Portal de Periódicos da CAPES

Oxygen and Iron Regulation of Iron Regulatory Protein 2

2003; Elsevier BV; Volume: 278; Issue: 41 Linguagem: Inglês

10.1074/jbc.m302798200

1083-351X

Eric S. Hanson, Mindy L. Rawlins, Elizabeth A. Leibold,

Mitochondrial Function and Pathology

Iron regulatory protein 2 (IRP2) is a central regulator of cellular iron homeostasis due to its regulation of specific mRNAs encoding proteins of iron uptake and storage. Iron regulates IRP2 by mediating its rapid proteasomal degradation, where hypoxia and the hypoxia mimetics CoCl2 and desferrioxamine (DFO) stabilize it. Previous studies showed that iron-mediated degradation of IRP2 requires the presence of critical cysteines that reside within a 73-amino acid unique region. Here we show that a mutant IRP2 protein lacking this 73-amino acid region degraded at a rate similar to that of wild-type IRP2. In addition, DFO and hypoxia blocked the degradation of both the wild-type and mutant IRP2 proteins. Recently, members of the 2-oxoglutarate (2-OG)-dependent dioxygenase family have been shown to hydroxylate hypoxia-inducible factor-1α (HIF-1α), a modification required for its ubiquitination and proteasomal degradation. Since 2-OG-dependent dioxygenases require iron and oxygen, in addition to 2-OG, for substrate hydroxylation, we hypothesized that this activity may be involved in the regulation of IRP2 stability. To test this we used the 2-OG-dependent dioxygenase inhibitor dimethyloxalylglycine (DMOG) and showed that it blocked iron-mediated IRP2 degradation. In addition, hypoxia, DFO and DMOG blocked IRP2 ubiquitination. These data indicate that the region of IRP2 that is involved in IRP2 iron-mediated degradation lies outside of the 73-amino acid unique region and suggest a model whereby 2-OG-dependent dioxygenase activity may be involved in the oxygen and iron regulation of IRP2 protein stability. Iron regulatory protein 2 (IRP2) is a central regulator of cellular iron homeostasis due to its regulation of specific mRNAs encoding proteins of iron uptake and storage. Iron regulates IRP2 by mediating its rapid proteasomal degradation, where hypoxia and the hypoxia mimetics CoCl2 and desferrioxamine (DFO) stabilize it. Previous studies showed that iron-mediated degradation of IRP2 requires the presence of critical cysteines that reside within a 73-amino acid unique region. Here we show that a mutant IRP2 protein lacking this 73-amino acid region degraded at a rate similar to that of wild-type IRP2. In addition, DFO and hypoxia blocked the degradation of both the wild-type and mutant IRP2 proteins. Recently, members of the 2-oxoglutarate (2-OG)-dependent dioxygenase family have been shown to hydroxylate hypoxia-inducible factor-1α (HIF-1α), a modification required for its ubiquitination and proteasomal degradation. Since 2-OG-dependent dioxygenases require iron and oxygen, in addition to 2-OG, for substrate hydroxylation, we hypothesized that this activity may be involved in the regulation of IRP2 stability. To test this we used the 2-OG-dependent dioxygenase inhibitor dimethyloxalylglycine (DMOG) and showed that it blocked iron-mediated IRP2 degradation. In addition, hypoxia, DFO and DMOG blocked IRP2 ubiquitination. These data indicate that the region of IRP2 that is involved in IRP2 iron-mediated degradation lies outside of the 73-amino acid unique region and suggest a model whereby 2-OG-dependent dioxygenase activity may be involved in the oxygen and iron regulation of IRP2 protein stability. Iron regulatory proteins 1 and 2 (IRP1 and IRP2) 1The abbreviations used are: IRP1 and IRP2, iron regulatory proteins 1 and 2; HIF-1α, hypoxia-inducible factor-1α; PA, ponasterone A; 2-OG, 2-oxoglutarate; DMOG, dimethyloxalylglycine; DFO, desferrioxamine; EGFP, enhanced green fluorescent protein; HA, hemagglutinin; mAb, monoclonal antibody; PHD, prolyl hydroxylase. are cytosolic RNA-binding proteins that have a central role in iron homeostasis (reviewed in Refs. 1Eisenstein R.S. Annu. Rev. Nutr. 2000; 20: 627-662Crossref PubMed Scopus (571) Google Scholar and 2Hentze M.W. Kühn L.C. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 8175-8182Crossref PubMed Scopus (1132) Google Scholar). IRP1 and IRP2 bind to stem-loop structures, known as iron-responsive elements, that are found in the untranslated regions of mRNAs encoding proteins important in iron homeostasis. The interaction of IRPs with iron-responsive elements represses translation of ferritin mRNA while increasing the stability of the transferrin receptor mRNA. Since IRP1 and IRP2 RNA binding activities are regulated by intracellular iron concentration, these proteins couple cellular iron with the regulation of expression of proteins important in iron homeostasis. In iron-replete cells, IRP1 RNA binding activity decreases as the protein assembles a [4Fe-4S] cluster having cytosolic aconitase activity, whereas in iron-depleted cells apo-IRP1 binds RNA (3Kaptain S. Downey W.E. Tang C. Philpott C. Haile D. Orloff D.G. Harford J.B. Rouault T.A. Klausner R.D. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 10109-10113Crossref PubMed Scopus (160) Google Scholar). In the case of IRP2, iron controls its RNA binding activity through the regulation of protein stability. This is achieved through iron stimulation of its ubiquitination and proteasomal degradation (4Guo B. Phillips J.D. Yu Y. Leibold E.A. J. Biol. Chem. 1995; 270: 21645-21651Abstract Full Text Full Text PDF PubMed Scopus (270) Google Scholar, 5Samaniego F. Chin J. Iwai K. Rouault T.A. Klausner R.D. J. Biol. Chem. 1994; 269: 30904-30910Abstract Full Text PDF PubMed Google Scholar, 6Iwai K. Drake S.K. Wehr N.B. Weissman A.M. LaVaute T. Minato N. Klausner R.D. Levine R.L. Rouault T.A. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 4924-4928Crossref PubMed Scopus (265) Google Scholar, 7Iwai K. Klausner R.D. Rouault T.A. EMBO J. 1995; 14: 5350-5357Crossref PubMed Scopus (193) Google Scholar). Thus, the amount of IRP2 is inversely related to cellular iron concentration. The mechanism by which iron accelerates IRP2 degradation has been reported to involve iron-catalyzed oxidation of the protein within a unique 73-amino acid region termed the “degradation domain” (6Iwai K. Drake S.K. Wehr N.B. Weissman A.M. LaVaute T. Minato N. Klausner R.D. Levine R.L. Rouault T.A. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 4924-4928Crossref PubMed Scopus (265) Google Scholar, 7Iwai K. Klausner R.D. Rouault T.A. EMBO J. 1995; 14: 5350-5357Crossref PubMed Scopus (193) Google Scholar). It has been recently shown that in vitro iron can oxidize cysteines in this domain (8Kang D.K. Jeong J. Drake S.K. Wehr N. Rouault T.A. Levine R.L. J. Biol. Chem. 2003; 278: 14857-14864Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar). Oxidatively modified IRP2 could then potentially interact with an E3 ubiquitin ligase, such as the recently identified HOIL-1 E3 ligase that has been implicated in IRP2 ubiquitination (9Yamanaka K. Ishikawa H. Megumi Y. Tokunaga F. Kanie M. Rouault T.A. Morishima I. Minato N. Ishimori K. Iwai K. Nat. Cell Biol. 2003; 5: 336-340Crossref PubMed Scopus (152) Google Scholar). Like IRP2, hypoxia-inducible factor-1α (HIF-1α) is regulated at the level of protein stability. HIF-1α is a heterodimeric transcription factor that is the principal mediator of transcriptional activation during hypoxia (reviewed in Refs. 10Semenza G.L. Annu. Rev. Cell Dev. Biol. 1999; 15: 551-578Crossref PubMed Scopus (1676) Google Scholar, 11Kaelin Jr., W.G. Genes Dev. 2002; 16: 1441-1445Crossref PubMed Scopus (147) Google Scholar, 12Semenza G.L. Cell. 2001; 107: 1-3Abstract Full Text Full Text PDF PubMed Scopus (790) Google Scholar). During normoxia, the HIF-1α subunit is hydroxylated on two prolines by three recently identified prolyl hydroxylases (PHD1, PHD2, and PHD3) (13Epstein A.C. Gleadle J.M. McNeill L.A. Hewitson K.S. O'Rourke J. Mole D.R. Mukherji M. Metzen E. Wilson M.I. Dhanda A. Tian Y.M. Masson N. Hamilton D.L. Jaakkola P. Barstead R. Hodgkin J. Maxwell P.H. Pugh C.W. Schofield C.J. Ratcliffe P.J. Cell. 2001; 107: 43-54Abstract Full Text Full Text PDF PubMed Scopus (2736) Google Scholar, 14Bruick R.K. McKnight S.L. Science. 2001; 294: 1337-1340Crossref PubMed Scopus (2114) Google Scholar, 15Ivan M. Kondo K. Yang H. Kim W. Valiando J. Ohh M. Salic A. Asara J.M. Lane W.S. Kaelin Jr., W.G. Science. 2001; 292: 464-468Crossref PubMed Scopus (3893) Google Scholar, 16Jaakkola P. Mole D.R. Tian Y.M. Wilson M.I. Gielbert J. Gaskell S.J. Kriegsheim A. Hebestreit H.F. Mukherji M. Schofield C.J. Maxwell P.H. Pugh C.W. Ratcliffe P.J. Science. 2001; 292: 468-472Crossref PubMed Scopus (4456) Google Scholar, 17Yu F. White S.B. Zhao Q. Lee F.S. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 9630-9635Crossref PubMed Scopus (647) Google Scholar). Proline hydroxylation is required for the interaction of HIF-1α with the von Hippel-Lindau protein E3 ubiquitin ligase, leading to its ubiquitination and proteasomal degradation (13Epstein A.C. Gleadle J.M. McNeill L.A. Hewitson K.S. O'Rourke J. Mole D.R. Mukherji M. Metzen E. Wilson M.I. Dhanda A. Tian Y.M. Masson N. Hamilton D.L. Jaakkola P. Barstead R. Hodgkin J. Maxwell P.H. Pugh C.W. Schofield C.J. Ratcliffe P.J. Cell. 2001; 107: 43-54Abstract Full Text Full Text PDF PubMed Scopus (2736) Google Scholar, 14Bruick R.K. McKnight S.L. Science. 2001; 294: 1337-1340Crossref PubMed Scopus (2114) Google Scholar, 15Ivan M. Kondo K. Yang H. Kim W. Valiando J. Ohh M. Salic A. Asara J.M. Lane W.S. Kaelin Jr., W.G. Science. 2001; 292: 464-468Crossref PubMed Scopus (3893) Google Scholar, 16Jaakkola P. Mole D.R. Tian Y.M. Wilson M.I. Gielbert J. Gaskell S.J. Kriegsheim A. Hebestreit H.F. Mukherji M. Schofield C.J. Maxwell P.H. Pugh C.W. Ratcliffe P.J. Science. 2001; 292: 468-472Crossref PubMed Scopus (4456) Google Scholar, 17Yu F. White S.B. Zhao Q. Lee F.S. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 9630-9635Crossref PubMed Scopus (647) Google Scholar, 18Masson N. Willam C. Maxwell P.H. Pugh C.W. Ratcliffe P.J. EMBO J. 2001; 20: 5197-5206Crossref PubMed Scopus (857) Google Scholar, 19Min J.-H. Yang H. Ivan M. Gertler F. Kaelin Jr., W.G. Pavletich N.P. Science. 2002; 296: 1886-1889Crossref PubMed Scopus (584) Google Scholar, 20Hon W.C. Wilson M.I. Harlos K. Claridge T.D. Schofield C.J. Pugh C.W. Maxwell P.H. Ratcliffe P.J. Stuart D.I. Jones E.Y. Nature. 2002; 417: 975-978Crossref PubMed Scopus (560) Google Scholar). Prolyl hydroxylases are members of the superfamily of 2-oxoglutarate (2-OG)-dependent dioxygenases that are enzymes requiring oxygen and Fe2+, in addition to 2-OG, for substrate hydroxylation. Due to the oxygen requirement of prolyl hydroxylases, HIF-1α hydroxylation is reduced as cellular oxygen decreases. This accounts for the inverse relationship between oxygen concentration and HIF-1α stabilization. The involvement of prolyl hydroxylases in HIF-1α protein degradation provides a mechanism for gene regulation by hypoxia. Our previous studies demonstrated that IRP2 RNA binding activity increases during hypoxia through altered protein degradation (21Hanson E.S. Foot L.M. Leibold E.A. J. Biol. Chem. 1999; 274: 5047-5052Abstract Full Text Full Text PDF PubMed Scopus (133) Google Scholar). In this study we show that hypoxia blocks IRP2 degradation, and that the previously designated degradation domain is dispensable for both oxygen and iron regulation of IRP2 stability. In addition, IRP2 iron-mediated degradation is blocked by the hypoxic mimetic CoCl2 and by the 2-OG-dependent dioxygenase inhibitor DMOG. Since hypoxia, iron chelation, CoCl2, and DMOG can inhibit 2-OG-dependent dioxygenases, these results suggest that IRP2 degradation may require the activity of a member(s) of this class of enzymes. Moreover, iron chelation, hypoxia, and DMOG all decrease in vivo IRP2 ubiquitination, while iron and 2-OG stimulate in vitro IRP2 ubiquitination. Cell Culture—Human 293 embryonic kidney cells stably expressing the ecdysone receptor (EcR293 cell line, Invitrogen) were grown in Dulbecco's modified essential medium containing 4.5 g/liter glucose, 10% fetal calf serum, and 400 μg/ml Zeocin. Human 293T cells were grown in the same medium lacking Zeocin. Human Hep3B hepatoma cells were from American Type Culture Collection and cultured in Earle's minimum essential medium containing 1 mm pyruvate, 2 mml-glutamine, and 0.1 mm non-essential amino acids. Plates were treated with polylysine before seeding EcR293 and 293T cells. To generate an IRP2myc stable cell line EcR293 cells were transfected using LipofectAMINE (Invitrogen) and 2 μg of the pIND(sp1) vector harboring a C-terminal myc-tagged human IRP2 cDNA (21Hanson E.S. Foot L.M. Leibold E.A. J. Biol. Chem. 1999; 274: 5047-5052Abstract Full Text Full Text PDF PubMed Scopus (133) Google Scholar). The ponasterone A (PA) inducible expressing IRP2myc clone (CII) was isolated following selection in 400 μg/ml G418. IRP2myc degradation assays were performed in CII cells by inducing IRP2myc synthesis with 10 μm PA for 20 h. After removal of PA, IRP2myc degradation was monitored by immunoblot analysis using the 9E10 anti-myc mAb (Santa Cruz). Degradation assays in transiently transfected EcR293 cells were performed by transfecting with 12 μg total DNA/100 mm plate. Following transfection, cells were split equally to 60-mm plates and induced with 10 μm PA prior to chase under the indicated conditions. Degradation was monitored by anti-HA immunoblots. All experiments were performed multiple times with similar results. Hypoxia was achieved by placing cells in a humidified Napco incubator maintained at 1% O2 and 5% CO2 with a balance of nitrogen. Normoxia was achieved by exposing cells to ambient air with 5% CO2. All cells were grown at 37 °C. Cells were labeled with 100 μCi/ml [35S]Met/Cys (ICN) for 60 min in methionine/cysteine free media. Clones—The C-terminal myc-tagged human IRP2 cDNA in pIND(sp1) (21Hanson E.S. Foot L.M. Leibold E.A. J. Biol. Chem. 1999; 274: 5047-5052Abstract Full Text Full Text PDF PubMed Scopus (133) Google Scholar) and pcDNA3.1(+) were used as parental vectors to generate pIND(sp1)/IRP2-HA, the 73-amino acid deletion constructs pIND(sp1)/Δ73IRP2-HA and pcDNA3.1(+)/Δ73IRP2myc, and the N-terminal myc-tagged pcDNA3.1(+)/mycIRP2. All PCR manipulations were performed using high fidelity Platinum Taq (Invitrogen), and all constructs were sequenced verified. The 73-amino acid deletion removes IRP2 amino acids 137 to 209 (ΔCAI-X67-PVP). The location of the myc and HA tags was found to have no effect on IRP2 degradation or ubiquitination. Immunoblot Analysis—Protein extracts were prepared by lysing cells in lysis buffer (20 mm Hepes, pH 8.0, 25 mm KCl, 0.5% Nonidet P-40). Protein (25 μg) was resolved by 8% SDS-PAGE, transferred to nitrocellulose membranes, and probed with rabbit anti-IRP2 antibody (22Guo B. Yu Y. Leibold E.A. J. Biol. Chem. 1994; 269: 24252-24260Abstract Full Text PDF PubMed Google Scholar), anti-myc mAb, anti-HIF-1α mAb (BD Transduction Labs), anti-ubiquitin (a gift from M. Rechsteiner, University of Utah), anti-HA mAb (University of Utah Protein Core Facility), or anti-EGFP mAb (Sigma). Visualization was achieved using Renaissance Western blot Chemiluminescence Reagent (PerkinElmer Life Sciences). Ubiquitination Assays—For in vivo ubiquitination assays of wild-type IRP2, 293T cells grown on 100 mm plates were cotransfected with 6 μg each of pcDNA3.1(+)/mycIRP2 and a vector that expresses HA-tagged ubiquitin (23Treier M. Staszewski L.M. Bohmann D. Cell. 1994; 78: 787-798Abstract Full Text PDF PubMed Scopus (846) Google Scholar). Following transfection, cells were split equally to 60-mm plates with one plate receiving 100 μm DFO where indicated. After 22 h, 5 mm DMOG was added where indicated. One h after addition of DMOG all plates received 10 μm lactacystin to inhibit the proteasome, and one plate was exposed to hypoxia. Cells were harvested 5.5 h after addition of lactacystin in lysis buffer containing 2.5 mm N-ethylmaleimide. Immunoprecipitations were performed in phosphate-buffered saline using 6 μg of anti-myc antibody and 300 μg of lysate. Immunoprecipitates were resolved on 8% SDS-PAGE and probed with anti-HA antibody to monitor IRP2myc ubiquitination. The blots were stripped and reprobed with anti-IRP2 to ensure equal precipitation from each sample. The same transfection protocol was used for monitoring ubiquitination of Δ73IRP2myc and its wild-type control, WtIRP2myc. For in vitro ubiquitination assays, extracts from 18-h DFO-exposed cells were lysed in 10 mm Tris, pH 7.8, 0.25% Triton X-100 and used in assays containing 0.5 μg/μl ubiquitin, 3.2 μm ubiquitin aldehyde, 1 mm ascorbic acid and an ATP regeneration system (1 mm ATP, 60 μg/ml creatine phosphokinase, 6.6 mm phosphocreatin) with or without 400 μm ferrous ammonium sulfate and 1 mm 2-OG. Assays for were performed in 25 μl at ∼4 μg/ml protein for 18 h at 30 °C. DMOG and Other Reagents—DMOG was synthesized by Frontier Scientific (Logan, UT) and verified by GC mass spec and nuclear magnetic resonance analysis. DMOG and lactacystin (Sigma) were dissolved in H2O and MG132 (Calbiochem) was dissolved in dimethyl sulfoxide. Ubiquitin aldehyde was from Boston Biochem Inc. Hypoxia Stabilizes IRP2myc in the CII Cell Line—To study IRP2 stability during hypoxia, we generated a stable ecdysone receptor 293 cell line (termed CII) that expresses IRP2myc only in the presence of the inducer PA. When PA is removed from the growth medium, IRP2myc synthesis ceases, and its degradation can be monitored by anti-myc immunoblot analysis. The CII cell line was used for time course experiments to determine the stability of IRP2myc under normoxic and hypoxic conditions. Fig. 1A reveals that IRP2myc decays with a t ½ of ∼5 h during normoxia that is increased to greater than 12 h during hypoxia. Stabilization of IRP2myc is also seen at 20 h of hypoxia (data not shown). The degradation of IRP2myc is completely blocked by the proteasome inhibitors lactacystin and MG132, demonstrating that its degradation recapitulates endogenous IRP2 proteasomal degradation (Fig. 1B) (4Guo B. Phillips J.D. Yu Y. Leibold E.A. J. Biol. Chem. 1995; 270: 21645-21651Abstract Full Text Full Text PDF PubMed Scopus (270) Google Scholar, 6Iwai K. Drake S.K. Wehr N.B. Weissman A.M. LaVaute T. Minato N. Klausner R.D. Levine R.L. Rouault T.A. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 4924-4928Crossref PubMed Scopus (265) Google Scholar, 7Iwai K. Klausner R.D. Rouault T.A. EMBO J. 1995; 14: 5350-5357Crossref PubMed Scopus (193) Google Scholar). To determine whether hypoxia irreversibly stabilizes IRP2myc, CII cells were exposed to hypoxia for the last 2 h of a 20-h PA induction and subsequently returned to normoxia for the chase phase. Fig. 1C shows that IRP2myc degrades at the same rate during a normoxic chase irrespective of previous exposure to normoxia or hypoxia. A similar result was found when CII cells were exposed to 16 h of hypoxia prior to the normoxic chase (data not shown). These data indicate that the signal responsible for IRP2 hypoxic stabilization is reversed upon the reintroduction of oxygen. Removal of the 73-Amino Acid Unique Region in IRP2 Does Not Affect IRP2 Degradation in EcR293 Cells—The 73-amino acid unique region in IRP2 has been reported to be required for iron-mediated ubiquitination and degradation (6Iwai K. Drake S.K. Wehr N.B. Weissman A.M. LaVaute T. Minato N. Klausner R.D. Levine R.L. Rouault T.A. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 4924-4928Crossref PubMed Scopus (265) Google Scholar, 7Iwai K. Klausner R.D. Rouault T.A. EMBO J. 1995; 14: 5350-5357Crossref PubMed Scopus (193) Google Scholar). The current model suggests that three critical cysteines in the 73-amino acid unique region directly sense iron resulting in localized oxidation and subsequent ubiquitination. We performed degradation assays to address the role of this 73-amino acid region in the regulation of IRP2 degradation. Degradation of full-length HA-tagged wild-type IRP2 (WtIRP2-HA) and its 73-amino acid deletion counterpart (Δ73IRP2-HA) were monitored in transiently transfected EcR293 cells in the presence and absence of iron. Unexpectedly, the Δ73IRP2-HA protein degraded at a rate similar to that of the wild-type protein (Fig. 2A). Moreover, iron accelerated the degradation of both proteins. It should be noted that the degradation rate of transiently transfected IRP2 is longer than that observed for stably expressed IRP2myc (Figs. 1A and 2A) due to greater expression achieved in transient transfections. To determine whether the degradation of Δ73IRP2-HA is iron-dependent, degradation was monitored in the presence of DFO. EcR293 cells were separately transfected with vectors expressing WtIRP2-HA, Δ73IRP2-HA, or EGFP. Fig. 2B shows that DFO blocks the degradation of both WtIRP2-HA and Δ73IRP2-HA, whereas the stability of EGFP was not affected. In addition, exposure of cells to hypoxia also resulted in the stabilization of Δ73IRP2-HA (Fig. 2C). These data indicate that the region of IRP2 that is involved in IRP2 iron- and oxygen-mediated degradation lies outside of the 73-amino acid unique region. Inhibitors of 2-OG-dependent Dioxygenase Activity Block IRP2 Degradation—Our results indicate that cysteine oxidation in the 73-amino acid unique region is not required for IRP2 regulation in 293 cells. Since IRP2 and HIF-1α are both stabilized by hypoxia and DFO, we considered an alternative model for IRP2 regulation that requires hydroxylation. To test this model, we used inhibitors of 2-OG-dependent dioxygenases. Cobaltous ion is a well known inducer of the hypoxic response due to its stabilizing effect on HIF-1α. Recent studies have attributed this to a decrease in HIF-1α hydroxylation by the direct inhibition of prolyl hydroxylase activity by cobalt through competition with iron (13Epstein A.C. Gleadle J.M. McNeill L.A. Hewitson K.S. O'Rourke J. Mole D.R. Mukherji M. Metzen E. Wilson M.I. Dhanda A. Tian Y.M. Masson N. Hamilton D.L. Jaakkola P. Barstead R. Hodgkin J. Maxwell P.H. Pugh C.W. Schofield C.J. Ratcliffe P.J. Cell. 2001; 107: 43-54Abstract Full Text Full Text PDF PubMed Scopus (2736) Google Scholar). We previously showed that CoCl2-treated cells resulted in IRP2 accumulation, however, the mechanism was not examined (21Hanson E.S. Foot L.M. Leibold E.A. J. Biol. Chem. 1999; 274: 5047-5052Abstract Full Text Full Text PDF PubMed Scopus (133) Google Scholar). To further investigate the effect of CoCl2 on IRP2, we measured IRP2myc degradation in CoCl2 treated cells. Fig. 3A illustrates that CoCl2, like hypoxia, completely inhibits IRP2myc degradation. Furthermore, the stabilization effect of CoCl2 on IRP2myc was prevented by iron (Fig. 3B). Taken together these results suggest the possibility that IRP2 degradation may be mediated by 2-OG-dependent dioxygenase activity. To further examine the potential role for 2-OG-dependent dioxygenase activity in IRP2 degradation we used the inhibitor DMOG. Intracellularly DMOG is converted to N-oxalylglycine (a structural analog of 2-OG), which is a competitive inhibitor with respect to 2-OG in the inhibition of 2-OG-dependent dioxygenases, where inhibition of collagen prolyl-4-hydroxylase occurs in the presence of excess iron (24Cunliffe C.J. Franklin T.J. Hales N.J. Hill G.B. J. Med. Chem. 1992; 35: 2652-2658Crossref PubMed Scopus (87) Google Scholar, 25Baader E. Tschank G. Baringhaus K.H. Burghard H. Günzler V. Biochem. J. 1994; 300: 525-530Crossref PubMed Scopus (61) Google Scholar). DMOG has been demonstrated to inhibit prolyl hydroxylation of HIF-1α, resulting in its normoxic accumulation in Hep3B cells and in Caenorhabditis elegans (13Epstein A.C. Gleadle J.M. McNeill L.A. Hewitson K.S. O'Rourke J. Mole D.R. Mukherji M. Metzen E. Wilson M.I. Dhanda A. Tian Y.M. Masson N. Hamilton D.L. Jaakkola P. Barstead R. Hodgkin J. Maxwell P.H. Pugh C.W. Schofield C.J. Ratcliffe P.J. Cell. 2001; 107: 43-54Abstract Full Text Full Text PDF PubMed Scopus (2736) Google Scholar, 16Jaakkola P. Mole D.R. Tian Y.M. Wilson M.I. Gielbert J. Gaskell S.J. Kriegsheim A. Hebestreit H.F. Mukherji M. Schofield C.J. Maxwell P.H. Pugh C.W. Ratcliffe P.J. Science. 2001; 292: 468-472Crossref PubMed Scopus (4456) Google Scholar, 26Chan D.A. Sutphin P.D. Denko N.C. Giaccia A.J. J. Biol. Chem. 2002; 277: 40112-40117Abstract Full Text Full Text PDF PubMed Scopus (228) Google Scholar, 27Ivan M. Haberberger T. Gervasi D.C. Michelson K.S. Günzler V. Kondo K. Yang H. Sorokina I. Conaway R.C. Conaway J.W. Kaelin Jr., W.G. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 13459-13464Crossref PubMed Scopus (491) Google Scholar). To examine the effect of DMOG on IRP2 stability, 293T cells were first treated with DFO for 16 h to maximally increase endogenous IRP2. This was necessary because under normal 293T cell growth conditions IRP2 is present at low amounts. Cells were then exposed to DMOG for 2 h prior to the removal of DFO, after which DMOG was added back to the medium with or without ferrous sulfate to promote IRP2 degradation. Cells were harvested at 1, 2, 3, and 4 h after the addition of iron. Fig. 4A illustrates significant IRP2 degradation after 4 h in the absence of DMOG, whereas IRP2 degradation was prevented in DMOG-treated cells. The blot was then probed for HIF-1α. As expected, HIF-1α was present in DFO-treated cells that were immediately harvested, whereas it degraded in the absence of DMOG, consistent with its rapid normoxic degradation (Fig. 4A). Like IRP2, the degradation of HIF-1α was blocked in the presence of DMOG. In addition, the inhibition of IRP2 and HIF-1α degradation by DMOG occurs at similar concentrations (Fig. 4B). DMOG stabilization of IRP2 is not unique to 293T cells, since IRP2 in Hep3B cells is similarly affected by DMOG (Fig. 4C). To determine whether the degradation of Δ73IRP2-HA is 2-OG-dependent, degradation was measured in the presence of DMOG using transient transfection assays. Fig. 4D shows that DMOG blocks the degradation of both WtIRP2-HA and Δ73IRP2-HA. This result is consistent with data demonstrating that the 73-amino acid unique region is dispensable for iron and oxygen regulation of IRP2 (Fig. 2), and suggests that WtIRP2-HA and Δ73IRP2-HA degradation proceed by a pathway requiring 2-OG-dependent dioxygenase activity. Control experiments were performed to demonstrate that DMOG does not affect overall protein ubiquitination or bulk protein degradation (Fig. 4, E and F). IRP2 Ubiquitination Is Blocked by DMOG, DFO, and Hypoxia in Vivo—Since IRP2 degradation requires ubiquitination (6Iwai K. Drake S.K. Wehr N.B. Weissman A.M. LaVaute T. Minato N. Klausner R.D. Levine R.L. Rouault T.A. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 4924-4928Crossref PubMed Scopus (265) Google Scholar, 28Cockman M.E. Masson N. Mole D.R. Jaakkola P. Chang G.W. Clifford S.C. Maher E.R. Pugh C.W. Ratcliffe P.J. Maxwell P.H. J. Biol. Chem. 2000; 275: 25733-25741Abstract Full Text Full Text PDF PubMed Scopus (922) Google Scholar), the effect of DMOG, DFO, and hypoxia on in vivo IRP2 ubiquitination was examined. 293T cells were cotransfected with vectors that express mycIRP2 and HA-ubiquitin. Following transfection, cells were split into four plates to allow for the accumulation of mycIRP2 and HA-ubiquitin, with one plate exposed to DFO. After 22 h, a second plate was pretreated to DMOG for 1 h prior to addition of 10 μm lactacystin to inhibit the proteasome. Untreated cells and DFO-exposed cells also received lactacystin. A third plate was exposed to hypoxia immediately following addition of lactacystin. After 5.5 h cell lysates were prepared, and untransfected and transfected extracts were sequentially probed with anti-HA and anti-IRP2 to determine expression (Fig. 5A, left panels). Immunoprecipitations were performed to capture mycIRP2 from each lysate. Immunoprecipitated mycIRP2 was then assessed for HA-ubiquitin conjugation in an anti-HA immunoblot (Fig. 5A, right panel). The blot was reprobed for IRP2 to show equal precipitations. The results demonstrate that mycIRP2 is ubiquitinated in 293T cells, and that DMOG, hypoxia, and DFO decrease ubiquitination. To examine whether 2-OG-dependent dioxygenase activity is required for Δ73IRP2 ubiquitination, the effect of DMOG on its ubiquitination was monitored in 293T cells. 293T cells were cotransfected with WtIRP2myc or Δ73IRP2myc along with the vector for HA-ubiquitin. IRP2myc and Δ73IRP2myc were immunoprecipitated and assessed for ubiquitination by an anti-HA immunoblot. Fig. 5B shows that Δ73IRP2myc is ubiquitinated in the absence of DMOG. In the presence of DMOG, ubiquitination of Δ73IRP2myc, like that of WtIRP2myc, is greatly reduced. These data suggest that the region of IRP2 that may be involved in 2-OG-dependent dioxygenase-mediated ubiquitination is outside of the 73-amino acid unique region. Iron and 2-OG Stimulate IRP2 Ubiquitination in Vitro—We determined the effect of iron and 2-OG on IRP2 ubiquitination in vitro. Extracts prepared from 293T cells exposed to DFO for 18 h were incubated with an ATP regenerating system containing ubiquitin, ubiquitin aldehyde, and ascorbic acid in the presence or absence of iron and 2-OG. IRP2 ubiquitination was measured by immunoblot analysis using anti-IRP2 antibody. The results illustrate an increase in high molecular weight IRP2 when iron is present that is further increased in the presence 2-OG (Fig. 5C). These data show that in the presence of cofactors required for 2-OG-dependent dioxygenase activity, IRP2 ubiquitination is stimulated in vitro. These results are consistent with the in vivo ubiquitination experiments, suggesting that 2-OG-dependent dioxygenase activity is required for IRP2 ubiquitination. Previously Iwai et al. proposed a model for iron-dependent IRP2 degradation where IRP2 is directly oxidized through an iron-catalyzed mechanism (6Iwai K. Drake S.K. Wehr N.B. Weissman A.M. LaVaute T. Minato N. Klausner R.D. Levine R.L. Rouault T.A. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 4924-4928Crossref PubMed Scopus (265) Google Scholar). In this model, the unique 73-amino acid degradation domain is inferred to directly sense iron through three cysteines resulting in localized oxidation. Oxidized IRP2 is then efficiently ubiquitinated. Recently, Kang et al. (8Kang D.K. Jeong J. Drake S.K. Wehr N. Rouault T.A. Levine R.L. J. Biol. Chem. 2003; 278: 14857-14864Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar) used an in vitro metal-catalyzed oxidation system and showed that iron induces the oxidation of any one of these three cysteines. Whether these specific oxidations are involved in the ubiquitination of IRP2 remains to be determined. Since the IRP2 73-amino acid domain has been implicated in iron-mediated degradation, we hypothesized that this region could be involved in its regulation by hypoxia. Our analysis of IRP2 lacking this region indicates that this 73-amino acid unique region is dispensable for oxygen- and iron-mediated degradation. This is supported by the finding that DFO, DMOG, and hypoxia block the degradation of Δ73IRP2-HA, whereas iron stimulates its degradation. Moreover, DMOG blocked Δ73IRP2-HA ubiquitination in vivo. Although it is not yet clear why our results differ from those of Iwai et al. (6Iwai K. Drake S.K. Wehr N.B. Weissman A.M. LaVaute T. Minato N. Klausner R.D. Levine R.L. Rouault T.A. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 4924-4928Crossref PubMed Scopus (265) Google Scholar, 7Iwai K. Klausner R.D. Rouault T.A. EMBO J. 1995; 14: 5350-5357Crossref PubMed Scopus (193) Google Scholar), it is possible that different cells employ distinct, as well as an overlapping mechanisms, in the regulation of IRP2 stability. The degradation of HIF-1α has been found to be critically dependent on the activity of 2-OG-dependent dioxygenases (13Epstein A.C. Gleadle J.M. McNeill L.A. Hewitson K.S. O'Rourke J. Mole D.R. Mukherji M. Metzen E. Wilson M.I. Dhanda A. Tian Y.M. Masson N. Hamilton D.L. Jaakkola P. Barstead R. Hodgkin J. Maxwell P.H. Pugh C.W. Schofield C.J. Ratcliffe P.J. Cell. 2001; 107: 43-54Abstract Full Text Full Text PDF PubMed Scopus (2736) Google Scholar, 14Bruick R.K. McKnight S.L. Science. 2001; 294: 1337-1340Crossref PubMed Scopus (2114) Google Scholar, 15Ivan M. Kondo K. Yang H. Kim W. Valiando J. Ohh M. Salic A. Asara J.M. Lane W.S. Kaelin Jr., W.G. Science. 2001; 292: 464-468Crossref PubMed Scopus (3893) Google Scholar, 16Jaakkola P. Mole D.R. Tian Y.M. Wilson M.I. Gielbert J. Gaskell S.J. Kriegsheim A. Hebestreit H.F. Mukherji M. Schofield C.J. Maxwell P.H. Pugh C.W. Ratcliffe P.J. Science. 2001; 292: 468-472Crossref PubMed Scopus (4456) Google Scholar, 17Yu F. White S.B. Zhao Q. Lee F.S. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 9630-9635Crossref PubMed Scopus (647) Google Scholar). The oxidative decarboxylation reaction catalyzed by 2-OG-dependent dioxygenases requires iron and oxygen in addition to 2-OG (reviewed in Ref. 29Kivirikko K.I. Pihlajaniemi T. Adv. Enzymol. Relat. Areas Mol. Biol. 1998; 72: 325-398PubMed Google Scholar). Cobaltous ions inhibit mammalian and C. elegans HIF-1α PHDs in addition to inhibiting the recently identified HIF-1α asparaginyl hydroxylase that catalyzes the hydroxylation of the HIF-1α C-terminal activation domain (13Epstein A.C. Gleadle J.M. McNeill L.A. Hewitson K.S. O'Rourke J. Mole D.R. Mukherji M. Metzen E. Wilson M.I. Dhanda A. Tian Y.M. Masson N. Hamilton D.L. Jaakkola P. Barstead R. Hodgkin J. Maxwell P.H. Pugh C.W. Schofield C.J. Ratcliffe P.J. Cell. 2001; 107: 43-54Abstract Full Text Full Text PDF PubMed Scopus (2736) Google Scholar, 16Jaakkola P. Mole D.R. Tian Y.M. Wilson M.I. Gielbert J. Gaskell S.J. Kriegsheim A. Hebestreit H.F. Mukherji M. Schofield C.J. Maxwell P.H. Pugh C.W. Ratcliffe P.J. Science. 2001; 292: 468-472Crossref PubMed Scopus (4456) Google Scholar, 30Hewitson K.S. McNeill L.A. Riordan M.V. Tian Y.M. Bullock A.N. Welford R.W. Elkins J.M. Oldham N.J. Bhattacharya S. Gleadle J.M. Ratcliffe P.J. Pugh C.W. Schofield C.J. J. Biol. Chem. 2002; 277: 26351-26355Abstract Full Text Full Text PDF PubMed Scopus (603) Google Scholar, 31Lando D. Peet D.J. Gorman J.J. Whelan D.A. Whitelaw M.L. Bruick R.K. Genes Dev. 2002; 16: 1466-1471Crossref PubMed Scopus (1229) Google Scholar). A probable mechanism for inactivation of these enzymes by cobalt is through direct competition with iron at the catalytic active site (13Epstein A.C. Gleadle J.M. McNeill L.A. Hewitson K.S. O'Rourke J. Mole D.R. Mukherji M. Metzen E. Wilson M.I. Dhanda A. Tian Y.M. Masson N. Hamilton D.L. Jaakkola P. Barstead R. Hodgkin J. Maxwell P.H. Pugh C.W. Schofield C.J. Ratcliffe P.J. Cell. 2001; 107: 43-54Abstract Full Text Full Text PDF PubMed Scopus (2736) Google Scholar). Our previous data showing that IRP2 accumulates in CoCl2-treated cells (21Hanson E.S. Foot L.M. Leibold E.A. J. Biol. Chem. 1999; 274: 5047-5052Abstract Full Text Full Text PDF PubMed Scopus (133) Google Scholar) is explained here by the demonstration that, like hypoxia, CoCl2 blocks IRP2 degradation. Similar to HIF-1α, the mechanism of CoCl2-induced IRP2 stabilization is through competition with iron. In addition, the 2-OG analog DMOG blocked HIF-1α degradation (through inhibition of HIF-1α PHDs) at the same concentration required to prevent IRP2 degradation. The relativity high concentration of DMOG (2.5-5 mm) needed to block IRP2 degradation may be due to inefficient conversion of DMOG to its active N-oxalylglycine form within the cell. It should be noted that IRP1 RNA binding activity was not affected in DMOG-treated cells, indicating that DMOG is not chelating iron (data not shown). In addition to CoCl2, DMOG and hypoxia, treating cells to iron chelators has long been known to stabilize IRP2. Therefore, when cellular 2-OG-dependent dioxygenase activity is predicted to be inhibited (exposure to CoCl2, DMOG, DFO, or hypoxia) there is a simultaneous increase in IRP2 stability. Conversely, when cellular iron is elevated, a condition that might be expected to increase 2-OG-dependent dioxygenase activity, IRP2 is more efficiently degraded. Based on these results, we suggest that 2-OG-dependent dioxygenase activity may be involved in the degradation of IRP2. To determine the step in the IRP2 degradation pathway that is affected by inhibitors of 2-OG-dependnet dioxygenase activity, we examined the effect of DMOG, DFO and hypoxia on in vivo mycIRP2 ubiquitination. Ubiquitination of mycIRP2 in vivo was blocked by DMOG and hypoxia. DFO also blocked in vivo mycIRP2 ubiquitination consistent with the results of others (6Iwai K. Drake S.K. Wehr N.B. Weissman A.M. LaVaute T. Minato N. Klausner R.D. Levine R.L. Rouault T.A. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 4924-4928Crossref PubMed Scopus (265) Google Scholar). Since DMOG, hypoxia, and DFO inhibit 2-OG-dependent dioxygenases, these results suggest that a hydroxylation event lies upstream of IRP2 ubiquitination. Extending our ubiquitination experiments to an in vitro assay, we show that high molecular weight IRP2 species (consistent with ubiquitination) maximally forms when iron and 2-OG are present. To a lesser extent, ubiquitination of IRP2 occurred when iron alone was added to the reaction. This may reflect some stimulation of 2-OG-dependent dioxygenase activity by 2-OG present in the extracts. These results suggest that we have recapitulated in vitro both the hydroxylation and subsequent ubiquitination of IRP2. Although hydroxylation of IRP2 would be a direct way to communicate cellular oxygen and iron to IRP2, our data do not show that IRP2 is directly hydroxylated. It is possible that hydroxylation could regulate the function of a protein in the IRP2 degradation pathway, such as an E3 ubiquitin ligase. Potential candidates for IRP2 hydroxylases are one of the three HIF-1α PHDs and the HIF-1α asparaginyl hydroxylase, all of whose activities are influenced by physiologic levels of oxygen (13Epstein A.C. Gleadle J.M. McNeill L.A. Hewitson K.S. O'Rourke J. Mole D.R. Mukherji M. Metzen E. Wilson M.I. Dhanda A. Tian Y.M. Masson N. Hamilton D.L. Jaakkola P. Barstead R. Hodgkin J. Maxwell P.H. Pugh C.W. Schofield C.J. Ratcliffe P.J. Cell. 2001; 107: 43-54Abstract Full Text Full Text PDF PubMed Scopus (2736) Google Scholar, 14Bruick R.K. McKnight S.L. Science. 2001; 294: 1337-1340Crossref PubMed Scopus (2114) Google Scholar, 30Hewitson K.S. McNeill L.A. Riordan M.V. Tian Y.M. Bullock A.N. Welford R.W. Elkins J.M. Oldham N.J. Bhattacharya S. Gleadle J.M. Ratcliffe P.J. Pugh C.W. Schofield C.J. J. Biol. Chem. 2002; 277: 26351-26355Abstract Full Text Full Text PDF PubMed Scopus (603) Google Scholar, 31Lando D. Peet D.J. Gorman J.J. Whelan D.A. Whitelaw M.L. Bruick R.K. Genes Dev. 2002; 16: 1466-1471Crossref PubMed Scopus (1229) Google Scholar, 32Hirsilä M. Koivunen P. Günzler V. Kivirikko K.I. Myllyharju J. J. Biol. Chem. 2003; 278: 30772-30780Abstract Full Text Full Text PDF PubMed Scopus (655) Google Scholar). However, our preliminary experiments showed that the HIF-1α PHDs did not accelerate IRP2 degradation. 2E. S. Hanson, unpublished observation. Whether the activity of these enzymes is also affected by physiologic changes in iron levels is not known. In summary, we propose an alternative model for the regulation of IRP2. We suggest that hydroxylation may be a critical event signaling IRP2 degradation by stimulating its interaction with the ubiquitination machinery. If true, the activity of the 2-OG-dependent dioxygenase catalyzing IRP2 hydroxylation would be expected to be responsive to physiologic changes in both oxygen and iron and therefore serve as an iron and oxygen sensor. We thank Dr. Martin Rechsteiner, Gregory Pratt, and Dr. Volkmar Günzler for helpful discussions and the anti-ubiquitin antibody; Dr. Jerry Kaplan, Dr. Dennis Winge, Brian Schneider, and Michelle Wallander for comments on the manuscript; Dr. Gary Allred (Frontier Scientific) for the DMOG; and all members of the Leibold laboratory for helpful discussions.