Nucleolar Nek11 Is a Novel Target of Nek2A in G 1 /S-arrested Cells*

We previously reported that Nek11, a member of the NIMA (never-in-mitosis A) family of kinases, is activated in G 1 /S-arrested cells. We provide herein several lines of evidence for a novel interaction between Nek11 and Nek2A. Both Nek11 and Nek2A, but not Nek2B, were detected at nucleoli, and the Nek2A-specific C-terminal end (amino acids 399–445) was responsible for nucleolar localization. Endogenous Nek11 coimmunoprecipitated with endogenous Nek2A, and non-catalytic regions of each kinase were involved in the complex formation. Nek11L interacted with phosphorylated Nek2A but barely with the kinase-inactive Nek2A (K37R) mutant. In addition, both Nek2A autophosphorylation activity and the Nek11L-Nek2A complex formation increased in G 1 /S-arrested cells. These results indicate that auto- phosphorylation of Nek2A could stimulate its interaction with Nek11L at the nucleolus. Moreover, Nek2 directly phosphorylated Nek11 in the C-terminal non-catalytic region and elevated Nek11 kinase activity. The non-catalytic region of Nek11 showed autoinhibitory activity through intramolecular interaction with its N-terminal catalytic domain. Nek2 dissociated this autoinhibitory interaction. Altogether, our studies dem-onstrate a unique mechanism of Nek11 activation by Nek2A was replaced by as esis kit CA) according to Synthetic primers (5 (cid:2) -CAAGATATTAGTTTGGAGAGAACTTGACTAT-GGC-3 (cid:2) and 5 (cid:2) -GCCATAGTCAAGTTCTCTCCAAACTAATATCTTG-3 (cid:2) ) were used with template pCR/nek2 plasmids. All mutated Nek2A and Nek2B were fully sequenced and termed pCR/nek2 (K37R). These Nek2A/B ORF containing cDNAs were subcloned into pD3HA plasmid pcDNA3-derived plasmid with an HA tag at the C terminus) at the BamHI site to construct a C-terminal HA-tagged Nek2. These plasmids were termed pD3Nek2A-HA, pD3Nek2A (K37R)-HA, pD3Nek2B-HA, pD3Nek2B (K37R)-HA, and pD3Nek2-(1–342)-HA. leucine Nek2B and PP1 non- Nek2A -GGTCGACGCTACTCATGGTATCCAAGG-3 -CGAATTCCTGTTCACCCCAGGACGAGGATGAAGA- G-3 5 -CGAATTCCCTCATGCGTTTCATCTTGGTCCTGGA-3 Nek11-(412–573)). PCR products subcloned into pCR®-Blunt II-TOPO plasmid, and their sequences were verified. EcoRI-digested in-sert fragments were subcloned into pGEX6P-3 or pGEX6P-2 (Amersham Biosciences) at an EcoRI site. Then GST fusion protein-expressing plasmids, termed pGEX-Nek11-(287–337), pGEX-Nek11-(385–467), and pGEX-Nek11-(412–573) obtained.

The NIMA 1 (never-in mitosis A) kinase was first identified in the filamentous fungus Aspergillus nidulans by genetic complementation of the nimA mutation and is essential for nuclear division cycle at G 2 /M transition (1). Protein kinases structurally related to fungal NIMA have been identified in various organisms (2). In the human genome, eleven NIMArelated kinases (Nek1-11) have been reported (3).
In human, Nek2, the most fungal NIMA-related kinase, is the best characterized (4). Two splice variants of Nek2, Nek2A and Nek2B that encode different C termini have been identi-fied (5,6). Non-catalytic C-terminal region of Nek2A but not that of Nek2B has the PP1 binding domain (6). PP1 represses Nek2A autophosphorylation and activation by dephosphorylation (7). Both Nek2A and Nek2B are cell cycle-regulated protein kinases detected at the centrosome (6), and Nek2A overexpression causes premature centrosome splitting (8). Although the role of Nek2B is unclear in somatic cells, Nek2B has an important function in zygotic centrosome assembly and maintenance in Xenopus oocytes (9 -11). The studies on NIMArelated kinases in lower organisms also support a model that the regulatory function in the spindle pole body/microtubules organization center/centrosome is a conserved activity of Nek2like kinases (12,13). Additionally, Nek2 could localize at nucleus in somatic cell lines albeit its physiological significance remains (8,14), and the functional significance of Nek2A specific extra coiled-coil domain at the C-terminal end has not been addressed. Concerning other member of human Neks, most of their functions have been largely unclear; however, recent studies are beginning to reveal their functions. Interesting findings have been reported that Nercc1/Nek9 activates Nek6/Nek7 during mitosis, representing a novel cascade of human NIMA-related kinases (15,16). These findings raise a possibility that diversity of human NIMA-related kinases may compose an unknown NIMA family cascade.
We previously identified new members of the mammalian NIMA family of kinases, termed Nek11L and Nek11S (NIMArelated kinase 11 long and short isoform) and showed activation of Nek11 kinase activity by various DNA-damaging agents and DNA replication inhibitors (17). The transient expression of wild-type Nek11L enhanced the aphidicolin-induced S-phase arrest. Conversely, this S-phase arrest was reduced in the U2OS cell lines expressing kinase-inactive Nek11L (K61R) and these cells were more sensitive to aphidicolin-induced cell lethality (17). Therefore, Nek11 appears to have a role in the G 1 /S-arrested cells, probably in the DNA replication checkpoint, although the molecular mechanism for Nek11 activation was undetermined.
In this study, we investigated the molecular mechanism of Nek11 activation in G 1 /S-arrested cells. Unexpected colocalization of Nek11 and Nek2A was observed at nucleoli, and association of Nek11 with Nek2A was enhanced especially in G 1 /Sarrested cells. Biochemical analysis suggested that Nek11 was regulated by Nek2A. Overall, our observations strongly suggest that Nek11 is a novel target of Nek2A at nucleolus in G 1 /Sarrested cells.

EXPERIMENTAL PROCEDURES
Immunofluorescence Microscopic Analysis-For immunofluorescence microscopic studies, cells were seeded on multichamber slides (1-4 ϫ 10 4 per well on 4-well multichamber slides, Lab-Tek II-CC2 chamber slide system, Nalgen Nunc International, Rochester, NY). The next day, cells were washed once with PBS, fixed with 4% formaldehyde/PBS buffer for 10 min, and treated with Triton buffer (0.5% Triton X-100, TBS, 10% glycerol, 1 mM EDTA) for 10 min at room temperature. The permeabilized cells were washed twice with PBS, and primary antibod-ies (1 g/ml) in dilution buffer (1% bovine serum albumin, PBS, 0.1% Tween-20) were placed in a multichamber under indicated conditions. The slides were then washed with PBS (10 min ϫ 3) and covered with secondary antibodies in dilution buffer (goat anti-rabbit IgG Alexa Fluor 594-conjugated (ϫ2000), goat anti-mouse IgG Alexa Fluor 488conjugated (ϫ2000) (Molecular Probes, Inc., Eugene, OR)) for 1 h in a dark box at room temperature. Alternatively, the anti-HA antibody FITC-conjugated (clone F-7, Santa Cruz Biotechnology Inc., Santa Cruz, CA) was used as a primary antibody. Slides were again washed three times with PBS and mounted with DAPI-containing antifade (Vectashield, Vector Laboratories, Inc., Burlingame, CA). Immunostained cells were analyzed by a confocal laser scanning microscope using a Carl Zeiss LSM 510 system or by a conventional fluorescence microscope (OLYMPUS IX70) equipped with fluorescence digital CCD camera (KEYENCE, VB-6010).
Western Blot Analysis-For preparation of whole cell lysates, cells were lysed in 1ϫ Laemmli SDS sample buffer and sonicated. Other soluble cell extracts were prepared using Nek11 lysis buffer as described above. Sample proteins from an equal number of cells were mixed with 4ϫ Laemmli SDS sample buffer, heat-denatured for 2 min, resolved by SDS-PAGE, and electrophoretically transferred onto Immo-bilon™ polyvinylidene difluoride membranes (Millipore Corp., Bedford, MA). The membranes were blocked in 5% low fat milk or 3% bovine serum albumin/TBS-T (20 mM Tris-HCl, pH 7.5, 137 mM NaCl, 0.1% Tween-20), and incubated with primary antibodies diluted with blocking buffer at 4°C for 15-20 h. After incubation at room temperature for 1 h with secondary antibodies conjugated with horseradish peroxidase, signals were detected by enhanced chemiluminescence using an ECL detection reagent (Amersham Biosciences) or a Western Lightning chemiluminescence reagent Plus (PerkinElmer Life Sciences, Inc., Boston, MA).
In Vitro Dephosphorylation Assay-For Nek2A dephosphorylation, Nek2A (F386A)-HA was coimmunoprecipitated with FLAG-Nek11L using anti-FLAG M2 agarose from 293T cells. Aliquots of immunocomplex beads were incubated at 25°C for 30 min in the presence or absence of phosphatases (PP2A (0.5 units/25 l), PP1 (0.5 units/25 l), and calf intestine alkaline phosphatase (50 units/25 l)). For Nek11L de-phosphorylation, Nek11L-FLAG was coexpressed with Nek2B and immunoprecipitated by anti-FLAG M2 agarose from 293T cells. Aliquots of immunocomplex beads were incubated at 37°C for 30 min in the presence or absence of phosphatases (PP2A and PP1, 0.2 units/20 l each). Dephosphorylation reaction was carried out using an attached phosphatase buffer (Upstate, Charlottesville, VA). Samples were subjected to SDS-PAGE, and electrophoretic mobility changes were examined by Western blotting.
Phosphoamino Acid Analysis-FLAG-Nek11L (K61R) protein was phosphorylated by HA-tagged Nek2 (F386A) or Nek2B in vitro as described above, separated by SDS-PAGE, and transferred to the polyvinylidene difluoride membrane. Corresponding bands were cut out and subjected to acid hydrolysis in 6 N HCl at 110°C for 1 h. Hydrolyzed phosphoamino acids were dried using a Speedvac concentrator, and resolved by pH 1.9 buffer (2.2% formic acid, 7.8% acetic acid). Sample was separated by two-dimensional electrophoresis (pH 1.9 and pH 3.5) on thin layer chromatography plate (MERCK 1.05716.) by the standard method, and results were recorded on x-ray film by autoradiography.
Biochemical Studies of Nek11 C-terminal Non-catalytic Region-For detection of Nek11 homo-oligomerization and intramolecular interaction, FLAG-Nek11-(289 -645) and Myc-Nek11-(1-337)-expressing cells were lysed in Nek11 lysis buffer. Immunoprecipitated complexes by anti-Myc agarose were washed with Nek11 lysis buffer (1 ml ϫ 5) and 50 mM Hepes-NaOH pH 7.5 (1 ml ϫ 1) as described above. To examine the effect of in vitro post-phosphorylation by Nek2 on Nek11 intramolecular interaction, Nek2A (F386A)-HA protein was transiently expressed in 293T cells, immunoprecipitated by anti-HA agarose, and eluted by HA peptide (200 g/ml in 50 mM Hepes-NaOH, pH 7.5). Then aliquots of Myc-agarose immunocomplex beads were subjected to in vitro kinase assay in the presence or absence of Nek2A (F386A)-HA at 37°C for 30 min. Following a wash by Nek11 lysis buffer (1 ml ϫ 3) and by 50 mM Hepes-NaOH pH 7.5 (1 ml ϫ 1), immunocomplex samples were subjected to Western blot analysis. To examine the effect of the C-terminal non-catalytic region of Nek11 on its kinase activity, equal amounts of FLAG-Nek11-(1-337)-beads were preincubated with GST-Nek11-(412-573) or control GST proteins (1-12 g/10 l) for 5 min at room temperature. Subsequently 2 ϫ kinase buffer (10 l) containing histone H2A substrate (4 g) was added to perform kinase reaction at 30°C for 15 min followed by SDS-PAGE. Quantitative analysis was performed as described above.
Cell Culture and Miscellaneous Materials-Human embryonic kidney transformed fibroblast HEK293T cells, cervical epithelioid carcinoma HeLaS3 cells, and osteosarcoma U2OS cells were grown in Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% heat-inactivated fetal bovine serum (Sigma) and 5 g/ml gentamicin (Invitrogen Life Technologies). For cell cycle block, cells were treated with aphidicolin (2 g/ml) and hydroxyurea (2 mM) or by nocodazole (1 M) and taxol (1 g/ml) for 18 h to arrest at the G 1 /S phase or at the G 2 /M phase. Anti-FLAG M2 agarose, anti-HA agarose, and anti-FLAG M2 antibodies were purchased from Sigma. Anti-Nucelophosmin/B23 antibody was from Zymed Laboratory Inc. (South San Francisco, CA), and anti-Myc antibody (9E10) from Oncogene Research Products (San Diego, CA). Anti-Nek2 antibody was from Transduction Laboratories (Lexington, KY). Anti-GFP, anti-HA polyclonal (Y-11), and anti-Myc agarose antibodies were purchased from Santa Cruz Biotechnology, Inc., and anti-GST antibody was from Amersham Biosciences. Affinitypurified anti-Nek11 polyclonal antibody was prepared as described (17). PP2A and PP1 were obtained from Upstate (Charlottesville, VA), and calf intestine alkaline phosphatase was from TaKaRa. Nocodazole, aphidicolin, and hydroxyurea were purchased from Sigma and okadaic acid was from Wako Chemicals. Histone H2A was from Roche Applied Science.

Association and Colocalization of Endogenous Nek11 with
Nek2A at Nucleolus-We observed enrichment of Nek11 at the nuclear granular structure in the indirect immunofluorescent analysis using anti-Nek11 antibody and U2OS cells. These nuclear small round granular structures appeared to be nucleoli, and we thus compared subnuclear localization of Nek11 with a typical nucleolar protein Nucleophosmin (NPM)/B23 (20). Most Nek11 colocalized with NPM at interphase, telophase, and M/G 1 transition phase, whereas mitotic Nek11 was detected at perichromosome, and colocalization was partial during metaphase (Fig. 1A). These dynamics of Nek11 resembled nucleolar component disassembly/reassembly during cell cycle (21), and thus we concluded that Nek11 is a novel nucleolar protein.
Another well studied NIMA family kinase Nek2 is shown to localize at the centrosome (8). In our preliminary experiments, however, overexpressed exogenous Nek11 and Nek2 were abundant at cytoplasm (probably due to the leptomycin Bsensitive nuclear export system) and some of them colocalized at the pericentrosomal aggresome-like structure in 293T cells (data not shown). Since a recent study reported a NIMA family cascade consisting of Nercc1/Nek9 and Nek6/7 (16), we analogously hypothesized interaction/colocalization of endogenous Nek11 with Nek2. We surveyed subnuclear localization of endogenous Nek2 protein using a confocal laser scanning microscope. Consistent with the previous report, Nek2 protein was detected at the centrosome in both U2OS and HeLaS3 cells (Fig. 1B, arrowheads). Moreover, Nek2 signals also showed colocalization with nucleolar Nek11 in both cell lines (Fig. 1B,  arrows). These results indicated that some of the endogenous Nek2 protein could colocalize with nucleolar Nek11. We next tested association of Nek11 with Nek2 by coimmunoprecipitation assay. Endogenous Nek11 was immunoprecipitated from U2OS and HeLaS3 cells, and we detected coimmunoprecipitation of Nek2A but not of Nek2B in both cell lines (Fig. 1C). Coimmunoprecipitation of Nek2A with Nek11 was only observed in the nuclear fraction (H), but not in the cytoplasmic fraction (L) (Fig. 1D). These results suggested that endogenous Nek11 could associate with some Nek2A at the nucleolus.
Specific Nucleolar Targeting of Nek2A Mediated by its Cterminal End Region-We next addressed why endogenous Nek11 coimmunoprecipitated with endogenous Nek2A, but not with Nek2B. As the anti-Nek2 antibody we used here could not distinguish Nek2A and Nek2B, we introduced HA-tagged Nek2A-and Nek2B-expressing plasmids into U2OS cells, and subnuclear localization of each kinase was compared by a confocal laser scanning microscope. Consistent with the results above, nuclear Nek2B did not colocalize with nucleolar Nek11 in interphase cells whereas some Nek2A merged with nucleolar Nek11 (Fig. 2A, left panels). Since Nek2A has an additional coiled-coil domain at the C-terminal end, we introduced a plasmid expressing GFP-fused Nek2A specific coiled-coil domain (amino acids 399 -445) into U2OS cells (Fig. 2B). Using a conventional fluorescence microscope, GFP-Nek2A-(399 -445) protein showed nuclear (94%, n ϭ 122) and nucleolus-like (89%, n ϭ 122) enrichment in living U2OS cells, but control GFP protein showed diffused distribution (Fig. 2C, left panels). Furthermore, analysis of fixed and permeabilized cells using a confocal laser scanning microscope confirmed that GFP-Nek2A-(399 -445) protein indeed colocalized with nucleolar Nek11 whereas GFP-Nek2B-(319 -384) did not (Fig. 2C, right  panels, arrows). Overall, these experiments demonstrated for the first time that Nek2A, but not Nek2B, has a unique nucleolar targeting/retention activity through its C-terminal end coiled-coil domain (amino acids 399 -445). In contrast, mitotic Nek2A and Nek2B proteins were detected on the centrosome and disengaged from perichromosomal Nek11 at metaphase ( Fig. 2A, right panels), suggesting that the Nek11-Nek2A complex dissociates during mitosis.
Nek11 Association With and Activation by Nek2-Endogenous Nek11 and Nek2A formed a complex in vivo, so we next addressed the interaction domain of each kinase. Transient overexpression assay is convenient to estimate the interaction FIG. 1. Nucleolar association of Nek11 with Nek2. A, subcellular localization of Nek11 during cell cycle. U2OS cells were fixed with 4% formalin/PBS and treated with 0.5% Triton X-100, TBS, 10% glycerol, and primary antibodies (anti-Nek11 polyclonal and anti-NPM monoclonal antibodies, 1 g/ml each) were reacted for 18 h at 4°C. Nek11 was visualized with anti-rabbit-IgG Alexa Fluor 594-conjugated (red), and Nucleophosmin (NPM) with anti-mouse IgG Alexa Fluor 488-conjugated (green) for 1 h of incubation at room temperature. DNA was costained by DAPI (1 g/ml). Indirect immunofluorescent analysis was performed with a confocal laser scanning microscope LSM510 system. B, detection of endogenous Nek2 at nucleoli. Asynchronous U2OS and HeLaS3 cells were fixed and permeabilized as in A. Endogenous Nek11 (red) and Nek2 (green) were probed with anti-Nek11 polyclonal and anti-Nek2 antibodies (1 g/ml each) for 4 -18 h at 4°C. DNA was stained with DAPI, and analyzed using a confocal laser scanning microscope LSM510 system. White arrows indicate nucleolar staining, and arrowheads indicate centrosomal staining. C, endogenous Nek11 coimmunoprecipitated with Nek2A. Endogenous Nek11 was immunoprecipitated (IP) with anti-Nek11 polyclonal antibody (anti-Nek11) from U2OS and HeLaS3 cells. Coimmunoprecipitated Nek2A was detected by Western blot analysis using anti-Nek2 monoclonal antibody. As a negative control, normal rabbit IgG (Control Ig) was used for immunoprecipitation. D, Nek11-Nek2A complex recovered from nuclear fraction. HeLaS3 cells were primarily extracted by low salt lysis buffer (0.1 M NaCl, 0.1% Nonidet P-40) (cytoplasmic fraction, L), and cell pellets were secondarily extracted by high salt lysis buffer (0.42 M NaCl, 0.1% Nonidet P-40) (nuclear fraction, H). Nek11 protein was immunoprecipitated by anti-Nek11 polyclonal antibody from each extract, and coimmunoprecipitated Nek2A protein was detected by Western blot analysis using anti-Nek2 antibody. Left panels show the immunoprecipitated Nek11 and Nek2, and right panels show the expression level of Nek11 and Nek2 proteins in each extract. domain of each kinase in vivo, although artificial nonphysiological interaction between Nek2B and Nek11 could be detected (probably due to their cytoplasmic abundance). FLAGtagged Nek11L protein was coexpressed with various HA-tagged Nek2 proteins (wild-type Nek2A/B, kinase-inactive Nek2A/B (K37R), PP1 binding motif lacking Nek2 (F386A) mutant, leucine zipper-deleted Nek2BdLZ (lacking amino acids 306 -334), and C-terminal coiled-coil-deleted Nek2-(1-342)) in 293T cells. Immunoprecipitation-Western blot analysis was carried out to determine the region of Nek2 involved in inter-action with Nek11. The results showed that PP1 binding motif lacking Nek2A (F386A) and wild-type Nek2B formed a tight complex with Nek11L, but others did not (Fig. 3A). In addition, in vitro dephosphorylation assay showed that PP2A, PP1, and calf intestine alkaline phosphatase treatments caused rapid electrophoretic migration of Nek11-associated Nek2A (F386A)-HA (Fig. 3B), indicating that Nek11-bound Nek2A (F386A) was phosphorylated. Previous studies have shown that binding of PP1, deletion of the leucine zipper region, and inactivation of mutation (K37R) suppress kinase activity and au-
Association of Nek11 with Nek2 Increased in G 1 /S-arrested Cells-Nek11 is activated in G 1 /S-arrested cells (17), and Nek2A protein is increased in these cells (18). Therefore, we examined Nek11-Nek2 interaction during cell cycle by coimmunoprecipitation analysis. Endogenous Nek11 protein was immunoprecipitated from HeLaS3 cells chemically synchronized at G 1 /S or G 2 /M phase, and we found that coimmunoprecipitation of endogenous Nek2A increased especially in G 1 /S-arrested cells (Fig. 4A, left panels). In addition, autophosphorylation activity of Nek2A-HA was stimulated in aphidicolin-treated 293T cells (Fig. 4B). Consistent with these data, association of exogenous Nek2A-HA with Nek11L-FLAG increased in G 1 /S-arrested 293T cells (Fig. 4C, middle panels). As association of Nek2A (F386A)-HA mutant with Nek11L-FLAG also increased in G 1 /S-arrested cells (Fig. 4C, middle panels), G 1 /S-arrest-induced increase of Nek2A-Nek11 interaction appeared to be independent of PP1 binding with Nek2A. These results suggested that Nek2A was activated through a PP1independent mechanism to increase its association with Nek11 in G 1 /S-arrested cells.
Modulation of Nek11 C-terminal Non-catalytic Function by Nek2-In our studies above, Nek2-HA coexpression caused slower migration of some Nek11L-FLAG on SDS-PAGE (as seen in Fig. 3, A and C), suggesting post-translational modification of Nek11 by Nek2. Thus, we examined whether Nek11 could be a phosphorylation substrate for Nek2. In vitro incubation of Nek11L-FLAG with Nek2B-HA in a kinase reaction buffer induced electrophoretic mobility shift of Nek11L (Fig.  5A). The band shift of Nek11L was reduced by in vitro phosphatase treatment, especially by PP2A but not by PP1 (Fig.  5B), indicating that Nek11L could be phosphorylated by Nek2 in vitro and in vivo. In vitro immunocomplex kinase assay using [␥-32 P]ATP showed that Nek2 phosphorylated both FLAG-Nek11L (K61R) and a non-catalytic region containing FLAG-Nek11-(285-645), but FLAG-Nek11 (K61R, 1-337) to a lesser extent (Fig. 5C). GST-Nek11-(287-337) and GST-Nek11-(412-573) but not GST also served as a good substrate for Nek2, while phosphorylation of GST-Nek11-(385-467) was low (Fig. 5D). Phosphoamino acid analysis of the FLAG-Nek11L (K61R) phosphorylated by Nek2A (F386A)-HA in vitro detected predominantly phosphoserine with traces of phosphothreonine, but no phosphotyrosine (Fig. 5E). These data indicated that the non-catalytic region of Nek11L was phosphorylated by Nek2 on multiple serine residues, at least within amino acid positions 287-337 and 468 -573 in vitro.
Non-catalytic regions of many kinases have autoregulatory functions. Actually, in vitro kinase assay indicated that FLAG-Nek11-(1-337) kinase activity was strongly suppressed by an addition of GST-Nek11-(412-573) but not by GST (Fig. 6A). Further, we found that Myc-Nek11-(1-337) coimmunoprecipitated with FLAG-Nek11-(289 -645) from asynchronous cells (Fig. 6B). Coimmunoprecipitation assay, however, showed that FLAG-Nek11L and Myc-Nek11L oligomer was not detected in asynchronous cells despite our being able to detect it in nocodazole-arrested M phase cells (Fig. 6C). These results indicated that C-terminal non-catalytic region of Nek11 would directly interact with its N-terminal catalytic domain during interphase probably in an intramolecular manner as an autorepressor domain. Importantly, Nek2-mediated phosphorylation in vitro caused a dissociation of FLAG-Nek11-(289 -645) from Myc-Nek11-(1-337) (Fig. 6D), indicating that Nek2 could disrupt autoinhibitory intramolecular interaction within Nek11. These experiments provided evidence that the C-terminal noncatalytic region of Nek11 has an autorepressive function, and that Nek2-association and/or Nek2-mediated phosphorylation would antagonize this autoinhibitory activity. Overall, these results strongly suggested that Nek2A could activate nucleolar Nek11 by modulating the non-catalytic region of Nek11 (summarized in Fig. 7).

DISCUSSION
In this study, we demonstrated for the first time several lines of evidence indicating a novel interaction between Nek11 and Nek2 in G 1 /S-arrested cells. Both Nek11 and Nek2 could localize at nucleolus, and Nek11-Nek2 complex formation increased in G 1 /S-arrested cells. In addition, Nek2 activated Nek11 through modulation of the autorepressive function of the Nek11 C-terminal non-catalytic region. These observations strongly pointed to the likelihood that nucleolar Nek11 is a novel target of Nek2A in G 1 /S-arrested cells.
In the indirect immunofluorescent analysis, cell preparation protocol and primary antibody are important factors affecting antibody accessibility and reactivity, and experimental protocols also should be adjusted to each cell type. Nek2A and Nek2B localize at centrosome through the cell cycle (6), and our indirect immunofluorescent analysis additionally showed unexpected nucleolar colocalization of endogenous Nek11 and Nek2. Importantly, we here discovered a novel difference between Nek2A and Nek2B. Nek2A has a nucleolar targeting/ retention activity via Nek2A specific coiled-coil domain at the C-terminal end, whereas Nek2B does not localize at nucleolus. Nek2A specific C-terminal end coiled-coil domain (amino acids 399 -445) contains a lysine/arginine stretch that may contribute to nuclear/nucleolar targeting. Consistent with data from our indirect immunofluorescent analysis, endogenous nucleolar Nek11 associated only with Nek2A, and endogenous Nek11-Nek2A complex was detected only in the nuclear fraction but not the cytosolic fraction. Although overexpressed ectopic Nek11 and Nek2B could interact with each other, probably due to their abundance at the cytoplasm, our observations suggest that nucleolar localization is an important factor for endogenous Nek11-Nek2A complex formation.
In addition, Nek11 formed a complex preferentially with PP1 binding motif-disrupted Nek2A (F386A) but hardly with wildtype Nek2A, suggesting that PP1 might repress the Nek11-Nek2A interaction in cells. Alternatively, as PP1-mediated dephosphorylation inactivates Nek2A (7,22) and kinase-inactive Nek2B (K37R) hardly formed a complex with Nek11 in cells, activation of Nek2A by autophosphorylation (or upstream regulator) might be required before its interaction with Nek11. The Mos-MAPK-p90RSK pathway activates Nek2 during meiosis in mouse pachytene spermatocytes (23). However, little is known concerning the endogenous Nek2A activation mecha- FIG. 4. Complex formation between Nek11 and Nek2A in G 1 /S-arrested cells. A, HeLaS3 cells were synchronized at G 1 /S phase by aphidicolin (APH, 2 g/ml) and hydroxyurea (HU, 2 mM) or at M phase by nocodazole (Noc, 1 M) and taxol (1 g/ml) for 18 h. Endogenous Nek11 was immunoprecipitated (IP) and coimmunoprecipitated endogenous Nek2 was detected by Western blot analysis (WB) using anti-Nek2 antibody (left panels). The amounts of immunoprecipitated endogenous Nek11 were also confirmed with anti-Nek11 antibody (lower panel). The expression levels of endogenous Nek11 and Nek2 proteins were confirmed by Western blot analysis using cell lysates (right panels). B, autophosphorylation activity of Nek2A in G 1 /S-arrested cells. Nek2A-HA transfected 293T cells were arrested at G 1 /S phase by aphidicolin (APH, 2 g/ml) or hydroxyurea (HU, 2 mM) for 16 h, and Nek2A-HA was immunoprecipitated (IP) by anti-HA agarose. The amounts of immunoprecipitated Nek2A-HA protein were examined by Western blot analysis (WB) using anti-HA antibody (upper panel) and autophosphorylation activity of Nek2A-HA was analyzed by in vitro kinase assay (middle panel). Lower graph shows the result from in vitro kinase assay performed in triplicate. C, association between Nek11L-FLAG and Nek2A-HA increased in G 1 /S-arrested cells. Nek2A-HA and Nek2A (F386A)-HA were coexpressed with Nek11L-FLAG in 293T cells, and transfected 293T cells were treated with aphidicolin (2 g/ml, APH) for 16 h. Nek2A-HA, Nek2A (F386A)-HA and Nek11L-FLAG were immunoprecipitated (IP) by anti-HA or anti-FLAG agarose from each cell lysate aliquot. The amounts of Nek2As immunoprecipitated by anti-HA-agarose, and the expression level of Nek2As in cell lysates were confirmed by Western blot analysis (WB). Autophosphorylation activities of anti-HA agarose-immunoprecipitated Nek2A-HA and Nek2A (F386A)-HA were examined by in vitro kinase assay. The amounts of Nek11L-FLAG immunoprecipitated by anti-FLAG agarose and coimmunoprecipitated Nek2A-HA were confirmed by Western blot analysis (WB), and kinase activity of immunoprecipitated Nek11L-FLAG was examined by in vitro kinase assay using histone H2A as a substrate. The amounts of coimmunoprecipitated Nek2A-HA were recorded on x-ray films with short and long exposure times. nism in human somatic cells. Because endogenous Nek2A protein levels and its autophosphorylation activity increased in G 1 /S-arrested cells, post-translational regulation of Nek2A would be upstream of the Nek11 activation in these cells. Further exploration of the Nek2A activation mechanism might provide a clue to understand nucleolar regulation of the Nek2A-Nek11 complex in G 1 /S-arrested cells.
We showed here that the C-terminal non-catalytic region of Nek11L could associate with its N-terminal catalytic domain most probably in an intramolecular manner. Nek2 could phosphorylate the Nek11 C-terminal non-catalytic region and antagonize its autoinhibitory function, which would cause Nek11 activation. Nek2A also autophosphorylates its C-terminal noncatalytic region, which appears to be important for its kinase FIG. 5. Nek2-mediated phosphorylation on Nek11 non-catalytic region. A, Nek2-mediated in vitro phosphorylation caused Nek11L electrophoretic mobility shift. Nek11L (K61R)-FLAG protein was immunoprecipitated and recovered by competitive peptide elution. Nek2B-HA was immunoprecipitated by anti-HA agarose from 293T cells as enzyme source. Nek11L (K61R)-FLAG protein was subjected to in vitro kinase assay in the presence or absence of Nek2B-HA-beads, and Western blotting using anti-Nek11 antibody confirmed electrophoretic mobility of Nek11L (K61R)-FLAG. Nek2B-HA used in this assay was also confirmed by Western blotting using anti-HA antibody. B, electrophoretic mobility change of Nek11 by PP2A treatment. Nek11L-FLAG was coexpressed with Nek2B-HA and immunoprecipitated (IP) by anti-FLAG agarose. Aliquots of immunocomplex were treated by either control buffer, PP2A or PP1. Subsequent Western blotting (WB) showed electrophoretic mobility of Nek11L-FLAG using anti-FLAG antibody. Coimmunoprecipitated Nek2B-HA was also confirmed with anti-HA antibody. C, Nek2 phosphorylated Nek11L in vitro. Nek2A (F386A)-HA immunocomplex beads recovered from 293T cells were used as enzyme source. FLAG-Nek11 proteins recovered by immunoprecipitation from 293T cells were used as substrates for in vitro kinase assay. Left and right panels show the result of autoradiography and Coomassie Blue staining of the gel. D, Nek2-mediated phosphorylation on Nek11 C-terminal region. GST-Nek11-(287-337), GST-Nek11-(385-467), GST-Nek11-(412-573), and GST (4 g/20 l each) were subjected to in vitro kinase assay using Nek2A (F386A)-HA immunocomplex. Upper and lower panels show the result of autoradiography and Coomassie Blue staining of the gel. E, phosphoamino acid analysis of phosphorylated Nek11L. FLAG-Nek11L (K61R) was phosphorylated by Nek2A (F386A)-HA in vitro as described above, separated by SDS-PAGE, blotted onto polyvinylidene difluoride membrane, and subjected to acid hydrolysis followed by two-dimensional separation on thin layer chromatography plate. The positions of phosphoamino acids are illustrated at left. activity on exogenous substrate (19), suggesting that the Nek11 C-terminal non-catalytic region might contribute to the Nek11 substrate recognition mechanism. Similar autoregulatory mechanisms by non-catalytic domain have been shown in several protein kinases such as Plk, Hsl1, and Chk1 (24 -26).
Considering these previous studies and our results, we presume that Nek11 kinase is kept latent by a non-catalytic regulatory region in interphase and activated by Nek2A through modulation of this region (as in Fig. 7). Although we did not identify phosphoacceptor serine residues on Nek11, additional experiments indicated that single alanine substitutions at Ser-97, -161, -185, -200, -299 -334, -372, and -380 did not affect Nek2-mediated Nek11 activation (data not shown). Since Nek11 seemed to be phosphorylated by Nek2 at multiple sites, we speculate that multiple phosphorylation on the Nek11 noncatalytic regulatory region might be required to neutralize its autoinhibitory activity or would change the total electric charge, which might affect its substrate recognition process.
Our data here suggest a novel possibility that Nek2A and Nek11 have unknown roles at nucleolus. Interestingly, two hybrid-based large scale comprehensive analysis of proteinprotein interactions in Saccharomyces cerevisiae showed that the yeast NIMA-related kinase Kin3 could interact with some proteins including nucleolar GTP-binding protein Nog1 and mitotic GTP-binding protein Tem1 (27) (also shown in the yeast protein-protein interaction data bases provided by The GRID and Yeast Resource Center). Nucleolar Nog1 has a role in ribosome biogenesis (28,29), and the protein interaction data base above also suggest that Nog1 could interact with MCM and Yph1/Nop7, both of which have abilities to interact with DNA replication origin recognition complex protein (30). In addition, that two-hybrid study (27) also showed that both Nog1 and Tem1 could interact with Fob1, a DNA replication fork blocking nucleolar protein (31), and that Fob1 could interact with Rad53 (27), a DNA damage checkpoint kinase associated with the DNA replication mechanism (32-34). These col- lective data suggest to us a possibility that the Kin3-Nog1/ Tem1-Fob1-Rad53 complex might be involved in unknown signaling between the DNA replication checkpoint mechanism and ribosome biogenesis at nucleolus. As mammalian nucleolar NIMA-related kinases Nek11 and Nek2A are activated by DNA replication inhibitor-induced G 1 /S-arrest, nucleolar function of NIMA-related kinase might be conserved between yeast and human. We have a working hypothesis that the Nek11-Nek2A complex might be associated with an unknown signaling pathway for synchronization between ribosome biogenesis and the DNA replication checkpoint mechanism.
Alternatively, yeast Tem1 activates Cdc14 release from nucleolus to trigger mitotic exit network (35), and Fob1 could interact with Spo12, a component of the Cdc14 early anaphase release (FEAR) network (36). If Kin3 actually interacts with Tem1, Kin3 might be involved in the Tem1-Fob1-Spo12 complex-mediated Cdc14 regulation during mitosis. Attractively, in human, both Nek2A and hCdc14A are centrosomal components and important for centrosome separation and chromosome segregation (8,(37)(38)(39). The possibility of interplay between Nek2A and hCdc14A was also discussed earlier (39). Another human Cdc14 isoform, hCdc14B is a nucleolar protein phosphatase and negatively regulates a mitotic phosphoprotein SIRT2 in an indirect manner (38 -40). Both hCdc14B and SIRT2 show perichromosomal localization during mitosis similar to Nek11, and it may be interesting to explore the functional interaction between Nek11 and hCdc14B/SIRT2.
Collectively, we demonstrated that Nek11 is a nucleolar NIMA-related kinase regulated by Nek2A. We are currently searching for a target of Nek11 among nucleolar components involved in both cell cycle progression and ribosome biogenesis. Future studies could provide significant information concerning unknown cell cycle-related functions of human NIMA-related kinases.