Phosphorylation of the Mitotic Regulator Protein Hec1 by Nek2 Kinase Is Essential for Faithful Chromosome Segregation*

Hec1 (highly expressed incancer) plays essential roles in chromosome segregation by interacting through its coiled-coil domains with several proteins that modulate the G2/M phase. Hec1 localizes to kinetochores, and its inactivation either by genetic deletion or antibody neutralization leads to severe and lethal chromosomal segregation errors, indicating that Hec1 plays a critical role in chromosome segregation. The mechanisms by which Hec1 is regulated, however, are not known. Here we show that human Hec1 is a serine phosphoprotein and that it binds specifically to the mitotic regulatory kinase Nek2 during G2/M. Nek2 phosphorylates Hec1 on serine residue 165, both in vitro and in vivo. Yeast cells are viable without scNek2/Kin3, a close structural homolog of Nek2 that binds to both human and yeast Hec1. When the same yeasts carry an scNek2/Kin3 (D55G) or Nek2 (E38G) mutation to mimic a similar temperature-sensitive nima mutation inAspergillus, their growth is arrested at the nonpermissive temperature, because the scNek2/Kin3 (D55G) mutant binds to Hec1 but fails to phosphorylate it. Whereas wild-type human Hec1 rescues lethality resulting from deletion of Hec1 in Saccharomyces cerevesiae, a human Hec1 mutant or yeast Hec1 mutant changing Ser165 to Ala or yeast Hec1 mutant changing Ser201 to Ala does not. Mutations changing the same Ser residues to Glu, to mimic the negative charge created by phosphorylation, partially rescue lethality but result in a high incidence of errors in chromosomal segregation. These results suggest that cell cycle-regulated serine phosphorylation of Hec1 by Nek2 is essential for faithful chromosome segregation.

Hec1 (highly expressed in cancer) plays essential roles in chromosome segregation by interacting through its coiled-coil domains with several proteins that modulate the G 2 /M phase. Hec1 localizes to kinetochores, and its inactivation either by genetic deletion or antibody neutralization leads to severe and lethal chromosomal segregation errors, indicating that Hec1 plays a critical role in chromosome segregation. The mechanisms by which Hec1 is regulated, however, are not known. Here we show that human Hec1 is a serine phosphoprotein and that it binds specifically to the mitotic regulatory kinase Nek2 during G 2 /M. Nek2 phosphorylates Hec1 on serine residue 165, both in vitro and in vivo. Yeast cells are viable without scNek2/Kin3, a close structural homolog of Nek2 that binds to both human and yeast Hec1. When the same yeasts carry an scNek2/Kin3 (D55G) or Nek2 (E38G) mutation to mimic a similar temperature-sensitive nima mutation in Aspergillus, their growth is arrested at the nonpermissive temperature, because the scNek2/Kin3 (D55G) mutant binds to Hec1 but fails to phosphorylate it. Whereas wild-type human Hec1 rescues lethality resulting from deletion of Hec1 in Saccharomyces cerevesiae, a human Hec1 mutant or yeast Hec1 mutant changing Ser 165 to Ala or yeast Hec1 mutant changing Ser 201 to Ala does not. Mutations changing the same Ser residues to Glu, to mimic the negative charge created by phosphorylation, partially rescue lethality but result in a high incidence of errors in chromosomal segregation. These results suggest that cell cycle-regulated serine phosphorylation of Hec1 by Nek2 is essential for faithful chromosome segregation.
Mitosis must be precisely regulated and checked for faithful partitioning of chromosomes during a short but crucial period of the cell division cycle. Since the basic mechanics of chromosome segregation and checkpoint control are conserved in all eukaryotes, yeast and other fungi are excellent tools for dissecting mechanisms that apply to mammalian cell cycle proteins with clear homologs in lower eukaryotes. The novel coiled-coil protein Hec1 plays important roles in chromosome condensation and cohesion by interacting with structural com-ponents of the mitotic chromosome, including Smc1 (structure of the mitotic chromosome 1) complexes (1)(2)(3) and the kinetochore protein Ctf19p (4). Hec1 also regulates 26 S proteasome activity through interaction with MSS1 (CIM5/subunit 7), p45/ Trip1 (Sug1/CIM3/subunit 8), and p44.5 (subunit 9) (5). These Hec1-interacting proteins were first identified in yeast twohybrid assays using the coiled-coil region of Hec1 as bait (1). The mechanisms by which Hec1 itself is regulated during G 2 and M phases are not yet known, however. Since Hec1 has a structural and functional homolog in S. cerevesiae (scHec1, also known as TID3, NDC80, and YIO4) (3,6), the consequences of its interactions with other proteins can be meaningfully explored in yeast.
Protein kinases have been shown to play important roles regulating G 2 /M phase progression. One such kinase, NimA (never in mitosis A), which phosphorylates specific proteins on serine and threonine residues, is vital in Aspergillus nidulans for entry into mitosis (7)(8)(9). Cells harboring temperature-sensitive mutations of nima arrest specifically in G 2 at nonpermissive temperature, but rapidly and synchronously enter mitosis upon shift to permissive temperature (10). The expression of NimA is tightly regulated during the nuclear division cycle, peaking in G 2 and M phases. NimA, like Hec1, has also been shown to influence faithful chromosome segregation (7)(8)(9).
Kinases with structural and functional homology to NimA exist in vertebrate cells (11)(12)(13)(14)(15)(16)(17)(18). These NimA-related kinases, or Neks, are purported to complement or function in manners similar to those of Cdc2 and other G 2 /M phase cyclin-dependent kinases. Nek2, the homolog in human cells with the greatest structural similarity to NimA within the catalytic domain (19), is regulated in yeast in a manner similar to regulation of NimA in A. nidulans; its expression and serine/threonine kinase activity are highest during late G 2 phase, when Nek2 is expected to function critically (7,17,20). Furthermore, a portion of Nek2 localizes to centrosomes and appears in mammalian cells to play a role similar to NimA in controlling entry into mitosis (17,21). Nek2 may have more diverse roles during several phases of the cell cycle, from S phase to multiple phases of mitosis, based on its dynamic expression and subcellular localization in cytosol, nucleus, and chromosome portions other than the centrosome (21,22). Nek2 may therefore have several functions in regulating cell proliferation, not only a role in G 2 /M phase progression. In budding yeast, a structural homolog of NimA and Nek2, scNek2/Kin3 (also known as Fun52 and Npk1), has been identified (23)(24)(25), but its functional similarity to human Nek2 has not yet been established.
In this report, we demonstrate how the function of Hec1 is regulated during G 2 /M phases by characterizing the previously reported interaction between Hec1 and Nek2 (5). We take advantage of Nek2 and Hec1 homologs, specific point mutants, and the ability to assay chromosomal segregation errors in budding yeast to show that Hec1 is phosphorylated by Nek2 kinase during G 2 and M phases. This specific modification is vital for Hec1 to coordinate faithful chromosome segregation.

EXPERIMENTAL PROCEDURES
Cell Culture and Synchronization-Human bladder carcinoma T24 cells (American Type Tissue Collection, Manassas, VA) grown in Dulbecco's modified Eagle's medium plus 10% fetal bovine serum were synchronized at G 1 by density arrest and then released at time zero by replating in Dulbecco's modified Eagle's medium plus 10% fetal calf serum at a density of 2 ϫ 10 6 cells per 10-cm plate. At various time points thereafter (18 h for G 1 /S, 22 h for S, 32 h for G 2 ), the cells were harvested. To obtain a cell population enriched in M phase, nocodazole (0.4 g/ml) was added to the culture medium for 8 h prior to harvest (26).
Yeast Strains, Reagents, and Media-Yeast strains are described in Table I. Strains used in this study were grown in complete medium (YPD; 1% yeast extract, 2% peptone, and 2% dextrose) or in supplemented minimal medium with appropriate amino acids missing. The chemicals and medium components were purchased from Sigma and BD Industries (Franklin Lakes, NJ).
Immunoprecipitation and Western Blot Analysis-T24 cells resuspended in ice-cold Lysis 250 buffer (50 mM Tris-HCl, pH 7.4, 250 mM NaCl, 5 mM EDTA, 0.1% Nonidet P-40, 50 mM NaF, 1 mM phenylmethylsulfonyl fluoride) were subjected to three freeze/thaw cycles (liquid nitrogen/37°C) and then centrifuged at 14,000 rpm for 2 min at room temperature. The supernatants were used for immunoprecipitation as described (27). Briefly, anti-Hec1 antibody mAb 1 9G3 (1 g) or mouse polyclonal anti-Nek2 antiserum (1 l) was added to each supernatant. After a 1-h incubation, protein A-Sepharose beads were added, and incubation continued for another 1 h. Beads were collected, washed five times with lysis buffer containing hypertonic NaCl, and then boiled in SDS-loading buffer for immunoblot analysis as described (27). For the double immunoprecipitation experiment, Hec1 was immunoprecipitated from 35 S-labeled T24 cells as above. The washed immunocomplex was then incubated at 100°C for 5 min in 200 l of disassociation buffer (20 mM Tris, pH 7.4, 50 mM NaCl, 1% SDS, and 5 mM dithiothreitol). The heated immunocomplex was then diluted with 1 ml of cold Lysis 250 and immunoprecipitated again with the same polyclonal anti-Hec1 antibody (26,27). For the co-immunoprecipitation experiments from yeast cells, yeast cell lysate was prepared as described (3). Briefly, yeast cells were collected by centrifugation and washed twice with cold distilled H 2 O. Glass beads were used to break the cells in lysis buffer (50 mM Tris, pH 7.5, 10 mM MgSO 4 , 1 mM EDTA, 10 mM KOAc, 1 mM dithiothreitol). Clarified yeast cell lysates were then used for co-immunoprecipitation by anti-Hec1 mAb 9G3 or anti-scHec1 polyclonal antiserum as described above. After a 4-h incubation with antibodies and protein A-Sepharose beads, the beads were collected, washed extensively with lysis buffer, and then boiled in SDS-loading buffer. After immunoblotting to Immobilon-P membrane (Millipore Corp., Bedford, MA), blots were probed with anti-scHec1 antibodies, anti-Hec1 mAb 9G3, or anti-scNek2 antibodies. All but one of the immunoblots were developed by 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium substrate for alkaline phosphatase-conjugated anti-mouse antibodies. Horseradish peroxidase-conjugated protein A was used to detect anti-scNek2 antibodies. Blots using that antibody were developed using an ECL chemiluminescence kit (Amersham Biosciences), according to the manufacturer's instructions.
Metabolic Labeling-T24 cells grown in Dulbecco's modified Eagle's medium plus 10% fetal calf serum were synchronized at G 1 by density arrest and then released at time 0 by replating in Dulbecco's modified Eagle's medium plus 10% fetal calf serum at a density of 2 ϫ 10 6 cells/10-cm plate. At various time points thereafter (18 h for G 1 /S, 22 h for S, 32 h for G 2 ), the cells were metabolically labeled with 100 Ci/ml [ 32 P]phosphoric acid (ICN, Costa Mesa, CA) for 3 h and harvested for immunoprecipitation with mAb 9G3 anti-Hec1 antibodies as described above.
Colony Sectoring Assay-Chromosome segregation errors were measured by colony sectoring assay as described (3,28), except that adenine was added at a concentration of 6 g/ml instead of 30 g/ml.
Purification of His-tagged Hec1 Protein-cDNA encoding full-length Hec1 was digested with XhoI and fused in-frame to His 6 at the N terminus and expressed in E. coli using the PET expression system (29). After induction with 0.1 mM isopropylthiogalactoside, cells were lysed and clarified by centrifugation. The clarified total soluble cellular protein was passed through a DEAE-Sepharose column (Amersham Biosciences). The flow-through from the column was passed through an SP-Sepharose column (Amersham Biosciences) and eluted with a 100 -750 mM NaCl gradient. His-Hec1 eluted at fractions between 200 and This study This study This study 300 mM NaCl. The fractions from the SP Sepharose column were loaded onto a nickel-Sepharose column (Amersham Biosciences) and eluted with 60 mM imidazole. The Hec1 protein was fractionated by Sephadex 300 to obtain purified protein.

Expression of His-tagged Nek2 in a Baculovirus System-Full-length
Nek2 was fused in-frame to His 6 and expressed in a baculovirus system as described (30). 36 h after infection, the infected Sf9 cells were lysed and immunoprecipitated with anti-Nek2 antisera. The resulting immune complexes were used in kinase assays.
In Vitro Kinase Assay-Immunoprecipitated recombinant Nek2 was washed with Lysis 250 buffer five times, followed by washing twice with Tris-buffered saline (10 mM Tris-HCl, pH 7.4, 20 mM NaCl) and once with distilled H 2 O. The kinase reactions were carried out for 20 min at 37°C in Nek2 kinase buffer (0.5 M Hepes, pH 7.5, 50 mM MnCl 2 , 50 mM NaF, 50 mM ␤-glycerol phosphate, 10 M okadaic acid, 10 g/ml heparin sulfate, 40 M ATP, and 10 mM dithiothreitol) supplemented with 10 Ci of [␥-32 P]ATP. Purified Hec1 proteins (5 g) were added to the kinase reactions as described (20). Kinase reactions were stopped by adding 2ϫ SDS sample buffer, and proteins were separated by SDS-PAGE. The resulting gel was dried and autoradiographed.
Phosphoamino Acid Analysis-T24 cells were labeled with [ 32 P]orthophosphoric acid for 2 h, followed by immunoprecipitation with mAb 9G3 anti-Hec1 antibody. Immune complexes were separated by SDS-PAGE and transferred to Immobilon-P membrane. Phosphoamino acid analysis was performed as described (31).
Antibody Production-For production of anti-Nek2 antibody, cDNA encoding aa 235-399 of Nek2 was fused to glutathione S-transferase (GST) in frame. Purified GST-Nek2 fusion protein was used as antigen to immunize a mouse to produce mouse polyclonal anti-Nek2 antisera.
For the anti-scNek2/Kin3 antibody, cDNA encoding full-length scNek2/ Kin3 was fused to GST in frame; the fusion protein was purified and used as an antigen. Anti-Hec1 and anti-scHec1 antibodies have been described (1,3). For anti-phosphorylated Hec1 antibody, a synthetic phosphopeptide (A439; Fig. 3A) was coupled to keyhole limpet hemacyanin (KLH) and used as antigen.

HEC1 Is a Serine Phosphoprotein, and Its Phosphorylation Is
Cell Cycle-dependent-To explore a potential mechanism by which Hec1 is regulated, we tested whether Hec1 is modified by phosphorylation. T24 cells were labeled with either [ 35 S]methionine or 32 P-orthophosphate and lysed. The lysates were immunoprecipitated with polyclonal anti-Hec1 serum, monoclonal anti-Hec1 antibodies (mAb9G3), or preimmune serum and then separated by SDS-PAGE. The 76-kDa Hec1 protein recognized by both polyclonal and monoclonal 9G3 antibodies was labeled by 32 P (Fig. 1A, lanes 5 and 6), showing Hec1 to be a phosphoprotein. Phosphoamino acid analysis showed Hec1 to be phosphorylated only on serine residues (Fig. 1B). To determine the cell cycle dependence of Hec1 phosphorylation, T24 cells released from density arrest at G 0 phase for different Deletion mutants containing the different coiled-coil domains of Hecl were fused in frame to a GAL4 DNA binding domain. Nek2 was expressed as a GAL4 transactivation domain fusion. Yeast transformants with these two hybrid proteins were grown in liquid cultures and used for O-nitrophenyl-␤-galactopyranosidase quantification of ␤-galactosidase activity. The -fold increase in activity compared with the host yeast strain Y153 is indicated. Assays were performed in triplicate for each transformation. C, cell cycle-dependent interaction between Hecl and Nek2. T24 bladder carcinoma cells were first density-arrested at G 1 (lanes 2) and then released for reentry into the cell cycle. At different time points after release from density arrest (indicated above the lanes), cells were collected and lysed. The clarified lysates were immunoprecipitated with mAb9G3 anti-Hecl monoclonal antibodies (upper two panels) or anti-Nek2 antisera (lower two panels). Hecl and Nek2 coimmunoprecipitated at G 2 and M phases (lanes 5 and 6).
periods of time were labeled with [ 32 P]orthophosphate and analyzed. The phosphorylation of Hec1 began during time periods corresponding to S phase and was most prominent during M phase (Fig. 1C). These results showed Hec1 to be phosphorylated on serine residues by a cell cycle-regulated serine kinase.
Hec1 Binds to Nek2 Specifically at G 2 /M Phase-A candidate kinase for phosphorylating Hec1 is Nek2, which we found in a yeast two-hybrid screen to be a specific interacting protein (1). The binding of Hec1 with Nek2 was further established using GST pull-down assays ( Fig. 2A). In vitro translated Nek2 interacted with the carboxyl-terminal portion of Hec1 (aa 251-618). Using Hec1 deletion constructs and yeast two-hybrid assays, we found that Hec1 interacts with Nek2 via the first (aa 251-431) or second (aa 361-547) coiled-coil domain (Fig. 2B).
To determine the cell cycle specificity of the Hec1-Nek2 interaction in living cells, T24 cells released from density arrest at G 0 phase were collected at various subsequent time points. Proteins from cell lysates were immunoprecipitated with either anti-Nek2 serum or with mAb 9G3, which recognizes Hec1. The expression of both Hec1 and Nek2 was regulated during progression of the cell cycle (Fig. 2C). Co-immunoprecipitation of Nek2 and Hec1 occurred specifically during G 2 and M phases (Fig. 2C, lanes 5 and 6). The initiation of Hec1 phosphorylation (Fig. 2C, G24, lane 6) corresponded to the same time period during which Nek2 was most abundant (Fig. 2C, lane 4), suggesting that Nek2 may phosphorylate Hec1 in vivo during G 2 /M phase.
Phosphorylation of Hec1 on Serine 165 in Vivo-We noted that Hec1 has a potential phosphorylation site at serine 165 for both NimA and Nek2 (16, 20, 32) (Fig. 3A). To test whether Ser 165 of Hec1 is the authentic site phosphorylated by Nek2, an   FIG. 3. Hec1 is phosphorylated on serine 165 in vivo. A, potential Nek2 recognition sequence including serine 165 of human Hec1 (hs-Hec1) and the chemically synthesized phosphopeptide (A439) used as antigen for generating specific antibodies. B, anti-A439 antibodies specifically recognize phosphorylated Hec1. T24 cells, either unsynchronized (T24-U, lanes 1 and 2) or treated with nocadozole to arrest them at G 2 /M (T24-M, lanes 3 and 4), were lysed and immunoprecipitated with mAb 9G3. The immunoprecipitates were not treated (lanes 1 and 3) or treated with calf intestine phosphatase (CIP) (lanes 2 and 4), separated by SDS-PAGE, and subjected to Western blotting probed with mAb 9G3 (upper panel) or anti-A439 antibodies (lower panel). Anti-A439 antibodies recognized the untreated but not calf intestine phosphatase-treated Hec1. C, expression of the phosphorylated Hec1 detected by anti-A439 antibodies during cell cycle progression. T24 cells in different synchronized stages of the cell cycle were prepared as described above. Hec1 was detected by straight Western blotting with either mAb 9G3 (upper panel) (G11, 11 h after release for G 1 ; G18, 18 h after release for G 1 /S; G24, S phase; G32, G 2 /M phase) or anti-A439 antibodies (lower panel).

FIG. 4. Nek2 phosphorylates Hec1 in vitro.
A, purification of His-tagged Hec1. His 6 -tagged full-length Hec1 was expressed in E. coli using the PET expression system (29). The total bacterial lysate (lane 1) was passed through a DEAE-Sepharose column, and the flow-through (lane 2) was then bound to an SP Sepharose column. Hec1 eluted with NaCl gradient fractions between 200 and 300 mM. The eluant was then loaded onto a nickel-Sepharose column and eluted with 60 mM imidazole (lane 4). This eluant was loaded onto a Sephadex 300 column (lane 5) to obtain nearly pure Hec1. Hec1 from different steps of the purification was subjected to SDS-PAGE and then stained with Coomassie Blue. Hec1S165A was purified by an identical scheme; the final purified product is shown in lane 6. B, expression of His-tagged Nek2 in a baculovirus system. Baculovirus carrying the His 6 full-length Nek2 was generated as described (30). Cell lysates from infected (lanes 2 and 3) or uninfected (lane 1) Sf9 cells were probed with anti-Nek2 antibodies to demonstrate the specificity of the anti-Nek2 antibodies. The antibodies specifically recognize the recombinant Nek2 protein but not proteins from uninfected Sf9 cells. C, Nek2 phosphorylates Hec1. Kinase reactions were performed with [␥-32 P]ATP to assess the activity of immunopurified Nek2 kinase, using either wild-type Hec1 antibody specifically recognizing a synthesized Hec1 phosphopeptide (Fig. 3A) was generated and used to examine the expression of phosphorylated Hec1 (Fig. 3B). Lysates from T24 cells, from cells synchronized at M phase, and from an unsynchronized population were immunoprecipitated with anti-Hec1 antibodies. The anti-A439 antibody recognized the phosphorylated form of Hec1 but did not recognize the unphosphorylated form from lysates treated with calf intestine phosphatase (Fig. 3B). In contrast, interaction between Hec1 and mAb 9G3 recognized both phosphorylated and unphosphorylated forms of Hec1 and was not affected by phosphatase treatment (Fig. 3B). The phosphorylated form of Hec1 was detected most abundantly by anti-439 in the lysates enriched for mitotic cells (Fig. 3C, lane 3). This finding is consistent with the 32 P labeling experiment shown in Fig. 1C, in which the phosphorylated form of Hec1 was most abundant at the G 2 /M phase. Together, the results suggest that human Hec1 is phosphorylated on serine 165 in vivo.
Yeast Kin3 Shares Similar Properties with Nek2-Human Hec1 (hsHec1) has a structural and functional homolog in yeast, scHec1/TID3/NDC80/YIO4, and is required for faithful chromosome segregation (3). Since Hec1 is specifically phosphorylated at the G 2 and M phases, phosphorylation of Hec1 by Nek2 may be critical for chromosome segregation. To address this possibility, a yeast model system was employed because well established methods are available to assay chromosome segregation (2, 3, 28, 37). However, we first needed to identify a homolog of Nek2 in yeast that may phosphorylate scHec1.
There is an open reading frame in the S. cerevesiae genome, Kin3/scNek2, which encodes a putative protein and could function as a serine/threonine kinase (23,24). This protein shares relatively high homology (36.4% identity) with NimA and human Nek2 in the catalytic domain (Fig. 5A) and contains a coiled-coil domain in its C-terminal region that is similar to the same domains in the other two proteins (Fig. 5B). To test whether these C-terminal regions share similar abilities to physically interact with hsHec1p or scHec1p, Nek2p and scNek2p were synthesized in vitro for GST pull-down assays with both GST-hsHec1 and GST-scHec1. Nek2p and scNek2p could bind both human and yeast Hec1p (Fig. 5C). These results suggested that scNek2 and Nek2 not only share homology at their N-terminal kinase domain sequences but that they also both have Hec1 binding activity at their C-terminal regions.
Changing glutamic acid 41 of NimA into glycine leads to a temperature-sensitive growth phenotype that arrests the cells in the G 2 phase at the nonpermissive temperature (8,9). Interestingly, similar acidic residues have been found by other researchers to be highly conserved in the other kinases: residue FIG. 5. Yeast Kin3 shares properties with Nek2. A, primary sequence homology comparison between NimA, Nek2, and scNek2/Kin3. B, comparison of the potential coiled-coil regions in NimA, Nek2, and scNek2/Kin3, predicted from primary sequences using a program found on the World Wide Web at www.isrec.isb-sib.ch/software/software.html. C, Nek2 or scNek2 binds to hsHec1 and scHec1 by GST pull-down assay. GST (lanes 2 and 7), GST fusions of a small N-terminal hsHec1 (GST-Hec1 aa 56 -128) (lanes 3 and 8), a longer portion of hsHec1 (aa 56 -618, GST-Hec1) (lanes 4 and 9), and a full-length of scHec1 (aa 1-691, GST-scHec1) (lanes 4 and 8) were prepared and used to bind to in vitro translated human Nek2 (lane 1) or scNek2/Kin3 (lane 6). Human Nek2 and scNek2/Kin3 both bind to hsHec1 and scHec1. D, partial sequence of NimA showing the glutamic acid residue at amino acid position 41. Comparison of homologous sequences in Nek2 and scNek2/Kin3, with glutamic acid at amino acid residue 38 (Nek2) and aspartic acid at amino acid residue 55 (scNek2/Kin3), is also shown. 38 (glutamic acid) in Nek2 (19), and residue 55 (aspartic acid) in scNek2 (23,24) (Fig. 5D). To test the functional similarity of these key regions among the kinases, glutamic acid 38 of Nek2 and aspartic acid 55 of scNek2 were each changed to glycine. Like the temperature-sensitive nima mutant (8,9), the scNek2D55G mutant grew at 25°C, arrested at 37°C, and reentered the cell cycle when shifted back to the 25°C (Fig. 6A). When the Nek2E38G mutant was introduced into scNek2 null cells, growth and propagation of the cells was temperaturesensitive as well (Fig. 6B). This temperature-sensitive phenotype was partially suppressed by expression of additional wildtype scNek2 or hsNek2 (Fig. 6C). Taken together, these results suggest that scNek2/Kin3 shares several similar functions with hsNek2 and that scNek2/Kin3 might function as an Nek2 homolog in S. cerevesiae.
Temperature-sensitive scNek2 Mutant Fails to Phosphorylate Hec1 at the Nonpermissive Temperature-To determine whether scNek2D55G may behave in a dominant negative fashion to arrest cells at nonpermissive temperature, we first generated specific antibodies and examined the physical interaction between scNek2p and scHec1p or scNek2p and hsHec1p, using co-immunoprecipitation (Fig. 7, A and B). In cells carrying the scNek2D55G mutant, the interaction between scNek2p and scHec1p or scNek2p and hsHec1p (Fig. 7, C and D) was intact, as it was in wild-type scNek2 cells (Fig. 7, C and D). Moreover, the phosphorylation of hsHec1p on serine 165 was detected by anti-A439, both in wild-type and in scNek2D55G mutant cells at the permissive temperature, but not in scNek2D55G mutant cells at the nonpermissive temperature (Fig. 7E). These results suggest that phosphorylation of Hec1 by scNek2p is essential for cells to continue cycling. scNek2D55G thus appears indeed to be a dominant negative mutant; it can bind to Hec1 but cannot phosphorylate it.
Phosphorylation of hsHec1 Serine 165 Is Critical for Its Function in Chromosome Segregation-To examine whether the phosphorylation of human Hec1 (hsHec1) on serine 165 is important for hsHec1 to function, yeast strains containing specific hsHec1 mutations were created. The homolog of human Hec1 in S. cerevesiae (scHec1/TID3/NDC80/YIO4) has been characterized extensively and shown, like its mammalian counterpart, to be essential for chromosome segregation and yeast survival (3,6). Furthermore, hsHec1 can complement the essential functions of scHec1 (3). Mutant constructs were created in which the critical serine residues phosphorylated by Nek2 in scHec1 (Ser 201 ) and hsHec1 (Ser 165 ) were mutated. scHec1S201A and hsHec1S165A substituted the neutral amino acid alanine for serine; scHec1S201E and hsHec1S165E substituted glutamic acid for serine to mimic the negative charge created by serine phosphorylation (Fig. 8A). To test whether these Hec1 mutants could complement scHec1 deficiency, they were introduced into the scHec1 null yeast strain. Both scHec1S201E and Hec1S165E were able to rescue yeast deficient in scHec1, but the scHec1S201A and hsHec1S165A mutants were not (Fig. 8B). These results suggest that phosphorylation of serine 165 (or serine 201 in yeast Hec1) is important for the function of Hec1.
To determine whether substituting glutamic acid for serine 165 in hsHec1 could rescue all essential functions of Hec1 in yeast, plating efficiency and chromosome segregation were examined in schec1 null yeast rescued by either wild-type hsHec1 or hsHec1S165E. The plating efficiency of the schec1⌬/ hsHec1S165E strain was only 75% of the efficiency for the strain rescued by wild-type hsHec1 (Fig. 8C). This result suggested that the hsHec1S165E mutant was not fully functional in allowing faithful mitosis. To address this possibility, colony sectoring assays (2,3,28) were performed in the two yeast strains to monitor chromosome segregation. Yeast cells null for scHec1 and rescued by the hsHec1S165E were 10 times more prone to segregation errors, especially chromosome loss (1:0) events, compared with cells rescued by wild-type hsHec1 (Table  II). Yeast cells lacking scNek2, although viable, were thought to have subtle errors in chromosomal segregation. To test this hypothesis directly, scNek2 null cells were examined by colony sectoring assays. They were found to have 50-fold higher rates of errors of chromosomal losses (1:0 events) and 6-fold higher rates of nondisjunction (2:0 events) (Table II). Taken together, the results suggest that precisely regulated phosphorylation of Hec1 by Nek2 is critical for accurate chromosome segregation during mitosis. DISCUSSION In this paper, we have shown that Hec1 binds to Nek2 both in vitro and in vivo at G 2 /M phase. Nek2 specifically phosphorylates human Hec1 on serine residue 165 in a cell cycle-dependent manner, with a peak activity during G 2 /M. The phos- 6. Growth properties of the scNek2D55G mutant. A, temperature-sensitive growth of yeast strain carrying scNek2D55G. Yeasts were spotted on triplicate plates and grown at two different temperatures. One of the plates originally maintained at 37°C was shifted to room temperature (25°C). Both strains carrying the scNek2D55G mutation were reversibly temperature-sensitive. The growth of these mutant strains was temperature-dependent regardless of the Hec1 status. The ⌬scNek2 mutant grew like the wild-type strain. B, expression of the Nek2E38G mutant in the ⌬scNek2 mutant leads to temperature-sensitive growth. Nek2E38G was introduced into the ⌬scNek2 mutant under control of the scNek2 promoter in a CEN.ARS construct to create strain Nek2E38G. The growth of this mutant strain was temperaturesensitive. C, overexpression of wild type scNek2 or hsNek2 partially suppressed the temperature sensitivity of the scNek2D55G mutant strain.
phorylation of Hec1 is necessary for faithful chromosome segregation and cell survival. Nek2 appears to be the primary kinase responsible for Hec1 phosphorylation, and scNek2/Kin3 is a functional homolog of Nek2 in yeast.
The Role of Hec1 in Chromosome Segregation Is Highly Conserved-Hec1, a coiled-coil protein highly expressed in most cancer cells, is crucial for faithful chromosome segregation. Cells microinjected with anti-Hec1 antibodies undergo aberrant mitosis, with grossly inequitable distribution of chromo-somes (1). Furthermore, Hec1 associates with several proteins required for G 2 /M phase progression, including components of the 26 S proteasome, Smc1/2, and the NimA-like protein kinase Nek2 (5). Nek2 has significant sequence homology with NimA, a serine/threonine kinase required for passage of fungi past G 2 into M phase, exit from M phase, and response to DNA damage (9). To facilitate analysis of the function of Hec1, a yeast homolog scHec1 has been identified and characterized (3,6). scHec1/NDC80 is essential for yeast survival and plays an important role in chromosome segregation. Moreover, human hsHec1 can serve all essential functions in S. cerevesiae null for scHec1, suggesting that fundamental mechanisms governing chromosome segregation are highly conserved in evolutionarily divergent species (3). G 2 /M Phase-specific Regulation of Hec1 by Phosphorylation-Regulation of proteins involved in cell cycle progression must occur rapidly and precisely. Such regulation is especially important during the dynamic processes of mitosis, during which replicated chromosomes must align and distribute equally between daughter cells. The many proteins involved in mitosis must be activated and deactivated during a very narrow time window. Phosphorylation is an excellent way to achieve such precise regulation. We have shown here that yeast strains carrying scHec1S201A or hsHec1S165A mutations, which cannot be phosphorylated by Nek2, are unable to rescue the lethal phenotype in scHec1-deficient yeast. Furthermore, an hsHec1S165E mutation that mimics constitutive phosphorylation of Hec1 can only partially rescue faithful chromosome segregation and subsequent viability of daughter cells.  1). The anti-scNek2 serum was specific for scNek2, since it failed to detect scNek2 in lysate prepared from scNek2 null cells (B, lane 2). C, scHeclp and scNek2 interact at nonpermissive temperature. Wild-type or scNek2D55G mutant strains were grown at 25°C for 4 h before shifting to 37°C for an additional 8 h. The cells were harvested and lysed for immunoprecipitation with anti-scHec1 antibodies (lanes 1-4) and Western blotted with anti-scHeclp antisera (upper panel) or anti-scNek2p antisera (lower panel). scHeclp interacts with scNek2p at permissive and nonpermissive temperature in both wildtype and scNek2D55G cells. D, interaction between hsHecl and scNek2 at nonpermissive temperature in cells carrying the scNek2D55G mutation. scHec1 null cells rescued by hsHec1 in scNek2 or scNek2D55G mutant background were cultured as described for C. Clarified lysates were immunoprecipitated with anti-hsHec1 mAb 9G3 antibodies (lanes [1][2][3][4] and Western blotted with anti-Hec1p mAb 9G3 antibodies (upper panel) or anti-scNek2p antisera (lower panel). hsHeclp interacted with scNek2p at permissive and nonpermissive temperatures in both yeast strains. E, phosphorylation of Hecl is abolished at nonpermissive temperature in scHec1 null cells rescued by hsHec1 in the scNek2D55G mutant. Cells were prepared as described for C. The cells were harvested and lysed in the presence of phosphatase inhibitors. The clarified lysates were immunoprecipitated with mAb 9G3 and Western blotted with mAb 9G3 (upper panel) or monoclonal anti-A439 antibodies (lower panel). hsHec1 was not phosphorylated in scNek2 mutant cells at the nonpermissive temperature.

FIG. 8. Phosphorylation of Hecl by Nek2 is critical for yeast survival.
A, Hec1 phosphorylation sites for Nek2. The potential Nek2 Ser phosphorylation sites in scHec1 (Ser 201 ) and hsHec1 (Ser 165 ) were mutated to Ala (S201A,S165A) or Glu (S201E,S165E) (37). B, both scHec1S201A and hsHeclS165A failed to rescue cells null for scHec1. Only wild type scHec1 or hsHec1 and the scHec1S201E and hsHec1S165E mutants, in which glutamic acid substitution for serine mimics the negative charge created by serine phosphorylation, were able to rescue yeast deficient for scHecl. C, plating efficiency of yeast rescued by wild-type hsHec1 or by hsHeclS165E. Two hundred cells from log phase cultures were plated onto solid plates. The surviving cells were scored for colonies formed on plates after 3 days in culture at 30°C. The results are shown as means Ϯ S.E. from three independent experiments.
Thus, phosphorylation of Hec1 must be tightly regulated and coordinated along with the cell cycle progression.
Hec1 Is a Substrate of Nek2-Specific phosphorylation during G 2 /M is required for Hec1 to function properly and for chromosome segregation to occur faithfully. Our studies have clearly shown Hec1 to be an authentic substrate of Nek2. Failure of Nek2 to phosphorylate Hec1 during G 2 /M leads to errors in segregation of chromosomes. Another potential substrate of Nek2, C-Nap1, localizes to centrioles in both mother and daughter cells and has coiled-coil structures appropriate for other protein-protein interactions (33). The structure of C-Nap1 might allow it to connect proximal ends of centrioles to each other, although this concept at present remains speculative (33). Nevertheless, it is interesting to note that both Nek2 substrates, C-Nap1 and Hec1, localize either to centrioles or kinectochores (1,3,6,21). They are therefore positioned precisely at the mitotic apparatus, along with the machinery responsible for chromosome segregation.
Potential Redundant Kinases for Hec1-scNek2/Kin3 has been purported to be the homolog of Nek2 and NimA because the three proteins have significant structural similarities. However, complete deletion of scNek2 had little influence on yeast survival and led to some suggestions that scNek2 may be functionally different from NimA in fungi. The scNek2D55G mutant in our experiments was generated specifically to mimic the characterized NimA mutants. The growth of cells carrying this scNek2D55G mutant arrested at the nonpermissive temperature, similar to the homologous nima mutant. Interestingly, the physical association between Hec1 and scNek2D55G or Nek2E38G remained intact at any temperature, although the kinase activity of the mutants, and therefore their ability to phosphorylate Hec1, was temperature-sensitive. Hec1 appears to be a crucial substrate of Nek2, and the phosphorylation of Hec1 by Nek2 is required for passage through mitosis and for faithful chromosome segregation. These results support the notion that scNek2/Kin3 is an important gene in yeast with functions similar to Nek2.
Surprisingly, a complete lack of Nek2 was not lethal in yeast; another kinase was apparently able to supplant the function of Nek2 in phosphorylating Hec1. To reconcile the apparent paradox of these observations (i.e. that precisely phosphorylated Hec1 is essential for yeast mitosis but that the kinase responsible for the phosphorylation is not), the existence other kinases with functions redundant for Nek2 must be postulated. Cdc5, based on its structural similarity with Nek2 and cell cycle expression pattern (34 -36), is a potential candidate. A search of GenBank TM showed that Cdc5 shares 35% similarity with NIMA, Nek2, and scNek2/Kin3 in the catalytic domain and, like Nek2, contains a coiled-coil domain near the catalytic domain. Our preliminary results have shown that Cdc5 phosphorylates Hec1 in vitro and specifically associates with Hec1 only when Nek2 is unavailable (data not shown). However, Cdc5 seems to have lower affinity for binding to Hec1. It is particularly interesting that the temperature-sensitive sc-Nek2/Kin3 mutants bind to Hec1 but fail to phosphorylate it at the nonpermissive temperature. These preliminary results suggest that binding and kinase activity are two distinct and potentially independent steps in the activation of Hec1 by Nek2. scNek2D55G thus serves as a dominant negative mutant that binds to Hec1 at the nonpermissive temperature. The secondary kinase, perhaps Cdc5, may fail to compete successfully for binding in the presence of wild-type or mutant Nek2. Cells carrying the dominant negative Nek2 mutation are markedly prone to segregation errors, however, particularly chromosomal losses (1:0 errors), perhaps because the redundant kinase for Hec1 is less efficient or less precisely regulated than Nek2. These data provide a reasonable explanation for why yeast cells completely lacking scNek2 are viable but those with the scNek2D55G mutation are growth-arrested at the nonpermissive temperature.
Implications in Higher Organisms and in Cancer-We have demonstrated that Hec1 is an important substrate of hsNek2 and scNek2/Kin3. Hec1 is specifically phosphorylated by these kinases, and such phosphorylation is required for faithful chromosome segregation. Without Hec1, chromosomes distribute to daughter cells in a disordered, ultimately lethal fashion. Without precise regulation through phosphorylation of Hec1 by Nek2 during G 2 /M phases, more subtle errors in chromosome segregation, similar to those involved commonly in the progression of cancer in humans, are likely. If we can extrapolate findings in yeast to similar systems controlling mitosis in humans, then phosphorylation of Hec1 by Nek2 in mammalian systems may be a focus for exploring chromosomal mechanisms of carcinogenesis and cancer progression. Because of the abundant expression of Hec1 in cancer cells (1), the specific phosphorylation of Hec1 by Nek2 may also be a potential target for drug development in the treatment of cancers.

TABLE II
Increased chromosome segregation errors Yeast cells null for scHec1 and rescued by the hsHec1S165E mutant (strain WHL4001) were susceptible to segregation errors, which were monitored by sectoring assays (2). Chromosome loss errors (1:0) in these hsHec1S165E mutant cells were especially prominent compared with wild-type cells (strain WHL2003) or scHec1 null cells rescued by wild-type hsHec1 (strain WHL4003). Yeast cells without scNek2 (strain WHL6502) had a much higher incidence of chromosome segregation errors compared to cells rescued by wild-type hsHec1 or hsHec1S165E. NS, not significant.