Identification of a 14-3-3 Binding Sequence in the Common β Chain of the Granulocyte-Macrophage Colony-Stimulating Factor (GM-CSF), Interleukin-3 (IL-3), and IL-5 Receptors That Is Serine-Phosphorylated by GM-CSF

HUMAN granulocyte-macrophage colony-stimulating factor (GM-CSF), interleukin-3 (IL-3), and IL-5 are pleiotropic cytokines capable of stimulating normal and transformed hematopoietic cells.1-4 With each, the initiating event for signal transduction is the binding of GM-CSF, IL-3, and IL-5 to their surface receptors that are composed of a cytokine-specific α chain and a common β chain (β_c).5-7 Engagement of β_c by the binding of GM-CSF, IL-3, and IL-5 to surface receptors results in the stimulation of cell survival, proliferation, and differentiation and mature cell effector function in the appropriate lineage, a fact that emphasises the major signaling role played by β_c in mediating GM-CSF, IL-3, and IL-5 biological activities.8 

One of the first events in GM-CSF, IL-3, and IL-5 activation of their receptors and in the initiation of the signaling cascade is tyrosine phosphorylation of β_c.9 This is a common theme among cytokine receptor signaling subunits and can be seen in homodimeric receptors such as the erythropoietin (EPO) receptor, thrombopoietin (TPO) receptor, and granulocyte colony-stimulating factor (G-CSF) receptor as well as in heterodimeric receptors such as in the IL-6 and IL-2 receptors.10-13 In the latter case, as in the GM-CSF, IL-3, and IL-5 receptor system, tyrosine phosphorylation is also restricted to its signaling subunit.14 

Tyrosine phosphorylation of cytokine receptor signaling subunits appears to be a critical step in the creation of docking sites for the association of signaling molecules. For example, ligand-induced tyrosine phosphorylation of β_c is considered critical for binding of SH2-containing signaling molecules such as STATs, SHP1, SHP2, and SHC. In the case of STAT5, this is recruited to phosphotyrosines in β_c and is then phosphorylated by JAK2, resulting in STAT activation, dimerization, and translocation to the nucleus, where it directly regulates gene transcription.15,16 Previously, it has been shown that 6 of the 8 tyrosine residues of β_c allow the docking of STAT5 and its subsequent activation, indicating a high degree of redundancy as to which tyrosine residues are critical for STAT5 activation.17 The specific phospho-tyrosine residue responsible for recruitment of SHC to β_c is⁵⁷⁷Tyr. The bound SHC recruits grb2 into the complex and grb2 binds sos, which then activates p21^ras nucleotide exchange. This leads to the activation of ras and downstream partners of the mitogen-activated protein (MAP) Kinase pathway, including Raf.17-20 SHP2, a prominent tyrosine phospho-protein, is able to bind 3 tyrosine residues of the activated β_c (⁵⁷⁷Tyr, ⁶¹²Tyr, and⁶⁹⁵Tyr), although only ⁵⁷⁷Tyr and⁶¹²Tyr allow SHP2 to interact with grb2.17,21 

Despite the well-documented importance of tyrosine phosphorylation of cytokine receptors, it is becoming apparent that signaling can proceed in its absence, suggesting that other mechanisms may be at work. This is demonstrated in the EPO and TPO receptors, in which the substitution of all tyrosines failed to abolish their activities.22,23In the case of β_c, mutation of all 8 cytoplasmic tyrosine residues does not abolish activation of JAK2 and induction of tyrosine phosphorylation of STATs and many other cellular proteins. The only significant defect in this mutation is the reduction in phosphorylation of SHP2 and SHC.17 In the case of mutant mice that have STAT5a and STAT5b genes deleted, there is no hematopoietic phenotype, suggesting that other pathways may substitute for STAT5 activation.24 We and others have also shown that FDCP-1 cells transduced with a β_c mutant in which all intracellular tyrosines were substituted for phenylalanines were able to proliferate in response to GM-CSF.16,17,25 

We show here the identification in the common β chain of the GM-CSF, IL-3, and IL-5 receptors of a sequence responsible for its association with the adaptor protein 14-3-3ζ. Furthermore, we show that GM-CSF regulates the phosphorylation of ⁵⁸⁵Ser within the 14-3-3ζ binding sequence. Given the conservation of this region in β_c of different species but not in other cytokine receptors, we suggest that it may be involved in GM-CSF, IL-3, and IL-5 receptor-specific functions.

Substitution mutations of 2 sequences within the cytoplasmic domain of the human β_c cDNA were constructed using oligonucleotide-directed mutagenesis (Altered-sites; Promega, Sydney, New South Wales, Australia), as described previously.26Both mutants were essentially poly-alanine substitutions. Mutagenesis oligonucleotides encoding nonalanine residues were included to facilitate restriction enzyme screening of mutant clones. The first motif was ⁵⁸²HSRSLP⁵⁸⁷ mutated to⁵⁸²EFAAAA⁵⁸⁷, and the second was⁸²⁰RSKPSSP⁸²⁶ mutated to⁸²⁰EFAAAAA⁸²⁶. The point mutant S585A was also constructed; however, this mutant created a cryptic proteolytic site in β_c and was not able to be used (see Results). The mutations were confirmed by nucleotide sequencing and the mutant β_c cDNAs subcloned into the eukaryotic expression vector pcDNA1 (Invitrogen, San Diego, CA). The β_cdeletion mutant cDNAs were a kind gift of Dr A. Miyajima (University of Tokyo, Tokyo, Japan). The 14-3-3 zeta clone was isolated from a human monocyte lambda gt11 cDNA library (Clontech Laboratories, Palo Alto, CA) and subcloned into a pGEX-2T vector.

Recombinant human IL-3 and GM-CSF were produced in Escherichia coli essentially as described before.27,28 Cytokine purity and quantitation was determined by high-performance liquid chromatography (HPLC) analysis and Coomasie staining of sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)–separated proteins. The activity of the cytokines based on the ED₅₀ values in a TF-1 proliferation assay29was 0.03 ng/mL for GM-CSF and 0.1 ng/mL for IL-3.

The monoclonal antibodies (MoAbs) 8E4 and 1C1 directed against the β_c were generated as previously described.30The anti–phospho-⁵⁸⁵Serβ_c antibody was raised by immunizing New Zealand White rabbits with the phosphorylated CLGPPHSRSLPDILG peptide conjugated to keyhole limpet hemocyanin (Sigma, St Louis, MO). The antipeptide antibody was firstly affinity purified with the immunizing peptide conjugated to sepharose and then absorbed with the nonphosphorylated CLGPPHSRSLPDILG peptide conjugated to sepharose. The specificity of the anti–phospho-⁵⁸⁵Serβ_c antibody was verified by dot immunoblots against the CLGPPHSRSLPDILG peptide in either the nonphosphorylated or serine-phosphorylated form and a scrambled peptide LPLSGPDSHIRGPL. The corresponding phosphorylated serine in each peptide is underlined. Peptides were synthesized by Chiron Mimotopes (Melbourne, Australia). The anti–14-3-3ζ antibody was kindly provided by Dr A. Aitken (Edinburgh University, Edinburgh, UK).

The HEK293T cell line was maintained in RPMI-1640 supplemented with 10% vol/vol fetal calf serum (FCS). On the day before transfection, 1.4 × 10⁶ cells were plated into 6-cm tissue culture dishes to adhere overnight. Four hours after a medium change, 6 μg of wild-type or mutated β_c cDNA was added to cells in the form of a calcium phosphate precipitate,31 and the cells were placed in an incubator for 4 to 6 hours to permit the uptake of the DNA-calcium phosphate precipitate. The cells were then washed, replated, and placed in the incubator for 48 hours before cytokine treatment. M1 cell line expressing GM-CSF receptor α chain and β_c wild-type was maintained in RPMI-1640 supplemented with 10% vol/vol FCS. The M1 cell line was kindly provided by Dr N. Nicola (Walter and Eliza Hall Institute of Medical Research, Australia).

Expression of receptors on transfected cells was verified by flow cytometry. Briefly, cells were incubated with the anti-β_cMoAb (1C1)32 or anti–GM-CSFRα MoAb (4H1)30for 20 minutes on ice, washed, and subsequently incubated with fluorescein isothiocyanate (FITC)-conjugated antimouse IgG antibody (Silenus Laboratories, Hawthorn, Victoria, Australia) for 20 minutes on ice. Cells were then washed and resuspended in FACS FIX and analyzed using a Profile II (Coulter Electronics, Hileah, FL).

Cells were lysed in lysis buffer (150 mmol/L NaCl, 10 mmol/L Tris-HCl [pH 7.4], 1% Digitonin with protease inhibitors [10 μg/mL leupeptin, 2 mmol/L phenylmethlysulfonyl fluoride, and 10 μg/mL aprotinin], and 2 mmol/L sodium vanadate) for 30 minutes at 4°C followed by centrifugation of the lysate for 15 minutes at 12,000g at 4°C. After 1 hour of preclearance with Protein-A-sepharose (Pierce, Rockford, IL) at 4°C, the supernatant was incubated for 2 hours with 5 μg/mL antibody. Protein-Ig complexes were captured by incubation for 1 hour with Protein-A-sepharose followed by 6 subsequent washes in lysis buffer. Samples were boiled for 5 minutes in SDS sample load buffer in the presence or absence of 2-mercaptoethanol before separating immunoprecipitated proteins by SDS-PAGE.

Cells were lysed in lysis buffer for 30 minutes at 4°C followed by centrifugation of the lysate for 15 minutes at 12,000g at 4°C. After 1 hour of preclearance with glutathione S-transferase (GST)-sepharose at 4°C, the supernatant was incubated for 2 hours with GST–14-3-3-sepharose followed by 3 subsequent washes in lysis buffer. Samples were boiled for 5 minutes in SDS sample load buffer in the presence of 2-mercaptoethanol before separating precipitated proteins by SDS-PAGE.

Cell lysates were precipitated or immunoprecipitated in the presence of various concentrations of the following peptides: β_cpeptide sequence CLGPPHSRSLPDILG in either the nonphosphorylated or serine-phosphorylated form and a scrambled peptide CLPLSGPDSHIRGPL. Raf1 peptides corresponding to the sequence CLSQRQRSTSTPNVHM were also used and were either nonphosphorylated or serine-phosphorylated. The corresponding phosphorylated serine in each peptide is underlined. Peptides were synthesized by Chiron Mimotopes. The presence of β_c in either the immunoprecipitation or precipitation experiment was determined by Western blotting with anti-β_c antibody (MoAb 1C1).

Immunoprecipitated proteins separated by SDS-PAGE were transferred to nitrocellulose membrane by electroblotting. Dot blots of peptides were performed by spotting either 10 or 100 ng of each peptide onto nitrocellulose membranes. Routinely, nitrocellulose membranes were blocked in a solution of phosphate-buffered saline (PBS)/0.05% (vol/vol) Tween 20 (PBT) containing 1% (wt/vol) blocking reagent 1096 176 (Boehringer Mannheim, Mannheim, Germany) and probed with anti-β_c (1C1), anti–14-3-3ζ,33 or anti–phospho-⁵⁸⁵Serβ_c, followed by either anti-mouse or anti-rabbit peroxidase-conjugated antibodies. Immunoreactive proteins were detected by chemiluminescence using the ECL kit (Amersham, Little Chalfont, UK) following the manufacturer’s instructions.

The recombinant 14-3-3ζ protein was ¹²⁵I-labeled using IODOBEADS (Pierce). The synthetic peptides were solubilized in distilled water and then diluted to 50 μg/mL in 0.1 mol/L NaHCO₃, pH 9.2. The peptides were coated onto microtiter wells (Immunolon II Removawells; Dynatech Laboratories, Chantilly, VA) by incubation at 22°C for 6 hours and then at 4°C overnight. The peptide-coated microtiter wells were blocked at 22°C with 5% bovine serum albumin for 2 hours and then with¹²⁵I-labeled 14-3-3ζ protein for 2 hours. After 3 washes, microtiter well-bound radioactivity were estimated in a γ-counter.

In examining the cytoplasmic region of β_c, we identified a small region encompassing amino acids 566 to 589 that exhibits unusually high sequence identity (72%) between mouse and human β_c when compared with the overall identity of the cytoplasmic domain of β_c (55% identity; Fig 1). Previous studies have implicated this region as being important for cell survival and differentiation.34 Closer examination of this region showed, in addition to Tyr⁵⁷⁷ (binding site for SHC), a possible 14-3-3 binding motif, ⁵⁸²HSRSLP⁵⁸⁷that is conserved in β_c from different species. This sequence closely resembles the prototypic 14-3-3 binding consensus (RSXSXP) identified in Raf-1 (where S is phosphorylated) and in other proteins involved in signal transduction (Fig 2).

To test the possibility that human β_c was able to associate with 14-3-3 and to determine the region of interaction, we transfected HEK 293T cells with wild-type β_c and a series of C-terminal truncation mutants and examined the ability of β_c to associate with 14-3-3 by coimmunoprecipitation. Whereas wild-type β_c and C-terminal truncation mutants up to amino acid 626 coimmunoprecipitated with 14-3-3, further truncations resulted in no detectable association of β_c and 14-3-3 (Fig 3A). Similar results were obtained in GST–14-3-3 pull-down experiments. When compared with GST controls, GST–14-3-3 interacted with wild-type β_c and C-terminal truncation mutants up to amino acid 626. However, additional truncations resulted in a loss of detectable association of β_c and GST–14-3-3 (Fig 3). These experiments indicate that (1) β_c interacts with 14-3-3 and (2) that the region of interaction lies between amino acids 544 and 626.

We then examined whether 14-3-3 interacted with β_c via the ⁵⁸²HSRSLP⁵⁸⁷ motif that lies within the 544-626 region identified in Fig 3. A substitution mutant (β_c-⁵⁸²HSRSLP⁵⁸⁷→EFAAAA) and a point mutant (β_c-585S→A) in the putative 14-3-3 binding site were constructed as well as a control mutant (β_c⁸²⁰RSKPSSP⁸²⁶→EFAAAAA). These mutants were expressed in HEK 293T cells and examined for their ability to interact with GST–14-3-3 in pull-down experiments. Whereas wild-type β_c and the control mutant β_c interacted with GST–14-3-3, no detectable interaction was observed for the β_c⁵⁸²HSRSLP⁵⁸⁷→EFAAAA mutant (Fig 4). These results indicate that 14-3-3 associates with β_c via the⁵⁸²HSRSLP⁵⁸⁵ sequence. Experiments examining the association of β_c-585S→A point mutant with GST–14-3-3 were not possible, because it is likely that this mutation introduced a cryptic proteolysis cleavage site. Flow cytometry and Western blot analysis indicated that this mutant was proteolysed and failed to be expressed on the cell surface (data not shown).

14-3-3 is known to be a phospho-serine binding protein that interacts with the RSXSXP motif, where S is phosphorylated. It would then be expected in the case of β_c that⁵⁸⁵Ser phosphorylation within the⁵⁸²HSRSLP⁵⁸⁷ motif would be required for 14-3-3 association. The inability to express full-length β_c-585S→A precluded using this mutant in either coimmunoprecipitation or pull-down experiments to examine the requirement of β_c⁵⁸⁵Ser phosphorylation for 14-3-3 interaction. As an alternative, we synthesized a β_c peptide containing a nonphosphorylated⁵⁸⁵Ser (C⁵⁷⁸LGPPHSRSLPDILG⁵⁹¹) and a β_c peptide containing a phosphorylated⁵⁸⁵Ser and examined their ability to inhibit β_c interaction with GST–14-3-3 in a pull-down experiment. Whereas the peptide containing phosphorylated⁵⁸⁵Ser inhibited β_c association with GST–14-3-3, no inhibition of association was observed for the peptide containing the nonphosphorylated ⁵⁸⁵Ser (Fig 5). As a comparison, nonphosphorylated and phosphorylated peptides corresponding to the 14-3-3 binding site in Raf-1 were also tested. We found that the serine-phosphorylated Raf-1 peptide was also able to inhibit β_c association with 14-3-3, whereas the nonphosphorylated peptide did not (Fig 5). Furthermore, the ability of the β_c-phosphorylated peptide to inhibit the association of β_c with 14-3-3 was dose-dependent and specific as another phosphorylated peptide with sequence corresponding to a different region of β_c failed to inhibit this association (Fig 6). Direct binding experiments and Scatchard analysis demonstrated that the phosphorylated peptide (C⁵⁷⁸LGPPHSRSLPDILG⁵⁹¹) bound to 14-3-3 with an affinity of approximately 150 nmol/L (Fig 7). This affinity is comparable to the affinities reported by other investigators (55 to 190 nmol/L)35,36 for 14-3-3 binding. Together, these experiments show that the ⁵⁸²HSRSLP⁵⁸⁷ sequence in β_c directly binds to 14-3-3 and that this binding is dependent on ⁵⁸⁵Ser being phosphorylated.

Having identified the requirement for β_c⁵⁸⁵Ser to be phosphorylated to allow 14-3-3 binding, we then examined whether ⁵⁸⁵Ser was phosphorylated in vivo and whether its phosphorylation was regulated by GM-CSF. These possibilities were initially addressed using³²P-orthophosphate–labeled HEK 293T cells transfected with the GM-CSF receptor (α and β_c subunits). These cells were stimulated with GM-CSF, and total β_c was immunoprecipitated and examined for ³²P-labeling. No detectable change in β_c serine/threonine phosphorylation in response to GM-CSF stimulation was observed (data not shown). This is most likely due to the presence of 60 serine and threonine residues in the intracellular region of β_c, some of which may be constitutively phosphorylated. To directly address the phosphorylation status of β_c⁵⁸⁵Ser in vivo, we raised a rabbit antiserum against a peptide containing the 14-3-3 binding site identified in β_c: C⁵⁷⁸LGPPHSRSLPDILG⁵⁹¹. This antibody preparation, termed anti–phospho-⁵⁸⁵Ser, specifically recognized the CLGPPHSRSLPDILG peptide containing a phosphorylated ⁵⁸⁵Ser but not a peptide containing a nonphosphorylated ⁵⁸⁵Ser or a peptide containing a scrambled 14-3-3 binding site (Fig 8). The specificity of the anti–phospho-⁵⁸⁵Ser antibody was further confirmed by Western blotting of immunoprecipitated β_c from GM-CSF receptor-transfected HEK 293T cells. In these experiments, the phosphorylated CLPPHSRSLPDILG peptide was able to inhibit anti–phospho-⁵⁸⁵Ser recognition of β_c, whereas the nonphosphorylated and scrambled peptides did not (Fig 8B). In addition, pretreatment of β_cimmunoprecipitates with calf intestinal phosphatase before Western blot analysis completely abolished the anti–phospho-⁵⁸⁵Ser signal, and immunoprecipitation of either the wild-type β_c or the⁵⁸²HSRSLP⁵⁸⁷→EFAAAA mutant from HEK293T transfected cells followed by Western blot analysis with the anti–phospho-⁵⁸⁵Ser antibody demonstrated that this antibody specifically recognized the wild-type but not the mutant receptor (data not shown).

Using these anti–phospho-⁵⁸⁵Ser specific antibodies, we then examined the regulation of β_c⁵⁸⁵Ser phosphorylation after GM-CSF stimulation of MI myeloid leukemic cells. M1 myeloid leukemic cells were stimulated with 2 ng/mL GM-CSF, β_c was immunoprecipitated, and immunoprecipitates were probed with the anti–phospho-⁵⁸⁵Ser antibody. As shown in Fig 8, GM-CSF stimulation strongly upregulated ⁵⁸⁵Ser phosphorylation of β_c.

We show here the identification of a 14-3-3–binding motif in the common β chain of the GM-CSF/IL-3/IL-5 receptors. It is shown that this motif mediates the association of β_c with 14-3-3 in vitro and in vivo and requires the phosphorylation of⁵⁸⁵Ser, but it diverges from the canonical 14-3-3 binding sequence, having an His at position 1 instead of the usual Arg. Furthermore, we obtained evidence that phosphorylation of⁵⁸⁵Ser is regulated by GM-CSF, demonstrating for the first time that not only tyrosine, but also serine phosphorylation of β_c is subjected to ligand-mediated regulation. Given the proximity of the 14-3-3 binding sequence (⁵⁸²HSRSLP⁵⁸⁷) to the SHC-binding sequence (⁵⁷⁷Tyr), we suggest that these form a novel motif perhaps involved in certain specialized functions associated with the GM-CSF, IL-3, and IL-5 receptors.

We have demonstrated β_c:14-3-3 interaction using several approaches. Firstly, we were able to coimmunoprecipitate β_c with 14-3-3, indicating that these proteins interact in vivo. Secondly, we have shown in pull-down experiments that GST–14-3-3 interacts with β_c. Using a series of C-terminally truncated mutants, we have localized this interaction to a region between amino acids 544 and 626. Closer examination of this region showed a possible 14-3-3–binding motif,⁵⁸²HSRSLP⁵⁸⁷. This sequence closely resembles the 14-3-3–binding consensus identified in Raf-1 and several other signaling molecules (⁵⁸²RSXSXP⁵⁸⁷, where S is phosphorylated). Thus, phosphorylation of Ser⁵⁸⁵ within ⁵⁸²HSRSLP⁵⁸⁷ was an expected prerequisite for 14-3-3 binding to β_c. A substitution mutant of this motif (⁵⁸²EFAAAA⁵⁸⁷) abolished the interaction of β_c with 14-3-3. A S585A mutant was not able to be tested for its ability to associate with 14-3-3 due to its poor expression (see Results); however, phosphopeptide competition experiments suggest that the association is dependent on Ser⁵⁸⁵phosphorylation.

Although the phosphorylation of the second serine in the HSRSLP sequence and the presence of a proline at the end are consistently found in 14-3-3 binding sequences, from the 14-3-3 crystal structure it is predicted that the proline in the docking protein produces a critical 90° bend in the chain, allowing the remainder of the protein to exit the binding cleft.36 The presence of an His at the beginning of the sequence is unusual, because the most frequent residue found at this position is Arg (Fig 2), although other amino acids can be found and, in fact, the sequence seen in IRS-1 contains an His. It is becoming clear that there is no fixed 14-3-3 binding sequence but instead a defined hierarchy in which certain amino acids are preferred over others.36 In terms of the sequence found in β_c, the affinity of GST–14-3-3 for a peptide encompassing the phosphorylated 14-3-3 binding motif was 152 nmol/L, which is in line with the affinity of other previously reported 14-3-3 binding motifs.35,36 

The notion that the 14-3-3 binding motif in β_c may be functionally important is supported by the ability of GM-CSF to regulate the phosphorylation of ⁵⁸⁵Ser in the M1 myeloid cell line (Fig 8). This was shown through the development of antibodies specific for the CLGPPHSRSLPDILG sequence in which the second serine was phosphorylated. In previous experiments, we have found that the overall serine/threonine phosphorylation of the cytoplasmic region of β_c (which contains 60 serines and threonines) was similar before and after stimulation by GM-CSF (data not shown), indicating a substantial basal level of phosphorylation that obscured any specific serine and threonine phosphorylation event regulated by ligand. These results make ⁵⁸⁵Ser more akin to tyrosine residues in β_c, because the phosphorylation of these is ligand-dependent. Analogous to tyrosine-phosphorylated sites in β_c that serve as docking sites for STAT-517, SHC,16,17,19 and PTP137 with clear functional consequences, the ligand-dependent phosphorylation of⁵⁸⁵Ser may also initiate distinct signaling events. In other experiments using a CTL-EN cell line expressing the human IL-3 receptor, we found that stimulation with IL-3 increased the association of β_c to 14-3-3 by approximately 5-fold (data not shown). The importance of serine/threonine phosphorylation in mediating the biological activity of cytokines and growth factors is underscored by the recent findings of several groups that several receptor (GHR, EPOR, TPOR, and G-CSF) mutants in which all cytoplasmic tyrosines have been substituted for phenylalanine are able to mediate most, if not all, of the biological activity of the wild-type receptors. We25and others16,17 have also shown that a mutant β_c in which all 8 cytoplasmic tyrosines are substituted for phenylalanine is able to support both proliferation and survival of FDCP1 and BaF3 cells in response to GM-CSF. Thus, it is tempting to speculate that this may be achieved, at least in part, by ligand-dependent ⁵⁸⁵Ser phosphorylation, which activates either the same or alternative and redundant pathways.

Although our results show that 14-3-3 binds ⁵⁸⁵Ser of β_c, little is known about the possible role of 14-3-3 in receptor signaling. The 14-3-3 family of proteins consists of 7 different isoforms and is expressed ubiquitously from yeast to humans.38,39 The ability of 14-3-3 to bind phosphoserine motifs in a wide range of signaling molecules40-50 suggests that 14-3-3 proteins participate in a number of cell signaling pathways that may include mitogenesis, transformation, and survival.

Although 14-3-3 has been shown to bind a number of signaling molecules, it has been more difficult to determine how 14-3-3 can regulate signaling events. However, the recent x-ray crystallographic structure of 14-3-3 has provided several hints.51,52 The 14-3-3 proteins dimerize through an N-terminal domain to form 2 large acidic grooves. It is proposed that this groove can bind 2 separate 14-3-3–binding motifs, raising the possibility that 14-3-3 could function as an adaptor, bringing together 2 different cellular proteins.41,53 Thus, one possible function of 14-3-3 could be to act as an adaptor linking β_c to other 14-3-3–binding molecules involved in GM-CSF signaling. Another possibility is that 14-3-3 could be involved in the regulation of β_c cytoplasmic domain contact with cellular membranes in a manner similar to that described for BAD and Raf-1.

Although the exact functions and signaling pathways involved remain to be elucidated, it is noteworthy that the⁵⁸²HSRSLP⁵⁸⁷ 14-3-3 binding sequence in β_c lies proximal to the ⁵⁷⁷Tyr binding site for SHC.16,17 Because these sequences form a distinct motif conserved in β_c of different species and not found in other cytokine receptors, we propose that they subserve specific functions associated to the GM-CSF, IL-3, and IL-5 receptors. In support of this concept, we note that this new motif lies within a region of β_c involved in myeloid cell survival and differentiation34; however, functional studies will be necessary to test this hypothesis.

The authors thank Anna Nitschke for excellent secretarial assistance and Dr A. Aitken for a gift of antibodies to 14-3-3 and useful discussions.