Báo cáo Y học: A functional role of the membrane-proximal extracellular domains of the signal transducer gp130 in heterodimerization with the leukemia inhibitory factor receptor

doi:10.1046/j.1432-1033.2002.02941.x

Eur. J. Biochem. 269, 2716–2726 (2002) (cid:1) FEBS 2002

A functional role of the membrane-proximal extracellular domains of the signal transducer gp130 in heterodimerization with the leukemia inhibitory factor receptor

Andreas Timmermann, Andrea Ku¨ ster, Ingo Kurth, Peter C. Heinrich and Gerhard Mu¨ ller-Newen

Institut fu¨r Biochemie, Rheinisch-Westfa¨lische Technische Hochschule Aachen, Germany

Replacement of the gp130 domains by the corresponding domains of the related GCSF receptor either restores weak LIF responsiveness (D4-GCSFR), leads to constitutive activation of gp130 (D5-GCSFR) or results in an inactive receptor (D6-GCSFR). Mutation of a specific cysteine in D5 of gp130 (C458A) leads to constitutive heterodimerization with the LIFR and increased sensitivity towards LIF stimulation. Based on these findings, a functional model of the gp130–LIFR heterodimer is proposed that includes contacts between D5 of gp130 and the corresponding domain D7 of the LIFR and highlights the requirement for both receptor dimerization and adequate receptor orienta- tion as a prerequisite for signal transduction.

Keywords: cytokines; receptors; signal transduction; leuke- mia inhibitory factor; gp130. gp130 is the common signal transducing receptor subunit of interleukin (IL)-6-type cytokines. gp130 either homodimer- izes in response to IL-6 and IL-11 or forms heterodimers with the leukemia inhibitory factor (LIF) receptor (LIFR) in response to LIF, oncostatin M (OSM), ciliary neurotrophic factor (CNTF), cardiotrophin-1 (CT-1) or cardiotrophin- like cytokine resulting in the onset of cytoplasmic tyrosine phosphorylation cascades. The extracellular parts of both gp130 and LIFR consist of several Ig-like and fibronectin type III-like domains. The role of the membrane-distal domains of gp130 (D1, D2, D3) and LIFR in ligand binding is well established. In this study we investigated the func- tional significance of the membrane-proximal domains of gp130 (D4, D5, D6) in respect to heterodimerization with LIFR. Deletion of each of the membrane-proximal domains of gp130 (D4, D5 and D6) leads to LIF unresponsiveness.

transcription)

the Janus kinase (Jak) family. Upon ligand binding the associated Jaks become activated by transphosphorylation and phosphorylate tyrosine residues in the cytoplasmic part of the receptor. These phosphotyrosines serve as docking sites for signalling molecules that, in most cases, also become phosphorylated. Most importantly, STAT (signal factors are transducer and activator of recruited to the receptor, dimerize upon phosphorylation and translocate into the nucleus to induce expression of target genes [5].

Secretion of mediators by cells that are recognized by specific receptors on target cells is a basic mechanism of intercellular communication. The molecular mechanism by which binding of the ligand to the receptor on the plasma membrane leads to the onset of cytoplasmic signal trans- duction cascades has gained considerable attention during recent years. In the case of receptors that span the membrane only once, ligand induced receptor dimerization has been accepted as the main mechanism for receptor activation [1]. Only recently, several reports suggested that some receptors may exist as preformed dimers or multimers that switch from an inactive to an active conformation upon ligand binding [2,3].

Hematopoietic cytokine receptors [4] consist of an extracellular part, a single transmembrane region, and a cytoplasmic part that is devoid of any intrinsic enzymatic activity but constitutively associates with tyrosine kinases of

Based on the architecture of the extracellular part, hematopoietic cytokine receptors can be subdivided into two groups. The extracellular parts of short cytokine receptors like erythropoetin recepter (EpoR), growth hormone receptor (GHR), prolactinR, IL-2Rb or IL-4R consist of only a single cytokine binding module (CBM). The CBM is made up of two fibronectin type III-like (FNIII) domains containing some characteristic conserved motifs in their primary structures. Several structures of CBMs of short cytokine receptors bound to their ligands have been solved showing that in the active receptor dimer the membrane-proximal domains are juxtaposed in a well- defined orientation [6,7].

Correspondence to G. Mu¨ ller-Newen, Institut fu¨ r Biochemie, Rheinisch-Westfa¨ lische Technische Hochschule Aachen, Pauwelsstr. 30, D-52057 Aachen, Germany. Fax: + 49 241 8082428, Tel.: + 49 241 8088860, E-mail: Mueller-Newen@RWTH-Aachen.de Abbreviations: CBM, cytokine binding module; FNIII, fibronectin type III-like; GCSF, granulocyte colony stimulating factor; GH, growth hormone; IL, interleukin; Jak, Janus kinase; LIF, leukemia inhibitory factor; OSM, oncostatin M; STAT, signal transducer and activator of transcription. Note: a web site is available at http://www.biochem.rwth-aachen.de (Received 28 February 2002, accepted 18 April 2002)

The extracellular parts of complex cytokine receptors like gp130, LIFR, leptinR or GCSFR contain at least one CBM and additional FNIII- and Ig-like domains. The cytokine receptor gp130 consists of an Ig-like domain (D1), followed by a CBM (D2, D3) and three FNIII-like domains (D4, D5, and D6) (Fig. 1) [8]. The role of the membrane-distal domains (D1–D3) in ligand binding has been well estab- lished by functional and structural studies. In response

Heterodimerization of gp130 with LIFR (Eur. J. Biochem. 269) 2717

(cid:1) FEBS 2002

rization. We proposed a particular role for D5 of gp130 in respect to proper receptor spacing and orientation [18]. In this study, using gp130 deletion mutants, point mutants and chimeric receptors, the role of the individual membrane- proximal domains of gp130 in heterodimerization with the LIFR is evaluated. A cysteine to alanine mutation in D5 of gp130 in combination with the LIFR leads to a weak constitutive activity and an elevated response to stimulation with LIF. A model for the gp130/LIFR interaction is proposed, in which D5 of gp130 contacts domain 7 of the LIFR.

M A T E R I A L S A N D M E T H O D S

Enzymes, proteins, antibodies, chemicals, and cell culture media

Enzymes were purchased from Roche (Mannheim, Ger- many) and protein A–Sepharose was obtained from Amersham (Freiburg, Germany). Fugene was obtained from Roche (Mannheim, Germany). DMEM and antibi- otics were obtained from Life Technologies (Eggenstein, Germany); fetal bovine serum was provided by Seromed (Berlin, Germany). [a-32P]deoxyATP was purchased from Hartmann Analytic (Braunschweig, Germany). Human rIL-6 was expressed in Escherichia coli, refolded, and purified as described by Arcone et al. [19]. The specific activity was 108 units per mg of protein in the B9 cell proliferation assay [20]. Soluble IL-6R (sIL-6R) [21] was expressed in insect cells as described previously. The gp130 mAbs B-P4 and B-P8 and the LIFR mAb 10B2 were generated as described elsewhere [22,23]. The polyclonal LIFR antiserum sc-659 was obtainded from New England Biolabs (Frankfurt/Main, Germany). All other Abs were purchased from Dako (Hamburg, Germany). NaCl/Pi buffer contained 200 mM NaCl, 2.5 mM KCl, 8 mM Na2HPO4, and 1.5 mM KH2PO4.

Fig. 1. Schematic representation of gp130, gp130c and LIFR. The predicted structural organizations of gp130, gp130c and LIFR are shown. Black lines in the CBM indicate conserved cysteine residues, black bars the WSXWS motifs. The Ig-like domains and the mem- brane-proximal FNIII domains are labelled. The extracellular domains of gp130 and LIFR are numbered from domain 1 (D1) to domain 6 (D6) or domain 1 (D1) to domain 8 (D8), respectively. In the cyto- plasmic part of gp130c, the amino-acid residues following the Jak binding sites (box1 and box2, gray boxes) were replaced by the strongly and specifically STAT1-activating motif YDKPH of the interferon-c receptor.

Cell culture

BaF3-cells, a murine pro-B lymphocyte line, were cultured in RPMI 1640 containing 10% fetal bovine serum, 100 mgÆL)1 streptomycin, 60 mgÆL)1 penicillin and 5% conditioned medium from X63Ag-653 BPV-mIL-3 myeloma cells as a source of IL-3. Simian monkey kidney cells (COS7) were cultured in DMEM supplemented with 10% fetal bovine serum, 100 mgÆL)1 streptomycin, and 60 mgÆL)1 penicillin. Cells were grown at 37 (cid:2)C in a water-saturated atmo- sphere at 5% CO2. BaF3 transfectants were cultured in the presence of 0.5 lgÆmL)1 hygromycin if transfected with the LIFR expression vector pSBC1/2-LIFR/Hygro and 1 mgÆmL)1 G418 if transfected with a pSVL-gp130-expres- sion vector together with pSV2-Neo.

All cells were regularly checked for the absence of mycoplasma infection using PCR detection of mycoplasma DNA. to cytokines like IL-6 and IL-11 that lead to gp130 homodimerization, two different epitopes of gp130 are involved in ligand binding: the Ig-like domain and the CBM [9–11]. In the case of LIF-induced heterodimerization of LIFR with gp130, the cytokine first binds to the Ig-like domain of the LIFR [12,13]. gp130 is recruited to the LIF– LIFR complex via its CBM without involvement of its Ig-like domain [14,15]. The cytokine oncostatin M (OSM) first binds to gp130 and then induces heterodimerization with LIFR [12] or OSMR [16].

Plasmid construction

Construction of gp130 wild-type and domain mutant expression vectors D4, D5, D6 and D5 has been described elsewhere [18]. The domain exchange mutants gp130 D4 and D6 were cloned analogously after amplifying the DNA Besides the CBM and Ig-like domains both gp130 and LIFR share three further membrane-proximal FNIII-like domains as a common structural feature [17]. In a previous study we established a functional role for each of the membrane-proximal domains of gp130 for receptor activa- tion in response to ligands that lead to gp130 homodime-

2718 A. Timmermann et al. (Eur. J. Biochem. 269)

(cid:1) FEBS 2002

Flow cytometry

Cells were collected, washed and resuspended in cold NaCl/ Pi containing 5% fetal bovine serum and 0.1% sodium azide. Subsequently, cells were incubated on ice with 10 lgÆmL)1 gp130 antibodies B-P4 or B-P8 or 10 lgÆmL)1 LIFR antibody 10B2. Cells were washed with cold NaCl/Pi/ azide and incubated with R-phycoerythrin-conjugated anti- (mouse IgG) Fab-fragment at a 1 : 50 dilution. Again, cells were washed with cold NaCl/Pi/azide and then resuspended in 400 lL NaCl/Pi/azide followed by flow cytometry analysis using a FACScalibur (Beckton Dickinson).

Electrophoretic mobility shift assay (EMSA)

Cells were incubated at 37 (cid:2)C for 15 min in the presence of IL-6/sIL-6R, LIF, OSM or left unstimulated. BaF3-cells were stimulated with 25 ngÆmL)1 IL-6 and 1 lgÆmL)1 sIL- 6R or 50 ngÆmL)1 LIF or 50 ngÆmL)1 OSM. COS7 cells were stimulated with 12.5 ngÆmL)1 IL-6 and 500 ngÆmL)1 sIL-6R, 20 ngÆmL)1 LIF and 4 ngÆmL)1 OSM. Where indicated, cells were preincubated for 2 h in the presence of 500 lM 2-mercaptoethanol prior to stimulation. Prepar- ation of nuclear extracts and EMSAs were performed as described previously [25]. A double stranded size-inducible element (SIE) oligonucleotide derived from the c-fos promoter (m67SIE; 5¢-GATCCGGGAGGGATTTACGG GGAAATGCTG-3¢) was used as 32P-labeled probe [26]. The protein–DNA complexes were separated on a 4.5% polyacrylamide gel containing 7.5% glycerol. The electro- phoresis was performed using 0.25 · NaCl/Tris/borate buffer at 20 VÆcm)1. encoding domains 4 and 6 of GCSFR using the oligo- nucleotides: 5¢-ACTACCGAACGGGCCCCCGGGGTC AGACTGGACACATGG-3¢ and 5¢-TCGGGCCATGGC ATGCCCGGGGGTCAGAGCTGGG-3¢ for amplifica- tion of D4 of GCSFR and 5¢-TACTCTCAAGAAATG CCCGGGTCCCATGCCCCAGAG-3¢ and 5¢-GCCCAG GATGATGTGTAGCTCCCCGGGCTCTGGGGTCAA GGT-3¢ for D6 of GCSFR (the XmaI sites are underlined) as PCR primers. Starting point for cloning of gp130 C458A, C466A and C491A was the full length human gp130 cDNA cloned into the XhoI and BamHI site of the eukaryotic expression vector pSVL lacking the EcoRI site (gp130- pSVLDEco). Using this vector as template, for each point mutant two fragments were amplified. In a first reaction, the DNA was amplified using the primer pSVL(sense) and an antisense primer containing the mutation. A second PCR- fragment was generated using the primers pSVL(antisense) and a sense primer with the corresponding mutation. These fragments were isolated, mixed and served as templates for a fusion PCR using the primers pSVL(sense) and pSVL(anti- sense). The reaction products were digested with the restriction enzymes Xho I and BstEII and cloned into the expression vector gp130-pSVLDEco. The primers used for the PCR reactions were: pSVL(sense) 5¢-GTGTTACTT CTGCTCT-3¢; pSVL(antisense) 5¢-TCTAGTTGTGGTT TGT-3¢; C458A(sense) 5¢-ATACTTGAGTGGGCTGTG TTATCAG-3¢; C458A(antisense) 5¢-ATCTGATAACAC AGCCCACTCAAGTAT-3¢; C466A(sense) 5¢-GATAAA GCACCCGCTATCACAGACTGG-3¢; C466A(antisense) 5¢-CCAGTCTGTGATAGCGGGTGCTTTATCTG-3¢; C491A(sense) 5¢-GCAGAGAGCAAAGCCTATTTGAT AACAG-3¢ and C491(antisense) 5¢-TGTTATCAAATAG GCTTTGCTCTCTG-3¢.

Coimmunoprecipitation of LIFR/gp130 complexes

PCRs were performed applying standard procedures. All plasmids were sequenced using an ABI Prism Automated sequencer (Applied Biosystems).

The full-length human LIFR cDNA was cloned into pSBC-1 to yield the mammalian expression vector pSBC- LIFR as previously described [15]. For the transfection of BaF3-cells, the bicistronic expression vector pSBC1/2- LIFR/Hygro was used [15,24].

Transfection of cells

Transiently transfected COS7 cells were stimulated for 15 min with IL-6/sIL-6R, LIF or OSM as described or left unstimulated. Where indicated, cells were preincubated for 2 h in the presence of 500 lM 2-mercaptoethanol prior to stimulation. Immediately after stimulation, cells were washed twice with ice-cold NaCl/Pi containing 100 lM vanadate. After addition of 600 lL lysis buffer (10% glycerol, 0.25% Brij-96, 50 mM Tris/HCl, 50 lM Na3VO4, 100 lM EDTA, 1 mM phenylmethanesulfonyl fluoride, 1 mgÆL)1 aprotinin, 1 mgÆL)1 leupeptin, pH 8.0) the cells were collected and lysed for 30 min in a microcentrifuge tube. The lysate was centrifuged for 1 min at 3000 r.p.m. in an Eppendorf centrifuge and the supernatant was transferred into a new centrifuge tube. Following incubation of the lysate with 1.6 lg sc-659 antiserum for 12 h at 4 (cid:2)C 15 mg protein A– sepharose was added. After incubation for 12 h at 4 (cid:2)C, the complexes were washed twice with NaCl/Tris/borate/Non- idet P40 buffer, resuspended in Laemmli-buffer, incubated at 95 (cid:2)C for 5 min and separated on a 7% SDS polyacrylamide gel under reducing conditions followed by electroblotting.

Immunoblotting and enhanced chemiluminescence (ECL) detection

Plasmid DNA was transfected into BaF3-cells by electro- poration. Thirty micrograms of the bicistronic LIFR expression vector pSBC1/2-LIFR-Hygro were electropo- rated into 3.5 · 106 cells in 0.8 mL medium applying a single 70-ms pulse at 200 V. Selection with hygromycin (0.5 mgÆmL)1) was initiated 24 h after transfection. Selected BaF3 clones were screened for the presence of membrane- bound LIFR proteins by flow cytometry. For transfection of gp130 constructs, 28 lg of the gp130 expression vector were coelectroporated with 2 lg of pSV2neo as described above. Selection with G418 (3 mgÆmL)1) was initiated 24 h after transfection. For transfection, either untransfected BaF3-cells or cells previously transfected with pSBC1/2- LIFR-Hygro were used. Selected BaF3 clones were screened for the presence of membrane-bound gp130 proteins by flow cytometry.

Immunoprecipitated proteins separated by SDS/PAGE were transferred to a poly(vinylidene difluoride) membrane by a semidry electroblotting procedure [27]. Poly(vinylidene difluoride) membranes were blocked in a solution of 20 mM COS7 cells were transiently transfected using the Fugene method. Efficiency of transfection was analysed by flow cytometry.

Heterodimerization of gp130 with LIFR (Eur. J. Biochem. 269) 2719

(cid:1) FEBS 2002

Tris/HCl (pH 7.6), 137 mM NaCl, 0.1% Nonidet-P40 containing 10% bovine serum albumin and probed with antibody, followed by incubation with horseradish peroxi- dase conjugated secondary antibody. Immunoreactive pro- teins were detected by chemiluminescence using the ECL-kit (Amersham, UK) following the manufacturer’s instruc- tions.

R E S U L T S A N D D I S C U S S I O N

Each of the three membrane-proximal domains of gp130 is required for signal transduction in response to LIF and OSM

Transfected cells were stimulated with IL-6/sIL-6R, LIF or OSM and subsequently activation of STAT1 was analysed by EMSA (Fig. 2B, right panel). In mock-trans- fected cells only a weak response to IL-6/sIL-6R and OSM was observed due to endogenously expressed receptors resulting in a low-level activation of mainly STAT3. Cells transfected with gp130c showed an increased STAT1 response to IL-6, and less pronounced to OSM. The LIF response remained unchanged. Transfection of LIFR alone did not significantly change the sensitivity of the cells to any of the cytokines. Transfection of both gp130c and LIFR led to a prominent response of the cells to all three cytokines. No STAT1-activation after cytokine stimulation could be detected when one of the gp130c deletion constructs was coexpressed with the LIFR. In COS7 cells under conditions of receptor overexpression, as in BaF3-cells, each of the membrane-proximal domains of gp130 is necessary for the formation of signal transducing heterodimeric complexes of gp130 and LIFR.

To investigate the role of the membrane-proximal FNIII- domains of gp130 in signal transduction through hetero- dimeric complexes with the LIFR, mutants of gp130, in which single FNIII-domains are deleted were generated lacking either D4 (gp130-D4), D5 (gp130-D5) or D6 (gp130- D6) [18]. These gp130 mutants were coexpressed with the LIFR in different cell types. The STAT-activation after stimulation with the cytokines IL-6, LIF or OSM was used as a measure of signal transduction through the analysed complexes.

Cells of the murine pre-B cell line BaF3 do not express endogenous gp130 or LIFR. After stably transfecting these cells with the respective cDNAs, cell surface expression of both receptors was detected. After stable transfection of the deletion constructs gp130D4, gp130D5 or gp130D6 together with the LIFR expression vector in BaF3-cells, the surface expressions of both receptors were similar to those detected for wild-type gp130/LIFR transfected cells (Fig. 2A, upper panel).

Two explanations for this finding can be discussed. The first is based on the identical domain architecture of LIFR and gp130 in the membrane-proximal six domains. This is likely to result in the same distance between the cell surface and the ligand-binding epitopes of both receptors. Deletion of a single domain in the membrane-proximal part of gp130 leads to a shift of the receptor areas involved in ligand binding closer to the membrane, resulting in the inability of the receptor chains to form an active receptor dimer. Additionally, the membrane-proximal domains can act as contact sites between the signal transducing receptor chains or can permit the signal competent conformation of gp130 homo- or heterodimers by adjusting a defined position towards each other. Thus, deletion of a membrane-proximal domain of gp130 may be without consequence on ligand binding but lead to a larger distance or a twist of the cytoplasmic parts of the receptors responsible for signal transduction.

Replacement of single membrane-proximal FNIII-domains of gp130 by corresponding domains of GCSFR leads to different effects on signal transduction

After stimulation of these cells with IL-6/sIL-6R, none of the analysed mutants showed a STAT activation similar to wild-type receptors (Fig. 2A, lower panel, right). This confirms the previously reported inactivity of the deletion mutants in response to the gp130-homodimerizing cytokine IL-6 [18]. Interestingly, also the formation of active hetero- dimers with wild-type LIFR in response to LIF or OSM is strongly reduced or abolished by deletion of individual membrane-proximal domains of gp130. Thus, in BaF3-cells, each of the membrane-proximal domains of gp130 is necessary for the efficient formation of a signal transducing heterodimeric complex of gp130 and the LIFR.

To investigate, if the function of the membrane-proximal domains of gp130 is limited to ensure the correct spacing between the CBM and the membrane, each of the domains was replaced by the corresponding domain of the GCSFR. The replacement is assumed to compensate for the shift of the ligand-binding epitopes of the receptors. The domain architecture of GCSFR is identical to that of gp130; moreover these receptors share 46% sequence homology. The gp130 constructs with exchanged individual FNIII- domains were introduced into the gp130c-construct result- ing in the mutants gp130D4c and gp130D6c, respectively. The construction of the corresponding mutant gp130D5c has been previously described [18].

To ensure that the measured receptor activation after cytokine stimulation does not depend on the analysed cellular environment, the deletion mutants were expressed together with the LIFR in COS7 cells. In a previous report [15], we established a system that allows the study of gp130 mutants together with the LIFR in COS7 cells despite the presence of low amounts of endogenous wild-type receptors. To achieve this, the cytoplasmic tyrosine motifs of gp130 that predominantly recruit STAT3 were replaced by the STAT1 recruiting motif of the interferon-c receptor result- ing in a chimeric protein designated gp130c. In order to investigate the role of the membrane-proximal domains of gp130 in receptor activation, the FNIII-domain deletions were introduced into the gp130c-construct (gp130-D4c, gp130-D5c or gp130-D6c) [18]. Each of these constructs was cotransfected with the LIFR into COS7 cells. Enhanced receptor surface expression was detected by flow cytometry (Fig. 2B, left panel). Upon co-expression of the gp130 chimeras with the LIFR in COS7 cells, both receptors were expressed on the cell surface as detected by flow cytometry in amounts similar to those of gp130 wild-type (data not shown). Signal trans- duction was measured by STAT1 activation in an EMSA (Fig. 3). The exchange of individual membrane-proximal FNIII-domains of gp130 by the corresponding domains of the GCSFR resulted in a complex signal transduction

2720 A. Timmermann et al. (Eur. J. Biochem. 269)

(cid:1) FEBS 2002

stimulation with cytokines. These results are in line with observations previously made in COS7 cells transfected with the gp130D5c-construct alone [18]. When coexpressed pattern. Cells transfected with gp130D5c together with the LIFR showed a prominent STAT1-activation independ- ently of stimulation. This activation was not enhanced by

Heterodimerization of gp130 with LIFR (Eur. J. Biochem. 269) 2721

(cid:1) FEBS 2002

with the LIFR, the gp130D6c chimera showed no STAT1-activation after stimulation with cytokines. After stimulation with IL-6/sIL-6R or OSM, cells that express LIFR together with gp130D4c showed no STAT1-activa- tion. In contrast, these cells showed a pronounced STAT1 activation upon stimulation with LIF.

Fig. 3. STAT1 activation in COS7 cells transiently transfected with LIFR and either gp130D4c, gp130D5c or gp130D6c in response to various cytokines. Forty-eight hours after transfection cells were sti- mulated as described in Fig. 1B as indicated. Nuclear extracts were prepared and activated STAT1 homodimers were detected by EMSA. A representative of three independent experiments is shown.

transduction of

In the heterodimeric LIFR/gp130 receptor system, different demands are posed to the individual FNIII domains for signal transduction. Because domain 4 of gp130 can be replaced by a similar domain of a different receptor without abrogation of signal transduction, this points to a spacer role of the domain. Intriguingly, signal transduction of the D4-GCSFR chimera occurs only after stimulation with LIF, while after OSM stimulation no STAT activation can be found in the transfected cells. In previous experiments [15] we were able to show that different epitopes in the gp130 CBM are required for after LIF- and OSM-induced STAT activation. The difference in signal the D4-GCSFR chimera after stimulation with LIF and OSM could point to further epitopes positioned C-terminally to the CBM that play specific roles in activation of the receptor by these two cytokines.

The ligand independent activation of the D5-GCSFR chimera occurs in both the absence and the presence of LIFR with identical intensities (compare with [18]). This suggests that the constitutive activation of this mutant receptor is due to the formation of homomeric gp130D5 complexes without involvement of LIFR. The observation of constitutive gp130 activation after replacement of D5 led to the proposal of a model for gp130 activation [18]. In this model, D5 is the site for direct contact of two gp130 molecules.

Analysis of dimerization of gp130 mutants: constructs lacking domains 5 or 6 do not heterodimerize with the LIFR in response to LIF

Fig. 2. STAT activation in cells expressing gp130 deletion mutants in response to various cytokines. (A) STAT activation in BaF3-cells stably transfected with LIFR and either gp130, gp130D4, gp130D5 or gp130D6 in response to various cytokines. (Upper panel) Cells were analysed for receptor surface expression by flow cytometry. Cells were incubated with gp130 antibody B-P8 (light gray histograms) or with LIFR antibody 10B2 (dark gray histograms) followed by phycoerythrin-conjugated secondary antibody. As a negative control, mock-transfected cells were treated in the same way (black histograms). The receptor surface expressions of the cells used for the EMSA in the lower panel are shown. After transfection of the cells with pSVL-gp130 or pSBC1/2-LIFR/Hygro, the encoded proteins can be detected on the cell surface in similar amounts. (Lower panel) Stably transfected cells were stimulated for 15 min with IL-6 (25 ngÆmL)1 in the presence of 1 lgÆmL)1 sIL-6R), LIF (50 ngÆmL)1) or OSM (50 ngÆmL)1) or left unstimulated (–) as indicated. Nuclear extracts were prepared and activated STAT3 and STAT1 homodimers as well as STAT1/3 heterodimers were detected by EMSA after binding to a labelled oligonucleotide probe (m67SIE). A representative of three independent experiments is shown. (B) STAT activation in COS7 cells transiently transfected with gp130c, LIFR, LIFR/gp130c, LIFR/gp130D4c, LIFR/ gp130D5c or LIFR/gp130D6c in response to various cytokines. (Left) Forty-eight hours after transfection cells were analysed for receptor surface expression by flow cytometry. Cells were incubated with gp130 antibody B-P8 (gp130) or with LIFR antibody 10B2 (LIFR) followed by phycoerythrin-conjugated secondary antibody (black histograms). As a negative control, mock-transfected cells were treated in the same way (gray histograms). The receptor surface expressions of the cells used for the EMSA in the right panel are shown. Surface expression of gp130c after transfection of pSVL-gp130c does not influence the LIFR surface expression. Transfection of the LIFR expression vector results in increased LIFR surface expression without affecting gp130 expression (upper row, left and central histograms). Consequently, transfection of both, gp130c and LIFR led to strongly increased surface expression of both receptors (upper row, right histograms). The surface expression of the gp130 deletion constructs was similar with the one of gp130c and did not interfere with LIFR-expression (lower histograms). (Right) Forty-eight hours after transfection cells were stimulated for 15 min with IL-6 (12.5 ngÆmL)1 in the presence of 500 ngÆmL)1 sIL-6R), LIF (20 ngÆmL)1) or OSM (4 ngÆmL)1) or left unstimulated (–) as indicated. Nuclear extracts were prepared and activated STAT1 homodimers were detected by EMSA after binding to a labelled oligonucleotide probe (m67SIE). A representative of three independent experiments is shown.

a second ligand binding or signal transducing receptor chain is a substantial point to judge its biological activity. To investigate the LIF dependence and structural requirements for dimerization of the signal transducing receptor chains gp130 and LIFR, a coprecipitation assay was established. The amounts of cell surface expressed receptors in stably transfected BaF3-cells were too small to achieve a copre- cipitation of gp130 with LIFR independently of the lysis buffer used for the disintegration of the cells (data not shown). Because of the high expression levels of the transiently transfected receptors in COS7 cells, they were used for the analysis of dimerization between LIFR and gp130 or gp130 mutants (Fig. 4A). In mock-transfected COS7 cells, a weak coprecipitation of LIFR and gp130 could be detected after stimulation of the cells with LIF (lane 2). This is due to the endogenous expression of both receptor chains on these cells. After transfection of both In addition to the ability of an extracellular receptor mutant to transduce a signal, its propensity to form a complex with

2722 A. Timmermann et al. (Eur. J. Biochem. 269)

(cid:1) FEBS 2002

molecular mass of 190 and 175 kDa appeared on the Western blot. They represented differently glycosylated forms of the LIFR in different steps of protein maturation. The slower migrating form (190 kDa) corresponds to the reported molecular mass of LIFR and is assumed to be the cell surface expressed receptor [12].

In order to analyse the ability of gp130 domain mutants to form heterodimers with LIFR, the deletion mutants gp130D5c and gp130D6c and the chimeric receptors gp130D5c and gp130D6c were coexpressed with the LIFR in COS7 cells (Fig. 4B). The only monoclonal antibody with sufficient sensitivity for the detection of gp130 in a coprecipitation experiment maps to domain 4 of the extracellular part of the receptor [28]. Therefore, analysis of the ability of the mutants gp130D4 and gp130D4 to form complexes with the LIFR by coprecipitation was not possible with the experimental procedure used in this study. As the formation of a high affinity complex of LIF, LIFR and gp130 is a prerequisite for signal transduction upon LIF stimulation, a coprecipitation analysis of LIFR and gp130D4 is not meaningful as this chimeric receptor is able to transduce a signal in response to LIF.

Fig. 4. Coprecipitation of LIFR with wild-type and mutant gp130. (A) Coprecipitation of gp130 with the LIFR is ligand dependent. COS7 cells were transfected with LIFR and wild-type gp130 or LIFR and gp130c or were left untransfected. Forty-eight hours after transfection cells were stimulated for 15 min with LIF (50 ngÆmL)1) or left unstimulated (–). After lysis of the cells with Brij buffer the LIFR was precipitated by addition of 1 lgÆmL)1 of the specific antiserum sc-659 that is directed against the 19 C-terminal amino acids of the receptor and thus does not interfere with ligand binding and extracellular receptor dimerization. The immunoprecipitated proteins were separ- ated by gel electrophoresis on a 7% SDS/PAGE gel followed by transfer to a poly(vinylidene difluoride) membrane. Detection of gp130 was performed with the monoclonal antibody B-P4. After removing of the antibodies from the blot, LIFR was detected using the specific LIFR-antiserum. (B) gp130 mutants lacking domains 5 or 6 do not heterodimerize with LIFR in response to LIF. Forty-eight hours after transfection of COS7 cells with LIFR and the gp130 chimeras copre- cipitation and detection of LIFR and gp130 was performed as des- cribed in (A).

Stimulation-independent formation of complexes with the LIFR could be not detected for any of the gp130 mutants analysed. All gp130c domain mutants showed a decreased coprecipitation with the LIFR compared to wild- type gp130c. Deletion of domain 6 in gp130 led to a complete loss of coprecipitation of this mutant with the LIFR. Compared with the deletion mutants gp130D5 and the respective domain replacement mutants gp130D6, gp130D5 and gp130D6 showed an increased coprecipitation of LIFR. In all cells transfected with gp130 mutants, endogenous wild-type gp130 was coprecipitated with LIFR after LIF stimulation. This served as a control for proper coprecipitation conditions.

In the case of LIF, ligand binding is a two step process. First, the cytokine is bound by the specific LIFR with low affinity (Kd ¼ 1–3 · 10)9 M). Then, engagement of gp130 to this low affinity complex leads to the formation of a signal transducing trimer, in which LIF is bound with high affinity (Kd ¼ 1–20 · 10)11 M) to both receptor chains [12]. Under the conditions used for the coprecipitation experiments there is no indication of ligand-independent preassociation of the receptor chains. Therefore, the dime- rization of LIFR and gp130 can be used as a measure of high affinity ligand binding. Deletion of

individual membrane-proximal FNIII domains of gp130 leads to the abrogation of ternary complex formation. Therefore the ligand cannot be bound with high affinity when one of the domains is missing, even though the receptor epitopes involved in ligand binding are unchanged.

LIFR and gp130, a prominent signal of coprecipitated gp130 was seen, when the cells were stimulated with LIF (lane 4). Even after overexpression of both receptor chains, in unstimulated cells gp130 did not coprecipitate with LIFR (lane 3). To distinguish between complexes of LIFR with the endogenous gp130 and those of LIFR and transfected gp130, gp130c was used for cotransfection with the LIFR in COS7 cells. After LIF stimulation of the LIFR/gp130c transfected cells, two gp130 bands could be seen in the Western blot following LIFR precipitation. The upper band was due to endogenous gp130, while the lower one represented the cytoplasmically truncated gp130c.

To control LIFR precipitation, the blots were also probed with LIFR antibody. In all precipitations performed with LIFR transfected cells, two LIFR bands of apparent Replacement of D6 of gp130 by the corresponding domain of the GCSFR leads to the abrogation of signal transduction of all tested cytokines. As coprecipitation of this gp130 mutant and LIFR after stimulation with LIF is still possible, it seems very unlikely that the high affinity binding of the ligand is abrogated in this chimera, indicating that the reason for the missing signal transduction is not an incorrect spacing between the ligand binding epitopes of gp130 and the membrane. Domain 6 therefore is believed to play a role in the formation of specific contacts between gp130 and the LIFR.

Heterodimerization of gp130 with LIFR (Eur. J. Biochem. 269) 2723

(cid:1) FEBS 2002

Fig. 5. Model of gp130/LIFR dimerization. After LIF (white cylinder) binding, gp130 (dark gray ovals) and LIFR (white ovals) dimerize and transduce a signal. In our model, the membrane proximal domains function to bring the cytoplasmic parts of the receptors in close proximity. The model is based on experiments performed with chimeric recep- tors, in which individual domains of gp130 are replaced by the corresponding domains of the GCSFR (black). The receptor binding sites of LIF are labelled with II and III.

protected against solvent contact by a loop of eight amino acids of D5, which is positioned by the disulfide bond between cysteines C458 and C466 [30].

Based on our model for gp130 activation [18], we propose here a model for a ternary gp130/LIFR/LIF complex that is consistent with the data presented in this work (Fig. 5). The formation of a signal transducing heterocomplex of gp130 and LIFR depends on the binding of a cytokine (e.g. LIF). Site II of the ligand contacts the CBM of gp130 [14,15], while site III contacts the Ig-like domain of the LIFR [13]. In the resulting complex, D3 of gp130 and D5 of the LIFR show the greatest distance between individual receptor domains. This distance does not allow domain contacts in activated receptor complexes. gp130 D4 and LIFR D6 point towards each other but have no contact, thus diminishing the distance of the C-terminally located domains. D5 of gp130 serves as a contact site with D7 of the LIFR. The membrane-proximal D6 of gp130 and D8 of the LIFR, respectively, get into close proximity, ensuring the correct spacing and orientation of the cytoplasmic parts of the receptors for signalling. An exchange of gp130 D6 with a domain that does not allow interactions with the corresponding domain 8 of the LIFR abrogates the receptor’s signalling capacity.

To analyse the role of the three cysteines in D5 of gp130 in heterodimerization with the LIFR, each of the cysteines was mutated to alanine. Considering the disulfide bond between C458 and C466, mutation of one of these amino acids leads to a free thiol group in the domain (e.g. C466-SH in the construct gp130C458A). These mutants were intro- duced into the gp130c construct and together with the LIFR transiently transfected into COS7 cells. The surface expres- sion of the mutant receptors were similar to that of wild- type gp130 (data not shown). Cotransfection of the gp130 mutants C466Ac and C491Ac together with the LIFR did not lead to a constitutive signal transduction in COS7 cells (Fig. 6A, upper panel). Also, stimulation of these cells with OSM did not result in STAT1 activation. In both cases, stimulation with LIF led to signal transduction, while a significant STAT1 activation after stimulation with IL-6 was detectable only for the C491A mutant. However, cotransfection of LIFR and the C458Ac mutant in COS7 cells lad to a weak STAT1 activation independent of cytokine stimulation. This activation was dramatically increased after stimulation of the cells with LIF, but not after stimulation with OSM. Signal transduction after stimulation with IL-6/sIL-6R was similar to that of wild- type gp130. The complex architecture of the three membrane prox- imal domains of gp130 and LIFR in the signal transducing complex correlates well with the observation that the exchange of all three membrane-proximal gp130 domains (D4–D6) with the corresponding domains of the GCSFR inhibits the formation of LIFR/gp130 complexes and high binding affinity for LIF [29].

A Cys to Ala point mutation in the domain 5 of gp130 leads to a weak constitutive activation of gp130/LIFR heterodimers and increased sensitivity towards LIF

To investigate whether the C458A mutation leads to the formation of gp130-homodimers or the LIFR is required for stimulation-independent signalling, the gp130C458Ac- construct was transfected into COS7 cells alone or together with LIFR (Fig. 6A, lower panel). Without LIFR expres- sion, no constitutive STAT activation was detectable. Stimulation with LIF or IL-6/sIL-6R did not lead to a STAT activation. Interestingly, after stimulation with OSM the C458A mutant can transduce a signal. Thus, the constitutive signal transduction and increased sensitivity towards LIF of the C458A mutant depends on the coexpression of LIFR on the cell surface.

In the study of Moritz et al. [30], cysteine C458 was proposed to become part of an intermolecular disulfide bond between gp130 chains upon receptor activation. The domain 5 of gp130 proposed to contact domain 7 of LIFR contains three cysteine residues. Based on a structural model of D5, these have been suggested to be involved in the dimerization and activation of gp130 by formation of disulfide bonds [30]. The analysis of cysteine residues in gp130 led to the finding that the two N-terminal cysteines of D5 (C458 and C466) form an intramolecular disulfide bond, while the third (C-terminal) cysteine (C491) in this domain contains a free thiol group. The latter was proposed to be

2724 A. Timmermann et al. (Eur. J. Biochem. 269)

(cid:1) FEBS 2002

Fig. 6. Analysis of signal transduction and coprecipitation of gp130 cysteine mutants in domain 5. (A) STAT1 activation in COS7 cells transiently transfected with LIFR and either gp130C458Ac, gp130C466Ac or gp130C491Ac in response to various cytokines. Forty-eight hours after transfection cells were stimulated as described in Fig. 1B as indicated. Nuclear extracts were prepared and activated STAT1 homodimers were detected by EMSA. A representative of three independent experiments is shown. (B) STAT1 activation in COS7 cells transiently transfected with LIFR and gp130c or gp130C458Ac in response to LIF in absence or presence of 2-mercaptoethanol. Forty-eight hours after transfection cells were incubated for 2 h in medium containing 500 lM 2-mercaptoethanol as indicated. Cells were then stimulated for 15 min with LIF (20 ngÆmL)1) or left unstimulated as indicated. Nuclear extracts were prepared and activated STAT1 homodimers were detected by EMSA as decribed in legend to Fig. 1B. (C) The gp130C458A mutant ligand-independently coprecipitates with the LIFR. Forty-eight hours after transfection coprecipitation of LIFR and gp130 was performed as described in legend to Fig. 3A. While in LIFR/gp130 transfected cells the coprecipitation of the receptors depends on stimulation with LIF, it is independent of stimulation in LIFR/gp130C458A transfected cells.

transduction is not

Exchange of this amino acid to alanine leads to the abrogation of IL-6 signal transduction in COS7 cells transiently transfected with this mutant. In contrast, upon coexpression of gp130C458A together with LIFR IL-6 impaired. Additionally, the signal gp130C458A mutant is able to transduce a signal in response to OSM. Therefore, there does not seem to be an absolute requirement for this cysteine in gp130 signal transduction.

activity and increased sensitivity to LIF of the C458Ac mutant was abrogated. Instead, the mutant receptor behaved like wild-type gp130c (lanes 9–12). These findings point to the formation of a new disulfide bond after mutation of cysteine 458 to alanine in gp130, that gives rise to constitutively active LIFR/gp130C458Ac heterodimers. How is this new disulfide bond positioned in the heterodimeric complex? One possibility is an intramolecular bond between the remaining cysteines C466 and C491 of gp130. This could result in a conformational change within the receptor chain enabling the activating interaction with the LIFR. Another explanation is that of an intermolecular bond between gp130C458Ac and the LIFR. The preformed complex would than enable the stimulation-independent STAT1 activation.

To distinguish between these possibilities, coprecipita- tions of LIFR with the gp130C458Ac mutant were performed (Fig. 6C). LIFR was expressed in COS7 cells either alone, together with gp130c or gp130C458Ac. In contrast to the LIF-dependent coprecipitation of LIFR and gp130c (lanes 1 and 2), coexpression of LIFR with gp130C458Ac led to the ligand-independent coprecipitation of both receptor chains independently of stimulation (lanes 3 and 4, respectively). A stable complex of LIFR and To investigate whether the constitutive activity of the C458A mutant relies on the formation of a disulfide bond not present in the wild-type receptor, the signal transduction of this mutant was measured under reducing conditions (Fig. 6B). These experiments were performed in analogy to activation and dimerization studies of different receptors [31,32]. Cells cotransfected with LIFR and gp130C458Ac or with wild-type gp130c were preincubated in the presence of 2-mercaptoethanol prior to stimulation with LIF. After incubation of the cells with 2-mercaptoethanol, the surface expression of both transfected receptors were similar to that of cells incubated under nonreducing conditions (data not shown). While the reducing conditions did not lead to a change in signal transduction in mock (lanes 1–4) and LIFR/gp130 transfected cells (lanes 5–8), the constitutive

Heterodimerization of gp130 with LIFR (Eur. J. Biochem. 269) 2725

(cid:1) FEBS 2002

Fig. 7. Position of the proposed disulfide bond (-S–S-) between LIFR and gp130C458A leading to constitutive dimerization of the receptors. Without LIF stimulation, only a weak signal occurs. After ligand binding, the receptor dimer adopts the conformation required for efficient signal transduction. Domain 7 of LIFR contains four cysteines that might function as binding partners.

discussed experiments suggest that this binding is inhibited in the preformed LIFR/gp130C458A complex, thereby abrogating OSM signalling.

In a recently published paper by Chow et al. [11] the solution structure of the membrane-distal three domains of gp130 (D1–D3) in complex with viral IL-6 was reported. In this complex, gp130 is believed to adopt the same three dimensional structure as in the complex with IL-6 and IL-6Ra. The presented structure revealed that the two D3 domains of the gp130 fragments point away from each other, as was proposed in our previously published model for gp130 activation by Kurth et al. [18]. The membrane- proximal FNIII domains of gp130 are therefore assumed to be arranged in such a way that the C-terminal parts of the membrane-proximal domain D6 are positioned in close vicinity to each other. In analogy, our current model based on the presented data proposes a similar domain architec- ture in the gp130/LIFR heterodimeric complex. Further- more, we propose that in all cytokine receptors that share structural homology with gp130 and LIFR like OSMR, GCSFR and IL-12R, ligand binding leads to a receptor dimer, in which the C-terminal domains of the CBM are separated from each other. The three membrane-proximal FNIII domains function in bringing the transmembrane and cytoplasmic regions in close proximity in order to enable signal transduction to occur. More structural data are required to substantiate our proposed model of gp130/ LIFR activation.

A C K N O W L E D G E M E N T S

We thank Dr John Wijdenes (DIACLONE, Besanc¸ on, France) for providing the gp130 mAbs B-P4 and B-P8 and Dr Vincent Pitard (CNRS-UMR 5540, Universite´ de Bordeaux 2, Bordeaux, France) for providing the LIFR mAb 10B2 used in this study. This work was supported by grants from the Deutsche Forschungsgemeinschaft (SFB 542) and the Fonds der Chemischen Industrie (Frankfurt, Germany).

R E F E R E N C E S

1. Heldin, C.-H. (1995) Dimerization of cell surface receptors in

signal transduction. Cell 80, 213–223.

2. Ballinger, M.D. & Wells, J.A. (1998) Will any dimer do? Nat.

Struct. Biol. 5, 938–940.

3. Jiang, G. & Hunter, T. (1999) Receptor signaling: when dimer-

ization is not enough. Curr. Biol. 9, R568–R571.

4. Wells, J.A. & de Vos, A.M. (1996) Hematopoietic receptor com-

plexes. Annu. Rev. Biochem. 65, 609–634.

5. Ihle, J.N. (1995) Cytokine receptor signaling. Nature 377, 591–594. 6. de Vos, A.M., Ultsch, M. & Kossiakoff, A.A. (1992) Human growth hormone and extracellular domain of its receptor: crystal structure of the complex. Science 255, 306–312.

7. 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., Finer- Moore, J., Elliott, S., Sitney, K., Katz, B.A., Matthews, D.J., Wendoloski, J.J., Egrie, J. & Stroud, R.M. (1998) Efficiency of signalling through cytokine receptors depends critically on receptor orientation. Nature 395, 511–516.

8. Hibi, M., Murakami, M., Saito, M., Hirano, T., Taga, T. & Kishimoto, T. (1990) Molecular cloning and expression of an IL-6 signal transducer, gp130. Cell 63, 1149–1157.

9. Kurth, I., Horsten, U., Pflanz, S., Dahmen, H., Ku¨ ster, A., Gro¨ tzinger, J., Heinrich, P.C. & Mu¨ ller-Newen, G. (1999) Activation of the signal transducer gp130 by both interleukin-6

gp130C458Ac is therefore formed independently of cytoki- ne stimulation. This constitutive complex formation points to a covalent bonding between the two receptor chains on the cell surface, as indicated in the activation model based on these results (Fig. 7). The complex does not have the conformation necessary for the effective induction of signalling pathways, although a weak constitutive receptor activation can be observed, probably due to the increased proximity of the cytoplasmic parts of the receptors. This finding highlights the importance of proper receptor orien- tation in addition to dimerization for receptor activation. Upon ligand binding, the complex adopts a conformation able to effectively induce signalling. The complex might function as a trap for the cytokine, as the dissociation of the ligand from the receptors is diminished. This would lead to a prolonged signal transduction via the receptor complex, resulting in the observed prominent STAT activation.

The elevated signal transduction of the gp130C458A- mutant together with the LIFR depends on ligand binding to the preformed receptor complex, which in case of LIF is a two step process of defined order. This binding order can explain the observed cytokine specificity of the gp130C458A mutant, as it can be assumed that also in the preformed receptor complex LIF first interacts with the LIFR via a site III, Ig-like domain contact and is than transferred into the position where it also contacts gp130 via a site II, CBM interaction. IL-6, which can also induce signal transduction via the gp130C458A mutant, is recruited by the signal transducer only after binding to its specific a-receptor IL-6Ra. The sIL-6Ra, which is not involved in signal transduction in the cytoplasm, was preincubated with IL-6 prior to stimulation. Therefore, the IL-6/sIL-6R complexes used in the discussed experiments could bind to the gp130C458A/LIFR complex and induce a conformational change of this receptor chain sufficient for the induction of a cytoplasmic signal. In contrast, the first step of OSM signal transduction is binding of the cytokine to gp130. The

2726 A. Timmermann et al. (Eur. J. Biochem. 269)

(cid:1) FEBS 2002

two distinct binding epitopes.

interleukin-6 receptor: expression in insect cells, purification and characterization. Eur. J. Biochem. 234, 661–669.

and interleukin-11 requires J. Immunol. 162, 1480–1487.

10. Pflanz, S., Kurth, I., Gro¨ tzinger, J., Heinrich, P.C. & Mu¨ ller- Newen, G. (2000) Two different epitopes of the signal transducer gp130 sequentially cooperate on IL-6-induced receptor activation. J. Immunol. 165, 7042–7049.

22. Wijdenes, J., Heinrich, P.C., Mu¨ ller-Newen, G., Roche, C., Zong- Jiang, G., Clement, C. & Klein, B. (1995) Interleukin-6 signal transducer gp130 has specific binding sites for different cytokines as determined by antagonistic and agonistic anti-gp130 mono- clonal antibodies. Eur. J. Immunol. 25, 3474–3481.

11. Chow, D., He, X., Snow, A.L., Rose-John, S. & Garcia, K.C. (2001) Structure of an extracellular gp130 cytokine receptor complex. Science 291, 2150–2155.

12. Gearing, D.P., Comeau, M.R., Friend, D.J., Gimpel, S.D., Thut, C.J., McGourty, J., Brasher, K.K., King, J.A., Gillis, S., Mosley, B., Ziegler, S.F. & Cosman, D. (1992) The IL-6 signal transducer, gp130: an oncostatin M receptor and affinity converter for the LIF receptor. Science 255, 1434–1437.

23. Pitard, V., Taupin, J.-L., Miossec, V., Blanchard, F., Cransac, M., Jollet, I., Vernallis, A., Hudson, K., Godard, A., Jacques, Y. & Moreau, J.-F. (1997) Production and characterization of monoclonal antibodies against the leukemia inhibitory factor low affinity receptor, gp190. J. Immunol. Methods 205, 177–190. 24. Thiel, S., Behrmann, I., Timmermann, A., Dahmen, H., Mu¨ ller- Newen, G., Schaper, F., Tavernier, J., Pitard, V., Heinrich, P.C. & Graeve, L. (1999) Identification of a Leu-Ile interalization motif within the cytoplasmic domain of the leukaemia inhibitory factor. Biochem. J. 339, 15–19.

13. Owczarek, C.M., Zhang, Y., Layton, M.J., Metcalf, D., Roberts, B. & Nicola, N.A. (1997) The unusual species cross-reactivity of the leukemia inhibitory factor receptor alpha-chain is determined primarily by the immunoglobulin-like domain. J. Biol. Chem. 272, 23976–23985.

25. Wegenka, U.M., Lu¨ tticken, C., Buschmann, J., Yuan, J., Lottspeich, F., Mu¨ ller-Esterl, W., Schindler, C., Roeb, E., Hein- rich, P.C. & Horn, F. (1994) The interleukin-6-activated acute- phase response factor is antigenically and functionally related to members of the signal transduction and transcription (Stat) factor family. Mol. Cell. Biol. 14, 3186–3196.

14. Hammacher, A., Richardson, R.T., Layton, J.E., Smith, D.K., Angus, L.J.L., Hilton, D.J., Nicola, N.A., Wijdenes, J. & Simp- son, R.J. (1998) The immunoglobulin-like module of gp130 is required for signaling by interleukin-6 but not by leukemia inhibitory factor. J. Biol. Chem. 273, 22701–22707.

26. Wagner, B.J., Hayes, T.E., Hoban, C.J. & Cochran, B.H. (1990) The SIF binding element confers sis/PDGF inducibility onto the c-fos promoter. EMBO J. 9, 4477–4484.

15. Timmermann, A., Pflanz, S., Gro¨ tzinger, J., Ku¨ ster, A., Kurth, I., Pitard, V., Heinrich, P.C. & Mu¨ ller-Newen, G. (2000) Different epitopes are required for gp130 activation by interleukin-6, oncostatin M and leukemia inhibitory factor. FEBS Lett. 468, 120–124.

27. Kyhse-Anderson, J. (1984) Electroblotting of multiple gels: a simple apparatus without buffer tank for rapid transfer of proteins form polyacrylamide to nitrocellulose. J. Biochem. Biophys. Methods 10, 203–209.

16. Mosley, B., De Imus, C., Friend, D., Boiani, N., Thoma, B., Park, L.S. & Cosman, D. (1996) Dual oncostatin M (OSM) receptors. Cloning and characterization of an alternative signaling subunit conferring OSM-specific receptor activation. J. Biol. Chem. 271, 32635–32643.

28. Pflanz, S., Kernebeck, T., Giese, B., Herrmann, A., Pachta-Nick, M., Stahl, J., Wollmer, A., Heinrich, P.C., Mu¨ ller-Newen, G. & Gro¨ tzinger, A. (2001) The signal transducer gp130: Biochemical characterization of the three membrane-proximal extracellular domains and studies on their oligomerization potential. Biochem. J. 356, 605–612.

17. Gearing, D.P., Thut, C.J., VandenBos, T., Gimpel, S.D., Delaney, P.B., King, J., Price, V., Cosman, D. & Beckmann, M.P. (1991) Leukemia inhibitory factor receptor is structurally related to the IL-6 signal transducer, gp130. EMBO J. 10, 2839–2848.

29. Hammacher, A., Wijdenes, J., Hilton, D.J., Nicola, N.A., Simpson, R.J. & Layton, J.E. (2000) Ligand-specific utilization of the extracellular membrane-proximal region of the gp130-related signalling receptors. Biochem. J. 345, 25–32.

Importance of

18. Kurth, I., Horsten, U., Pflanz, S., Timmermann, A., Ku¨ ster, A., Dahmen, H., Tacken, I., Heinrich, P.C. & Mu¨ ller-Newen, G. (2000) the membrane-proximal extracellular domains for activation of the signal transducer gp130. J. Immunol. 164, 273–282.

30. Moritz, R.L., Hall, N.E., Connolly, L.M. & Simpson, R.J. (2001) Determination of the disulfide structure and N-glycosylation sites of the extracellular domain of the human signal transducer gp130. J. Biol. Chem. 276, 8244–8253.

19. Arcone, R., Pucci, P., Zappacosta, F., Fontaine, V., Malorni, A., Marino, G. & Ciliberto, G. (1991) Single-step purification and structural characterization of human interleukin-6 produced in Escherichia coli from a T7 RNA polymerase expression vector. Eur. J. Biochem. 198, 541–547.

31. Robertson, S.C., Meyer, A.N., Hart, K.C., Galvin, B.D., Webster, M.K. & Donoghue, D.J. (1998) Activating mutations in the extracellular domain of the fibroblast growth factor receptor 2 function by disruption of the disulfide bond in the third immunoglobulin-like domain. Proc. Natl Acad. Sci. USA 95, 4567–4572.

20. Aarden, L.A., De Groot, E.R., Schaap, O.L. & Lansdorp, P.M. (1987) Production of hybridoma growth factor by human monocytes. Eur. J. Immunol. 17, 1411–1416.

21. Weiergra¨ ber, O., Hemmann, U., Ku¨ ster, A., Mu¨ ller-Newen, G., Schneider, J., Rose-John, S., Kurschat, P., Brakenhoff, J.P.J., Hart, M.H.L., Stabel, S. & Heinrich, P.C. (1995) Soluble human

32. Siegel, P.M. & Muller, W.J. (1996) Mutations affecting conserved cysteine residues within the extracellular domain of Neu promote receptor dimerization and activation. Proc. Natl Acad. Sci. USA 93, 8878–8883.