A Cell Type-specific Constitutive Point Mutant of the Common β-Subunit of the Human Granulocyte-Macrophage Colony-stimulating Factor (GM-CSF), Interleukin (IL)-3, and IL-5 Receptors Requires the GM-CSF Receptor α-Subunit for Activation*

The high affinity receptor for human granulocyte-macrophage colony-stimulating factor (GM-CSF) consists of a cytokine-specific α-subunit (hGMRα) and a common signal-transducing β-subunit (hβc) that is shared with the interleukin-3 and -5 receptors. We have previously identified a constitutively active extracellular point mutant of hβc, I374N, that can confer factor independence on murine FDC-P1 cells but not BAF-B03 or CTLL-2 cells (Jenkins, B. J., D’Andrea, R. J., and Gonda, T. J. (1995) EMBO J. 14, 4276–4287). This restricted activity suggested the involvement of cell type-specific signaling molecules in the activation of this mutant. We report here that one such molecule is the mouse GMRα (mGMRα) subunit, since introduction of mGMRα, but not hGMRα, into BAF-B03 or CTLL-2 cells expressing the I374N mutant conferred factor independence. Experiments utilizing mouse/human chimeric GMRα subunits indicated that the species specificity lies in the extracellular domain of GMRα. Importantly, the requirement for mGMRα correlated with the ability of I374N (but not wild-type hβc) to constitutively associate with mGMRα. Expression of I374N in human factor-dependent UT7 cells also led to factor-independent proliferation, with concomitant up-regulation of hGMRα surface expression. Taken together, these findings suggest a critical role for association with GMRα in the constitutive activity of I374N.

GM-CSF 1 is a potent cytokine that promotes the survival, proliferation, differentiation, and functional activity of a wide variety of hemopoietic cell types including monocytes/macrophages, granulocytes, and myeloid progenitor cells (reviewed in Ref. 1). Like other cytokines, GM-CSF exerts its biological activities through binding to specific receptors on the surface of target cells. The high affinity receptor for human GM-CSF (hGMR) is composed of a cytokine-specific ␣-subunit (hGMR␣) associated with a common signal-transducing ␤-subunit (h␤c) that is also utilized by the IL-3 and IL-5 receptors (2)(3)(4)(5)(6), all of which belong to the cytokine receptor family (reviewed in Ref. 7). Members of this family are characterized by a structurally conserved extracellular cytokine receptor module (CRM) of about 200 amino acids that consists of two fibronectin type III-like domains (8). The ␤-subunit has two CRMs, whereas the ␣-subunits contain one CRM and an additional N-terminal domain of about 100 amino acids.
Although the stoichiometry of subunits in active hGMR, hIL-3R, and hIL-5R complexes remains unresolved, it has become clear that ligand-induced ␣-␤-subunit heterodimerization is a key step in the formation of these complexes (9,10). More recently, it has been shown that ␤-subunit homodimers are found in active hGMR (11) and human IL-3R (12) complexes and that the functional hGMR complex may contain at least two ␣-subunits (13). Taken together, these results suggest that the ␣and ␤-subunits may form higher order receptor complexes, and indeed it has been proposed that the GMR/IL-3R/ IL-5R normally functions as an ␣ 2 ␤ 2 tetramer (10,12,13).
The isolation of constitutively active cytokine receptor mutants has provided a useful tool for examining the normal activation process of some receptors (e.g. erythropoietin receptor and c-Mpl (14,15)), since these mutant receptors most likely mimic the structure of the normal cytokine-activated receptors. With regard to the GMR/IL-3R/IL-5R system, we have previously combined random mutagenesis with retroviral expression cloning to identify constitutively activating point mutations in h␤c by virtue of their ability to confer factorindependent proliferation on mouse factor-dependent FDC-P1 cells (16,17). One of these mutations, V449E, is located in the transmembrane domain of h␤c and is similar to an activating mutation in the neu/c-erbB-2 oncogene (18,19). By analogy, this mutant most likely acts by inducing h␤c homodimerization. Another group of activating point mutations, exemplified by I374N, lies in the extracellular region of h␤c; however, it is unclear precisely how this group might affect receptor function. Interestingly, only certain transmembrane mutants, such as V449E, were able to confer factor independence on mouse factordependent BAF-B03 cells, suggesting that the I374N mutation activates h␤c in a cell type-specific manner.
One possible explanation for the cell type specificity of the I374N mutant is that a molecule that is present in FDC-P1 * This work was supported by a research grant from the National Health and Medical Research Council (NHMRC) of Australia. 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 with 18 1 The abbreviations used are: GM-CSF granulocyte-macrophage colony-stimulating factor; mGM-CSF, mouse GM-CSF; GMR, GM-CSF receptor; hGMR, human GMR; mGMR, mouse GMR; GMR␣, GMR ␣-subunit; IL, interleukin; hIL, human IL; IL-3R and IL-5R, interleukin-3 and -5 receptors, respectively; ␤c, common ␤-subunit of the GM-CSF, IL-3 and IL-5 receptors; h␤c, human ␤c; CRM, cytokine receptor module; PCR, polymerase chain reaction; kb, kilobase pair(s); wt, wild type; HSV, herpes simplex virus.
(and other myeloid) cells is required for its constitutive activity. We report here the use of retroviral expression cloning to identify the mouse GMR␣ (mGMR␣) subunit as one such molecule and show that one effect of the I374N mutation is to induce constitutive association with mGMR␣.

EXPERIMENTAL PROCEDURES
Cell Lines-BOSC 23 (20) and ⌿2 (21) ecotropic retroviral packaging cell lines were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. The BING amphotropic retroviral packaging cell line was kindly provided by Prof. Suzanne Cory (Walter and Eliza Hall Institute, Melbourne, Victoria, Australia) with permission from Dr. Warren Pear (MIT, Cambridge, MA) and was maintained as described above. The CTL-EN subline of the mouse IL-2-dependent cell line, CTLL-2 (22), was kindly provided by Dr. John Norton (Paterson Institute for Cancer Research, Manchester) and was maintained as described previously for CTLL-2 cells (16). Mouse IL-3-dependent BAF-B03 cells (23) were maintained as described previously (16). Human factor-dependent UT7 cells (24) were maintained in Dulbecco's modified Eagle's medium plus 10% fetal calf serum supplemented with 2 ng/ml human GM-CSF.
Construction of the FDC-P1 cDNA Library-cDNA library construction was performed essentially as described by Rayner and Gonda (25). Briefly, cDNA was synthesized from the mouse IL-3/GM-CSF-dependent myeloid cell line FDC-P1 (26) and size-selected for cDNA fragments greater than 500 base pairs. Following digestion with BamHI and XhoI, the size-selected cDNA was ligated directionally into the pRUFNeo retroviral expression vector (25). The library was amplified in Escherichia coli by electroporation of aliquots of the ligated FDC-P1 cDNA. The resultant colonies from each electroporation were harvested, and plasmid DNA was prepared from each pool.
Infection of Target Cells with the FDC-P1 cDNA Library-Retroviral DNA was used to generate a library of retroviruses by a modification of the method described by Rayner and Gonda (25). Briefly, amphotropic BING packaging cells were transiently transfected using the procedure described by Jenkins et al. (27) with 10 g of retroviral plasmid per 60-mm culture dish (seeded 18 h previously with 2 ϫ 10 6 cells). At 48 h post-transfection, virus-containing supernatants were filtered and used to infect ecotropic ⌿2 packaging cells. Infected ⌿2 cells were harvested and selected in medium containing G418 (400 g/ml) to generate the stable G418-resistant ⌿2 retroviral library. BAF-B03 cells expressing the I374N h␤c mutant were infected with the ⌿2 retroviral library by co-cultivating 3.75 ϫ 10 5 BAF/I374N cells with 1.2 ϫ 10 6 irradiated (30 grays) ⌿2 cells for 48 h in each of eight 25-cm 2 culture flasks. The BAF/I374N cells were then harvested, washed, and selected for factorindependent growth in 24-well multidishes (204 wells, each seeded with 10 5 cells) in liquid culture medium without factor.
PCR Recovery and Sequencing of cDNAs from Factor-independent Cells-PCR was performed on 100 ng of genomic DNA (prepared essentially as described by Hughes et al. (28)) with an XL PCR kit (Perkin-Elmer) under conditions recommended by the manufacturer. The primers used for amplification were RCF1 (25), which corresponds to the vector gag sequence approximately 80 base pairs 5Ј of the polylinker in the pRUFNeo vector and RCR2 (5Ј-ATAGCCTCTCCACCCAAGCG-3Ј), which corresponds to the MC1Neo sequence 364 base pairs 3Ј of the polylinker. PCR products were agarose gel-purified, and the 5Ј-and 3Ј-ends were sequenced with PCR primers. Internal primers corresponding to cDNA sequences obtained from initial sequencing with PCR primers were subsequently used to fully sequence PCR products. Sequencing reactions were performed using a Taq DyeDeoxy Terminator Cycle Sequencing kit (Perkin-Elmer), and sequence data were obtained by running reactions on an ABI Prism 377 DNA Sequencer.
Receptor Expression Constructs-The pRUFNeo/mGMR␣ expression construct was generated by subcloning the full-length mGMR␣ cDNA recovered from factor-independent BAF/I374N infectants into the BamHI and HindIII restriction sites of pRUFNeo. The pRUFNeo/ hGMR␣ expression construct was generated by inserting the cDNA for hGMR␣ into the XhoI site of pRUFNeo.
To introduce the 8-amino acid DYKDDDDK FLAG polypeptide (Eastman Kodak Co.) at the N terminus of mGMR␣ ( F mGMR␣), a 5Ј BamHI/ NaeI fragment encoding the signal sequence and first 8 structural residues of mGMR␣ was excised from pRUFNeo/mGMR␣ and replaced in frame with a PCR-generated BamHI/NaeI fragment from pcDNA1Neo/ F hIL-3R␣ (kindly provided by Richard D'Andrea, Hanson Center for Cancer Research, Adelaide, South Australia, Australia) encoding the hIL-3R␣ signal sequence, FLAG octapeptide, and first 6 structural residues of hIL-3R␣. The sense primer corresponded to the T7 promoter sequence and included a BamHI site, and the antisense primer corresponded to codons 19 -24 (as numbered by Kitamura et al. (5)) of hIL-3R␣ and included a NaeI site. The pRUFPuro/ F mGMR␣ expression vector was constructed by inserting the BamHI/EcoRI F mGMR␣ cDNA from pRUFNeo/ F mGMR␣ into the BamHI and EcoRI sites of the pRUFPuro retroviral expression vector (16).
The HSV-derived 11-amino acid QPELAPEDPED polypeptide (Novagen) was inserted after the signal sequence of the wild-type and I374N mutant ␤-subunits (between residues Cys 16 and Trp 17 as numbered by Hayashida et al. (4)) by site-directed mutagenesis using the pAlter-1 system (Promega) in accordance with the manufacturer's instructions. The modified ␤-subunit cDNAs were subcloned into the BamHI and HindIII restriction sites of pRUFNeo.
The following GMR␣ chimeras were generated by PCR amplification and ligation of the relevant portions of human and mouse GMR␣: (i) the pRUFNeo/h␣m␣1 chimera encoding the extracellular and transmembrane domains of hGMR␣ (346 amino acids) and the cytoplasmic domain of mGMR␣ (38 amino acids); (ii) the pRUFNeo/h␣m␣2 chimera encoding the extracellular N-terminal domain of hGMR␣ (117 amino acids) and the extracellular CRM, transmembrane, and cytoplasmic domains of mGMR␣ (262 amino acids); (iii) the pRUFNeo/ F m␣h␣1 chimera encoding the extracellular and transmembrane domains of F mGMR␣ (335 amino acids) and the cytoplasmic domain of hGMR␣ (54 amino acids); and (iv) the pRUFNeo/ F m␣h␣2 chimera encoding the extracellular FLAG-tagged N-terminal domain of F mGMR␣ (111 amino acids) and the extracellular CRM, transmembrane, and cytoplasmic domains of hGMR␣ (283 amino acids). A full description of the templates and primers used is available upon request.
Extracellular truncations of mGMR␣ were generated by PCR on the pRUFNeo/ F mGMR␣ construct with primers designed to amplify the entire construct except for the desired extracellular sequence to be removed while leaving the N-terminal signal sequence and FLAG octapeptide intact. Each PCR was performed with different sense primers corresponding to codons 97-102 (for m␣D1) and codons 195-200 (for m␣D2) and the same antisense primer corresponding to codons 9 -14 of mGMR␣. The blunt ends of each PCR fragment were then ligated together in frame.
The cytoplasmic truncation mutant of mGMR␣ was generated by PCR on the pRUFNeo/ F mGMR␣ construct with RCF1 as the sense primer and an antisense primer that contained codons 344 -339 of the mGMR␣ cytoplasmic domain together with a HindIII restriction site and termination codon. The PCR products were subcloned into the BamHI and HindIII restriction sites of pRUFNeo.
All PCRs were performed on 20 ng of plasmid DNA with Pfu DNA polymerase (Stratagene) under conditions recommended by the manufacturer. The structures of all mutated or chimeric cDNAs were verified by sequencing.
Infection of Hemopoietic Cells-Retroviral infection of mouse BAF-B03 cells and CTL-EN cells was performed using either stably transfected ⌿2 packaging cells (16) or transiently transfected BOSC 23 packaging cells as described previously (27). Infected BAF-B03 cells were selected in liquid culture medium containing growth factor and either G418 (1.5 mg/ml) or puromycin (2 g/ml). Infected CTL-EN cells were selected as described previously for CTLL-2 cells (16).
Retroviral infection of human UT7 cells was performed using amphotropic BING packaging cells based on the method for infecting mouse hemopoietic cells with BOSC 23-derived retroviruses (27). Briefly, BING cells were transiently transfected with 10 g of retroviral DNA, following which infections were performed by co-cultivating 3 ϫ 10 5 UT7 cells with the BING cells for 48 h in growth medium supplemented with 4 g/ml polybrene. Cells were harvested and selected in liquid culture medium containing growth factor and G418 at 1.5 mg/ml.

Analysis of Receptor Subunit Expression by Flow
Cytometry-Expression of receptor subunits on the surface of infected cells was detected by high sensitivity immunofluorescence followed by flow cytometry on an Epics-Profile II analyzer (Coulter). High sensitivity immunofluorescence was performed by incubating cells with primary antibody followed by biotinylated anti-mouse IgG (Vector Laboratories) and streptavidin-phycoerythrin (Caltag Laboratories). Expression of FLAG epitope-tagged mGMR␣ subunits was detected by staining with the anti-FLAG monoclonal antibody M2 (Kodak), and expression of hGMR␣ subunits was detected by staining with the anti-hGMR␣ monoclonal antibody 8G6 (29). Expression of wild-type and I374N mutant ␤-subunits on the surface of infected BAF-B03 cells was detected by staining with the anti-h␤c monoclonal antibody 1C1 (10), whereas HSV epitope-tagged wild-type and I374N mutant ␤-subunits expressed on the surface of human UT7 cells were detected by staining with an HSV tag monoclonal antibody (Novagen).
Cell Proliferation Assays-Infected cells were washed twice, and triplicate samples of equal cell number (5 ϫ 10 3 ) were cultured in a 96-well microtiter plate with or without appropriate growth factor for 72 h. Cell proliferation was measured by the CellTiter 96 nonradioactive cell proliferation assay (Promega).
Immunoprecipitation and Immunoblotting-Cells (2 ϫ 10 7 ) were cultured overnight in the absence of growth factor and left unstimulated. Cells were washed with cold PBS containing 20 mM sodium orthovanadate and lysed on ice in lysis buffer (50 mM Hepes (pH 7.5), 150 mM NaCl, 10% glycerol, 1% Nonidet P-40, 2 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 1 mM EDTA, 1 mM EGTA, 2 mg/ml iodoacetamide, 0.2 mg/ml trypsin inhibitor (Boehringer Mannheim), and Complete TM protease inhibitor (Boehringer Mannheim)) for 15 min. Insoluble material was removed by centrifugation, and cell lysates were incubated with primary antibody for 2 h at 4°C. Antibodies used for immunoprecipitation were the anti-h␤c antibody 8E4 (30) and the anti-FLAG antibody M2 (Kodak). Immune complexes were precipitated with 75 l of protein A-Sepharose (Amersham Pharmacia Biotech) for 1 h at 4°C, washed three times with lysis buffer, and boiled in 1ϫ reducing SDS sample buffer. In the case of whole cell protein analyses, samples were lysed in buffer without 10% glycerol, and insoluble material was removed and boiled in 1ϫ reducing SDS sample buffer.
Immunoprecipitates were separated by SDS-polyacrylamide gel electrophoresis on 10% gels and electrophoretically transferred to Poly-Screen R polyvinylidene difluoride membranes (NEN Life Science Products). Membranes were then incubated with the anti-h␤c antibody 1C1 (10), the anti-hGMR␣ antibody 8D10 (29), or the biotinylated anti-FLAG antibody BIOM2 (Kodak), as indicated, following which the membranes were washed and incubated with either an alkaline phosphatase-conjugated anti-mouse antibody (Amersham Pharmacia Biotech) or a streptavidin-conjugated alkaline phosphatase antibody (Molecular Probes, Inc., Eugene, OR), as appropriate. Membranes were washed and subjected to enhanced chemifluorescence detection (Amersham Pharmacia Biotech) as per the manufacturer's instructions, following which they were scanned on a FluorImager (Molecular Dynamics, Inc., Sunnyvale, CA). For reprobing, membranes were stripped in 50 mM Tris (pH 7.4), 2% SDS, 100 mM ␤-mercaptoethanol at 55°C for 20 min; washed; and subsequently probed with the indicated antibodies.

Isolation of Factor-independent BAF/I374N Cells Infected with an FDC-P1 cDNA Retroviral Expression Library-We
have previously identified a constitutively activating point mutation, I374N, in the extracellular region of h␤c by virtue of its ability to confer factor-independent growth on FDC-P1 cells (16). Surprisingly, this mutant was unable to confer factor independence on mouse IL-3-dependent BAF-B03 cells, leading us to suggest that the cell type-specific activity of this mutant may reflect the presence of a ␤-subunit-associated signaling molecule in FDC-P1 cells, but not in BAF-B03 cells, that is required by this mutant for constitutive activation (16). We therefore reasoned that the introduction of such a molecule from FDC-P1 cells into BAF-B03 cells expressing the I374N mutant should lead to its constitutive activity and thus render these cells factor-independent.
Using procedures described previously (25), an FDC-P1 cDNA library (ϳ8.5 ϫ 10 5 independent plasmid clones, with an average insert size of 1.1 kb) was generated in the pRUFNeo retroviral expression vector. As described under "Experimental Procedures," the plasmid DNA was used to generate a stable ⌿2 retroviral library estimated to contain ϳ3.5 ϫ 10 6 independent viral producer clones, which should adequately represent all cDNA species present in the plasmid library.
BAF-B03 cells expressing I374N (BAF/I374N) were infected by co-cultivation with the virus-producing ⌿2 cells at an infection frequency of 18% (estimated by colony assays in the presence of G418). As a control, parallel infections were also performed on uninfected BAF-B03 cells and BAF-B03 cells expressing wild-type h␤c. Cells were then selected for factorindependent growth in 24-well multidishes. After 1 week in the absence of factor, 37 of 204 wells seeded with 10 5 infected BAF/I374N cells contained viable, proliferating cells, while no such cells were present in control cultures. Factor independence was not the result of autocrine growth factor production, since conditioned medium from the factor-independent cell cultures did not support the growth of uninfected BAF-B03 cells (data not shown).
PCR Recovery of Mouse GMR␣ cDNA from Factor-independent BAF/I374N Infectants-To identify the cDNA sequence carried by the provirus in the factor-independent BAF/I374N infectants, long range PCR was performed with retroviral primers on genomic DNA samples from 17 of the 37 factorindependent cell populations. This revealed a common fragment of approximately 2.3 kb that was amplified from all 17 genomic DNA samples (data not shown); considering the positions of the PCR primers relative to the cloning sites in pRUF-Neo, the size of the cDNA insert was estimated to be 1.9 kb. For 8 of the 17 samples, the 2.3-kb fragment was the only PCR product generated, suggesting that these factor-independent cell populations contained only one retroviral insertion and that its presence was responsible for factor independence. Sequence analysis of the 1.9-kb cDNA insert recovered from two of the factor-independent cell populations revealed that it cor- responded to the full-length cDNA for the mGMR␣ subunit (31).

Expression of mGMR␣ with I374N in BAF-B03 and CTL-EN Cells Results in Factor
Independence-To confirm that mGMR␣ would allow the constitutive activation of I374N, we expressed the recovered mGMR␣ subunit in BAF/I374N cells and then tested these cells for factor independence. In order to monitor cell surface expression of mGMR␣, a FLAG epitopetagged mGMR␣ ( F mGMR␣) was generated in the pRUFNeo vector (see "Experimental Procedures"). This was introduced into puromycin-resistant BAF/I374N cells as well as wild-type h␤c-expressing and uninfected BAF-B03 cells. Following selection for G418 resistance, flow cytometric analysis with a FLAGspecific monoclonal antibody indicated that the F mGMR␣ subunit was efficiently expressed on the surface of these cells (Fig.  1A). Upon selection for growth in medium without factor, only BAF-B03 cells co-expressing F mGMR␣ and I374N exhibited factor-independent growth (Fig. 1B). The ability of F mGMR␣ to behave as wild-type mGMR␣ was demonstrated by the proliferation of all F mGMR␣-infected BAF-B03 cells in response to mGM-CSF (Fig. 1B).
The observation that the mouse GMR␣ subunit was required for the activity of I374N raised the possibility that another component(s) of the mouse GMR or IL-3R (i.e. mIL-3R␣ m␤c or m␤IL-3) present in FDC-P1 and BAF-B03 cells might also be needed. We therefore introduced I374N and, as a control, wildtype h␤c with F mGMR␣ into mouse IL-2-dependent CTL-EN cells, which do not express any receptor components belonging to the GMR or IL-3R. CTL-EN cells are a derivative of CTLL-2 cells engineered for increased expression of the ecotropic retroviral receptor (41), 2 thereby rendering them more susceptible to retroviral infection. We also included the V449E transmembrane h␤c mutant in this experiment, since it is inactive when expressed in CTLL-2 cells, although, unlike the I374N mutant, it does confer factor independence on BAF-B03 cells (16). The expression of these subunits was confirmed by flow cytometry (data not shown), following which these cells were tested for factor-independent proliferation. As shown in Fig. 2, only CTL-EN cells expressing both F mGMR␣ and I374N were factor-independent, thereby indicating that components of the mouse IL-3R are not required for the constitutive activity of I374N. In view of this result, all subsequent experiments were performed in BAF-B03 cells.
The I374N Mutation Induces Constitutive Association of h␤c with mGMR␣ in BAF-B03 Cells-To examine whether the requirement for mGMR␣ by I374N might reflect a physical association between these two subunits, BAF-B03 cells co-expressing F mGMR␣ with I374N or, as a control, wild-type h␤c were subjected to immunoprecipitation with an anti-h␤c antibody, followed by immunoblot analysis with an anti-FLAG antibody. As shown in Fig. 3A, a protein of 60-kDa, consistent with the predicted size of mGMR␣, was detected only in immunoprecipitates from cell lysates expressing F mGMR␣ and the I374N mutant. Importantly, the converse immunoprecipitation (with anti-FLAG antibody) and immunoblot analysis (with anti-h␤c antibody) confirmed the physical association between mGMR␣ and the I374N mutant (data not shown). Reprobing the immunoblot with an anti-h␤c antibody indicated that both wild-type 2 J. Norton, personal communication. and I374N ␤-subunits were immunoprecipitated from the appropriate cell lysates (Fig. 3B). Furthermore, immunoblot analysis of whole cell lysates with an anti-FLAG antibody indicated that the total levels of F mGMR␣ protein present in lysates from all cell populations were comparable (Fig. 3C). Together, these observations indicate that the I374N mutation acts, at least in part, by inducing constitutive association of h␤c with mGMR␣.
The constitutive association of mGMR␣ with the I374N mutant was reminiscent of the ability of human GMR␣ to associate with wild-type h␤c in the absence of GM-CSF (32). We therefore examined the ability of I374N to associate with hGMR␣ in the absence of ligand, since a failure to do so could explain our previous observation that co-expression of hGMR␣ did not allow constitutive activity of I374N in BAF-B03 cells ( Refs. 16 and 27; Fig. 7). The experiment illustrated in Fig. 4A shows, however, that both mutant and wild-type h␤c could associate equally well with hGMR␣ in the absence (or presence) of ligand, as judged by co-immunoprecipitation from BAF-B03 cells expressing both subunits. Equivalent levels of expression of the ␤and ␣-subunits are confirmed by the analyses of Fig.  4, B and C, respectively.
Both the N-terminal and C-terminal Regions of mGMR␣ Are Essential for Activation of and Association with the I374N Mutant-To broadly define the regions of the mGMR␣ extracellular domain required for the constitutive activation of I374N, two FLAG-tagged extracellular truncation mutants were generated. One of these, F m␣D1, lacked residues Leu 15 -Ala 96 , which comprise the N-terminal domain, whereas the other, F m␣D2, lacked residues Leu 15 -Glu 194 , which also includes domain 1 of the cytokine receptor module (CRM; Fig.  4A). Although these truncation mutants (and full-length F mGMR␣) were efficiently expressed on the surface of G418resistant cells (Fig. 4B), neither truncation mutant was able to confer factor independence on BAF/I374N cells (Fig. 4C), indicating that the N-terminal domain of mGMR␣ is required for constitutive activation of I374N. Furthermore, the inability of BAF/I374N cells expressing the m␣D1 mutant to proliferate in the presence of mGM-CSF suggests that the N-terminal domain of mGMR␣ is also important in normal mGMR function.
Considering that the cytoplasmic domain of GMR␣ is essential for normal GM-CSF-mediated cell growth (33), we also investigated whether the cytoplasmic domain of mGMR␣ was required for constitutive signaling by I374N. We therefore generated a cytoplasmic truncation mutant, F m␣t3, which lacked the C-terminal 14 amino acids of mGMR␣ (Fig. 5A). Although G418-resistant BAF/I374N infectants efficiently expressed F m␣t3 (Fig. 5B), these cells failed to grow in the absence of factor (Fig. 5C) or in response to mGM-CSF. This implies that the C-terminal 14 amino acids of mGMR␣ are essential for mediating factor-independent growth conferred by I374N and also for normal mGM-CSF-mediated growth.
We next examined whether the inability of the extracellular and cytoplasmic truncation mGMR␣ mutants to confer factor independence on BAF/I374N cells was due to a failure to associate with I374N. Lysates from BAF/I374N cells expressing the F m␣D1 extracellular truncation and the F m␣t3 cytoplasmic truncation were therefore subjected to immunoprecipitation with an anti-FLAG antibody, followed by immunoblot analysis with an anti-h␤c antibody. As shown in Fig. 6A, the I374N mutant was precipitated when co-expressed with the fulllength F mGMR␣ subunit but not with the truncated F mGMR␣ subunits. Reprobing with an anti-FLAG antibody demonstrated that both full-length and truncated F mGMR␣ subunits were themselves immunoprecipitated (Fig. 6B), and immunoblot analysis of whole cell lysates with an anti-h␤c antibody confirmed that comparable levels of the I374N mutant were expressed in the cells (Fig. 6C). Thus, these data demonstrate that both the N-terminal and C-terminal regions of mGMR␣ are essential for the association with I374N and, together with the data presented in Fig. 5, that the constitutive activity of I374N is dependent upon this association.
Species Specificity of GMR␣ for the Constitutive Activation of I374N Lies in Its Extracellular and/or Transmembrane Domains-In view of our previous observations that co-expression of the human GMR␣ subunit with I374N in BAF-B03 and CTLL-2 cells did not lead to factor-independent growth (16,27), the ability of the mouse GMR␣ subunit to facilitate constitutive activity of I374N in BAF-B03 and CTL-EN cells was somewhat surprising. To define which region(s) of the GMR␣ subunit govern this apparent species specificity, we constructed a series of chimeric GMR␣ subunits containing regions from both species (Fig. 7A). These chimeras, along with the normal F mGMR␣ and hGMR␣ subunits, were then introduced into BAF/I374N cells and tested for their ability to confer factor independence. Flow cytometric analyses confirmed that while the chimeric GMR␣ subunits were co-expressed with the I374N mutant (Fig.  7B), only cells co-expressing the F m␣h␣1 chimera or, as expected, the normal F mGMR␣ subunit with the I374N mutant exhibited factor-independent proliferation (Fig. 7C). Thus, the species specificity lies in the extracellular and/or transmembrane domains of mGMR␣. Furthermore, since chimeras containing only the mouse N-terminal domain ( F m␣h␣2) or the mouse extracellular CRM and transmembrane domain (h␣m␣2) were unable to confer factor independence on BAF-B03 cells, it is likely that both of the mGMR␣ regions present in these chimeras contribute to the species-specific requirement for mGMR␣ for I374N activity.
The I374N Mutant Confers Factor Independence on Human Hemopoietic Cells: A Possible Role for hGMR␣ in the Constitutive Activity of I374N in Human Cells-Although the human GMR␣ subunit was unable to facilitate the constitutive activity of I374N in mouse BAF-B03 and CTLL-2 cells (16,27) (see also Fig. 7), it was conceivable that the I374N mutant might be constitutively active in human cells expressing hGMR␣. We therefore introduced this mutant and, as a control, wild-type h␤c into human GM-CSF/IL-3/erythropoietin-dependent UT7 cells and tested these cells for factor-independent proliferation. To distinguish between the introduced ␤-subunits and the endogenous ␤-subunits expressed by UT7 cells, we inserted an 11-amino acid HSV-derived epitope at the N terminus of both wild-type and I374N ␤-subunits. Cells infected with these modified ␤-subunits were then selected for G418 resistance or growth in medium without factor. The surface expression of the introduced subunits was confirmed by flow cytometric analysis of infected cells stained with both anti-h␤c and anti-HSV antibodies (Fig. 8A). In two independent experiments, one of which is shown in Fig. 8B, the I374N mutant allowed factorindependent proliferation of UT7 cells. Factor independence was not the result of low level autocrine growth factor production, since conditioned medium from factor-independent cell pools did not support the growth of uninfected UT7 cells (data not shown).
Unfortunately, to the best of our knowledge, no human factordependent hemopoietic cell lines "equivalent" to BAF-B03 cells, i.e. that lack human GMR␣, are available; thus, we could not directly test the requirement for human GMR␣ by I374N in human hemopoietic cells. Notably, however, flow cytometric analysis with an anti-hGMR␣ antibody revealed that the expression of hGMR␣ was significantly up-regulated on the surface of factor-independent cells expressing I374N (FI *I374N cells) compared with uninfected cells or G418-resistant cells (expressing wild-type h␤c or I374N) that were not selected for factor independence (Fig. 8C). Importantly, the increase in hGMR␣ expression specifically correlated with the factor independence of I374N-expressing cells. This increase in hGMR␣ expression was not simply a function of high level ␤-subunit expression (see FI *I374N histogram in Fig. 8A), since infected UT7 cells that were sorted for comparably high levels of HSVtagged wild-type h␤c exhibited a similar low level of hGMR␣ expression to the unsorted cells (*wt) shown in Fig. 8C (data not shown). BAF-B03 or CTLL-2 cells (16), raising the possibility that cell type-specific signaling molecules are involved in its activation. In this study, we have employed retroviral expression cloning to identify the mGMR␣ subunit as one such molecule, since its introduction into BAF-B03 and CTL-EN (a derivative of CTLL-2) cells expressing the I374N mutant conferred factor independence. Importantly, the absence of the mouse GMR and IL-3R in CTL-EN cells indicates that the mechanism of activation of I374N does not require any subunits, apart from mGMR␣, of these receptors. In contrast, another h␤c mutant, V449E, that confers factor independence on both FDC-P1 and BAF-B03 cells (16) is not constitutively active when co-expressed with mGMR␣ in CTL-EN cells. This suggests that the I374N and V449E mutants are activated by fundamentally different mechanisms.
Physical Association of I374N and mGMR␣-Co-immunoprecipitation experiments demonstrated that one effect of the I374N mutation in h␤c is to induce constitutive association with mGMR␣. The constitutive association between these subunits is reminiscent of a recent report in which hGMR␣ and wild-type h␤c were co-immunoprecipitated from cell lines in the absence of GM-CSF (32). Factor-independent association with h␤c appears to be a unique property of GMR␣, since similar preformed complexes could not be detected with hIL-3R␣ or hIL-5R␣ (32). This may in part explain the specific requirement for mGMR␣, as opposed to mIL-3R␣, for constitutive activity of I374N.
We observed that deletions in the extracellular N-terminal domain of mGMR␣ abolished both the constitutive activity of I374N and the association between I374N and mGMR␣, as well as mGM-CSF-induced proliferative signaling. While the corresponding domains of the hIL-3R␣ and hIL-5R␣ subunits have been reported to play a critical role in ligand binding (34 -36), our demonstration that the N-terminal domain of mGMR␣ is required for association with the h␤c mutant suggests that this domain may also play a role in receptor subunit assembly.
Our observation that the cytoplasmic domain of GMR␣ is needed for the activity of I374N was not unexpected, since deletion of the cytoplasmic domains of GMR␣, IL-3R␣, and IL-5R␣ renders these receptors inactive in proliferative signal-ing (33,34,37). Normally, however, ␣-subunit cytoplasmic truncations do not detectably affect the association of ␣and ␤-subunits, since truncated ␣-subunits still form high affinity ligand-binding receptors (33,34,37), and a cytoplasmic truncation of hGMR␣ could still associate with h␤c in the preformed hGMR complex described by Woodcock et al. (32). Thus, it is surprising that deletion of the C-terminal 14 amino acids of mGMR␣ also abolished the association between mGMR␣ and I374N. Nevertheless, this observation suggests that there may be a degree of interaction between the intracellular domains of ␣and ␤-subunits and that the effect of such an interaction may only be detectable in the context of weaker extracellular interactions between mGMR␣ and I374N as compared with those between wild-type h␤c and hGMR␣.
Most importantly, however, the fact that (i) mGMR␣ associates with the I374N mutant but not with wild-type h␤c and (ii) association of mGMR␣ mutants with I374N correlates with their ability to allow constitutive receptor activity suggests that induction of this association is essential for h␤c activation. However, constitutive association of hGMR␣ with h␤c per se is  Fig. 1A. B, proliferation assay of the UT7 cells depicted in A in the presence and absence of human GM-CSF (2 ng/ml). C, flow cytometric analysis of hGMR␣ expression on the surface of the UT7 cells depicted in A. Cells were stained with the anti-hGMR␣ antibody 8G6 by high sensitivity immunofluorescence. The axes are as in Fig. 1A.