The IL-2 Promotes Lymphocyte Proliferation and Induction of the c-, bcl-2, and bcl-x Through the trans-Activation Domain of Stat51

James D. Lord,*†‡ Bryan C. McIntosh,* Philip D. Greenberg,†‡§ and Brad H. Nelson2*‡

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 on the IL-2R ␤-chain (IL-2R␤). Therefore, to assess the role of Stat5 in mitogenic signaling independent of any redun- dant signals, all cytoplasmic tyrosines were deleted from IL-2R␤ 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 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 Journal of Immunology, 2000, 164: 2533–2541.

he T cell mitogen IL-2 binds to a heterotrimeric receptor Stat5 refers to either of two highly homologous members of the complex consisting of IL-2R␣, IL-2R␤, and ␥c chains (1), larger Stat family of signal-transducing activators of , T with the consequent heterodimerization of IL-2R␤ and ␥c Stat5a and Stat5b (16–18). Stat activation conventionally com- inducing the proliferative signal (2, 3). ␥c contributes to prolifer- mences with selective binding of a Stat SH2 domain to a phos- ative signaling by recruiting the Jak3(4–8), the phorylated tyrosine motif on a ligated receptor (19–22), approxi- activation of which also requires a membrane-proximal S region of mating the Stat molecule with a receptor-associated, activated IL-2R␤ (9, 10). Additionally, at least one of three cytoplasmic which phosphorylates a key tyrosine residue on the tyrosine phosphorylation sites (Tyr338, Tyr393,orTyr510) distal to Stat molecule (23, 24). These Stat phosphotyrosines preferentially the S region on IL-2R␤ must be present for proliferation, suggest- bind to Stat SH2 domains (21), causing Stat molecules to dimerize, ing that obligate mitogenic signaling molecules interact with these release from receptors, enter the nucleus, and bind palindromic phosphotyrosines (11, 12). Indeed, the adapter molecule Shc in- DNA sequences (22) in certain promoters (25–27), inducing teracts with Tyr338 (11, 13) and can deliver a proliferative signal transcription via a C-terminal trans-activation domain (TAD)3 (14). The other two tyrosines, Tyr510 and Tyr393, provide docking (28–31). Some members of the Stat family also demonstrate less sites for the Stat5 (11), which has led to the conventional activities. Stat3, for example, serves as an adapter hypothesis that Stat5 mediates a proliferative signal parallel to the molecule between PI3 kinase and type I IFN receptors (32). Fur- one involving Shc. However, Shc and Stat5 have very different thermore, Stat1 and, in cooperation with the glucocorticoid recep- biochemical properties, which raises the question of whether they tor, Stat5 mediate the expression of certain genes through a TAD- could indeed generate redundant proliferative signals from IL- independent mechanism (33–35). Thus, not all signaling functions 2R␤. In addition, the fact that a single phosphotyrosine on a re- of Stat molecules involve the conventional trans-activation ceptor chain can interact with multiple signaling molecules, as domain. exemplified by studies of the platelet-derived growth factor recep- A critical role for Stat5 in IL-2-mediated mitogenesis has been tor (15), raises the possibility that a factor other than Stat5 medi- suggested by the failure of TCR-stimulated T cells from mice lack- ates the proliferative signal from Tyr393 and/or Tyr510. ing 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 mito- *Virginia Mason Research Center, Seattle, WA 98101; †Fred Hutchinson Research Center, Seattle, WA 98104; and Departments of ‡Immunology and §Med- genic signaling molecules, such as Shc (11, 14). Indeed, IL-2R␤ icine, University of Washington, Seattle, WA 98195 mutants that fail to activate Stat5 mediate proliferation in cultured Received for publication September 3, 1999. Accepted for publication December cell lines (11, 37, 38), and primary T cells (39). Thus, the inability 10, 1999. of Stat5-null T cells to proliferate may reflect a recently described The costs of publication of this article were defrayed in part by the payment of page essential role for Stat5 in TCR signaling (40) rather than a requisite charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. role in IL-2 signaling. 1 This work was supported by Grant GM57931 from the National Institutes of Health. This report assesses the role of Stat5 in signals derived from the J.D.L. was supported by a National Defense Science and Engineering Graduate fel- IL-2R in cell lines that do not require TCR stimulation for prolif- lowship from the U.S. Department of Defense and by a grant from the Poncin Schol- eration. A receptor mutagenesis and rescue strategy was used to arship 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.

Copyright © 2000 by The American Association of Immunologists 0022-1767/00/$02.00 2534 IL-2 MEDIATES PROLIFERATION THROUGH THE TAD OF Stat5 study signaling by Stat5 in the absence of any redundant signals resed on acrylamide gels and transferred to nitrocellulose. Nitrocellulose from Shc or other molecules potentially associating with IL-2R␤. blots were blocked with 0.1 M Tris base (pH 7.5), 0.9% sodium chloride, We demonstrate that Tyr510 of IL-2R␤ mediates proliferation 0.05% Tween 20 (TTBS) containing 1% BSA (for antiphosphotyrosine probes) or 5% powdered skim milk (Carnation, Glendale, CA), and probed through the activation of Stat5 and that either the Stat5a or the with rabbit antisera recognizing Jak1 (Santa Cruz Biotechnology) or Jak3 Stat5b isoform can transduce the mitogenic signal. Furthermore, (Upstate Biotechnology, Lake Placid, NY) or murine Abs recognizing proliferative signaling by Stat5 is critically dependent on its TAD, phosphotyrosine (4G10; Upstate Biotechnology) or Stat5 (Transduction indicating that Stat5a mediates proliferation through its conven- Laboratories, Lexington, KY). Blots were then washed with TTBS, probed with peroxidase-conjugated goat anti-rabbit or anti-mouse Abs (Life Tech- tional role as a transcription factor. Indeed, IL-2-induced transcrip- nologies), and washed again with TTBS. Bound Abs were detected by tion of the proliferative and survival genes c-myc, bcl-2, and bcl-x enhanced chemiluminescence (DuPont NEN, Boston, MA). Blots were is mediated by Stat5a through the TAD. stripped between probings with a 30-min, 50°C incubation in 62.6 mM Tris-HCl (pH 6.7), 0.1 M ␤-mercaptoethanol, and 2% SDS. Materials and Methods Plasmid construction Electrophoretic mobility shift assays Expression vectors encoding the chimeric ␣␥- and ␤␤-chains (formerly Cells stimulated as indicated were washed once with buffer H (20 mM denoted GM␣/2␥ and GM␤/2␤) under the control of the human ␤-actin HEPES (pH 7.9), 1 mM EDTA, 0.1 mM EGTA, 2 mM magnesium chlo- promoter have been described previously (2, 9, 14, 41). Mutants of ␤␤ ride, 1 mM sodium o-vanadate, 20 mM sodium fluoride, 1 mM DTT, 0.1 were generated by annealing sense and antisense oligonucleotides encod- mM 4-(2-aminoethyl)benzenesulfonyl fluoride, and 1 mg/ml leupeptin) ing novel C-terminal sequences for ␤␤ and/or premature stop codons, and and lysed in buffer H plus 0.2% Nonidet P-40 at 0°C. Nuclei were pelleted cloning overhanging ends of these annealed primers between a unique AflII by centrifugation and extracted with buffer K (buffer H plus 0.42 M sodium site in the cytoplasmic domain of ␤␤ and a unique XbaI site immediately chloride and 20% v/v glycerol). To generate a probe for Stat activity, 3Ј to the stop codon of ␤␤. For analyses in BA/FG cells, the full length, incompletely overlapping oligonucleotides corresponding to the sense and nonchimeric human IL-2R␤ cDNA was cloned into the expression vector antisense strands of the Stat-responsive DNA element from the Fc␥R1 used above, and the C terminus of this chain, between the AflII site and a promoter were annealed, radiolabeled with [␣-32P]dCTP by an end-filling unique 3Ј ScaI site, was then replaced with the respective mutated se- T4 polymerase reaction, and purified with a MicroSpin G-25 column (Phar- quences. The ⌬713 mutant of Stat5a was generated by PCR with an anti- macia, Piscataway, NJ). The probe was added to nuclear extracts in 50 mM sense oligonucleotide encoding residues 708 to 712 of Stat5a followed by potassium chloride, 10 mM HEPES (pH 7.9), 10% glycerol, 1 mM DTT, the flag epitope, a stop codon, and an EcoRV site. This PCR product was and 87.5 mg/ml dITP/dCTP at room temperature for 30 min in the presence then cut with XhoI and EcoRV enzymes and ligated between these same or absence of either an Ab recognizing the flag epitope (Sigma) or a com- sites in a vector containing murine Stat5a, described previously (29). ⌬713, bination of Abs recognizing Stat5a and Stat5b (Santa Cruz Biotechnology). as well as C-terminally flag epitope-tagged wild-type (wt) murine Stat5a Reaction mixtures were electrophoresed on a nonreducing 0.25ϫ Tris- and Stat5b, were next excised with EcoRI and HindIII and cloned between buffered EDTA-acrylamide gel which was then dried and subjected to SalI and HindIII sites in the ␤-actin promoter-driven plasmid used above autoradiography. for receptor expression. All mutated Stat5a and IL-2R␤ regions were se- quenced with the ABI Prism dye terminator cycle sequencing (Perkin- Elmer, Norwalk, CT). Northern blots Cell culture and transfection Cells stimulated as indicated were pelleted by centrifugation and flash- frozen in a dry ice-ethanol bath. RNA was harvested from thawed pellets The murine T cell line CTLL-2 was obtained from the American Type with the RNA Stat 60 kit (Tel Test, Friendswood, TX); denatured for 10 Culture Collection (Manassas, VA) and maintained as described previously min at 65°C in 20 mM MOPS, 5 mM sodium acetate, 0.5 mM EDTA, 2.4 (14). BA/F3 pro-B cells were obtained from Immunex (Seattle, WA) and M formaldehyde, and 50% formamide; and run on a 1.2% agarose gel maintained in RPMI supplemented with 10% FCS, 2 mM L-glutamine, (containing 20 mM MOPS, 5 mM sodium acetate, 0.5 mM EDTA, and 1.1 50U/ml penicillin, and 50 mg/ml streptomycin, as well as 10% WEHI3- M formaldehyde) in 20 mM MOPS, 5 mM sodium acetate, and 0.5 mM conditioned medium as a source of murine IL-3. To grow BA/F3 cells in EDTA at pH 7.0. RNA was passively transferred to Zetabind membranes the absence of Stat5 activation, the cells were transfected with a chimeric (Cuno, Meriden, CT) with 10ϫ SSC (1.5 M sodium chloride, 0.15 M G-CSFR/gp130 receptor chain described previously (42) and maintained sodium citrate, pH 7.0), and UV cross-linked. Blots were prehybridized at with 100 ng/ml recombinant human G-CSF (Amgen, Thousand Oaks, CA) 43°C in hybridization buffer (1 M sodium phosphate (pH 7.1), 2 mM instead of IL-3. The resultant cell line is referred to hereafter as BA/FG. EDTA, 2% BSA, 10% SDS, 50% formamide, and 0.16 mg/ml yeast tRNA Linearized plasmids were introduced into cells by electroporation, and or herring sperm DNA). To generate nucleic acid probes, a 0.4-kb PstI transfectants were selected for resistance to G418 (Life Technologies, fragment of murine c-myc, a 0.9-kb PstI fragment of murine bcl-2, a 1-kb Gaithersburg, MD) in 96-well plates at limiting dilution to isolate inde- EcoRI fragment of murine bcl-x, and a 1.2-kb PstI fragment of murine pendent subclones. Receptor expression was assessed by flow cytometry GAPDH cDNA were radiolabeled with [␣-32P]dCTP using a random- with Abs to human GM-CSFR␣ or ␤c (Santa Cruz Biotechnology, Santa primed labeling kit (Boehringer Mannheim, Indianapolis, IN) and purified Cruz, CA), or to IL-2R␤ (PharMingen, San Diego, CA). Stat5 expression with Centri-Sep spin columns (Princeton Separations, Adelphia, NJ). was assessed by Western blot with an Ab to Stat5 (Transduction Labs, Probes were boiled for 5 min and added to blots in hybridization buffer. Lexington, KY), or by flow cytometry of cells stained intracellularly with After overnight incubation at 43°C, blots were washed 2–3 times with 2ϫ an Ab to the flag epitope (Sigma, St. Louis, MO). Subclones with compa- SSC, 0.1% SDS; once with 0.2ϫ SSC, 0.1% SDS at room temperature; and rable receptor and/or Stat5 expression were chosen for further analyses. 0–2 times with 0.2ϫ SSC, 0.1% SDS at 55°C before autoradiography. Blots were stripped between probings with a 2-min immersion in boiling Proliferative assays water. 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 ␮Ci/well) present during the last 4 h. Results Cells were harvested onto glass fiber filters, and DNA synthesis was quan- Proliferative signaling by Tyr510 of IL-2R␤ correlates tified by liquid scintillation counting. The data presented in Figs. 1–4 used quantitatively with Stat5 activation 100 ng/ml GM-CSF or 3000 U/ml IL-2, which were found to elicit a - imal response from all functional receptor mutants in CTLL-2 or BA/F3 Structure/function analyses of the cytoplasmic domain of IL-2R␤ cells, respectively. were performed in the IL-2-dependent murine T cell line CTLL-2. Immunoprecipitaion and immunoblotting To avoid activation of the endogenous IL-2R, we introduced a previously described chimeric GM-CSF/IL-2 receptor consisting Jak1 and Jak3 were immunoprecipitated and subjected to kinase assays as ␣␥ ␤␤ described previously (9). Whole cell lysates were generated by boiling cells of two chains, and , containing the extracellular domains of in 62.5 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, 50 mM DTT, and the human GM-CSF receptor ␣- and ␤-chains fused, respectively, 0.1% bromphenol blue. Lysates or immunoprecipitates were electropho- to the transmembrane and intracellular regions of ␥c and IL-2R␤. The Journal of Immunology 2535

FIGURE 2. Activation of Stat5 by Tyr510 of IL-2R␤ shows a quantita- tive 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. FIGURE 1. An IL-2 proliferative signal is mediated through a con- 510 B, Schematic depicting point mutations (bold face) introduced C-terminal served 13-residue motif flanking Tyr (Y510). A, Schematic depicting wt to Tyr510 (underlined) in ␤␤⌬325 ϩ Tyr510. C, Stat5 activation in CTLL-2 ␤ ␤␤ and mutant versions of chimeric GM-CSF/IL-2R -chain ( ). TM, trans- cells mediated by mutant chimeric receptors, assessed as per Fig. 1D. D, membrane. B, Thymidine incorporation mediated by the indicated chimeric Thymidine incorporation in CTLL-2 cells mediated by the indicated chi- F receptors. , Individual CTLL-2 clones stimulated with 100 ng/ml human meric receptors, assessed and depicted as per Fig. 1B. GM-CSF; E, the same clones cultured without . Data for each clone are expressed as a percentage of the thymidine incorporated after stimulation with 100 U/ml human IL-2. C, ␤␤⌬325 retains the ability to activate Janus . Representative CTLL-2 clones bearing the indicated ble levels were chosen for further study. The wt version of ␤␤ 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), delivered a proliferative signal in response to GM-CSF similar to or no cytokine (0). The indicated Janus kinases were then immunoprecipi- that mediated by the endogenous IL-2R. By contrast, a version of tated (IP) and either subjected to an in vitro autokinase assay or a Western ␤␤ from which all cytoplasmic tyrosines had been removed by blot for phosphorylated tyrosines (pTyr). The phosphotyrosine Western introduction of a premature stop codon at codon 325 (relative to blots were then stripped and reprobed for Jak1 or Jak3, as indicated, to the human IL-2R␤ sequence) (␤␤⌬325 (Fig. 1A)) failed to mediate assess loading. D, ␤␤⌬325 ϩ Tyr510 activates Stat5 through Tyr510. Rep- proliferation (Fig. 1B), in accordance with prior reports (11, 12, resentative CTLL-2 clones bearing the indicated chimeric receptors were 14). However, both ␤␤wt and ␤␤⌬325 induced phosphorylation deprived of cytokine for 4 h and then stimulated for 10 min with 100 ng/ml and activation of Jak1 and Jak3 (Fig. 1C), confirming that Janus human GM-CSF (G), 100 U/ml human IL-2 (2), or no cytokine (0). Nu- kinase activation, whereas critical for Stat activation, is not suffi- clear extracts were harvested and subjected to an electrophoretic mobility cient for proliferation. shift assay using a Stat5-binding probe derived from the Fc␥R1 promoter. In lanes marked Ab, extracts from GM-CSF-stimulated cells were treated To investigate the role of Stat5 in proliferation, 13 amino acids ␤ 510 with anti-Stat5 Abs to specifically supershift Stat5 complexes. F510, from the C terminus of IL-2R that encompass Tyr , the most Phe510. efficient Stat5-activating tyrosine on IL-2R␤ (11), were attached to the C terminus of the ␤␤⌬325 receptor (␤␤⌬325 ϩ Tyr510 (Fig. 1A)). Like ␤␤wt and the endogenous IL-2R, this receptor induced When coexpressed, ␣␥ and ␤␤ deliver in response to human GM- strong Stat5 activation (Fig. 1D), as well as cell proliferation (Fig. CSF a signal that is biochemically and physiologically indistin- 1B). Thymidine incorporation was 30–40% less than that induced guishable from that induced in the same cell by the wt IL-2 re- by ␤␤wt or the endogenous IL-2R, likely due to the absence of a ceptor (2, 9, 14, 41). A CTLL-2 clone demonstrating stable high concurrent proliferative signal mediated through Shc (11, 14). expression of ␣␥ was generated and transfected with wt or mutated Nonetheless, ␤␤⌬325 ϩ Tyr510 supported the long term prolifer- derivatives of ␤␤ (Figs. 1A and 2B). Subclones of transfectants ation and survival of CTLL-2 cells in culture (Fig. 3, D and E; data were analyzed for receptor expression by flow cytometry (data not not shown). However, if Tyr510 was mutated to in shown), and those expressing both ␣␥ and ␤␤-chains at compara- this receptor context (␤␤⌬325 ϩ Phe510 (Fig. 1A)), neither Stat5 2536 IL-2 MEDIATES PROLIFERATION THROUGH THE TAD OF Stat5

activation nor proliferation was induced (Fig. 1, B and D), indi- cating 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.In- deed, such a molecule has been previously proposed to explain how Tyr510 mediates IL-2R␣ expression (44). If a molecule other than Stat5 mediates the proliferative signal from Tyr510, specific disruption of the interaction between Stat5 and ␤␤⌬325 ϩ Tyr510 should not eliminate mitogenesis. To test this possibility, point mutations were introduced near Tyr510 in ␤␤⌬325 ϩ Tyr510 to specifically disrupt the consensus Stat5-binding motif without eliminating Tyr510. Analysis of Stat5-binding sites in several cy- tokine receptors suggested that the SH2 domain of Stat5 prefer- entially associates with a phosphorylated tyrosine followed by hy- drophobic amino acids one and three positions C-terminal to the tyrosine (i.e., YLSL in the case of IL-2R␤) (Fig. 2A). These res- idues, in the context of ␤␤⌬325 ϩ Tyr510, were point mutated individually or concomitantly to arginine or glutamic acid to gen- erate four mutant receptors (Fig. 2B). Compared with ␤␤⌬325 ϩ Tyr510, these receptors demonstrated a spectrum of Stat5 activation potencies (Fig. 2C) which correlated directly with proliferation (Fig. 2D).

IL-2 delivers a proliferative signal through Stat5a or Stat5b Although designed to specifically disrupt Stat5 binding, these mu- tations near Tyr510 could also disrupt the binding of other putative mitogenic molecules. Therefore, to directly test whether Stat5 de- livers the proliferative signal from Tyr510, the effect of overex- pressing Stat5 on the marginal Stat5-activating potential of ␤␤⌬325 ϩ YRSL was assessed. By mass action, overexpression of Stat5 was expected to overcome the affinity barrier between Stat5 and Tyr510 created by the of Leu511 to arginine. A flag epitope-tagged version of murine Stat5a was cotransfected with ␤␤⌬325 ϩ YRSL into CTLL-2 cells bearing ␣␥, and stable subclones that coexpressed ␣␥, ␤␤, and flag-Stat5a were identified by flow cytometry (data not shown). Expression levels of trans- fected Stat5a were estimated by Western blot to be 5- to 10-fold higher than endogenous Stat5 (Fig. 3A). As intended, overexpres- sion of Stat5a restored the ability of ␤␤⌬325 ϩ YRSL to induce Stat5 DNA-binding activity in response to GM-CSF to levels equal to or greater than that mediated by ␤␤⌬325 ϩ Tyr510 (Fig. 3B). Rescue of Stat5 activation in turn restored the ability of ␤␤⌬325 ϩ YRSL to induce thymidine incorporation (Fig. 3C), and cell pro- liferation (Fig. 3D) and maintain cell viability (Fig. 3E). Thus, the proliferative signal from Tyr510 of IL-2R␤ can be mediated by Stat5a. Overexpression of Stat5a also enhanced the level of Stat5 DNA-binding activity induced by the endogenous IL-2R (Fig. 3B) but had no observable effect on IL-2-mediated proliferation or vi- FIGURE 3. Overexpression of Stat5a or Stat5b rescues the ability of ability of CTLL-2 cells (Fig. 3, D and E). This suggests that the ␤␤⌬325 ϩ YRSL to mediate proliferation. A, Stat5 expression in trans- fected CTLL-2 cells. Whole cell lysates of CTLL-2 clones transfected with a vector encoding ␤␤⌬325 ϩ YRSL either alone (0) or with a vector expression rescues cell expansion mediated by ␤␤⌬325 ϩ YRSL. Repre- encoding either Stat5a (6 lanes on left) or Stat5b (4 lanes on right) were sentative CTLL-2 clones expressing the indicated chimeric receptor/Stat5 assessed for Stat5 content by Western blot using a Stat5-specific Ab. The combinations were cultured for several days in the presence of 100 ng/ml blot was stripped and reprobed with a Jak1-specific Ab to demonstrate even human GM-CSF (GM), 100 U/ml human IL-2 or no cytokine (0). Every 2 loading. B, Stat5 overexpression rescues Stat5 activation by ␤␤⌬325 ϩ days, cells were stained with trypan blue and live cells were counted. Cells YRSL. Representative CTLL-2 clones expressing ␤␤⌬325 ϩ Tyr510 were split as needed to maintain a density of Ͻ5 ϫ 105 cells/ml. All clones (Y510), ␤␤⌬325 ϩ YRSL (ϩ0), or ␤␤⌬325 ϩ YRSL plus the indicated expanded similarly in response to IL-2, and all died within 2 days in the Stat5 isotype were assessed by electrophoretic mobility shift assay for the absence of cytokine (data not shown). E, Stat5 overexpression rescues ability to activate Stat5 as per Fig. 1D. C, Stat5 overexpression rescues viability mediated by ␤␤⌬325 ϩ YRSL. Representative CTLL-2 clones thymidine incorporation mediated by ␤␤⌬325 ϩ YRSL. The ability of were cultured and counted as in D. Cells excluding and incorporating multiple CTLL-2 clones expressing ␤␤⌬325 ϩ Tyr510 or ␤␤⌬325 ϩ trypan blue were counted as live and dead, respectively. All clones re- YRSL with or without the indicated Stat5 isotype to induce thymidine mained comparably viable in the presence of IL-2 and died at a similar rate incorporation was assessed and depicted as in Fig. 1B. D, Stat5 over- in the absence of cytokine (data not shown). The Journal of Immunology 2537

degree of Stat5 activation normally induced by the IL-2R is satu- rating 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 ab- normalities (26, 47–50). Thus, Stat5a and Stat5b may serve non- redundant functions. To determine whether Stat5b can also medi- ate a proliferative signal from Tyr510, a flag epitope-tagged version of Stat5b was cotransfected with ␤␤⌬325 ϩ YRSL, as above. Although we were unable to overexpress Stat5b to the levels achieved with Stat5a (Fig. 3A), Stat5b overexpression nevertheless enhanced the ability of ␤␤⌬325 ϩ YRSL to activate Stat5 (Fig. 3B) induce thymidine incorporation (Fig. 3C), support cell prolif- eration (Fig. 3D), and inhibit cell death (Fig. 3E).

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 transcrip- tion (23, 24, 29). However, the TAD of Stat5 has been reported to be dispensable for some transcriptional events, such as ␤-casein induction mediated by Stat5 in cooperation with the (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 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-2R␤. ⌬713 is a naturally occurring isoform of Stat5a that lacks the TAD (Fig. 4A) but retains all other functional do- mains and therefore can become tyrosine phosphorylated and bind DNA in response to cytokine stimulation (29). If overexpressed in CTLL2 cells, ⌬713 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 ⌬713 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 ␥c chain but do not respond to IL-2 due to the absence of IL-2R␤, which made it unnecessary to use the chimeric ␣␥ and ␤␤ receptor chains used in CTLL-2 cells. Instead, IL-2R␤ mutants were tested using the normal ectodomain of IL-2R␤. Stable transfectants of BA/FG cells were generated that expressed the different IL-2R␤ mutants at similar levels, as assessed by flow cytometry (data

FIGURE 4. Stat5-mediated proliferation requires the TAD of Stat5. A, Schematic depiction of wt or mutant versions of IL-2R␤ and flag epitope- Stat5a variants were stimulated with 3000 U/ml human IL-2 (F), 10% tagged Stat5a used in these studies. B, Stat5 DNA binding activity is en- IL-3-conditioned medium (Œ), or no cytokine (E). Data for each clone are hanced by wtStat5a or ⌬713. Representative BA/FG clones bearing the expressed as a percentage of the thymidine incorporated by that clone when indicated versions of IL-2R␤ and Stat5a were deprived of cytokine for 4 h stimulated with 100 ng/ml human G-CSF. D, IL-2-mediated expansion of and then stimulated for 10 min with 3000 U/ml human IL-2 (2), 10% BA/FG expressing ⌬325 ϩ Tyr510 is inhibited by ⌬713. A representative IL-3-conditioned media (3), or no cytokine (0). Nuclear extracts were then BA/FG clone expressing ⌬325 ϩ Tyr510 with or without ⌬713 was cul- harvested and assessed for Stat5 DNA-binding activity as in Fig. 1D. Lanes tured in the presence of 3000 U/ml human IL-2, 100 ng/ml human G-CSF, marked Ab represent extracts from IL-2-stimulated cells treated with an or no cytokine (0). Cell proliferation was assessed and depicted as in Fig. anti-flag Ab to specifically supershift transfected Stat5. C, Thymidine in- 3D. Cells were split as needed to maintain a density of Ͻ106 cells/ml. All corporation is enhanced by wtStat5a and inhibited by ⌬713. Three BA/FG clones proliferated similarly in response to G-CSF and died at a similar rate clones expressing each of the indicated combinations of IL-2R␤ and/or in the absence of cytokine (data not shown). 2538 IL-2 MEDIATES PROLIFERATION THROUGH THE TAD OF Stat5 not shown). Experiments were performed with a high dose of IL-2 (3000 U/ml) sufficient to saturate IL-2R␤ and ␥c irrespec- tive of IL-2R␣ 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-2R␤. As expected, BA/FG expressing ⌬325 ϩ Tyr510 acti- vated Stat5 and proliferated in response to IL-2, unless Tyr510 was point mutated to phenylalanine (⌬325 ϩ Phe510 (Fig. 4)). Moreover, BA/FG cells expressing ⌬325 ϩ 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 ⌬325 ϩ Tyr510 and the ⌬713 version of Stat5a were generated. Expression of ⌬713 did not appear to im- pose adverse selective pressure on BA/FG cells, inasmuch as 1) cells expressing ⌬713 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) ⌬713 had no effect on the proliferative response of BA/FG cells to G-CSF (Fig. 4D). Similar to wtStat5a, overexpression of ⌬713 dramatically enhanced Stat5 DNA-binding activity induced by ⌬325 ϩ Tyr510, as well as by the endogenous IL-3 receptor (Fig. 4B), 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, ⌬713 severely reduced the ability of either ⌬325 ϩ 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-2R␤ 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 sug- gests 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 inter- nal control for Stat5 activity (26). ⌬325 ϩ YRSL, which induces minimal Stat5 activation, induced c-myc to a moderate level and FIGURE 5. Stat5 mediates c-myc, bcl-2, and bcl-x induction through its only weakly induced CIS, bcl-2, and bcl-x in comparison with the TAD. A and B, BA/FG cells expressing the indicated versions of IL-2R␤ endogenous IL-3 receptor (Fig. 5A). Overexpression of Stat5a with and Stat5a were deprived of cytokine for 8 h and then stimulated with 3000 ⌬325 ϩ YRSL enhanced induction not only of CIS but also of U/ml IL-2, 10% IL-3 conditioned medium, or 100 ng/ml G-CSF for the c-myc, bcl-2, and bcl-x (Fig. 5A), thereby implicating Stat5 in the indicated times. RNA was harvested from cells and assessed by Northern regulation of all four genes. To test the role of the Stat5 TAD in the blot with serial probing for expression of CIS, c-myc, bcl-2, bcl-x, and induction of these genes, the effects of ⌬713 were assessed. Al- GAPDH. C, The ability of BA/FG expressing ⌬325 ϩ Tyr510 (Y510) to though expression of ⌬713 enhanced the ability of ⌬325 ϩ Tyr510 induce c-myc, bcl-2 and bcl-x with (ϩCHX) or without 20 ␮g/ml cyclo- to promote Stat5 DNA-binding activity (Fig. 4B), it prevented in- heximide administered 30 min before cytokine stimulation was assessed as in A and B. duction of the CIS gene by this receptor, as well as the IL-3 re- ceptor (Fig. 5B). Thus, ⌬713 exerted dominant negative inhibition of conventional Stat5 transcriptional activity, consistent with prior assessment of this construct (29). In so doing, ⌬713 also inhibited To determine whether the induction of these genes by Stat5 is an the ability of ⌬325 ϩ Tyr510, as well as the IL-3 receptor, to immediate-early event or requires the de novo synthesis of inter- induce c-myc, bcl-2, and bcl-x (Fig. 5B), indicating that Stat5 reg- mediate transcription factors, the transcriptional consequences of ulates expression of these genes by a TAD-dependent mechanism. ⌬325 ϩ Tyr510 signaling were assessed in the presence of the The Journal of Immunology 2539 synthesis inhibitor cycloheximide. Although cyclohexi- Stat5 is absolutely essential for IL-2-induced proliferation in pri- mide enhanced the basal expression of genes in the absence of mary T cells. It is unclear why no Stat5-independent proliferative cytokine as reported for c-myc (59, 60), IL-2 and IL-3 still induced pathway was operational in T cells from the latter mice, as it ap- expression of c-myc and bcl-x above this basal level (Fig. 5C). By parently was in the cells studied by Fujii et al. (39). One possible contrast, induction of bcl-2, which occurred with slower kinetics explanation for the discrepant results is that the mutant form of than seen with c-myc and bcl-x (Fig. 5, A and B), was blocked by IL-2R␤ used by Fujii et al. might have been able to activate Stat5 cycloheximide treatment (Fig. 5C). Thus, whereas bcl-2 induction to a physiologically relevant but experimentally undetectable level. requires de novo protein synthesis, c-myc and bcl-x are induced by However, this seems an unlikely explanation, in that we show here Stat5 in an immediate-early fashion, suggesting that Stat5 may that even detectable, low level Stat5 activity is insufficient for the interact directly with the promoters of these genes rather than via high level of T cell proliferation observed by Fujii et al. (39) (Figs. an intermediary transcription factor. 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 Discussion essential components of the IL-2R complex, such as Jak3, render- This report investigates the mechanism of mitogenic signaling by ing the receptor globally dysfunctional in the total absence of a single cytoplasmic tyrosine (Tyr510) on IL-2R␤ which has pre- Stat5. There is precedent for such a mechanism, because the basal viously been shown to interact with the transcription factors Stat5a activity of Stat1 has recently been shown to be required for ex- and Stat5b (11, 43). By eliminating all other tyrosines from IL- pression of caspases, which in turn are required to sensitize cells to 2R␤ and then introducing point mutations at or near Tyr510,we TNF-␣-mediated (33). Finally, it is also possible that the have established a quantitative correlation between the level of severe proliferative defect seen in Stat5a/Stat5b-deficient T cells in Stat5 activation mediated by IL-2R␤ and subsequent proliferation part reflects impaired signaling through the TCR, which has re- of the murine T cell line CTLL-2. Moreover, overexpression of cently been show to induce transient activation of Stat5 (40). Stat5a or Stat5b rescued the proliferative signal mediated by a The molecular mechanism by which Stat5 mediates proliferative receptor mutant with a low affinity Stat5-binding site, thereby di- signaling has also remained obscure. A truncated version of Stat5, rectly implicating Stat5a and Stat5b as mediators of mitogenic sig- which was unable to dimerize and hence bind DNA, inhibited IL- naling in T cells. Proliferative signaling by Stat5a required a C- 3-mediated proliferation by a dominant negative mechanism (51), terminal TAD, which was similarly required for Stat5 to induce thus implicating Stat5 in the IL-3 proliferative signal. However, in expression of the c-myc, bcl-x, and bcl-2 genes, indicating that a second study by Wang et al. (29), a less severely truncated ver- Stat5 mediates proliferation by trans-activation of both promito- sion of Stat5 lacking only the TAD had no impact on the IL-3- genic and antiapoptotic genes. proliferative signal. Together, these studies suggested that Stat5 Previous studies of the role of Stat5 in proliferative signaling may mediate proliferative signaling by a TAD-independent mech- have yielded somewhat contradictory results. Mui et al. (51) re- anism. Consistent with this notion, a version of Stat5 capable of ported that a dominant-negative mutant of Stat5 significantly in- binding DNA, but lacking its TAD, can nevertheless interact with hibited the proliferation of the pro-B cell line BA/F3 in response to the glucocorticoid receptor to induce transcription of the ␤-casein IL-3, another cytokine which activates Stat5. This occurred in the gene (34, 35), thereby demonstrating that Stat5 is capable of TAD- face of normal c-myc induction, suggesting that Stat5 was critical independent transcriptional regulation. However, our results define for some, but not all, aspects of the proliferative signal. By con- a clear requirement for the C-terminal TAD of Stat5 for the pro- trast, Wang et al. (29) reported that dominant-negative inhibition liferative response of lymphocytes to both IL-2 and IL-3. This of Stat5 activity in the promyeloid cell line 32D had no impact on indicates that, despite its capacity for TAD-independent signaling, proliferative responses to IL-3 and Epo. The residual proliferative Stat5 induces cell proliferation through the conventional TAD at responses seen in these two studies may have been mediated by its C terminus. The discrepancy between our findings and those of one or more redundant proliferative pathways operating in parallel Wang et al. may reflect the fact that their cells were cultured with with Stat5, such as the Shc pathway (14, 61). Indeed, the Stat5- IL-3 in the face of constitutive dominant negative Stat5 expression, binding sites on IL-2R␤ (Tyr510 and Tyr393) are only critical for which would be expected to impose a strong selective pressure in proliferation when Shc activity is abrogated by mutation of its favor of Stat5-independent signaling mechanisms. In the present binding site at Tyr338 (11). Consequently, ⌬713 or other dominant study, we avoided this potential problem by creating a cell line, negative versions of Stat5 would not be expected to block prolif- BA/FG, that does not use Stat5 for growth and instead responds to eration mediated by full length IL-2R␤. For this reason, the ex- a Stat3-dependent proliferative signal delivered by the gp130 re- periments in the present study utilized a mutated version of IL- ceptor subunit. This allowed for expression of the TAD-deficient 2R␤ that lacked the Shc-binding site and contained only the Stat5 form of Stat5 at a level sufficient for complete dominant negative activation site at Tyr510. By eliminating potentially redundant activity, which likely accounts for the more pronounced impact on pathways, a definitive role for Stat5 in proliferative signaling was IL-2- and IL-3-mediated proliferation observed in this study com- revealed. pared with Wang et al. (29) and Mui et al. (51). Although both Stat5 and Shc are clearly able to mediate cell In addition to cell proliferation, Stat5 was also found to induce proliferation, the extent to which each contributes to the IL-2- expression of the growth-related genes c-myc, bcl-2, and bcl-x induced proliferation of primary T cells remains controversial. Fu- through a TAD-dependent mechanism (Fig. 5, A and B). Induction jii et al. (39) reconstituted IL-2R␤-deficient mice with a mutated of the c-myc and bcl-x genes by Stat5 did not require de novo version of IL-2R␤ lacking the ability to detectably activate Stat5, protein synthesis (Fig. 5C), suggesting that Stat5a interacts directly but containing an intact Shc-binding site. They found that this with the promoters of these genes. Although no Stat5 binding sites mutated receptor promoted a normal proliferative response to IL-2 have yet been functionally defined in these promoters, other Stat in primary T cells (39), consistent with prior studies in cell lines family members have been shown to regulate these genes. Stat1 (11, 37, 38, 62). By contrast, Moriggl et al. reported that T cells can interact with a TTCGGAGAA motif in the human bcl-x pro- from mice lacking both the Stat5A and Stat5B genes show no moter, and this site is involved in the regulation of bcl-x expression proliferative response to IL-2 (36), which led them to conclude that (63). Stat3 has also been proposed to regulate bcl-x expression (64) 2540 IL-2 MEDIATES PROLIFERATION THROUGH THE TAD OF Stat5 and can associate with and trans-activate the c-myc promoter (65). 9. Nelson, B. H., B. C. McIntosh, L. L. Rosencrans, and P. D. Greenberg. 1997. However, the site in the c-myc promoter with which Stat3 interacts Requirement for an initial signal from the membrane-proximal region of the receptor ␥ c chain for Janus kinase activation leading to T cell is not appreciably bound by Stat5(65) (J.D.L., unpublished obser- proliferation. Proc. Natl. Acad. Sci. USA 94:1878. vations), suggesting that Stat3 and Stat5 mediate c-myc induction 10. Witthuhn, B. A., O. Silvennoinen, O. Miura, K. S. Lai, C. Cwik, E. T. Liu, and J. N. Ihle. 1994. Involvement of the Jak-3 janus kinase in signalling by interleu- by binding distinct promoter sites. kins 2 and 4 in lymphoid and myeloid cells. Nature 370:153. In contrast to c-myc and bcl-x, induction of bcl-2 by Stat5 ap- 11. Friedmann, M. C., T. S. Migone, S. M. Russell, and W. J. Leonard. 1996. Dif- pears to be a delayed early event, because it occurs with slower ferent interleukin 2 receptor beta-chain tyrosines couple to at least two signaling pathways and synergistically mediate interleukin 2-induced proliferation. Proc. kinetics and requires de novo protein synthesis (Fig. 5). This sug- Natl. Acad. Sci. USA 93:2077. gests that Stat5 induces expression of one or more secondary tran- 12. Goldsmith, M. A., S. Y. Lai, W. Xu, M. C. Amaral, E. S. Kuczek, L. J. Parent, scription factors that in turn bind to and trans-activate the bcl-2 G. B. Mills, K. L. Tarr, G. D. Longmore, and W. C. Greene. 1995. Growth signal transduction by the human interleukin-2 receptor requires cytoplasmic tyrosines promoter. Such a model, in which Stat5 initiates a cascade of gene of the ␤ chain and non-tyrosine residues of the ␥ c chain. J. Biol Chem 270: activation events, would explain how a single transcription factor 21729. could promote the diverse array of cellular changes that are nec- 13. Ravichandran, K. S., V. Igras, S. E. Shoelson, S. W. Fesik, and S. J. Burakoff. 1996. Evidence for a role for the phosphotyrosine binding domain of Shc in essary for cell cycle progression, including protooncogene induc- interleukin-2 signalling. Proc. Natl. Acad. Sci. USA 93:5275. tion, up-regulation of cell metabolism, and activation of the cell 14. Lord, J. D., B. C. McIntosh, P. D. Greenberg, and B. H. Nelson. 1998. The IL-2 cycle machinery. receptor promotes proliferation, bcl-2 and bcl-x induction, but not cell viability through the adapter molecule Shc. J. Immunol. 161:4627. ⌬ ϩ 510 Although the 325 Tyr receptor was able to mediate Stat5 15. Claesson-Welsh, L. 1994. Platelet derived growth factor receptor signals. J. Biol. activation at least as potently as ␤␤wt and the endogenous wt IL-2 Chem. 269:32023. receptor (Fig. 1D), it was ϳ40% weaker in its ability to promote 16. Mui, A. L.-F., H. Wakao, A.-M. O’Farrell, N. Harada, and A. Miyajima. 1995. Interleukin-3, granulocyte-macrophage colony stimulating factor and interleu- cell proliferation (Fig. 1B). This suggests that Stat5 is not sufficient kin-5 transduce signals through two STAT5 homologs. 1995 14:1166. to mediate all aspects of the IL-2 proliferative signal. Indeed, ac- 17. Lin, J.-X., J. Mietz, W. S. Modi, S. John, and W. J. Leonard. 1996. Cloning of tivation of the ras/receptor-activated factor/mitogen-activated pro- human Stat5B. J. Biol. Chem. 271:10738. 18. Hou, J., U. Schindler, W. J. Henzel, S. C. Wong, and S. L. McKnight. 1995. tein kinase signaling pathway, culminating in the induction of c- Identification and purification of human stat activated in response to fos,c-jun, and other genes, is mediated exclusively by Shc through interleukin-2. Immunity 2:321. a Stat5-independent mechanism (14, 66–68). Conversely, Stat5 19. Heim, M. H., I. M. Kerr, G. R. Stark, and J. E. Darnell, Jr. 1995. Contribution of STAT SH2 groups to specific signaling by the Jak-STAT pathway. uniquely induces expression of the CIS and CD25 genes (26, 27, Science 267:1347. 44) and mediates a bcl-2-independent survival signal that cannot 20. Lin, J.-X., T.-S. Migone, M. Tsang, M. Friedmann, J. A. Weatherbee, L. Zhou, A. Yamaguchi, E. T. Bloom, J. Mietz, S. John, et al. 1995. The role of shared be delivered through Shc (38). Perhaps STAT5 acts cooperatively receptor motifs and common Stat proteins in the generation of cytokine pleotropy with Shc, or potentially other proliferative signals from the IL-2R, and redundancy by IL-2, IL-4, IL-7, IL-13, and IL-15. Immunity 2:331. to deliver a signal of maximum mitogenic potential. In vivo, where 21. Greenlund, A. C., M. O. Morales, B. L. Viviano, H. Yan, J. Krolewski, and R. D. Schreiber. 1995. Stat recruitment by tyrosine-phosphorylated cytokine re- receptor and/or cytokine concentrations may be limiting, synergy ceptors: an ordered reversible affinity-driven process. Immunity 2:677. between Stat5- and Shc-mediated mitogenic signals from the 22. Schindler, U., P. Wu, M. Rothe, M. Brasseur, and S. L. McKnight. 1995. Com- IL-2R may be essential for rapid T cell expansion, and hence the ponents of a Stat recognition code: evidence for two layers of molecular selec- tivity. Immunity 2:689. generation of an effective immune response to an invading 23. Darnell, J. E. 1997. STATs and gene regulation. Science 277:1630. pathogen. 24. Ihle, J. N. 1996. STATs: signal transducers and activators of transcription. Cell 84:331. 25. Yoshimura, A., M. Ichihara, I. Kinjyo, M. Moriyama, N. G. Copeland, Acknowledgments D. J. Gilbert, N. A. Jenkins, T. Hara, and A. Miyajima. 1996. Mouse oncostatin We thank David Hockenbery, Lee James, and Pierre Rollini for providing M: an immediate early gene induced by multiple cytokines through the JAK- STAT5 pathway. EMBO J. 15:1055. Northern blot probes; James Ihle for providing Stat5 cDNAs; and Mark 26. Matsumoto, A., M. Masuhara, K. Misui, M. Yokouchi, M. Ohtsubo, H. Mitsawa, Benson, Laurel Hickock, Meghan Parsons, and Ben Wittenstein for tech- A. Miyajima, and A. Yoshimura. 1997. CIS, a cytokine inducible SH2 protein, is nical assistance. a target of the JAK-STAT5 pathway and modulates STAT5 activation. Blood 89:48. 27. John, S., C. M. Robbins, and W. J. Leonard. 1996. An IL-2 in References the human IL-2 receptor alpha chain promoter is a composite element that binds Stat5, Elf-1, HMG-I(Y) and a GATA family protein. EMBO J. 15:5627. 1. Takeshita, T., K. Ohtani, H. Asao, S. Kumaki, M. Nakamura, and K. Sugamura. 28. Minami, M., M. Inoue, S. Wei, K. Takeda, M. Matsumoto, T. Kishimoto, and 1992. An associated molecule, p64 with IL-2 receptor ␤ chain: its possible in- S. Akira. 1996. STAT3 activation is a critical step in gp130-mediated terminal volvement in the formation of the functional intermediate-affinity IL-2 receptor differentiation and growth arrest of a myeloid cell line. Proc. Natl. Acad. Sci. USA complex. J. Immunol. 148:2154. 93:3963. 2. Nelson, B. H., J. D. Lord, and P. D. Greenberg. 1994. Cytoplasmic domains of 29. Wang, D., D. Stravdopodis, S. Teglund, J. Kitazawa, and J. N. Ihle. 1996. Nat- the interleukin-2 receptor ␤ and ␥ chains mediate the signal for T-cell prolifer- urally occurring dominant negative variants of Stat5. Mol. Cell. Biol. 16:6141. ation. Nature 369:333. 3. Nakamura, Y., S. Russell, S. Mess, M. Friedmann, M. Erdos, C. Francois, 30. Azam, M., C. Lee, I. Strehlow, and C. Schindler. 1997. Functionally distinct Y. Jacques, S. Adelstein, and W. Leonard. 1994. Heterodimerization of the IL-2 isoforms of STAT5 are generated by protein processing. Immunity 6:691. receptor ␤- and ␥-chain cytoplasmic domains is required for signalling. Nature 31. Moriggl, R., V. Gouilleux-Gruart, R. Jahne, S. Berchtold, C. Gartmann, X. Liu, 369:330. L. Hennighausen, A. Sotiropoulos, B. Groner, and F. Gouilleux. 1996. Deletion 4. Russell, S., J. Johnston, M. Noguchi, M. Kawamura, C. Bacon, M. Friedmann, of the carboxy-terminal transactivation domain of MGF-Stat5 results in sustained M. Berg, D. McVicar, B. Witthuhn, and O. Silvennoinen. 1994. Interaction of DNA binding and a dominant negative phenotype. Mol. Cell. Biol. 16:5691. IL-2R ␤ and ␥ c chains with Jak1 and Jak3: implications for XSCID and XCID. 32. Pfeffer, L. M., J. E. Mullersman, S. R. Pfeffer, A. Murti, W. Shi, and C. H. Yang. Science 266:1042. 1997. STAT3 as an adapter to couple phosphatidylinositol 3-kinase to the IF- 5. Miyazaki, T., A. Kawahara, H. Fujii, Y. Nakagawa, Y. Minami, Z. Liu, I. Oishi, NAR1 chain of the type I interferon receptor. Science 276:1418. O. Silvennoinen, B. Witthuhn, J. Ihle, et al. 1994. Functional activation of Jak1 33. Kumar, A., M. Commane, T. W. Flickinger, C. M. Horvath, and G. R. Stark. and Jak3 by selective association with IL-2 receptor subunits. Science 266:1045. 1997. Defective TNF-␣-induced apoptosis in STAT1-null cells due to low con- 6. Kawahara, A., Y. Minami, T. Miyazaki, J. N. Ihle, and T. Taniguchi. 1995. stitutive levels of caspases. Science 278:1630. Critical role of the interleukin 2 (IL-2) receptor ␥-chain-associated Jak3 in the 34. Lechner, J., T. Welte, J. K. Tomasi, P. Bruno, C. Cairns, J.-A. Gustafsson, and IL-2-induced c-fos and c-myc, but not bcl-2, gene induction. Proc. Natl. Acad. W. Doppler. 1997. Promoter-dependent synergy between glucocorticoid receptor Sci. USA 92:8724. and Stat5 in the activation of ␤-casein gene transcription. J. Biol. Chem. 272: 7. Park, S. Y., K. Saijo, K. Takahashi, M. Osawa, H. Arase, N. Hirayama, 20954. K. Miyake, H. Nakauchi, T. Shirasawa, and T. Saito. 1995. Developmental de- 35. Stoecklin, E., M. Wissler, R. Moriggl, and B. Groner. 1997. Specific DNA bind- fects of lymphoid cells in Jak3 kinase-deficient mice. Immunity 3:771. ing of Stat5, but not of glucocorticoid receptor, is required for their functional 8. Thomis, D. C., C. B. Gurniak, E. Tivol, A. H. Sharpe, and L. J. Berg. 1995. cooperation in the regulation of gene transcription. Mol. Cell. Biol. 17:6708. Defects in B lymphocyte maturation and T lymphocyte activation in mice lacking 36. Moriggl, R., C. J. Topham, S. Teglund, V. Sexl, C. McKay, D. Wang, Jak3. Science 270:794. A. Hoffmeyer, J. van Deursen, M. Y. Sangster, K. D. Bunting, et al. 1999. Stat5 The Journal of Immunology 2541

is required for IL-2-induced cell cycle progression of peripheral T cells. Immunity 52. Zhong, Z., Z. Wen, and J. E. Darnell. 1994. Stat3: a STAT family member 10:249. activated by tyrosine phosphorylation in response to epidermal growth factor and 37. Fujii, H., Y. Nakagawa, U. Schindler, A. Kawahara, H. Mori, F. Gouilleux, interleukin-6. Science 264:95. B. Groner, J. Ihle, Y. Minami, T. Miyazaki, et al. 1995. Activation of Stat5 by 53. Shibuya, H., M. Yoneyama, J. Ninomiya-Tsuji, K. Matsumoto, and T. Taniguchi. interleukin 2 requires a carboxyl-terminal region of the interleukin 2 receptor ␤ 1992. IL-2 and EGF receptors stimulate the hematopoietic cell cycle via different chain but is not essential for the proliferative signal transmission. Proc. Natl. signaling pathways: demonstration of a novel role for c-myc. Cell 70:57. Acad. Sci. USA 92:5482. 54. Heikkila, R., G. Schwab, E. Wickstrom, S. L. Loke, D. H. Pluznik, R. Watt, and 38. Zamorano, J., H. Y. Wang, R. Wang, Y. Shi, G. D. Longmore, and A. D. Keegan. L. M. Neckers. 1987. A c-myc antisense oligodeoxynucleotide inhibits entry into 1998. Regulation of cell growth by IL-2: role of Stat5 in protection from apo- S phase but not progress from G0 to G1. Nature 328:445. ptosis but not in cell cycle progression. J. Immunol. 160:3502. 55. Akbar, A. N., N. J. Borthwick, R. G. Wickremasinghe, P. Panayiotidis, D. Pilling, 39. Fujii, H., K. Ogasawara, H. Otsuka, M. Suzuki, K.-I. Yamamura, T. Yokochi, M. Bofill, S. Krajewski, J. C. Reed, and M. Salmon. 1996. Interleukin-2 receptor T. Miyazaki, H. Suzuki, T. W. Mak, S. Taki, et al. 1998. Functional dissection of common gamma chain signaling cytokines regulate activated T cell apoptosis in the cytoplasmic subregions of the IL-2 receptor ␤ c chain in primary lymphocyte response to growth factor withdrawal: selective induction of anti-apoptotic (bcl-2, populations. EMBO J. 17:6551. bcl-xL) but not pro-apoptotic (bax, bcl-xS) . Eur. J. Immunol. 40. Welte, T., D. Leitenberg, B. N. Dittel, B. K. al-Ramadi, B. Xie, Y. E. Chin, 26:294. C. A. Janeway, Jr., A. L. M. Bothwell, K. Bottomly, and X.-Y. Fu. 1999. STAT5 56. Deng, G., and E. R. Podack. 1993. Suppression of apoptosis in a cytotoxic T-cell interaction with the T cell receptor complex and stimulation of T cell prolifera- line by interleukin-2-mediated gene transcription and deregulated expression of tion. Science 283:222. the protooncogene bcl-2. Proc. Natl. Acad. Sci. USA 90:2189. 41. Nelson, B. H., J. D. Lord, and P. D. Greenberg. 1996. A membrane-proximal 57. Boise, L. H., A. J. Minn, P. J. Noel, C. H. June, M. A. Accavitti, T. Lindsten, and ␥ region of the interleukin-2 receptor c chain sufficient for Jak kinase activation C. B. Thompson. 1995. CD28 costimulation can promote T cell survival by and induction of proliferation in T cells. Mol. Cell. Biol. 16:309. enhancing the expression of bcl-xL. Immunity 3:87. 42. Ziegler, S. F., T. A. Bird, K. M. Morella, B. Mosley, D. P. Gearing, and 58. Miyazaki, T., Z.-J. Liu, A. Kawahara, Y. Minami, K. Yamada, Y. Tsujimoto, H. Baumann. 1993. Distinct regions of the human granulocyte-colony-stimulat- E. L. Barsoumian, R. M. Perlmutter, and T. Taniguchi. 1995. Three distinct IL-2 ing factor receptor are required for proliferation and gene induction. Mol. Cell. signaling pathways mediated by bcl-2, c-myc, and Ick cooperate in hematopoietic Biol. 13:2384. cell proliferation. Cell 81:223. 43. Gaffen, S. L., S. Y. Lai, M. Ha, X. Liu, L. Hennighausen, W. C. Greene, and 59. Kelley, K., B. H. Cochran, S. C. D., and P. Leder. 1983. Cell-specific regulation M. A. Goldsmith. 1996. Distinct tyrosine residues within the interleukin-2 re- of the c-myc gene by lymphocyte mitogens and platelet-derived growth factor. ceptor ␤ chain drive signal transduction specificity, redundancy, and diversity. Cell 35:603. J. Biol. Chem. 271:21381. 60. Dani, C., J. M. Blanchard, M. Piechaczyk, S. El Sabouty, L. Marty, and 44. Ascherman, D. P., T.-S. Migone, M. C. Friedmann, and W. J. Leonard. 1997. P. Jeanteur. 1984. Extreme instability of myc mRNA in normal and transformed Interleukin-2 (IL-2)-mediated induction of the IL-2 receptor ␣ chain gene. J. Biol. human cells. Proc. Natl. Acad. Sci. USA 81:7046. Chem. 272:8704. 45. Boucheron, C., S. Dumon, S. C. Santos, R. Moriggl, L. Hennighausen, 61. Gotoh, N., M. Toyoda, and M. Shibuya. 1997. Tyrosine phosphorylation sites at S. Gisselbrecht, and F. Gouilleux. 1998. A single in the DNA-binding amino acids 239 and 240 of Shc are involved in epidermal growth factor-induced regions of STAT5A and STAT5B confer distinct DNA-binding specificities. mitogenic signalling that is distinct from ras/mitogen-activated protein kinase J. Biol. Chem. 273:33936. activation. Mol. Cell. Biol. 17:1824. 62. Hatakeyama, M., H. Mori, T. Doi, and T. Taniguchi. 1989. A restricted cyto- 46. Verdier, F., R. Rabionet, F. Gouilleux, C. Beisenherz-Huss, P. Varlet, O. Muller, ␤ P. Mayeux, C. Lacombe, S. Gisselbrecht, and S. Chretien. 1998. A sequence of plasmic region of IL-2 receptor chain is essential for growth signal transduction the CIS gene promoter interacts preferentially with two associated STAT5A but not for binding and internalization. Cell 59:837. dimers: a distinct biochemical difference between STAT5A and STAT5B. Mol. 63. Fujio, Y., K. Kunisada, H. Hirota, K. Yamauchi-Takihara, and T. Kishimoto. Cell. Biol. 18:5852. 1997. Signals through gp130 upregulate bcl-x gene expression via Stat1-binding 47. Udy, G. B., R. P. Towers, R. G. Snell, R. W. Wilkins, S.-H. Park, P. A. Ram, cis-element in cardiac myocytes. J. Clin. Invest. 99:2892. D. J. Waxman, and H. W. Davey. 1997. Requirement of STAT5b for sexual 64. Catlett-Falcone, R., T. H. Landowski, M. M. Oshiro, J. Turkson, A. Levitzski, dimorphism of body growth rates and liver gene expression. Proc. Natl. Acad. S. R., G. Ciliberto, L. Moscinski, J. L. Fernandez-Luna, G. Nunez, W. S. Dalton, Sci. USA 94:7239. et al. 1999. Constitutive activation of Stat3 signaling confers resistance to apo- 48. Nakajima, H., X.-W. Liu, A. Wynshaw-Boris, L. A. Rosenthal, K. Imada, ptosis in human U266 myeloma cells. Immunity 10:105. D. S. Finbloom, L. Hennighausen, and W. J. Leonard. 1997. An indirect effect of 65. Kiuchi, N., K. Nakajima, M. Ichiba, T. Fukada, M. Narimatsu, K. Mizuno, Stat5a in IL-2-induced proliferation: a critical role for Stat5a in IL-2-mediated M. Hibi, and T. Hirano. 1999. Stat3 is required for the gp130-mediated full IL-2 receptor alpha chain induction. Immunity 7:691. activation of the c-myc gene. J. Exp. Med. 189:63. 49. Imada, K., E. T. Bloom, H. Nakajima, J. A. Horvath-Arcidiacono, G. B. Udy, 66. Satoh, T., Y. Minami, T. Kono, K. Yamada, A. Kawahara, T. Taniguchi, and H. W. Davey, and W. J. Leonard. 1998. Stat5b is essential for natural killer Y. Kaziro. 1992. Interleukin 2-induced activation of ras requires two domains of cell-mediated proliferation and cytolytic activity. J. Exp. Med. 188:2067. interleukin 2 receptor ␤ subunit, the essential region for growth stimulation and 50. Teglund, S., C. McKay, E. Schuetz, J. M. v. Deursen, D. Stravopodis, D. Wang, lck-binding domain. J. Biol Chem 267:25423. M. Brown, S. Bodner, G. Grosveld, and J. N. Ihle. 1998. Stat5a and Stat5b 67. Hatakeyama, M., A. Kawahara, H. Mori, H. Shibuya, and T. Taniguchi. 1992. proteins have essential and nonessential, or redundant, roles in cytokine re- c-fos gene induction by interleukin 2: identification of the critical cytoplasmic sponses. Cell 93:841. regions within the interleukin 2 receptor ␤ chain. Proc. Natl. Acad. Sci. USA 51. Mui, A. L., H. Wakao, T. Kinoshita, T. Kitamura, and A. Miyajima. 1996. Sup- 89:2022. pression of interleukin-3-induced gene expression by a C-terminal truncated 68. Bonfini, L., E. Miggliaccio, G. Pelicci, L. Lanfrancone, and P. G. Pelicci. 1996. Stat5: role of Stat5 in proliferation. EMBO J. 15:2425. Not all Shc’s roads lead to ras. Trends Biol. Sci. 21:257.