Inhibitor-2 Regulates Protein Phosphatase-1 Complexed with NimA-related Kinase to Induce Centrosome Separation
2002; Elsevier BV; Volume: 277; Issue: 46 Linguagem: Inglês
10.1074/jbc.m208035200
ISSN1083-351X
AutoresMasumi Eto, Elizabeth Elliott, Todd D. Prickett, David L. Brautigan,
Tópico(s)Fungal and yeast genetics research
ResumoCentrosome separation is regulated by balance ofin situ protein kinase/phosphatase activities during the cell cycle. The mammalian NimA-related kinase Nek2 forms a complex with the catalytic subunit of protein phosphatase-1 (PP1C). This complex is located at centrosomes and has been implicated in regulation of the cycle of duplication and separation. Inhibitor-2 (Inh2) is an inhibitor protein specific for PP1C, and its expression level fluctuates during the cell cycle. Here we report cellular regulation of the Nek2·PP1C complex by Inh2. PP1C-binding segments of Nek2 were isolated by yeast two-hybrid screening using Inh2 bait. Inh2 indirectly associates with Nek2 via PP1C, which binds to both proteins, forming a bridged heterotrimeric complex. Double Ala mutation of the PP1C-binding site (KVHF) in Nek2 eliminated both PP1C and Inh2 interactions in both a yeast conjugation assay and an in vitro binding assay. The kinase activity of Nek2·PP1C was enhanced 2-fold by addition of recombinant Inh2, with EC50 = 10 nm. Immunofluorescence showed concentration of endogenous Inh2 at centrosomes and in a region surrounding the centrosomes. Transient expression of wild-type Inh2 increased by 5-fold dispersed/split centrosomes in fibroblasts, mimicking the phenotype produced by overexpression of Nek2. Deletion of the Inh2 C-terminal domain yielded Inh2-(1–118), which failed to interact with or activate the Nek2·PP1C complex, suggesting that the C-terminal region of Inh2 is required for regulation of the Nek2·PP1C complex. Thus, Inh2 can enhance the kinase activity of the Nek2·PP1C complex via inhibition of phosphatase activity to initiate centrosome separation. Centrosome separation is regulated by balance ofin situ protein kinase/phosphatase activities during the cell cycle. The mammalian NimA-related kinase Nek2 forms a complex with the catalytic subunit of protein phosphatase-1 (PP1C). This complex is located at centrosomes and has been implicated in regulation of the cycle of duplication and separation. Inhibitor-2 (Inh2) is an inhibitor protein specific for PP1C, and its expression level fluctuates during the cell cycle. Here we report cellular regulation of the Nek2·PP1C complex by Inh2. PP1C-binding segments of Nek2 were isolated by yeast two-hybrid screening using Inh2 bait. Inh2 indirectly associates with Nek2 via PP1C, which binds to both proteins, forming a bridged heterotrimeric complex. Double Ala mutation of the PP1C-binding site (KVHF) in Nek2 eliminated both PP1C and Inh2 interactions in both a yeast conjugation assay and an in vitro binding assay. The kinase activity of Nek2·PP1C was enhanced 2-fold by addition of recombinant Inh2, with EC50 = 10 nm. Immunofluorescence showed concentration of endogenous Inh2 at centrosomes and in a region surrounding the centrosomes. Transient expression of wild-type Inh2 increased by 5-fold dispersed/split centrosomes in fibroblasts, mimicking the phenotype produced by overexpression of Nek2. Deletion of the Inh2 C-terminal domain yielded Inh2-(1–118), which failed to interact with or activate the Nek2·PP1C complex, suggesting that the C-terminal region of Inh2 is required for regulation of the Nek2·PP1C complex. Thus, Inh2 can enhance the kinase activity of the Nek2·PP1C complex via inhibition of phosphatase activity to initiate centrosome separation. Centrosomes are structures in the eucaryotic cell that function as centers for organizing microtubules (see review in Ref. 1Doxsey S. Nat. Rev. Mol. Cell. Biol. 2001; 2: 688-698Crossref PubMed Scopus (316) Google Scholar). During the cell cycle, centrosomes are first duplicated coincident with S phase and then undergo separation at the onset of mitosis and become the spindle poles for chromosome segregation. Recent discoveries have shown that protein phosphorylation is involved in regulation of the centrosome cycle (2Meraldi P. Nigg E.A. FEBS Lett. 2002; 521: 9-13Crossref PubMed Scopus (136) Google Scholar). Kinases such as cyclin-dependent kinase (CDK) 1The abbreviations used are: CDK, cyclin-dependent kinase; PP, protein phosphatase; PP1C, protein phosphatase-1 catalytic subunit; Inh2, inhibitor-2; GST, glutathione S-transferase; HA, hemagglutinin; MOPS, 4-morpholinepropanesulfonic acid; DMEM, Dulbecco's modified Eagle's medium; MBP, myelin basic protein; PIPES, 1,4-piperazinediethanesulfonic acid; PBS, phosphate-buffered saline; CTD, C-terminal domain; IBF, Inh2-binding fragment; MAPK, mitogen-activated protein kinase 1The abbreviations used are: CDK, cyclin-dependent kinase; PP, protein phosphatase; PP1C, protein phosphatase-1 catalytic subunit; Inh2, inhibitor-2; GST, glutathione S-transferase; HA, hemagglutinin; MOPS, 4-morpholinepropanesulfonic acid; DMEM, Dulbecco's modified Eagle's medium; MBP, myelin basic protein; PIPES, 1,4-piperazinediethanesulfonic acid; PBS, phosphate-buffered saline; CTD, C-terminal domain; IBF, Inh2-binding fragment; MAPK, mitogen-activated protein kinaseand NimA-related kinase (Nek2) have been implicated in regulation of the centrosome cycle. In particular, Nek2 is the human relative of the NimA protein Ser/Thr kinase first described fromAspergillus and localized at centrosomes (3Schultz S.J. Fry A.M. Sutterlin C. Ried T. Nigg E.A. Cell Growth Differ. 1994; 5: 625-635PubMed Google Scholar). Ectopic expression of Nek2 gives separation without spindle formation, consistent with a loss of centrosome cohesion (4Fry A.M. Meraldi P. Nigg E.A. EMBO J. 1998; 17: 470-481Crossref PubMed Scopus (342) Google Scholar). In addition to kinases, the protein phosphatase-1 (PP1) α-isoform is localized at centrosomes (5Andreassen P.R. Lacroix F.B. Villa-Moruzzi E. Margolis R.L. J. Cell Biol. 1998; 141: 1207-1215Crossref PubMed Scopus (172) Google Scholar). Indeed, the catalytic subunit of PP1 (PP1C) binds to a PP1-binding motif (-KVHF-) in the C-terminal non-catalytic region of Nek2 (6Helps N.R. Luo X. Barker H.M. Cohen P.T. Biochem. J. 2000; 349: 509-518Crossref PubMed Scopus (163) Google Scholar). Coexpression of PP1 antagonizes Nek2 function via dephosphorylation of Nek2 itself and its substrates, indicating that PP1 acts a negative regulator for centrosome separation (6Helps N.R. Luo X. Barker H.M. Cohen P.T. Biochem. J. 2000; 349: 509-518Crossref PubMed Scopus (163) Google Scholar, 7Meraldi P. Nigg E.A. J. Cell Sci. 2001; 114: 3749-3757Crossref PubMed Google Scholar). On the other hand, phosphorylation of PP1C by Nek2 reduces the phosphatase activity of the complex (6Helps N.R. Luo X. Barker H.M. Cohen P.T. Biochem. J. 2000; 349: 509-518Crossref PubMed Scopus (163) Google Scholar). Thus, the kinase·phosphatase complex of Nek2·PP1C is expected to function as a bistable switch described by Ferrell (8Ferrell Jr., J.E. Curr. Opin. Cell Biol. 2002; 14: 140-148Crossref PubMed Scopus (849) Google Scholar), potentially creating a double-negative feedback circuit for regulation of centrosome dynamics. Phosphatase inhibitor-2 (Inh2) was first described in 1976 as a heat-stable protein that inhibited protein phosphatase activity (9Huang F.L. Glinsmann W. FEBS Lett. 1976; 62: 326-329Crossref PubMed Scopus (53) Google Scholar). The specificity of Inh2 was later used for functionally defining type 1versus type 2 Ser/Thr phosphatases (10Ingebritsen T.S. Cohen P. Eur. J. Biochem. 1983; 132: 255-261Crossref PubMed Scopus (345) Google Scholar). Several lines of evidence suggest involvement of Inh2 in cell cycle regulation. The expression level of Inh2 fluctuates during the cell cycle and is enhanced at mitosis (11Brautigan D.L. Sunwoo J. Labbe J.C. Fernandez A. Lamb N.J. Nature. 1990; 344: 74-78Crossref PubMed Scopus (71) Google Scholar). Nuclear localization of Inh2 occurs in parallel to S phase entry, and Inh2 is retained in the cytoplasm of cells at high density (12Kakinoki Y. Somers J. Brautigan D.L. J. Biol. Chem. 1997; 272: 32308-32314Abstract Full Text Full Text PDF PubMed Scopus (35) Google Scholar, 13Leach C. Eto M. Brautigan D.L. J. Cell Sci. 2002; 115: 3739-3745Crossref PubMed Scopus (14) Google Scholar). Glc8, a yeast homolog of Inh2, compensates for mutation of Ipl1 kinase in Saccharomyces cerevisiae, which suffers severe missegregation of chromosomes during mitosis (14Tung H.Y. Wang W. Chan C.S. Mol. Cell. Biol. 1995; 15: 6064-6074Crossref PubMed Scopus (87) Google Scholar). Thus, Inh2 seems to function in cell cycle regulation; however, the physiological targets of Inh2 are unknown. The N-terminal IKGI sequence of Inh2 is essential for potent inhibition of monomeric PP1C (15Huang H.-b. Horiuchi A. Watanabe T. Shih S.-R. Tsay H.-J., Li, H.-C. Greengard P. Nairn A.C. J. Biol. Chem. 1999; 274: 7870-7878Abstract Full Text Full Text PDF PubMed Scopus (118) Google Scholar). PP1C associated with regulatory subunits such as the myosin-targeting subunit MYPT1 is insensitive to Inh2, and PP1C is thought to bind either to a targeting subunit or to Inh2, but not to both (16Alessi D. Macdougall L.K. Sola M.M. Ikebe M. Cohen P. Eur. J. Biochem. 1992; 210: 1023-1035Crossref PubMed Scopus (328) Google Scholar, 17Yang J. Hurley T.D. DePaoli-Roach A.A. J. Biol. Chem. 2000; 275: 22635-22644Abstract Full Text Full Text PDF PubMed Scopus (54) Google Scholar, 18Eto M. Wong L. Yazawa M. Brautigan D.L. Cell Motil. Cytoskeleton. 2000; 46: 222-234Crossref PubMed Scopus (41) Google Scholar). The docking sites on PP1C for the IKGI sequence of Inh2 and the (R/K)(V/I)X(F/W) motif of other PP1C regulatory subunits have been mapped adjacent to one another on the back side, opposite the active site, suggesting a competition between Inh2 and regulatory subunits (19Egloff M.P. Johnson D.F. Moorhead G. Cohen P.T. Cohen P. Barford D. EMBO J. 1997; 16: 1876-1887Crossref PubMed Scopus (531) Google Scholar, 20Connor J.H. Frederick D. Huang H.-b. Yang J. Helps N.R. Cohen P.T. Nairn A.C. DePaoli-Roach A. Tatchell K. Shenolikar S. J. Biol. Chem. 2000; 275: 18670-18675Abstract Full Text Full Text PDF PubMed Scopus (52) Google Scholar). Here we report the discovery of binding of Inh2 to the Nek2·PP1 complex and show that this association activates Nek2 kinase. Our results implicate Inh2 as a regulator of the Nek2·PP1C complex and centrosome separation in cells, creating new potential links for coordination of cell cycle signals. Human Inh2 residues 1–197 2The Inh2 constructs used residues 1–197 of a total of 204 to facilitate PCR amplifications by eliminating the GC-rich coding region at the C-terminal end of the protein. The recombinant Inh2-(1–197) protein is indistinguishable in functions from the full-length protein purified from rabbit skeletal muscle and is referred to herein as the recombinant wild-type protein. were cloned into pGBT10 and used to transform yeast strain HF7c, which was sequentially transformed with a randomly primed and size-selected mouse day 9 embryo DNA library cloned into vector pVP16 (Stan Hollenberg, Fred Hutchinson Cancer Center, Seattle, WA). As previously described (21Liu J. Prickett T.D. Elliott E. Meroni G. Brautigan D.L. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 6650-6655Crossref PubMed Scopus (86) Google Scholar), positives were selected by growth on triple drop-out medium (leu −/trp −/his −) and confirmed by expression of β-galactosidase from an alternate promoter. The DNAs isolated from a two-hybrid screen were cloned into theBamHI-EcoRI sites of pGEX4T to produce glutathione S-transferase (GST) fusion proteins in theEscherichia coli BL21 strain. The DNA for human Inh2 residues 1–197 (12Kakinoki Y. Somers J. Brautigan D.L. J. Biol. Chem. 1997; 272: 32308-32314Abstract Full Text Full Text PDF PubMed Scopus (35) Google Scholar) was cloned into the pET30a vector (Novagen) for expression as a fusion protein with an N-terminal hexahistidine plus 44-residue S tag™ (His6S) and into theXbaI site of the pRK7-Myc vector for expression in mammalian cells with an N-terminal Myc epitope tag (sequence EQKLISEEDL) (18Eto M. Wong L. Yazawa M. Brautigan D.L. Cell Motil. Cytoskeleton. 2000; 46: 222-234Crossref PubMed Scopus (41) Google Scholar). The DNA fragment of Inh2 encoding residues 1–118 or 14–197 was prepared by PCR methods and inserted into the pET30 or pRK7-Myc vector. Full-length Nek2 was prepared by reverse transcription-PCR using RNA prepared from HeLa cells and was subcloned into theBamHI-EcoRI sites of the pRK7-HA3mammalian expression vector, which added an N-terminal triple-hemagglutinin (HA) epitope tag (sequence YPYDVPDYA). Proteins were detected by immunoblotting using anti-pan PP1C antibody (Transduction Laboratories), affinity-purified sheep anti-Inh2 antibody (13Leach C. Eto M. Brautigan D.L. J. Cell Sci. 2002; 115: 3739-3745Crossref PubMed Scopus (14) Google Scholar), anti-Myc monoclonal antibody 9E10, and anti-HA monoclonal antibody 12CA5 as previously described (18Eto M. Wong L. Yazawa M. Brautigan D.L. Cell Motil. Cytoskeleton. 2000; 46: 222-234Crossref PubMed Scopus (41) Google Scholar). Rabbit lung extracts were prepared by homogenizing tissue in 50 mm MOPS (pH 7.0), 0.1m NaCl, 1 mm EGTA, 1% Nonidet P-40 (IGEPAL CA-630), 0.4 mm Pefabloc, 1 μg/ml leupeptin, 10 μg/ml lima bean trypsin inhibitor, and 0.2% 2-mercaptoethanol; and the supernatant recovered after centrifugation was used as a source of native PP1C in binding assays. African green monkey kidney cells (COS-7) and human retinal pigment epithelial cells (ARPE19) were maintained in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 10% newborn calf serum (Invitrogen) and in DMEM/nutrient mixture F-12 with 10% fetal bovine serum (Invitrogen), respectively, in a humidified incubator at 37 °C with 5% CO2. For binding and kinase assays with N-terminally triple HA-tagged Nek2 (referred to as HA3-Nek2), each 150-mm dish of COS-7 cells (∼106 cells) was transfected with 8 μg of DNA in 16 μl of NovaFECTORTM (VennNova, LLC, Pompano Beach, FL) for 24 h and then lysed in 2 ml of buffer containing 0.1% IGEPAL CA-630. Aliquots of 0.25 ml were mixed with 5 μg of His6S-Inh2, or 1.0 ml was mixed with 8 μg of monoclonal antibody 12CA5; and bound proteins were recovered by binding to 10 μl of S-proteinTM-agarose (Novagen) or 10 μl of protein-A agarose (Sigma), respectively. Beads were recovered and washed by centrifugation. Nek2 assay used the immunoprecipitate in 20 μl of reaction mixture consisting of 25 mm MOPS (pH 7.2), 10 mm MgCl2, 1 mm dithiothreitol, and 0.4 mm Pefabloc plus 0.2 mg/ml myelin basic protein (MBP) (Sigma) and 0.1 mm [γ-32P]ATP (1 μCi/nmol) as substrates. After incubation at room temperature for 30 min, samples were analyzed by SDS-PAGE and PhosphorImager (AmershamBiosciences) analysis using ImageQuant software. A 100-mm stock plate of COS-7 cells was split 1:8 onto fibronectin-coated 22 × 22-mm coverslips seated in 35-mm tissue culture dishes. Cultures were incubated at 37 °C for at least 6 h before transfection. Each 35-mm culture was transiently transfected with 1 μg of either empty green fluorescent protein vector plasmid or plasmid encoding HA3-Nek2 or N-terminally Myc-tagged Inh2 (referred to as Myc-Inh2) using 3 μl of FuGENE 6 transfection reagent (Roche Molecular Biochemicals) in serum-free DMEM following the manufacturer's instructions. Cultures were allowed to express protein at least overnight (12 h) before fixing as described to preserve microtubule structure. Transfected cells were stained with antibodies directed against either HA and Myc as well as endogenous γ-tubulin. Untransfected cultures of COS-7 cells in 35-mm dishes were prepared, and the growth medium was replaced either with fresh 10% newborn calf serum-supplemented DMEM or with 10 μm Taxol or 20 μm nocodazole solutions prepared in 10% newborn calf serum-supplemented DMEM. Cultures were incubated at 37 °C for 30 min for Taxol treatment and for 60 min for nocodazole treatment before fixation and staining. Cells were stained with antibodies directed against endogenous Inh2, γ-tubulin, and β-tubulin. A stage micrometer was used to calibrate the measurement tool in Openlab software (at ×60, 48 pixels = 10 μm; Improvision, Coventry, United Kingdom). Intercentriolar distances in transfected cells were measured with the calibrated measuring tool directly from contrast-enhanced images of γ-tubulin staining. Only those cells with two clearly distinguished foci of γ-tubulin staining were included in the analyses. Cells were rinsed once with 1.2× PEM (1 m PIPES, 50 mm EGTA, and 20 mm MgCl2) at 37 °C, fixed with methanol at −20 °C for 3 min, rinsed twice again with 1.2× PEM, and permeabilized with 0.1% Triton X-100 in 1.2× PEM for 5 min at room temperature. Cells were rinsed three times with phosphate-buffered saline (PBS) and incubated in 3% bovine serum albumin in PBS blocking solution for 1 h at room temperature. Various primary monoclonal and polyclonal antibodies, including sheep anti-Inh2, anti-Myc (9E10), mouse anti-HA (F7; Santa Cruz Biotechnology), rabbit anti-γ-tubulin (Sigma), and Cy3-conjugated mouse anti-β-tubulin (Sigma) antibodies, were diluted in 3% bovine serum albumin-containing PBS and applied to the coverslips for at least 1 h at room temperature or overnight at 4 °C. Cells were rinsed three times with PBS for 5 min each before staining with the appropriate secondary antibodies, including rhodamine-conjugated goat anti-rabbit, Oregon Green 488-conjugated goat anti-mouse, and Alexa 488-conjugated goat anti-sheep antibodies (Molecular Probes, Inc., Eugene, OR), diluted in 3% bovine serum albumin-containing PBS and 1 μg/ml Hoechst 33342 nuclear stain for 1 h at room temperature. Coverslips were rinsed again three times with PBS as described above and mounted onto glass slides with 10 μl of Vectashield mounting medium (Vector Laboratories, Inc., Burlingame, CA). Background staining was determined by preparing identical coverslips without primary antibody. Images of fixed cells were captured using Openlab software with a Nikon Microphot-SA epifluorescence microscope equipped with a Nikon Plan Apo ×60/1.4 oil immersion objective; filter sets for fluorescein isothiocyanate, Texas Red, and 4,6-diamidino-2-phenylindole fluorophores; and a Hamamatsu Orca C4742-95 digital camera. Raw data images were converted to 8-bit tiff images in Openlab and further processes using Adobe Photoshop Version 5.5. Human phosphatase Inh2 was used as bait in a yeast two-hybrid screen, and five distinct positives from 3 × 106 clones in a mouse embryo library were retrieved. Among the clones were fragments of the known proteins Nek2 and spinophilin/neurabin II. Fig.1 A shows a schematic of the Nek2 structure with an N-terminal kinase domain and a C-terminal domain (CTD). The fragments of Nek2 recovered in the screen are shown asbars, and we note they both lie within the CTD and contain the coiled-coil domain (residues 306–330) plus the PP1C-binding motif, KVHF (residues 383–386). Other epitope-tagged full-length and CTD proteins that were used in this study, both wild-type and Ala-substituted, are shown in Fig. 1 A. A yeast conjugation assay was employed to assess interactions between various forms of Inh2, the Nek2 CTD, and PP1C (Fig. 1 B). Transformation and mating of various yeast MATα andMATa strains were verified on a control double drop-out plate, on which all formed colonies (right panel). On triple drop-out medium, colonies could form only if the specific pair of fusion proteins interacted. Each row is a different protein fused to a DNA-binding domain, and each column is a protein fused to the VP16 transcription activation domain. The entire first row and first column (left panel) showed no growth, as expected for the negative control, which was the PP2A α4 subunit and which did not dimerize or interact with any of the other proteins. In the second row, wild-type Inh2 interacted with the wild-type Nek2 CTD, as expected from the screen, and interacted also with PP1C. The Nek2 CTD (second column) interacted strongly with PP1C, consistent with a previous report (6Helps N.R. Luo X. Barker H.M. Cohen P.T. Biochem. J. 2000; 349: 509-518Crossref PubMed Scopus (163) Google Scholar). Thus, in this assay, the Nek2 CTD interacted with both Inh2 and PP1C. However, if Val383 and Phe386 were substituted with Ala in the VXF motif to give Nek2 CTD(AA), interactions with both PP1C and Inh2 were lost entirely (third column). Likewise, there was no interaction of Nek2 CTD(AA) with any of the other forms of Inh2. Eliminating PP1C association with the VXF motif also completely prevented Inh2 interaction with the Nek2 CTD. The results suggest that Inh2 associated with PP1C, which at the same time was bound to Nek2. We imagine that the yeast version of PP1C, called Glc7, bridged the interaction of Inh2 with the targets recovered in the two-hybrid screen. Different domains of Inh2 were required for interactions with different partners. With PP1C (Fig. 1 B, left panel,fifth column), wild-type Inh2 and Inh2-(1–118), lacking 40% of the protein, interacted equally well, whereas Inh2-(14–197), without the 10IKGI13 sequence for PP1C recognition at the N terminus, interacted weakly. On the other hand, with the Nek2 CTD (second column), wild-type Inh2 interacted strongly, and Inh2-(14–197) interacted weakly, but Inh2-(1–118) did not interact, showing that the C-terminal region of Inh2 is required for this association. We compared the Nek2 CTD with another PP1 regulatory subunit, the N-terminal domain of the muscle glycogen-targeting subunit GM. There was only weak interaction with PP1C and no interactions with any of the forms of Inh2 in this assay. Our conclusions are that Inh2 interaction with the Nek2 CTD requires binding of PP1C and the C-terminal domain of the Inh2. Thus, the Inh2-PP1C interaction is different with monomeric PP1C compared with PP1C bound to the Nek2 CTD or with PP1C bound to GM. We used recombinant GST fusion proteins in pull-down assays to show interactions of the Nek2 CTD with Inh2 and PP1C. The Nek2 CTD and double Ala-substituted Nek2 CTD(AA) were expressed in bacteria as GST fusion proteins attached to glutathione-agarose beads and mixed with a rabbit lung extract as a source of native PP1C and endogenous Inh2. Bound proteins were eluted and analyzed by immunoblotting (Fig.2 A,upper and middle panels). PP1C bound to GST-Nek2 CTD, but not to GST-Nek2 CTD(AA) (Fig. 2 A, upper panel). Inh2 binding exactly followed PP1C binding (Fig.2 A, middle panel). Substitution of only 2 residues in a motif known to interact directly with PP1C completely eliminated binding of Inh2. Equivalent amounts of the GST fusion proteins and GST itself as a negative control were bound to the beads and stained with Coomassie Blue (Fig. 2 A, lower panel). We interpret these results as further evidence for formation of a Nek2 CTD·PP1C·Inh2 trimeric complex, with PP1C acting as a bridge between Nek2 and Inh2. Association of Inh2 with full-length Nek2 was tested by transient transfection and affinity purification. Full-length Nek2 with an N-terminal triple-tandem hemagglutinin epitope tag was transiently expressed in COS-7 cells. The wild-type and Ala-substituted versions of Nek2 (Fig. 1 A) were expressed at similar levels. Samples of these cell extracts and a control extract from untransfected cells were subjected to anti-HA immunoblotting to show the relative expression levels of HA3-Nek2 proteins (Fig. 2 B,upper panel). Human recombinant Inh2 was produced in bacteria as a His6S fusion protein and purified to homogeneity. This His6S-Inh2 protein was bound to S-proteinTM-agarose and distributed into three parallel columns for affinity chromatography with the different cell extracts. The bound proteins from each column were eluted and analyzed by immunoblotting using anti-HA antibody (Fig. 2 B, lower panel). The HA3-tagged wild-type Nek2 protein bound to Inh2, but Ala-substituted HA3-Nek2 did not. Only Nek2 with an intact VXF motif for interaction with PP1C could bind to Inh2. Fig. 2 C shows the indirect interaction of His6S-Inh2 with GST-Nek2 Inh2-binding fragment (IBF). His6S-Inh2 (0.1 μm) was mixed with GST-Nek2 IBF on glutathione-agarose beads. Bound His6S-Inh2 was analyzed by Western blotting. GST-α4 was used as a negative control. His6S-Inh2 failed to associate with GST-Nek2 IBF, whereas the GST-Nek2 IBF·PP1C complex prepared by preincubation with a lung extract coprecipitated with His6S-Inh2 (Fig. 2 C). The results of binding assays and the yeast conjugation assay indicate that Inh2 is capable of interaction with PP1C on Nek2. Nek2 is known to be localized at centrosomes, and the association of Inh2 with Nek2 predicts that Inh2 should appear at centrosomes in living cells. By immunofluorescence microscopy, we found that endogenous Inh2 was concentrated into a cloud in the perinuclear region in COS-7 cells that were fixed under conditions to preserve microtubules (Fig.3 A). Centrosomes were embedded within this cloud of Inh2 and were located by staining with anti-γ-tubulin antibody (inset, yellow arrow). The cloud of Inh2 was centered at the microtubule-organizing center and depended on microtubule organization. Treatment of COS-7 cells briefly with nocodazole to disrupt microtubules dispersed the cloud of endogenous Inh2 and the fibrillar staining for β-tubulin (Fig.3 B, panels c and d). Longer exposure to nocodazole made the Inh2 staining even more diffuse (data not shown). On the other hand, stabilizing microtubules by adding Taxol to the cells enlarged the zone of bright staining for Inh2; and upon closer inspection, the Inh2 staining appeared in part to overlap microtubules emanating from the centrosome (Fig. 3 B,panels a and b). Association of Inh2 with centrosomes was more evident in ARPE19 cells that did not have a cloud of Inh2 (Fig. 3, C and D). Instead, staining of Inh2 was concentrated into discrete foci near the nucleus (Fig.3 C), which exactly colocalized with γ-tubulin as a marker for the location of the centrosomes in double-stained cells (data not shown). These results show that endogenous Inh2 was localized at centrosomes and, in some cells, was concentrated into a cloud surrounding the centrosomes by a microtubule-dependent mechanism. We assayed the effects of added Inh2 on the kinase activity of a Nek2·PP1C complex. In this complex, PP1C suppressed kinase activity by dephosphorylation and inactivation of Nek2. The HA3-Nek2·PP1C complex was prepared by transient transfection and anti-HA immunoprecipitation from COS-7 cells. Nek2 kinase activity was assayed by incorporation of32P from [γ-32P]ATP into the exogenous substrate MBP. Addition of His6S-tagged recombinant Inh2 produced a dose-dependent activation, with EC50∼ 10 nm (Fig. 4). The maximum activity given by addition of 1 μmHis6S-Inh2 was the same as the activity induced by addition of 1 μm microcystin LR, a potent PP1/PP2A inhibitor (data not shown), suggesting the full activation of the HA3-Nek2·PP1C complex induced by Inh2. This effect is consistent with the high affinity binding of His6S-Inh2 to PP1C complexed to HA3-Nek2. We propose that inhibition of PP1C by Inh2 resulted in higher specific activity of Nek2. The anti-HA immunocomplex might be a mixture of HA3-Nek2·PP1C plus HA3-Nek2 unbound to PP1C; therefore, basal activity without His6S-Inh2 is due to free HA3-Nek2. We tested for binding and activation of Nek2·PP1C complexes with 1) Inh2 (residues 1–198), essentially the full-length wild-type protein; 2) C-terminally truncated Inh2-(1–118), which lacks two sites for interaction with PP1C (17Yang J. Hurley T.D. DePaoli-Roach A.A. J. Biol. Chem. 2000; 275: 22635-22644Abstract Full Text Full Text PDF PubMed Scopus (54) Google Scholar, 22Helps N.R. Vergidou C. Gaskell T. Cohen P.T.W. FEBS Lett. 1998; 438: 131-136Crossref PubMed Scopus (14) Google Scholar); and 3) N-terminally truncated Inh2-(14–197), which lacks the IKGI motif (residues 10–13) required for binding to monomeric PP1C (15Huang H.-b. Horiuchi A. Watanabe T. Shih S.-R. Tsay H.-J., Li, H.-C. Greengard P. Nairn A.C. J. Biol. Chem. 1999; 274: 7870-7878Abstract Full Text Full Text PDF PubMed Scopus (118) Google Scholar). These His6S-Inh2 proteins were produced in bacteria, purified to homogeneity, and immobilized on S-proteinTM-agarose for affinity chromatography. An extract of COS-7 cells expressing HA3-Nek2 was split and subjected to affinity chromatography using parallel columns of the different recombinant Inh2 proteins. As shown in Fig.5 A, all three forms of His6S-Inh2 bound both HA3-Nek2 and PP1C from extracts, but the relative recovery of these proteins was quite different. The His6S-tagged wild-type Inh2 protein bound a substantial amount of PP1C relative to its abundance in the extract. By comparison, the His6S-tagged wild-type Inh2 protein bound a much smaller fraction of total HA3-Nek2 from the extract. This hinted that of all the endogenous PP1C that could bind to immobilized Inh2, only a minor pool was associated with HA3-Nek2. Deletion of the IKGI sequence severely reduced binding of PP1C to His6S-Inh2-(14–197) to levels that were barely detectable, as expected. Likewise, only low levels of HA3-Nek2 bound to His6S-Inh2-(14–197), consistent with PP1C participating in the interaction. Compared with these results, we were surprised by the robust recovery of HA3-Nek2 by immobilized His6S-Inh2-(1–118), which yielded much more than even wild-type Inh2. The level of PP1C recovered was lower than with wild-type Inh2, making the ratio of HA3-Nek2 to PP1C very high. This suggests that His6S-Inh2-(1–118) preferentially bound to HA3-Nek2·PP1C complexes compared with either wild-type Inh2 or Inh2-(14–197). We assayed activation of Nek2 kinase in HA3-Nek2·PP1C complexes recovered by immunoprecipitation in response to addition of the different His6S-Inh2 proteins used for affinity purification (Fig. 5 B). Addition of 100 nm His6S-Inh2 in these assays gave a >2-fold increase in Nek2 kinase activity (n = 3;p < 0.001), measured by phosphorylation of MBP as substrate. T
Referência(s)