Abstract
Studies assessing the role of Stat5 in the IL-2 proliferative signal have produced contradictory, and thus inconclusive, results. One factor confounding many of these studies is the ability of IL-2R to deliver redundant mitogenic signals from different cytoplasmic tyrosines on the IL-2R β-chain (IL-2Rβ). Therefore, to assess the role of Stat5 in mitogenic signaling independent of any redundant signals, all cytoplasmic tyrosines were deleted from IL-2Rβ except for Tyr510, the most potent Stat5-activating site. This deletion mutant retained the ability to induce Stat5 activation and proliferation in the T cell line CTLL-2 and the pro-B cell line BA/F3. A set of point mutations at or near Tyr510 that variably compromised Stat5 activation also compromised the proliferative signal and revealed a quantitative correlation between the magnitude of Stat5 activation and proliferation. Proliferative signaling by a receptor mutant with a weak Stat5 activating site could be rescued by overexpression of wt Stat5a or b. Additionally, the ability of this receptor mutant to induce c-myc, bcl-x, and bcl-2 was enhanced by overexpression of wt Stat5. By contrast, overexpression of a version of Stat5a lacking the C-terminal trans-activation domain inhibited the induction of these genes and cell proliferation. Thus, Stat5 is a critical component of the proliferative signal from Tyr510 of the IL-2R and regulates expression of both mitogenic and survival genes through its trans-activation domain.
The T cell mitogen IL-2 binds to a heterotrimeric receptor complex consisting of IL-2Rα, IL-2Rβ, and γc chains (1), with the consequent heterodimerization of IL-2Rβ and γc inducing the proliferative signal (2, 3). γc contributes to proliferative signaling by recruiting the tyrosine kinase Jak3(4, 5, 6, 7, 8), the activation of which also requires a membrane-proximal S region of IL-2Rβ (9, 10). Additionally, at least one of three cytoplasmic tyrosine phosphorylation sites (Tyr338, Tyr393, or Tyr510) distal to the S region on IL-2Rβ must be present for proliferation, suggesting that obligate mitogenic signaling molecules interact with these phosphotyrosines (11, 12). Indeed, the adapter molecule Shc interacts with Tyr338 (11, 13) and can deliver a proliferative signal (14). The other two tyrosines, Tyr510 and Tyr393, provide docking sites for the transcription factor Stat5 (11), which has led to the hypothesis that Stat5 mediates a proliferative signal parallel to the one involving Shc. However, Shc and Stat5 have very different biochemical properties, which raises the question of whether they could indeed generate redundant proliferative signals from IL-2Rβ. In addition, the fact that a single phosphotyrosine on a receptor chain can interact with multiple signaling molecules, as exemplified by studies of the platelet-derived growth factor receptor (15), raises the possibility that a factor other than Stat5 mediates the proliferative signal from Tyr393 and/or Tyr510.
Stat5 refers to either of two highly homologous members of the larger Stat family of signal-transducing activators of transcription, Stat5a and Stat5b (16, 17, 18). Stat activation conventionally commences with selective binding of a Stat SH2 domain to a phosphorylated tyrosine motif on a ligated receptor (19, 20, 21, 22), approximating the Stat molecule with a receptor-associated, activated Janus kinase which phosphorylates a key tyrosine residue on the Stat molecule (23, 24). These Stat phosphotyrosines preferentially bind to Stat SH2 domains (21), causing Stat molecules to dimerize, release from receptors, enter the nucleus, and bind palindromic DNA sequences (22) in certain promoters (25, 26, 27), inducing gene transcription via a C-terminal trans-activation domain (TAD)3 (28, 29, 30, 31). Some members of the Stat family also demonstrate less conventional activities. Stat3, for example, serves as an adapter molecule between PI3 kinase and type I IFN receptors (32). Furthermore, Stat1 and, in cooperation with the glucocorticoid receptor, Stat5 mediate the expression of certain genes through a TAD-independent mechanism (33, 34, 35). Thus, not all signaling functions of Stat molecules involve the conventional trans-activation domain.
A critical role for Stat5 in IL-2-mediated mitogenesis has been suggested by the failure of TCR-stimulated T cells from mice lacking both the Stat5a and Stat5b genes to proliferate in response to IL-2 (36). However, the IL-2R can mediate proliferation in the absence of detectable Stat5 activation through alternative mitogenic signaling molecules, such as Shc (11, 14). Indeed, IL-2Rβ mutants that fail to activate Stat5 mediate proliferation in cultured cell lines (11, 37, 38), and primary T cells (39). Thus, the inability of Stat5-null T cells to proliferate may reflect a recently described essential role for Stat5 in TCR signaling (40) rather than a requisite role in IL-2 signaling.
This report assesses the role of Stat5 in signals derived from the IL-2R in cell lines that do not require TCR stimulation for proliferation. A receptor mutagenesis and rescue strategy was used to study signaling by Stat5 in the absence of any redundant signals from Shc or other molecules potentially associating with IL-2Rβ. We demonstrate that Tyr510 of IL-2Rβ mediates proliferation through the activation of Stat5 and that either the Stat5a or the Stat5b isoform can transduce the mitogenic signal. Furthermore, proliferative signaling by Stat5 is critically dependent on its TAD, indicating that Stat5a mediates proliferation through its conventional role as a transcription factor. Indeed, IL-2-induced transcription of the proliferative and survival genes c-myc, bcl-2, and bcl-x is mediated by Stat5a through the TAD.
Materials and Methods
Plasmid construction
Expression vectors encoding the chimeric αγ- and ββ-chains (formerly denoted GMα/2γ and GMβ/2β) under the control of the human β-actin promoter have been described previously (2, 9, 14, 41). Mutants of ββ were generated by annealing sense and antisense oligonucleotides encoding novel C-terminal sequences for ββ and/or premature stop codons, and cloning overhanging ends of these annealed primers between a unique AflII site in the cytoplasmic domain of ββ and a unique XbaI site immediately 3′ to the stop codon of ββ. For analyses in BA/FG cells, the full length, nonchimeric human IL-2Rβ cDNA was cloned into the expression vector used above, and the C terminus of this chain, between the AflII site and a unique 3′ ScaI site, was then replaced with the respective mutated sequences. The Δ713 mutant of Stat5a was generated by PCR with an antisense oligonucleotide encoding residues 708 to 712 of Stat5a followed by the flag epitope, a stop codon, and an EcoRV site. This PCR product was then cut with XhoI and EcoRV enzymes and ligated between these same sites in a vector containing murine Stat5a, described previously (29). Δ713, as well as C-terminally flag epitope-tagged wild-type (wt) murine Stat5a and Stat5b, were next excised with EcoRI and HindIII and cloned between SalI and HindIII sites in the β-actin promoter-driven plasmid used above for receptor expression. All mutated Stat5a and IL-2Rβ regions were sequenced with the ABI Prism dye terminator cycle sequencing kit (Perkin-Elmer, Norwalk, CT).
Cell culture and transfection
The murine T cell line CTLL-2 was obtained from the American Type Culture Collection (Manassas, VA) and maintained as described previously (14). BA/F3 pro-B cells were obtained from Immunex (Seattle, WA) and maintained in RPMI supplemented with 10% FCS, 2 mM l-glutamine, 50U/ml penicillin, and 50 mg/ml streptomycin, as well as 10% WEHI3-conditioned medium as a source of murine IL-3. To grow BA/F3 cells in the absence of Stat5 activation, the cells were transfected with a chimeric G-CSFR/gp130 receptor chain described previously (42) and maintained with 100 ng/ml recombinant human G-CSF (Amgen, Thousand Oaks, CA) instead of IL-3. The resultant cell line is referred to hereafter as BA/FG. Linearized plasmids were introduced into cells by electroporation, and transfectants were selected for resistance to G418 (Life Technologies, Gaithersburg, MD) in 96-well plates at limiting dilution to isolate independent subclones. Receptor expression was assessed by flow cytometry with Abs to human GM-CSFRα or βc (Santa Cruz Biotechnology, Santa Cruz, CA), or to IL-2Rβ (PharMingen, San Diego, CA). Stat5 expression was assessed by Western blot with an Ab to Stat5 (Transduction Labs, Lexington, KY), or by flow cytometry of cells stained intracellularly with an Ab to the flag epitope (Sigma, St. Louis, MO). Subclones with comparable receptor and/or Stat5 expression were chosen for further analyses.
Proliferative assays
Thymidine incorporation assays were conducted in triplicate wells with 104 cells/well exposed to the indicated doses of GM-CSF, IL-2, G-CSF, or IL-3 for 24 h, with [3H]thymidine (2.5 μCi/well) present during the last 4 h. Cells were harvested onto glass fiber filters, and DNA synthesis was quantified by liquid scintillation counting. The data presented in Figs. 1–4⇓⇓⇓⇓ used 100 ng/ml GM-CSF or 3000 U/ml IL-2, which were found to elicit a maximal response from all functional receptor mutants in CTLL-2 or BA/F3 cells, respectively.
An IL-2 proliferative signal is mediated through a conserved 13-residue motif flanking Tyr510 (Y510). A, Schematic depicting wt and mutant versions of chimeric GM-CSF/IL-2R β-chain (ββ). TM, transmembrane. B, Thymidine incorporation mediated by the indicated chimeric receptors. •, Individual CTLL-2 clones stimulated with 100 ng/ml human GM-CSF; ○, the same clones cultured without cytokine. Data for each clone are expressed as a percentage of the thymidine incorporated after stimulation with 100 U/ml human IL-2. C, ββΔ325 retains the ability to activate Janus kinases. Representative CTLL-2 clones bearing the indicated chimeric receptors were deprived of cytokine for 4 h and then stimulated for 10 min with 100 ng/ml human GM-CSF (G), 100 U/ml human IL-2 (2), or no cytokine (0). The indicated Janus kinases were then immunoprecipitated (IP) and either subjected to an in vitro autokinase assay or a Western blot for phosphorylated tyrosines (pTyr). The phosphotyrosine Western blots were then stripped and reprobed for Jak1 or Jak3, as indicated, to assess loading. D, ββΔ325 + Tyr510 activates Stat5 through Tyr510. Representative CTLL-2 clones bearing the indicated chimeric receptors were deprived of cytokine for 4 h and then stimulated for 10 min with 100 ng/ml human GM-CSF (G), 100 U/ml human IL-2 (2), or no cytokine (0). Nuclear extracts were harvested and subjected to an electrophoretic mobility shift assay using a Stat5-binding probe derived from the FcγR1 promoter. In lanes marked Ab, extracts from GM-CSF-stimulated cells were treated with anti-Stat5 Abs to specifically supershift Stat5 complexes. F510, Phe510.
Activation of Stat5 by Tyr510 of IL-2Rβ shows a quantitative correlation with proliferation. A, Schematic comparison of the residues surrounding Stat5-activating tyrosines (Y) (underlined) in the cytoplasmic domains of various receptors. Conserved hydrophobic residues are in bold. B, Schematic depicting point mutations (bold face) introduced C-terminal to Tyr510 (underlined) in ββΔ325 + Tyr510. C, Stat5 activation in CTLL-2 cells mediated by mutant chimeric receptors, assessed as per Fig. 1⇑D. D, Thymidine incorporation in CTLL-2 cells mediated by the indicated chimeric receptors, assessed and depicted as per Fig. 1⇑B.
Overexpression of Stat5a or Stat5b rescues the ability of ββΔ325 + YRSL to mediate proliferation. A, Stat5 expression in transfected CTLL-2 cells. Whole cell lysates of CTLL-2 clones transfected with a vector encoding ββΔ325 + YRSL either alone (0) or with a vector encoding either Stat5a (6 lanes on left) or Stat5b (4 lanes on right) were assessed for Stat5 content by Western blot using a Stat5-specific Ab. The blot was stripped and reprobed with a Jak1-specific Ab to demonstrate even loading. B, Stat5 overexpression rescues Stat5 activation by ββΔ325 + YRSL. Representative CTLL-2 clones expressing ββΔ325 + Tyr510 (Y510), ββΔ325 + YRSL (+0), or ββΔ325 + YRSL plus the indicated Stat5 isotype were assessed by electrophoretic mobility shift assay for the ability to activate Stat5 as per Fig. 1⇑D. C, Stat5 overexpression rescues thymidine incorporation mediated by ββΔ325 + YRSL. The ability of multiple CTLL-2 clones expressing ββΔ325 + Tyr510 or ββΔ325 + YRSL with or without the indicated Stat5 isotype to induce thymidine incorporation was assessed and depicted as in Fig. 1⇑B. D, Stat5 over expression rescues cell expansion mediated by ββΔ325 + YRSL. Representative CTLL-2 clones expressing the indicated chimeric receptor/Stat5 combinations were cultured for several days in the presence of 100 ng/ml human GM-CSF (GM), 100 U/ml human IL-2 or no cytokine (0). Every 2 days, cells were stained with trypan blue and live cells were counted. Cells were split as needed to maintain a density of <5 × 105 cells/ml. All clones expanded similarly in response to IL-2, and all died within 2 days in the absence of cytokine (data not shown). E, Stat5 overexpression rescues viability mediated by ββΔ325 + YRSL. Representative CTLL-2 clones were cultured and counted as in D. Cells excluding and incorporating trypan blue were counted as live and dead, respectively. All clones remained comparably viable in the presence of IL-2 and died at a similar rate in the absence of cytokine (data not shown).
Stat5-mediated proliferation requires the TAD of Stat5. A, Schematic depiction of wt or mutant versions of IL-2Rβ and flag epitope-tagged Stat5a used in these studies. B, Stat5 DNA binding activity is enhanced by wtStat5a or Δ713. Representative BA/FG clones bearing the indicated versions of IL-2Rβ and Stat5a were deprived of cytokine for 4 h and then stimulated for 10 min with 3000 U/ml human IL-2 (2), 10% IL-3-conditioned media (3), or no cytokine (0). Nuclear extracts were then harvested and assessed for Stat5 DNA-binding activity as in Fig. 1⇑D. Lanes marked Ab represent extracts from IL-2-stimulated cells treated with an anti-flag Ab to specifically supershift transfected Stat5. C, Thymidine incorporation is enhanced by wtStat5a and inhibited by Δ713. Three BA/FG clones expressing each of the indicated combinations of IL-2Rβ and/or Stat5a variants were stimulated with 3000 U/ml human IL-2 (•), 10% IL-3-conditioned medium (▴), or no cytokine (○). Data for each clone are expressed as a percentage of the thymidine incorporated by that clone when stimulated with 100 ng/ml human G-CSF. D, IL-2-mediated expansion of BA/FG expressing Δ325 + Tyr510 is inhibited by Δ713. A representative BA/FG clone expressing Δ325 + Tyr510 with or without Δ713 was cultured in the presence of 3000 U/ml human IL-2, 100 ng/ml human G-CSF, or no cytokine (0). Cell proliferation was assessed and depicted as in Fig. 3⇑D. Cells were split as needed to maintain a density of <106 cells/ml. All clones proliferated similarly in response to G-CSF and died at a similar rate in the absence of cytokine (data not shown).
Immunoprecipitaion and immunoblotting
Jak1 and Jak3 were immunoprecipitated and subjected to kinase assays as described previously (9). Whole cell lysates were generated by boiling cells in 62.5 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, 50 mM DTT, and 0.1% bromphenol blue. Lysates or immunoprecipitates were electrophoresed on acrylamide gels and transferred to nitrocellulose. Nitrocellulose blots were blocked with 0.1 M Tris base (pH 7.5), 0.9% sodium chloride, 0.05% Tween 20 (TTBS) containing 1% BSA (for antiphosphotyrosine probes) or 5% powdered skim milk (Carnation, Glendale, CA), and probed with rabbit antisera recognizing Jak1 (Santa Cruz Biotechnology) or Jak3 (Upstate Biotechnology, Lake Placid, NY) or murine Abs recognizing phosphotyrosine (4G10; Upstate Biotechnology) or Stat5 (Transduction Laboratories, Lexington, KY). Blots were then washed with TTBS, probed with peroxidase-conjugated goat anti-rabbit or anti-mouse Abs (Life Technologies), and washed again with TTBS. Bound Abs were detected by enhanced chemiluminescence (DuPont NEN, Boston, MA). Blots were stripped between probings with a 30-min, 50°C incubation in 62.6 mM Tris-HCl (pH 6.7), 0.1 M β-mercaptoethanol, and 2% SDS.
Electrophoretic mobility shift assays
Cells stimulated as indicated were washed once with buffer H (20 mM HEPES (pH 7.9), 1 mM EDTA, 0.1 mM EGTA, 2 mM magnesium chloride, 1 mM sodium o-vanadate, 20 mM sodium fluoride, 1 mM DTT, 0.1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, and 1 mg/ml leupeptin) and lysed in buffer H plus 0.2% Nonidet P-40 at 0°C. Nuclei were pelleted by centrifugation and extracted with buffer K (buffer H plus 0.42 M sodium chloride and 20% v/v glycerol). To generate a probe for Stat activity, incompletely overlapping oligonucleotides corresponding to the sense and antisense strands of the Stat-responsive DNA element from the FcγR1 promoter were annealed, radiolabeled with [α-32P]dCTP by an end-filling T4 polymerase reaction, and purified with a MicroSpin G-25 column (Pharmacia, Piscataway, NJ). The probe was added to nuclear extracts in 50 mM potassium chloride, 10 mM HEPES (pH 7.9), 10% glycerol, 1 mM DTT, and 87.5 mg/ml dITP/dCTP at room temperature for 30 min in the presence or absence of either an Ab recognizing the flag epitope (Sigma) or a combination of Abs recognizing Stat5a and Stat5b (Santa Cruz Biotechnology). Reaction mixtures were electrophoresed on a nonreducing 0.25× Tris-buffered EDTA-acrylamide gel which was then dried and subjected to autoradiography.
Northern blots
Cells stimulated as indicated were pelleted by centrifugation and flash-frozen in a dry ice-ethanol bath. RNA was harvested from thawed pellets with the RNA Stat 60 kit (Tel Test, Friendswood, TX); denatured for 10 min at 65°C in 20 mM MOPS, 5 mM sodium acetate, 0.5 mM EDTA, 2.4 M formaldehyde, and 50% formamide; and run on a 1.2% agarose gel (containing 20 mM MOPS, 5 mM sodium acetate, 0.5 mM EDTA, and 1.1 M formaldehyde) in 20 mM MOPS, 5 mM sodium acetate, and 0.5 mM EDTA at pH 7.0. RNA was passively transferred to Zetabind membranes (Cuno, Meriden, CT) with 10× SSC (1.5 M sodium chloride, 0.15 M sodium citrate, pH 7.0), and UV cross-linked. Blots were prehybridized at 43°C in hybridization buffer (1 M sodium phosphate (pH 7.1), 2 mM EDTA, 2% BSA, 10% SDS, 50% formamide, and 0.16 mg/ml yeast tRNA or herring sperm DNA). To generate nucleic acid probes, a 0.4-kb PstI fragment of murine c-myc, a 0.9-kb PstI fragment of murine bcl-2, a 1-kb EcoRI fragment of murine bcl-x, and a 1.2-kb PstI fragment of murine GAPDH cDNA were radiolabeled with [α-32P]dCTP using a random-primed labeling kit (Boehringer Mannheim, Indianapolis, IN) and purified with Centri-Sep spin columns (Princeton Separations, Adelphia, NJ). Probes were boiled for 5 min and added to blots in hybridization buffer. After overnight incubation at 43°C, blots were washed 2–3 times with 2× SSC, 0.1% SDS; once with 0.2× SSC, 0.1% SDS at room temperature; and 0–2 times with 0.2× SSC, 0.1% SDS at 55°C before autoradiography. Blots were stripped between probings with a 2-min immersion in boiling water.
Results
Proliferative signaling by Tyr510 of IL-2Rβ correlates quantitatively with Stat5 activation
Structure/function analyses of the cytoplasmic domain of IL-2Rβ were performed in the IL-2-dependent murine T cell line CTLL-2. To avoid activation of the endogenous IL-2R, we introduced a previously described chimeric GM-CSF/IL-2 receptor consisting of two chains, αγ and ββ, containing the extracellular domains of the human GM-CSF receptor α- and β-chains fused, respectively, to the transmembrane and intracellular regions of γc and IL-2Rβ. When coexpressed, αγ and ββ deliver in response to human GM-CSF a signal that is biochemically and physiologically indistinguishable from that induced in the same cell by the wt IL-2 receptor (2, 9, 14, 41). A CTLL-2 clone demonstrating stable high expression of αγ was generated and transfected with wt or mutated derivatives of ββ (Figs. 1⇑A and 2B). Subclones of transfectants were analyzed for receptor expression by flow cytometry (data not shown), and those expressing both αγ and ββ-chains at comparable levels were chosen for further study. The wt version of ββ delivered a proliferative signal in response to GM-CSF similar to that mediated by the endogenous IL-2R. By contrast, a version of ββ from which all cytoplasmic tyrosines had been removed by introduction of a premature stop codon at codon 325 (relative to the human IL-2Rβ sequence) (ββΔ325 (Fig. 1⇑A)) failed to mediate proliferation (Fig. 1⇑B), in accordance with prior reports (11, 12, 14). However, both ββwt and ββΔ325 induced phosphorylation and activation of Jak1 and Jak3 (Fig. 1⇑C), confirming that Janus kinase activation, whereas critical for Stat activation, is not sufficient for proliferation.
To investigate the role of Stat5 in proliferation, 13 amino acids from the C terminus of IL-2Rβ that encompass Tyr510, the most efficient Stat5-activating tyrosine on IL-2Rβ (11), were attached to the C terminus of the ββΔ325 receptor (ββΔ325 + Tyr510 (Fig. 1⇑A)). Like ββwt and the endogenous IL-2R, this receptor induced strong Stat5 activation (Fig. 1⇑D), as well as cell proliferation (Fig. 1⇑B). Thymidine incorporation was 30–40% less than that induced by ββwt or the endogenous IL-2R, likely due to the absence of a concurrent proliferative signal mediated through Shc (11, 14). Nonetheless, ββΔ325 + Tyr510 supported the long term proliferation and survival of CTLL-2 cells in culture (Fig. 3⇑, D and E; data not shown). However, if Tyr510 was mutated to phenylalanine in this receptor context (ββΔ325 + Phe510 (Fig. 1⇑A)), neither Stat5 activation nor proliferation was induced (Fig. 1⇑, B and D), indicating both events were dependent on Tyr510.
Like previous reports (11, 43), these data show a correlation between Stat5 activation and proliferation but do not exclude the participation of an undefined molecule interacting with Tyr510. Indeed, such a molecule has been previously proposed to explain how Tyr510 mediates IL-2Rα expression (44). If a molecule other than Stat5 mediates the proliferative signal from Tyr510, specific disruption of the interaction between Stat5 and ββΔ325 + Tyr510 should not eliminate mitogenesis. To test this possibility, point mutations were introduced near Tyr510 in ββΔ325 + Tyr510 to specifically disrupt the consensus Stat5-binding motif without eliminating Tyr510. Analysis of Stat5-binding sites in several cytokine receptors suggested that the SH2 domain of Stat5 preferentially associates with a phosphorylated tyrosine followed by hydrophobic amino acids one and three positions C-terminal to the tyrosine (i.e., YLSL in the case of IL-2Rβ) (Fig. 2⇑A). These residues, in the context of ββΔ325 + Tyr510, were point mutated individually or concomitantly to arginine or glutamic acid to generate four mutant receptors (Fig. 2⇑B). Compared with ββΔ325 + Tyr510, these receptors demonstrated a spectrum of Stat5 activation potencies (Fig. 2⇑C) which correlated directly with proliferation (Fig. 2⇑D).
IL-2 delivers a proliferative signal through Stat5a or Stat5b
Although designed to specifically disrupt Stat5 binding, these mutations near Tyr510 could also disrupt the binding of other putative mitogenic molecules. Therefore, to directly test whether Stat5 delivers the proliferative signal from Tyr510, the effect of overexpressing Stat5 on the marginal Stat5-activating potential of ββΔ325 + YRSL was assessed. By mass action, overexpression of Stat5 was expected to overcome the affinity barrier between Stat5 and Tyr510 created by the mutation of Leu511 to arginine. A flag epitope-tagged version of murine Stat5a was cotransfected with ββΔ325 + YRSL into CTLL-2 cells bearing αγ, and stable subclones that coexpressed αγ, ββ, and flag-Stat5a were identified by flow cytometry (data not shown). Expression levels of transfected Stat5a were estimated by Western blot to be 5- to 10-fold higher than endogenous Stat5 (Fig. 3⇑A). As intended, overexpression of Stat5a restored the ability of ββΔ325 + YRSL to induce Stat5 DNA-binding activity in response to GM-CSF to levels equal to or greater than that mediated by ββΔ325 + Tyr510 (Fig. 3⇑B). Rescue of Stat5 activation in turn restored the ability of ββΔ325 + YRSL to induce thymidine incorporation (Fig. 3⇑C), and cell proliferation (Fig. 3⇑D) and maintain cell viability (Fig. 3⇑E). Thus, the proliferative signal from Tyr510 of IL-2Rβ can be mediated by Stat5a. Overexpression of Stat5a also enhanced the level of Stat5 DNA-binding activity induced by the endogenous IL-2R (Fig. 3⇑B) but had no observable effect on IL-2-mediated proliferation or viability of CTLL-2 cells (Fig. 3⇑, D and E). This suggests that the degree of Stat5 activation normally induced by the IL-2R is saturating with respect to cell proliferation.
Functional differences have been described between the DNA-binding activities of Stat5a and Stat5b (45, 46). Additionally, mice lacking either Stat5a or Stat5b genes have distinct phenotypic abnormalities (26, 47, 48, 49, 50). Thus, Stat5a and Stat5b may serve nonredundant functions. To determine whether Stat5b can also mediate a proliferative signal from Tyr510, a flag epitope-tagged version of Stat5b was cotransfected with ββΔ325 + YRSL, as above. Although we were unable to overexpress Stat5b to the levels achieved with Stat5a (Fig. 3⇑A), Stat5b overexpression nevertheless enhanced the ability of ββΔ325 + YRSL to activate Stat5 (Fig. 3⇑B) induce thymidine incorporation (Fig. 3⇑C), support cell proliferation (Fig. 3⇑D), and inhibit cell death (Fig. 3⇑E).
The TAD of Stat5 is critical for Stat5 to mediate proliferation
Stat molecules conventionally require both a conserved tyrosine phosphorylation site and a C-terminal TAD to mediate transcription (23, 24, 29). However, the TAD of Stat5 has been reported to be dispensable for some transcriptional events, such as β-casein induction mediated by Stat5 in cooperation with the glucocorticoid receptor (34, 35). Similarly, the TAD of Stat5 appears dispensable for induction of c-myc by IL-3 (51) and for proliferation of the promyeloid line 32D in response to the Stat5-activating cytokines Epo and IL-3 (29).
To test the requirement for the TAD of Stat5 in IL-2R signaling, we assessed whether overexpression of a TAD-deficient version of Stat5 would enhance or inhibit the proliferative signal mediated by Tyr510 of IL-2Rβ. Δ713 is a naturally occurring isoform of Stat5a that lacks the TAD (Fig. 4⇑A) but retains all other functional domains and therefore can become tyrosine phosphorylated and bind DNA in response to cytokine stimulation (29). If overexpressed in CTLL2 cells, Δ713 could potentially exert a dominant negative blockade on some Stat5-mediated transcriptional events (29, 31, 51) and impose undesirable selective pressure before analysis. Therefore, experiments with Δ713 were performed in a cytokine-dependent lymphocyte line that can grow in the absence of any Stat5 activity. The IL-3-dependent pro-B cell line BA/F3 was transfected with a chimeric receptor containing the extracellular domain of the human G-CSF receptor and the transmembrane and cytoplasmic domains of the gp130 receptor chain (42), to produce a cell line termed BA/FG. gp130 activates Stat3, but not Stat5 (52), so this chimeric receptor promoted the proliferation of BA/F3-derived cells in response to human G-CSF (42) without activation of Stat5 (data not shown). BA/FG cells express an endogenous γc chain but do not respond to IL-2 due to the absence of IL-2Rβ, which made it unnecessary to use the chimeric αγ and ββ receptor chains used in CTLL-2 cells. Instead, IL-2Rβ mutants were tested using the normal ectodomain of IL-2Rβ. Stable transfectants of BA/FG cells were generated that expressed the different IL-2Rβ mutants at similar levels, as assessed by flow cytometry (data not shown). Experiments were performed with a high dose of IL-2 (3000 U/ml) sufficient to saturate IL-2Rβ and γc irrespective of IL-2Rα expression, in that the latter is dependent on Stat5 activity (27).
We first confirmed that BA/FG cells, like CTLL2 T cells, were dependent on Stat5 to mediate the proliferative signal from Tyr510 of IL-2Rβ. As expected, BA/FG expressing Δ325 + Tyr510 activated Stat5 and proliferated in response to IL-2, unless Tyr510 was point mutated to phenylalanine (Δ325 + Phe510 (Fig. 4⇑)). Moreover, BA/FG cells expressing Δ325 + YRSL demonstrated dramatically impaired Stat5 activation and proliferation, and both defects could be rescued by overexpression of wtStat5a (Fig. 4⇑, B and C).
To test the role of the Stat5 TAD in signaling, stable BA/FG clones coexpressing Δ325 + Tyr510 and the Δ713 version of Stat5a were generated. Expression of Δ713 did not appear to impose adverse selective pressure on BA/FG cells, inasmuch as 1) cells expressing Δ713 and wtStat5a were generated with similar cloning efficiencies, 2) the two versions of Stat5 were expressed at similar levels as revealed by flow cytometry (data not shown), and 3) Δ713 had no effect on the proliferative response of BA/FG cells to G-CSF (Fig. 4⇑D). Similar to wtStat5a, overexpression of Δ713 dramatically enhanced Stat5 DNA-binding activity induced by Δ325 + Tyr510, as well as by the endogenous IL-3 receptor (Fig. 4⇑B), confirming that the TAD is not required for receptor-mediated nuclear translocation or DNA binding by Stat5(29). Despite the capacity to inducibly bind DNA, Δ713 severely reduced the ability of either Δ325 + Tyr510 or the IL-3 receptor to induce BA/FG proliferation (Fig. 4⇑, C and D). Thus, the transcriptional activation domain is essential for Stat5 to mediate proliferative signals from the Tyr510 site of IL-2Rβ and the full length IL-3 receptor.
Stat5 mediates induction of c-myc, bcl-2, and bcl-x through its TAD
The obligate role for the TAD in Stat5 mitogenic signaling suggests that Stat5 may mediate proliferation through induction of genes critical for cell cycle progression and/or survival. c-myc, a protooncogene induced by IL-2 (53), is essential for T cells to enter S phase in response to stimulation (54). bcl-2 and bcl-xL, also induced by IL-2 (55), have been shown to support lymphocyte survival (56, 57). Constitutive expression of both c-myc and bcl-2 is sufficient to support factor-independent growth of the lymphoid cell line BAF/B03 (58).
To assess the role of Stat5 in c-myc, bcl-2, and bcl-x induction, Northern blot analyses were performed with the different BA/FG transfectants described above. The CIS gene was used as an internal control for Stat5 activity (26). Δ325 + YRSL, which induces minimal Stat5 activation, induced c-myc to a moderate level and only weakly induced CIS, bcl-2, and bcl-x in comparison with the endogenous IL-3 receptor (Fig. 5⇓A). Overexpression of Stat5a with Δ325 + YRSL enhanced induction not only of CIS but also of c-myc, bcl-2, and bcl-x (Fig. 5⇓A), thereby implicating Stat5 in the regulation of all four genes. To test the role of the Stat5 TAD in the induction of these genes, the effects of Δ713 were assessed. Although expression of Δ713 enhanced the ability of Δ325 + Tyr510 to promote Stat5 DNA-binding activity (Fig. 4⇑B), it prevented induction of the CIS gene by this receptor, as well as the IL-3 receptor (Fig. 5⇓B). Thus, Δ713 exerted dominant negative inhibition of conventional Stat5 transcriptional activity, consistent with prior assessment of this construct (29). In so doing, Δ713 also inhibited the ability of Δ325 + Tyr510, as well as the IL-3 receptor, to induce c-myc, bcl-2, and bcl-x (Fig. 5⇓B), indicating that Stat5 regulates expression of these genes by a TAD-dependent mechanism.
Stat5 mediates c-myc, bcl-2, and bcl-x induction through its TAD. A and B, BA/FG cells expressing the indicated versions of IL-2Rβ and Stat5a were deprived of cytokine for 8 h and then stimulated with 3000 U/ml IL-2, 10% IL-3 conditioned medium, or 100 ng/ml G-CSF for the indicated times. RNA was harvested from cells and assessed by Northern blot with serial probing for expression of CIS, c-myc, bcl-2, bcl-x, and GAPDH. C, The ability of BA/FG expressing Δ325 + Tyr510 (Y510) to induce c-myc, bcl-2 and bcl-x with (+CHX) or without 20 μg/ml cycloheximide administered 30 min before cytokine stimulation was assessed as in A and B.
To determine whether the induction of these genes by Stat5 is an immediate-early event or requires the de novo synthesis of intermediate transcription factors, the transcriptional consequences of Δ325 + Tyr510 signaling were assessed in the presence of the protein synthesis inhibitor cycloheximide. Although cycloheximide enhanced the basal expression of genes in the absence of cytokine as reported for c-myc (59, 60), IL-2 and IL-3 still induced expression of c-myc and bcl-x above this basal level (Fig. 5⇑C). By contrast, induction of bcl-2, which occurred with slower kinetics than seen with c-myc and bcl-x (Fig. 5⇑, A and B), was blocked by cycloheximide treatment (Fig. 5⇑C). Thus, whereas bcl-2 induction requires de novo protein synthesis, c-myc and bcl-x are induced by Stat5 in an immediate-early fashion, suggesting that Stat5 may interact directly with the promoters of these genes rather than via an intermediary transcription factor.
Discussion
This report investigates the mechanism of mitogenic signaling by a single cytoplasmic tyrosine (Tyr510) on IL-2Rβ which has previously been shown to interact with the transcription factors Stat5a and Stat5b (11, 43). By eliminating all other tyrosines from IL-2Rβ and then introducing point mutations at or near Tyr510, we have established a quantitative correlation between the level of Stat5 activation mediated by IL-2Rβ and subsequent proliferation of the murine T cell line CTLL-2. Moreover, overexpression of Stat5a or Stat5b rescued the proliferative signal mediated by a receptor mutant with a low affinity Stat5-binding site, thereby directly implicating Stat5a and Stat5b as mediators of mitogenic signaling in T cells. Proliferative signaling by Stat5a required a C-terminal TAD, which was similarly required for Stat5 to induce expression of the c-myc, bcl-x, and bcl-2 genes, indicating that Stat5 mediates proliferation by trans-activation of both promitogenic and antiapoptotic genes.
Previous studies of the role of Stat5 in proliferative signaling have yielded somewhat contradictory results. Mui et al. (51) reported that a dominant-negative mutant of Stat5 significantly inhibited the proliferation of the pro-B cell line BA/F3 in response to IL-3, another cytokine which activates Stat5. This occurred in the face of normal c-myc induction, suggesting that Stat5 was critical for some, but not all, aspects of the proliferative signal. By contrast, Wang et al. (29) reported that dominant-negative inhibition of Stat5 activity in the promyeloid cell line 32D had no impact on proliferative responses to IL-3 and Epo. The residual proliferative responses seen in these two studies may have been mediated by one or more redundant proliferative pathways operating in parallel with Stat5, such as the Shc pathway (14, 61). Indeed, the Stat5-binding sites on IL-2Rβ (Tyr510 and Tyr393) are only critical for proliferation when Shc activity is abrogated by mutation of its binding site at Tyr338 (11). Consequently, Δ713 or other dominant negative versions of Stat5 would not be expected to block proliferation mediated by full length IL-2Rβ. For this reason, the experiments in the present study utilized a mutated version of IL-2Rβ that lacked the Shc-binding site and contained only the Stat5 activation site at Tyr510. By eliminating potentially redundant pathways, a definitive role for Stat5 in proliferative signaling was revealed.
Although both Stat5 and Shc are clearly able to mediate cell proliferation, the extent to which each contributes to the IL-2-induced proliferation of primary T cells remains controversial. Fujii et al. (39) reconstituted IL-2Rβ-deficient mice with a mutated version of IL-2Rβ lacking the ability to detectably activate Stat5, but containing an intact Shc-binding site. They found that this mutated receptor promoted a normal proliferative response to IL-2 in primary T cells (39), consistent with prior studies in cell lines (11, 37, 38, 62). By contrast, Moriggl et al. reported that T cells from mice lacking both the Stat5A and Stat5B genes show no proliferative response to IL-2 (36), which led them to conclude that Stat5 is absolutely essential for IL-2-induced proliferation in primary T cells. It is unclear why no Stat5-independent proliferative pathway was operational in T cells from the latter mice, as it apparently was in the cells studied by Fujii et al. (39). One possible explanation for the discrepant results is that the mutant form of IL-2Rβ used by Fujii et al. might have been able to activate Stat5 to a physiologically relevant but experimentally undetectable level. However, this seems an unlikely explanation, in that we show here that even detectable, low level Stat5 activity is insufficient for the high level of T cell proliferation observed by Fujii et al. (39) (Figs. 2⇑ and 3⇑, B--E). An alternative explanation is that deletion of both the Stat5a and Stat5b genes by Moriggl et al. had unanticipated detrimental effects on the expression or function of one or more essential components of the IL-2R complex, such as Jak3, rendering the receptor globally dysfunctional in the total absence of Stat5. There is precedent for such a mechanism, because the basal activity of Stat1 has recently been shown to be required for expression of caspases, which in turn are required to sensitize cells to TNF-α-mediated apoptosis (33). Finally, it is also possible that the severe proliferative defect seen in Stat5a/Stat5b-deficient T cells in part reflects impaired signaling through the TCR, which has recently been show to induce transient activation of Stat5 (40).
The molecular mechanism by which Stat5 mediates proliferative signaling has also remained obscure. A truncated version of Stat5, which was unable to dimerize and hence bind DNA, inhibited IL-3-mediated proliferation by a dominant negative mechanism (51), thus implicating Stat5 in the IL-3 proliferative signal. However, in a second study by Wang et al. (29), a less severely truncated version of Stat5 lacking only the TAD had no impact on the IL-3-proliferative signal. Together, these studies suggested that Stat5 may mediate proliferative signaling by a TAD-independent mechanism. Consistent with this notion, a version of Stat5 capable of binding DNA, but lacking its TAD, can nevertheless interact with the glucocorticoid receptor to induce transcription of the β-casein gene (34, 35), thereby demonstrating that Stat5 is capable of TAD-independent transcriptional regulation. However, our results define a clear requirement for the C-terminal TAD of Stat5 for the proliferative response of lymphocytes to both IL-2 and IL-3. This indicates that, despite its capacity for TAD-independent signaling, Stat5 induces cell proliferation through the conventional TAD at its C terminus. The discrepancy between our findings and those of Wang et al. may reflect the fact that their cells were cultured with IL-3 in the face of constitutive dominant negative Stat5 expression, which would be expected to impose a strong selective pressure in favor of Stat5-independent signaling mechanisms. In the present study, we avoided this potential problem by creating a cell line, BA/FG, that does not use Stat5 for growth and instead responds to a Stat3-dependent proliferative signal delivered by the gp130 receptor subunit. This allowed for expression of the TAD-deficient form of Stat5 at a level sufficient for complete dominant negative activity, which likely accounts for the more pronounced impact on IL-2- and IL-3-mediated proliferation observed in this study compared with Wang et al. (29) and Mui et al. (51).
In addition to cell proliferation, Stat5 was also found to induce expression of the growth-related genes c-myc, bcl-2, and bcl-x through a TAD-dependent mechanism (Fig. 5⇑, A and B). Induction of the c-myc and bcl-x genes by Stat5 did not require de novo protein synthesis (Fig. 5⇑C), suggesting that Stat5a interacts directly with the promoters of these genes. Although no Stat5 binding sites have yet been functionally defined in these promoters, other Stat family members have been shown to regulate these genes. Stat1 can interact with a TTCGGAGAA motif in the human bcl-x promoter, and this site is involved in the regulation of bcl-x expression (63). Stat3 has also been proposed to regulate bcl-x expression (64) and can associate with and trans-activate the c-myc promoter (65). However, the site in the c-myc promoter with which Stat3 interacts is not appreciably bound by Stat5(65) (J.D.L., unpublished observations), suggesting that Stat3 and Stat5 mediate c-myc induction by binding distinct promoter sites.
In contrast to c-myc and bcl-x, induction of bcl-2 by Stat5 appears to be a delayed early event, because it occurs with slower kinetics and requires de novo protein synthesis (Fig. 5⇑). This suggests that Stat5 induces expression of one or more secondary transcription factors that in turn bind to and trans-activate the bcl-2 promoter. Such a model, in which Stat5 initiates a cascade of gene activation events, would explain how a single transcription factor could promote the diverse array of cellular changes that are necessary for cell cycle progression, including protooncogene induction, up-regulation of cell metabolism, and activation of the cell cycle machinery.
Although the Δ325 + Tyr510 receptor was able to mediate Stat5 activation at least as potently as ββwt and the endogenous wt IL-2 receptor (Fig. 1⇑D), it was ∼40% weaker in its ability to promote cell proliferation (Fig. 1⇑B). This suggests that Stat5 is not sufficient to mediate all aspects of the IL-2 proliferative signal. Indeed, activation of the ras/receptor-activated factor/mitogen-activated protein kinase signaling pathway, culminating in the induction of c-fos, c-jun, and other genes, is mediated exclusively by Shc through a Stat5-independent mechanism (14, 66, 67, 68). Conversely, Stat5 uniquely induces expression of the CIS and CD25 genes (26, 27, 44) and mediates a bcl-2-independent survival signal that cannot be delivered through Shc (38). Perhaps STAT5 acts cooperatively with Shc, or potentially other proliferative signals from the IL-2R, to deliver a signal of maximum mitogenic potential. In vivo, where receptor and/or cytokine concentrations may be limiting, synergy between Stat5- and Shc-mediated mitogenic signals from the IL-2R may be essential for rapid T cell expansion, and hence the generation of an effective immune response to an invading pathogen.
Acknowledgments
We thank David Hockenbery, Lee James, and Pierre Rollini for providing Northern blot probes; James Ihle for providing Stat5 cDNAs; and Mark Benson, Laurel Hickock, Meghan Parsons, and Ben Wittenstein for technical assistance.
Footnotes
-
↵1 This work was supported by Grant GM57931 from the National Institutes of Health. J.D.L. was supported by a National Defense Science and Engineering Graduate fellowship from the U.S. Department of Defense and by a grant from the Poncin Scholarship Fund.
-
↵2 Address correspondence and reprint requests to Dr. Brad Nelson, Virginia Mason Research Center, 1201 9th Avenue, Seattle, WA 98101-2795.
-
↵3 Abbreviations used in this paper: TAD, trans-activation domain; wt, wild type.
- Received September 3, 1999.
- Accepted December 10, 1999.
- Copyright © 2000 by The American Association of Immunologists