Delineation of the Regions of Interleukin-2 (IL-2) Receptor b Chain Important for Association of Jak1 and Jak3 Jak1-INDEPENDENT FUNCTIONAL RECRUITMENT OF Jak3 TO IL-2Rb*

Interleukin-2 (IL-2) induces heterodimerization of the IL-2 receptor b (IL-2Rb) and gc chains of its receptor and activates the Janus family tyrosine kinases, Jak1 and Jak3. Whereas Jak1 associates with IL-2Rb, Jak3 associates primarily with gc but also with IL-2Rb. We analyzed four IL-2Rb mutations that diminish IL-2-induced proliferation and found that each also decreased IL-2induced signal transducer and activator of transcription (STAT) activation. For this reason, and because the mutations were in the IL-2Rb membrane-proximal region, we investigated and found that each mutation diminished IL-2Rb association with both Jak1 and Jak3. This suggested that these Jaks might interact with the same region of IL-2Rb; however, certain IL-2Rb internal deletions and C-terminal truncations differentially affected the association of Jak1 and Jak3. Interestingly, just as Jak1-IL-2Rb association is Jak3-independent and functionally important, we show that Jak3-IL-2Rb association is Jak1-independent and implicate this association as being important for IL-2-induced Stat5 activation. Moreover, Jak1 and Jak3 could associate only in the presence of IL-2Rb, suggesting that these kinases can simultaneously bind to IL-2Rb. Thus, our data not only demonstrate that somewhat more distal as well as membrane-proximal cytoplasmic regions of a type I cytokine receptor are important for Jak kinase association but also suggest that two IL-2Rb-Jak kinase interactions are important for IL-2 signaling.

Interleukin-2 (IL-2) 1 is the principal growth factor for T lymphocytes and is responsible for regulating the magnitude and duration of the T cell immune response following antigen encounter (1)(2)(3)(4). Three classes of IL-2 receptors exist, binding IL-2 with low (K d ϭ 10 Ϫ8 M), intermediate (K d ϭ 10 Ϫ9 M), and high (K d ϭ 10 Ϫ11 M) affinity. The low affinity receptors contain only the IL-2 receptor ␣ chain (IL-2R␣); intermediate affinity receptors contain IL-2R␤ and the common cytokine receptor ␥ chain, ␥ c ; and high affinity receptors contain all three chains (3,4). The intermediate and high affinity receptors are the functional forms, and heterodimerization of the IL-2R␤ and ␥ c cytoplasmic domains is necessary and sufficient for IL-2 signaling (5)(6)(7). The highly inducible ␣ chain has a very short cytoplasmic domain (8,9) and presumably mainly functions to increase the affinity for IL-2, allowing cellular responsiveness to the low levels of IL-2 that are physiologically present in vivo. In contrast, IL-2R␤ and ␥ c have longer cytoplasmic domains that can associate with a number of signaling molecules, allowing the activation of signaling pathways (2)(3)(4). Stimulation of lymphocytes with IL-2 results in the rapid activation of the Janus family tyrosine kinases, Jak1 and Jak3 (10 -14). Activated Jaks are critical for inducing rapid tyrosine phosphorylation of downstream substrates, including STATs (signal transducers and activators of transcription), which then dimerize, translocate into the nucleus, and regulate the transcription of target genes (4,(13)(14)(15).
It has been reported that IL-2R␤ and ␥ c constitutively associate with two of the four Jak family kinases in a selective manner, IL-2R␤ with Jak1 and ␥ c with Jak3 (10,16,17). The S region (amino acids 267-322) of IL-2R␤ has been shown to be important for Jak1 association (17). In addition to its ability to constitutively interact with Jak1, although it is not well appreciated, IL-2R␤ can also associate with Jak3 following IL-2 stimulation of lymphoid cells (10), but the regions of interaction between IL-2R␤ and Jak3 have not previously been investigated.
A number of membrane proximal cytoplasmic point mutants of IL-2R␤ that diminish IL-2-induced proliferation have been identified (18 -20). We found that these mutants also diminish IL-2-induced STAT protein activation and the association of both Jak1 and Jak3 with IL-2R␤. This led us to further characterize the regions of IL-2R␤ required for the binding of Jak1 and Jak3, and we demonstrate that membrane distal as well as membrane proximal regions of IL-2R␤ are vital for Jak kinase interaction. Moreover, the association between Jak3 and IL-2R␤ is Jak1-independent and both Jak3 and Jak1 can be coprecipitated only in the presence of IL-2R␤. Finally, we provide evidence indicating that the association between IL-2R␤ and Jak3 is important for potent Stat5 activation in response to IL-2 and, thus, that more than one IL-2R␤-Jak kinase interaction is involved in IL-2 signaling.
Cell Lines and Transfections-COS-7 cells (ATCC), 293T ϩ cells (provided by Dr. N. Rice, National Cancer Institute), and E1C3 cells (Jak1deficient HeLa cells, provided by Dr. R. Flavell, Yale University) were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 2 mM glutamine, 100 units/ml each of penicillin and streptomycin. Transient transfections were performed using either DEAE-dextran (for COS-7 cells) or calcium phosphate (for 293T ϩ and E1C3 cells) methods. For immunoprecipitation experiments, cells were transfected in 150-mm dishes using 2-3 g of each plasmid. Transfec-tants were harvested 36 -48 h later. For experiments in which IL-2induced STAT DNA binding activity was reconstituted, cells were transfected in 100-mm dishes using 1-2 g of each plasmid, and nuclear extracts were made 36 -48 h later.
Preparation of Nuclear Extracts and Electrophoresis Mobility Shift Assays (EMSAs)-Extracts were prepared from 293T ϩ or COS-7 transfectants (cells from one 100-mm culture dish) or from 32D cells (1 ϫ 10 7 cells) that were starved of growth factor for 4 h in RPMI 1640 medium and treated with 2 nM IL-2 for 30 min at 37°C. Cells were washed with ice-cold phosphate-buffered saline, nuclear extracts were prepared as described previously (26), and 1 g of protein from 293T ϩ or COS-7 transfectants or 5-10 g of protein from 32D cells were used in EMSAs. For EMSAs, 1 g of poly(dI-dC) was used as a nonspecific competitor and 15,000 cpm of 32 P-labeled double-stranded oligonucleotide containing a trimer of the GAS sequence (5Ј-AGATTTCTAGGAATTC-3Ј) from the ␤-casein promoter (a motif capable of binding IL-2-activated STAT proteins) was used as the probe. The reactions were separated on 6% polyacrylamide gels in 0.5 ϫ Tris borate-EDTA and autoradiographed.
Thymidine Incorporation Assays-32D cells were washed and starved of growth factor for 4 h in RPMI medium. Cells were aliquoted at 2-4 ϫ 10 4 cells/well in a 96-well plate in triplicate in 200 l of medium or medium containing 2 nM IL-2 or 5% WEHI-CM. After 20 h of incubation at 37°C, 1 Ci of 3 H-labeled thymidine (NEN Life Science Products) was added, and the cells were incubated at 37°C for 4 h. Cells were harvested using a cell harvester (Tom Tec), and thymidine incorporation was assayed using a Betaplate 1205 counter (Wallac). For each transfectant, at least three clones with similar IL-2R␤ expression were assayed.

IL-2R␤ Point Mutants That Affect Proliferation Diminish
IL-2-induced Stat5 DNA Binding Activity-Four IL-2R␤ mutants, including P257S (proline 257 replaced by serine), D258A, W277G, and L299A, have been reported to impair IL-2-induced proliferation in Ba/F3 or MOLT4 cells (18 -20) even though they exhibit similar surface expression and IL-2 binding affinities (19). We sought to investigate the basis for the decreased proliferation of these mutants. We first made stable transfectants of each of these mutants in 32D cells and confirmed similar cell surface expression by flow cytometry (Fig. 1A). As expected, we confirmed that these mutants mediated greatly diminished proliferation, as compared with wild-type IL-2R␤, in 32D cells, which lack IL-2R␤ but can proliferate in response to IL-2 after IL-2R␤ is transfected and expressed (Refs. 21 and 27; Fig. 1B). Moreover, each of these mutations also diminished IL-2-induced STAT binding activity in transfected 32D cells (Fig. 1C) as well as in transiently transfected COS-7 cells (Fig.  1D). In 32D cells, previous studies indicate that the IL-2induced STAT binding activity is due to Stat5 rather than Stat3 (28). For the COS-7 cell experiments, cells were transfected with ␥ c , Jak3, Stat5a, Stat5b, and the different IL-2R␤ constructs using a system previously shown to reconstitute IL-2-induced Stat5 DNA binding activity with wild-type IL-2R␤ (24).
IL-2R␤ Point Mutants Also Exhibit Diminished Association with Both Jak1 and Jak3-Given the diminished STAT activation and that each of these mutations are contained in a region of IL-2R␤ where Jak kinase interactions might be affected (Box 1/Box 2 region, see Refs. 29 -32), we tested if these mutations diminished the association of Jak1 or Jak3 as a possible explanation for the decreased IL-2 signaling. Because the IL-2R␤-Jak3 interaction is only well seen in T cells following IL-2 stimulation, we used an overexpression system to map the regions of IL-2R␤ that mediate association with Jak1 and Jak3. COS-7 cells were transfected with Jak1 or Jak3 and IL-2R␤ mutants, cells were lysed, and lysates were immunoprecipitated with hMik␤1 antibody to IL-2R␤ and then blotted with antibodies to IL-2R␤ ( Fig. 2A), Jak1 (Fig. 2B), or Jak3 (Fig. 2C). Jak1 and Jak3 each exhibited less binding to each of the IL-2R␤ mutants than to wild-type IL-2R␤ (Fig. 2, B and C).
The Regions of IL-2R␤ Required for Jak1 and Jak3 Binding Partially Overlap-Because each of the IL-2R␤ point mutations interfered with the association of both Jak1 and Jak3, we hypothesized that the regions of IL-2R␤ that were important for Jak kinase interaction might be similar. To investigate this possibility and to map the regions of IL-2R␤ involved in the binding of both Jak kinases, COS-7 cells were transiently transfected with Jak1, Jak3, and wild-type IL-2R␤ or a series of IL-2R␤ truncation mutants (Fig. 3A). Cells were lysed, and lysates were immunoprecipitated with anti-IL-2R␤ antibody, followed by Western blotting with antibodies to IL-2R␤ (to control for expression and the efficiency of immunoprecipitation, Fig. 3B), Jak1 (Fig. 3C), or Jak3 (Fig. 3D). IL-2R␤ truncation mutants retaining 350 (␤350 construct) or more residues (␤362, ␤371, and ␤379 constructs) could bind efficiently to Jak1 (Fig. 3C); ␤330 and ␤313 bound to Jak1 weakly; whereas ␤300, ␤290, and ␤267 could not bind to Jak1. Therefore, the region between residues 300 and 350 of IL-2R␤ is important for its interaction with Jak1. In contrast, the region between residues 330 and 362 was important for the Jak3-IL-2R␤ interaction, given that there was efficient coprecipitation of Jak3 with ␤362

FIG. 2. IL-2R␤ mutants containing individual substitutions at Pro-257, Asp-258, Trp-277, or Leu-299 decreased interaction
with Jak1 and Jak3. A, similar expression levels for the mutant and wild-type IL-2R␤. Cell lysates were immunoprecipitated with hMik␤1 and then Western blotted with anti-IL-2R␤. Cell surface expression of each IL-2R␤ construct was confirmed by flow cytometry (data not shown). B, the P257S, D258A, W277G, and L299A IL-2R␤ point mutants exhibit decreased association with Jak1. COS-7 cells were transfected with Jak1, Jak3, and either mutant or wild-type IL-2R␤, were immunoprecipitated with hMik␤1, and then were Western blotted with anti-Jak1 (top panel). Lysates were Western blotted with anti-Jak1 to confirm the expression of Jak1 in different transfectants (bottom panel). C, the P257S, D258A, W277G, and L299A point mutations in IL-2R␤ also resulted in decreased association of Jak3. The blots in panel C were stripped and reblotted with anti-Jak3. but no detectable coprecipitation of Jak3 with ␤330 even at longer exposure times ( Fig. 3D and data not shown). Wild-type IL-2R␤ and IL-2R␤ mutants retaining the first 362, 371, or 379 amino acids could associate with both Jak1 and Jak3. These results in COS-7 cells were confirmed using 293T ϩ cells (data not shown). Thus, the 300 -350 and 330 -362 regions of IL-2R␤ are important for Jak1 and Jak3 binding, respectively (summarized below in Fig. 8).
We next tested the effect of internal deletions of the S region (residues 267 to 322) or the A region (residues 313 to 382) on the binding of Jak1 (Fig. 4A) and Jak3 (Fig. 4B). Deletion of the S region (␤⌬S) resulted in a dramatic decrease in IL-2R␤ association with Jak1, consistent with previously reported results (17), whereas deletion of the A region only modestly decreased Jak1 association (Fig. 4A). In contrast to the findings for Jak1, deletion of the A region had a much greater effect on the association of Jak3 than did deletion of the S region (Fig. 4B). Thus, the A region of IL-2R␤ is more important for Jak3 association, whereas the S region is more important for Jak1 association. Consistent with the data in Fig. 3, these data indicate that Jak3 binding extends to a more distal region of the IL-2R␤ cytoplasmic domain than does Jak1. Therefore, the data con-tained in Figs. 3 and 4 demonstrate that Jak1 and Jak3 interact with different, albeit overlapping regions of IL-2R␤.
As the A region contains four tyrosines (Tyr-338, Tyr-355, Tyr-358, and Tyr-361), we evaluated the ability of Jak3 to associate with IL-2R␤ containing mutations in these tyrosines (IL-2R␤FFFFYY). As shown in Fig. 4D, Jak3 efficiently associated with this mutant, indicating that the interaction does not depend on phosphorylated tyrosine residues.
Jak3 Can Bind to IL-2R␤ in Jak1-deficient HeLa Cells-Given that Jak1 is ubiquitously expressed, it was possible that the interaction of Jak3 with IL-2R␤ required Jak1. To investigate this possibility, we transfected Jak1-deficient HeLa cells (E1C3 cells) with Jak3 ϩ wild-type IL-2R␤ Ϯ Jak1. Transfected cells were lysed and immunoprecipitated with hMik␤1, followed by blotting with an antiserum to Jak3. We found that IL-2R␤ and Jak3 could interact even in the absence of Jak1, and the presence of Jak1 did not enhance this interaction (Fig.  5A, first two lanes). The uniformity of expression of Jak3, Jak1, and IL-2R␤ was verified by immunoblotting with appropriate antibodies (Fig. 5B). We also used E1C3 cells to map the region of IL-2R␤ required for its interaction with Jak3, and confirmed with Jak1, Jak3, and either wild-type IL-2R␤ (␤wt) or truncated mutants of IL-2R␤, were immunoprecipitated with hMik␤1, and then were Western blotted with anti-IL-2R␤. Cell surface expression of each IL-2R␤ construct was confirmed by flow cytometry (data not shown). C, importance of the amino acids 300 to 350 region of IL-2R␤ for Jak1 binding. Top panel, lysates from COS-7 cells transfected with Jak1, Jak3, and either wild-type or mutant forms of IL-2R␤ were immunoprecipitated with hMik␤1 and then Western blotted with anti-Jak1. Bottom panel, the lysates were Western blotted with anti-Jak1 to confirm the expression of Jak1 in different transfectants. D, importance of the amino acids 330 to 362 region of IL-2R␤ for Jak3 binding. The blots described in panel C were stripped and reblotted with anti-Jak3.
FIG. 4. The A and S regions of IL-2R␤ are differentially important for binding Jak3 and Jak1, respectively. A, deletion of either the A or S regions of IL-2R␤ diminished the association of Jak1. COS-7 cells were transfected with Jak1, Jak3, and either mutant or wild-type IL-2R␤, were immunoprecipitated with hMik␤1, and then were Western blotted with anti-Jak1. Lysates were Western blotted with anti-Jak1 to confirm the expression of Jak1 in different transfectants. B, deletion of the A region of IL-2R␤ more greatly diminished the association with Jak3 than did the S region. The blots described in panel A were stripped and reblotted with anti-Jak3. C, similar expression levels for ␤wt, ␤⌬A, and ␤⌬S. The cell lysates were immunoprecipitated with hMik␤1 and then Western blotted with anti-IL-2R␤. Cell surface expression of each IL-2R␤ construct was confirmed by flow cytometry (data not shown). D, association of IL-2R␤ with Jak1 and Jak3 does not depend on the phosphorylation of the tyrosine residues located on the A region. Cells were transfected with Jak1 ϩ Jak3 ϩ either ␤wt, ␤FFFFYY, or pME18S. Cell lysates were immunoprecipitated with hMik␤1 and then blotted with either anti-Jak3 (top panel), anti-Jak1 (middle panel), or anti-IL-2R␤ (bottom panel). The bands corresponding to Jak3, Jak1, and IL-2R␤ are indicated. the findings reported above in Figs. 3 and 4 (data not shown).
Jak1 and Jak3 Can Only Be Coprecipitated in the Presence of IL-2R␤-Because the association between Jak3 and IL-2R␤ was Jak1-independent, and Jak1 could be coprecipitated with IL-2R␤ in the absence of Jak3, we next investigated whether Jak1 and Jak3 could be coprecipitated through IL-2R␤. COS-7 cells were transfected with Jak1, Jak3, and either pME18S, wild-type IL-2R␤, or IL-2R␤ deletion constructs (␤⌬A, ␤⌬S, and ␤350) that were missing regions important for the interaction of either Jak1 and/or Jak3 (see Figs. 3 and 4). Coprecipitation of Jak3 and Jak1 required IL-2R␤ (Fig. 6A, lane 2 versus lane  1); this association was markedly decreased when the ␤⌬A, ␤⌬S, or ␤350 mutants were used instead of wild type IL-2R␤ (lanes 3-5), further confirming that the association between Jak1 and Jak3 is dependent on the presence of IL-2R␤.
Association between IL-2R␤ and Jak3 Is Required for IL-2induced Stat5 DNA Binding Activity-It has previously been shown that disruption of the Jak1-IL-2R␤ interaction diminished IL-2 signaling (2). To investigate the functional significance of the association between IL-2R␤ and Jak3, we used Jak3-deficient 293T ϩ cells in which IL-2-induced Stat5 DNA binding activity could be reconstituted following transfection with IL-2R␤, ␥ c , Jak3, Stat5a, and Stat5b (Fig. 7A, lanes 5 and  6; Fig. 7B, lanes 1 and 2). Previous studies indicate the vital role of Jak3 for IL-2-induced STAT activation (24). Both IL-2R␤ and ␥ c were required since little, if any, IL-2-induced Stat5 DNA binding activity was seen in the absence of either ␥ c (Fig.  7A, lanes 1 and 2) or IL-2R␤ (Fig. 7A, lanes 3 and 4). However, a truncated form of ␥ c (␥ c -⌬CT) that is missing 80 of 86 amino acids of the ␥c cytoplasmic domain and contributes to IL-2 binding (22, 23) but does not interact with Jak3 (10) still allowed partial IL-2-induced DNA binding activity (Fig. 7B,  lanes 3 and 4). This activity was diminished when Jak3 (Fig.  7B, lanes 5 and 6) was deleted, implicating the IL-2R␤-Jak3 interaction as being important for STAT activation. DISCUSSION IL-2 signaling requires the dimerization of both IL-2R␤ and ␥ c . As Jak1 has been shown to associate with IL-2R␤ and Jak3 with ␥ c , an attractive model has been that each receptor chain associates with a different Jak family kinase in a selective manner and that IL-2-mediated activation of Jak1 and Jak3 initiates a signaling cascade(s). It is well established that the ␥ c -Jak3 interaction (10) and Jak3 activation (33,34) are vital for signaling. We now provide evidence that Jak3 and IL-2R␤ can associate with each other in a Jak1-independent fashion. The fact that IL-2R␤ provides interaction sites for Jak3 as well as Jak1 (see Figs. 8 and 9) suggests that a function of ␥ c might be not only to recruit Jak3 but also to facilitate the "delivery" of Jak3 to IL-2R␤. Moreover, the ability of Jak3 to associate with both IL-2R␤ and ␥ c suggests that Jak3 might stabilize the receptor complex and promote downstream signaling. Our studies on the reconstitution of IL-2-induced Stat5 activation in 293T ϩ cells provide evidence that the full activation of Stat5 requires IL-2R␤ association with both Jak1 and Jak3, and that the heretofore poorly appreciated IL-2R␤-Jak3 association has physiological significance.
We have now delineated regions on IL-2R␤ that are impor- FIG. 5. Jak3 can interact with IL-2R␤ in the absence of Jak1. A, HeLa cells lacking Jak1 (E1C3 cells) were transfected with ␤wt or pME18S and Jak3, Jak1, or Jak3 ϩ Jak1. Cell lysates were immunoprecipitated with hMik␤1 and then blotted with either anti-Jak3. B, lysates of the E1C3 transfectants were Western blotted with anti-Jak3, anti-Jak1, or anti-IL-2R␤ to confirm the expression levels of transfected cDNAs.
FIG. 6. Jak3 can associate with Jak1 in the presence of wildtype IL-2R␤. COS-7 cells were transfected with Jak1, Jak3, and either wild-type IL-2R␤, mutant IL-2R␤, or pME18S. Cell lysates were immunoprecipitated with anti-Jak3 and then blotted with anti-Jak1 (A). The blot was stripped and reblotted with anti-Jak3 to control the immunoprecipitation of Jak3 (B). The expression of Jak1 is shown in (C). tant for the interaction of Jak1 and Jak3. We show that four point mutations in the Box 1/Box 2 region of IL-2R␤ that diminished proliferation also decreased the binding of both Jak1 and Jak3. This is consistent with the important role of this region of a number of type I cytokine receptors for Jak interaction (14, 29 -32). Interestingly, however, analysis of a series of deletion and truncation mutants not only demonstrated differences in the regions of IL-2R␤ that mediate recruitment of Jak1 versus Jak3, but unexpectedly also provided evidence that regions more distal than previously suspected play major roles in the recruitment of the Jak kinases (see Fig.  8). To our knowledge, these data represent the most detailed mapping on a cytokine receptor of the region/residues involved in Jak kinase association. Previously, for all cytokine receptors studied, including IL-2R␤, only the membrane proximal and Box1/Box2 regions have been shown to be important for the association of Jak kinases; thus, our findings have implications regarding the interaction sites of Jak kinases for other type I cytokine receptors as well. Although some receptor chains, such as ␥ c , appear to be uniquely associated with a single Jak, the gp130 signal transducing receptor component that is shared by the receptors for IL-6, IL-11, leukemia inhibitory factor, ciliary neurotrophic factor, oncostatin M, and cardiotrophin-1, can associate with more than one Jak. gp130 has been reported to associate with Jak1, Jak2, and Tyk2 (35,36), but it remains unknown whether these three Jak family kinases serve completely distinctive roles and how they associate with gp130. Our data therefore provide the first example wherein more than one Jak (Jak1 and Jak3) can independently interact with a single receptor molecule (IL-2R␤) via overlapping but non-identical regions.  1 and 2); ␥ c , Jak3, Stat5a, and Stat5b (lanes 3 and 4); or IL-2R␤, ␥ c , Stat5a and Stat5b (lanes 5 and 6). 36 -48 h after transfection, cells were either not treated (lanes 1, 3, and 5) or treated with 2 nM IL-2 for 30 min (lanes 2, 4, and 6), and nuclear extracts were made. EMSAs were performed using the ␤-casein probe. B, 293T ϩ cells were transfected with IL-2R␤, ␥ c , Jak3, Stat5a and Stat5b (lanes 1 and 2); IL-2R␤, ␥ c -⌬CT, Jak3, Stat5a, and Stat5b (lanes 3 and 4); or IL-2R␤, ␥ c -⌬CT, Stat5a, and Stat5b (lanes 5 and 6). IL-2 treatment, nuclear extracts, and EMSAs were performed as described in panel A.
FIG. 8. Regions of IL-2R␤ important for the association of Jak1 and Jak3. The extracellular, transmembrane (TM), and cytoplasmic domains are shown. In the cytoplasmic domain, the Box 1, Box 2, S region, and A region are shown on the left. On the right are shown the regions and residues important for the association of Jak1 and Jak3.

FIG. 9. Schematic model showing an important role for Jak3 association with IL-2R␤ for IL-2-induced Stat5 activation in both transfected 293T ؉ cells (A) and IL-2-dependent cells (B).
A, in 293T ϩ cells, Jak3 associates with both IL-2R␤ and ␥ c in order to achieve full Stat5 activation upon IL-2 stimulation (shown as ϩϩϩϩ on the left). When the truncated ␥ c , which cannot bind Jak3, was present instead of wild-type ␥ c , Jak3 associates with IL-2R␤; in this setting, there is still Stat5 activation, albeit decreased (shown as ϩϩ in the middle). Finally, in the absence of Jak3, there is very little Stat5 activation (shown as ϩ/Ϫ on the right). B, in IL-2-dependent cells, IL-2R␤ and ␥ c form a heterodimer after IL-2 stimulation. Jak3 binds both IL-2R␤ and ␥ c , perhaps stabilizing the receptor complex, allowing for potent downstream signaling.