Proteasomes Regulate Erythropoietin Receptor and Signal Transducer and Activator of Transcription 5 (STAT5) Activation

Cis is an Src homology 2 domain-containing protein, which binds to the erythropoietin receptor and decreases erythropoietin-stimulated cell proliferation. We show that Cis associates with the second tyrosine residue of the intracellular domain of the erythropoietin receptor (Tyr401). Two forms of Cis with molecular masses of 32 and 37 kDa were detected, and we demonstrate that the 37-kDa protein resulted from post-translational modifications of the 32-kDa form. Anti-ubiquitin antibodies recognized the 37-kDa form of Cis and the proteasome inhibitorsN-acetyl-leucyl-leucyl-norleucinal and lactacystin inhibited its degradation, showing that the 37-kDa form of Cis is a ubiquitinated protein, which seems to be rapidly degraded by the proteasome. In erythropoietin-stimulated UT-7 cells, the activation of the erythropoietin receptor and signal transducer and activator of transcription 5 (STAT5) was transient and returned to basal levels after 30–60 min of erythropoietin stimulation. In contrast, these proteins remained strongly phosphorylated, and STAT5 remained activated for at least 120 min in the presence of proteasome inhibitors. These experiments demonstrate that the proteasomes are involved in the down-regulation of the erythropoietin receptor activation signals. Because the proteasome inhibitors induced the accumulation of both the ubiquitinated form of Cis and the Cis-erythropoietin receptor complexes, our results suggest that the ubiquitinated form of Cis could be involved in the proteasome-mediated inactivation of the erythropoietin receptor.

Survival, proliferation, and differentiation of hematopoietic cells are controlled by multiple cytokines that exert their effects through specific membrane receptors. Except for SCF, 1 macrophage colony-stimulating factor, and FLT3 ligand recep-tors, which are class III tyrosine kinase receptors, most receptors for hemopoietic cytokines belong to the class I cytokine receptor family. Although these receptors are devoid of tyrosine kinase activity, they are associated with intracellular tyrosine kinases that belong to several families. Among these tyrosine kinases, the Jak kinases appear to play a key role in the initiation of the intracellular signaling mechanisms (for review, see Refs. 1,2). The Epo receptor belongs to this cytokine receptor superfamily, which also includes receptors for many interleukins and colony-stimulating factors. Epo is essential for the survival and proliferation of the late erythroid progenitors and for their differentiation into erythrocytes. The Epo receptor is preassociated with the tyrosine kinase Jak2, and Epo binding to its receptor induces the dimerisation of the Epo receptor. This dimerisation activates the Jak2 kinases most likely by allowing their transphosphorylation (3). The activated kinases then phosphorylate several proteins, including the Epo receptor itself (4 -7). Several intracellular signaling pathways are subsequently activated, including mitogen-activated protein kinases (8 -10), phosphatidylinositol 3-kinase (11)(12)(13)(14), and STAT5 (10,(15)(16)(17)(18)(19); see Ref. 20 for a recent review concerning the mechanisms of intracellular Epo signaling). Besides these positive regulatory pathways, the activation of the Epo receptor also turns on negative regulatory pathways, which are involved in the termination of the signal transduction. One of these negative regulatory pathways is the tyrosine phosphatase SHP-1, which binds both to the phosphorylated third tyrosine residue of the intracellular domain of the Epo receptor (21) and to the Jak2 kinase (22). SHP-1 then dephosphorylates and inactivates Jak2 and probably the Epo receptor.
Another negative regulatory pathway of Epo receptor signaling involves the protein Cis. Cis (cytokine-inducible Src homology 2-containing protein) gene expression is rapidly induced in hematopoietic cells by IL-2, IL-3, GM-CSF, and Epo (23). All of these cytokines activate STAT5, and it has been shown that Cis is a target gene of STAT5 (24). Forced expression of Cis suppresses IL-3-and Epo-induced cell proliferation. Cis contains an Src homology 2 domain in the central part of the molecule, and it was previously shown to associate to the cytoplasmic domain of the tyrosine-phosphorylated Epo receptor and the tyrosine-phosphorylated ␤ chain of the IL-3 receptor. No catalytic domain was detected in the Cis protein, and Cis does not seem to be tyrosine phosphorylated or to contain another known protein association module. Thus, Cis appears to share neither the characteristics of an effector protein nor those of an adapter protein. Cis has a calculated molecular mass of 29 kDa but migrates as multiple bands in polyacrylamide gels. The Cis protein has a rapid turnover, because it is quickly expressed in cytokine-stimulated cells and disappears rapidly after cytokine starvation (23). The mechanism of action of Cis is currently unknown, although it has been shown that Cis overexpression reduced STAT5 activation (24).
Recently a family of proteins that are structurally related to Cis has been cloned (25)(26)(27). SOCS1 (suppressor of cytokine signaling 1), also named JAB (for Jak-binding protein) and SSI-1 (for STAT-induced STAT inhibitor 1) inhibits the differentiation of M1 cells into macrophages in response to IL-6. Its expression is regulated by STATs, and it can reduce Jak tyrosine kinase activity by direct interaction with Jak family members, suggesting that SOCS1 could act in a classic negative feedback loop in the regulation of cytokine signal transduction. Two relatives of SOCS1 named SOCS2 and SOCS3 have also been cloned. Their roles have not been determined.
In this report, we show that Cis associates to the phosphorylated second tyrosine residue (Tyr 401 ) of the Epo receptor, which is one of the two major STAT5 activation sites. We also observed that Cis is a ubiquitinated protein and that proteasome inhibitors induced the accumulation of the ubiquitinated form of Cis and the accumulation of the Cis-Epo receptor complexes.

EXPERIMENTAL PROCEDURES
Chemicals and Reagents-Highly purified recombinant human Epo (specific activity, 120 000 units/mg) used throughout this study was a generous gift from Dr. M. Brandt (Boehringer Mannheim). Recombinant SCF was kindly provided by Dr. A. Shimosaka (Kirin Brewery Co., Tokyo, Japan). The proteasome inhibitors LLnL and lactacystin were purchased from Sigma and Calbiochem, respectively. Anti-Cis antibodies were produced as described previously (23). During the achievement of this work, commercial anti-Cis antibodies (SC1569; Santa Cruz Biotechnology, Santa Cruz, CA) were made available and were used for some Western blot experiments. Anti-phosphotyrosine antibodies (4G10) were a generous gift from Dr B. Druker (Dana Farber Cancer Institute, Boston, MA). Anti-Epo receptor antibodies were produced by immunizing rabbits with a recombinant protein composed of the full intracellular domain of the human Epo receptor fused to GST or to MalE. Anti-GST and anti-MalE antibodies were also prepared in our laboratory using the same protocol. Whole anti-GST, anti-MalE, and anti-Epo receptor sera were used. We produced anti-STAT5A and anti-STAT5B antibodies by immunizing rabbits with peptides corresponding to the last 12 amino acids (AGLFTSARSSLS) of STAT5A or the last 8 amino acids of STAT5B (QWIPHAQS) coupled to keyhole limpet hemocyanin. These peptidic sequences are specific for STAT5A and STAT5B, respectively. A mixture of both antibodies was used in the experiments reported in this paper. Anti-ubiquitin antibodies were purchased from Santa Cruz (catalogue number SC6085) and from Zymed Laboratories Inc. (San Francisco, CA; catalogue number 13-1600).
Cell Lines and Cell Culture-A subclone of the human leukemic cell line UT-7 (28) able to grow in SCF, GM-CSF, or Epo was established. These cells were cultured in ␣-minimum essential medium containing 5% fetal calf serum complemented with 2 units/ml Epo, 2.5 ng/ml GM-CSF, or 50 ng/ml SCF. A clone of these cells overexpressing the Cis protein was established by transfecting the Cis cDNA in pME vector (pME-Cis) (23) together with a pBabe plasmid carrying a neomycin resistance gene. Transfected cells were selected for neomycin resistance in the presence of SCF and tested for Cis expression by Western blot analysis. Before each experiment, the cells were serum and growth factor deprived by incubation overnight in Iscove's Dulbecco's modified Eagle's medium (Life Technologies, Inc.; catalogue number 31980-022) containing 0.4% deionized bovine serum albumin. The same pMe-Cis plasmid was also used to transiently transfect COS7 cells (ECACC number 87021302) using DEAE-dextran (29). Mo7E cells (30) and TF-1 cells (31,32) were cultivated in ␣-minimum essential medium containing 5% fetal calf serum and complemented with 2.5 ng/ml GM-CSF. BaF3 (33) and FDCP1 (34) cells were cultivated in RPMI 1640 medium and ␣-minimum essential medium-5% fetal calf serum, respectively, complemented with 3% WEHI-conditioned medium as a source of IL-3. WEHI-conditioned medium is the supernatant of end-logarithmic phase cultures of WEHI-3B cells (ATCC number TIB-68). BaF3 cells expressing dexamethasone-inducible Cis were described previously (BF-Cis) (23). These cells were transfected with expression vectors encoding wild-type or mutated Epo receptors and selected for their ability to grow in Epo as described previously (35). Before each experiment, BF-Cis cells were cultured 15 h in RPMI 1640 medium complemented with 5% fetal calf serum, 0.3% WEHI, and 100 nM dexamethasone to induce Cis expression.
Epo Receptor Mutants-Several mutants of the Epo receptor were used throughout this study (see Fig. 1). These mutants have been previously described (17,35). All of these constructs were inserted in a modified pCDNA 3 vector in which the cytomegalovirus promoter was exchanged for a Rous sarcoma virus promoter.
Cis Detection in Whole-cell Extracts-Exponentially growing hematopoietic cells were washed twice with phosphate-buffered saline and solubilized by boiling in electrophoresis sample buffer at a cell density of 5 ϫ 10 6 cells/ml. Samples corresponding to 5 ϫ 10 5 cells were separated by SDS-polyacrylamide gel electrophoresis using 12.5% polyacrylamide gels and analyzed by Western blot.

Cis Associates with the Second Tyrosine Residue of the Epo
Receptor-To determine which tyrosine of the Epo receptor binds to Cis, BF-Cis cells were transfected with the wild-type Epo receptor or with several mutants of the Epo receptor ( Fig.  1). Because available anti-Epo receptor antibodies are poorly efficient in Western blot experiments, and because the tyrosine phosphorylation of some of the Epo receptor mutants is difficult to detect, we first used an indirect method to test the association of Cis with the Epo receptor (12). The cells were stimulated for 10 min with 2 units/ml 125 I-Epo and solubilized. Cellular extracts were immunoprecipitated with anti-Epo receptor antibodies, with anti-GST control antibodies, or with anti-Cis antibodies, and immunoprecipitated radioactivity was measured by ␥ counting. Immunoprecipitations using anti-Epo receptor antibodies showed that all transfected cells bound high amounts of 125 I-Epo. In contrast, these antibodies immunoprecipitated a very low amount of 125 I-Epo in nontransfected cells. Because the same amount of 125 I-Epo was also precipitated by anti-GST control antibodies, this low binding in nontransfected cells corresponded to nonspecific binding. As shown in Fig. 2, anti-Cis antibodies immunoprecipitated 125 I-Epo bound to the wild-type Epo receptor, but they did not precipitate Epo bound to an Epo receptor devoid of tyrosine residues in its intracellular domain (F1⌬Y2-8). Moreover, these experiments showed an absolute requirement for the second tyrosine residue of the Epo receptor for Cis association. Indeed, Cis also associated with the F1Y2⌬Y3-8 mutant receptor, but it did not associate with mutant Epo receptors, which did not possess this tyrosine residue. Interestingly, the single mutation of the second tyrosine residue of the intracellular domain of the Epo receptor (Tyr 401 ) fully inhibited Cis binding to the Epo receptor (see mutant Y1F2Y3-8). To confirm this result, Epo receptor-transfected BF-Cis cells were stimulated with 10 units/ml unlabeled Epo and solubilized. Cellular extracts were immunoprecipitated with anti-Cis antibodies and analyzed by Western blot using anti-phosphotyrosine antibodies. As shown in Fig. 3, Cis bound to the wild-type Epo receptor but not to the Y1F2Y3-8 Epo receptor mutant. Control immunoprecipitations made using anti-Epo receptor antibodies revealed that the wild-type Epo receptor and the Y1F2Y3-8 Epo receptor mutant were equally tyrosine phosphorylated after Epo stimulation.
Then, we investigated the association of Cis with the Epo receptor in the human UT-7 cells, which naturally express the Epo receptor. However, because Cis gene expression is regulated by STAT5, unstimulated cells do not contain any Cis protein. To induce Cis gene expression, UT-7 cells were prestimulated for 1.5 h with a low level (2.5 ng/ml) of GM-CSF. Cells were then stimulated for 10 min with 10 units/ml Epo, and the cell extracts were immunoprecipitated with anti-Cis antibodies. Western blot using anti-phosphotyrosine antibodies (Fig. 4A) showed that the Epo receptor was also immunoprecipitated with Cis in UT-7 cells. To overcome the requirement of STAT5 activation by a cytokine receptor to induce the expression of the endogenous Cis gene, we produced a UT-7 cell line expressing Cis under the control of a constitutive SR␣ promoter. These cells were selected and grown in the presence of SCF, which does not activate STAT5 in UT-7 cells (38). UT-7 cells expressing Cis (Cl8 cells) as well as UT-7 cells that did not express Cis (Cl20 cells) were stimulated for 10 min with 10 units/ml Epo and solubilized. As shown in Fig. 4B, the Epo receptor was also immunoprecipitated by anti-Cis antibodies in cellular extracts of UT-7 Cl8 cells.
Cis Is a Ubiquitinated Protein That Seems to Be Degraded by the Proteasome-Anti-Cis antibodies recognized several pro-teins in both transfected cells and in cells naturally expressing the Cis protein. Indeed, although the theoretical molecular mass of Cis is 29 kDa, Western blots revealed the expression of two main Cis proteins with apparent molecular masses of 32 and 37 kDa and a minor species of 45-47 kDa molecular mass (23). To study the relationship between these proteins, we performed a transcription-translation experiment, and we compared the size of the in vitro Cis translated protein with that of the Cis proteins expressed in COS cells. The in vitro translated protein corresponded to the 32-kDa Cis protein (Fig. 5A), thereby indicating that the 37-kDa Cis protein resulted from post-translation modifications. Considering these molecular masses and the quick degradation of Cis, an attractive hypothesis would be that Cis was a ubiquitinated protein. To test this hypothesis, Cis proteins were immunoprecipitated from Cistransfected COS cells using anti-Cis antibodies and analyzed by Western blot using anti-ubiquitin antibodies. As shown in Fig. 5B, anti-ubiquitin antibodies recognized the 37-kDa form

FIG. 4. Cis associates with the Epo receptor in UT-7 cells.
A, starved UT-7 cells were incubated for 1.5 h with 2.5 ng/ml GM-CSF to induce endogenous Cis gene expression. The cells were then stimulated for 10 min with 10 units/ml Epo and lysed using 1% Brij 96. The lysates were cleared by centrifugation (27,000 ϫ g, 15 min) and immunoprecipitated with anti-Epo receptor antibodies (ER), anti-Cis antibodies (C), or anti-GST control antibodies (G). Immunoprecipitates were analyzed by Western blot (WB) using anti-phosphotyrosine antibodies (anti-PY). B, UT-7 cells were stably transfected with a Cis expression vector, and a clone of cells overexpressing Cis (Cl8) was selected. Another clone, which does not express the transfected Cis protein (Cl20), was used as control. The cells were serum and growth factor starved for 15 h and incubated for 10 min with 10 units/ml Epo (ϩ) or with vehicle alone (Ϫ). After cell solubilization, the cellular extracts were immunoprecipitated (IP) using anti-Cis, and immunoprecipitates were analyzed by Western blot (WB) using anti-phosphotyrosine (anti-PY) antibodies. of Cis. Reprobing the blot with anti-Cis antibodies evidenced both the 37-and 32-kDa forms of Cis. Similar results were obtained using another anti-ubiquitin antibody from Zymed (data not shown). Because ubiquitinated proteins are often degraded through the proteasome pathway, we tested the effects of the proteasome inhibitors LLnL and lactacystin (39) on the expression of the Cis protein. As shown in Fig. 5, C and D, the proteasome inhibitors induced the accumulation of the 37-kDa Cis protein in UT-7 Cl8 cells and in BF-Cis cells. To confirm that the accumulation of the 37-kDa form of Cis was attributable to the inhibition of Cis degradation and not to an increase of its synthesis induced by the proteasome inhibitors, we blocked the protein synthesis in UT-7 Cl8 cells using 0.5 mM cycloheximide (40). As shown in Fig. 5E, the 37-kDa form of Cis was undetectable in cells incubated for 30 min with cycloheximide, and the amount of the 32-kDa form of Cis was strongly decreased. These results confirm that the half-life of Cis was very short. In contrast, when the cells were incubated with both proteasome inhibitors and cycloheximide, the amount of Cis protein remained unchanged during the 30 min of incubation, demonstrating that proteasome inhibitors inhibited the degradation of Cis.
Proteasome Inhibitors Inhibit the Inactivation of the Epo Receptor and of STAT5-Because Cis is a ubiquitinated protein that is protected from degradation by proteasome inhibitors, we tested the effects of LLnL on the activation of the target of Cis, the Epo receptor. As reported in Fig. 6A, LLnL did not modify the activation of the Epo receptor. Indeed, LLnL neither induced the tyrosine phosphorylation of the Epo receptor in the absence of Epo nor modified the level of activation of the Epo receptor during the first 10 min of Epo stimulation. In contrast, LLnL strongly inhibited the dephosphorylation of the Epo receptor after a long-term incubation. As shown in Fig. 6A, Epo receptor phosphorylation was not decreased after 90 min of Epo stimulation in cells treated with the proteasome inhibitor, whereas it returned to basal level at this time in untreated cells. Similar effects were observed concerning the kinetics of STAT5 activation (Fig. 6B). Lactacystin, another proteasome inhibitor, also protected the Epo receptor and STAT5 from inactivation with identical efficiency as LLnL (data not shown). Thus, the proteasome is involved in the inactivation of both the Epo receptor and STAT5.

Proteasome Inhibitors Protect Cis-Epo Receptor Complexes-
The association kinetics of Cis with the tyrosine-phosphorylated Epo receptor was studied using UT-7 Cl8 cells (Fig. 7A). In the absence of proteasome inhibitors, the association between Cis and the Epo receptor was maximal after 10 min of Epo stimulation and started to decrease thereafter to become nearly undetectable after 2 h of Epo stimulation. In the presence of proteasome inhibitors, the association between Cis and the Epo receptor remained constant throughout the two hours of Epo stimulation. Fig. 7A also shows that the level of Epo receptor associated with Cis was strongly increased by proteasome inhibitors. To confirm this result, UT-7 Cl8 cells were stimulated for 10 min with Epo, and the Cis-Epo receptor complexes were immunoprecipitated using a saturating concentration of anti-Cis antibodies. Cis-depleted extracts were then immunoprecipitated using anti-Epo receptor antibodies. Fig. 7B shows that ϳ10% of the tyrosine-phosphorylated Epo receptors were found associated with Cis in the absence of proteasome inhibitors, whereas 50% of the Epo receptors were associated with Cis in the presence of LLnL. Similar results were obtained using lactacystin (data not shown). Thus, proteasome inhibitors induced the accumulation of Cis-Epo receptor complexes.

DISCUSSION
Cis belongs to a family of proteins that to date comprises four members in addition to Cis (SOCS1, also called JAB or SS1, SOCS2, also called Cis2, SOCS3 also called Cis3, and Cis4). Cis, SOCS1, SOCS2, and SOCS3 are composed of 210 -260 amino acids, and they exhibit a similar structure with an Src homology 2 domain located in the middle of the protein and a conserved motif of ϳ40 amino acids (the SOCS box) located at the C-terminal part of the molecules. Cis4 possesses a longer N-terminal region of ϳ380 amino acids (41). At least two of these proteins, Cis and SOCS1, appear to be involved in a negative feedback regulation process that controls the activation of several cytokine receptors. Indeed, the syntheses of Cis and SOCS1 are controlled by STAT5 and STAT3 transcription factors, respectively, and overexpression of these proteins leads to a decreased activation of these transcription factors in cytokine-stimulated cells. Moreover, Cis overexpression decreases the cell proliferation induced by Epo or IL-3, and SOCS1 overexpression inhibits the macrophage differentiation of M1 cells induced by IL-6. However, Cis and SOCS1 appear to have different targets. Indeed, it has been shown that SOCS1 associates with the tyrosine-phosphorylated Jak1, Jak2, Jak3, and Tyk2 tyrosine kinases (25,26). This association involves the kinase domain (JH1) of the Jak kinases and induces a decrease of the kinase activity. In contrast, Cis associates with the Epo and the IL-3 receptors (23) and not with Jaks (25). Moreover, SOCS1 rapidly reduced IL-6-induced gp130 tyrosine phosphorylation (26), whereas we did not observe a decrease of Epo receptor phosphorylation in UT-7 Cl8 cells at the early times of Epo stimulation (data not shown).
Our results show that the 37-kDa form of Cis is a ubiquiti-nated protein, which seems to be quickly degraded by the proteasome pathway (Fig. 5). Indeed, the 37-kDa form was directly recognized by anti-ubiquitin antibodies, and two different proteasome inhibitors induced the intracellular accumulation of the 37-kDa form of Cis. Although protein degradation by the proteasome usually requires the polyubiquitination of the target protein, previous reports have shown that monoubiquitinated ␣-globin was also degraded by the proteasome (42,43). Thus, the 37-kDa form of Cis could be another example of a monoubiquitinated protein degraded through the proteasome pathway. A 45-kDa form of Cis was also previously described (23). Although we did not detect this protein in most experiments, we sometimes observed a faint band at 45 kDa in anti-Cis antibody Western blots using proteins from transfected COS cells. Interestingly, this band was also faintly stained with anti-ubiquitin antibodies, suggesting that this minor form was also ubiquitinated and that polyubiquitinated forms of Cis could also occur. A plausible explanation would be that these polyubiquitinated forms could be the target of the proteasomes but are trimmed down by ubiquitin isopeptidases when the proteolytic activity of the proteasome was blocked. The ubiquitination of Cis and its degradation by the proteasome could account for the previously reported short half-life of the protein. In contrast to Cis synthesis, which requires the activation of STAT5, Cis degradation does not seem to be regulated by the activation of the cytokine receptors, because it was evident both in nonstimulated cells (Fig. 5) and in Epostimulated cells (Fig. 7). We show that Cis associates with the second tyrosine residue . The cells were then stimulated with 10 units/ml Epo for the indicated times and solubilized. Cellular extracts were then immunoprecipitated with anti-Cis antibodies, and immunoprecipitates were analyzed by Western blot (WB) using anti-phosphotyrosine (anti-PY) and anti-Cis antibodies successively. B, starved UT-7 Cl8 cells were incubated for 15 min with 50 M LLnL (ϩ) or with vehicle alone (0.1% dimethylsulfoxide; Ϫ). The cells were then stimulated with 10 units/ml Epo for 10 min, and cellular extracts were immunoprecipitated (IP) with a saturating concentration of anti-Cis antibodies (C). Nonprecipitated material was then immunoprecipitated with anti-Epo receptor antibodies (ER). Immunoprecipitates were analyzed by Western blot (WB) using anti-phosphotyrosine (anti-PY) antibodies. The faint band at 116 kDa corresponds to the Epo receptor-associated Jak2 protein. and 3). This tyrosine residue is one of the major activation sites for STAT5 (17,18), and it has been shown that Cis decreased the transcription of STAT5 target genes (24). A possible explanation would be that STAT5 and Cis compete for binding to the same tyrosine residue of the Epo receptor. However, this explanation seems unlikely, because in the absence of Tyr 401 , complete STAT5 activation could also be achieved through Tyr 343 of the Epo receptor (10,(17)(18)(19)44), and this tyrosine residue does not associate with Cis. Moreover, in the absence of proteasome inhibitors, the fraction of tyrosine-phosphorylated Epo receptors associated with Cis is rather low (Fig. 7B), suggesting that most Tyr 401 residues are probably available for STAT5 activation. Alternatively, Cis could decrease the activation of STAT5 by inducing the inactivation of the Epo receptor through the proteasome pathway. Indeed, the proteasome inhibitors LLnL and lactacystin inhibited the inactivation of the Epo receptor and STAT5 that naturally occurred after a longterm Epo stimulation. The fraction of the Epo receptor associated with Cis was strongly increased in the presence of proteasome inhibitors, and these complexes accumulated in the cells treated with proteasome inhibitors (Fig. 7). An attractive hypothesis would be that Cis targets the Epo receptor for degradation. It has been previously reported that the degradation of the IL-3 receptor ␤ chain required its ligand-induced phosphorylation (45). This mechanism is consistent with the hypothesis that Cis, which binds to the tyrosine-phosphorylated Epo and IL-3 receptors, could be responsible for the degradation of these receptors and for the termination of the activation process. In a recent paper, Yu and Burakoff (46) have shown that the proteasome pathway regulates the IL-2-stimulated Jak-STAT pathways, although neither STAT nor Jak appears to be ubiquitinated. Because IL-2 was also shown to induce Cis gene expression (23), it would be interesting to determine whether Cis also binds to the activated IL-2 receptor and whether it could be responsible for the reported effects.
Part of the dephosphorylation of the Epo receptor has been previously attributed to the activation of the tyrosine phosphatase SHP-1 and to its binding to both the Epo receptor and the Jak2 tyrosine kinase (21,22). Our experiments show that another Epo receptor inactivation process involves the proteasomes. Thus, at least two mechanisms could account for the termination of the Epo receptor signaling; the dephosphorylation of Jak2 by the tyrosine phosphatase SHP-1 and the inactivation of the receptor by the proteasomes. The relationship between each inactivation process has now to be studied in the aim of understanding the relative importance of each mechanisms.