Growth Signal Transduction by the Human Interleukin-2 Receptor Requires Cytoplasmic Tyrosines of the β Chain and Non-tyrosine Residues of the γc Chain

To evaluate the possible role for receptor-based tyrosine phosphorylation in growth signaling induced by interleukin-2 (IL-2), a series of substitution tyrosine mutants of the IL-2 receptor β and γc chains was prepared and analyzed. Concurrent mutation of all six of the cytoplasmic tyrosines present in the β chain markedly inhibited IL-2-induced growth signaling in both pro-B and T cell lines. Growth signaling in a pro-B cell line was substantially reconstituted when either of the two distal tyrosines (Tyr-392, Tyr-510) was selectively restored in the tyrosine-negative β mutant, whereas reconstitution of the proximal tyrosines (Tyr-338, Tyr-355, Tyr-358, Tyr-361) did not restore this signaling function. Furthermore, at least one of the two cytoplasmic tyrosines that is required for β chain function was found to serve as a phosphate acceptor site upon induction with IL-2. Studies employing a chimeric receptor system revealed that tyrosine residues of the β chain likewise were important for growth signaling in T cells. In contrast, although the γc subunit is a target for tyrosine phosphorylation in vivo, concurrent substitution of all four cytoplasmic tyrosines of this chain produced no significant effect on growth signaling by chimeric IL-2 receptors. However, deletion of either the Box 1, Box 2, or intervening (V-Box) regions of γc abrogated receptor function. Therefore, tyrosine residues of β but not of γc appear to play a pivotal role in regulating growth signal transduction through the IL-2 receptor, either by influencing cytoplasmic domain folding or by serving as sites for phosphorylation and subsequent association with signaling intermediates. These findings thus highlight a fundamental difference in the structural requirements for IL-2Rβ and γc in receptor-mediated signal transduction.

Interleukin-2 (IL-2) 1 is a helical cytokine that induces the proliferation of T and B lymphocytes as well as the expression of a number of immune effector functions by binding to the heterotrimeric IL-2 receptor complex (IL-2R). The 70 -75-kDa ␤ (IL-2R␤) and 64-kDa ␥ (␥ c ) subunits of the IL-2R share structural homology with other members of a cytokine receptor superfamily (1) and together form a receptor complex that is competent to bind IL-2 with intermediate affinity and to transduce growth and differentiation signals (reviewed in Ref. 2). As in other receptor systems, evidence has accumulated indicating that signal transduction is initiated upon ligand-induced heterodimerization of the ␤ and ␥ c cytoplasmic tails (3,4). Interestingly, IL-2R␤ is also employed in the receptor for IL-15 (5,6), whereas ␥ c participates in the formation of the receptors for IL-4 (7), IL-7 (8,9), IL-9 (10), and IL-15 (6).
Among the earliest biochemical changes induced by ligation of the IL-2 receptor is activation of cytoplasmic tyrosine kinases resulting in the phosphorylation of certain recognized and unrecognized cellular substrates. The biologic relevance of IL-2-induced tyrosine kinase activity is supported by the finding that selective tyrosine kinase inhibitors (herbimycin A and genistein) concomitantly block these intracellular phosphorylation events as well as growth signal transduction (11,12). Although none of the known IL-2R subunits contain recognizable kinase catalytic domains, tyrosine kinase activity has been coimmunoprecipitated with the IL-2R (13)(14)(15)(16)(17)(18). Recent evidence indicates that the Janus kinases JAK1 and JAK3 (19 -21) as well as various src family kinases (13,15,18,22) are among the signaling molecules that are physically and functionally linked to the IL-2R. However, the specific role of each of these kinases and their substrates in IL-2R signal transduction remains to be defined.
Like many growth factor receptors containing intrinsic tyrosine kinase activity (for review, see Ref. 23), the cytoplasmic domains of the ␤ and ␥ subunits of the interleukin-2 receptor itself undergo inducible tyrosine phosphorylation upon engagement by IL-2 (24 -26). The biological significance of such receptor phosphorylation is poorly defined for cytokine receptors lacking intrinsic tyrosine kinase activity. Since the IL-2 receptor itself is a major substrate of tyrosine phosphorylation following the binding of IL-2, the present investigation was undertaken to determine the potential regulatory role played by the cytoplasmic tyrosine residues of the IL-2R␤ and ␥ c subunits. Our results demonstrate that tyrosines within the cyto-plasmic tail of IL-2R␤ are critical for full growth signaling in pro-B and T cells. In contrast, the tyrosine residues of the ␥ c chain are dispensable for this function, revealing an important distinction between the IL-2R␤ and ␥ c subunits. These findings, along with a delineation of essential membrane-proximal domains of ␥ c , may have general implications for the functional design of cytokine receptors, particularly those employing the common ␥ c subunit.

MATERIALS AND METHODS
Cell Lines-The cell line BA/F3 (27), an IL-3-dependent murine pro-B cell line, was maintained as described previously (28). Supernatant from WEHI-3 cells (ATCC) was used as a source of IL-3. HT-2, an IL-2-dependent murine helper T cell line (ATCC), was maintained in RPMI 1640 supplemented with 10% fetal bovine serum, 55 M ␤-mercaptoethanol, 2 mM L-glutamine, and 200 units/ml recombinant human IL-2 (a gift of the Chiron Corp.). Transfection of either BA/F3 or HT-2 cells was performed by electroporation as described previously (28); stable transfectants were obtained by selection in G418 (Geneticin, 1 mg/ml, Life Technologies, Inc.) and clones isolated by limiting dilution were screened by radioligand binding analysis with 125 I-IL-2 or 125 I-EPO (see below) or by Northern blot analysis to identify clones expressing the transfected receptor (see text and figure legends). HT-2EPO␤␥ was established by transfecting HT-2EPO␤ cells with pEPO␥neo and culturing in recombinant human EPO (10 units/ml, Amgen, Inc.) without IL-2. The COS-7 cell line (ATCC) was maintained as described (29).
Proliferation Assays-Conventional 24-h [ 3 H]thymidine incorporation assays and transfection proliferation assays were performed essentially as described previously (28). In transfection studies using the chimeric receptors, HT-2 cells and their derivatives (see text) were transfected with expression plasmids encoding chimeric receptors and were then selected for approximately 10 days in EPO (50 units/ml) in the absence of IL-2; cell growth was assessed by [ 3 H]thymidine incorporation on the indicated days.
Plasmid Constructs-All receptor cDNAs were subcloned into the expression vectors pCMV4 (30), pCMV4Neo (28), or pCMV4⌬ (a pCMV4 derivative containing a deletion of a vestigial second polylinker downstream of the cytomegalovirus expression cassette). For all constructs requiring synthetic oligonucleotides or PCR reactions, sequences were confirmed by DNA sequencing. The murine EPOR cDNA from pXM-nEPOR (31) was inserted into the KpnI/XbaI sites of pCMV4Neo to yield pEPORneo, and the human IL-2R␤ cDNA from pIL2R30 (provided by T. Taniguchi) was inserted into the HindIII/BamHI sites of pCMV4Neo to yield p␤neo.
The tyrosine substitution mutants of IL-2R␤ and ␥ c (tyrosine (TAC) to phenylalanine (TTC)) were prepared by a combination of oligonucleotide-directed mutagenesis in M13 bacteriophage and PCR-based methods. For constructs involving the ␥ c cytoplasmic tail, a full-length cDNA was obtained by reverse transcription PCR based on the IL-2R␥ sequence reported by Takeshita et al. (32). Deletion and substitution mutants described under "Results" (see figure legends) were prepared by PCR using IL-2R␤ or ␥ c cDNAs as templates.
pEPO␤neo, constructed by PCR using an NheI site at the fusion junction, encodes a chimeric receptor (see Fig. 6A) containing the extracellular domain of the EPOR fused just above the transmembrane segment to the human IL-2R␤ transmembrane and cytoplasmic segments (resulting sequence: . . . (EPOR-T-A-S)-(G-K-D-IL-2R␤) . . . ). pEPO␥neo, also constructed by PCR using the NheI site, encodes a receptor (see Fig. 6A) containing the extracellular domain of the EPOR fused to the human ␥ c transmembrane and cytoplasmic segments (resulting sequence: . . . (EPOR-T-A-S)-(S-K-E-␥ c ) . . . ). Expression plasmids encoding the mutants described in the text were prepared by subcloning appropriate DNA fragments spanning the indicated mutations into the parental pEPO␤neo and pEPO␥neo plasmids.
Protein Expression and Phosphorylation Studies-COS-7 cells (ATCC) were transfected with the indicated plasmids (see text) using Lipofectamine (Life Technologies, Inc.) as per the manufacturer's instructions. For expression analysis of chimeric receptors, immunoblotting analyses were performed on cell lysates using an anti-EPOR Nterminal peptide antiserum and 125 I-protein A as described previously (31). For phosphorylation analyses, the indicated cell lines were stripped of bound ligands by a 1-min acidic wash (10 mM sodium citrate, 0.14 M NaCl, pH 4) and then were rested in medium without serum or cytokines for 4 h. Cells were then stimulated with either IL-2 (10 nM) or EPO (50 units/ml) for 10 min at 37°C, lysed (1% Nonidet P-40, 150 mM NaCl, 20 mM Tris, pH 8.0, 50 mM NaF, 100 M sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml aprotinin, 1 g/ml pepstatin A) and immunoprecipitated with either the anti-IL-2R␤ monoclonal antibody 561 (kindly provided by Dr. R. Robb) or an anti-JAK1 antiserum (Upstate Biotechnology, Inc.) and protein A-Sepharose. Immunoblotting studies were performed with anti-phosphotyrosine antibody (4G10, Upstate Biotechnology, Inc.) per the manufacturer's instructions followed by ECL (Amersham Corp.) signal development.

Substitution Mutation of all Six Cytoplasmic Tyrosine Residues in IL-2R␤ Impairs Growth Signal Transduction in a Transient Assay
System-The cytoplasmic tail of the human interleukin-2 receptor (IL-2R) ␤ chain contains six tyrosine residues (37) (Fig. 1), including four in the "acidic" region (A) (38) and one in each of two distal segments (B, C) (28). To investigate the possibility that growth signaling through the IL-2R is regulated by tyrosine phosphorylation, a mutant IL-2R␤ chain (␤YF) containing concurrent substitutions of phenylalanine at all six cytoplasmic tyrosine positions was prepared and analyzed in a transient assay of lymphocyte growth signal transduction. In this method (28), IL-3-dependent murine pro-B cells (BA/F3) (27) containing endogenous IL-2R␥ chains are transfected with expression plasmids encoding wild type or mutant IL-2R␤ and selected in medium containing IL-2 in the absence of IL-3. Cells transfected with wild type IL-2R␤ (␤WT) chains proliferated vigorously as indicated by substantial incorporation of [ 3 H]thymidine within 7 to 9 days, whereas cells receiving the vector control died in culture (Fig. 1A). Using this assay system, lymphocytes transfected with the all tyrosine-negative IL-2R␤ mutant (␤YF) demonstrated a dramatically impaired proliferative response to IL-2 (Fig. 1B). Thus, one or more of these cytoplasmic tyrosines of IL-2R␤ appeared to be critically required for full growth signal transduction through the IL-2R.
Two independent types of experiments were performed to ensure that the impaired function of ␤YF was not simply the result of ineffective surface expression or faulty binding of ligand. First, to monitor surface expression COS cells were transiently transfected with expression vectors encoding the IL-2R␣ chain and either native IL-2R␤ or ␤YF, followed by incubation with 125 I-IL-2, chemical cross-linking with disuccinimidyl suberate, and immunoprecipitation with the anti-␤ monoclonal antibody DU-2 (14). Following SDS-polyacrylamide gel electrophoresis, bands of comparable intensity and migration were observed for cells transfected with the wild type ␤ and ␤YF, indicating the unimpaired surface expression of the mutant ␤YF receptor (Fig. 1C). To investigate potential changes in receptor affinity, radioligand binding analyses were performed with 125 I-IL-2 in COS cells transfected with IL-2R␥ and ␤WT or ␤YF. These studies revealed the expected single class of intermediate affinity IL-2 binding sites for both ␤WT and ␤YF (K d 300 -400 pM) (Fig. 1, D and E). Thus, surface expression and ligand binding by ␤YF appeared indistinguishable from wild type ␤ and therefore do not account for its impaired signaling function in the transfection assay system.
The Tyrosine-negative Mutant of IL-2R␤ Demonstrates Impaired Responsiveness to IL-2 in a Stable Transfectant-To confirm the phenotype of ␤YF, stable sublines of BA/F3 were prepared by transfection with the plasmid p␤YFNeo. Radioligand analysis demonstrated that the Baf␤WT and Baf␤YF cell lines expressed receptors that bound IL-2 with comparable intermediate affinities (data not shown), although for unknown reasons the Baf␤YF lines consistently expressed the receptor at somewhat lower levels than did Baf␤WT (Baf␤WT, 3000 receptors/cell; Baf␤YF, 700 receptors/cell). Nevertheless, in analyses of numerous sublines we have seen no correlation between expression levels in this range and proliferative signaling capacity.
Analysis of [ 3 H]thymidine incorporation in response to IL-2 revealed marked unresponsiveness of the stable Baf␤YF cell line to IL-2 compared with Baf␤WT ( Fig. 2). As expected, the Baf␤WT cell line demonstrated detectable proliferation even at very low doses of IL-2 (10 pM) well below the K d of IL-2 binding to IL-2R␤␥ complexes, whereas the Baf␤YF line demonstrated no response even at very high doses of IL-2 (100 nM) vastly exceeding the measured K d . These findings confirmed the impaired proliferation signaling exhibited by the ␤YF mutant initially detected in the transient system.

Selective Mutation of Individual Tyrosine Residues Does Not Alter IL-2R␤ Growth Signaling in a Pro-B Cell Line-
The results in both transient and stable assay systems indicated that at least one tyrosine residue contributes importantly to IL-2R growth signaling competence in pro-B cells. To identify the relevant functional tyrosine residue(s), IL-2R␤ mutants containing selective phenylalanine for tyrosine substitutions were constructed and characterized using the BA/F3 transient assay system. Surprisingly, substitution of phenylalanine at Tyr-338 (␤Y1F), Tyr-355/Tyr-358/Tyr-361 (␤Y234F), Tyr-392 (␤Y5F), or Tyr-510 (␤Y6F) had little or no effect on growth signal transduction in response to IL-2 ( Fig. 3). In contrast to ␤YF, each of these selective tyrosine mutants mediated substantial proliferation; only a subtle compromise in receptor function was intermittently observed with ␤Y5F and ␤Y6F. These results revealed that no single cytoplasmic tyrosine is essential to growth signaling function, implying that a functional redundancy may exist involving two or more of these residues.
Either Tyr-392 or Tyr-510 Alone Is Sufficient to Permit IL-2R␤ Growth Signaling Function in Pro-B Cells-Previous reports with stable transfectants expressing IL-2R␤ mutants had demonstrated that the "A" segment spanning the first four cytoplasmic tyrosine residues is dispensable for growth signaling function (38), an observation confirmed in our previous studies employing the transient assay system in BA/F3 cells (28). This finding implied that the C-terminal tyrosines (Tyr-392 and Tyr-510) may be sufficient for full growth signaling. To evaluate this possibility, a mutant was prepared (␤YF:56Y) containing substitutions of phenylalanine for the proximal four tyrosines, leaving the distal tyrosines intact; this mutant mediated a full proliferative response to IL-2 in BA/F3 cells (Fig.  3). In contrast, a mutant with phenylalanines replacing exclusively these two distal tyrosines (␤Y56F) was substantially impaired in its growth signal transduction capacity in the BA/F3 cells, further demonstrating the importance of Tyr-392 and Tyr-510 to growth signaling by IL-2R␤ (Fig. 3).
We further observed that internal deletion of a 119-amino acid cytoplasmic region of IL-2R␤ spanning the A region as well as the contiguous "B" segment exhibited fully preserved growth signaling (Fig. 3, ␤⌬AB), suggesting that the first five tyrosines are dispensable. In contrast, extension of this deletion to include the C-terminal region containing the sixth tyrosine (␤⌬ABC) abrogated receptor function (Fig. 3). These results suggested that the sixth tyrosine (Tyr-510) is sufficient to permit growth signal transduction. Indeed, an IL-2R␤ mutant in which only this single tyrosine was restored in the ␤YF background (␤YF:6Y) exhibited substantial IL-2 growth signaling (Fig. 4A).
Although Tyr-510 alone is sufficient for receptor competence, selective substitution of phenylalanine at this position had little effect on the signaling function (Fig. 3). These results strongly implied that at least one other tyrosine site also could support growth signal transduction, a hypothesis that was tested by evaluating additional tyrosine add-back mutants. Interestingly, reconstitution of Tyr-392 (␤YF:5Y) substantially restored the IL-2R␤ signaling function (Fig. 4B). In contrast, restoration of tyrosines in the first four positions in two additional add-back mutants (␤YF:234Y and ␤YF:1Y) failed to reconstitute receptor function (Fig. 4, C and D, respectively). Importantly, the ␤YF:56Y, ␤Y56F, ␤YF:1Y, ␤YF234Y, ␤YF:5Y, and ␤YF:6Y proteins were all expressed abundantly as detected by immunoblotting analysis (data not shown). Thus, either the fifth tyrosine (Tyr-392) or sixth tyrosine (Tyr-510) is necessary and sufficient for IL-2 growth signaling in BA/F3 cells.
Tyrosine 392 of IL-2R␤ Is Phosphorylated upon Engagement of the IL-2R-The present findings indicating a functional role for certain cytoplasmic tyrosine residues of IL-2R␤ raised the important question of whether or not these tyrosine residues serve as phosphate acceptor sites, a possibility suggested by the recognition that this chain undergoes rapid tyrosine phosphorylation during receptor activation (24,25). To address this question, stable transfectants of the BA/F3 line were prepared using expression plasmids encoding tyrosine add-back mutants (p␤YF:5YNeo and p␤YF:6YNeo). Both of the resulting cell lines (Baf␤YF:5Y and Baf␤YF:6Y) proliferated vigorously in IL-2 despite the unresponsiveness of the Baf␤YF line (Fig. 5). These results confirmed in permanent BA/F3 cell lines the reconstitution of growth signaling function upon restoration of either Tyr-392 or Tyr-510.
Phosphorylation studies were next performed using these stable transfectants. In these experiments, cell lines were rested without growth factors and then exposed to IL-2. Stimulated cells were lysed, immunoprecipitated with anti-IL-2R␤ monoclonal antibody, and then subjected to immunoblot analysis with anti-phosphotyrosine antibody. Upon induction with IL-2 the Baf␤WT line yielded a strong phosphotyrosine signal at the appropriate molecular weight for IL-2R␤ chains, whereas the Baf␤YF line yielded no discernible signal (Fig. 5). Like Baf␤WT, Baf␤YF:5Y cells also yielded a phosphotyrosinecontaining protein band (Fig. 5B). Since this add-back cell line expresses IL-2R␤ chains containing only a single cytoplasmic tyrosine residue (Tyr-392) with all others replaced by phenylalanine, a phosphotyrosine signal generated in the immunoblot experiment is clearly attributable to this tyrosine. These results thus indicated that Tyr-392 of IL-2R␤ serves as a phosphate acceptor site during receptor activation.
Similar experiments were performed with the Baf␤YF:6Y line to assess the role of Tyr-510 in receptor phosphorylation. Surprisingly, no IL-2R␤ chain tyrosine phosphorylation was detectable in experiments with cells expressing the Tyr-510 add-back mutant (data not shown). Such experiments were performed with multiple, independently derived lines, and stimulations were performed for various lengths of time ranging from 3 to 30 min. It remains possible that this functional tyrosine residue of IL-2R␤ does indeed undergo phosphorylation and that this site is perhaps particularly sensitive to phosphatase attack after detergent solubilization of the cells. Nonetheless, phosphorylation of this tyrosine has not yet been detected (see "Discussion").

Establishment of EPOR/IL-2R Chimeric Receptors to Study the Cytoplasmic Domains of the IL-2R␤ and ␥ c Receptor Subunits in T Cells-To permit study of the functional interactions
of the IL-2R␤ and ␥ c cytoplasmic domains in T lymphocytes already expressing endogenous IL-2 receptors, we developed a chimeric receptor system in which the intracellular domains of interest (derived from IL-2R␤ and ␥ c ) were fused to an extracellular ligand binding domain not present in the host cell lines (Fig. 6A). Extracellular domains of the homodimeric EPOR extracellular domain were employed for this purpose, since the EPOR, IL-2R␤, and ␥ c subunits are all members of the cytokine receptor superfamily. Because the EPOR homodimerizes in the presence of EPO, these chimeric receptors were expected to promote dimerization of the IL-2R␤ and/or ␥ c cytoplasmic domains following ligand binding. Plasmids encoding the chimeric EPO␤ and EPO␥ receptors expressed proteins of the predicted masses as detected by immunoblot analysis of lysates from transfected COS-7 cells (Fig. 6B): the native EPOR and wild type EPO␤ and EPO␥ constructs yielded bands of approximately 70, 75, and 40 kDa, respectively. Frequently protein doublets were observed with all of these constructs, which result from variable glycosylation.
The IL-2-dependent murine helper T cell line, HT-2, was employed for analysis of EPO␤ and EPO␥ signaling. Initially, stable HT-2 transfectants expressing the EPOR, EPO␤, or EPO␥ subunits were established. In 24-h [ 3 H]thymidine incorporation assays, the EPOR was found to mediate a modest response to EPO, whereas neither of the chimeric receptor subunits alone produced a detectable response in multiple transfected clones (Fig. 6C). The failure of EPO␤ and EPO␥ to mediate a response was not due to lack of expression, since Northern blotting, Western blotting, and radioligand binding analyses with 125 I-EPO confirmed the expression and ligand binding competence of these chimeras in the HT-2EPO␤ and HT-2EPO␥ cell lines (data not shown).
Since neither chimera alone (EPO␤ or EPO␥) demonstrated detectable growth signal transduction, combinations of these chimeras in HT-2 cells were tested for growth signaling in response to EPO as a means of promoting heterodimerization of the IL-2R␤ and ␥ c cytoplasmic tails. For these studies the transfection assay originally described for BA/F3 cells (28) was adapted to HT-2 cells. When the EPO␥ expression plasmid was introduced by electroporation into multiple HT-2 clones stably expressing EPO␤ (HT-2EPO␤), addition of EPO without IL-2 produced marked proliferation and vigorous incorporation of [ 3 H]thymidine during the 12-day assay (Fig. 7A). Similarly, multiple HT-2 clones stably expressing EPO␥ (HT-2EPO␥) displayed marked proliferative responses to EPO following introduction of the EPO␤ expression plasmid (Fig. 7B) chimeras is required for effective growth signaling, as has been reported in studies with other chimeric receptors (3,4).
Tyrosine Residues of IL-2R␤ Are Required for Full Growth Signaling in Mature T Cells-The functional contribution of IL-2R␤ cytoplasmic tyrosines in T cells was assessed using the chimeric receptor system and the HT-2 cell line. HT-2EPO␥ cells transfected with expression plasmids encoding either wild type EPO␤ or a mutant, tyrosine-negative EPOR/IL-2R␤ chimera (EPO␤YF) were selected in EPO and assessed for proliferation. Unlike the parental EPO␤ (Fig. 7A), the tyrosinenegative EPO␤YF exhibited no detectable growth response to EPO (Fig. 7C). Similarly, stable double transfectants of HT-2 expressing both EPO␤YF and EPO␥ demonstrated no proliferation response to EPO (data not shown). These findings demonstrated that the cytoplasmic tyrosines of the IL-2R␤ chain strongly influence receptor growth signaling independently of ligand specificity in both pro-B and mature T cells.
To analyze further the disruption in signal transduction by the ␤YF mutant, Janus kinase induction in response to receptor engagement was assessed. Lysates prepared from HT-2 cells stimulated with no cytokine, IL-2, or EPO were subjected to immunoprecipitation with an anti-JAK1 antiserum followed by immunoblot analysis with an anti-phosphotyrosine antibody. Cells expressing chimeric ␥ c chains and either wild type chimeric ␤ chains (HT-2EPO␤␥) or tyrosine-negative ␤ chains (HT-2EPO␤YF/␥) both exhibited strong induction of JAK1 phosphorylation in response to either ligand (Fig. 7D). Likewise, preserved induction of JAK3 phosphorylation by receptor complexes containing EPO␤YF was observed in parallel experiments employing an anti-JAK3 antiserum (data not shown). Therefore, at least one early phase of receptor-mediated signaling by the ␤YF mutant is intact despite the failure to achieve full growth signaling.
Characterization of Cytoplasmic ␥ c Mutant Function in T Lymphocytes-Development of the chimeric receptor system also permitted an examination in T cells of the functional contributions of tyrosine residues and other elements within the ␥ c cytoplasmic tail. We therefore introduced EPO␥ mutants into the HT-2EPO␤ stable cell line for functional analysis in the transfection assay. Protein expression from the various mutant EPO␥ chimeric constructs was first verified by immunoblot analysis of lysates from transiently transfected COS-7 cells (Fig. 6B). As predicted, the substitution mutant construct (EPO␥YF, see below) produced protein comparable with that of the wild type EPO␥ construct, and the deletion mutants (EPO␥336, EPO␥294, EPO␥⌬Box1, and EPO␥⌬V-Box) produced slightly faster migrating species.
Since tyrosine phosphorylation of the ␥ c subunit upon ligand binding has been well described (26), we investigated the putative role of the tyrosine residues present in the ␥ c subunit by phenylalanine substitution of all four tyrosine residues (EPO␥YF). Surprisingly, growth signal transduction by EPO␥YF was nearly indistinguishable from that by EPO␥ both in transfection assays (Fig. 8, A and B) and in 24-h [ 3 H]thymidine incorporation assays of stable transfectants arising from transfection of HT-2EPO␤ cells with the EPO␥YF expression plasmid (Fig. 8, C and D). Thus, the cytoplasmic tyrosine residues of ␥ c appeared to be dispensable for growth signaling, which stands in sharp contrast to their importance in the IL-2R␤ subunit.
Although the tyrosine residues are non-essential, other regions of the ␥ c cytoplasmic tail proved important for growth signaling. EPO␥ mutants truncated at the cell membrane (EPO␥TM) or at the end of the Box 1 (39) homology region (EPO␥294) mediated no detectable proliferation signaling (Fig.  9). Similarly, internal deletion of Box 1 (EPO␥⌬Box1), of a segment with distant relationship to the Box 2 motif (EPO␥⌬Box2), or of the segment connecting Box 1 to Box 2 (EPO␥⌬V-Box), also abolished proliferation signaling. However, truncation of the ␥ c subunit at the C-terminal end of the Box 2 region (EPO␥336) resulted in levels of growth signaling similar to that obtained with the wild type subunit. Thus, unlike the IL-2R␤ subunit, the distal portion of the ␥ c subunit is dispensable for proliferation signal transduction, and full growth-signaling function resides in the proximal 53 amino acids containing the Box 1, Box 2, and intervening (V-Box) segments.

DISCUSSION
Like many other cytokine receptor systems, the binding of IL-2 to the IL-2R induces the tyrosine phosphorylation of a variety of intracellular substrates, including the IL-2R␤ and ␥ c chains (24 -26). Although no tyrosine kinase domain is identifiable within the recognized ligand-binding subunits of the IL-2R, the Janus kinases JAK1 and JAK3 as well as the src family kinase p56 lck and p59 fyn are now recognized to associate noncovalently with the cytoplasmic tails of IL-2R subunits (10,15,19,40). The activation of such receptor-associated kinases may represent a mechanism for signal transduction that is fundamentally the same as that for receptors containing intrinsic kinase activity. Indeed, as in such kinase-containing receptors, some evidence has accumulated from mutagenesis and in vitro analyses that certain tyrosine residues of the IL-4 and interferon receptors are crucial for signal transduction competence (41)(42)(43)(44).
The present studies were undertaken to evaluate the poten-  Fig. 6 were transfected with the EPO␤ expression plasmid, selected in EPO (50 units/ml) without other cytokines, and assayed for growth by measuring [ 3 H]thymidine incorporation on the indicated days. B, parental HT-2 cells or HT-2EPO␤ cells described in Fig. 6 were transfected with the EPO␥ expression plasmid, selected in EPO, and assayed for growth. C, HT-2EPO␥ cells were transfected with the EPO␤YF expression plasmid, selected in EPO, and assayed for growth. Each experiment shown was performed multiple times with similar results. D, to assess phosphorylation of JAK1 during receptor activation, the stable transfectants HT-2EPO␤␥ and HT-2EPO␤YF/␥ were stimulated with no cytokine (Ϫ), EPO (E), or IL-2 (2) followed by immunoprecipitation with the anti-JAK1 antiserum and immunoblotting with the anti-phosphotyrosine antibody. tial regulatory role of cytoplasmic tyrosines of the IL-2R␤ and ␥ c chains. In these studies employing both native and chimeric receptors, substitution of phenylalanine for all six cytoplasmic tyrosine residues of IL-2R␤ substantially impaired growth signaling in both a pro-B and a mature T cell line (Figs. 1 and 7). A panel of add-back mutants revealed that both Tyr-392 and Tyr-510 individually exhibit signaling potential in the BA/F3 pro-B cell line while the four more proximal tyrosines demonstrate no functional capacity in this specific cellular environment (Fig. 4). We conclude from these experiments that, in BA/F3 cells, the two C-terminal cytoplasmic tyrosines serve important but redundant functions in determining the signal transduction competence of the IL-2R␤ chain.
The finding that C-terminal tyrosines of IL-2R␤ influence growth signaling in this system appears to contrast with an earlier report that the IL-2R␤ segment encompassing these tyrosines is dispensable for proliferative signaling (38). However, point substitutions and deletions of identical regions may have different phenotypic consequences, particularly if the protein region in question exerts regulatory effects via conformational changes. For example, the C terminus of IL-2R␤ may negatively regulate proximal domains through steric hindrance, which might be relieved by receptor activation. Such a model would also explain the negative regulatory domain identified within the EPOR C terminus (45). A deletion mutant thus may obscure a role of tyrosine residues within this region. Therefore, we conclude that tyrosines within the IL-2R␤ cytoplasmic tail are indeed important for the growth signaling competence of IL-2R␤.
The mechanism(s) underlying the importance of Tyr-392 and Tyr-510 to IL-2R function remain uncertain. In the plateletderived growth factor receptor system, several distinct signaling pathways are activated selectively by individual phosphotyrosine residues through interactions with proteins via SH2 domains (46,47). Recent reports have described the inducible binding of p52 shc to the IL-2R␤ chain upon the binding of IL-2 (48,49), although the molecular basis of this interaction is unknown. Similarly, phosphatidylinositol 3-kinase has also been found to associate with the IL-2R␤ chain in the presence of IL-2 (50,51), an event which may be facilitated by phosphorylation of IL-2R␤ Tyr-392 as revealed in studies with phosphopeptides (51). Finally, following completion of the present FIG. 9. Functional analyses of EPOR/␥ c chimeras in transfection assays of proliferation. Transfection growth assays using the HT-2EPO␤ line as a host to assess the responses of the the indicated EPO␥ mutants. The ␥336, ␥294, and ␥TM mutants are truncated immediately after amino acids 336, 294, and 286, respectively, in the mature ␥ c protein. ␥⌬Box 1 is deleted of residues 281-294, ␥⌬V-Box is deleted of residues 295-320, and ␥⌬Box 2 is deleted of residues 321-334. Results are expressed as the incorporation of [ 3 H]thymidine for each line relative to that of the wild type (␥wt) cytoplasmic tail, with standard errors of the mean (n Ն 3). work, we (52) and others (53) have demonstrated that phosphopeptides encompassing either Tyr-392 or Tyr-510 are potent and specific inhibitors of the in vitro DNA binding activity of STAT-5, a STAT factor that is regulated by the IL-2R (52)(53)(54). Interestingly, tyrosine residues of IL-2R␤ are dispensable for Janus kinase activation by the IL-2R (Fig. 7D) but are essential for the effective induction of STAT-5 (52). Together, these findings are consistent with the popular model of cytokine receptor function (55) in which ligand-induced phosphorylation of certain tyrosine residues of the receptor is a critical step in the generation of downstream intracellular signals.
Convincing demonstration of the significance of this model for IL-2R function requires identification of the sites of IL-2induced tyrosine phosphorylation of IL-2R␤ in vivo. The present studies demonstrated that Tyr-392 serves as a phosphate acceptor site upon exposure of BA/F3 transfectants to IL-2 (Fig.  5). Unexpectedly we failed to detect phosphorylation of Tyr-510 in parallel experiments. It is possible that this lack of detection results from technical problems, such as insensitivity of the assay method or contaminating phosphatase activity released during cell lysis. Alternatively, this observation may indicate that Tyr-510 function is entirely independent of its phosphorylation status. Indeed, the published evidence supporting a critical role for receptor phosphotyrosines in the JAK-STAT pathway is largely circumstantial. For example, experimental demonstration of direct interactions between STAT factors and phosphotyrosine-containing receptor segments has proven difficult in most circumstances, and heavy emphasis has been placed instead on in vitro peptide approaches (44). Therefore, the lack of detectable phosphorylation of Tyr-510 in the present studies raises the possibility that this and perhaps other tyrosine residues of IL-2R␤ exert crucial influences on the tertiary conformation of IL-2R␤ independently of their phosphorylation status. Although we tend to favor the tyrosine phosphorylation model, rigorous consideration of the published data demands further studies to distinguish effectively between these interpretations.
Other cytokine receptor superfamily members (1) may similarly be influenced by tyrosines. Functionally important tyrosine residues within the cytoplasmic domains of the IL-4 and interferon-␥ receptors have been described recently (41)(42)(43), although the significance of IL-4R phosphorylation has been disputed (56). The functional redundancy described here for the distal IL-2R␤ tyrosines may also be a feature of the human IL-4 receptor that could explain the incomplete impairment of function reported upon substitution of phenylalanine for Tyr-497 in the IL-4 receptor (41). Further investigation is needed to clarify these events within the IL-2R.
The EPOR/IL-2R chimeric system also permitted an assessment of the role of tyrosine and other residues within the ␥ c cytoplasmic tail for growth signaling in T cells. In contrast to the IL-2R␤ chain, the ␥ c subunit functioned fully in the absence of all four of its cytoplasmic tyrosine residues (Fig. 8). This finding indicates that growth signaling intermediates interacting with the ␥ c tail do so independently of phosphotyrosine docking sites, even though one or more of these tyrosine sites is phosphorylated after IL-2 stimulation in vivo. In view of the fact that both the IL-4 and IL-2 receptors employ the ␥ c subunit, these observations raise the intriguing possibility that the longer, unique chain in each receptor provides the docking sites for the specific signaling intermediates engaged by each receptor complex. In this arrangement, the shared ␥ c subunit would participate in general initiation of the signaling process, whereas the specialized subunits would contain unique sites for the inducible binding of specific components, such as STAT factors. Other cytokine receptors might employ a similar func-tional configuration. Of course, it remains possible that components involved in other pathways not measured here (such as differentiation) do indeed depend upon these ␥ c tyrosine sites.
Although the tyrosines of ␥ c proved to be dispensable for growth signaling by the IL-2R, a panel of truncation and internal deletion mutants revealed other elements within ␥ c that are critical for growth signaling in the T cell line. Remarkably, the C-terminal 33 amino acids of ␥ c are fully dispensable for growth signaling (Fig. 9), indicating that the proximal 53 amino acids are sufficient for full growth signal transduction. Mutations within this membrane-proximal region abrogated signaling function. For example, extension of the truncation N-terminal to a vestigial "Box 2" motif (39) abolished the signaling function, as did internal deletion of the 14 amino acids constituting a "Box 1" motif, the 14 amino acids constituting this vestigial Box 2 motif, or the 26 amino acids connecting Box 1 to Box 2 (V-Box) (Fig. 5). These observations in T cells extend the studies by others which employed certain truncated ␥ c subunits expressed in heterologous cell types (57)(58)(59) and demonstrate clearly that the ␥ c tail is needed for growth signal transduction by IL-2R heterodimers in T cells. Importantly, the impairment of these ␥ c domains undoubtedly contributes to the pathologic effects manifested in the X-linked severe combined immunodeficiency syndrome (60).
The recognition that the growth signaling function of ␥ c resides in a relatively small portion of the cytoplasmic tail and that this segment functions independently of tyrosine residues is consistent with the receptor model described above. The essential, membrane-proximal region of ␥ c has been shown to be crucial for the assembly of the Janus kinase JAK3 with ␥ c (10,40). Perhaps the primary function of ␥ c in the IL-2, IL-4, and other receptors is to convey JAK3 into the receptor complex upon engagement of the appropriate ligand, which would thus allow trans-activation of JAK1 and JAK3 bound to their respective receptor subunits. Subsequent signaling activities may focus primarily upon the extended cytoplasmic tail of the unique IL-2R␤ chain, including the inducible binding and activation of specific factors. Further studies are needed to determine whether or not the ␥ c chain has additional functions in addition to its conveyance role. One or both of the Janus kinases may be involved in phosphorylation substrates within the receptor complex. The present findings provide a rationale for further investigation of these intracellular events.