CENP-E Kinesin Interacts with SKAP Protein to Orchestrate Accurate Chromosome Segregation in Mitosis*

Background: CENP-E is a kinetochore-associated kinesin responsible for chromosome congression in mitosis. Results: CENP-E interacts with SKAP to orchestrate kinetochore-microtubule interaction. Conclusion: The SKAP-CENP-E interaction links kinetochore structural components to the spindle microtubule attachment in the centromere. Significance: SKAP cooperates with CENP-E to ensure chromosome stability in cell division. Mitotic chromosome segregation is orchestrated by the dynamic interaction of spindle microtubules with the kinetochore. Although previous studies show that the mitotic kinesin CENP-E forms a link between attachment of the spindle microtubule to the kinetochore and the mitotic checkpoint signaling cascade, the molecular mechanism underlying dynamic kinetochore-microtubule interactions in mammalian cells remains elusive. Here, we identify a novel interaction between CENP-E and SKAP that functions synergistically in governing dynamic kinetochore-microtubule interactions. SKAP binds to the C-terminal tail of CENP-E in vitro and is essential for an accurate kinetochore-microtubule attachment in vivo. Immunoelectron microscopic analysis indicates that SKAP is a constituent of the kinetochore corona fibers of mammalian centromeres. Depletion of SKAP or CENP-E by RNA interference results in a dramatic reduction of inter-kinetochore tension, which causes chromosome mis-segregation with a prolonged delay in achieving metaphase alignment. Importantly, SKAP binds to microtubules in vitro, and this interaction is synergized by CENP-E. Based on these findings, we propose that SKAP cooperates with CENP-E to orchestrate dynamic kinetochore-microtubule interaction for faithful chromosome segregation.

Mitotic chromosome segregation is orchestrated by the dynamic interaction of spindle microtubules with the kinetochore. Although previous studies show that the mitotic kinesin CENP-E forms a link between attachment of the spindle microtubule to the kinetochore and the mitotic checkpoint signaling cascade, the molecular mechanism underlying dynamic kinetochore-microtubule interactions in mammalian cells remains elusive. Here, we identify a novel interaction between CENP-E and SKAP that functions synergistically in governing dynamic kinetochore-microtubule interactions. SKAP binds to the C-terminal tail of CENP-E in vitro and is essential for an accurate kinetochore-microtubule attachment in vivo. Immunoelectron microscopic analysis indicates that SKAP is a constituent of the kinetochore corona fibers of mammalian centromeres. Depletion of SKAP or CENP-E by RNA interference results in a dramatic reduction of inter-kinetochore tension, which causes chromosome mis-segregation with a prolonged delay in achieving metaphase alignment. Importantly, SKAP binds to microtubules in vitro, and this interaction is synergized by CENP-E. Based on these findings, we propose that SKAP cooperates with CENP-E to orchestrate dynamic kinetochore-microtubule interaction for faithful chromosome segregation.
The inheritance of genetic material depends on the consistent segregation of chromosomes in mitosis. The physical connection between the centromeres and spindle microtubules is orchestrated by the kinetochore, a protein supercomplex assembled onto the centromere (1).
The stochastic nature by which highly dynamic spindle microtubules attach to the chromosome requires that the centromere-kinetochore protein machine be able to capture microtubules and ensure the quality of the kinetochore-microtubule interaction (2). Once all chromosomes achieved bipolar attachment and alignment at the equator of the spindle, the spindle assembly checkpoint was satisfied and resulted in subsequent equal segregation of parental genomes into two daughter cells (3). The identification of numerous kinetochore outer proteins in yeast and worms and the molecular characterization of their functions offer a new view of the structure (4,5). However, the precise molecular identity of the kinetochore-microtubule interface in human cells has remained elusive as not all yeast and worm kinetochore proteins are conserved in the human. Our previous studies show that mitotic kinesin CENP-E forms a link between attachment of spindle microtubules to kinetochores and the mitotic checkpoint (6,7). Most significantly, components that interact with CENP-E and dynamic spindle microtubule plus-ends essential for chromosome segregation have not been identified.
Recent studies reveal that several kinetochore protein complexes exhibit weak but cooperative associations with spindle microtubule in mitosis (2,8). Those include kinetochore core complex KMN (9,10), mitotic kinesin CENP-E (7), microtubule plus-end tracking protein complexes such as EB1 and TIP150 (11), and other kinetochore proteins. CENP-E participates in the chromosome movements from prometaphase to anaphase (12) by tethering the kinetochores to microtubule "plus" ends and moving toward the equator (7,13), thereby helping monooriented chromosomes align at the metaphase plate before biorientation (14). Using the chemical inhibitor syntelin, it has been recently demonstrated that CENP-E function is essential for bi-orientation as syntelin-treated cells exhibit syntelic chromosomes (15).
To illustrate the molecular network and dynamics underlying CENP-E-microtubule orchestration, we carried out a combination of CENP-E immunoisolations of the kinetochore protein complex and yeast two-hybrid screen for CENP-E-interacting proteins. This screen has identified Nuf2 and SEPT7 as two of several dozen positive clones (16,17). Here, we report our characterization of CENP-E interaction with SKAP, a recently identified kinetochore protein (18 -20). Our biochemical studies revealed that SKAP directly interacts with CENP-E via the C-terminal coiled-coil region. This SKAP-CENP-E interaction is critical for achieving accurate chromosome alignment at the equator during mitosis.

MATERIALS AND METHODS
Cell Culture-HeLa cells from the American Type Culture Collection (Manassas, VA) were maintained as subconfluent monolayers in advanced DMEM (Invitrogen) with 10% FCS (Hyclone) and 100 units/ml penicillin plus 100 g/ml streptomycin (Invitrogen).
Affinity Purification-The CENP-E complex was isolated from mitotic HeLa cells using an anti-CENP-E antibody affinity matrix as described previously (6,16). After extensive washing, the CENP-E immunoprecipitates were fractionated on SDS-PAGE. Coomassie Blue-stained bands were removed for in-gel digestion; this was done essentially as described previously followed by mass spectrometric identification (26). Positive hits were validated using specific antibodies.
Yeast Two-hybrid Assay-Yeast two-hybrid assays were performed as described previously (16,17). Briefly, a CENP-E bait containing amino acids 2131-2701 was inserted into the BamHI-EcoRI sites of pGBKT7 individually to create a fusion with amino acids 1-147 of the Gal4 DNA-binding domain. The resultant pGBKT7/CENP-E C termini were transformed into strain AH109 along with the GAL4 reporter plasmid pCL and the negative control plasmid pGBKT7-Lam. Protein expression was validated by Western blotting using Gal4 and an anti-CENP-E antibody. Specificity of the interaction was independently verified by retransforming the candidate cDNAs back into AH109 along with BD-CENP-E(2131-2701). Those cDNAs that can form colonies grown from Leu Ϫ , Trp Ϫ , His Ϫ , Ade Ϫ SD plates were sequenced. From the candidate-positive clones, we selected SKAP to perform further analyses.
Recombinant Protein Production-Purification of recombinant proteins was carried out as described previously (21,22). Briefly, the GST fusion protein in bacteria in the soluble fraction was purified by using glutathione-agarose chromatogra-phy, and histidine-tagged protein was purified using nickel-nitrilotriacetic acid-agarose beads.
Antibodies-Antibodies against CENP-E (mAb177 and HpX) were generated previously (7,12). Both rabbit and mouse antibodies to SKAP were generated by using full-length recombinant proteins from bacteria using the standard protocol as described previously (21). Anti-tubulin antibody (DM1A) was purchased from Sigma. Anti-SKAP antibody was described previously (18).
Immunofluorescence Microscopy-For immunofluorescence, cells synchronized by mitotic shake-off were seeded onto sterile, acid-treated 18-mm coverslips in 6-well plates (Corning Glass Works, Corning, NY). Two hours after replating, synchronized HeLa cells were transfected with 2 g/ml Lipofectamine premixed with various oligonucleotides as described previously (6).
At various times after transfection with siRNA or scrambled (control) oligonucleotides, cells were rinsed for 1 min with PHEM buffer (100 mM PIPES, 20 mM HEPES, pH 6.9, 5 mM EGTA, 2 mM MgCl 2 , and 4 M glycerol) and permeabilized for 1 min with PHEM plus 0.1% Triton X-100 as described previously (6,7). Extracted cells were then fixed in freshly made 4% paraformaldehyde plus 0.05% glutaraldehyde in PHEM and rinsed three times in PBS. Coverslips were blocked with 0.05% Tween 20 in PBS (TPBS) with 1% BSA (Sigma). Cells were incubated with various primary antibodies in a humidified chamber for 1 h and then washed three times in TPBS. To visualize CENP-E and SKAP simultaneously, cells were incubated with mouse anti-SKAP antibody in a humidified chamber for 1 h and then washed three times in TPBS. Rabbit polyclonal antibodies bound to CENP-E were visualized using rhodamine-conjugated goat anti-rabbit immunoglobulin G (IgG), and binding of anti-SKAP antibody was visualized using fluorescein-conjugated goat anti-mouse IgG. DNA was stained with DAPI (Sigma). Slides were examined with a Zeiss Axiovert-200 fluorescence microscope, and images were collected and analyzed with Image-5 (Carl Zeiss, Germany).
Deconvolution Microscopy-Deconvolution images were collected using a Deltavision wide field deconvolution microscope system built on an Olympus IX-71 inverted microscope base. For imaging, a 100 ϫ 1.35 NA lens was used, and optical sections were taken at intervals of 0.2 m. Images for display were generated by projecting single optical sections as described previously (3).
Measurement of Inter-kinetochore Distance-The distance between sister kinetochores was measured using LSM-5 imaging (Carl Zeiss), ACA 4 -marked centromeres, and a Zeiss LSM510 confocal microscope as described previously (6). In some cases, an aliquot of syntelin, CENP-E inhibitor (1 M), was added to scramble transfected or SKAP siRNA-treated HeLa cells culture for 30 min prior to fixation and immunofluorescence staining (15). Only sister kinetochores that were in the same focal plane were measured.
Immunoelectron Microscope-For immunoelectron microscopic assay, HeLa cells were seeded onto sterile, acid-treated 18-mm coverslips for 12 h, pre-extracted, and fixed using a common protocol (7). After fluorescent examination to verify fixation and antibody binding, SKAP was visualized by 10-nm colloidal gold. The ultrastructural localization of SKAP was carried out as described previously (7,23). Thin serial sections (silver-gold) were then cut, placed on copper grids, and stained with uranyl acetate and lead citrate. The sections were examined by a JEOL 1200 EM (Peabody, MA).
Transient Transfection and Immunoprecipitation-Transient transfection and immunoprecipitation were performed as described previously (6). After SDS-PAGE, proteins were transferred to a nitrocellulose membrane, probed with appropriate antibodies, and detected with the ECL system (22).
Microtubule Cosedimentation Assays-Microtubule assembly and measurement of polymeric tubulin concentrations were performed as described previously (24,25). For microtubulebinding reactions, purified recombinant GST-SKAP and MBP-CENP-E(2132-2701) proteins were incubated with various concentrations of taxol-stabilized microtubules at room temperature for 10 min followed by sedimentation for 10 min at 80,000 rpm in a TLA100 rotor at 25°C. For visualization and quantification of SKAP, pellets and supernatants were solubilized in SDS gel sample buffer and subjected to electrophoresis. Some gels were stained with Coomassie Blue and dried between sheets of cellulose for visualization and quantitative scanning. Others were transblotted to nitrocellulose and probed with SKAP antibodies.
Real Time Image Acquisition-Cells were maintained at 35-37°C using a heated stage. Images of cells expressing GFPhistone H2B and GFP-CENP-A were collected on an inverted microscope (Olympus IX-70) and a 60ϫNA 1.4 PlanApo objective. 0.2-m step sections were acquired using a 100ϫNA 1.3 U-PlanApo objective (Olympus) with 1 ϫ 1 binning. Acquisition parameters, including exposure, focus, and illumination, were controlled by SoftWorks (Applied Precision). Z stack projection, subsequent analysis, and processing of images were performed using SoftWorks (Applied Precision). For analysis of microtubule attachments, images were deconvolved using the DeltaVision software (Applied Precision). Measurements of the intensity of kinetochore localization were conducted on nondeconvolved images. All images for a specific experiment used identical exposure settings and scaling. For live cell imaging, medium was replaced with CO 2 -independent medium supplemented with 10% FBS, penicillin/streptomycin, and L-glutamine (Invitrogen) was covered with mineral oil with a 60ϫNA 1.40 PlanApo objective lens together with a filter wheel. Images were analyzed with DeltaVision deconvolution system (Applied Precision).
Data Analyses-All distance and fluorescence intensity measurements were made using MetaMorph software. Analysis of kinetochore movements was performed by measuring the distance from the center of sister kinetochores to a point on the spindle equator along the trajectory of chromosome movement. The bulk of metaphase-aligned chromosomes was used as a reference point for the spindle equator, which occasionally shifted during our analysis. Inter-kinetochore distances were measured using the centers of the paired CENP-A dots. Kinetochore fluorescence intensities were determined by measuring the integrated fluorescence intensity within a 7 ϫ 7-pixel square positioned over a single kinetochore and subtracting the background intensity of a 7 ϫ 7-pixel square positioned in a region of cytoplasm lacking kinetochores. Maximal projected images were used for these measurements. To determine significant differences between means, unpaired t tests assuming unequal variance were performed; differences were considered significant when p Ͻ 0.05.

SKAP Is a Novel CENP-E-interacting
Protein-To study the molecular association of CENP-E with other kinetochore proteins, we carried out an affinity isolation of a CENP-E-containing protein complex followed by mass spectrometric identification of tryptic peptides derived from the complex as described previously (26). As shown in Fig. 1A, anti-CENP-E affinity beads coupled to an anti-CENP-E antibody isolated a major polypeptide of 330 kDa in addition to two visible bands of polypeptides with an approximate mass of 125 and 32 kDa, respectively. The 330-kDa polypeptide is CENP-E based on the analyses of mass spectrometry and Western blotting. Our mass spectrometric analyses identified the 125-kDa polypeptide as BubR1, a characterized mitotic kinase interacting with CENP-E in mitosis (6,27). In addition, our analyses had identified the 32-kDa polypeptide as a previously uncharacterized open reading frame of FLJ14502 (c15orf23; data not shown). Western blotting analyses validated that BubR1, Nuf2, and Septin7 exist in the immunoprecipitates of CENP-E but to a lesser extent.
During the preparation of this paper, a number of studies have named the 32-kDa protein as SKAP (small kinetochoreassociated protein) (18 -20). Computational analysis shows that SKAP encodes a polypeptide of 316 amino acids and contains two coiled-coil regions in its C-terminal half. SKAP exists in mammals but is absent from lower eukaryotes. Rabbit antibody raised against bacterially recombinant protein of SKAP recognizes a 32-kDa polypeptide in CENP-E immunoprecipitation (Fig. 1B). In addition, our immunoprecipitation of SKAP using a FLAG tag antibody confirmed that SKAP binds to the C-terminal of CENP-E (GFP-CENP-E(2131-2701), Fig. 1C, lane 4). The FLAG immunoprecipitation did not pull down GFP tag alone from a parallel transfection (data not shown), validating an independent yeast two-hybrid screen for proteins interacting with the tail of CENP-E (16). To demonstrate a direct interaction between SKAP and CENP-E, we have carried out a pulldown assay in which GST-tagged SKAP was expressed in bacteria and used as an affinity matrix to absorb MBP-CENP-E(2131-2701). Coomassie Blue-stained SDS-PAGE confirmed a direct interaction between SKAP and CENP-E in vitro (Fig.  1D, lane 3). Thus, we conclude that SKAP binds to CENP-E in vitro and in vivo.
SKAP Located to Fibrous Corona of Kinetochore-CENP-E is a mitotic kinesin that accumulates in G 2 /M and translocates to kinetochore upon the nuclear envelop breakdown (7). We sought to examine if SKAP distributes with CENP-E to the kinetochore of mitotic cells. As shown in Fig. 2A, SKAP signal is readily evident at the kinetochore where is superimposed to that of CENP-E (merge; red dots, see both insets). In addition, SKAP immunoreactivity is apparent in centrosomes ( Fig. 2A,  arrows). As expected, exogenously expressed GFP-SKAP is indeed localized to the kinetochore (Fig. 2B). Thus, we conclude that SKAP colocalizes with CENP-E to kinetochore corona in mitosis.
CENP-E is shown to be an integral component of the kinetochore corona fibers (7). To better understand the exact molecular function of SKAP, we carried out immunoelectron microscopic analysis of SKAP in mitotic HeLa cells. One bi-oriented kinetochore is highlighted in Fig. 2C (asterisk marks centriole). At a higher magnification, seven 10-nm gold particles representing specific labeling of SKAP can be seen clearly on the interface between spindle microtubules and the kinetochores (Fig. 2D, arrows). This is consistent with the previously reported position for CENP-E (7, 29). Thus, we conclude that SKAP is located in the interface between kinetochores and spindle microtubules.
SKAP Is Essential for Accurate Chromosome Alignment in Mitosis-To investigate the influence of SKAP on mitotic progression, we introduced siRNA oligonucleotides of SKAP into HeLa cells synchronized in early G 1 . As SKAP protein is quantitatively degraded at the end of mitosis (data not shown), suppression of new synthesis in early G 1 cells allowed us to examine SKAP function without the complication of a pre-existing pool of SKAP. We confirmed by immunoblotting that our RNAi protocol caused an ϳ90% reduction in the amount of SKAP protein after 24 h, although the levels of both tubulin and CENP-E showed no fluctuations in siRNA-treated cells (Fig. 3A). If SKAP is essential for mitotic progression, reduction of SKAP should result in a mitotic arrest. Indeed, in cells with normal SKAP levels (transfected with a siRNA duplex of scrambled sequence), a wave of mitosis was seen at 20 h, at which time 79 Ϯ 3% of cells were in mitosis, as scored by the number of rounded cells (Fig. 3B). In contrast, although cells with diminished SKAP levels proceeded through interphase with roughly the same cell cycle time, they retained a rounded conformation that was indicative of mitotic arrest. At 20 h, the mitotic index was 81 Ϯ 3% and thereafter remained at a high level (77 Ϯ 4% at 22 h), which is consistent with an almost total mitotic arrest in the transfected cell population. The mitotic index for scramble-transfected control cells is 17 Ϯ 2% at 22 h. Thus, we conclude that SKAP is essential for mitotic progression.
To determine whether the prometaphase arrest that arises from elimination of CENP-E reflects an underlying defect in spindle assembly or maintenance, we visualized spindles with antibodies against tubulin and identified centromeres with ACA (a classical centromere marker for immunofluorescence). The reduction in total SKAP content was also reflected in loss of SKAP at kinetochores, although ACA levels were largely unaffected, as demonstrated by double indirect immunofluorescence with antibodies against SKAP and ACA, respectively (Fig. 3C). Careful examination revealed that some chromosomes in the SKAP-suppressed cells failed to align at the equator (Fig. 3C, arrows). Although the knockdown of SKAP does not alter the localization of CENP-E, the suppression of CENP-E results in a moderate reduction of SKAP in kinetochores (Fig. 3D). Quantitative analyses of SKAP relative fluorescence intensity indicated that depletion of CENP-E causes an average of 30% reduction in SKAP levels at the kinetochore (Fig. 3E). To evaluate the precise defects in chromosome movement in the SKAP-depleted cells, we conducted time-lapse imaging of chromosome movements in HeLa cells stably expressing GFP-histone 2B. We first treated the cells with the Eg5 inhibitor monastrol for 1 h and then released cells into normal medium. Although chromosomes reach a metaphase alignment within 20 min in scramble siRNA-transfected prophase cells (Fig. 3F), the chromosomes never achieve metaphase alignment in SKAP-suppressed cells even 1 h later (Fig.  3G, arrows). Thus, consistent with other groups (19,20), the absence of SKAP does not affect progression through interphase, but it does cause a long term prometaphase arrest with chromosome alignment defects.
SKAP Is a Microtubule-binding Protein and Cooperates with CENP-E in Kinetochore Microtubule Attachment-Because SKAP is located at the kinetochore-microtubule interface (Fig.  2D), we asked whether depletion of SKAP affects the attachment and stability of spindle microtubules to kinetochores. The control and SKAP siRNA-treated cells were briefly chilled at 4°C before being extracted, fixed, and stained with anti-tubulin antibody and ACA. Although most non-kinetochore-associated microtubules were lost, images obtained by optical sectioning with deconvolution microscopy showed that centromeres labeled by ACA remained associated with spindle microtubule ends in control cells (Fig. 4A, red dots, insets 1 and 2). In SKAP-depleted cells, many chromosomes were misaligned and scattered around the spindle pole (Fig. 4B, inset 3), with kinetochore fibers relatively stable as compared with those of control cells. We also found several examples in which both kinetochores were apparently attached to the same microtubule bundle in cells with reduced SKAP levels (Fig. 4B, inset 3), reminiscent of what was seen in the kinetochore-microtubule attachment in CENP-E inhibitor syntelin-treated cells (15). Although cold-resistant attachment of kinetochores to spindle microtubules can be achieved for a majority of chromosomes when SKAP levels are diminished, the abundance of chromosomes without a stable bi-oriented kinetochore attachment even after long periods of mitotic arrest suggests that SKAP must have an important function in the bi-orientation and/or subsequent orchestration of dynamic microtubule plus-end association with kinetochore. Thus, the perturbation of accurate bi-oriented attachment and subsequent chromosome congression/alignment is a specific consequence of the reduction in SKAP.
To determine whether the SKAP has a direct role in forming kinetochore-microtubule attachments, we expressed recombinant SKAP in bacteria and assessed for its binding activity on microtubules. To this end, samples of SKAP alone, tubulin alone, or varied molar ratios of SKAP/tubulin were incubated for 10 min in polymerization buffer and centrifuged as described previously (11,16). Equal volumes of supernatant (S) and pellet (P) fractions were resolved by electrophoresis and visualized by Coomassie Blue. SKAP concentration was fixed (0.5 M), and tubulin was varied (2-20 M). The Coomassie Blue-stained gel in Fig. 4C shows that recombinant SKAP cosedimented with taxol-stabilized microtubules in a concentration-dependent manner. No pellet of SKAP was observed in the absence of microtubules or with a histidine tag (GFP-histidine; data not shown). Quantifying the fraction of SKAP pelleted with microtubules indicates an apparent dissociation constant of K d ϳ0.68 Ϯ 0.05 M from four separate experiments. GFP-SKAP is readily apparent at the kinetochores as pairs of resolved double dots are similar to those of CENP-E (arrows). High magnification of selected kinetochore demonstrates super-imposition of CENP-E and SKAP labeling. Bar, 10 m. C, SKAP is located to the interface between spindle microtubules and the kinetochore. HeLa cells grown on coverslips were pre-extracted and fixed. The visualization of SKAP was achieved by 10-nm gold-conjugated goat anti-rabbit IgG. Low magnification of a prometaphase HeLa cell is shown. Bar, 1 m. D, at a higher magnification, it is readily apparent that seven 10-nm gold particles representing SKAP position decorate the interface between kinetochores and attached spindle microtubules (arrows). Bar, 100 nm.
Therefore, we conclude that SKAP binds to microtubules directly with relatively low affinity.
Because the C-terminal CENP-E possesses microtubule binding activity (37) and interacts with SKAP (Fig. 1), we tested whether CENP-E facilitates the binding of SKAP to microtubules. The Coomassie Blue-stained gel in Fig. 4D shows that both recombinant CENP-E and SKAP are cosedimented with polymerized microtubules. No pelleting of either protein was observed in the absence of microtubules. Addition of CENP-E synergizes microtubule binding of SKAP (Fig. 4D, compare  inset 1 with 2). Quantifying the fraction of SKAP pelleted with microtubules in the presence of CENP-E demonstrates a synergism in microtubule binding activity, with an apparent dissociation constant of K d ϳ0.17 Ϯ 0.03 M (Fig. 4E). If SKAP cooperates with CENP-E to constitute the kinetochore-microtubule interface in vivo, depletion of CENP-E should further release the tension exerted on the sister kinetochores. As the interkinetochore distance on sister chromatids has previously been proposed as an accurate reporter for judging tension developed across the kinetochore pair (6, 38), we measured this distance in Ͼ100 kinetochore pairs in which both kinetochores were in the same focal plane, in SKAP-depleted, CENP-E-SKAP-depleted, and control cells (Table 1). Control kinetochores exhibited a separation of 1.67 Ϯ 0.16 m, whereas the distance between misaligned kinetochores in SKAP cells was 1.37 Ϯ 0.21 m, which is significantly reduced compared with those of scramble-transfected cells (p Ͻ 0.05). The inter-kinetochore distance of misaligned kinetochores in CENP-E-and SKAP-depleted cells was 1.19 Ϯ 0.15 m (p Ͻ 0.05 compared with those of SKAP-depleted cells). We have also measured the distance in SKAP-depleted and syntelin-treated kinetochores. The interkinetochore distance of misaligned kinetochores in CENP-Einhibited and SKAP-depleted cells was 1.19 Ϯ 0.15 m (p Ͻ 0.05 compared with those of SKAP-depleted cells). The distance in SKAP-depleted and nocodazole-treated cells, in which kinetochore pairs were presumably under no tension, was 1.01 Ϯ 0.13 m from four separate experiments. We conclude that SKAP interacts with CENP-E to link kinetochore to spindle microtubules.

DISCUSSION
The kinetochore is a complex structure that functions as a molecular machine to power chromosome movement and as a signaling device to govern chromosome segregation and cell cycle control. However, the precise molecular composition of the mammalian kinetochore proteins and their molecular organization of the higher order kinetochore structure have remained elusive. We have identified SKAP that cooperates with the mitotic motor CENP-E in linking the kinetochore to spindle microtubules. Our studies demonstrated that in the absence of SKAP, cells remain in mitosis for extended periods, fail to establish stable bipolar attachments to chromosomes, and develop syntelic attachment to spindles, a similar phenotype seen in the absence of CENP-E motor activity either by siRNA or chemical inhibition by syntelin (6,15). A central characteristic of the kinetochore-spindle interface is its ability to orchestrate stable and dynamic associations while bound microtubules are polymerizing or depolymerizing. Such properties would be best coordinated by several distinct but cooperative kinetochore-microtubule-binding sites (21).
One critical feature of this SKAP-CENP-E interaction serves as a fine molecular switch of kinetochore-microtubule attach- As a control, a control, lanes 11 and 12 contain tubulin only without SKAP. D, synergy in association with SKAP with microtubules after interaction with CENP-E. Coomassie-stained gel of microtubule binding assay of recombinant SKAP, CENP-E, and a CENP-E⅐SKAP complex is shown. S and P denote supernatant and pellet fractions, respectively. E, determination of binding affinity for SKAP bound to microtubules after interaction with CENP-E. Equal molar amount of SKAP was mixed with CENP-E before incubation with microtubules for cosedimentation assay. SKAP in supernatant and pellet fractions was judged by Western blot with an anti-SKAP antibody. Binding was quantified and expressed as a percentage of SKAP found in the pellet fraction in the presence (red line) and absence of CENP-E (blue line). Data were plotted from four independent experiments as mean Ϯ S.D.

TABLE 1 SKAP cooperates with CENP-E to control the tension between sister kinetochores
Data were obtained from Ͼ100 kinetochore pairs in which kinetochores were in the same focal plane. ments to enable the simultaneous satisfaction of the spindle assembly checkpoint and accurate chromosome segregation. On the one hand, reduced kinetochore-microtubule attachment stability delays the spindle assembly checkpoint satisfaction by preventing a stable localization of protein phosphatase 1␥ (e.g. 15). On the other hand, static and hyperstable attachments would compromise the correction of aberrant kinetochore-microtubule attachments and lead to chromosome missegregation. Recent studies show that SKAP interacts with astrin in the kinetochore essential for accurate mitosis (18,19). Astrin-SKAP-depleted cells fail to maintain proper chromosome alignment, resulting in a spindle assembly checkpoint-dependent mitotic delay (18). Astrin and SKAP bind directly to microtubules and are required for CLASP localization to kinetochores (18). It was surprising that astrin was not recovered from our CENP-E immunoprecipitates ( Fig. 1A and data not shown). Our preliminary gel filtration analysis suggests that astrin is not fractionated with CENP-E, although a fraction of SKAP is enriched in astrin-rich fractions. It would be of great interest in the future to delineate whether and how these regulators of kinetochore microtubule dynamics coordinate to couple microtubule polymerization and depolymerization dynamics to force generation at kinetochores during mitosis. Cheeseman et al. (9) analyzed the biochemical properties of the Caenorhabditis elegans KMN complex and demonstrated that KNL and Ndc80 subcomplexes bind to microtubules independently with low affinity. However, a greater microtubule binding capacity was realized when these two low affinity binding sites were combined in the in vitro reconstitution assay. It has been shown that the Ndc80 complex is required for a stable localization of CENP-E to the kinetochore (30, 31) via a direct CENP-E-Nuf2 contact (16). Thus, it is tempting to speculate that the CENP-E⅐SKAP complex binds to the microtubule plus ends and leads to a synergistic enhancement of the microtubule binding capacity of the kinetochore by interacting with KMN complex, whereas CENP-E exerts its plus-end directed motor activity required for chromosome alignment to the equator (14,32). It is worth noting that SKAP contains a classic SXIP motif that functions as a microtubule plus-end tracking protein (11). Therefore, it would be of great interest to evaluate whether and how SKAP interacts with plus-end microtubules of mitotic spindle during cell division.
Although microtubules appear to terminate at kinetochores in many electron microscopic images, which leads to the view that kinetochores capture microtubules by end binding but not side attachment, CENP-E binds to the lateral side of microtubules based on the crystal study in which the two CENP-E motors in a CENP-E dimer can only attach to one or two of the 13 proto-filaments of a microtubule (33). Thus, the synergistic interaction between SKAP and CENP-E in vitro may represent the cooperative nature of kinetochore-microtubule interface in vivo. In the future, it will also be important to identify and categorize different microtubule binding activities that exist at the kinetochore and to determine how they orchestrate the dynamic kinetochore-microtubule interactions in mitosis. Perturbation of this orchestration would mostly affect chromosome alignment and stability during bipolar spindle formation, thereby inducing a loss of chromosomes, as seen in CENP-E deficient cells (34 -36). Given the fact that multiple proteins such as BubR1, Nuf2, and Septin7 interact with the CENP-E tail, future studies should map the precise binding interfaces for the aforementioned interactions so that perturbation with competitive peptides can be applied for targeted interventions of specific protein-protein interactions. In addition, it would of great importance to examine whether the CENP-E-SKAP interaction regulates spindle checkpoint silencing as CENP-E operates the checkpoint satisfactorily by targeting PP1␥ to the kinetochore during mitosis (15,39). This can be done by examining the temporal dynamics of kinetochore-localized kinases in real time mitosis (40).
In summary, we show that SKAP interacts with CENP-E, and this interaction is essential for temporal regulation of kinetochore microtubule attachments. We propose that the functional collaboration of CENP-E and SKAP is necessary in governing dynamic microtubule-kinetochore attachment, maintaining tension, and satisfying the spindle assembly checkpoint upon correction of aberrant attachments (Fig. 5). It is likely that all of the kinetochore outer plate proteins interact to orchestrate a functional kinetochore during chromosome segregation. This includes a recently identified PLK1-elicited phospho-regulation of MCAK in mitotic chromosome move- Cytoplasmic dynein mediates initial lateral attachment of mono-oriented chromosomes to spindle microtubules and rapid translocation of chromosomes to the adjacent pole. Ndc80 complex is a major constituent between spindle microtubule and kinetochore link (9,10). SKAP-CENP-E interaction is required for efficient and consistent alignment of the chromosome to the equator of a spindle for achieving and/or maintaining alignment due mainly to its cooperative regulation of microtubule plasticity and dynamics. CENP-E is responsible for efficient localization of SKAP to the kinetochore, which is required for bi-orientation and coordinated oscillation of kinetochore pairs near the equator by regulating kinetochore microtubule plus-end dynamics. Mitotic kinases such as PLK1 and BubR1 govern kinetochore microtubule dynamics and spindle checkpoint satisfaction by phosphorylation of microtubule regulators such as MCAK and SKAP. Our recent study shows that polo-like kinase phosphorylates MCAK by which the microtubule polymerase activity is promoted for efficient kinetochore attachment error correction (28). The PLK1mediated phospho-regulation is also implicated in regulation of kinetochoremicrotubule attachment and SKAP distribution (20).
ments. The SKAP-CENP-E interaction established here is a core of this huge complex, which links kinetochore structural components to spindle microtubule attachment in the centromere. The interaction of SKAP with astrin and Aurora B, which has been reported recently, also supports our working model (19,20).