc X-ray crystal structure [31] to clarify the role of GM-CSFRα D1. In light of our recent studies establishing critical roles for the IL-3Rα subunit Ig-like domain in ligand binding, receptor activation and downstream signalling to control haemopoietic cell differentiation [29], we describe in the present studies the essential role that the related hGM-CSFRα D1 plays in mediating hGM-CSF binding and governing receptor activation.

Initially, we set out to determine whether GM-CSFRα D1 is critically required for: (i) directly binding hGM-CSF with low affinity; (ii) high-affinity binding of hGM-CSF in the presence of hβc; (iii) receptor activation and the initiation of intracellular signalling pathways; and/or (iv) intracellular trafficking and cell-surface localization, a function attributed previously to D1 of IL-6Rα [43]. We deleted exons 2 and 3 of hGM-CSFRα, to eliminate D1 (residues 25–113) without disrupting the signal peptide encoded by exon 1, expressed the resulting construct in COS7 cells in the absence or presence of the shared receptor subunit hβc, and assessed binding to 125I-labelled hGM-CSF. Our data clearly demonstrate the necessity for GM-CSFRα D1 for direct binding to hGM-CSF, since no hGM-CSF binding was detectable in these experiments, despite confirmation that the mutant receptor was expressed on the cell surface at wild-type levels (Table 1 and Figure 4). In the presence of hβc, low-affinity GM-CSF binding was again not observed, but rather a single binding site corresponding to reduced hGM-CSF binding relative to wild-type high-affinity binding (Table 2 and Figure 2). The existence of a single binding site is illustrated by the linearity of fit in Scatchard plots for binding experiments when hGM-CSFRα ΔD1 and hβc were co-expressed, compared with the curvilinearity of wild-type hGM-CSFRα and hβc Scatchard plots, characteristic of a two-site fit (Figure 2). Thus additional binding sites provided by hβc partially compensate for the inability of the hGM-CSFRα ΔD1 to bind directly to hGM-CSF. Similar compensation effects by the presence of hβc have been reported previously. For example, mutation of Asp112 in GM-CSF abrogated detectable binding to hGM-CSFRα alone, yet this mutant exhibited near wild-type activity in cells co-expressing hGM-CSFRα and hβc [30]. Notably, the affinity of hGM-CSF binding by the hGM-CSFRα ΔD1 and hβc subunits was reduced compared with the wild-type hGM-CSFRα and hβc complex, thereby indicating a critical role for the hGM-CSFRα Ig-like domain in stabilizing GM-CSF binding within the signalling complex. To the best of our knowledge, these data represent the first evidence establishing the importance of the Ig-like domain of hGM-CSFRα in hGM-CSF binding.

Although COS7 cells co-expressing the hGM-CSFRα ΔD1 and hβc showed medium-affinity binding of hGM-CSF, when these subunits were stably expressed in the factor-dependent CTLL2 cell line, a defective growth response was observed. The observation of hGM-CSF binding by cells co-expressing hGM-CSFRα ΔD1 and hβc confirms their cell-surface expression, in agreement with the FACS analysis, indicating that deletion of D1 did not disrupt the hGM-CSFRα translation, folding or transport to the membrane. This result emphasizes the importance of GM-CSFRα D1 not only in ligand binding, but also in the formation of an efficient signalling receptor complex, and underscores the complexity of the receptor system, where high-affinity GM-CSF binding is a necessary, but not sufficient, step to instigate receptor activation. The data reported in the present study suggest that GM-CSF binding by the hGM-CSFRα–hβc complex is necessary to reorientate the receptor subunits to initiate intracellular signal transduction and that this process critically relies on the Ig-like domain of the hGM-CSFRα subunit. The concept that activation of receptor signalling requires appropriate receptor orientation was originally established for the EPO receptor [44]. Recent FRET (fluorescence resonance energy transfer) studies of IL-5 binding to its receptor suggest binding by a pre-existing receptor hetero-oligomer induces reorientation of receptor ectodomains and rearrangement of the transmembrane and cytoplasmic regions to initiate signal transduction [45].

In order to deduce the epitope within hGM-CSFRα D1 that plays a key role in GM-CSF recognition and receptor activation, we generated a structural model of the hGM-CSFRα ectodomain by homology with the recently reported IL-13Rα1–IL-13 complex crystal structure [26]. IL-13Rα1 and IL-13 are suitable structural templates for hGM-CSFRα and hGM-CSF respectively, as both α chains share the Ig-like–CRM topology in their ectodomains and both cytokines are short-chain four-helix-bundle proteins with an up-up-down-down topology [26,46]. To test the validity of our approach we initially carried out an alignment with hIL-5Rα as the critical residues in D1 of this receptor have been identified. The alignment showed correct positioning of key structural features as well as the key D1-binding residues Asp55 and Asp56 within the predicted C′ strand and the adjacent Tyr57 (Figure 3C) [27]. Interestingly, the alignment also indicated that hIL-5Rα may utilize a surface-bound disulfide in D3 analogous to that in IL-13Rα1, which is involved in IL-13 binding. From a similar alignment with GM-CSFRα, 12 candidate GM-CSF contact residues were identified in the region of the predicted hGM-CSFRα D1 C′ strand (Figures 3A and 3B), a region established previously to be important in the related IL-5Rα, as alanine substitution of Asp55 and Asp56 within the predicted C′ strand and the adjacent Tyr57 disrupted IL-5 binding [27]. We examined the contributions of these 12 residues in the region of the putative hGM-CSFRα D1 C′ strand to hGM-CSF binding and receptor activation by characterizing their alanine mutant analogues. Interestingly, alanine substitution of hGM-CSFRα Val51 and Cys60 resulted in the complete loss of detectable low-affinity GM-CSF binding, and mutation of the neighbouring Arg54 and Leu55 caused respective 8- and 7-fold reductions in low-affinity GM-CSF binding. When these mutant hGM-CSFRα subunits were co-expressed with hβc, the most severe disruption of high-affinity hGM-CSF binding was observed for C60A GM-CSFRα, which exhibited a 10-fold lower affinity for hGM-CSF, comparable with the effect of D1 deletion. The reduction in GM-CSF high-affinity binding observed for hGM-CSFRα C60A was borne out in studies of its capacity to deliver a proliferative signal when installed into a factor-dependent cell line, CTLL2, stably expressing wild-type hβc. These cells showed a retarded growth response in a hGM-CSF titration and a low maximum stimulation. This growth assay confirmed that Cys60 within hGM-CSFRα D1 is a key determinant of both GM-CSF recognition and receptor activation, whereas analogous CTLL2 cell lines co-expressing hβc and V51A, R54A and L55A mutant hGM-CSFRα subunits exhibited near wild-type proliferative responses in a hGM-CSF titration. These data indicate that Val51, Arg54 and Leu55 of hGM-CSFRα D1 individually play minor roles in facilitating hGM-CSF binding, but do not appear to play important roles in receptor activation. The role played by the critical residue Cys60 in ligand binding and subunit reorientation within the signalling complex is not completely clear. Alignment of GM-CSF ectodomains from a number of species indicated that Cys60 is highly conserved (Figure 3E), but gives no indication of other conserved cysteine residues that could partner it in disulfide bond formation. Similarly, it does not align with cysteine residues involved in disulfides in hIL-13Rα1 D1 (Supplementary Figure S1). Thus no clear evidence was obtained of involvement of Cys60 in disulfide bond formation. A postulated mechanism for hGM-CSFR activation involves disulfide cross-linking between an unidentified cysteine residue in hGM-CSFRα and hβc D1 D–E loop [47]. It seems unlikely that Cys60 would be involved in such a cross-link as it plays a critical role in ligand binding. A more complete understanding of the role of Cys60 awaits elucidation of an X-ray crystal structure with electron density for the hGM-CSFRα D1 ligand interaction.

Abbreviations

     
  • (h)βc

    (human) common β-subunit

  •  
  • CRM

    cytokine-receptor homology module

  •  
  • EPO

    erythropoietin

  •  
  • FBS

    fetal bovine serum

  •  
  • GH

    growth hormone

  •  
  • GHR

    GH receptor

  •  
  • (h)GM-CSF

    (human) granulocyte/macrophage colony-stimulating factor

  •  
  • (h)GM-CSFR

    (human) GM-CSF receptor

  •  
  • IL

    interleukin

  •  
  • IL-3R

    IL-3 receptor

  •  
  • (h)IL-5R

    (human) IL-5 receptor

  •  
  • IL-6R

    IL-6 receptor

  •  
  • (h)IL-13R

    (human) IL-13 receptor

  •  
  • D1

    the N-terminal Ig-like domain of IL-13Rα1, GM-CSFRα, IL-5Rα or IL-3Rα, α subunits of GM-CSFR, IL-5R or IL-3R respectively

  •  
  • Kd

    dissociation constant

  •  
  • mIL-2

    murine IL-2

AUTHOR CONTRIBUTION

Shamaruh Mirza, Jinglong Chen, James Murphy and Ian Young designed the research. Shamaruh Mirza, Andrew Walker and James Murphy performed the research. Shamaruh Mirza, Jinglong Chen and Ian Young analysed the data. Shamaruh Mirza, James Murphy and Ian Young wrote the paper.

We thank J. Olsen for excellent technical advice; Dr Steve Hobbs for providing the pEF-IRES expression vectors; and Tracy Willson (The Walter and Eliza Hall Institute for Medical Research, Parkville, Australia) for providing the hGM-CSFRα cDNA. We are also grateful to the John Curtin School of Medical Research Microscopy & Cytometry Facility for FACS and ACRF Biomolecular Resource Facility, the Australian National University for automated sequencing of plasmids.

FUNDING

This work was supported by the National Health and Medical Research Council of Australia (NHMRC) [grant numbers 471481]. J.M.M. is the recipient of a NHMRC CJ Martin (Biomedical) Fellowship.

References

References
1
Kopf
M.
Brombacher
F.
Hodgkin
P. D.
Ramsay
A. J.
Milbourne
E. A.
Dai
W. J.
Ovington
K. S.
Kohler
G.
Young
I. G.
Matthaei
K. I.
IL-5-deficient mice have a developmental defect in CD5+ B-1 cells and lack eosinophilia but have normal antibody and cytotoxic T cell responses
Immunity
1996
, vol. 
4
 (pg. 
15
-
24
)
2
Foster
P. S.
Hogan
S. P.
Ramsay
A. J.
Matthaei
K. I.
Young
I. G.
Interleukin-5 deficiency abolishes eosinophilia, airways hyperreactivity, and lung damage in a mouse asthma model
J. Exp. Med.
1996
, vol. 
183
 (pg. 
195
-
201
)
3
Lantz
C. S.
Boesiger
J.
Song
C. H.
Mach
N.
Kobayashi
T.
Mulligan
R. C.
Nawa
Y.
Dranoff
G.
Galli
S. J.
Role for interleukin-3 in mast-cell and basophil development and in immunity to parasites
Nature
1998
, vol. 
392
 (pg. 
90
-
93
)
4
Hamilton
J. A.
Colony-stimulating factors in inflammation and autoimmunity
Nat. Rev. Immunol.
2008
, vol. 
8
 (pg. 
533
-
544
)
5
Notarengelo
L. D.
Pessach
I.
Out of breath: GM-CSFRα mutations disrupt surfactant homeostasis
J. Exp. Med.
2008
, vol. 
205
 (pg. 
2693
-
2697
)
6
Hayashida
K.
Kitamura
T.
Gorman
D. M.
Arai
K.
Yokota
T.
Miyajima
A.
Molecular cloning of a second subunit of the receptor for human granulocyte-macrophage colony-stimulating factor (GMCSF): reconstitution of a high-affinity GM-CSF receptor
Proc. Natl. Acad. Sci. U.S.A.
1990
, vol. 
87
 (pg. 
9655
-
9659
)
7
Kitamura
T.
Sato
N.
Arai
K.
Miyajima
A.
Expression cloning of the human IL-3 receptor cDNA reveals a shared β subunit for the human IL-3 and GM-CSF receptors
Cell
1991
, vol. 
66
 (pg. 
1165
-
1174
)
8
Tavernier
J.
Devos
R.
Cornelis
S.
Tuypens
T.
Van der Heyden
J.
Fiers
W.
Plaetinck
G.
A human high affinity interleukin-5 receptor (IL5R) is composed of an IL5-specific α chain and a β chain shared with the receptor for GM-CSF
Cell
1991
, vol. 
66
 (pg. 
1175
-
1184
)
9
Takaki
S.
Murata
Y.
Kitamura
T.
Miyajima
A.
Tominaga
A.
Takatsu
K.
Reconstitution of the functional receptors for murine and human interleukin-5
J. Exp. Med.
1993
, vol. 
177
 (pg. 
1523
-
1529
)
10
Sakamaki
K.
Miyajima
I.
Kitamura
T.
Miyajima
A.
Critical cytoplasmic domains of the common β subunit of the human GMCSF, IL-3 and IL-5 receptors for growth, signal transduction and tyrosine phosphorylation
EMBO J.
1992
, vol. 
11
 (pg. 
3541
-
3549
)
11
Quelle
F. W.
Sato
N.
Witthuhn
B. A.
Inhorn
R. C.
Eder
M.
Miyajima
A.
Griffin
J. D.
Ihle
J. N.
JAK2 associates with the βc chain of the receptor for granulocyte-macrophage colony-stimulating factor, and its activation requires the membrane-proximal region
Mol. Cell. Biol.
1994
, vol. 
14
 (pg. 
4335
-
4341
)
12
Itoh
T.
Muto
A.
Watanabe
S.
Miyajima
A.
Yokota
T.
Arai
K.
Granulocyte-macrophage colony-stimulating factor provokes RAS activation and transcription of c-fos through different modes of signaling
J. Biol. Chem.
1996
, vol. 
271
 (pg. 
7587
-
7592
)
13
Itoh
T.
Liu
R.
Yokota
T.
Arai
K.
Watanabe
S.
Definition of the role of tyrosine residues of the common β subunit regulating multiple signaling pathways of granulocyte-macrophage colony-stimulating factor receptor
Mol. Cell. Biol.
1998
, vol. 
18
 (pg. 
742
-
752
)
14
de Groot
R. P.
Coffer
P. J.
Koenderman
L.
Regulation of proliferation, differentiation and survival by the IL-3/IL-5/GM-CSF receptor family
Cell. Signalling
1998
, vol. 
10
 (pg. 
619
-
628
)
15
Murphy
J. M.
Young
I. G.
IL-3, IL-5, and GM-CSF signaling: crystal structure of the human β-common receptor
Vitam. Horm.
2006
, vol. 
74
 (pg. 
1
-
30
)
16
Somers
W.
Ultsch
M.
De Vos
A. M.
Kossiakoff
A. A.
The X-ray structure of a growth hormone-prolactin receptor complex
Nature
1994
, vol. 
372
 (pg. 
478
-
481
)
17
Syed
R. S.
Reid
S. W.
Li
C.
Cheetham
J. C.
Aoki
K. H.
Liu
B.
Zhan
H.
Osslund
T. D.
Chirino
A. J.
Zhang
J.
, et al. 
Efficiency of signaling through cytokine receptors depends critically on receptor orientation
Nature
1998
, vol. 
395
 (pg. 
511
-
516
)
18
Tamada
T.
Honjo
E.
Maeda
Y.
Okamoto
T.
Ishibashi
M.
Tokunaga
M.
Kuroki
R.
Homodimeric cross-over structure of the human granulocyte colony-stimulating factor (GCSF) receptor signaling complex
Proc. Natl. Acad. Sci. U.S.A.
2006
, vol. 
103
 (pg. 
3135
-
3140
)
19
Chow
D.
He
X.
Snow
A. L.
Rose-John
S.
Garcia
K. C.
Structure of an extracellular gp130 cytokine receptor signaling complex
Science
2001
, vol. 
291
 (pg. 
2150
-
2155
)
20
Yoon
C.
Johnston
S. C.
Tang
J.
Stahl
M.
Tobin
J. F.
Somers
W. S.
Charged residues dominate a unique interlocking topography in the heterodimeric cytokine interleukin-12
EMBO J.
2000
, vol. 
19
 (pg. 
3530
-
3541
)
21
Hage
T.
Sebald
W.
Reinemer
P.
Crystal structure of the interleukin-4/receptor α chain complex reveals a mosaic binding interface
Cell
1999
, vol. 
97
 (pg. 
271
-
281
)
22
Carr
P. D.
Gustin
S. E.
Church
A. P.
Murphy
J. M.
Ford
S. C.
Mann
D. A.
Woltring
D. M.
Walker
I.
Ollis
D. L.
Young
I. G.
Structure of the complete extracellular domain of the common β subunit of the human GM-CSF, IL-3, and IL-5 receptors reveals a novel dimer configuration
Cell
2001
, vol. 
104
 (pg. 
291
-
300
)
23
Murphy
J. M.
Ford
S. C.
Wiedemann
U. M.
Carr
P. D.
Ollis
D. L.
Young
I. G.
A novel functional epitope formed by domains 1 and 4 of the human common β-subunit is involved in receptor activation by granulocyte-macrophage colony-stimulating factor and interleukin-5
J. Biol. Chem.
2003
, vol. 
278
 (pg. 
10572
-
10577
)
24
Murphy
J. M.
Ford
S. C.
Olsen
J. E.
Gustin
S. E.
Jeffrey
P. D.
Ollis
D. L.
Young
I. G.
Interleukin-3 binding to the murine βIL-3 and human βc receptors involves functional epitopes formed by domains 1 and 4 of different protein chains
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
26500
-
26508
)
25
Boulanger
M. J.
Chow
D. C.
Brevnova
E. E.
Garcia
K. C.
Hexameric structure and assembly of the interleukin-6/IL-6 α-receptor/gp130 complex
Science
2003
, vol. 
300
 (pg. 
2101
-
2104
)
26
LaPorte
S. L.
Juo
Z. S.
Vaclavikova
J.
Colf
L. A.
Qi
X.
Heller
N. M.
Keegan
A. D.
Garcia
K. G.
Molecular and structural basis of cytokine receptor pleiotropy in the interleukin-4/13 system
Cell
2008
, vol. 
132
 (pg. 
259
-
272
)
27
Cornelis
S.
Plaetinck
G.
Devos
R.
Van der Heyden
J.
Tavernier
J.
Sanderson
C. J.
Guisez
Y.
Fiers
W.
Detailed analysis of the IL-5–IL-5Rα interaction: characterization of crucial residues on the ligand and the receptor
EMBO J.
1995
, vol. 
14
 (pg. 
3395
-
3402
)
28
Ishino
T.
Pasut
G.
Scibek
J.
Chaiken
I.
Kinetic interaction analysis of human interleukin 5 receptor α mutants reveals a unique binding topology and charge distribution for cytokine recognition
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
9547
-
9556
)
29
Chen
J.
Olsen
J.
Ford
S.
Mirza
M.
Walker
A.
Murphy
J. M.
Young
I. G.
A new isoform of interleukin-3 receptor α with novel differentiation activity and high affinity binding mode
J. Biol. Chem.
2009
, vol. 
284
 (pg. 
5763
-
5773
)
30
Rajotte
D.
Cadieux
C.
Haman
A.
Wilkes
B. C.
Clark
S. C.
Hercus
T.
Woodcock
J. A.
Lopez
A.
Hoang
T. D.
Crucial role of the residue R280 at the F′-G′ loop of the human granulocyte/macrophage colony-stimulating factor receptor α chain for ligand recognition
J. Exp. Med.
1997
, vol. 
185
 (pg. 
1939
-
1950
)
31
Hansen
G.
Hercus
T. R.
McClure
B. J.
Stomski
F. C.
Dottore
M.
Powell
J.
Ramshaw
H.
Woodcock
J. M.
Xu
Y.
Guthridge
M.
, et al. 
The structure of the GM-CSF receptor complex reveals a distinct mode of cytokine receptor activation
Cell
2008
, vol. 
134
 (pg. 
496
-
507
)
32
Miller
J.
Malek
T. R.
Leonard
W. J.
Greene
W. C.
Shevach
E. M.
Germain
R. N.
Nucleotide sequence and expression of a mouse interleukin 2 receptor cDNA
J. Immunol.
1985
, vol. 
134
 (pg. 
4212
-
4217
)
33
Hobbs
S.
Jitrapakdee
S.
Wallace
J. C.
Development of a bicistronic vector driven by the human polypeptide chain elongation factor 1α promoter for creation of stable mammalian cell lines that express very high levels of recombinant proteins
Biochem. Biophys. Res. Commun.
1998
, vol. 
252
 (pg. 
368
-
372
)
34
Gustin
S. E.
Church
A. P.
Ford
S. C.
Mann
D. A.
Carr
P. D.
Ollis
D. L.
Young
I. G.
Expression, crystallization and derivatization of the complete extracellular domain of the βc subunit of the human IL-5, IL-3 and GM-CSF receptors
Eur. J. Biochem.
2001
, vol. 
268
 (pg. 
2905
-
2911
)
35
Murphy
J. M.
Soboleva
T. A.
Mirza
S.
Ford
S. C.
Olsen
J. E.
Chen
J.
Young
I. G.
Clarification of the role of N-glycans on the common β-subunit of the human IL-3, IL-5 and GM-CSF receptors and the murine IL-3 β-receptor in ligand-binding and receptor activation
Cytokine
2008
, vol. 
42
 (pg. 
234
-
242
)
36
McPherson
G. A.
Analysis of radioligand binding experiments. A collection of computer programs for the IBM PC
J. Pharmacol. Methods
1985
, vol. 
14
 (pg. 
213
-
228
)
37
Munson
P. J.
Rodbard
D.
LIGAND: a versatile computerized approach for characterization of ligand-binding systems
Anal. Biochem.
1980
, vol. 
107
 (pg. 
220
-
239
)
38
Gearing
D. P.
King
J. A.
Gough
N. M.
Nicola
N. A.
Expression and cloning of a receptor for human granulocyte-macrophage colony stimulating factor
EMBO J.
1989
, vol. 
8
 (pg. 
3667
-
3676
)
39
Barry
S. C
Korpelainen
E.
Sun
Q.
Stomski
F. C.
Moretti
P. A. B.
Wakao
H.
D'Andrea
R. J.
Vadas
M. A.
Lopez
A. F.
Goodall
G. J.
Roles of the N and C terminal domains of the interleukin-3 receptor α chain in receptor function
Blood
1997
, vol. 
89
 (pg. 
842
-
852
)
40
Shi
J.
Blundell
T. L.
Mizuguchi
K.
FUGUE: sequence-structure homology recognition using environment-specific substitution tables and structure-dependent gap penalties
J. Mol. Biol.
2001
, vol. 
310
 (pg. 
243
-
257
)
41
Emsley
P.
Cowtan
K.
Coot: model-building tools for molecular graphics
Acta Crystallogr. D Biol. Crystallogr.
2004
, vol. 
60
 (pg. 
2126
-
2132
)
42
Devos
R.
Guisez
Y.
Plaetinck
G.
Cornelis
S.
Tavernier
J.
van der Heyden
J.
Foley
L. H.
Scheffler
J. E.
Covalent modification of the interleukin-5 receptor by isothiazolones leads to inhibition of the binding of interleukin-5
Eur. J. Biochem.
1994
, vol. 
225
 (pg. 
635
-
640
)
43
Vollmer
P.
Oppmann
B.
Voltz
N.
Fischer
M.
Rose-John
S.
A role for the immunoglobulin-like domain of the human IL-6 receptor: intracellular protein transport and shedding
Eur. J. Biochem.
1999
, vol. 
263
 (pg. 
438
-
446
)
44
Livnah
O.
Johnson
D. L.
Stura
E. A.
Farrell
F. X.
Barbone
F. P.
You
Y.
Liu
K. D.
Goldsmith
M. A.
He
W.
Krause
C. D.
, et al. 
An antagonist peptide-EPO receptor complex suggests that receptor dimerization is not sufficient for activation
Nat. Struct. Biol.
1998
, vol. 
5
 (pg. 
993
-
1004
)
45
Zaks-Zilberman
M.
Harrington
A. E.
Ishino
T.
Chaiken
I. M.
Interleukin-5 receptor subunit oligomerization and rearrangement revealed by fluorescence resonance energy transfer imaging
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
13398
-
13406
)
46
Walter
M. R.
Cook
W. J.
Ealick
S. E.
Nagabhushan
T. L.
Trotta
P. P.
Bugg
C. E.
Three-dimensional structure of recombinant human granulocyte-macrophage colony-stimulating factor
J. Mol. Biol.
1992
, vol. 
224
 (pg. 
1075
-
1085
)
47
Stomski
F. C.
Woodcock
J. M.
Zacharakis
B.
Bagley
C. J.
Sun
Q.
Lopez
A. F.
Identification of a Cys motif in the common β chain of the interleukin 3, granulocyte-macrophage colony-stimulating factor, and interleukin 5 receptors essential for disulfide-linked receptor heterodimerization and activation of all three receptors
J. Biol. Chem.
1998
, vol. 
273
 (pg. 
1192
-
1199
)

Supplementary data