Involvement of Proteasomes in Regulating Jak-STAT Pathways upon Interleukin-2 Stimulation

Interleukin-2 (IL-2) activates the receptor-associated Janus family tyrosine kinases, Jak1 and Jak3, which in turn phosphorylate and activate specific STAT proteins (signal transducers and activators of transcription), such as STAT5. Activation of Jak and STAT proteins by IL-2 is transient and the mechanism for the subsequent down-regulation of their activity is largely unknown. We report here that IL-2-induced DNA-binding activity and tyrosine phosphorylation of STAT5 are stabilized by a proteasome inhibitor MG132; however, no detectable ubiquitination of the STAT proteins is observed. This sustained STAT5 activation can be blocked by protein kinase inhibitors, which is consistent with the ability of the proteasome inhibitor to stabilize IL-2-induced tyrosine phosphorylation of Jak1 and Jak3. These results suggest that proteasome-mediated protein degradation modulates protein-tyrosine phosphatase activity that negatively regulates the Jak-STAT signaling pathways. Interleukin-2 (IL-2) activates the receptor-associated Janus family tyrosine kinases, Jak1 and Jak3, which in turn phosphorylate and activate specific STAT proteins (signal transducers and activators of transcription), such as STAT5. Activation of Jak and STAT proteins by IL-2 is transient and the mechanism for the subsequent down-regulation of their activity is largely unknown. We report here that IL-2-induced DNA-binding activity and tyrosine phosphorylation of STAT5 are stabilized by a proteasome inhibitor MG132; however, no detectable ubiquitination of the STAT proteins is observed. This sustained STAT5 activation can be blocked by protein kinase inhibitors, which is consistent with the ability of the proteasome inhibitor to stabilize IL-2-induced tyrosine phosphorylation of Jak1 and Jak3. These results suggest that proteasome-mediated protein degradation modulates protein-tyrosine phosphatase activity that negatively regulates the Jak-STAT signaling pathways. Important insights into gene regulation by cytokines have been gained from recent studies of the Janus kinase (Jak) 1The abbreviations used are:JakJanus kinaseSTATsignal transducer and activator of transcriptionIFN-γinterferon-γIL-2interleukin-2mAbmonoclonal antibodyEMSAelectrophoretic mobility shift assayPBSphosphate-buffered saline. family of nonreceptor protein-tyrosine kinases and the signal transducers and activators of transcription (STATs) (1Darnell Jr., J.E. Kerr I.M. Stark G.R. Science. 1994; 264: 1415-1421Crossref PubMed Scopus (5062) Google Scholar, 2Ihle J.N. Nature. 1995; 377: 591-594Crossref PubMed Scopus (1145) Google Scholar, 3Taniguchi T. Science. 1995; 268: 251-255Crossref PubMed Scopus (676) Google Scholar). Members of the cytokine receptor superfamily recruit the Jak family kinases to phosphorylate and activate downstream STAT proteins. STAT proteins are latent cytoplasmic transcription factors that, upon activation by tyrosine phosphorylation, translocate to the nucleus and bind to specific regulatory elements that control gene expression. Diversity in cellular responses to different cytokines is determined in part by selective association of cytokine receptors with different Jak kinases and subsequent phosphorylation of different STAT proteins (4Leonard W.J. Nat. Med. 1996; 2: 968-969Crossref PubMed Scopus (44) Google Scholar, 5Stahl N. Farruggella T.J. Boulton T.G. Zhong Z. Darnell Jr., J.E. Yancopoulos G.D. Science. 1995; 267: 1349-1353Crossref PubMed Scopus (869) Google Scholar, 6Pernis A. Gupta S. Yopp J. Garfein E. Kashleva H. Schindler C. Rothman P. J. Biol. Chem. 1995; 270: 14517-14522Abstract Full Text Full Text PDF PubMed Scopus (30) Google Scholar). Janus kinase signal transducer and activator of transcription interferon-γ interleukin-2 monoclonal antibody electrophoretic mobility shift assay phosphate-buffered saline. Similar to many other mitogen-induced signaling events (7Wange R.L. Samelson L.E. Immunity. 1996; 5: 197-205Abstract Full Text Full Text PDF PubMed Scopus (462) Google Scholar, 8Kruijer W. Cooper J.A. Hunter T. Verma I.M. Nature. 1984; 312: 711-716Crossref PubMed Scopus (570) Google Scholar, 9Hipskind R.A. Baccarini M. Nordheim A. Mol. Cell. Biol. 1994; 14: 6219-6231Crossref PubMed Scopus (137) Google Scholar), cytokine-induced activation of the Jak-STAT signal transduction pathway is both rapid and transient (10Shuai K. Schindler C. Prezioso V.R. Darnell Jr., J.E. Science. 1992; 259: 1808-1812Crossref Scopus (659) Google Scholar, 11Shuai K. Ziemiecki A. Wilks A.F. Harpur A.G. Sadowski H.B. Gilman M.Z. Darnell Jr., J.E. Nature. 1993; 366: 580-583Crossref PubMed Scopus (411) Google Scholar, 12Silvennoinen O. Ihle J.N. Schlessinger J. Levy D.E. Nature. 1993; 366: 583-585Crossref PubMed Scopus (290) Google Scholar). While the mechanisms involved in Jak and STAT activation are being delineated, the mechanisms underlying their subsequent deactivation are still largely unknown. Based on the observation that there is a good correlation between tyrosine dephosphorylation and inactivation of both Jak and STAT proteins, the roles of protein-tyrosine phosphatases in down-regulating the Jak-STAT pathways have been suggested. Indeed, there is genetic and biochemical evidence indicating that protein-tyrosine phosphatases are critical in regulating distinct Jak kinases within specific cytokine receptor complexes (13David M. Chen H.E. Goelz S. Larner A.C. Neel B.G. Mol. Cell. Biol. 1995; 15: 7050-7058Crossref PubMed Scopus (318) Google Scholar, 14Klingmuller U. Lorenz U. Cantley L.C. Neel B.G. Lodish H.F. Cell. 1995; 80: 729-738Abstract Full Text PDF PubMed Scopus (842) Google Scholar, 15Yi T. Mui A.L.-F. Krystal G. Ihle J.N. Mol. Cell. Biol. 1993; 13: 7577-7586Crossref PubMed Google Scholar). The evidence for the involvement of specific protein-tyrosine phosphatases in dephosphorylating STAT proteins, however, is more indirect (16David M. Grimley P.M. Finbloom D.S. Larner A.C. Mol. Cell. Biol. 1993; 13: 7515-7521Crossref PubMed Scopus (102) Google Scholar, 17Haque S.J. Flati V. Deb A. Williams B.R.G. J. Biol. Chem. 1995; 270: 25709-25714Abstract Full Text Full Text PDF PubMed Scopus (92) Google Scholar, 18Haspel R.L. Salditt-Georgieff M. Darnell Jr., J.E. EMBO J. 1996; 15: 6262-6268Crossref PubMed Scopus (272) Google Scholar). Another important mechanism in regulating signal transduction is proteolysis (19Ciechanover A. Cell. 1994; 79: 13-21Abstract Full Text PDF PubMed Scopus (1602) Google Scholar, 20Goldberg A.L. Rock K.L. Nature. 1992; 357: 375-379Crossref PubMed Scopus (506) Google Scholar). The ubiquitin-proteasome pathway, in particular, plays an important role in degrading a number of cellular proteins, including transcription factors (21Palombella V.J. Rando O.J. Goldberg A.L. Maniatis T. Cell. 1994; 78: 773-785Abstract Full Text PDF PubMed Scopus (1922) Google Scholar, 22Trier M. Staszewski L.M. Bohmann D. Cell. 1994; 78: 787-798Abstract Full Text PDF PubMed Scopus (847) Google Scholar, 23Stanscovski I. Gpmem H. Orian A. Schwartz A.L. Ciechanover A. Mol. Cell. Biol. 1995; 15: 7106-7116Crossref PubMed Scopus (159) Google Scholar). The target proteins are first tagged with multiple molecules of the small protein ubiquitin and then destroyed by the multisubunit proteasome complex. Recently, there have been reports indicating the potential role of proteasome-mediated proteolysis in down-regulating STAT1 activity following interferon-γ (IFN-γ) stimulation. However, one report provided evidence that STAT1 itself is ubiquitinated and degraded (24Kim T.K. Maniatis T. Science. 1996; 273: 1717-1719Crossref PubMed Scopus (361) Google Scholar), while another report suggested that its upstream signaling pathway is the target of proteasomes (18Haspel R.L. Salditt-Georgieff M. Darnell Jr., J.E. EMBO J. 1996; 15: 6262-6268Crossref PubMed Scopus (272) Google Scholar). In this report, we address the possible role of proteasome-mediated protein degradation in modulating STAT5 activity following stimulation with interleukin-2 (IL-2) to determine if proteasomes are involved in regulating other STAT proteins in response to diverse cytokines. Furthermore, we investigate whether the STAT proteins themselves or the upstream signaling molecules, such as the Jak kinases, are modulated through the ubiquitin-proteasome pathway. Maintenance of the IL-2-dependent cell line CTLL-20 has been described earlier (25Ravichandran K.S. Igras V. Shoelson S.E. Fesik S.W. Burakoff S.J. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 5275-5280Crossref PubMed Scopus (92) Google Scholar). For IL-2 stimulation experiments, exponentially growing CTLL-20 cells were washed three times in RPMI 1640 medium, resuspended in medium without IL-2, and starved for 4 h at 37 °C. IL-2-deprived CTLL-20 cells were then stimulated with 30 units/ml of recombinant human IL-2 for various times as indicated in the figure legends. Recombinant human IL-2 was a gift of Dr. J. Sepinwall (Hoffmann-La Roche Inc., Nutley, NJ). The oligonucleotides specific for STAT5 binding were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA) (sc-2565). The proteasome inhibitor MG132 and calpain inhibitor II were from Peptide Institute, Inc. (Louisville, KY) and Boehringer Mannheim, respectively. Me2SO, genistein, and staurosporine were from Sigma. Anti-STAT5 antibodies used for immunoprecipitation and immunoblotting were from Santa Cruz Biotechnology, Inc. (sc-835) and Transduction Laboratories (Lexington, KY) (S21520), respectively. Anti-Jak1 antibodies used for immunoprecipitation and immunoblotting were from Upstate Biotechnology, Inc. (Lake Placid, NY) (06-272) and Transduction Laboratories (J24320), respectively. Anti-Jak3 antibodies used for immunoprecipitation and immunoblotting were from Upstate Biotechnology (06–342) and Santa Cruz Biotechnology, Inc. (sc-1079), respectively. Horseradish peroxidase-conjugated anti-phosphotyrosine antibody RC20H was obtained from Transduction Laboratories. CTLL-20 cells were washed once with PBS supplemented with 1 mmNa3VO4 (ortho) and once with PBS supplemented with 1 mm Na3VO4 (ortho) and 5 mm NaF. Nuclear extracts were prepared from washed cell pellets essentially as described earlier (26Yu C.-L. Meyer D.J. Campbell G.S. Larner A.C. Carter-Su C. Schwartz J. Jove R. Science. 1995; 269: 81-83Crossref PubMed Scopus (824) Google Scholar), except for the modification of the hypotonic buffer (10 mm HEPES, pH 7.9, 1.5 mm MgCl2, and 10 mm KCl). STAT5 gel shift oligonucleotides were labeled with T4 polynucleotide kinase in the presence of γ-32PATP. EMSA was performed as described previously (26Yu C.-L. Meyer D.J. Campbell G.S. Larner A.C. Carter-Su C. Schwartz J. Jove R. Science. 1995; 269: 81-83Crossref PubMed Scopus (824) Google Scholar). CTLL-20 cells were washed once with PBS supplemented with 1 mmNa3VO4 (ortho) and once with PBS supplemented with 1 mm Na3VO4 (ortho) and 10 mm NaF. Preparation of cell lysates, immunoprecipitation, and immunoblotting were done essentially as described elsewhere (24Kim T.K. Maniatis T. Science. 1996; 273: 1717-1719Crossref PubMed Scopus (361) Google Scholar,27Chen Z. Hagler J. Palombella V.J. Melandri F. Scherer D. Ballard D. Maniatis T. Genes Dev. 1995; 9: 1586-1597Crossref PubMed Scopus (1172) Google Scholar). Dilutions of different antibodies for immunoprecipitation and immunoblotting and subsequent detection by the enhanced chemiluminescence (ECL) system were performed as recommended by the manufacturers. To determine if proteasome-mediated proteolysis is involved in modulating STAT activity in response to cytokines, such as IL-2, we examined the effects of a proteasome inhibitor (MG132) on STAT5 activation induced by IL-2 (28Gaffen S.L. Lai S.Y. Xu W. Gouilleux F. Groner B. Goldsmith M.A. Greene W.C. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7192-7196Crossref PubMed Scopus (86) Google Scholar, 29Wakao H. Harada N. Kitamura T. Mui A.L. Miyajima A. EMBO J. 1995; 14: 2527-2535Crossref PubMed Scopus (214) Google Scholar). An IL-2-dependent cell line CTLL-20 was deprived of IL-2, pretreated with the carrier Me2SO, MG132, or a cysteine protease inhibitor (calpain inhibitor II) for 1 h and then stimulated with IL-2 for 15 min to 2 h. We analyzed nuclear extracts by EMSA using a labeled probe specific for STAT5 DNA binding activity (Fig. 1). In Me2SO-treated cells, IL-2 induced rapid and transient STAT5 DNA binding activity (lanes 1–6). This DNA binding activity involves both STAT5a and STAT5b, as determined by supershift assays using specific antibodies (data not shown). This IL-2-induced STAT5 DNA binding activity was sustained in the presence of MG132 (lanes 7–12), but not calpain inhibitor II (lanes 13–15). This result demonstrates that proteasome-mediated protein degradation, but not proteolysis in general, is involved in modulating IL-2-induced STAT5 activity. In many cases, phosphorylation of target proteins is required for their ubiquitination and subsequent degradation by proteasomes (24Kim T.K. Maniatis T. Science. 1996; 273: 1717-1719Crossref PubMed Scopus (361) Google Scholar, 27Chen Z. Hagler J. Palombella V.J. Melandri F. Scherer D. Ballard D. Maniatis T. Genes Dev. 1995; 9: 1586-1597Crossref PubMed Scopus (1172) Google Scholar, 30Deshaies R.J. Chau V. Kirschner M. EMBO J. 1995; 14: 303-312Crossref PubMed Scopus (188) Google Scholar, 31Yaglom J. Linskens M.H.K. Sadis S. Rubin D.M. Futcher B. Finley D. Mol. Cell. Biol. 1995; 15: 731-741Crossref PubMed Google Scholar). To determine whether active STAT5 is directly degraded by proteasomes, we prepared cell lysates from IL-2-stimulated cells in the absence or presence of MG132 under conditions that should preserve ubiquitinated proteins (24Kim T.K. Maniatis T. Science. 1996; 273: 1717-1719Crossref PubMed Scopus (361) Google Scholar, 27Chen Z. Hagler J. Palombella V.J. Melandri F. Scherer D. Ballard D. Maniatis T. Genes Dev. 1995; 9: 1586-1597Crossref PubMed Scopus (1172) Google Scholar). Anti-STAT5 immunoprecipitates were prepared from cell lysates and analyzed by protein immunoblotting with antibody to phosphotyrosine or anti-STAT5 mAb (Fig. 2). The anti-STAT5 antibodies used for immunoprecipitation and immunoblotting recognize both STAT5a and STAT5b. As a result, STAT5 proteins appear as broad bands on immunoblots, and the bands become even broader when some of the STAT5 proteins are phosphorylated and migrate slower in SDS-polyacrylamide gels (32Lin J.X. Mietz J. Modi W.S. John S. Leonard W.J. J. Biol. Chem. 1996; 271: 10738-10744Abstract Full Text Full Text PDF PubMed Scopus (195) Google Scholar). Consistent with its DNA binding activity, IL-2-induced tyrosine phosphorylation of STAT5 is prolonged by MG132 (Fig. 2, lanes 9–14 compared with lanes 3–8). Proteins conjugated with multiple molecules of ubiquitin appear as a ladder of higher molecular mass proteins in SDS-polyacrylamide gels (24Kim T.K. Maniatis T. Science. 1996; 273: 1717-1719Crossref PubMed Scopus (361) Google Scholar, 27Chen Z. Hagler J. Palombella V.J. Melandri F. Scherer D. Ballard D. Maniatis T. Genes Dev. 1995; 9: 1586-1597Crossref PubMed Scopus (1172) Google Scholar). We could not detect any ubiquitinated STAT5 in the presence of MG132 either in phosphotyrosine or STAT5 immunoblots. This result suggests that inactivation of STAT5 following IL-2-induced activation is mostly due to dephosphorylation, but not removal of active STAT5 through a ubiquitin-proteasome pathway. To further confirm that tyrosine-phosphorylated STAT5 is not the direct target of proteasomes, we examined the effects of protein kinase inhibitors on IL-2-induced DNA-binding activity (Fig.3 A) and tyrosine phosphorylation (Fig.3 B) of STAT5. Two protein kinase inhibitors were tested: genistein, a protein-tyrosine kinase inhibitor, and staurosporine, a broad spectrum protein kinase inhibitor. Both inhibitors have been shown to block cytokine-induced STAT activation (18Haspel R.L. Salditt-Georgieff M. Darnell Jr., J.E. EMBO J. 1996; 15: 6262-6268Crossref PubMed Scopus (272) Google Scholar, 24Kim T.K. Maniatis T. Science. 1996; 273: 1717-1719Crossref PubMed Scopus (361) Google Scholar, 33Bhat G.J. Thekkumkara T.J. Thomas W.G. Conrad K.M. Baker K.M. J. Biol. Chem. 1994; 269: 31443-31449Abstract Full Text PDF PubMed Google Scholar). We first stimulated CTLL-20 cells, which were pretreated with either Me2SO or MG132, with IL-2 to accumulate a pool of active STAT5 (lanes 2 and 9). Genistein, staurosporine, or carrier Me2SO were then added to block further upstream signaling that would activate STAT5. As expected, in the absence of MG132, treatment with both protein kinase inhibitors results in rapid dephosphorylation of STAT5 with a rapid decrease in STAT5 DNA binding activity (lanes 5–8). If active STAT5 proteins are directly destroyed by proteasomes, MG132 should still be able to stabilize these pre-activated proteins, even without continuous generation of active STAT5. Consistent with our previous conclusion that proteasomes do not directly degrade tyrosine-phosphorylated, active STAT5, both protein kinase inhibitors almost completely abolished STAT5 phosphorylation and activation in the presence of MG132 (lanes 12–15). Moreover, this result indicates that protein kinases are important in maintaining IL-2-induced prolonged STAT5 activation by MG132. Jak1 and Jak3 are the two Janus family nonreceptor tyrosine kinases that become tyrosine-phosphorylated and activated by IL-2 stimulation, which in turn activate the STAT proteins (34Johnston J.A. Kawamura M. Kirken R.A. Chen Y.Q. Blake T.B. Shibuya K. Ortaldo J.R. McVicar D.W. O'Shea J.J. Nature. 1994; 370: 151-153Crossref PubMed Scopus (510) Google Scholar, 35Johnston J.A. Bacon C.M. Finbloom D.S. Rees R.C. Kaplan D. Shibuya K. Ortaldo J.R. Gupta S. Chen Y.Q. Giri J.D. O'Shea J.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 8705-8709Crossref PubMed Scopus (342) Google Scholar). To determine if sustained Jak activation contributes to prolonged STAT5 activation by IL-2 in the presence of MG132, anti-Jak1 and anti-Jak3 immunoprecipitates were prepared from cell lysates and analyzed by immunoblotting with antibody to phosphotyrosine. As shown in Fig.4, IL-2-induced tyrosine phosphorylation of both Jak1 (panel A) and Jak3 (panel B) was prolonged in the presence of MG132 (compare lanes 3 and 6) and correlated well with sustained STAT5 activity. Without IL-2 stimulation, however, MG132 itself does not enhance tyrosine phosphorylation of either Jak1 or Jak3 above background (lane 7). Furthermore, under conditions that preserve protein ubiquitination (24Kim T.K. Maniatis T. Science. 1996; 273: 1717-1719Crossref PubMed Scopus (361) Google Scholar,27Chen Z. Hagler J. Palombella V.J. Melandri F. Scherer D. Ballard D. Maniatis T. Genes Dev. 1995; 9: 1586-1597Crossref PubMed Scopus (1172) Google Scholar), we were unable to detect any ubiquitinated Jak1 or Jak3 in the presence of MG132 (lanes 5 and 6). This observation suggests that inactivation of Jak kinases following IL-2-induced activation is mostly due to tyrosine dephosphorylation instead of removal of active Jak proteins through the ubiquitin-proteasome pathway. Consistent with this explanation, there was no detectable loss of either Jak proteins after IL-2 stimulation in the absence of MG132 (lanes 2 and 3, lower panels). These findings suggest that MG132 may sustain IL-2-induced STAT5 activation by maintaining tyrosine phosphorylation of Jak1 and Jak3. Recent studies have shown that proteasome-mediated proteolysis down-regulates STAT1 activity subsequent to IFN-γ-induced activation either by degrading active STAT1 (24Kim T.K. Maniatis T. Science. 1996; 273: 1717-1719Crossref PubMed Scopus (361) Google Scholar) or by blocking upstream signaling (18Haspel R.L. Salditt-Georgieff M. Darnell Jr., J.E. EMBO J. 1996; 15: 6262-6268Crossref PubMed Scopus (272) Google Scholar). Our results that the proteasome inhibitor MG132 stabilizes IL-2-induced STAT5 activation further support the notion that the ubiquitin-proteasome pathway does play an important role in regulating STAT activity in response to multiple cytokines. Even though active STAT1 may be removed through this pathway in some cases (24Kim T.K. Maniatis T. Science. 1996; 273: 1717-1719Crossref PubMed Scopus (361) Google Scholar), we provide further evidence indicating that this mechanism is not generalizable to all STAT proteins in response to different cytokines. First, Haspel et al. (18Haspel R.L. Salditt-Georgieff M. Darnell Jr., J.E. EMBO J. 1996; 15: 6262-6268Crossref PubMed Scopus (272) Google Scholar) demonstrated that STAT1 is quantitatively preserved through the activation-inactivation cycle following IFN-γ stimulation. Second, as demonstrated by our findings, proteasomes modulate IL-2-initiated signaling pathways upstream of STAT5, instead of directly destroying active STAT5 proteins. Haspel et al. (18Haspel R.L. Salditt-Georgieff M. Darnell Jr., J.E. EMBO J. 1996; 15: 6262-6268Crossref PubMed Scopus (272) Google Scholar) also reported that proteasome inhibitors prolonged signaling from the IFN-γ receptors after ligand stimulation by showing sustained tyrosine phosphorylation of the receptors in the presence of MG132. They proposed that proteasome inhibitors may prevent the internalization and subsequent proteolysis of the ligand or receptor or both. The alternative, but not mutually exclusive explanation, is that the proteasome inhibitor MG132 may stabilize active Jak kinases, which are the kinases phosphorylating both the cytokine receptors and the downstream STAT proteins. Consistent with this possibility, we clearly demonstrate here that IL-2-induced tyrosine phosphorylation of Jak1 and Jak3 is maintained in the presence of MG132. It has become increasingly evident that protein-tyrosine phosphatases play important roles in modulating Jak-STAT signaling pathways in response to different cytokines (13David M. Chen H.E. Goelz S. Larner A.C. Neel B.G. Mol. Cell. Biol. 1995; 15: 7050-7058Crossref PubMed Scopus (318) Google Scholar, 14Klingmuller U. Lorenz U. Cantley L.C. Neel B.G. Lodish H.F. Cell. 1995; 80: 729-738Abstract Full Text PDF PubMed Scopus (842) Google Scholar, 15Yi T. Mui A.L.-F. Krystal G. Ihle J.N. Mol. Cell. Biol. 1993; 13: 7577-7586Crossref PubMed Google Scholar). Even though we cannot completely rule out the possibility that small amounts of tyrosine-phosphorylated, active Jak kinases are subjected to proteasome-mediated protein degradation, our results are more consistent with the involvement of protein-tyrosine phosphatases in modulating Jak activity. Therefore, we propose that the ubiquitin-proteasome pathway down-regulates a phosphatase inhibitor(s), which normally inhibits protein-tyrosine phosphatase activity directed toward the Jak kinases. The proteasome inhibitor MG132 stabilizes the phosphatase inhibitor(s), prevents dephosphorylation of the Jak kinases, and results in prolonged signaling by activating downstream STAT proteins. We thank Dr. H. L. Ploegh for generously providing ZL3-H (MG132) to initiate our experiments. We thank J. C. Pratt for her critical review of the manuscript and Y.-J. Jin, J.-H. Chang, and S. Sawasdikosol for their stimulating discussions and helpful comments.