Signal Transducer Glycoprotein 130 Extracellular Domains for Activation

Home , Cytokine receptor, Glycoprotein 130, Oncostatin M receptor, Thrombopoietin receptor

Importance of the Membrane-Proximal Extracellular Domains for Activation of the Signal Transducer Glycoprotein 130

This information is current as Ingo Kurth, Ursula Horsten, Stefan Pflanz, Andreas of September 28, 2021. Timmermann, Andrea Küster, Heike Dahmen, Ingrid Tacken, Peter C. Heinrich and Gerhard Müller-Newen J Immunol 2000; 164:273-282; ; doi: 10.4049/jimmunol.164.1.273

http://www.jimmunol.org/content/164/1/273 Downloaded from

References This article cites 40 articles, 21 of which you can access for free at: http://www.jimmunol.org/content/164/1/273.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 28, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2000 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Importance of the Membrane-Proximal Extracellular Domains for Activation of the Signal Transducer Glycoprotein 1301

Ingo Kurth, Ursula Horsten,2 Stefan Pﬂanz, Andreas Timmermann, Andrea Ku¨ster, Heike Dahmen, Ingrid Tacken, Peter C. Heinrich,3 and Gerhard Mu¨ller-Newen

The transmembrane glycoprotein gp130 is the common signal transducing receptor subunit of the IL-6-type cytokines. The gp130 extracellular part is predicted to consist of six individual domains. Whereas the role of the three membrane-distal domains (D1–D3) in binding of IL-6 and IL-11 is well established, the function of the membrane-proximal domains (D4–D6) is unclear. Mapping of a neutralizing mAb to the membrane-proximal part of gp130 suggests a functional role of D4–D6 in receptor activation. Individual deletion of these three domains differentially interferes with ligand binding of the soluble and membrane- bound receptors. All deletion mutants do not signal in response to IL-6 and IL-11. The deletion mutants ⌬4 and, to a lesser extent,

⌬6 are still activated by agonistic monoclonal gp130 Abs, whereas the deletion mutant ⌬5 does not respond. Because membrane- Downloaded from bound ⌬5 binds IL-6/soluble IL-6R as does wild-type gp130, but does not transduce a signal in response to various stimuli, this domain plays a prominent role in coupling of ligand binding and signal transduction. Replacement of the ﬁfth domain of gp130 by the corresponding domain of the homologous G-CSF receptor leads to constitutive activation of the chimera upon overexpression in COS-7 cells. In HepG2 cells this mutant responds to IL-6 comparable to wild-type gp130. Our ﬁndings suggest a functional role of the membrane-proximal domains of gp130 in receptor activation. Thus, within the hematopoietic receptor family the mechanism of receptor activation critically depends on the architecture of the receptor ectodomain. The Journal of Immu- http://www.jimmunol.org/ nology, 2000, 164: 273–282.

he coordination and regulation of immune responses is Subsequently, the activated Jaks phosphorylate tyrosine residues mainly mediated by cytokines that specifically bind to cell of the receptor. These phosphotyrosine residues are docking sites T surface receptors. Cytokine receptors are usually classi- for latent transcription factors of the STAT family, which also fied due to structural similarities. Class I cytokine receptors are become phosphorylated at the receptor. The phosphorylated characterized by the presence of at least one cytokine binding STATs dimerize and translocate into the nucleus to regulate target module (CBM)4 that consists of two fibronectin-type III-like gene expression (2). by guest on September 28, 2021 (FNIII) domains. The N-terminal domain contains a set of four Whereas the ectodomains of the receptors for growth hormone conserved cysteine residues, and the C-terminal domain contains a (GH), erythropoietin (Epo), or prolactin solely consist of a single WSXWS motif or a closely related sequence. Receptors belonging CBM, other members of the class I cytokine receptor family, such to this family are engaged by helical cytokines consisting of four as gp130, G-CSFR, or the leptin receptor show a more complex tightly packed ␣-helices (1). The cytoplasmic parts of cytokine architecture, which in respect to their biological functions is not receptors do not contain a kinase domain, but constitutively asso- understood (3). The glycoprotein gp130 is the common signal ciate with Janus tyrosine kinases (Jaks). Binding of the ligand transducing receptor subunit of the IL-6-type cytokines IL-6, IL- leads to receptor dimerization and juxtaposition of the associated 11, leukemia inhibitory factor (LIF), ciliary neurotrophic factor Jaks, which results in their activation by tyrosine phosphorylation. (CNTF), oncostatin M (OSM), and cardiotrophin-1 (4). The extracellular part of gp130 contains an Ig-like domain (D1) followed by a single CBM (D2 and D3) and three FNIII domains (D4, D5, and Institut fu¨r Biochemie, Rheinisch-Westfa¨lische Technische Hochschule Aachen, Aachen, Germany D6) (5). Signal transduction is achieved by either homodimerization of gp130 in response to IL-6 (6) and IL-11 (7) or heterodimer- Received for publication April 15, 1999. Accepted for publication October 8, 1999. ization of gp130 with the LIFR in response to LIF (8), CNTF (9), The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance OSM (8), or CT1 (10). Alternatively, OSM induces the het- with 18 U.S.C. Section 1734 solely to indicate this fact. erodimerization of gp130 with the recently cloned OSMR (11). 1 This work was supported by grants from the Deutsche Forschungsgemeinschaft The cytokines IL-6, IL-11, and CNTF alone do not efficiently en- (Bonn, Germany) and the Fonds der Chemischen Industrie (Frankfurt am Main, gage the signal transducing receptor chains. They first bind to their Germany). specific ␣ receptors that can functionally be replaced by the re- 2 Current address: Institut fu¨r Experimentelle und Klinische Pharmakologie und Tox- ikologie der Universita¨t Erlangen-Nu¨rnberg, Universitätsstrasse 22, D-91054 Erlan- spective soluble counterparts lacking the transmembrane and cy- gen, Germany. toplasmic regions (12). In addition to at least one CBM, all recep- 3 Address correspondence and reprint requests to Dr. Peter C. Heinrich, Institut fu¨r tors involved in IL-6-type cytokine signaling contain an Ig-like Biochemie, Rheinisch-Westfa¨lische Technische Hochschule Aachen, Pauwelsstrasse domain located N-terminally of the most membrane-proximal 30, D-52057 Aachen, Germany. E-mail address: [email protected] CBM. Besides the Ig-like domain and CBM(s), those receptors 4 Abbreviations used in this paper: CBM, cytokine binding module; FNIII, fibronectin type III-like; Jak, Janus kinase; GH, growth hormone; GHR, GH receptor; Epo, eryth- triggering the cytoplasmic signal transduction cascade (gp130, ropoietin; EpoR, Epo receptor; G-CSFR, G-CSF receptor; LIF, leukemia inhibitory LIFR, and OSMR) all contain three additional membrane-proximal factor; LIFR, LIF receptor; CNTF, ciliary neurotrophic factor; OSM, oncostatin M; OSMR, OSM receptor; SIE, sis-inducible element; sIL-6, soluble IL-6; sgp130, sol- FNIII domains whose function is still unknown. Mutagenesis stud- uble gp130. ies using deletion mutants of gp130 revealed that the CBM of

gp130 as well as the Ig-like domain are required for the interaction plasmids were sequenced using an ABI Prism Automated sequencer (Per- with IL-6/IL-6R (13–15), IL-11/IL-11R (16), and OSM (17). A kin-Elmer, Norwalk, CT). For expression in HepG2 cells, the cDNAs were more detailed analysis of the binding epitope revealed that the subcloned into the expression vector pRcCMV. To express the deletion constructs and chimera in the context of CBM of gp130 binds the ligand in a way similar to the GH/GHR gp130␥, respective cDNAs were subcloned into the vector pSVLEg-Y␥440 interaction (15, 18). These findings were supported by the crystal (27) using the restriction endonucleases XhoI and BstEII. The resulting structure of the gp130 CBM (19) as well as the solution structure vector encodes the extracellular domain, the transmembrane, and a trun- of gp130-D3 (20). cated intracellular part of gp130 (containing boxes 1 and 2) fused to the IFN-␥ receptor motif YDKPH, which predominantly activates STAT1. For Ligand-induced dimerization of cytokine receptors is widely ac- construction of plasmids encoding the soluble gp130 proteins the deletion cepted as a prerequisite for receptor activation (21, 22). In this constructs were subcloned into a vector (pSVLsgp130-Flag) (18) encoding respect, the functional role of the membrane-proximal domains of the extracellular domain of gp130 fused to a Flag epitope at the C-terminus. gp130 (D4–D6) is not understood. Because these three domains represent a common structural feature of the signal transducing Transfection of cells IL-6-type cytokine receptors, D4–D6 of gp130 may be essential Plasmid DNA was transfected into Ba/F3 cells by electroporation. Twenty- for receptor activation. The studies presented here using deletion eight micrograms of the gp130 expression vector were coelectroporated ␮ ϫ 6 mutants of gp130 reveal that the membrane-proximal domains with 2 g of pSV2neo into 3.5 10 cells in 0.8 ml of medium applying a single 70-ms pulse at 200 V. Selection with G418 (3 mg/ml) was initiated play an essential role in the coupling of ligand binding and signal 24 h after transfection. Selected Ba/F3 clones were screened for the pres- transduction. Our findings suggest a mechanism of receptor acti- ence of membrane-bound gp130 proteins by flow cytometry. COS-7 cells vation that differs from the mechanism of activation established for were transiently transfected using the DEAE-dextran method. The effi-

short cytokine receptors such as GHR, EpoR, or prolactin receptor. ciency of transfection was analyzed by FACS. HepG2 cells were trans- Downloaded from fected by the calcium phosphate coprecipitation method as described previously (28). Materials and Methods Enzymes, proteins, Abs, chemicals, and cell culture Soluble radioactive binding assay Enzymes were purchased from Roche Molecular Biochemicals (Mann- The binding assay was performed using IL-6 radiolabeled with 125I ac- heim, Germany), and protein A-Sepharose was obtained from Pharmacia cording to the procedure of Bolton and Hunter (29). Sixty nanograms of

(Freiburg, Germany). DMEM and antibiotics were obtained from Life sgp130 in 500 ␮l of TNET buffer (20 mM Tris-HCl (pH 7.5), 140 mM http://www.jimmunol.org/ Technologies (Eggenstein, Germany), and FCS from Seromed (Munich, NaCl, 5 mM Na2EDTA, 1% Triton X-100, 2 mM methionine, and 0.01% ␮ Germany). Radiochemicals were purchased from Amersham International NaN3) was incubated with 0.5 g of the respective Ab (B-P4 or B-R3). (Aylesbury, U.K.). Recombinant human IL-6 was expressed in Escherichia Five nanograms of 125I-labeled IL-6 and 100 ng of sIL-6R were added. coli, refolded, and purified as described by Arcone et al. (23). Soluble After incubation for3hatroom temperature, the complexes formed were IL-6R (sIL-6R) (24), soluble gp130 (25), as well as soluble IL-11R (16) immunoprecipitated by the addition of protein A-Sepharose. Coprecipi- were expressed in insect cells as described previously. The monoclonal tated 125I-labeled IL-6 was quantified using a gamma counter. gp130 Abs B-R3, B-S12-G7, B-P4, and B-P8 were generated as described previously (26). All other Abs were purchased from Dako (Hamburg, Ger- Immunofluorescence staining many). The PBS buffer contained 200 mM NaCl, 2.5 mM KCl, 8 mM COS-7 cells were transiently transfected with the respective expression Na2HPO4, and 1.5 mM KH2PO4. Simian monkey kidney cells (COS-7) were cultured in DMEM, and Ba/F3 cells and Ba/F3 transfectants were vectors (13) encoding gp130 deletion mutants. Forty-eight hours after by guest on September 28, 2021 cultured in DMEM containing 5% (v/v) conditioned medium from transfection, the cells were fixed with paraformaldehyde, incubated with X63Ag8-653 BPV-mIL-3 myeloma cells (as a source of IL-3) in a water- the gp130 mAbs B-P4 and B-R3, respectively, and stained with a rhodamine-conjugated secondary Ab. Subsequently, the cells were mounted with saturated atmosphere containing 5% CO2. These media were supplemented with 10% (v/v) FCS, streptomycin (100 ␮g/ml), and penicillin (60 ␮g/ml). Mowiol and analyzed by fluorescence microscopy. Ba/F3 transfectants were cultured in the presence of 1 mg/ml G418. HepG2 cells were cultivated in DMEM/F12. Flow cytometry Cells were collected, washed, and resuspended in cold PBS containing 5% Plasmid construction FCS and 0.1% sodium azide. Subsequently, cells were incubated on ice ␮ The starting point for cloning of ⌬4, ⌬5, and ⌬6 was the full-length human with 4 g/ml gp130 Abs B-S12-G7 or B-P8. Cells were washed with cold gp130 cDNA cloned into the XhoI and BamHI sites of the eukaryotic PBS/azide and incubated with PE-conjugated anti-mouse IgG Fab at a 1/50 expression vector pSVL lacking the EcoRI site (gp130-pSVL⌬Eco). Using dilution. Again, cells were washed with cold PBS/azide and then resus- ␮ this vector as template, for each deletion mutant two fragments were am- pended in 400 l of PBS/azide followed by flow cytometric analysis using plified. The first fragment encodes the N-terminal region of gp130 down to a FACScalibur (Becton Dickinson, Mountain View, CA). the deletion. The sense primer (5Ј-GTGTT ACTTC TGCTC T-3Ј) anneals upstream of the gp130 cDNA within the vector. The respective antisense Electrophoretic mobility shift assay primer introduces an XmaI site (underlined) that determines the deletion Cells were incubated at 37°C for 15 min or for the periods of time indicated Ј Ј ⌬ Ј 5 -CTTGG TGCTT TCCCG GGTCT ATCTT CATAG G-3 ( 4), 5 -TT in the figures in the presence of IL-6/sIL-6R or IL-11/sIL-11R or were left Ј ⌬ Ј TAA GATCC ATCCC GGGGT GAGTA G-3 ( 5), or 5 -GTAGG TC unstimulated. COS-7 cells were stimulated with 12.5 ng/ml IL-6 and 500 Ј ⌬ CTT TCCCG GGTGG AGCTT GTTTA AGG-3 ( 6). The second frag- ng/ml sIL-6R. Ba/F3 cells were stimulated with 33 ng/ml IL-6 and 500 ment encodes the C-terminal part of gp130 following the deleted domain. ng/ml sIL-6R or 150 ng/ml IL-11 and 700 ng/ml sIL-11R or ␮g/ml BS- Again, the respective sense primer introduces an XmaI site (underlined) 12-G7 and ␮g/ml B-P8. Preparation of nuclear extracts and EMSAs were Ј Ј ⌬ Ј 5 -GCTAC TCACC CCGGG ATGGA TCTTA AAG-3 ( 4), 5 -CCACC performed as previously described (30). A double-stranded sis-inducible Ј ⌬ Ј TTCCC CCGGG CCTAC TGTTC GG-3 ( 5), or 5 -CCAAA GTTTG element (SIE) oligonucleotide derived from the c-Fos promoter (m67SIE; Ј ⌬ Ј CTCCC GGGGA AATTG AAG-3 ( 6). The antisense primer (5 - 5Ј-GATCC GGGAG GGATT TACGG GGAAA TGCTG-3Ј) was used as Ј TCTAG TTGTG GTTTG T-3 ) anneals within the vector. The first frag- the 32P-labeled probe (31). The protein-DNA complexes were separated on ment was digested with XhoI and XmaI; the second fragment was digested a 4.5% polyacrylamide gel containing 7.5% glycerol. Electrophoresis was with XmaI and BstEII. Together, both fragments were cloned in the XhoI- performed using 0.25ϫ TBE buffer at 72 V/cm. and BstEII-digested gp130-pSVL⌬Eco, yielding the construct encoding the deletion mutant. To construct the D5-GCSFR chimera the corresponding Proliferation assay domain of the G-CSFR was amplified by PCR and cloned into the newly introduced XmaI site of the vector encoding the deletion mutant ⌬5. The Ba/F3 cells (20,000 cells/well) were plated on 96-well plates and stimu- human G-CSFR cDNA served as a template. The following oligonucleo- lated with varying amounts of IL-6 in the presence of 500 ng/ml of sIL-6R tides were used to amplify the DNA encoding domain 5 of G-CSFR (XmaI or with Abs as indicated in the diagrams. After 72 h of incubation, viable site underlined): 5Ј-GGCCC AGCTC TGACC CCCGG GCATG CCATG and metabolically active cells were quantified using a colorimetric assay GCCCG-3Ј and 5Ј-TAGAT GCAGC TCTGG CCCGG GGGAG GGAGC based on the Cell Proliferation Kit II (XTT; Roche Molecular CATTT C-3Ј. PCRs were performed applying standard procedures. All Biochemicals). The Journal of Immunology 275

Soluble ternary complex formation assay Soluble gp130-Flag proteins from supernatants of transfected COS-7 cells were incubated with 1 ␮g/ml of ﬂag-Ab and subsequently 1 mg/ml protein A-Sepharose was added. After an overnight incubation at 4°C, immuno- precipitates were washed twice with TNET (20 mM Tris-HCl (pH 7.5), 140 mM NaCl, 5 mM EDTA, and 1%Triton X-100) and incubated with 4, 10, or 30 nM IL-6 and 70 nM sIL-6R. After 16 h the complexes were washed twice with TNET, resuspended in Laemmli buffer, incubated at 95°C for 5 min and separated on a 12.5% SDS-polyacrylamide gel under reducing conditions followed by electroblotting. Binding of 125I-labeled IL-6 to stably transfected Ba/F3 cells Ba/F3 cells (5 ϫ 106) stably transfected with gp130, ⌬4, ⌬5, or ⌬6 were incubated in the presence of sIL-6R (200 nM) with varying concentrations of 125I-labeled IL-6 (1340 cpm/fmol) ranging from 0.15–10 nM. After incubation for 16 h at 4°C the cells were centrifuged through a mixture of dinonyl- and dibutylphthalate oil (1.020 g/ml). Cell-associated and free radioactivity were measured using a gamma counter. Speciﬁc binding was obtained by subtracting the radioactivity associated with untransfected Ba/F3 cells.

Immunoprecipitation of Jak1 Downloaded from Transfected COS-7 cells were stimulated for 15 min with 12.5 ng/ml IL-6 and 500 ng/ml sIL-6R as indicated in the figure. Cells were treated with lysis buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40, 1 mM sodium orthovanadate, and 1 mM NaF in 15% glycerol) in the presence of the protease inhibitors pepstatin, leupeptin, apro- tinin, and PMSF for 30 min at 4°C. After centrifugation the supernatants were incubated with 4 ␮g/ml Jak1 Ab and 5 mg/ml protein A-Sepharose. http://www.jimmunol.org/ Reporter gene assay using transfected HepG2 cells HepG2 cells were transiently transfected with expression vectors encoding ␣ wild-type gp130 or D5-GCSFR, an 2-macrogloblulin promotor luciferase gene reporter construct (28), and a ␤-galactosidase control vector. Twenty- four hours after transfection cells were stimulated with IL-6 (10 ng/ml) or left unstimulated. After an additional 16 h, luciferase activity was measured using the luciferase kit from Promega (Madison, WI) and normalized to ␤-galactosidase activity to correct for transfection efficiency. by guest on September 28, 2021 Immunoblotting and enhanced chemiluminescence detection Immunoprecipitated proteins separated by SDS-PAGE were transferred to a polyvinylidene difluoride membrane by a semidry electroblotting procedure (32). Polyvinylidene difluoride membranes were blocked in a solution of 20 mM Tris-HCl (pH 7.6), 137 mM NaCl, and 0.1% Nonidet-P40 containing 10% BSA and probed with Ab, followed by incubation with HRP- conjugated secondary Ab. Immunoreactive proteins were detected by FIGURE 1. The antagonistic mAb B-P4 does not interfere with ligand chemiluminescence using the enhanced chemiluminescence kit (Amer- binding and maps to the membrane-proximal half of gp130. A, Soluble sham) following the manufacturer’s instructions. gp130 was first incubated with the mAbs B-P4 or B-R3 and subsequently with 125I-labeled IL-6 and sIL-6R as indicated. Complexes formed (see Results scheme) were immunoprecipitated, and coprecipitated radioactivity was The IL-6 and IL-11 neutralizing gp130 mAb B-P4 maps to the measured. B, COS-7 cells were transfected with expression vectors encod- membrane-proximal part of gp130 ing gp130 deletion constructs as indicated schematically on the left. Forty- eight hours after transfection, cells were fixed and incubated with mAb In previous studies several mAbs directed against the ectodomain B-P4 or B-R3. Bound Abs were visualized by fluorescence microscopy of gp130 were characterized. Some of these Abs act agonistically, using a rhodamine-conjugated secondary Ab, and stained cells were pho- such as the combination of B-S12 and B-P8, while others, such as tographed at ϫ100 magnification. B-R3, were neutralizing (26). The mAb B-P4 has been described to specifically inhibit IL-11 (26) and, depending on the cells in- vestigated, IL-6 responses (33). The antagonistic mAbs B-P4 and transmembrane and cytoplasmic regions of human gp130. Staining B-R3 were used as precipitating Abs in a ternary complex forma- of the cells by immunofluorescence revealed that B-P4 maps to the tion assay. Soluble gp130 (sgp130) lacking the transmembrane and membrane-proximal domains (D4–D6), whereas B-R3 recognizes cytoplasmic parts was incubated with B-P4 or B-R3. Subse- the membrane-distal part of gp130 (Fig. 1B). Because B-P4 inter- quently, iodinated IL-6 was added in the absence or presence of feres with activation of gp130, but not with ligand binding, and soluble IL-6R (sIL-6R). Precipitation of sgp130 bound to B-P4 via maps to the membrane-proximal domains, we conclude that these protein A-Sepharose in the presence of sIL-6R leads to coprecipi- domains are involved in activation of the signal transducer. tation of IL-6, whereas B-R3 fails to coprecipitate significant amounts of IL-6 (Fig. 1A) indicating that, in contrast to B-R3, The membrane-proximal domains are not required for soluble binding of B-P4 to gp130 does not interfere with ligand binding. ternary complex formation, but contribute to binding of To confine the epitopes of B-P4 and B-R3, COS-7 cells were trans- IL-6/sIL-6R complexes to membrane-bound gp130 fected with deletion constructs encoding either the membrane-dis- To further assess their functional role in ligand binding and recep- tal or the membrane-proximal half of the ectodomain fused to the tor activation, the domains D4, D5, and D6 of human gp130 were 276 MEMBRANE-PROXIMAL DOMAINS OF gp130

surface to a similar extent compared with wild-type gp130 (Fig. 3B). The binding capabilities of the cell surface expressed receptor mutants were studied by an equilibrium binding assay using radiolabeled IL-6 (125I-labeled IL-6). In the presence of a large excess of sIL-6R, the transfected cells were incubated with 125I-labeled IL-6 for 16 h at 4°C. Subsequently, cell-bound and free radioactivity were separated by centrifugation of the cells through a dinonylphthalate/dibutylphthalate cushion (18). Unspeciﬁc binding was determined by incubation of untransfected Ba/F3 cells with 125I-labeled IL-6/sIL-6R. As in the soluble ternary complex formation assay, binding of 125I-labeled IL-6 to ⌬5 was indistin- guishable from binding to wild-type gp130 (Fig. 3C). Analysis of the binding data by Scatchard transformation revealed that the af- ﬁnities of both wild-type gp130 and ⌬5 ranged from 2.5–4 nM. This value is in agreement with those determined previously for Ba/F3 cells expressing gp130 (18). The mutants ⌬4 and ⌬6, however, showed reduced binding of 125I-labeled IL-6. For practical reasons saturating concentrations of 125I-labeled IL-6 could not be

⌬ ⌬ Downloaded from FIGURE 2. Schematic representation of gp130 and the deletion mutants applied. Therefore, analysis of the data for 4 and 6 by Scatchard ⌬4, ⌬5, and ⌬6. The domains of the gp130 extracellular part (D1–D6) are transformation was not possible. Because reduced binding capa- depicted followed by the transmembrane (black bar) and cytoplasmic part. bility of ⌬4 and ⌬6 is not due to reduced surface expression (see Sites of deletions of individual domains in the mutants ⌬4, ⌬5, and ⌬6 are Fig. 3B), we conclude that these domains are required for optimal marked by arrowheads. ligand binding by membrane-bound gp130. Each membrane-proximal domain is required for activation of http://www.jimmunol.org/ deleted individually, applying PCR-based methodology. Domain gp130 by IL-6 as well as IL-11 borders were chosen according to the suggestions of Hibi et al. (5) The ability of the receptor mutants to activate STAT3 was ana- (schematic representation of deletion mutants in Fig. 2). To per- lyzed by EMSA after stimulation of the stably transfected Ba/F3 form ternary complex formation assays using soluble receptor pro- cells with IL-6/sIL-6R or IL-11/sIL-11R (Fig. 4A). Cells trans- teins, the transmembrane and cytoplasmic parts of wild-type gp130 fected with wild-type gp130 responded with a strong activation of and of the deletion mutants were replaced by a flag epitope STAT3, whereas all deletion mutants were unable to induce STAT (sgp130-flag). From supernatants of transfected COS-7 cells the activation. Ba/F3 cells, which normally proliferate IL-3-depen- soluble recombinant receptor proteins were collected by immuno- dently, are known to grow in response to various cytokines after precipitation using a flag-Ab. After addition of sIL-6R in combi- transfection of the corresponding receptor chains. Therefore, cells by guest on September 28, 2021 nation with varying amounts of IL-6, ternary complexes formed transfected with wild-type gp130 proliferate after IL-6/sIL-6R were precipitated via the Sepharose-bound sgp130-flag. The com- stimulation (Fig. 4B). Again, the response to IL-6/sIL-6R is totally plexes were separated by SDS-PAGE followed by immunoblotting abolished in cells stably transfected with the deletion mutants (Fig. using IL-6 and flag Abs. The different electrophoretic mobilities of 4B), although these cells still proliferate in response to IL-3 (not sgp130-flag and the deletion mutants (Fig. 3A, upper panel) shown). roughly mirror the loss of molecular mass due to deletion of the In previous studies we established a system that allows the anal- individual domains and covalently linked N-glycans. Deletion of ysis of STAT activation induced by gp130 mutants in transiently domains 4 (101 aa) and 6 (96 aa) is accompanied by a loss of three transfected COS-7 cells even in the presence of endogenous gp130 and two potential N-glycosylation sites, respectively, whereas do- (15, 18). Because endogenous gp130 preferentially activates main 5 (95 aa) lacks potential N-glycosylation sites. The s⌬6-flag STAT3 (see Fig. 4A), the STAT recruiting modules of gp130 mu- appears as a broader band upon SDS-PAGE. This microheteroge- tants were replaced by the strongly and specifically STAT1 acti- neity is probably due to uneven glycosylation. In the presence of vating module YDKPH derived from the IFN-␥ receptor (gp130␥). IL-6/sIL-6R complexes, coprecipitation of IL-6 is achieved with The use of gp130␥ constructs enabled us to discriminate between all the deletion constructs as detected by immunoblotting (Fig. 3A, STAT activation of the endogenous and the transfected gp130 in lower panel). The IL-6 coprecipitation is indicative of ternary COS-7 cells. Even under conditions in which the gp130␥ deletion complex formation, because incubation of wild-type sgp130 with mutants were overexpressed in COS-7 cells, no significant STAT1 IL-6 in the absence of sIL-6R as well as precipitation using su- activation was observed in response to IL-6/sIL-6R (Fig. 4C). pernatants from untransfected cells does not lead to coprecipitation Thus, each of the membrane-proximal domains of gp130 is re- of IL-6 (control lanes). The amount of coprecipitated IL-6 does not quired for signal transduction by IL-6-type cytokines that induce substantially differ between sgp130-flag and the deletion con- homodimerization of gp130. structs. Even at low IL-6 concentrations no significant difference in ternary complex formation is observed, indicating that the soluble The deletion mutants show differential responses upon gp130 deletion mutants bind the ligand with comparable affinities. stimulation by agonistic gp130 mAbs Thus, the ligand binding capability of soluble gp130 is not affected gp130 can efficiently be activated by combination of the mAbs by deletion of individual membrane-proximal domains. B-S12-G7 and B-P8 (26) (unpublished observations). To investi- The deletion mutants ⌬4, ⌬5, and ⌬6 were stably transfected gate how the gp130 deletion mutants respond to stimulation by the into Ba/F3 cells, a mouse pro-B cell line that does not express agonistic gp130 Abs, Ba/F3 cells stably expressing the deletion gp130 endogenously. FACS analyses using the two gp130 mAbs mutants were incubated with equimolar concentrations of B-P8 and B-S12-G7, which map to the membrane-distal part of B-S12-G7 and B-P8. Interestingly, the mutants ⌬4 and, to a lesser gp130, revealed that all deletion mutants are expressed on the cell extent, ⌬6 respond to the mAbs with activation of STAT3, The Journal of Immunology 277 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 3. Binding of soluble and membrane-bound wild-type gp130 and deletion mutants to IL-6/sIL-6R complexes. A, The gp130 and the deletion mutants ⌬4, ⌬5, and ⌬6 were secreted by transfected COS-7 cells as soluble proteins containing the ectodomain fused to a C-terminal Flag epitope (sgp130-Flag, s⌬4-Flag, s⌬5-Flag, and s⌬6-Flag). Seventy-two hours after transfection the recombinant proteins were immunoprecipitated from cell supernatants using a Flag Ab. Precipitates were incubated with an excess of sIL-6R (70 nM) and varying concentrations of IL-6 as indicated in the figure. Ternary complexes formed (see scheme) were coprecipitated and separated by SDS-PAGE. Subsequently, the proteins were transferred to a polyvinylidene difluoride membrane for immunoblot analysis. Proteins were visualized by enhanced chemiluminescence using the Abs indicated. Supernatants from untransfected COS-7 cells served as a control (co). B, Cell surface expression of gp130 and the deletion constructs ⌬4, ⌬5, and ⌬6 in stably transfected Ba/F3 cells was analyzed by FACS. Untransfected cells served as a control (co). Cells were incubated with the gp130 mAbs B-P8 or B-S12-G7, respectively, and stained with a PE-labeled secondary Ab (open peaks). The filled peaks correspond to cells incubated with secondary Ab alone. C, Ba/F3 cells (5 ϫ 106) stably transfected with gp130 or the deletion mutants ⌬4, ⌬5, and ⌬6 as well as untransfected cells were incubated with 200 nM sIL-6R and different concentrations of 125I-labeled IL-6 (1340 cpm/fmol) for 16 h at 4°C. Radioactivity bound to untransfected cells was subtracted from radioactivity bound to transfected cells to obtain specific binding. Specifically bound radioactivity is presented as a function of the 125I-labeled IL-6 concentration.

whereas the ⌬5 mutant did not induce signiﬁcant STAT activation Ab concentrations no maximal response was achieved. The weak (Fig. 5A). Furthermore, the proliferative response of the trans- STAT activation induced by ⌬6 shown in Fig. 5A was not sufﬁ- fected Ba/F3 cells upon stimulation by agonistic Abs was ana- cient to induce proliferation of Ba/F3 cells. In line with the ob- lyzed. ⌬4 transduces a proliferative signal, although much less served lack of STAT activation, ⌬5 did not induce a proliferative pronounced than that of wild-type gp130 (Fig. 5B). Even at high response. In summary, domains 4 and 6 are strictly required for 278 MEMBRANE-PROXIMAL DOMAINS OF gp130 Downloaded from http://www.jimmunol.org/

FIGURE 5. Activation of gp130 and deletion mutants by agonistic mAbs. A, Ba/F3 cells stably transfected with gp130 and the deletion mutants ⌬4, ⌬5, and ⌬6 were stimulated with equimolar amounts of the agonistic mAbs B-S12-G7 and B-P8 (1 ␮g/ml each) for 15 min. Subse-

quently, nuclear extracts were prepared, and STAT activation was analyzed by guest on September 28, 2021 by EMSA. B, Ba/F3 cells (4 ϫ 105/ml) were seeded in a 96-well plate and incubated with increasing equimolar amounts of the agonistic mAbs B-S12-G7 and B-P8 as indicated in the diagram. After 72 h proliferation of the cells was measured as described in Fig. 3B (XTT assay).

FIGURE 4. Biological activity of gp130 and the deletion mutants ⌬4, ⌬5, and ⌬6. A, Ba/F3 cells stably transfected with gp130 or the deletion mutants ⌬4, ⌬5, or ⌬6 were stimulated with IL-6/sIL-6R or IL-11/sIL-11R, Replacement of D5 of gp130 by the corresponding domain of respectively, as indicated. After 15 min, nuclear extracts were prepared, the G-CSFR and the protein concentrations were quantified. Ten micrograms of nuclear ⌬ 32 Although ligand binding capability is fully retained, the 5 dele- protein was analyzed by EMSA using a P-labeled m67SIE probe derived tion mutant is unable to transduce a signal in response to various from the c-Fos promotor providing a binding site for STAT1 and STAT3. stimuli. Thus, D5 plays a central role in coupling ligand binding to Protein-DNA complexes were separated by PAGE and visualized by au- toradiography. STAT3 and STAT1 homodimers as well as STAT3/1 het- the membrane-distal part of gp130 with intracellular signal trans- erodimers are indicated by arrowheads. B, Ba/F3 cells (4 ϫ 105/ml) stably duction. To assess the specificity of D5 function in the formation transfected with gp130 or the deletion mutants ⌬4, ⌬5, and ⌬6 were seeded of the active gp130 homodimer, D5 of gp130 was replaced by the in a 96-well plate and incubated with 0.5 ␮g/ml sIL-6R and increasing corresponding domain of the G-CSFR. The G-CSFR was chosen amounts of IL-6 as indicated in the diagram. After 72 h a tetrazolium because its domain architecture is identical with that of gp130, and compound was added as a substrate and incubated for5hat37°C. Sub- moreover, this receptor shares 46% sequence similarity with sequently, the absorbances at 450 and 690 nm were measured. The differ- gp130. A fragment of the human G-CSFR cDNA encoding the ence in absorbances corresponds to the number of metabolically active corresponding domain was amplified by PCR and cloned into the cells (XTT proliferation assay). C, COS-7 cells were transfected with ex- ⌬5 deletion constructs to obtain D5-GCSFR (schematically repre- pression vectors encoding gp130␥, ⌬4␥, ⌬5␥, and ⌬6␥, respectively. For- sented in Fig. 6A) and D5-GCSFR␥. The chimera D5-GCSFR␥ ty-eight hours after transfection cells were stimulated with IL-6/sIL-6R for 15 min (ϩ) or were left unstimulated (Ϫ). Nuclear extracts were prepared was analyzed using the STAT1 activation assay in COS-7 cells, and analyzed for STAT activation as described above. Untransfected cells because several attempts to stably transfect Ba/F3 cells with the served as a control (co). The arrowhead indicates the STAT1 homodimer. D5-GCSFR construct failed. Unexpectedly, transfection of D5- GCSFR␥ resulted in an activation of STAT1 that is independent from cytokine stimulation, suggesting a constitutive activation of activation of gp130 by IL-6 and IL-11, but a partial activation can this gp130/G-CSFR chimera (Fig. 6B, upper panel). Analysis of be achieved by stimulation with agonistic Abs. In the absence of the time course of constitutive receptor activation revealed that D5, neither treatment with agonistic mAbs nor stimulation by cy- STAT activation reaches its maximal level within 40 h after trans- tokines leads to the formation of an active gp130 homodimer. fection and is sustained for at least 65 h (Fig. 6B, lower panel). The Journal of Immunology 279 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 6. Functional characterization of the gp130/G-CSFR chimera. A, Schematic representation of the chimera D5-GCSFR in comparison to gp130 (white) and G-CSFR (gray). B, COS-7 cells were transfected with expression vectors encoding gp130␥ or D5-GCSFR␥. Forty-eight hours after transfection cells were stimulated with IL-6/sIL-6R for 15 min (ϩ) or were left unstimulated (Ϫ, upper panel). Time course of constitutive STAT activation was analyzed by preparation of nuclear extracts at the time points after transfection indicated (lower panel). Nuclear extracts were prepared and analyzed by EMSA as described in Fig. 4. C, COS-7 cells were transfected with expression vectors encoding gp130␥ or D5-GCSFR␥. For FACS analyses, 48 h after transfection COS-7 cells were incubated with the gp130 mAb B-S12 and stained with a PE-labeled secondary Ab. Untransfected cells served as a control (co). Fluorescence in the range of marker 1 (M1) indicates gp130 overexpression due to transfection of cells. D, COS-7 cells were transfected with expression vectors encoding gp130␥ or D5-GCSFR␥. Forty-eight hours after transfection cells were stimulated with IL-6/sIL-6R for 15 min (ϩ) or were left unstimulated (Ϫ). Jak1 was immunoprecipitated from cell lysates, and precipitated proteins were separated by SDS-PAGE followed by immunoblotting. Tyrosine phosphorylation was detected using a phosphotyrosine Ab (upper panel). The blot was reprobed using a Jak1 Ab to control equal loading of the lanes (lower panel).

Surface expression of D5-GCSFR␥ and gp130␥ was compared by stimulation of the reporter gene by endogenous gp130 (control FACS analysis. Cells transfected with gp130␥ or D5-GCSFR␥ lanes). Cells additionally transfected with gp130 are more sensitive show a partial shift of the peak to higher ﬂuorescence intensity to IL-6 stimulation, and therefore, strong induction of the lucif- compared with untransfected cells, indicating that the chimeric erase gene is observed. In cells transfected with D5-GCSFR no protein appears at the cell surface (Fig. 6C). Tyrosine phosphor- constitutive activation of the reporter gene is detectable. However, ylation of Jak1 is the most upstream event in gp130-dependent these cells respond to IL-6 comparably to cells transfected with activation of the Jak/STAT pathway. As shown in Fig. 6D, Jak1 is wild-type gp130. Therefore, in HepG2 cells replacement of D5 of constitutively phosphorylated in cells transfected with D5- gp130 by the corresponding domain of the GCSFR retains the GCSFR␥, whereas phosphorylation of Jak1 in gp130␥-transfected functionality of gp130. cells is observed only after IL-6 stimulation. To analyze whether the observed constitutive activation of the chimera depends on the cellular environment, HepG2 cells were Discussion transiently transfected with wild-type gp130 or D5-GCSFR. To gp130 is the common signal transducing receptor subunit of the ␣ measure induction of the acute phase protein 2-macroglobulin, a IL-6-type cytokines. Signal transduction is triggered either by ␣ reporter gene construct was cotransfected containing the 2-mac- gp130 homodimerization in response to IL-6 and IL-11 or het- roglobulin promoter followed by the luciferase gene. Luciferase erodimerization of gp130 with a second signal transducing recep- activity was measured in stimulated as well as unstimulated cells tor (LIFR or OSMR) in response to LIF, CNTF, cardiotrophin-1, (Fig. 7). Cells were stimulated with low amounts of IL-6 to prevent or OSM. An intriguing structural feature of the signal transducing 280 MEMBRANE-PROXIMAL DOMAINS OF gp130

bind IL-6/sIL-6R complexes indistinguishably from wild-type gp130. Moreover, previous work deﬁned the ligand-binding epitopes in the membrane-distal domains of gp130 (13, 15). Each deletion mutant was unable to induce signal transduction in response to stimulation by both IL-6 and IL-11 suggesting a functional role of each of the membrane-proximal domains in receptor activation. Intriguingly, using different cellular systems no activation of ⌬5 was observed, although this deletion mutant, even in its membrane-bound form, binds IL-6/sIL-6R comparable to wild-type gp130. In the case of membrane-bound ⌬4 and ⌬6 the lack of biological activity may partially be due to the reduced afﬁnity to IL-6/sIL-6R complexes. Therefore, responsiveness of the deletion mutants to agonistic mAbs was studied. The Abs are believed to activate cytokine receptors by virtue of their bivalency that enables them to dimerize the receptor. Whereas deletion of D5 led to a loss of activity even by stimulation with the Abs, the ⌬ ⌬ FIGURE 7. Induction of the ␣ -macrogloblulin promotor in HepG2 mutants 4 and, to a lesser extent, 6 retained residual activity. 2 ⌬ cells transfected with wild-type gp130 or D5-GCSFR. HepG2 cells were The total loss of activity of 5 is not due to participation of D5 in Downloaded from transiently transfected with expression vectors encoding wild-type gp130 the Ab epitope because this deletion mutant is recognized by both ␣ or D5-GCSFR. Cells were cotransfected with an 2-macroglobulin promo- agonistic Abs in FACS analysis. Furthermore, in a previous study tor luciferase gene reporter construct and a ␤-galactosidase control vector. the epitopes of the agonistic Abs were mapped to the membrane- Twenty-four hours after transfection cells were stimulated (ϩ) with IL-6 at distal part of gp130 (26). Thus, there seems to exist an absolute Ϫ suboptimal concentration (10 ng/ml) or were left unstimulated ( ). After requirement for domain 5 to achieve activation of gp130. None of 16 h luciferase activity was measured and normalized to ␤-galactosidase the deletion mutants is stimulated by the Abs to an extent that is

activity to correct for transfection efficiency. Error bars show mean devi- http://www.jimmunol.org/ ations derived from three independent experiments. achieved with wild-type gp130. We conclude from these findings that dimerization of gp130 by the natural ligand or by mAbs is not sufficient for receptor activation. A well-defined active conformation of the receptor has to be adjusted to allow a productive jux- receptor subunits is the presence of three membrane-proximal taposition of the cytoplasmic parts that leads to activation of the FNIII domains in addition to the Ig-like domain and the CBM(s). associated kinases and finally to cytoplasmic signal transduction. In the case of gp130 (13) and LIFR (34) it has been shown that the These conformational requirements cannot be fulfilled in the ab- ligands bind to the CBMs and Ig-like domains of the receptors. To sence of D5. date, the functional relevance of the membrane-proximal domains

To decide whether there is either a specific functional require- by guest on September 28, 2021 of gp130 was unclear. Mapping of a gp130 neutralizing Ab to the ment for D5 of gp130 or whether it plays a rather unspecific role membrane-proximal part of gp130 suggested a functional role of (e.g., as a spacer), this domain was replaced by the corresponding these domains in receptor activation. domain of the homologous G-CSFR. Most surprisingly, upon To achieve a first assessment of the importance of these domains for receptor activation, three deletion mutants of human gp130 overexpression the D5 chimera was ligand independently acti- were generated, each lacking one individual FNIII domain (⌬4, vated, suggesting that the membrane-proximal half of receptors ⌬5, and ⌬6). As shown by FACS analysis, each of the deletion such as gp130 and G-CSFR has an intrinsic propensity to dimerize mutants was expressed in amounts comparable to that of wild-type that might normally be activated upon ligand binding, possibly by gp130. Therefore, deletion of the domains does not lead to struc- enhancing the accessibility of dimer interface(s). Our assumption tural alterations that would impair expression of the mutants. As is that in the D5-GCSFR chimera a kind of disturbance has been soluble proteins all deletion mutants retained the ability to bind introduced that leads to ligand-independent interaction of these IL-6/sIL-6R complexes to the same extent as gp130, confirming proposed dimer interfaces. This idea is supported by the fact that that the membrane-proximal part of wild-type gp130 is not in- chimeric receptors consisting of the ectodomain of the G-CSFR volved in ligand binding. Furthermore, the functionality of the mu- fused to the transmembrane and cytoplasmic parts of gp130 are tants with respect to ligand binding shows that deletion of single only activated in response to G-CSF (35). Thus, D5 of G-CSFR in domains of gp130 does not lead to disintegration of gp130 struc- the context of the complete G-CSFR ectodomain does not lead to ture. However, upon expression of the mutants as transmembrane ligand-independent activation of gp130. It was not possible to sta- proteins on the cell surface, deletion of the membrane-proximal bly transfect this chimera into Ba/F3 cells to test its mitogenic domains differentially interfered with ligand binding. Deletions of activity, suggesting that expression of this chimera is inconsistent D4 and D6 reduced the affinity of gp130 to IL-6/sIL-6R com- with viability of the cells. We therefore transiently expressed the plexes, whereas deletion of D5 had no significant effect on ligand chimera in HepG2 cells and assessed its activity using a reporter binding. The soluble receptor proteins are not restricted with re- gene assay. In these cells, no constitutive activation of D5-GCSFR spect to their relative orientations. For membrane-bound proteins was observed, but the chimera responded normally to IL-6 stim- only two of three translational and one of three rotational degrees ulation. Thus, in HepG2 cells replacement of gp130 D5 by the of freedom are available, leading to strong restrictions of the rel- corresponding domain of the G-CSFR restores a functional recep- ative orientation of the two gp130 molecules in the gp130 dimer. tor. Possibly, constitutive activation of this mutant is only ob- Reduced binding of membrane-bound ⌬4 and ⌬6 may therefore served after strong overexpression, as seen in transfected COS-7 originate from the inability of these mutants to adjust the correct cells (see FACS, Fig. 6C) or critically depends on the cell type. relative orientation required for high affinity ligand binding. Direct Cell type-dependent constitutive activation has also been described involvement in ligand binding in a sense that D4 and D6 contact for activating mutations in ␤c, the common ␤-chain of the recep- the ligand is unlikely, because as soluble proteins both mutants tors for IL-3, IL-5, and GM-CSF (36). The Journal of Immunology 281

FIGURE 8. Proposed mechanism for activation of cytokine receptors. In the case of short cytokine receptors such as EpoR or GHR (left panel) binding of the natural ligand to the CBMs adjusts the active conformation of the homodimer. High afﬁnity binding of the ligand to the CBM and Ig-like domain of gp130 is not sufﬁcient for receptor activation, because in the case of ⌬5(right panel) this does not lead to signal transducing receptor complexes. The

gp130 membrane-proximal domains Downloaded from are necessary for adjustment of the active dimer (central panel). The scheme highlights the functional role of domain 5 (striped) for productive spacing of the cytoplasmic parts of gp130. http://www.jimmunol.org/

We conclude from our studies that gp130 is activated by a pected due to formation of gp130 dimers as a result of random mechanism that differs from the activation mechanism of short interactions. It is tempting to speculate that in addition to the rigid by guest on September 28, 2021 cytokine receptors such as the EpoR or GHR (Fig. 8A, left panel). model presented in Fig. 8, a conformational change in the gp130 From the structure of the GH/sGHR complex it can be deduced ectodomain necessary for receptor activation is induced upon li- that receptor activation is achieved by binding of two receptor gand binding. By this hypothetical mechanism, signal transduction molecules to a single bivalent ligand (37). Most of the dimerization in the absence of the ligand is prevented. The insertion of the energy is contributed by the receptor ligand interaction, which is to analogous domain of the G-CSFR into the completely inactive some extent supported by receptor/receptor interactions. In the deletion mutant ⌬5 leads to a ligand-independent mechanism of case of gp130 we observe a functional dichotomy of the ectodo- receptor activation in COS-7 cells. In the family of cytokine re- main. The membrane-distal half binds the ligand, whereas the ceptors constitutively active variants of the EpoR (39), thrombo- membrane-proximal half mediates receptor activation. Both parts poietin receptor (40), and ␤c (36) have been described (for review, of the molecule must be functionally coupled to ensure that in the see Ref. 41). The ectodomains of these receptors consist solely of wild-type proteins receptor activation occurs only after ligand one (EpoR) or two (thrombopoietin receptor, ␤c) CBMs and lack binding. We present a model of gp130 activation that is in line with additional FNIII domains. For gp130 and receptors of similar the ﬁndings described above (Fig. 8, central panel). D5 adjusts the ectodomain architecture (LIFR, OSMR, G-CSFR, IL-12R, and the correct spacing of the cytoplasmic parts required for signaling. In leptin receptor), the membrane-proximal domains are promising the absence of D5, the ligand can still be bound, but a nonproduc- targets for rational design of constitutively active cytokine receptor tive conformation is adjusted (Fig. 8, right panel). Consequently, variants that can be used in basic research on the physiological partial restoration is achieved by adding back D5 of the G-CSFR. consequences of constitutive receptor activation. In the previously Our model proposes an interaction of domains 5 of two gp130 described cases of constitutively activated cytokine receptors the molecules. Possibly, in the homologous G-CSFR this interaction is oncogenic potential of ligand independent receptor activation has somewhat stronger, leading to constitutive activation of the chi- been established. A role for constitutively active gp130 in the de- mera in COS-7 cells. All members of the cytokine receptor family velopment of human malignancies can also be imagined. that contain additional membrane-proximal domains (LIFR, OSMR, G-CSFR, IL-12R, and leptin receptor) may be activated by Acknowledgments a similar mechanism. Indeed, a role of the membrane-proximal We thank John Wijdenes (Diaclone, Besanc¸on, France) for providing the domains of the G-CSFR in receptor activation has been established gp130 mAbs used in this study, and Drs. Iris Behrmann and Lutz Graeve by showing that simultaneous deletion of the three membrane- for critical reading of the manuscript. proximal domains strongly impairs signal transduction in response to G-CSF, whereas ligand binding is largely unaffected (38). References Even after overexpression in COS-7 cells no constitutive acti- 1. Bazan, J. F. 1990. Structural design and molecular evolution of a cytokine re- vation of wild-type gp130 is observed, although this could be ex- ceptor superfamily. Proc. Natl. Acad. Sci. USA 87:6934. 282 MEMBRANE-PROXIMAL DOMAINS OF gp130

2. Ihle, J. N. 1995. Cytokine receptor signaling. Nature 377:591. man interleukin-6 produced in Escherichia coli from a T7 RNA polymerase 3. Wells, J. A., and A. M. de Vos. 1996. Hematopoietic receptor complexes. Annu. expression vector. Eur. J. Biochem. 198:541. Rev. Biochem. 65:609. 24. Weiergra¨ber, O., U. Hemmann, A. Ku¨ster, G. Mu¨ller-Newen, J. Schneider, 4. Heinrich, P. C., I. Behrmann, G. Mu¨ller-Newen, F. Schaper, and L. Graeve. 1998. S. Rose-John, P. Kurschat, J. P. J. Brakenhoff, M. H. L. Hart, S. Stabel, et al. IL-6-type cytokine signaling through the gp130/JAK/STAT pathway. Biochem. J. 1995. Soluble human interleukin-6 receptor: expression in insect cells, purifica- 334:297. tion and characterization. Eur. J. Biochem. 234:661. 5. Hibi, M., M. Murakami, M. Saito, T. Hirano, T. Taga, and T. Kishimoto. 1990. 25. Mu¨ller-Newen, G., A. Ku¨ster, U. Hemmann, R. Keul, U. Horsten, A. Martens, Molecular cloning and expression of an IL-6 signal transducer, gp130. Cell 63: L. Graeve, J. Wijdenes, and P. C. Heinrich. 1998. Soluble interleukin-6 receptor 1149. potentiates the antagonistic activity of soluble gp130 on interleukin-6 responses. 6. Murakami, M., M. Hibi, N. Nakagawa, T. Nakawaga, K. Yasukawa, J. Immunol. 161:6347. K. Yamanishi, T. Taga, and T. Kishimoto. 1993. IL-6-induced homodimerization 26. Wijdenes, J., P. C. Heinrich, G. Mu¨ller-Newen, C. Roche, G. Zong-Jiang, of gp130 and associated activation of a tyrosine kinase. Science 260:1808. C. Clement, and B. Klein. 1995. Interleukin-6 signal transducer gp130 has spe- 7. Yin, T., T. Taga, M. L. Tsang, K. Yasukawa, T. Kishimoto, and Y. C. Yang. cific binding sites for different cytokines as determined by antagonistic and ag- 1993. Involvement of IL-6 signal transducer gp130 in IL-11-mediated signal onistic anti-gp130 mAbs. Eur. J. Immunol. 25:3474. transduction. J. Immunol. 151:2555. 27. Gerhartz, C., B. Heesel, J. Sasse, U. Hemmann, C. Landgraf, 8. Gearing, D. P., M. R. Comeau, D. J. Friend, S. D. Gimpel, C. J. Thut, J. Schneider-Mergener, F. Horn, P. C. Heinrich, and L. Graeve. 1996. Differential J. McGourty, K. K. Brasher, J. A. King, S. Gillis, B. Mosley, et al. 1992. The IL-6 activation of acute phase response factor/STAT3 and STAT1 via the cytoplasmic signal transducer, gp130: an oncostatin M receptor and affinity converter for the domain of the interleukin-6 signal transducer gp130. I. Definition of a novel LIF receptor. Science 255:1434. phosphotyrosine motif mediating STAT1 activation. J. Biol. Chem. 271:12991. 9. Davis, S., T. H. Aldrich, N. Stahl, L. Pan, T. Taga, T. Kishimoto, N. Y. Ip, and 28. Yuan, J., U. M. Wegenka, C. Lu¨tticken, J. Buschmann, T. Decker, C. Schindler, G. D. Yancopoulos. 1993. LIFR␤ and gp130 as heterodimerizing signal trans- P. C. Heinrich, and F. Horn. 1994. The signaling pathways of interleukin-6 and ducers of the tripartite CNTF receptor. Science 260:1805. interferon-␥ converge by the activation of different transcription factors which 10. Pennica, D., K. J. Shaw, T. A. Swanson, M. W. Moore, D. L. Shelton, bind to common responsive DNA elements. Mol. Cel. Biol. 14:1657. K. A. Zioncheck, A. Rosenthal, T. Taga, N. F. Paoni, and W. I. Wood. 1995. 29. Bolton, A. E., and W. M. Hunter. 1973. The labelling of proteins to high specific Cardiotrophin-1: biological activities and binding to the leukemia inhibitory fac- 125 tor receptor/gp130 signaling complex. J. Biol. Chem. 270:10915. radioactivities by conjugation to a I-containing acylating agent. Biochem. J. Downloaded from 11. Mosley, B., C. De Imus, D. Friend, N. Boiani, B. Thoma, L. S. Park, and 133:529. D. Cosman. 1996. Dual oncostatin M (OSM) receptors: cloning and character- 30. Wegenka, U. M., C. Lu¨tticken, J. Buschmann, J. Yuan, F. Lottspeich, W. Mu¨ller- ization of an alternative signaling subunit conferring OSM-specific receptor ac- Esterl, C. Schindler, E. Roeb, P. C. Heinrich, and F. Horn. 1994. The interleukin- tivation. J. Biol. Chem. 271:32635. 6-activated acute-phase response factor is antigenically and functionally related 12. Taga, T., M. Hibi, Y. Hirata, K. Yamasaki, K. Yasukawa, T. Matsuda, T. Hirano, to members of the signal transduction and transcription (Stat) factor family. Mol. and T. Kishimoto. 1989. Interleukin-6 triggers the association of its receptor with Cell. Biol. 14:3186. a possible signal transducer, gp130. Cell 58:573. 31. Wagner, B. J., T. E. Hayes, C. J. Hoban, and B. H. Cochran. 1990. The SIF binding element confers sis/PDGF inducibility onto the c-fos promoter. EMBO J. 13. Horsten, U., H. Schmitz-Van de Leur, J. Mu¨llberg, P. C. Heinrich, and http://www.jimmunol.org/ S. Rose-John. 1995. The membrane distal half of gp130 is responsible for the 9:4477. formation of a ternary complex with IL-6 and the IL-6 receptor. FEBS Lett. 32. Kyhse-Anderson, J. 1984. Electroblotting of multiple gels: a simple apparatus 360:43. without buffer tank for rapid transfer of proteins form polyacrylamide to nitro- 14. Hammacher, A., R. T. Richardson, J. E. Layton, D. K. Smith, L. J. L. Angus, cellulose. J. Biochem. Biophys. Methods 10:203. D. J. Hilton, N. A. Nicola, J. Wijdenes, and R. J. Simpson. 1998. The immuno- 33. Lebeau, B., F. A. Montero-Julian, J. Wijdenes, G. Mu¨ller-Newen, H. Dahmen, M. globulin-like module of gp130 is required for signaling by interleukin-6 but not Che´rel, P. C. Heinrich, H. Brailly, M.-M. Hallet, A. Godard, et al. 1997. Recon- by leukemia inhibitory factor. J. Biol. Chem. 273. stitution of two isoforms of the human interleukin-11 receptor and comparison of 15. Kurth, I., U. Horsten, S. Pflanz, H. Dahmen, A. Ku¨ster, J. Gro¨tzinger, their functional properties. FEBS Lett. 407:141. P. C. Heinrich, and G. Mu¨ller-Newen. 1999. Activation of the signal transducer 34. Owczarek, C. M., Y. Zhang, M. J. Layton, D. Metcalf, B. Roberts, and gp130 by both interleukin-6 and interleukin-11 requires two distinct binding N. A. Nicola. 1997. The unusual species cross-reactivity of the leukemia inhib- epitopes. J. Immunol. 162:1480. itory factor receptor ␣-chain is determined primarily by the immunoglobulin-like

16. Dahmen, H., U. Horsten, A. Ku¨ster, Y. Jacques, S. Minvielle, I. M. Kerr, domain. J. Biol. Chem. 272:23976. by guest on September 28, 2021 G. Ciliberto, G. Paonessa, P. C. Heinrich, and G. Mu¨ller-Newen. 1998. Activa- 35. Baumann, H., A. J. Symes, M. R. Comeau, K. K. Morella, Y. Wang, D. Friend, tion of the signal transducer gp130 by interleukin-11 and interleukin-6 is medi- S. F. Ziegler, J. S. Fink, and D. P. Gearing. 1993. Multiple regions within the ated by similar molecular interactions. Biochem. J. 331:695. cytoplasmic domains of leukemia inhibitory factor receptor and gp130 cooperate 17. Staunton, D., K. R. Hudson, and J. K. Heath. 1998. The interactions of the in signal transduction in hepatic and neuronal cells. Mol. Cell. Biol. 14:138. cytokine-binding homology region and immunoglobulin-like domains of gp130 36. Jenkins, B. J., R. J. D‘Andrea, and T. J. Gonda. 1995. Activating point mutations with oncostatin M: implications for receptor complex formation. Protein Eng. in the common ␤-subunit of the human GM-CSF, IL-3 and IL-5 receptors suggest 11:1093. the involvement of ␤-subunit dimerisation and cell type-specific molecules in 18. Horsten, U., G. Mu¨ller-Newen, C. Gerhartz, A. Wollmer, J. Wijdenes, signaling. EMBO J. 14:4276. P. C. Heinrich, and J. Gro¨tzinger. 1997. Molecular modeling-guided mutagenesis 37. de Vos, A. M., M. Ultsch, and A. A. Kossiakoff. 1992. Human growth hormone of the extracellular part of gp130 leads to the identification of contact sites in the and extracellular domain of its receptor: crystal structure of the complex. Science interleukin-6 (IL-6)/IL-6 receptor/gp130 complex. J. Biol. Chem. 272:23748. 255:306. 19. Bravo, J., D. Staunton, J. K. Heath, and E. Y. Jones. 1998. Crystal structure of a cytokine-binding region of gp130. EMBO J. 17:1665. 38. Fukunaga, R., E. Ishizaka-Ikeda, C.-X. Pan, Y. Seto, and S. Nagata. 1991. Func- 20. Kernebeck, T., S. Pflanz, G. Mu¨ller-Newen, G. Kurapkat, R. M. Scheek, tional domains of the granulocyte colony-stimulating factor receptor. EMBO J. K. Dijkstra, P. C. Heinrich, A. Wollmer, S. Grzesiek, and J. Gro¨tzinger. 1999. 10:2855. The signal transducer gp130: solution structure of the carboxy-terminal domain 39. Yoshimura, A., G. Longmore, and H. F. Lodish. 1990. Point mutation in the of the cytokine receptor homology region. Protein Sci. 8:1. exoplasmic domain of the erythropoietin receptor resulting in hormone-indepen- 21. Lemmon, M. A., and J. Schlessinger. 1994. Regulation of signal transduction and dent activation and tumorigenicity. Nature 348:647. signal diversity by receptor oligomerization. Trends Biochem. Sci. 19:459. 40. Souyri, M., I. Vigon, J. Pencioletti, J. Heard, P. Tambourin, and F. Wendling. 22. Heldin, C.-H. 1995. Dimerization of cell surface receptors in signal transduction. 1990. A putative truncated cytokine receptor gene transduced by the myelopro- Cell 80:213. liferative leukaemia virus immortalises hematopoietic progenitors. Cell 63:1137. 23. Arcone, R., P. Pucci, F. Zappacosta, V. Fontaine, A. Malorni, G. Marino, and 41. Gonda, T. J., and R. J. D’Andrea. 1997. Activating mutations in cytokine recep- G. Ciliberto. 1991. Single-step purification and structural characterization of hu- tors: implications for receptor function and role in disease. Blood 89:355.