WNK1 Activates ERK5 by an MEKK2/3-dependent Mechanism
2004; Elsevier BV; Volume: 279; Issue: 9 Linguagem: Inglês
10.1074/jbc.m313465200
ISSN1083-351X
AutoresBing-e Xu, Steve Stippec, Lisa Y. Lenertz, Byung‐Hoon Lee, Wei Zhang, Youn-Kyoung Lee, Melanie H. Cobb,
Tópico(s)Protein Tyrosine Phosphatases
ResumoWNK1 belongs to a unique protein kinase family that lacks the catalytic lysine in its normal position. Mutations in human WNK1 and WNK4 have been implicated in causing a familial form of hypertension. Here we report that overexpression of WNK1 led to increased activity of cotransfected ERK5 in HEK293 cells. ERK5 activation was blocked by the MEK5 inhibitor U0126 and expression of a dominant negative MEK5 mutant. Expression of dominant negative mutants of MEKK2 and MEKK3 also blocked activation of ERK5 by WNK1. Moreover, both MEKK2 and MEKK3 coimmunoprecipitated with endogenous WNK1 from cell lysates. WNK1 phosphorylated both MEKK2 and -3 in vitro, and MEKK3 was activated by WNK1 in 293 cells. Finally, ERK5 activation by epidermal growth factor was attenuated by suppression of WNK1 expression using small interfering RNA. Taken together, these results place WNK1 in the ERK5 MAP kinase pathway upstream of MEKK2/3. WNK1 belongs to a unique protein kinase family that lacks the catalytic lysine in its normal position. Mutations in human WNK1 and WNK4 have been implicated in causing a familial form of hypertension. Here we report that overexpression of WNK1 led to increased activity of cotransfected ERK5 in HEK293 cells. ERK5 activation was blocked by the MEK5 inhibitor U0126 and expression of a dominant negative MEK5 mutant. Expression of dominant negative mutants of MEKK2 and MEKK3 also blocked activation of ERK5 by WNK1. Moreover, both MEKK2 and MEKK3 coimmunoprecipitated with endogenous WNK1 from cell lysates. WNK1 phosphorylated both MEKK2 and -3 in vitro, and MEKK3 was activated by WNK1 in 293 cells. Finally, ERK5 activation by epidermal growth factor was attenuated by suppression of WNK1 expression using small interfering RNA. Taken together, these results place WNK1 in the ERK5 MAP kinase pathway upstream of MEKK2/3. More than 500 protein kinases have been recognized in the human genome, ∼1.7% of all human genes (1Kostich M. English J. Madison V. Gheyas F. Wang L. Qiu P. Greene J. Laz T.M. Genome Biol. 2002; http://www.genomebiology.com/2002/3/9/Research/0043PubMed Google Scholar, 2Manning G. Whyte D.B. Martinez R. Hunter T. Sudarsanam S. Science. 2002; 298: 1912-1934Crossref PubMed Scopus (6312) Google Scholar). These enzymes play crucial roles in regulating cellular processes and participate in most, if not all, of the signal transduction pathways in cells. Among them, WNKs 1The abbreviations used are: WNK, with no lysine (K); ERK, extracellular signal-regulated kinase; MAPK, mitogen-activated protein kinase; MEK, mitogen-activated protein kinase/extracellular signal-regulated kinase kinase; MEKK, MEK kinase; EGF, epidermal growth factor; GST, glutathione S-transferase; HA, hemagglutinin; KM, kinase-dead mutant. (with no lysine (K)) comprise a newly described subfamily with a unique placement of the catalytic lysine responsible for binding to ATP (3Xu B. English J.M. Wilsbacher J.L. Stippec S. Goldsmith E.J. Cobb M.H. J. Biol. Chem. 2000; 275: 16795-16801Abstract Full Text Full Text PDF PubMed Scopus (417) Google Scholar). WNKs exist in multicellular organisms including plants, Caenorhabditis elegans, Drosophila, and mammals but not in unicellular organisms such as Saccharomyces cerevisiae (3Xu B. English J.M. Wilsbacher J.L. Stippec S. Goldsmith E.J. Cobb M.H. J. Biol. Chem. 2000; 275: 16795-16801Abstract Full Text Full Text PDF PubMed Scopus (417) Google Scholar, 4Verissimo F. Jordan P. Oncogene. 2001; 20: 5562-5569Crossref PubMed Scopus (226) Google Scholar, 5Murakami-Kojima M. Nakamichi N. Yamashino T. Mizuno T. Plant Cell Physiol. 2002; 43: 675-683Crossref PubMed Scopus (75) Google Scholar, 6Nakamichi N. Murakami-Kojima M. Sato E. Kishi Y. Yamashino T. Mizuno T. Biosci. Biotechnol. Biochem. 2002; 66: 2429-2436Crossref PubMed Scopus (49) Google Scholar, 7Morrison D.K. Murakami M.S. Cleghon V. J. Cell Biol. 2000; 150: F57-F62Crossref PubMed Scopus (138) Google Scholar, 8Wilson F.H. Disse-Nicodeme S. Choate K.A. Ishikawa K. Nelson-Williams C. Desitter I. Gunel M. Milford D.V. Lipkin G.W. Achard J.M. Feely M.P. Dussol B. Berland Y. Unwin R.J. Mayan H. Simon D.B. Farfel Z. Jeunemaitre X. Lifton R.P. Science. 2001; 293: 1107-1112Crossref PubMed Scopus (1233) Google Scholar). The well studied plant Arabidopsis thaliana has the largest number of WNK family members known with at least nine, some of which are implicated in controlling circadian rhythm (5Murakami-Kojima M. Nakamichi N. Yamashino T. Mizuno T. Plant Cell Physiol. 2002; 43: 675-683Crossref PubMed Scopus (75) Google Scholar, 6Nakamichi N. Murakami-Kojima M. Sato E. Kishi Y. Yamashino T. Mizuno T. Biosci. Biotechnol. Biochem. 2002; 66: 2429-2436Crossref PubMed Scopus (49) Google Scholar). In contrast, both C. elegans and Drosophila have only one WNK in their genomes. There are at least four mammalian WNK family members with high sequence identity within their catalytic domains (4Verissimo F. Jordan P. Oncogene. 2001; 20: 5562-5569Crossref PubMed Scopus (226) Google Scholar, 8Wilson F.H. Disse-Nicodeme S. Choate K.A. Ishikawa K. Nelson-Williams C. Desitter I. Gunel M. Milford D.V. Lipkin G.W. Achard J.M. Feely M.P. Dussol B. Berland Y. Unwin R.J. Mayan H. Simon D.B. Farfel Z. Jeunemaitre X. Lifton R.P. Science. 2001; 293: 1107-1112Crossref PubMed Scopus (1233) Google Scholar). Rat WNK1, the first mammalian member of this protein kinase subfamily identified, contains an N-terminal kinase domain followed by an autoinhibitory domain and a long C-terminal tail (3Xu B. English J.M. Wilsbacher J.L. Stippec S. Goldsmith E.J. Cobb M.H. J. Biol. Chem. 2000; 275: 16795-16801Abstract Full Text Full Text PDF PubMed Scopus (417) Google Scholar, 9Xu B. Min X. Stippec S. Lee B.H. Goldsmith E.J. Cobb M.H. J. Biol. Chem. 2002; 277: 48456-48462Abstract Full Text Full Text PDF PubMed Scopus (124) Google Scholar). Disruption of the WNK1 gene in mice leads to lethality before embryonic day 13 (10Zambrowicz B.P. Abuin A. Ramirez-Solis R. Richter L.J. Piggott J. BeltrandelRio H. Buxton E.C. Edwards J. Finch R.A. Friddle C.J. Gupta A. Hansen G. Hu Y. Huang W. Jaing C. Key Jr., B.W. Kipp P. Kohlhauff B. Ma Z.Q. Markesich D. Payne R. Potter D.G. Qian N. Shaw J. Schrick J. Shi Z.Z. Sparks M.J. Van Sligtenhorst I. Vogel P. Walke W. Xu N. Zhu Q. Person C. Sands A.T. Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 14109-14114Crossref PubMed Scopus (303) Google Scholar). Lifton and colleagues (8Wilson F.H. Disse-Nicodeme S. Choate K.A. Ishikawa K. Nelson-Williams C. Desitter I. Gunel M. Milford D.V. Lipkin G.W. Achard J.M. Feely M.P. Dussol B. Berland Y. Unwin R.J. Mayan H. Simon D.B. Farfel Z. Jeunemaitre X. Lifton R.P. Science. 2001; 293: 1107-1112Crossref PubMed Scopus (1233) Google Scholar) showed that mutations in WNK1 and WNK4 could lead to a familial type of human hypertension pseudohypoaldosteronism type II (PHAII). The mutations in WNK1 are large deletions in the first intron that increase its expression up to 5-fold. The WNK4 mutations are missense mutations located in regions near the two coiled-coil domains that are highly conserved among the four WNKs (8Wilson F.H. Disse-Nicodeme S. Choate K.A. Ishikawa K. Nelson-Williams C. Desitter I. Gunel M. Milford D.V. Lipkin G.W. Achard J.M. Feely M.P. Dussol B. Berland Y. Unwin R.J. Mayan H. Simon D.B. Farfel Z. Jeunemaitre X. Lifton R.P. Science. 2001; 293: 1107-1112Crossref PubMed Scopus (1233) Google Scholar). Recently, WNK4 has been shown to inhibit the activity of the sodium chloride cotransporter by reducing its membrane expression (11Wilson F.H. Kahle K.T. Sabath E. Lalioti M.D. Rapson A.K. Hoover R.S. Hebert S.C. Gamba G. Lifton R.P. Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 680-684Crossref PubMed Scopus (361) Google Scholar, 12Yang C.L. Angell J. Mitchell R. Ellison D.H. J. Clin. Investig. 2003; 111: 1039-1045Crossref PubMed Scopus (401) Google Scholar). This inhibition was reported to be dependent on the catalytic activity of WNK4, and the disease-causing mutation Q562E has less inhibitory effect. WNK1, on the other hand, was found to have no effect on sodium chloride cotransporter activity on its own. However, WNK1 relieved sodium chloride cotransporter inhibition by WNK4 (12Yang C.L. Angell J. Mitchell R. Ellison D.H. J. Clin. Investig. 2003; 111: 1039-1045Crossref PubMed Scopus (401) Google Scholar). These new findings provide a potential mechanism for regulation of blood pressure by WNKs and suggest an explanation for the finding that loss of WNK regulation causes hypertension. MAP kinase (MAPK) cascades are involved in many signal transduction pathways including those regulating cell cycle, transcription, apoptosis, and proliferation (13Chen Z. Gibson T.B. Robinson F. Silvestro L. Pearson G. Xu B. Wright A. Vanderbilt C. Cobb M.H. Chem. Rev. 2001; 101: 2449-2476Crossref PubMed Scopus (794) Google Scholar, 14Lewis T.S. Shapiro P.S. Ahn N.G. Adv. Cancer Res. 1998; 74: 49-139Crossref PubMed Google Scholar). A typical MAPK cascade consists of a MAPK kinase kinase (MAP3K or MEKK), and a MAPK kinase (MAP2K or MEK) which act in series on a MAPK. There are several MAPK pathways in mammals. One of them, the ERK5 MAPK pathway, contains MEK5 (MAP2K) and MEKK2/3 (MAP3K) as its upstream regulators (15Zhou G. Bao Z.Q. Dixon J.E. J. Biol. Chem. 1995; 270: 12665-12669Abstract Full Text Full Text PDF PubMed Scopus (540) Google Scholar, 16English J.M. Vanderbilt C.A. Xu S. Marcus S. Cobb M.H. J. Biol. Chem. 1995; 270: 28897-28902Abstract Full Text Full Text PDF PubMed Scopus (187) Google Scholar, 17Yan C. Luo H. Lee J.D. Abe J. Berk B.C. J. Biol. Chem. 2001; 276: 10870-10878Abstract Full Text Full Text PDF PubMed Scopus (133) Google Scholar, 18Chao T.H. Hayashi M. Tapping R.I. Kato Y. Lee J.D. J. Biol. Chem. 1999; 274: 36035-36038Abstract Full Text Full Text PDF PubMed Scopus (153) Google Scholar, 19Sun W. Kesavan K. Schaefer B.C. Garrington T.P. Ware M. Johnson N.L. Gelfand E.W. Johnson G.L. J. Biol. Chem. 2001; 276: 5093-5100Abstract Full Text Full Text PDF PubMed Scopus (124) Google Scholar). ERK5 contains a long C-terminal tail, and its kinase domain most resembles ERK2. ERK5 can be activated by proliferative stimuli such as epidermal growth factor (EGF), serum, lysophosphatidic acid, neurotrophins, and phorbol ester, as well as by stress stimuli such as sorbitol, H2O2, and UV irradiation (20English J.M. Pearson G. Hockenberry T. Shivakumar L. White M.A. Cobb M.H. J. Biol. Chem. 1999; 274: 31588-31592Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar, 21Kamakura S. Moriguchi T. Nishida E. J. Biol. Chem. 1999; 274: 26563-26571Abstract Full Text Full Text PDF PubMed Scopus (458) Google Scholar, 22Karihaloo A. O'Rourke D.A. Nickel C.H. Spokes K. Cantley L.G. J. Biol. Chem. 2001; 276: 9166-9173Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar, 23Kato Y. Kravchenko V.V. Tapping R.I. Han J. Ulevitch R.J. Lee J.D. EMBO J. 1997; 16: 7054-7066Crossref PubMed Scopus (498) Google Scholar, 24Cavanaugh J.E. Ham J. Hetman M. Poser S. Yan C. Xia Z. J. Neurosci. 2001; 21: 434-443Crossref PubMed Google Scholar, 25Diaz-Meco M.T. Moscat J. Mol. Cell. Biol. 2001; 21: 1218-1227Crossref PubMed Scopus (60) Google Scholar, 26Dong F. Gutkind J.S. Larner A.C. J. Biol. Chem. 2001; 276: 10811-10816Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar, 27Pearson G.W. Cobb M.H. J. Biol. Chem. 2002; 277: 48094-48098Abstract Full Text Full Text PDF PubMed Scopus (17) Google Scholar, 28Watson F.L. Heerssen H.M. Bhattacharyya A. Klesse L. Lin M.Z. Segal R.A. Nat. Neurosci. 2001; 4: 981-988Crossref PubMed Scopus (384) Google Scholar). The known ERK5 substrates include the MEF2 family members, MEF2A, -C, and -D, and the ETS-like transcription factor SAP1a (21Kamakura S. Moriguchi T. Nishida E. J. Biol. Chem. 1999; 274: 26563-26571Abstract Full Text Full Text PDF PubMed Scopus (458) Google Scholar, 23Kato Y. Kravchenko V.V. Tapping R.I. Han J. Ulevitch R.J. Lee J.D. EMBO J. 1997; 16: 7054-7066Crossref PubMed Scopus (498) Google Scholar). The upstream kinase MEK5 contains two alternatively spliced forms, α and β (16English J.M. Vanderbilt C.A. Xu S. Marcus S. Cobb M.H. J. Biol. Chem. 1995; 270: 28897-28902Abstract Full Text Full Text PDF PubMed Scopus (187) Google Scholar). MEK5 is the only known MAP2K in the ERK5 pathway (21Kamakura S. Moriguchi T. Nishida E. J. Biol. Chem. 1999; 274: 26563-26571Abstract Full Text Full Text PDF PubMed Scopus (458) Google Scholar). Its sequence similarity to MEK1/2 renders it susceptible to the pharmacological inhibitors PD98059 and U0126, originally identified as selective blockers of MEK1/2 (21Kamakura S. Moriguchi T. Nishida E. J. Biol. Chem. 1999; 274: 26563-26571Abstract Full Text Full Text PDF PubMed Scopus (458) Google Scholar, 29Dudley D.T. Pang L. Decker S.J. Bridges A.J. Saltiel A.R. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7686-7689Crossref PubMed Scopus (2595) Google Scholar, 30Favata M.F. Horiuchi K.Y. Manos E.J. Daulerio A.J. Stradley D.A. Feeser W.S. Van Dyk D.E. Pitts W.J. Earl R.A. Hobbs F. Copeland R.A. Magolda R.L. Scherle P.A. Trzaskos J.M. J. Biol. Chem. 1998; 273: 18623-18632Abstract Full Text Full Text PDF PubMed Scopus (2754) Google Scholar). MEKK2 and MEKK3 are two closely related MAP3Ks with extremely high sequence identity within their catalytic domains, although less so within their N-terminal regulatory domains (31Blank J.L. Gerwins P. Elliott E.M. Sather S. Johnson G.L. J. Biol. Chem. 1996; 271: 5361-5368Abstract Full Text Full Text PDF PubMed Scopus (207) Google Scholar). Both kinases have been shown to interact with MEK5 directly and activate the MEK5-ERK5 pathway (18Chao T.H. Hayashi M. Tapping R.I. Kato Y. Lee J.D. J. Biol. Chem. 1999; 274: 36035-36038Abstract Full Text Full Text PDF PubMed Scopus (153) Google Scholar, 19Sun W. Kesavan K. Schaefer B.C. Garrington T.P. Ware M. Johnson N.L. Gelfand E.W. Johnson G.L. J. Biol. Chem. 2001; 276: 5093-5100Abstract Full Text Full Text PDF PubMed Scopus (124) Google Scholar). A dominant negative mutant of MEKK3 is capable of blocking ERK5 activation by growth factors and H2O2 (18Chao T.H. Hayashi M. Tapping R.I. Kato Y. Lee J.D. J. Biol. Chem. 1999; 274: 36035-36038Abstract Full Text Full Text PDF PubMed Scopus (153) Google Scholar, 19Sun W. Kesavan K. Schaefer B.C. Garrington T.P. Ware M. Johnson N.L. Gelfand E.W. Johnson G.L. J. Biol. Chem. 2001; 276: 5093-5100Abstract Full Text Full Text PDF PubMed Scopus (124) Google Scholar). In the present study, we demonstrate that overexpression of WNK1 in HEK293 cells causes activation of ERK5 but not ERK2. The activation is blocked by the MEK5 inhibitor U0126 and is dependent on MEKK2/3. Furthermore, WNK1 is required for activation of ERK5 by EGF. These results place WNK1 in the ERK5 MAPK pathway upstream of MEKK2/3. Plasmids, Subcloning, and Mutagenesis—pCEP4-HA-ERK5 and pCEP4-HA-ERK2 were as described by Xu et al. (3Xu B. English J.M. Wilsbacher J.L. Stippec S. Goldsmith E.J. Cobb M.H. J. Biol. Chem. 2000; 275: 16795-16801Abstract Full Text Full Text PDF PubMed Scopus (417) Google Scholar). pCMV5-Myc-WNK1-(1–491) and full-length WNK1 were described previously (3Xu B. English J.M. Wilsbacher J.L. Stippec S. Goldsmith E.J. Cobb M.H. J. Biol. Chem. 2000; 275: 16795-16801Abstract Full Text Full Text PDF PubMed Scopus (417) Google Scholar, 9Xu B. Min X. Stippec S. Lee B.H. Goldsmith E.J. Cobb M.H. J. Biol. Chem. 2002; 277: 48456-48462Abstract Full Text Full Text PDF PubMed Scopus (124) Google Scholar). pCMV5–3xFlag vector was purchased from Sigma. pCMV5-HA-MEKK2 and -3 were kindly provided by Dr. Gary Johnson (University of Colorado, Denver, CO/University of North Carolina, Chapel Hill, NC) (31Blank J.L. Gerwins P. Elliott E.M. Sather S. Johnson G.L. J. Biol. Chem. 1996; 271: 5361-5368Abstract Full Text Full Text PDF PubMed Scopus (207) Google Scholar). To make FLAG-MEK5α K195M and β K106M, FLAG-MEKK2, and FLAG-MEKK3 constructs, the open reading frames were amplified by PCR and subcloned into pCMV5–3xFlag. To make constructs expressing fusions of MEKK2 and -3, cDNAs encoding residues 1–350 and 1–354 (N constructs) and 332–619 K385M and 350–626 K391M (C constructs) of MEKK2 and MEKK3, respectively, were amplified by PCR and subcloned into pGEX-KG. pCMV5-Myc-TAO2-(1–993) D169A was as described by Chen et al. (32Chen Z. Raman M. Chen L. Lee S.F. Gilman A.G. Cobb M.H. J. Biol. Chem. 2003; 278: 22278-22283Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar). Site-directed mutagenesis was performed using the QuikChange kit (Stratagene) according to the manufacturer's directions. Proteins and Antibodies—GST·MEF2C-(204–321) and His6-MEK6 K82M were as described in Refs. 27Pearson G.W. Cobb M.H. J. Biol. Chem. 2002; 277: 48094-48098Abstract Full Text Full Text PDF PubMed Scopus (17) Google Scholar and 33Chen Z. Cobb M.H. J. Biol. Chem. 2001; 276: 16070-16075Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar. His6-WNK1-(198–491) was described previously (9Xu B. Min X. Stippec S. Lee B.H. Goldsmith E.J. Cobb M.H. J. Biol. Chem. 2002; 277: 48456-48462Abstract Full Text Full Text PDF PubMed Scopus (124) Google Scholar). Myelin basic protein was purchased from Sigma. GST·MEKK2/3-N/C proteins were expressed using standard protocols. The anti-HA antibody (12CA5) was from Berkeley Antibody Company. The anti-Myc antibody (9E10) was from the National Cell Culture Center. The monoclonal anti-FLAG antibody was obtained from Sigma. The polyclonal anti-WNK1 antibody Q256 and its preimmune serum were as described by Xu et al. (3Xu B. English J.M. Wilsbacher J.L. Stippec S. Goldsmith E.J. Cobb M.H. J. Biol. Chem. 2000; 275: 16795-16801Abstract Full Text Full Text PDF PubMed Scopus (417) Google Scholar). The anti-ERK1/2 antibody Y691 was described previously (34Boulton T.G. Cobb M.H. Cell Regul. 1991; 2: 357-371Crossref PubMed Scopus (283) Google Scholar). The anti-ERK5 antibody was purchased from Sigma. Cell Culture, Transfections, and Harvesting—HEK293 cells were maintained, transfected, and harvested as described by Xu et al. (35Xu B. Wilsbacher J.L. Collisson T. Cobb M.H. J. Biol. Chem. 1999; 274: 34029-34035Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar). Lysis buffer with 1% Triton X-100 or Nonidet P-40 was used, except for coimmunoprecipitation experiments, in which case detergent was omitted. HeLa cells were grown in Dulbecco's modified Eagle's medium with 10% fetal bovine serum and 1% l-glutamine at 37 °C under 10% CO2. For RNA interference experiments, HeLa cells were grown in 6-well plates to 30% confluence on the day of transfection. The cells were transfected with luciferase double-stranded RNA as a control or WNK1 double-stranded RNA oligonucleotides using Oligofectamine (Invitrogen) according to the manufacturer's instructions. After 72 h, the cells were deprived of serum for 4 h before treatment. Immunoblotting and Immunoprecipitation—For immunoblotting, proteins from the cell lysates were separated by SDS-PAGE followed by electrotransference to nitrocellulose membranes. The membranes were incubated with the indicated antibodies and developed using enhanced chemiluminescence. For immunoprecipitation, cell lysates were incubated with the respective antibody and protein A-Sepharose beads for 2 h at 4 °C. The beads were washed with either 1 m NaCl, 20 mm Tris-HCl (pH 7.4) (for the kinase assay) or 0.5 m NaCl, 10 mm Tris-HCl (pH 7.4), 5 mm MgCl2 (for coimmunoprecipitation). In Vitro Kinase Assays—Kinase assays were performed as described by Xu et al. (35Xu B. Wilsbacher J.L. Collisson T. Cobb M.H. J. Biol. Chem. 1999; 274: 34029-34035Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar). Most kinase reactions were in buffer containing 50 μm ATP; reactions with FLAG-MEKK3 included 10 μm ATP. To examine phosphorylation of FLAG-MEKK2/3 (kinase-dead) phosphorylation by His6-WNK1-(198–491), FLAG-MEKK2/3 (kinase-dead) proteins immunoprecipitated from cell lysates were preincubated in 1× kinase buffer for 30 min at 30 °C prior to addition of His6-WNK1-(198–491) and ATP. Overexpression of WNK1 Activates ERK5 in HEK293 Cells— We reported previously that overexpression of WNK1-(1–555) did not activate cotransfected MAPKs in HEK293 cells (3Xu B. English J.M. Wilsbacher J.L. Stippec S. Goldsmith E.J. Cobb M.H. J. Biol. Chem. 2000; 275: 16795-16801Abstract Full Text Full Text PDF PubMed Scopus (417) Google Scholar). Subsequently, we identified an autoinhibitory domain in WNK1 located between residues 491 and 555 (9Xu B. Min X. Stippec S. Lee B.H. Goldsmith E.J. Cobb M.H. J. Biol. Chem. 2002; 277: 48456-48462Abstract Full Text Full Text PDF PubMed Scopus (124) Google Scholar). Because WNK1-(1–491) had much higher catalytic activity than WNK1-(1–555) in vitro, we thought it was necessary to examine the more active form, WNK1-(1–491), for its possible effects on the activities of MAPKs. Thus, HA-tagged ERK2 or ERK5 was transfected into HEK293 cells along with either the vector or Myc-tagged WNK1-(1–491). Remarkably, ERK5 activity assayed using its substrate GST·MEF2C was increased significantly by WNK1-(1–491) compared with the vector control (Fig. 1A). In contrast, ERK2 activity was not affected by WNK1-(1–491) when assayed with myelin basic protein (Fig. 1B). Enhanced ERK5 autophosphorylation is another indicator of its activation state. As shown in Fig. 1C, WNK1-(1–491) increased ERK5 autophosphorylation as well as its activity toward GST·MEF2C. Activation of ERK5 was greater with wild type WNK1; a kinase defective mutant, WNK1-(1–491) K233M, did not stimulate ERK5 autophosphorylation or activity toward GST·MEF2C as well, although a small increase in ERK5 autophosphorylation was detected (Fig. 1C). We also found that expression of full-length WNK1 activated ERK5 (Fig. 1D). These results suggest that WNK1 might act as an upstream kinase in the ERK5 MAPK pathway. Activation of ERK5 by WNK1 Requires MEK5 Activity—We next examined whether activation of ERK5 by WNK1 required the known MEK in the ERK5 pathway, MEK5. To address this question, we first used kinase-dead mutants (KM) of the two MEK5 isoforms, α and β. MEK5α K195M fully blocked activation of ERK5 by WNK1 (Fig. 2A). Both ERK5 autophosphorylation and activity toward GST·MEF2C were reduced to the control value. Unexpectedly, MEK5β K106M had no effect in the same experiment. In addition, we utilized the pharmacological compound U0126, which specifically inhibits the activities of MEK1, MEK2, and MEK5 but not other MAP2Ks (30Favata M.F. Horiuchi K.Y. Manos E.J. Daulerio A.J. Stradley D.A. Feeser W.S. Van Dyk D.E. Pitts W.J. Earl R.A. Hobbs F. Copeland R.A. Magolda R.L. Scherle P.A. Trzaskos J.M. J. Biol. Chem. 1998; 273: 18623-18632Abstract Full Text Full Text PDF PubMed Scopus (2754) Google Scholar). At 10 μm, U0126 efficiently blocked the activation of ERK5 by WNK1 in transfected cells treated for 1 h prior to harvest (Fig. 2B). These findings indicate that MEK5 activity is crucial for activation of ERK5 by WNK1. Activation of ERK5 by WNK1 Is Dependent on MEKK2 and MEKK3—MEKK2 and MEKK3, the two known MAP3Ks in the ERK5 MAPK pathway, interact with MEK5 and activate the MEK5-ERK5 cascade (18Chao T.H. Hayashi M. Tapping R.I. Kato Y. Lee J.D. J. Biol. Chem. 1999; 274: 36035-36038Abstract Full Text Full Text PDF PubMed Scopus (153) Google Scholar, 19Sun W. Kesavan K. Schaefer B.C. Garrington T.P. Ware M. Johnson N.L. Gelfand E.W. Johnson G.L. J. Biol. Chem. 2001; 276: 5093-5100Abstract Full Text Full Text PDF PubMed Scopus (124) Google Scholar). To determine whether WNK1 acts upstream or downstream of MEKK2/3 in the ERK5 pathway, we examined the effects of dominant negative mutants of MEKK2/3 on ERK5 activation by WNK1. MEKK3 K391M prevented ERK5 activation by WNK1 (Fig. 3A). Both ERK5 autophosphorylation and activity toward GST·MEF2C decreased to basal levels if the MEKK3 K391M mutant was coexpressed. Similar results were also observed for MEKK2 K385M (Fig. 3B). On the other hand, a kinase-dead form of a MAP3K not known to be involved in the ERK5 pathway, TAO2-(1–993) D169A, did not affect the activation of ERK5 by WNK1 (Fig. 3B). Likewise, kinase-dead mutants of the MAP4K PAK1 and the MAP3K MEKK1 also failed to block ERK5 activation by WNK1 (data not shown). These data suggest that WNK1 most likely acts upstream of MEKK2 and MEKK3 in the ERK5 pathway. Endogenous WNK1 Coimmunoprecipitates with MEKK2 and MEKK3—Because dominant negative mutants of MEKK2 and MEKK3 could block ERK5 activation by WNK1, we tested whether WNK1 and MEKK2/3 might exist in a complex in cells. HEK293 cells were transfected with Myc-tagged WNK1-(1–491) and HA-tagged MEKK3. WNK1 was immunoprecipitated with anti-Myc, and the precipitates were blotted with anti-HA. HA-MEKK3 was readily detected in the precipitates from lysates containing both WNK1 and MEKK3, but not from those containing only MEKK3 (Fig. 4A). The converse also revealed an association: MEKK3 was immunoprecipitated with anti-HA, and the precipitates were blotted with anti-Myc. Again, Myc-WNK1 was only detected in the precipitates from lysates containing both MEKK3 and WNK1 and not from those containing only WNK1 (Fig. 4B). To address the question of whether MEKK2 might also coimmunoprecipitate with WNK1-(1–491), HEK293 cells were transfected with either the vector or FLAG-tagged MEKK2/3 along with Myc-WNK1-(1–491). MEKK2/3 proteins were immunoprecipitated with anti-FLAG antibody and the precipitates were blotted with anti-Myc antibody to detect the presence of WNK1. MEKK2, like MEKK3, pulled down WNK1-(1–491) (Fig. 4C). The catalytic activities of MEKK2/3 were not required because the kinase-dead mutants also coimmunoprecipitated with WNK1 (Fig. 4C). Next we wanted to examine whether the endogenous WNK1 could coimmunoprecipitate with MEKK2/3. HEK293 cells were transfected with FLAG-tagged MEKK2 or MEKK3. Endogenous WNK1 was immunoprecipitated from the cell lysates with the Q256 antibody, and the precipitates were blotted with anti-FLAG. As a control, the preimmune serum of Q256 was also used for immunoprecipitation. As shown in Fig. 4D, WNK1 immunoprecipitates contained significant amounts of MEKK2/3. Neither the preimmune serum nor a control antibody Y691 (anti-ERK1/2) pulled down comparable amounts of MEKK2. These results support the conclusion that MEKK2/3 and WNK1 are associated in cells. WNK1 Phosphorylates the N Terminus of MEKK2 and MEKK3 in Vitro—Because WNK1 lies upstream of MEKK2/3 and they can exist in a complex in cells, we reasoned that MEKK2 and MEKK3 might be direct substrates of WNK1. To test this possibility, we used FLAG-tagged MEKK2/3 (KM) proteins immunoprecipitated from 293 cells as in vitro substrates for His-WNK1-(198–491). As shown in Fig. 5A, both MEKK2 and MEKK3 were phosphorylated by WNK1. We also expressed GST fusion proteins of MEKK2 and MEKK3 in bacteria to test as WNK1 substrates. N-terminal and C-terminal (KM) fragments of the two MAP3Ks were expressed, and in both cases, the N-terminal fragment, but not the C-terminal fragment, was phosphorylated by recombinant WNK1 (Fig. 5B) and also by endogenous WNK1 immunoprecipitated from cells treated with NaCl (Fig. 5C). These results are consistent with the idea that WNK1 is a MAP4K in the ERK5 pathway. However, the phosphorylation of MEKK2/3 by WNK1 does not appear to directly alter their catalytic activities because preincubation of MEKK2/3 with WNK1 in the presence of ATP neither increased nor decreased their kinase activities (data not shown). Overexpression of WNK1 Activates Cotransfected MEKK3 in HEK293 Cells—To test whether WNK1 could affect the activity of MEKK3 in cells, we transfected 293 cells with either vector control or Myc-tagged WNK1-(1–491) along with FLAG-tagged MEKK3. MEKK3 was immunoprecipitated from cell lysates and assayed using His6-MEK6 K82M as the substrate. Overexpression of WNK1-(1–491) increased the autophosphorylation of MEKK3 as well as its activity toward His6-MEK6 K82M (Fig. 5D), showing that coexpression with WNK1 in cells stimulates MEKK3 activity. Interestingly, this activation is independent of the catalytic activity of WNK1, because a kinasedefective mutant of WNK1 also activated cotransfected MEKK3 to a similar extent (data not shown). WNK1 Is Required for Activation of ERK5 by EGF in HeLa Cells—We have presented evidence that overexpression of WNK1 is sufficient to activate the ERK5 MAPK pathway. However, whether WNK1 is required for ERK5 activation under any physiological circumstances is still unknown. Growth factors such as EGF and stress stimuli are well documented activators of ERK5. Therefore, we tested whether WNK1 is required for ERK5 activation by EGF using RNA interference to knock down expression of endogenous WNK1 in HeLa cells. ERK5 activation by many agents including EGF leads to a decrease in its electrophoretic motility on gels (Fig. 6A), providing a simple assessment of its activity. We successfully reduced expression of endogenous WNK1 as shown in Fig. 6B. ERK5 activation by 1 ng/ml EGF was partially reduced by suppression of WNK1 expression, suggesting that WNK1 is required for ERK5 activation by EGF under these circumstances. With higher concentrations of EGF, the requirement for WNK1 became less pronounced (data not shown). Since it was first cloned and characterized three years ago, the atypical protein kinase WNK1 has attracted considerable interest primarily because it has been linked to the regulation of blood pressure (3Xu B. English J.M. Wilsbacher J.L. Stippec S. Goldsmith E.J. Cobb M.H. J. Biol. Chem. 2000; 275: 16795-16801Abstract Full Text Full Text PDF PubMed Scopus (417) Google Scholar, 8Wilson F.H. Disse-Nicodeme S. Choate K.A. Ishikawa K. Nelson-Williams C. Desitter I. Gunel M. Milford D.V. Lipkin G.W. Achard J.M. Feely M.P. Dussol B. Berland Y. Unwin R.J. Mayan H. Simon D.B. Farfel Z. Jeunemaitre X. Lifton R.P. Science. 2001; 293: 1107-1112Crossref PubMed Scopus (1233) Google Scholar). Although progress was made in the last 2 years in defining how WNK1 is regulated and in suggesting an effect of WNK1 on ion channels, the biochemical pathways regulated by WNK1 remain unknown. Here we provide evidence that WNK1 can activate the MAPK ERK5 through MEKK2/3. The following observations suggest that WNK1 is a MAP4K in the ERK5 pathway: 1) its activation of ERK5 can be blocked by kinase-dead mutants of MEKK2/3; 2) WNK1 interacts with MEKK2/3 in cells; 3) WNK1 phosphorylates MEKK2/3 in vitro; and 4) coexpression of WNK1 with MEKK3 activates it in cells. Nevertheless, in vitro experiments suggest that phosphorylation of MEKK2/3 by WNK1 apparently does not directly activate them. In cells, WNK1 activated MEKK3; however, WNK1 kinase activity was not required, because a kinase-dead mutant of WNK1 had the same stimulatory effect as the wild type protein. One possible explanation comes from the observation that WNK1 exists as a tetramer (9Xu B. Min X. Stippec S. Lee B.H. Goldsmith E.J. Cobb M.H. J. Biol. Chem. 2002; 277: 48456-48462Abstract Full Text Full Text PDF PubMed Scopus (124) Google Scholar). The kinase-dead protein may interact with endogenous wild type WNK1, promoting activation of the MAP3Ks. Alternatively, protein-protein interactions could change the conformation of MEKK3 or its ability to form complexes. In contrast, activation of ERK5 by wild type WNK1 is greater than activation by kinase-dead WNK1. These apparently contradictory findings may be reconciled if WNK1 contributes two events toward ERK5 activation: stimulation of the MAP3K by a noncatalytic mechanism coupled with assembly of an ERK5 activation complex that requires phosphorylation of the MAP3K or some other component in the complex. If so, phosphorylation of MEKK2/3 by WNK1 may be important for transmitting the activation signal to ERK5 even though it is not required for activation of MEKK2/3. Interestingly, our data suggest a role for only one of two isoforms of MEK5 in activation of ERK5 by WNK1. MEK5α has an additional 89 residues at its N terminus compared with MEK5β due to alternative splicing, whereas MEK5β appears to be more widely expressed (16English J.M. Vanderbilt C.A. Xu S. Marcus S. Cobb M.H. J. Biol. Chem. 1995; 270: 28897-28902Abstract Full Text Full Text PDF PubMed Scopus (187) Google Scholar). The inactive form of MEK5α was a very effective inhibitor, whereas the comparable mutant of MEK5β was not. A recent study has suggested that wild type MEK5β can act as a dominant negative to block ERK5 activation by EGF or by a constitutively active form of MEK5α (36Cameron S.J. Abe J.I. Malik S. Che W. Yang J. J. Biol. Chem. 2004; 279: 1506-1512Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar). The mechanisms underlying this discrepancy are not known. Significantly, we have demonstrated that WNK1 is required for ERK5 activation by EGF. We also found a partial requirement for WNK1 in the activation of ERK5 by H2O2 (data not shown). In contrast, WNK1 was not required for activation of ERK5 by either sorbitol or NaCl (data not shown). Surprisingly, there is a poor correlation of the requirement for WNK1 in ERK5 activation by these agents and the sensitivity of WNK1 itself to be activated by them. NaCl and sorbitol are significantly stronger WNK1 activators than H2O2, and we have thus far failed to reveal activation of WNK1 by EGF. Thus, the potential involvement of WNK1 in a pathway may not be obvious based on changes in its kinase activity alone. These results indicate that WNK1 participates in ERK5 activation by a subset of physiological ERK5 regulators. There are four WNK homologs in mammals that share a high degree of identity within their kinase domains but significantly lower identity outside the kinase domains (4Verissimo F. Jordan P. Oncogene. 2001; 20: 5562-5569Crossref PubMed Scopus (226) Google Scholar, 8Wilson F.H. Disse-Nicodeme S. Choate K.A. Ishikawa K. Nelson-Williams C. Desitter I. Gunel M. Milford D.V. Lipkin G.W. Achard J.M. Feely M.P. Dussol B. Berland Y. Unwin R.J. Mayan H. Simon D.B. Farfel Z. Jeunemaitre X. Lifton R.P. Science. 2001; 293: 1107-1112Crossref PubMed Scopus (1233) Google Scholar). A recent study suggests that WNK1 and WNK4 might have opposite effects on channel regulation (12Yang C.L. Angell J. Mitchell R. Ellison D.H. J. Clin. Investig. 2003; 111: 1039-1045Crossref PubMed Scopus (401) Google Scholar). Thus, it will be important to determine whether any of the other WNK family members also activate ERK5. Currently, we are testing WNK2 and WNK4 for their effects on the ERK5 pathway. In summary, we demonstrated in this study that WNK1 activates the ERK5 MAPK pathway through MEKK2/3. The identification of potential WNK1 substrates other than myelin basic protein provides a starting point to map biochemical pathways that WNK1 regulates. Future studies should shed more light on the important roles WNK1 plays in various cellular processes including ion transport. We thank Anthony Anselmo and Tara Beers Gibson for critical suggestions and comments about the manuscript, Gray Pearson for helpful discussions, Wei Chen, Kathy McGlynn, and Svetlana Earnest for valuable reagents, and Dionne Ware for administrative assistance.
Referência(s)