Advertisement

A Membrane Proximal Domain of the Human Interleukin-3 Receptor βc Subunit That Signals DNA Synthesis in NIH 3T3 Cells Specifically Binds a Complex of Src and Janus Family Tyrosine Kinases and Phosphatidylinositol 3-Kinase (∗)

  • Padmini Rao
    Affiliations
    Holland Laboratory for BioMedical Science, American Red Cross, Rockville, Maryland 20855
    Search for articles by this author
  • R.Allan Mufson
    Correspondence
    To whom correspondence and reprint requests should be addressed: Holland Laboratory for BioMedical Science, American Red Cross, Immunology Dept., 15601 Crabbs Branch Way, Rockville, MD 20855 . Tel.: 301-738-0736; Fax: 301-738-0794.
    Affiliations
    Holland Laboratory for BioMedical Science, American Red Cross, Rockville, Maryland 20855
    Search for articles by this author
  • Author Footnotes
    ∗ This work was supported by National Institutes of Health Grant Ca 53609 and Grant 3134R2 from the Council for Tobacco Research. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore by hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Open AccessPublished:March 24, 1995DOI:https://doi.org/10.1074/jbc.270.12.6886
      The high affinity human interleukin-3 receptor is a heterodimeric protein consisting of an α and βc subunit. The βc subunit is responsible for receptor signal transduction. We have shown that a membrane proximal domain of the cytoplasmic tail of the human βc subunit (amino acids 451-517) is minimally required for human IL-3 to signal DNA synthesis in quiescent transfected NIH 3T3 cells. Glutathione S-transferase (GST) fusion proteins of this 451-517 region and another region 451-562 that includes an acidic domain previously shown in other receptors to bind Src family kinases were constructed. Purified Lyn and Lck kinase, but not Fes, could phosphorylate tyrosines in both domains. Adsorption with lysates from the human IL-3-dependent hematopoietic cell line (TF-1) or 3T3 cells and in vitro phosphorylation showed that both these domains were intensely phosphorylated. Phosphoamino acid analysis, however, revealed that the majority of phosphorylation was on serine and threonine rather than tyrosine. Adsorption of these domains with 3T3 or TF-1 cell lysates, followed by immunoblotting, showed that cytoplasmic tyrosine kinases Lyn, Fes, and JAK-2 could also stably associate with both domains; however, Src family kinases are more strongly recognized by both regions than JAK-2 kinase. In addition, phosphatidylinositol 3-kinase from cell lysates was also found stably associated with these domains, but GTPase activating protein, Vav, Sos1, or Grb2 were not.

      INTRODUCTION

      Human interleukin-3 (IL-3)
      The abbreviations used are: IL-3
      interleukin-3
      GM-CSF
      granulocyte macrophage colony stimulating factor
      IL-5
      interleukin-5
      GST
      glutathione S-transferase
      PCR
      polymerase chain reaction
      CAPS
      3-(cyclohexylamino)propanesulfonic acid
      PI 3-kinase
      phosphatidylinositol 3-kinase
      FSBA
      5′-p-fluorosulfonylbenzoyladenosine
      GAP
      GTPase activating protein.
      and granulocyte-macrophage colony stimulating factor (GM-CSF) are hematopoietins that support the survival and differentiation of a large number of myeloid progenitor cells(
      • Metcalf D.
      ). Human interleukin-5 (IL-5) is a more restricted hematopoietin whose target is more mature cells, especially eosinophils and some B-cells(
      • Arai K.I.
      • Lee L.
      • Miyajima A.
      • Miyatake S.
      • Arai N.
      • Yokata T.
      ). The receptors for human IL-3, GM-CSF, and IL-5 comprise a family of heterodimeric proteins that share a common structure, and their ligands can cross-compete for receptor binding(
      • Gesner T.
      • Mufson R.A.
      • Turner K.J.
      • Clark S.C.
      ,
      • Lopez A.
      • Vadas M.A.
      • Woodcock J.M.
      • Milton S.E.
      • Lewis A.
      • Elliott M.J.
      • Gillis D.
      • Ireland R.
      • Olwell E.
      • Park L.S.
      ). The high affinity receptor for each of these growth factors consists of a unique α subunit coupled to a common β subunit (βc)(
      • Miyajima A.
      • Kitamura T.
      • Harada N.
      • Yokata T.
      • Arai K.I.
      ). The α subunits are cytokine-specific and bind their respective ligand with low affinity. The βc subunit does not bind any ligand but allows for the formation of a high affinity receptor and signal transduction(
      • Miyajima A.
      • Kitamura T.
      • Harada N.
      • Yokata T.
      • Arai K.I.
      ). Transfection of the human receptor GM-CSF α and βc subunits into murine hematopoietic cells has reproducibly allowed human GM-CSF to support survival and proliferation of these cells(
      • Miyajima A.
      • Kitamura T.
      • Harada N.
      • Yokata T.
      • Arai K.I.
      ,
      • Hayashida K.
      • Kitamura T.
      • Gorman D.
      • Arai K.
      • Yokata T.
      • Miyajima A.
      ). Data on whether the human GM-CSF receptor transfected into murine fibroblasts allows these nonhematopoietic cells to proliferate in response to human GM-CSF are conflicting(
      • Watanabe S.
      • Mui A.F.
      • Muto A.
      • Chem J.X.
      • Hayashida K.
      • Yakota T.
      • Miyajima A.
      • Arai K.I.
      ,
      • Eder M.
      • Griffin J.D.
      • Ernst T.J.
      ). Thus, it is not yet clear whether nonhematopoietic cells contain all the downstream effector molecules required to allow the βc subunit to signal cell proliferation in such cells.
      The human βc subunit is a 120-kDa glycoprotein with a large cytoplasmic domain(
      • Hayashida K.
      • Kitamura T.
      • Gorman D.
      • Arai K.
      • Yokata T.
      • Miyajima A.
      ). The cytoplasmic domain contains neither consensus sequences for intrinsic tyrosine kinase activity nor SH2/SH3 domains; however, IL-3, GM-CSF, and IL-5 all induce similar patterns of tyrosine phosphorylation(
      • Morla A.O.
      • Schreurs J.
      • Miyajima A.
      • Wang J.Y.
      ,
      • Kanakura Y.
      • Druker B.
      • Cannistra S.A.
      • Furukawa Y.
      • Torimoto Y.
      • Griffin J.D.
      ,
      • Murata Y.
      • Yamaguchi N.
      • Hitashi Y.
      • Tominager A.
      • Takatsu K.
      ). This tyrosine phosphorylation appears to be required for signaling cell proliferation(
      • Miyajima A.
      • Kitamura T.
      • Harada N.
      • Yokata T.
      • Arai K.I.
      ,
      • Murata Y.
      • Yamaguchi N.
      • Hitashi Y.
      • Tominager A.
      • Takatsu K.
      ,
      • Toyo A.
      • Kasuga M.
      • Urabe A.
      • Takaku F.
      ,
      • Pierce J.H.
      • Di Fiore P.P.
      • Aaronson S.A.
      • Potter M.
      • Pumphrey J.
      • Scott A.
      • Ihle J.N.
      ,
      • Cleaveland J.L.
      • Dean M.
      • Rosenberg N.
      • Wang J.Y.
      • Rapp U.R.
      ,
      • Torigoe T.
      • O'Connor R.
      • Santoli D.
      • Reed J.C.
      ,
      • Hanazono Y.
      • Chiba S.
      • Sasaki K.
      • Mano H.
      • Miyajima A.
      • Arai K.
      • Yazaki Y.
      • Hirai H.
      ,
      • Silvennoinen O.
      • Witthuhn B.
      • Zuella F.W.
      • Cleveland J.L.
      • Yi T.
      • Ihle J.N.
      ). The cytoplasmic tail of the βc protein contains 431 amino acids(
      • Lopez A.
      • Vadas M.A.
      • Woodcock J.M.
      • Milton S.E.
      • Lewis A.
      • Elliott M.J.
      • Gillis D.
      • Ireland R.
      • Olwell E.
      • Park L.S.
      ), and it has been suggested that the carboxyl terminus of the protein can be truncated at amino acid 517 (βc517), leaving a membrane proximal 67 amino acids, without loss of signal transduction capability in hematopoietic cells; however, the βc517 signals a much reduced level of tyrosine phosphorylation(
      • Sakamaki K.
      • Miyajima I.
      • Kitamura T.
      • Miyajima A.
      ,
      • Sato N.
      • Sakamaki K.
      • Terada N.
      • Arai K.
      • Miyajima A.
      ). Interestingly, the cell proliferation signal transduced by this membrane proximal βc mutant was blocked by the tyrosine kinase inhibitor herbimycin A(
      • Hanazono Y.
      • Chiba S.
      • Sasaki K.
      • Mano H.
      • Miyajima A.
      • Arai K.
      • Yazaki Y.
      • Hirai H.
      ). The βc517 domain contains a Pro-X-Pro amino acid motif (where X is Asn in βc) that is shared with the p130 subunit of the IL-6 receptor and is important in signaling tyrosine phosphorylation and cell proliferation(
      • Murakomi M.
      • Hibi M.
      • Nakazawa N.
      • Nakazawa T.
      • Yasukawa K.
      • Taga T.
      • Kishimoto T.
      ). Adjacent to amino acid 517 is a region between amino acids 520 and 562 that is rich in acidic amino acids and homologous to an acidic region in the p75 IL-2 receptor subunit that appears to be required for binding p56lck(
      • Hatakeyama M.
      • Kono T.
      • Kobayashi N.
      • Kawalia A.
      • Levin S.D.
      • Perlmutter R.
      • Taniguchi T.
      ). It is thus possible that these regions may contribute to binding cytoplasmic tyrosine kinases and other signal transduction molecules. No data are available on the interactions of these regions with downstream signaling proteins. Thus, although deletion mutagenesis and sequence analysis have helped define the organization of the βc subunit and the presence of motifs capable of associating with signaling proteins, no molecular studies are available that directly investigate the physical interaction of βc subunit cytoplasmic domains with signaling effector proteins from hematopoietic and nonhematopoietic cell lines. In this study, we have investigated whether the intact βc or membrane proximal mutants of this subunit can signal quiescent murine fibroblasts to enter DNA synthesis. Further, we have prepared glutathione S-transferase (GST) fusion proteins containing these βc membrane proximal domains to analyze their interaction with purified effector proteins and effector proteins in the cytoplasm of human hematopoietic (TF-1) and murine fibroblast (NIH 3T3) cells.

      EXPERIMENTAL PROCEDURES

      DNA Construction

      cDNAs for the α and βc subunits of the human IL-3 receptor were cloned from TF-1 cells by reverse transcription and polymerase chain reaction (PCR). DNA sequencing was performed to confirm the fidelity of the amplified DNA products. They were then cloned into mammalian cell expression vectors pREP10 and pcDNA1neo (Invitrogen), respectively, as described previously by Weiss et al.(
      • Weiss M.
      • Yokoyama C.
      • Shikama Y.
      • Wangle C.
      • Druker B.
      • Seiff C.A.
      ). Mutants of the human βc, βc562, βc517, and βc455 were prepared by the polymerase chain reaction method also as described previously(
      • Weiss M.
      • Yokoyama C.
      • Shikama Y.
      • Wangle C.
      • Druker B.
      • Seiff C.A.
      ). Primers were synthesized by phosphoramidite chemistry on an Applied Biosystems Automated DNA synthesizer. All PCR products were sequenced to confirm the fidelity of amplification. The mutant βc subunits were also cloned into the pcDNA1neo expression vector. To construct fusion proteins, DNA fragments encoding the IL-3 receptor βc subunit domains were amplified from βc cDNA by using the following primer pairs in polymerase chain reactions (PCR): βc517, 5′ ATC GAA TTC CCC TCG CTG TCC CCG AA 3′ and 5′ ATC GGA TCC TAT ACG GGT ACA GGC TGC G 3′; βc562, 5′ATC GAA TTC CCC TCA GGT GTC TGG GA 3′ and 5′ ATC CGA TCC TAT ACG GGT ACA GGC TGC G 3′. The primers encode the unique restriction sites underlined (BamHI GGATCC and EcoRI GAATTC) which facilitate subsequent cloning and 2 extra base pairs at the 5′ end for stability and increased efficiency of restriction digestion. First strand cDNA synthesis was performed by previously described methods(
      • Weiss M.
      • Yokoyama C.
      • Shikama Y.
      • Wangle C.
      • Druker B.
      • Seiff C.A.
      ). The PCR was continued for 30 cycles at: 95°C for 1 min, 50°C or 43°C for 2 min (for βc517 and βc562, respectively), and 72°C for 3 min in a buffer containing 2 mM MgCl2. The fidelity of all PCR products was determined by cloning the PCR products into the vector PCR 1000 (Invitrogen), as recommended by the manufacturer, and sequencing the cloned fragment using an Applied Biosystems automated DNA sequencer (Applied Biosystems). All PCR sequences were found to be 100% faithfully amplified.

      Expression of GST-βc517 and GST βc562 and Coupling to Glutathione-Sepharose Beads

      The cloned PCR fragments were restricted with EcoRI and BamHI and ligated into the vector pGEX3X (Pharmacia Biotech Inc.) to produce GST fusion proteins. The ligated pGEX3X vector was used to transform Escherichia coli strain DH5α for fusion protein expression. Transformants were grown, induced for fusion protein expression, and lysed by methods recommended by the vector manufacturer (Pharmacia). Clarified lysates were incubated with a 50:50 (v/v) slurry of glutathione-Sepharose beads (Pharmacia) in equilibration buffer (20 mM HEPES, 100 mM KCl, 0.5 mM EDTA, 1 mM dithiothreitol, pH 7.5) and then washed extensively with phosphate-buffered saline containing 1% Triton X-100. Washed beads were stored in equilibration buffer.

      Transient Transfection of NIH 3T3 Cells

      The human IL-3 receptor α subunit cloned into pREP10 or βc subunit mutants cloned into pcDNA1neo (5 μg of each) were transiently transfected into subconfluent cultures of NIH 3T3 cells using DEAE-dextran essentially as described previously by Kriegler (
      • Kriegler M.
      ) and Watanabe(
      • Watanabe S.
      • Mui A.F.
      • Muto A.
      • Chem J.X.
      • Hayashida K.
      • Yakota T.
      • Miyajima A.
      • Arai K.I.
      ). The method can yield transfection efficiencies up to 80% (
      • Murakomi M.
      • Hibi M.
      • Nakazawa N.
      • Nakazawa T.
      • Yasukawa K.
      • Taga T.
      • Kishimoto T.
      ). The percentage of IL-3 receptor positive cells was determined and used as an index of transfection efficiency. This was done for each experiment, and transfection efficiencies of 60-80% were obtained. We have also found a constant level of IL-3 receptor expression for up to 72 h after transfection.

      Measurement of DNA Synthesis

      Twenty four h after transfection cells were trypsinized, washed in serum-free Dulbecco's modified Eagle's medium, and replated in Dulbecco's modified Eagle's medium containing 0.5% calf serum in 96-well plates (104 cells/well). In some experiments, the cells were cultured in the complete absence of serum. The cells were serum-deprived for 24 h and then stimulated with IL-3 (20 ng/ml) or 3% calf serum for 15 h. [3H]Thymidine (1 μCi/well) was then added, and the cells were incubated for an additional 4 h. Cells were then harvested in a Packard Filtermate 196, and radioactivity was determined by direct β counting in a Packard Matrix 9600 Direct Beta Counter.

      Mammalian Cell Culture and Preparation of Cell Lysates

      TF-1 and NIH 3T3 cells were propagated as described previously(
      • Watanabe S.
      • Mui A.F.
      • Muto A.
      • Chem J.X.
      • Hayashida K.
      • Yakota T.
      • Miyajima A.
      • Arai K.I.
      ). For preparation of lysates, 2 × 107cells were collected by centrifugation and resuspended in 1.0 ml of lysis buffer (0.5% Nonidet P-40, 150 mM NaCl, 10 mM Tris-HCl, pH 7.3, 2.0 mM sodium orthovanadate, 10 mM sodium pyrophosphate, 0.4 mM EDTA, 10 mM NaF, 2 μg/ml each of aprotinin and leupeptin, and 1 mM phenylmethylsulfonyl fluoride) and incubated on ice for 30 min. Following centrifugation, the clarified cell supernatants were used in further adsorption experiments.

      In Vitro Phosphorylation of IL-3 Receptor βc Subunit Domains by Cell Lysate Adsorbates

      Aliquots of the clarified supernatants were incubated overnight at 4°C with 50 μl of packed beads preadsorbed with 25-35 μg of fusion protein or GST alone. The beads were then washed as described previously and resuspended in kinase buffer (10 mM HEPES, pH 7.4, 10 mM MgCl2, 2 mM sodium orthovanadate, and 1 mM phenylmethylsulfonyl fluoride) containing 10 μCi of [ γ-32P]ATP(
      • Weiss M.
      • Yokoyama C.
      • Shikama Y.
      • Wangle C.
      • Druker B.
      • Seiff C.A.
      ). The kinase reaction was incubated for 10 min at 30°C. Samples were then washed with lysis buffer, and proteins were solubilized in 2 × Laemmli's sample buffer and boiled for 5 min. Proteins were resolved by 7.5% SDS-polyacrylamide gel electrophoresis under reducing conditions. After the gels were dried, phosphoproteins were detected by autoradiography using Kodak XAR film at −70°C.

      Factor Xa Cleavage of GST Fusion Proteins

      Fusion proteins coupled to beads were adsorbed with cell lysates and phosphorylated in vitro as described above. The phosphorylated fusion proteins were washed three times with lysis buffer, twice in phosphate-buffered saline containing 1% Triton X-100, once in wash buffer (50 mM Tris-Cl, 150 mM NaCl, pH 7.5), and finally once in cleavage buffer (wash buffer containing 1 mM CaCl2). Factor Xa (New England Biolabs) was used on a 2% (w/w of fusion protein) basis, and digestion was carried out at 25°C on an end over end rotator. Aliquots were drawn at hourly intervals, boiled in sample buffer, and resolved by SDS-polyacrylamide gel electrophoresis under reducing conditions (12.5% polyacrylamide, 8 M urea). Very low molecular weight standards were used to calibrate the protein molecular weights. Proteins were detected in dried gels by autoradiography at −70°C using Kodak XAR film.

      Phosphorylation of βc Cytoplasmic Domains with Purified Src Family Kinases

      Purified Src family kinases were purchased from Upstate Biotechnology (Lake Placid, NY). For phosphorylation by the purified Fes kinase, beads coupled to fusion protein (25-35 μg) or GST were added to a buffer composed of 200 mM HEPES, pH 7.4, 100 mM MgCl2, 100 mM MnCl2, and, for Lyn and Lck, the buffer was 50 mM Tris, pH 7.0, 25 mM MgCl2, 5 mM MnCl2, and 0.05 mM Na3VO4. To each reaction 0.5 mM ATP containing 2 μCi of [ γ-32P]ATP and 30 units of purified tyrosine kinase were added in a total volume of 50 μl. The reaction mixture was incubated for 30 min at 30°C. At the end of this incubation, the reaction mixture was solubilized in 2 × Laemmli's sample buffer. The solubilized protein was loaded on to 7.5% SDS-polyacrylamide gels for electrophoretic separation under reducing conditions. Phosphoproteins were detected by autoradiography at −70°C after drying the gels.

      Association of Tyrosine Kinases and Signal Transduction Proteins in Cell Lysates with IL-3 Receptor βc Subunit Cytoplasmic Domains

      Cytoplasmic domain fusion proteins or GST (25-35 μg) coupled to beads were incubated with cell lysate (2 × 107 cells) overnight at 4°C. After incubation, the beads were extensively washed as described previously(
      • Clark M.
      • Campbell K.S.
      • Kazlauskas A.
      • Johnson S.A.
      • Harty M.A.
      • Potter T.A.
      • Pleiman C.
      • Cambier J.C.
      ), and the washed beads were then boiled for 5 min in 2 × Laemmli's buffer. Solubilized proteins were resolved in 10% SDS-polyacrylamide gels under reducing conditions and transferred to nitrocellulose membranes (Amersham). Immunoblotting was performed using commercially characterized polyclonal and monoclonal antibodies by procedures prescribed by the manufacturers. For all primary polyclonal antibodies, the secondary antibody was sheep anti-rabbit IgG. When a rat monoclonal antibody (anti-v-Fes) was used, a mouse anti-rat antibody was used as a bridging antibody followed by horseradish peroxidase-labeled goat anti-mouse IgG. Immunoreactive bands were revealed by chemiluminescence (ECL System, Amersham Corp.). All antibodies were purchased from Upstate Biotechnology except the anti-v-Fes monoclonal antibody that was purchased from Oncogene Sciences (Uniondale, NY).

      Phosphoamino Acid Analysis of Phosphorylated Cytoplasmic Domains

      The procedure of Cooper et al.(
      • Cooper J.A.
      • Sefton B.M.
      • Hunter T.
      ) was used for phosphoamino acid analysis. Fusion protein adsorbates from cell lysates were phosphorylated by in vitrokinase assays, and the proteins were resolved by SDS-polyacrylamide gel electrophoresis. The separated proteins were transferred from the gel to polyvinylidene difluoride membranes in CAPS buffer (150 mM CAPS, pH 11.0, and 10% methanol), and detection of phosphoproteins was by autoradiography. The portions of the membranes containing the phosphorylated cytoplasmic domains were cut out and digested using constant boiling 6 N HCl (Pierce). Phosphoamino acids were resolved by two-dimensional thin layer electrophoresis using a pH 1.9 buffer (2.5% formic acid, 7.8% acetic acid, 89.7% H2O (v/v)) and a pH 3.5 buffer (15% acetic acid, 0.5% pyridine, 94.5% H2O (v/v/v)) in the first and second dimensions, respectively. Detection of phosphoamino acids was by autoradiography of the thin layer plate. Densitometric analysis of autoradiograms was performed using a flat bed Epson ES 300C scanner and the NIH software package Image 1.57.

      RESULTS

      Transient Transfection of NIH 3T3 Cells with Mutated Human IL-3 Receptors

      Watanabe and co-workers (
      • Watanabe S.
      • Mui A.F.
      • Muto A.
      • Chem J.X.
      • Hayashida K.
      • Yakota T.
      • Miyajima A.
      • Arai K.I.
      ) have reported that transient transfection of NIH 3T3 with both subunits of the human GM-CSF receptor allows human GM-CSF to drive these cells into DNA synthesis. Eder et al.(
      • Eder M.
      • Griffin J.D.
      • Ernst T.J.
      ), however, were unable to observe GM-CSF stimulation of DNA synthesis in NIH 3T3 cells after stable transfection and selection of cell lines stably expressing the receptor. We therefore chose to examine the response of NIH 3T3 cells to human IL-3 after transfection of human IL-3 receptor constructs using a DEAE-dextran transient transfection protocol similar to that used successfully by Watanabe et al.(
      • Watanabe S.
      • Mui A.F.
      • Muto A.
      • Chem J.X.
      • Hayashida K.
      • Yakota T.
      • Miyajima A.
      • Arai K.I.
      ).Table 1 shows that NIH 3T3 cells transiently transfected with the intact IL-3 receptor α and βc subunits and subsequently serum-deprived (0.5% calf serum) will enter DNA synthesis in response to stimulation with human IL-3. In addition, truncation of the βc subunit at either amino acid 562 or 517 still allowed the reconstructed human IL-3 vector to signal DNA synthesis in quiescent cells. Truncation of the βc subunit membrane proximal domain to amino acid 455, however, completely abrogated the response to human IL-3. Thus, the membrane proximal 67 amino acids of the βc subunit are a minimal domain for signaling DNA synthesis. All transfectants showed a similar increase in [3H] thymidine incorporation in response to 3% calf serum. Neither the full length βc517 nor βc562 mutations required the presence of 0.5% serum to signal DNA synthesis in NIH 3T3 cells. These latter three βc constructions could signal DNA synthesis even if the transfected cells were cultured in the complete absence (0%) of serum and the IL-3 was added in serum-free medium (data not shown).
      Tabled 1
      Table thumbnail ft1

      Construction of βc Subunit Membrane Proximal Domain Fusion Proteins and In Vitro Phosphorylation Assay of Domains βc517 and βc562

      Having established that these membrane proximal domains of the IL-3 receptor can signal DNA synthesis, we wished to determine how they might interact with downstream effector molecules. We therefore constructed GST fusion proteins containing membrane proximal domains of the βc subunit (Fig. 1). The domain βc517 contains a motif PNPSKSH that is shared with gp130, the signal-transducing subunit of the IL-6 receptor(
      • Murakomi M.
      • Hibi M.
      • Nakazawa N.
      • Nakazawa T.
      • Yasukawa K.
      • Taga T.
      • Kishimoto T.
      ). Within this motif, the sequence PNP appears to be important for signaling in the IL-6 receptor. The βc562 domain includes the Pro-X-Pro motif, but also contains a region homologous to the p75 subunit of the IL-2 receptor that mediates binding of p56c-1ck to this IL-2 receptor subunit(
      • Hatakeyama M.
      • Kono T.
      • Kobayashi N.
      • Kawalia A.
      • Levin S.D.
      • Perlmutter R.
      • Taniguchi T.
      ). We expressed these domains as fusion proteins with GST and coupled them to glutathione-Sepharose beads. These coupled constructs were then used as probes to identify molecules in cell lysates that can bind to the βc subunit domains. Fusion protein (25-35 μg) coupled to beads was incubated with aliquots of Nonidet P-40 lysates from unstimulated TF-1 cells equal to 2 × 107 cells/ml. The beads were extensively washed and subjected to an in vitro kinase assay with [ γ-32P]ATP. At the end of the assay, beads were further washed to remove free ATP, and the phosphorylated proteins were analyzed by polyacrylamide gel electrophoresis and autoradiography. In Fig. 2A, lane 1, shows that GST protein alone (50 μg) coupled to beads was not phosphorylated; however, GST fusion proteins containing either the domain βc517 (lane 3) or βc562 (lane 5) were intensely phosphorylated in the in vitro kinase assay. In addition to the domain themselves, several kinases or kinase substrates with molecular masses of 46-69 kDa or greater were also bound and phosphorylated. It should be noted that in this experiment unequal amounts of fusion protein were used in the in vitro kinase assay, and this accounts for the difference in phosphorylation intensity between the lanes (see Fig. 2B). Boiling the TF-1 lysate before adsorption with either fusion protein (lanes 2 and 6) reduced the in vitro phosphorylation to undetectable levels demonstrating that the fusion proteins themselves have no intrinsic protein kinase activity and their phosphorylation is due to protein kinase activity. This was verified by addition of 100 μM FSBA, a protein kinase inhibitor to the in vitro kinase assays for βc517 (lane 4) and βc562 (lane 7) adsorbates, respectively. The inhibitor completely blocked phosphorylation of the adsorbed fusion proteins. Thus, the phosphorylation is enzymatically mediated by protein kinases. Fig. 2B shows that about equal intensities of phosphorylation are achieved when equal amounts (35 μg) of either fusion protein are incubated with TF-1 cell lysate. Preadsorption of the lysates with an excess of GST-Sepharose did not result in any qualitative alteration in the precipitation of kinase activity by the fusion proteins. Similar results were obtained when phosphorylation experiments were performed in vitro with NIH 3T3 cell lysates (data not shown). Finally, to verify that the βc subdomains themselves were being phosphorylated, we cleaved both fusion proteins with Factor Xa after adsorption with TF-1 cell lysate and in vitro kinase assay. SDS-polyacrylamide gel electrophoresis revealed that only those proteolytic fragments representing βc cytoplasmic domain were significantly phosphorylated (data not shown). These experiments indicate that the βc517 and βc562 domains can both associate with protein kinases and are also themselves substrates for phosphorylation by the associating kinases.
      Figure thumbnail gr1
      Figure 1:Structure of the IL-3 receptor βc subunit. A, the organization of the IL-3 receptor β subunit is represented in schematic form: EC, extracellular domain; WSXWS, Trp-Ser-variable-Trp-Ser box; TM, transmembrane domain; gp130 homology, region containing Pro-Asp-Pro motif; and Acid rich, region containing 11 acidic and 1 basic amino acid IL-2 receptor homology region. The amino acids (a.a.) within each domain are numbered. B, schematic representation of the βc subunit mutants βc455, βc517, βc562, and wild type (WT) used to transfect NIH 3T3 cells. Amino acid 451 is the first amino acid in the cytoplasmic domain. C, amino acid sequences from the βc subunit that are expressed as fusion proteins. The starred residues identify a motif homologous in the IL-6 receptor gp130, the boxed motif is shared between the human and murine IL-3 receptor βc subunits, and the underlined amino acids are the acidic residues homologous to the acidic rich region in the p75 subunit of the IL-2 receptor.
      Figure thumbnail gr2
      Figure 2:In vitro phosphorylation of βc cytoplasmic domains after adsorption with TF-1 cell lysates. A, cytoplasmic domain fusion proteins or GST alone were coupled to GST-Sepharose beads. The beads were loaded with 20-60 μg of GST or GST fusion protein without equalizing the protein loadings. GST alone was incubated with native TF-1 cell lysate (lane 1), GST βc517 with boiled cell lysate (lane 2) or native TF-1 lysate (lanes 3 and 4), and GST βc562 with native TF-1 lysate (lanes 5 and 7) or boiled TF-1 lysate (lane 6) overnight at 4°C. After washing, the absorbed beads were resuspended in kinase buffer for in vitro kinase assay with [ γ-32P]ATP. In some of the kinase assays the protein kinase inhibitor FSBA was included (lanes 4 and 7). Samples were washed five times after in vitro kinase assay and solubilized in sample buffer for resolution on reducing SDS polyacrylamide gels. The dried gels were autoradiographed at −70°C using Kodak XAR film to visualize phosphorylated proteins. B, equal amounts of each protein (GST alone or GST fusion protein) were coupled to beads, adsorbed with TF-1 cell lysate, and subjected to in vitro kinase assay as described in A.

      Phosphorylation of IL-3 Receptor βc Cytoplasmic Domains by Purified Src Family Kinases

      Since residues 451 and 463 of the IL-3 receptor βc cytoplasmic domain are tyrosine residues, we determined whether these tyrosines are available for phosphorylation by Src family tyrosine kinases. We chose the Src family kinases, Lyn and Fes, reported to be associated with and activated by the IL-3 receptor in hematopoietic cells, and Lck, a related kinase present specifically in lymphocytic cells. Fig. 3 shows that both Lyn and Lck are capable of phosphorylating the tyrosines present in both βc517 and βc562, but Fes was only weakly active or inactive with these fusion protein substrates. None of the kinases phosphorylated GST alone. Thus, some but not all Src family kinases are capable of phosphorylating these domains. Finally, it is important to note in Fig. 3 that βc562 was reproducibly a better substrate for Src kinases than βc517.
      Figure thumbnail gr3
      Figure 3:Phosphorylation of βc cytoplasmic domains by purified Src family kinases. Cytoplasmic domain fusion proteins or GST alone (25-35 μg) were coupled to GST-Sepharose beads and suspended in the appropriate reaction buffer with 30 units of either purified p56lyn (A), p56lck (B), or p93fes (C), and [ γ-32P]ATP. The reaction mixture was incubated for 30 min at 30°C and then solubilized in reducing sample buffer. Proteins were resolved by SDS-polyacrylamide gel electrophoresis. Phosphoproteins were detected by autoradiography at −70°C using Kodak XAR film.

      Phosphoamino Acid Analysis of βc Cytoplasmic Domain after TF-1 Lysate Adsorption and in Vitro Phosphorylation

      Although it appears that purified Src family kinases can phosphorylate these βc cytoplasmic domains, previous investigations have shown that the major sites of βc tyrosine phosphorylation is not in this region(
      • Sakamaki K.
      • Miyajima I.
      • Kitamura T.
      • Miyajima A.
      ). Therefore, we further investigated the nature of the residues phosphorylated after adsorption of these cytoplasmic domains with TF-1 lysates and in vitro kinase assay. Phosphoamino acid analysis was performed on acid hydrolysates of gel-purified phosphorylated cytoplasmic domains using two-dimensional thin layer electrophoresis. Fig. 4 compares the phosphoamino acid patterns obtained from domains βc517 and βc562. In βc562, about 77% of the phosphorylation is on threonine, 13% is on serine, and 2% on tyrosine. In βc517, about 85% of the phosphorylation is on serine, 15% on threonine, and phosphotyrosine was undetectable. Thus, serine/threonine kinase activity accounts for a major proportion of the total phosphorylation of these domains, although the region between 517 and 562 appeared to enhance the phosphorylation of tyrosine residues 451 and 453. The increased availability of 5 more threonines in the 520-562 region may account for the shift to threonine phosphorylation in βc562 compared to βc517, or the additional residues could alter peptide conformation and contribute to the longer domain being a better substrate. The phosphorylation of βc517 by the purified Src kinases but not by TF-1-adsorbed tyrosine kinases is probably due to the greater amounts of tyrosine kinase enzyme activity added to the in vitro assays compared to the low levels of tyrosine kinase actually available in cell lysates.
      Figure thumbnail gr4
      Figure 4:Phosphoamino acid analysis of βc cytoplasmic domains phosphorylated by TF-1 cell lysates. Cytoplasmic domains of the βc subunit were subjected to in vitro kinase assays after adsorption with TF-1 cell lysates and resolved by gel electrophoresis as described in the legend to . The separated proteins were transferred to polyvinylidene difluoride membranes and detected by autoradiography. Portions of the membrane containing phosphorylated cytoplasmic domain fusion proteins were cut out and hydrolyzed using constant boiling HCl. Phosphoamino acids were resolved by two-dimensional thin layer electrophoresis and detected by autoradiography of the thin layer plates.

      IL-3 Receptor Cytoplasmic βc Domains βc517 and βc562 Associate with Src Family and Janus Family Kinases in TF-1 and 3T3 Cell Lysates

      Although the βc517 domain was not phosphorylated on tyrosine after adsorption with TF-1 cell lysates, this domain could still be capable of physically associating with cytoplasmic tyrosine kinases important in signal transduction. To characterize complexes of cytoplasmic tyrosine kinases with our βc subunit cytoplasmic domains, immunoblotting of NIH 3T3 and TF-1 adsorbates was performed. We chose antibodies against kinases that had been shown to be activated by IL-3 or associated with the wild type β subunit of the IL-3 receptor. p56lyn is a hematopoietic cell-specific kinase that had previously been shown to be activated by engagement of the IL-3 receptor(
      • Torigoe T.
      • O'Connor R.
      • Santoli D.
      • Reed J.C.
      ). Immunoblotting of proteins associated with βc517 and βc562 after incubation with TF-1 lysates using rabbit anti-human Lyn polyclonal antibody revealed the presence of Lyn immunoreactivity at the 56-kDa molecular mass (Fig. 5A). No Lyn immunoreactivity was associated with GST alone. If the TF-1 lysates were immunodepleted of Lyn protein by pre-immunoprecipitation with anti-Lyn antibodies, prior to incubation with cytoplasmic domain fusion proteins, then no Lyn immunoreactivity could be detected associated with the cytoplasmic domains (Fig. 5B). Immunoprecipitation of TF-1 lysates with preimmune serum did not alter Lyn association with the fusion protein (data not shown). Thus, we have confirmed that p56c-lyn is associated with the β subunit minimal signaling domain. Other studies have shown that GM-CSF and IL-3 induce phosphorylation and activation of p93fes, another hematopoietic cell-specific non-receptor tyrosine kinase(
      • Hanazono Y.
      • Chiba S.
      • Sasaki K.
      • Mano H.
      • Miyajima A.
      • Arai K.
      • Yazaki Y.
      • Hirai H.
      ). We have thus performed similar experiments to determine whether p93fes associates with the β subunit cytoplasmic domains using a rat monoclonal antibody against v-Fes which recognizes all mammalian c-Fes antigens. An immunoreactive band at the 93-kDa position characteristic of c-Fes was physically associated with both cytoplasmic domains of the βc subunit, but not with GST alone (Fig. 5C). Immunodepletion of the TF-1 lysates with anti-Fes antibody prior to adsorption with fusion proteins also abrogated the detection of Fes immunoreactivity confirming the identity of the Fes band associated with the fusion proteins. Finally, it has recently been suggested that the Janus family kinase JAK-2 associates with the β subunit and is crucial to hematopoietic receptor signaling(
      • Silvennoinen O.
      • Witthuhn B.
      • Zuella F.W.
      • Cleveland J.L.
      • Yi T.
      • Ihle J.N.
      ). Therefore, we also probed for the association of JAK-2 with these cytoplasmic domains using rabbit anti-mouse JAK2 polyclonal antibody that recognizes the human protein. Again, 25-35 μg of fusion protein was incubated with aliquots of cell lysate equivalent to 2 × 107 TF-1 cells/50 ml, and the adsorbed proteins were probed for the presence of p130 JAK-2 with JAK-2 polyclonal antibody. We could reproducibly detect JAK-2 at the appropriate molecular weight associated with both the βc517 region and the βc562 region, but not with GST alone in lysates of TF-1 or 3T3 cells (Fig. 5D). The immunological signal with JAK-2, however, was always significantly weaker than with the Src family kinase antibody probes. This is probably not due to a paucity of JAK-2 in our TF-1 lysates, because immunoblots of 5 μl of lysate from 2 × 107 TF-1 cells/ml yielded a strong signal with the same antibody (Fig. 5D). Again, pre-immunodepletion of the 3T3 or TF-1 lysates with anti JAK-2 antibody but not preimmune serum abrogated the detection of JAK-2 activity after adsorption with fusion protein.
      Figure thumbnail gr5
      Figure 5:Stable complexes of Src and Janus family kinases from TF-1 or 3T3 cell lysates with βc cytoplasmic domains. Cytoplasmic domain fusion proteins or GST alone (25-35 μg) coupled to beads were incubated overnight at 4°C, with TF-1 cell lysates. After incubation, the beads were extensively washed and solubilized in reducing sample buffer. Solubilized proteins were resolved in 10% SDS-polyacrylamide gels. Resolved proteins were transferred to nitrocellulose, and immunoblotting was performed with rabbit polyclonal anti-p56lynantibody (A), rat monoclonal anti-93fes antibody (C), and rabbit polyclonal anti-p130 JAK-2 antibodies (D). In B, the lane labeled +- is a non-preimmunoprecipitated control while lane ++ is preimmunoprecipitated with anti-Lyn antibody before gel electrophoresis and immunoblotting. Immunoreactive bands were detected by chemiluminescence.

      IL-3 Receptor βc Cytoplasmic Domains Bind Phosphatidylinositol 3-Kinase (PI 3-Kinase) but Not GTPase Activating Protein (GAP), Vav, Sos1, or Grb2

      Engagement of the IL-3 receptor results in the tyrosine phosphorylation of a number of proteins involved in signal transduction. Among these proteins are p21ras-GAP(
      • Satoh T.
      • Nakafuku M.
      • Miyajima A.
      • Kaziro Y.
      ,
      • Satoh T.
      • Uehara Y.
      • Kaziro Y.
      ,
      • Duronio V.
      • Welham M.J.
      • Abraham S.
      • Dryden P.
      • Schrader J.W.
      ), PI 3-kinase(
      • Corey S.
      • Equinoa A.
      • Puyana-Theall K.
      • Bolen J.B.
      • Cantley L.
      • Mollinedo F.
      • Jackson T.R.
      • Hawkins P.T.
      • Stephens L.R.
      ), and Vav and Shc(
      • Cutler R.L.
      • Liu L.
      • Damon J.E.
      • Krystal G.
      ). In addition, tyrosine-phosphorylated Shc has been shown to associate with the docking protein Grb2 and the guanine nucleotide exchange protein Sos1(
      • Corey S.
      • Equinoa A.
      • Puyana-Theall K.
      • Bolen J.B.
      • Cantley L.
      • Mollinedo F.
      • Jackson T.R.
      • Hawkins P.T.
      • Stephens L.R.
      ,
      • Cutler R.L.
      • Liu L.
      • Damon J.E.
      • Krystal G.
      ). We therefore determined whether either the β-subunit cytoplasmic domain βc517 or βc562 participates in the formation of such signaling complexes. To examine this possibility, fusion protein adsorbates obtained after incubation of cytoplasmic domains βc517 and βc562 with TF-1 lysates were analyzed by polyacrylamide gel electrophoresis and immunoblotting. Rabbit polyclonal antibodies against the p85 subunit of PI 3-kinase that recognize the human protein detected an 85-kDa protein adsorbed to the βc517 and βc562 fusion proteins (Fig. 6A). No protein immunoreactivity was detected in adsorbates from either beads coupled to GST alone or fusion protein adsorbed with TF-1 cell lysate and probed with preimmune serum. In contrast, adsorbates from βc517 or βc562 showed no immunoreactivity when a rabbit anti-GAP polyclonal antibody that can recognize human and mouse antigens was used as a probe (Fig. 6B). Both the anti-PI 3-kinase antibody and the anti-GAP antibody readily detected their respective antigens in total cell lysate from TF-1 cells. We also probed both cytoplasmic domain adsorbates for immunoreactivity associated with the signal transducing proteins Vav, Grb2, or Sos1 but did not detect the presence of these molecules in adsorbates from either cytoplasmic domain using commercially available antibodies (data not shown).
      Figure thumbnail gr6
      Figure 6:Stable complex of PI 3-kinase from TF-1 or 3T3 cell lysates with βc cytoplasmic domains. Adsorption of fusion protein coupled beads with TF-1 or 3T3 cell lysates, and immunoblot analysis was performed as in the legend to . A, immunoblotting was performed with antibody against the p85 subunit of PI 3-kinase. B, immunoblotting was performed with rabbit polyclonal antibody against human GAP residues 171-448 that also recognizes mouse GAP. Immunoreactive bands were detected by chemiluminescence.

      DISCUSSION

      Whether the human IL-3 receptor βc requires lineagespecific downstream effector molecules or can signal in a nonhematopoietic fibroblast background has been controversial. Watanabe et al.(
      • Watanabe S.
      • Mui A.F.
      • Muto A.
      • Chem J.X.
      • Hayashida K.
      • Yakota T.
      • Miyajima A.
      • Arai K.I.
      ) have reported that the human GM-CSF receptor reconstituted in NIH 3T3 cells induced tyrosine phosphorylation, immediate early response gene induction, and DNA synthesis in response to human GM-CSF. Eder and co-workers (
      • Eder M.
      • Griffin J.D.
      • Ernst T.J.
      ) also observed tyrosine phosphorylation and immediate early response gene induction in response to human GM-CSF in NIH 3T3 cells transfected with the human GM-CSF receptor complex, but were unable to measure any increases in DNA synthesis. Thus, it has been unclear if this hematopoietin receptor system could signal cell proliferation ubiquitously in nonhematopoietic cells. Previous work from our laboratory has shown that NIH 3T3 cells transiently transfected with the human IL-3 receptor respond to human IL-3 by increases in tyrosine phosphorylation and phosphatidylcholine-specific phospholipase C activity, as well as translocation of protein kinase C(
      • Rao P.
      • Kitamura T.
      • Miyajima A.
      • Mufson R.A.
      ). The experiments with transient transfection reported here extend these findings and confirm that the human IL-3 receptor indeed can induce DNA synthesis in NIH 3T3 cells in response to human IL-3. It is thus possible that the inability of GM-CSF to signal cell proliferation in the cell lines selected by Eder and co-workers (
      • Eder M.
      • Griffin J.D.
      • Ernst T.J.
      ) for stable expression of the GM-CSF receptor may represent a loss of cellular function acquired during the selection for stable receptor expression. The transient transfection experiments that we have performed show that both the βc562 and βc517 membrane proximal domain truncations could also signal cell proliferation in response to human IL-3 after transfection into NIH 3T3 cells. Thus, these truncations function in a nonhematopoietic environment as they do in murine hematopoietic BaF3 cells. In addition, our experiments also showed that IL-3 could signal DNA synthesis through the wild type βc or βc517 even in the complete absence of serum. Thus, the effector molecules activated by the IL-3 receptor alone are sufficient to complete the signal pathway required to move the cells from G1 through to S phase.
      Using GST fusion proteins containing the membrane proximal regions of the cytoplasmic tail of the βc subunit, we showed that both the βc517 and βc562 were intensely phosphorylated in kinase assays after adsorption with hematopoietic (TF-1) and fibroblast (NIH 3T3) cell lysates. The cytoplasmic domains were also phosphorylated by the Lyn and Lck tyrosine kinases. Interestingly, βc562 was reproducibly a better substrate for tyrosine phosphorylation by the purified Src family kinases. This is consistent with our phosphoamino acid analysis of cytoplasmic domain phosphorylation which revealed that tyrosines 451 and 453 were poorly if at all phosphorylated by kinases after adsorption with hematopoietic cell lysates and in vitro phosphorylation. Addition of the 518-562 region enhanced the phosphorylation of the two tyrosines in the 451-517 region to 2% of the total phosphoamino acid content. The fact that the βc517 domain could be phosphorylated by purified kinases is probably due to the far greater amounts of enzyme activity units added to the in vitro kinase assay compared to those present in the cell lysates. Thus, although the major site for tyrosine phosphorylation on the βc subunit is between Ser683 and Ser763(
      • Hanazono Y.
      • Chiba S.
      • Sasaki K.
      • Mano H.
      • Miyajima A.
      • Arai K.
      • Yazaki Y.
      • Hirai H.
      ), the tyrosines at positions 451 and 453 may play a role in signaling after the binding of tyrosine kinases in the full length βc subunit.
      The phosphoamino acid analysis also revealed that both cytoplasmic domains can also bind serine/threonine kinases. In fact, the major proportion of total phosphorylation observed after adsorption with hematopoietic cell lysates was serine/threonine phosphorylation. Such phosphorylation of the βc subunit has not been well investigated in in vivo studies, and it is not clear which kinase(s) contribute to this phosphorylation. Similar results have been reported, however, for the B-cell antigen receptor(
      • Clark M.
      • Campbell K.S.
      • Kazlauskas A.
      • Johnson S.A.
      • Harty M.A.
      • Potter T.A.
      • Pleiman C.
      • Cambier J.C.
      ). GST fusion proteins containing the antigen receptor homology I (ARH1) motif bind both Src family and serine/threonine kinases, and the majority of phosphorylation of the ARH1 domains is on serine and threonine after cell lysate adsorption and in vitro phosphorylation. In IgM-associated Igα and Igβ chains of the B-cell antigen receptor there is also constitutive in vivo phosphorylation of serine and threonine residues(
      • Clark M.
      • Campbell K.S.
      • Kazlauskas A.
      • Johnson S.A.
      • Harty M.A.
      • Potter T.A.
      • Pleiman C.
      • Cambier J.C.
      ). Further, the gp130 subunit of the IL-6 receptor is constitutively phosphorylated on serine and threonine residues, and activation of the IL-6 receptor increases the amount of serine/threonine phosphorylation of this protein(
      • Murakomi M.
      • Hibi M.
      • Nakazawa N.
      • Nakazawa T.
      • Yasukawa K.
      • Taga T.
      • Kishimoto T.
      ). Thus, both the IL-3 and IL-6 receptor signal transducing subunits can associate with serine/threonine kinases, although the role of serine/threonine phosphorylation in signaling through the βc subunit is unclear.
      Further investigation of the ability of these domains to associate with cytoplasmic signaling molecules revealed that at least two Src family kinases, p93fes and p56lyn, can form stable associations with βc517 and βc562 in cytosolic preparations from TF-1 cells. The presence of Src family kinases in a complex with this region is consistent with recent data showing that the Src family kinases p53/p56lyn, p62yes, and p93fes are activated in vivo during biological signaling through the IL-3/GM-CSF receptor βc subunit (
      • Cleaveland J.L.
      • Dean M.
      • Rosenberg N.
      • Wang J.Y.
      • Rapp U.R.
      ,
      • Torigoe T.
      • O'Connor R.
      • Santoli D.
      • Reed J.C.
      ,
      • O'Connor R.
      • Torigoe T.
      • Reed J.C.
      • Santoli D.
      ). Further evidence that Src family kinases are critical to coupling IL-3/GM-CSF receptor βc subunits to biological responses comes from the work of Linnekin et al.(
      • Linnekin D.W.
      • Howard O.M.
      • Park L.
      • Farrar D.
      • Longo D.W.
      ) examining the effects of the Src family kinase HCK in GM-CSF signaling in HL-60 cells. HL-60 cells express high affinity receptors for GM-CSF, but GM-CSF cannot signal cell proliferation in these cells unless the hck gene is induced with dimethyl sulfoxide or overexpressed after transfection. Thus, our finding of Src family kinases complexed with the membrane proximal domain accounts for the activation of these kinases during biological signaling observed by other laboratories.
      It has also been shown recently that the Janus family kinase JAK-2 associates with the IL-3 receptor βc subunit, and it has been suggested that JAK-2 is central to cytokine receptor signaling (
      • Argetsinger L.S.
      • Campbell G.S.
      • Yang X.
      • Witthuhn B.A.
      • Silvennoinen O.
      • Ihle J.
      • Carter-Su C.
      ,
      • Witthuhn B.A.
      • Quelle F.W.
      • Silvennoinen O.
      • Yi T.
      • Tang B.
      • Miura O.
      • Ihle J.
      ). We could reproducibly detect JAK-2 binding to both IL-3 receptor βc cytoplasmic domains. The association, however, appeared to be significantly weaker than the association with Src family kinases. It is possible that a stronger association with JAK-2 may occur at other sites in the IL-3 receptor subunit that provide better structural recognition than the domains examined here. This suggestion is completely consistent with the data of Quelle and co-workers (
      • Quelle F.W.
      • Sato N.
      • Witthuhn B.A.
      • Inhorn R.C.
      • Eder M.
      • Miyajima A.
      • Griffin J.D.
      • Ihle J.N.
      ) who have shown that JAK-2 is strongly activated by mutants of the IL-3 receptor βc subunit truncated at amino acids 763 or 626 in transfected BaF3 cells; however, the βc subunit truncated at amino acid 517 merely produced a weak activation of JAK-2. The weak association of JAK-2 with the βc517 region in vitro thus correlates well with the weak activation of JAK-2 observed with the βc517 mutant in vivo. Although Silvennoinen et al.(
      • Silvennoinen O.
      • Witthuhn B.
      • Zuella F.W.
      • Cleveland J.L.
      • Yi T.
      • Ihle J.N.
      ) detected only JAK-2 associated with the IL-3 receptor βc subunit, our work and that of several other groups would support the association of multiple tyrosine kinases from both the Src and Janus family with the βc subunit.
      Recent studies of the initial events in GM-CSF receptor signal transduction have shown that GM-CSF receptor occupancy additionally results in the formation of a complex between p53/p56lyn/p62yes and an 85-kDa protein (immunologically related to the 85-kDa subunit of PI 3-kinase) with an accompanying increase in PI 3-kinase activity. The tyrosine kinase•PI 3-kinase complex can be immunoprecipitated with antiphosphotyrosine antibodies(
      • Corey S.
      • Equinoa A.
      • Puyana-Theall K.
      • Bolen J.B.
      • Cantley L.
      • Mollinedo F.
      • Jackson T.R.
      • Hawkins P.T.
      • Stephens L.R.
      ). Our experiments directly showed that the minimal signaling domain βc517 indeed forms a stable complex with Src family kinases and PI 3-kinase. We could find no immunological evidence for the presence of GAP, Sos1, Vav, or Grb2 complexed with these domains. Thus, the interaction of the minimal signaling domain with tyrosine kinases and PI 3-kinase appears specific. The βc517 domain does not contain any PI 3-kinase consensus binding sequence. It is, therefore, likely that PI 3-kinase does not bind directly to the peptide, but to specific domains on the tyrosine kinases that have associated directly with the βc cytoplasmic domains. Interactions of effector molecules with such domains on Lyn, Fyn, and Blk have recently been directly demonstrated in in vitro fusion protein binding studies(
      • Pleiman C.M.
      • Clark M.
      • Gaven L.
      • Winity S.
      • Coggeshall M.
      • Johnson G.
      • Shaw A.
      • Cambier J.
      ). PI 3-kinase appears to bind specifically to p56lyn or p60fyn, and this binding appears to involve amino-terminal SH3 domains on these protein kinase molecules(
      • Coughlin S.R.
      • Lee W.M.F.
      • Williams P.W.
      • Giels G.M.
      • Williams L.T.
      ). Our data also correlate well with recent studies in transfected BaF3 cells showing that PI 3-kinase is activated after stimulation of GM-CSF receptors containing a βc subunit truncated at amino acid 517, and that this kinase may play a role in signaling cell proliferation(
      • Sakamaki K.
      • Miyajima I.
      • Kitamura T.
      • Miyajima A.
      ). In fact, activation of PI 3-kinase and tyrosine phosphorylation are the only measurable biochemical response to activation of this βc517 mutant in transfected BaF3 cells(
      • Sato N.
      • Sakamaki K.
      • Terada N.
      • Arai K.
      • Miyajima A.
      ). Our data would suggest that the activation of PI 3-kinase occurs in a direct physical association with the β subunit domain. The association of this region βc517 with PI 3-kinase may in fact be quite important to signaling cell proliferation because mutants of the platelet-derived growth factor receptor(
      • Coughlin S.R.
      • Lee W.M.F.
      • Williams P.W.
      • Giels G.M.
      • Williams L.T.
      ), CSF-1 receptor(
      • Reedjik M.
      • Liu X.
      • Dawson T.
      ,
      • Shurtleff S.A.
      • Downing J.R.
      • Rock C.O.
      • Hawkins S.A.
      • Roussel M.F.
      • Sherr C.J.
      ), and insulin receptor (
      • Kapeller R.
      • Chen K.S.
      • Yoakim M.
      • Shaffhausen B.S.
      ) that have reduced association with PI 3-kinase are defective in their ability to induce mitogenesis.

      Acknowledgments

      We thank Dr. Steven Ullrich for his assistance with the phosphoamino acid analysis and Tammy Krogman for technical assistance.

      REFERENCES

        • Metcalf D.
        Blood. 1986; 67: 257-267
        • Arai K.I.
        • Lee L.
        • Miyajima A.
        • Miyatake S.
        • Arai N.
        • Yokata T.
        Annu. Rev. Biochem. 1990; 56: 783-836
        • Gesner T.
        • Mufson R.A.
        • Turner K.J.
        • Clark S.C.
        Blood. 1989; 74: 2652-2656
        • Lopez A.
        • Vadas M.A.
        • Woodcock J.M.
        • Milton S.E.
        • Lewis A.
        • Elliott M.J.
        • Gillis D.
        • Ireland R.
        • Olwell E.
        • Park L.S.
        J. Biol. Chem. 1991; 266: 24741-24747
        • Miyajima A.
        • Kitamura T.
        • Harada N.
        • Yokata T.
        • Arai K.I.
        Annu. Rev. Immunol. 1992; 10: 295-331
        • Hayashida K.
        • Kitamura T.
        • Gorman D.
        • Arai K.
        • Yokata T.
        • Miyajima A.
        Proc. Nat. Acad. Sci. U. S. A. 1989; 87: 9655-9659
        • Watanabe S.
        • Mui A.F.
        • Muto A.
        • Chem J.X.
        • Hayashida K.
        • Yakota T.
        • Miyajima A.
        • Arai K.I.
        Mol. Cell Biol. 1993; 13: 1440-1448
        • Eder M.
        • Griffin J.D.
        • Ernst T.J.
        EMBO. J. 1993; 12: 1647-1656
        • Morla A.O.
        • Schreurs J.
        • Miyajima A.
        • Wang J.Y.
        Mol. Cell. Biol. 1988; 8: 2214-2218
        • Kanakura Y.
        • Druker B.
        • Cannistra S.A.
        • Furukawa Y.
        • Torimoto Y.
        • Griffin J.D.
        Blood. 1990; 76: 706-715
        • Murata Y.
        • Yamaguchi N.
        • Hitashi Y.
        • Tominager A.
        • Takatsu K.
        Biochem. Biophys. Res. Commun. 1990; 173: 1102-1108
        • Toyo A.
        • Kasuga M.
        • Urabe A.
        • Takaku F.
        Exp. Cell Res. 1987; 171: 16-23
        • Pierce J.H.
        • Di Fiore P.P.
        • Aaronson S.A.
        • Potter M.
        • Pumphrey J.
        • Scott A.
        • Ihle J.N.
        Cell. 1985; 41: 685-693
        • Cleaveland J.L.
        • Dean M.
        • Rosenberg N.
        • Wang J.Y.
        • Rapp U.R.
        Mol. Cell. Biol. 1989; 9: 5685-5695
        • Torigoe T.
        • O'Connor R.
        • Santoli D.
        • Reed J.C.
        Blood. 1992; 80: 617-624
        • Hanazono Y.
        • Chiba S.
        • Sasaki K.
        • Mano H.
        • Miyajima A.
        • Arai K.
        • Yazaki Y.
        • Hirai H.
        EMBO. J. 1993; 12: 1641-1646
        • Silvennoinen O.
        • Witthuhn B.
        • Zuella F.W.
        • Cleveland J.L.
        • Yi T.
        • Ihle J.N.
        Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 8429-8433
        • Sakamaki K.
        • Miyajima I.
        • Kitamura T.
        • Miyajima A.
        EMBO. J. 1992; 11: 3541-3549
        • Sato N.
        • Sakamaki K.
        • Terada N.
        • Arai K.
        • Miyajima A.
        EMBO J. 1993; 12: 4181-4189
        • Murakomi M.
        • Hibi M.
        • Nakazawa N.
        • Nakazawa T.
        • Yasukawa K.
        • Taga T.
        • Kishimoto T.
        Science. 1993; 260: 1808-1810
        • Hatakeyama M.
        • Kono T.
        • Kobayashi N.
        • Kawalia A.
        • Levin S.D.
        • Perlmutter R.
        • Taniguchi T.
        Science. 1991; 252: 1523-1526
        • Weiss M.
        • Yokoyama C.
        • Shikama Y.
        • Wangle C.
        • Druker B.
        • Seiff C.A.
        Blood. 1993; 82: 3298-3306
        • Kriegler M.
        Gene Transfer Expression. Stockton Press, New York1990: 85-102
        • Clark M.
        • Campbell K.S.
        • Kazlauskas A.
        • Johnson S.A.
        • Harty M.A.
        • Potter T.A.
        • Pleiman C.
        • Cambier J.C.
        Science. 1992; 258: 123-126
        • Cooper J.A.
        • Sefton B.M.
        • Hunter T.
        Methods Enzymol. 1983; 99: 387-392
        • Satoh T.
        • Nakafuku M.
        • Miyajima A.
        • Kaziro Y.
        Proc. Natl. Acad. Sci. U. S. A. 1992; 87: 5993-5997
        • Satoh T.
        • Uehara Y.
        • Kaziro Y.
        J. Biol. Chem. 1992; 267: 2537-2541
        • Duronio V.
        • Welham M.J.
        • Abraham S.
        • Dryden P.
        • Schrader J.W.
        Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 1587-1591
        • Corey S.
        • Equinoa A.
        • Puyana-Theall K.
        • Bolen J.B.
        • Cantley L.
        • Mollinedo F.
        • Jackson T.R.
        • Hawkins P.T.
        • Stephens L.R.
        EMBO J. 1993; 12: 2681
        • Cutler R.L.
        • Liu L.
        • Damon J.E.
        • Krystal G.
        J. Biol. Chem. 1993; 268: 21463-21465
        • Rao P.
        • Kitamura T.
        • Miyajima A.
        • Mufson R.A.
        J. Immunol. 1995; 154: 1664-1674
        • O'Connor R.
        • Torigoe T.
        • Reed J.C.
        • Santoli D.
        Blood. 1991; 80: 1017-1025
        • Linnekin D.W.
        • Howard O.M.
        • Park L.
        • Farrar D.
        • Longo D.W.
        Blood. 1994; 84: 94-103
        • Argetsinger L.S.
        • Campbell G.S.
        • Yang X.
        • Witthuhn B.A.
        • Silvennoinen O.
        • Ihle J.
        • Carter-Su C.
        Cell. 1993; 74: 237-244
        • Witthuhn B.A.
        • Quelle F.W.
        • Silvennoinen O.
        • Yi T.
        • Tang B.
        • Miura O.
        • Ihle J.
        Cell. 1993; 74: 227-236
        • Quelle F.W.
        • Sato N.
        • Witthuhn B.A.
        • Inhorn R.C.
        • Eder M.
        • Miyajima A.
        • Griffin J.D.
        • Ihle J.N.
        Mol. Cell. Biol. 1994; : 4335-4341
        • Pleiman C.M.
        • Clark M.
        • Gaven L.
        • Winity S.
        • Coggeshall M.
        • Johnson G.
        • Shaw A.
        • Cambier J.
        Mol. Cell. Biol. 1993; 13: 5877-5887
        • Coughlin S.R.
        • Lee W.M.F.
        • Williams P.W.
        • Giels G.M.
        • Williams L.T.
        Science. 1989; 243: 1191-1194
        • Reedjik M.
        • Liu X.
        • Dawson T.
        Mol. Cell. Biol. 1990; 10: 5601-5608
        • Shurtleff S.A.
        • Downing J.R.
        • Rock C.O.
        • Hawkins S.A.
        • Roussel M.F.
        • Sherr C.J.
        EMBO J. 1990; 9: 2415-2421
        • Kapeller R.
        • Chen K.S.
        • Yoakim M.
        • Shaffhausen B.S.
        Mol. Endocrinol. 1991; 5: 769-777