Oncogene (2006) 25, 6901–6914 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc ORIGINAL ARTICLE Hec1 sequentially recruits Zwint-1 and ZW10 to for faithful segregation and spindle checkpoint control

Y-T Lin1,2, Y Chen2,GWu1 and W-H Lee1

1Department of Biological Chemistry, University of California, Irvine, CA, USA and 2Department of Molecular Medicine, University of Texas Health Science Center, San Antonio, TX, USA

Faithful is essential for main- Introduction taining the genomic integrity, which requires coordination among , kinetochores, centrosomes and Accurate chromosome segregation is essential for spindles during . Previously, we discovered a novel maintaining the integrity of genome. In eukaryotic cells, coiled-coil , highly expressed in cancer 1 (Hec1), this process is precisely controlled by coordination which is indispensable for this process. However, the among the centrosomes, spindles, kinetochores and precise underlying mechanism remains unclear. Here, we chromosomes during the M phase progression. If the show that Hec1 directly interacts with human ZW10 process goes awry, chromosome segregation in indivi- interacting protein (Zwint-1), a binding partner of Zeste dual cellswill proceed with poor fidelity and the cellscan White 10 (ZW10) that is required for chromosome gain or lose the whole chromosomes. Such cells will motility and spindle checkpoint control. In mitotic cells, either die or survive with aneuploidy. Surviving cells will Hec1 transiently forms complexes with Zwint-1 and ultimately proliferate without a proper regulation and ZW10 in a temporal and spatial manner. Although the progressively accumulate more gross chromosomal three have variable cycle-dependent expres- abnormalitiesto form a tumor (Zhou et al., 1998; sion profiles, they can only be co-immunoprecipitated Dobles et al., 2000; Fodde et al., 2001; Kaplan et al., during M phase. Immunofluorescent study showed that 2001; Michel et al., 2001; Cleveland et al., 2003). Indeed, Hec1 and Zwint-1 co-localize at kinetochores beginning at most solid tumors are aneuploidy and have chromosome and that ZW10 joins them later at prometa- instability (CIN) phenotype (Lengauer et al., 1997; phase. Depletion of Hec1 impairs the recruitment of both Takahashi et al., 1999). Understanding the molecular Zwint-1 and ZW10 to kinetochores, while depletion of mechanism involved in accurate chromosome segrega- Zwint-1 abrogates the localization of ZW10 tion will provide an insight into the generation of the but not Hec1. The results suggest that the localization of chromosome instability phenotype in tumor cells. Hec1 at kinetochores is required for the sequential Kinetochoresare composedof multiple proteins recruitment of Zwint-1 and ZW10. Disrupting this associated with the centromeric heterochromatin recruitment by inhibiting the expression of Hec1 or DNA. They physically link chromosomes to the spindle Zwint-1 causes chromosome missegregation, spindle apparatusto mediate a poleward chromosome motion. checkpoint failure, and eventually cell death upon A fully functional kinetochore requiresa complicated cytokinesis. Taken together, these results, at least in yet highly organized assembly of distinct subcomplexes part, provide a molecular basis to explain how Hec1 plays (Cheeseman et al., 2002b). The kinetochore proteins a crucial role for spindle checkpoint control and faithful include intrinsic components such as CENP-A, -B, and chromosome segregation. -C, which bind to the throughout the cell Oncogene (2006) 25, 6901–6914. doi:10.1038/sj.onc.1209687; cycle (Earnshaw and Rothfield, 1985; Earnshaw et al., published online 29 May 2006 1987; Palmer et al., 1987; Craig et al., 1999) and transient components, which come to kinetochores Keywords: mitosis; kinetochores; spindle checkpoint; transiently during different mitotic phases (Craig et al., chromosome segregation and 1999; Cleveland et al., 2003). Kinetochoresplay a key role in establishing bipolar-microtubule attachment, which depends on those proteins involved in stable kinetochore (kMTs) formation, such as highly expressed in cancer 1 (Hec1), Nuf2, Spc24 and Spc25 (Desai et al., 2003; McCleland et al., 2003; Bharadwaj et al., 2004; Deluca et al., 2005). The lack of Correspondence: Dr W-H Lee, Department of Biological Chemistry, bipolar-microtubule attachment or tension on kineto- University of California, Irvine, 124 Sprague Hall, Irvine, CA 92697- chorescan activate the spindle checkpoint to delay or 4037, USA. E-mail: [email protected] arrest cell cycle progression until every chromosome has Received 14 December 2005; revised 3 April 2006; accepted 12 April an appropriate bipolar attachment (Li and Nicklas, 2006; published online 29 May 2006 1995; Rieder et al., 1995). A group of evolutionarily Hec1 recruits Zwint-1 and ZW10 to kinetochores Y-T Lin et al 6902 conserved proteins, including Bub1, Bub3, BubR1, Nuf2 complex serves as a scaffold at the kinetochore for Mad1, Mad2 and Mps1 that are known to be essential binding of these proteins (Martin-Lluesma et al., 2002; for spindle checkpoint functions, also localize at McCleland et al., 2003; Meraldi et al., 2004). kinetochores(Hoyt et al., 1991; Li and Murray, 1991; However, how Hec1 playsa critical role at kineto- Weiss and Winey, 1996; Chan et al., 1999; Cleveland chores in mediating faithful chromosome segregation et al., 2003). Furthermore, higher eukaryotic cellshave remainsunclear. Interestingly,one of Hec1 interacting additional spindle checkpoint proteins, such as human proteins, Zwint-1, isolated from the original yeast two- ZW10 interacting protein (Zwint-1), Zeste White 10 hybrid screening, was subsequently shown to interact (ZW10) and Rod, to ensure high fidelity of chromosome with ZW10, the human homologue of Zeste segregation (Basto et al., 2000; Chan et al., 2000; White 10 (Starr et al., 2000; Wang et al., 2004). Zwint-1 Williams et al., 2003; Wang et al., 2004). appearsto localize at kinetochoresduring prophase, Hec1 wasoriginally identified through itsinteraction followed by the relocation of ZW10, which either with retinoblastoma protein (Rb) and it localizes at recruitsdynactin and to facilitate chromosome kinetochores(Chen et al., 1997a). Inactivation of Hec1 movement or regulatesspindle checkpoint (Starr et al., by microinjection with a specific antibody leads to a 1998; Chan et al., 2000; Savoian et al., 2000; Sharp et al., severe chromosome missegregation and, consequently, 2000). Thus, it is likely that Hec1 serves as a recruiter at the death of daughter cellsin the following cycle (Chen kinetochores for facilitating chromosome segregation et al., 1997a). Hec1 isconservedevolutionarily from and monitoring spindle checkpoint. In this report, we yeast to human (Zheng et al., 1999; He et al., 2001). The show that Hec1 interacts directly with Zwint-1, which in Saccharomyces cerevisiae homologue of Hec1, also turn bindsZW10. Furthermore, Hec1 isrequired for the known asNdc80/TID3/YIO4, wasinitially character- recruitmentsof Zwint-1 and ZW10 to kinetochores. ized as a protein associated with the spindle pole body Disruption of Zwint-1 and ZW10 recruitment results in and later found to be a kinetochore protein (Wigge chromosome missegregation, spindle checkpoint failure et al., 1998; Zheng et al., 1999). The scHec1 null and subsequently cell death upon cytokinesis. These mutation is lethal but can be rescued by the expression results suggest that Hec1 may participate in the dynamic of human Hec1 (hsHec1) (Zheng et al., 1999). Inactiva- recruitmentsfor capturing key regulatory proteinsonto tion of hsHec1 by a temperature sensitive mutation in kinetochores for faithful chromosome segregation. budding yeast leads to a chromosome missegregation and cell death at the non-permissive temperature; the phenotypes observed in these yeast cells are very similar Results to those observed in mammalian cells injected with neutralizing anti-Hec1 antibodies(Chen et al., 1997a; Hec1 binds ZW10 through Zwint-1 Zheng et al., 1999). These results strongly suggested that Using the coiled-coil region of Hec1 as bait, we Hec1 is an essential protein for proper chromosome previously identified a panel of Hec1-interacting pro- segregation. teins through a yeast two-hybrid screening (Zheng et al., How Hec1 executesitsfunction in chromosome 1999). One such protein was Zwint-1, a known human segregation has been an interesting subject to pursue. ZW10 binding partner (Chan et al., 2000). To examine AsHec1 containsthree coiled-coil domainsthat are the molecular domainsmediating the Zwint-1/Hec1 frequently involved in protein–protein interaction, we interaction, we employed two complementary ap- have used yeast two-hybrid screening and identified proaches. We first performed yeast two-hybrid assays several Hec1-interacting proteins, including structural by using various coiled-coil containing Hec1 derivatives components of mitotic chromosomes, Smc1 and Smc2, fused with the Gal4 DNA-binding domain (DBD) and which have important roles in chromosome cohesion full-length Zwint-1 fused with the Gal4 transactivation and condensation; three subunits of the 26S proteasome domain (TA). Fusion protein containing the first (aa including MSS1 (CIM5/subunit 7), p45/Trip1 (Sug1/ 251–431) or second (aa 361–547) coiled-coil region of CIM3/subunit 8), and p44.5 (subunit 9) and Nek2A, a Hec1, wassufficient for Zwint-1 binding (Figure 1A). In mitotic kinase related to fungal NimA (Chen et al., contrast, the third coiled-coil region (aa 547–618) was 1997a, 2002; Zheng et al., 1999). The functionsof these not able to bind Zwint-1. Full-length Hec1 wasnot associated proteins strongly suggest that Hec1 plays a tested due to the auto-transactivation activity in this critical role in regulating mitosis. As expected, Hec1 assay (data not shown). To test whether Zwint-1 modulatesthe proteolysisofmitotic cyclinsby nega- interactswith Hec1 or ZW10 directly, we then tively regulating MSS1 (Chen et al., 1997b) and binds performed the in vitro glutathione S-tansferase (GST) Smc1p and Smc2p to facilitate sister chromatid cohesion pull down assays by using a panel of GST-tagged Hec1 and chromosome condensation (Zheng et al., 1999). or ZW10 variantspurified from bacteria (Figure 1B–D). Interestingly, Nek2A phosphorylates Hec1 on serine165 As shown in Figure 1C, Zwint-1 specifically binds GST- that iscritical for the cellular function of Hec1 sincethe Hec1-C (aa 251–618) and GST-Hec1-D (aa 56–642) but phosphorylation-defective mutant fails to rescue the not GST-Hec1-A (aa 56–128) or GST-Hec1-B (aa 1– scHec1 null phenotype (Chen et al., 2002). Furthermore, 360). Taken together, these results suggest that the Hec1 formscomplex with Nuf2 that isrequired for the Zwint-1/Hec1 interaction isprimarily mediated by the kinetochore localization of spindle checkpoint proteins second coiled-coil region of Hec1 (aa 361–431). In such as Mad1, Mad2 and Mps1, suggesting that Hec1- addition, Zwint-1 also bound GST-ZW10 in a similar

Oncogene Hec1 recruits Zwint-1 and ZW10 to kinetochores Y-T Lin et al 6903 in vitro binding assay, consistent with the previous ZW10 bind the domain A of Zwint-1. Zwint-1 failed in report (Starr et al., 2000) (Figure 1C). binding itself in the in vitro binding assay, suggesting We next performed additional experimentsto confirm that a domain A monomer might be sufficient to bind the binding capability of Zwint-1 towardsHec1 and both Hec1 and ZW10 (Figure 1D). ZW10. GST tagged full-length Zwint-1 and itscoiled- To test whether Hec1 directly binds ZW10, purified coil domainsA (aa 80–155) and B (aa 170–220) were GST-Hec1-D and GST-ZW10 were incubated with 35S- prepared and incubated with the in vitro 35S-methionine- methionine-labeled ZW10 and Hec1, respectively. No labeled Hec1, Zwint-1, and ZW10 (Figure 1B and D). detectable binding between Hec1 and ZW10 was The coiled-coil domain A exhibited stronger binding observed. As Zwint-1 interacts with both Hec1 and activity towardsHec1 or ZW10 than domain B ZW10 (Figure 1C and D), it islikely that Hec1 (Figure 1D). We further tested whether the dimerization associates with ZW10 through Zwint-1. To test this or multimerization of Zwint-1 isrequired for its possibility, 35S-methionine-labeled ZW10 wasmixed interaction with Hec1 and ZW10 since both Hec1 and with the purified His-Zwint-1 followed by the addition of GST-Hec1-D. Asshownin Figure 1E, ZW10 bound GST-Hec1-D only in the presence of His-Zwint-1. Reciprocally, 35S-methionine-labeled Hec1 bound GST- ZW10 only in the presence of His-Zwint-1. These results suggest that Zwint-1 bridges Hec1 and ZW10 for complex formation, at least in vitro.

Hec1, Zwint-1 and ZW10 co-localize at kinetochores and co-immunoprecipitate in cell lysates prepared from M phase The in vitro interaction of the above three proteins suggests that they may interact in vivo. To test this possibility, mitotic spreads of T24 cells were prepared for immunostaining with anti-Hec1, anti-Zwint-1, and anti-ZW10 antibodies. Hec1, Zwint-1 and ZW10 were co-stained at kinetochores, suggesting that they are co- localized at kinetochores of M phase chromosomes, consistent with the previous observation (Figure 2A) (Chen et al., 1997a; Starr et al., 2000). To test whether

Figure 1 Mapping interaction domainsof Hec1, Zwint-1 and ZW10. (A) Determination of the Zwint-1-binding region of Hec1 by yeast two-hybrid assay. Deletion mutants of Hec1 were tagged with the GAL4 DNA binding domain (Gal4-DBD) and the full length Zwint-1 with the GAL4 transactivation domain (Gal4- TAD). The fold-increase in b-galactosidase activity compared to the host strain Y153 was indicated. Assays were performed in triplicate for each transformation. (B) Schematic diagram of a panel of GST fusions used in the in vitro binding assay. (C) In vitro binding assay. Comparable amount of GST (lane 2), GST-tagged Hec1 fragments(lanes3–6) or ZW10 (lane 7) were prepared from bacteria and incubated with the 35S-methionine-labeled Zwint-1, and the results of autoradiograph after SDS-PAGE analysis were shown. Lane 1, input of the in vitro translated Zwint-1. (D). Comparable amount of purified GST (lane 2) and GST-Zwint-1 (wt or variants, lanes 3–5) shown in the top panel were used to bind the in vitro translated 35S-methionine-labeled Hec1, ZW10 or Zwint-1. Lane 1, input of the in vitro translated Hec1, ZW10, or Zwint-1. (E) Zwint-1 bridgesHec1 and ZW10 for interaction. GST- Hec1–15D or GST-ZW10 were used to bind the in vitro translated 35S-methionine-labeled ZW10 and Hec1, respectively, in the presence ( þ ) or absence (À) of His-Zwint-1. (a) 1/3 input amounts of the GST fusions shown in Coomassie blue staining. (b) Western blot of His-Zwint-1 with anti-His antibody. Lane 1, 1/40 of the total input of His-Zwint-1. (c) Autoradiography to detect 35S- methionine-labeled-ZW10 and -Hec1 pulled down by GST-Hec1 and GST-ZW10, respectively, in the absence or presence of His- Zwint-1. Note that GST alone did not bind to His-Zwint-1 or Hec1 (lane 6). Lane 2 and 5, 1/10 of the total input of the translated ZW10 and Hec1, respectively.

Oncogene Hec1 recruits Zwint-1 and ZW10 to kinetochores Y-T Lin et al 6904 they actually interact in cells, we performed co- anti-Zwint-1 or anti-ZW10 antibody brought down immunoprecipitation (co-IP) assays by using unsyn- Hec1, albeit less evidently (Figure 2B). To further chronized T24 cells. Immunoprecipitation with either clarify the in vivo interactions, T24 cells synchronized at various cell cycle stages were used for the co-IP assays. The interactions of Hec1 with either ZW10 or Zwint-1 were most evident during the G2/M phases (Figure 2B). During the cell cycle progression, the levels of Hec1 and Zwint-1 gradually increase from S to M phase and peaks at G2/M, while ZW10 is expressed ubiquitously throughout the entire cell cycle. The approximate cell cycle stages were indicated by the RB phosphorylation pattern as described previously (Chen et al., 1989). To consolidate the above observation, HeLa cells were arrested at prometaphase and then allowed to enter the next cell cycle upon release. Cells were then harvested at various time points for the co-IP experi- ment. Asshownin Figure 2C, Zwint-1 and ZW10 were co-IPed with Hec1 only during the early time pointspost release when the cells were presumably still in M phase asjudged by the phosphorylation patternsof Rb and Histone H3 and the presence of cyclin B (Figure 2C). Taken together, these results suggest that Hec1, Zwint-1 and ZW10 interact specifically during the G2/M phases of the cell cycle, probably at, but not limited to the kinetochores. This observation is consistent with the Xenopus study for Hec1 and Zwint-1 (Emanuele et al., 2005).

Sequential localization of Hec1, Zwint-1 and ZW10 at kinetochores during M phase progression To further examine the dynamic nature of Hec1 at kinetochoresduring cell cycle progression,we marked cellswith an antibody againstH3P, enabling the distinction among G1–S, G2 and M phases (Wei et al., 1999; Hsu et al., 2000). Cells stained with H3P showed

Figure 2 Interactionsamong Hec1, Zwint-1, and ZW10 in cellsby immunostaining and co-IP. (A) Hec1, Zwint-1, and ZW10 co- localize at kinetochoresof metaphasechromosomesin T24 cellsby immunostaining. Chromosome spreads were co-stained with the indicated antibodiesand the merged imageswere shownin the right column. Higher magnification of the selected kinetochore images were shown (inset). (B) co-IP assay for Hec1, Zwint-1, and ZW10 in T24 cells. T24 cells enriched at G0 by density arrest were allowed to enter the cell cycle upon re-splitting and then harvested for co-IP assay at various time points after release. Anti-Zwint-1 and anti- ZW10 antibodieswere usedfor co-IP followed by immunoblotting against Hec1, Zwint-1 and ZW10. Straight Western blotting displayed the protein level of Hec1 and p84 (a nuclear matrix protein, loading control). Rb phosphorylation serves as a convenient marker for cell cycle progression. For each lane, B107 cellswere usedfor co-IP reaction and the experimentswere repeated twice. (C) Hec1, Zwint-1, and ZW10 were co-immuno- precipitated from HeLa cells synchronized at M phase. HeLa cells arrested at prometaphase were allowed to resume the cell cycle and then collected at varioustime pointspostthe nocodazole with- drawal for co-IP assay with the anti-Hec1 antibody (top panel). Zwint-1 (top down 2nd) or ZW10 (top down 3rd) in the Hec1 immune complex were shown by Western blot. Hec1, Zwint-1, and ZW10 expression profiles and the cell cycle markers (Rb and Histone H3 phosphorylation patterns and cyclin B protein level) were also shown by probing with respective antibodies.

Oncogene Hec1 recruits Zwint-1 and ZW10 to kinetochores Y-T Lin et al 6905 discrete speckles in their nuclei at G2 phase, homo- nor ZW10 wasdetected at kinetochores(Figure 4e). In genous nuclear staining at prophase, and homogenous cellstreated with Zwint-1 siRNA, Hec1 remained at chromosomal staining at prometaphase to . kinetochores, but neither Zwint-1 nor ZW10 did Hec1 appeared at kinetochoresfrom late prophaseto (Figure 4f). In mitotic cellstreated with control Luc anaphase but disappeared at telophase (Figure 3a). siRNA, similar to untreated mitotic cells, Hec1, Zwint-1 Interestingly, a minor fraction of Hec1 was also and ZW10 were all localized at kinetochores observed at spindle poles and the proximal spindle (Figure 4d). A polyclonal CREST antibody that microtubules (Figure 3a and b), in part consistent with recognizesthe inner core kinetochore proteinssuchas the observations in yeast and chicken cells (Wigge et al., CENP-A, -B, and -C (Earnshaw and Rothfield, 1985) 1998; Hori et al., 2003). During mitosis, Zwint-1 was co- wasable to mark kinetochoresin all siRNAexperi- localized with Hec1 at kinetochoresfrom prophaseto ments, indicating these three proteins are not required anaphase in more than 100 mitotic cells examined for the primary assembly of kinetochores. Taken (Figure 3b and c), while ZW10 wasobservedat together, these results suggest that Hec1 is essential for kinetochores from prometaphase to only. the sequential recruitment of Zwint-1 and ZW10 to Intriguingly, ZW10 decorated the spindle poles as well kinetochores. as their proximal microtubules at metaphase, suggesting distinct yet potentially interrelated roles for each Inactivation of Hec1 and Zwint-1 by siRNA leads to individual protein at various stages of mitosis chromosome missegregation and eventual cell death (Figure 3c) (Starr et al., 1997). Taken together, Hec1 Previous studies showed that the inactivation of Hec1 and Zwint-1 appear to arrive at kinetochoresprior to leads to severe chromosome missegregation and cell ZW10, suggesting that Hec1 and Zwint-1 may play a death (Chen et al., 1997a; Zheng et al., 1999). Never- role in recruiting ZW10 to kinetochores. theless, the details of chromosome behavior in live cells have not yet been shown. To achieve this goal, we Localization of Zwint-1 and ZW10 at kinetochores established HeLa cells stably expressing histone 2B requires the presence of Hec1 fused with a green fluorescence protein (named HeLa/ The assembly of yeast kinetochores appears to be built H2B-GFP) by retrovirusinfection. TheseH2B-GFP upon a hierarchy of dependencies(He et al., 2001). Asa tagged cellsallowed usto directly monitor the nuclear result of the evolutionary conservation, mammalian dynamicsand the chromosome movementsduring M kinetochoresmay alsohave thischaracteristic(Kitaga- phase under fluorescence microscope (Kanda et al., wa and Hieter, 2001). To determine the hierarchy for the 1998). The effectsof individual siRNAon aberrant recruitment of these three proteins to kinetochores, we chromosome behaviors (misaligned chromosomes at used short interfering RNA (siRNA) to knockdown metaphase and premature chromosome segregation at Hec1 and Zwint-1 protein expressions. Depletion anaphase), formations of multinuclei and micronuclei, of Hec1 and Zwint-1 expression was accomplished and cell death were recorded. A high percentage of cells by transfection with 21 nucleotides, double-stranded with lagging chromosomes and incompletely segregated siRNAs targeting to unique sequence of Hec1 or Zwint- chromosomes were observed, most prominently at 56 h 1 (Elbashir et al., 2001). Transfection of specific siRNA after Hec1 siRNA transfection (Figure 5A). Micronuclei into HeLa cellssignificantly reduced Hec1 and Zwint-1 and multinuclei formation were prominent at later time protein expression, while the control siRNA targeting pointsabout 68–80 h (Figure 5B). At 92 h, 51% of Hec1- firefly luciferase (Luc) had little or no influence on the depleted cellswere dead (Figure 5C). Cellsdepleted with expression of either Hec1 or Zwint-1 (Figure 4a and b). Zwint-1 had similar but much less severe phenotypes The depletion effect started at 20 h, was most prominent (Figure 5A–C). In both parental or transfected with at 44–56 h (data not shown) and maintained as long as luciferase siRNA cells, a background level of abnormal 80 h after transfection. The amount of proteins not mitosis and interphase was seen. These results indicate targeted by the specific siRNA remained comparable, that depletion of Hec1 or Zwint-1 leadsto misaligned except that the Zwint-1 protein level wasreduced to chromosomes at metaphase, lagging chromosomes at 57% in Hec1 siRNA-targeted cells. However, in Zwint-1 anaphase, and eventual cell death. The effects of Hec1- siRNA-treated cells, Hec1 protein level was not reduced and Zwint-1-depletion were consistent with those (Figure 4c). To further test whether Zwint-1 is necessary observed in Drosophila mutantsthat the lagging for the interaction between Hec1 and ZW10, we chromosomes and premature chromosome segregation performed the co-IP experiment with anti-Hec1 anti- arise from the attenuated poleward chromosome move- body using lysates from Zwint-1 siRNA-treated cell. As ment and spindle checkpoint defects (Williams et al., shown in Figure 4c right panel, ZW10 failed to co- 1992; Williamsand Goldberg, 1994; Starr et al., 1997; immunoprecipiate with Hec1. These data further sup- Basto et al., 2000). port the notion that Hec1’sbinding to ZW10 isbridged by Zwint-1. Next, the individual siRNA-treated cells were stained Depletion of Hec1 prolongs M phase duration in HeLa with corresponding antibodies to locate Hec1, Zwint-1 cells and ZW10. Asexpected, the localization of Hec1 at As Hec1-depleted cells had severe chromosome segrega- kinetochoreswasabolished in cellstreated with Hec1 tion errors, we examined the mitotic details of each siRNA. Interestingly, in the same cells, neither Zwint-1 individual cell by performing time-lapse imaging of live

Oncogene Hec1 recruits Zwint-1 and ZW10 to kinetochores Y-T Lin et al 6906

Oncogene Hec1 recruits Zwint-1 and ZW10 to kinetochores Y-T Lin et al 6907 HeLa/H2B-GFP cellstreated with Hec1 siRNA. Of all prometaphase-like stage, which may have transiently the control cellstreated with Luc siRNA( n ¼ 60), 73% provoked the spindle checkpoint control to arrest cells completed mitosis within 2 h and divided into two at M phase. However, these cells underwent cytokinesis normal viable daughter cells(Figure 5D). Two repre- in spite of having many lagging chromosomes (Figure sentative cases were shown in Figure 5E (a and b). Hec1- 5Fa) or even without any chromosome congression and depleted cells( n ¼ 30), however, exhibited a prolonged segregation (Figure 5Fb). After exiting from M phase, M phase ranging 2–8 h, and 83% of them completed cellsexhibited micronuclei or multinuclei and later mitosis in a period of longer than 2 h (Figure 5D). As appeared as dense debris – typical of dying cells. The shown in Figure 5F, depletion of Hec1 caused the defect of chromosome congression was not evident in chromosome congression defect as well as a prolonged Zwint-1- and ZW10-depleted cells, suggesting that Hec1

Figure 4 Hec1 isrequired for Zwint-1 and ZW10 to localize at kinetochores.( a and b) Hec1 and Zwint-1 protein expressions in HeLa cells treated with Luciferase, Hec1 or Zwint-1 siRNA. Cells were harvested at different time points after transfection and analyzed by straight western blot using anti-Hec1, Zwint-1 and p84 antibodies. (c) Binding of Hec1 to ZW10 in cellsrequiresZwint-1. Protein amounts of Hec1, Zwint-1 and ZW10 in individual siRNA-transfected cells were examined at 44 h after transfection (left panel). Zwint- 1 was reduced to 57% in Hec1-depleted cells (lane 2), and the other proteins remained comparable in all three siRNA treated cells. The Luc and Zwint-1 siRNA-treated cells were synchronized at M phase and used for co-IP by anti-Hec1 antibody (right panel). Depletion of Zwint-1 eliminated the co-IP of ZW10 with Hec1 (lane 2). (d–f) Cells at 44 h after transfection were subjected to immunostaining with antibodies against CREST, Hec1, Zwint-1 and ZW10, as indicated. Cells at prometaphase or metaphase were examined for the localization of Hec1, Zwint-1 and ZW10 at kinetochores. DNA was visualized by DAPI staining. Signals from the co-stained subject proteins, along with DAPI, were merged to determine their co-localization (lower panels in d–f). (d) Co-localizationsof Hec1, Zwint-1, and ZW10 at kinetochores were not affected in Luciferase siRNA-transfected cells. Higher magnification images of the kinetochores indicated by arrows were shown (inset). (e) The localizationsof Hec1, Zwint-1 and ZW10 at kinetochoreswere abolishedin Hec1- depleted cells. (f) Hec1, but not ZW10, remained localized at kinetochoresin Zwint-1-depleted cells.

Figure 3 Sequential localization of Hec1, Zwint-1 and ZW10 at kinetochores during M phase progression. Asynchronized HeLa cells were seeded on cover slips and grown for two days. Cells were fixed and stained with a pair of different primary antibodies. Cells at different stages of G2 and M phases are shown (left to right). (a) The H3P staining showed a speckled pattern at G2 phase and homogenous staining of chromosomes during mitosis (middle panel). The kinetochore localization of Hec1 appeared from the late prophase through anaphase (upper panel). (b) The kinetochore localization of Zwint-1 appeared from the late prophase through anaphase (middle panel). (c) The kinetochore localization of ZW10 appeared from the prometaphase through metaphase (upper panel). The localizations of Hec1 and ZW10 at spindle poles and proximal spindle microtubules were also observed at metaphase (arrows, in a–c). DNA was visualized by DAPI staining and shown at lower panels in a–c.

Oncogene Hec1 recruits Zwint-1 and ZW10 to kinetochores Y-T Lin et al 6908 may have additional function for chromosome congres- cells at M phase (Figure 6a). In Luc siRNA treated sion. HeLa/H2B-GFP cells, the number of mitotic cells increased to 80% of total cell population and remained at that level from 18 to 22 h after nocodazole treatment. Depletions of Hec1 and Zwint-1 lead to defects in spindle In Zwint-1 siRNA treated cells, the mitotic index peaked checkpoint at 18 h with 66% of total population but declined Instead of arresting at M phase, Hec1- and Zwint-1- afterwards, concomitantly with the increase of the depleted cells prematurely progressed through M phase multinuclear interphase cells. More severe phenotypes with lagging chromosomes, suggesting that these cells were observed in the Hec1 siRNA-treated cells; where are defective in the spindle checkpoint control mitotic index peaked at 12 h with only 51% of total (Figure 5A). To further support this possibility, we population and the number of multinuclei cellsin- measured mitotic index and the number of interphase creased higher than that of the Zwint-1-depleted cells with multinuclei at subsequent time points (0–24 h) (Figure 6c). The incomplete activation and failure in after treating cellswith 80 ng/ml nocodazole, which is the maintenance of the spindle checkpoint induced by known to depolymerize b-tubulin, disrupt spindle nocodazole in these cells can be attributable to the formation, and activate the spindle checkpoint to arrest reduction of Hec1 or Zwint-1 protein level (Figure 6b), suggesting that Hec1 as well as Zwint-1 are critical for the spindle checkpoint control. It was also noted that only a few cellstreated with Zwint-1 or Hec1 siRNA alone showed multi-nuclei at the early time points (Figure 5b), suggesting that nocodazole treatment enhanced such abnormality seen in Hec1- or Zwint-1- depleted cells.

Zwint-1 is partially required for Mad2 to localize at kinetochores, but not for BubR1 Besides ZW10, other spindle checkpoint proteins are also recruited to kinetochores during M phase, and

Figure 5 Characterization of cellular phenotypesin Hec1 or Zwint-1 siRNA-treated HeLa/H2B-GFP cells. Live HeLa/H2B- GFP cells were monitored under a fluorescent microscope at the indicated time points after siRNA transfection. Chromosome behavior during M phase and nuclear abnormality at interphase were examined and scored. (A) Increase in the percentage of cells with the aberrant mitosis, including misaligned chromosome(s) at metaphase and lagging chromosome(s) at anaphase, after transfec- tion with Hec1 or Zwint-1 siRNA. The metaphase cells (n>200) and anaphase cells (n>200) were examined at each time point for individual siRNA-treated cells. Represented images were shown at the top of 2 histograms: left and right images represented the Zwint-1 and Hec1 siRNA treated cells, respectively. Scale bar, 5 mm. Error bars, s.d. (B) Increase in the percentage of cells with micronuclei and multinuclei after transfection with Hec1 or Zwint- 1 siRNA. As the following consequences of chromosome segrega- tion errorsare micronuclei and multinuclei, cells( n>1000) with micronuclei or multinuclei were scored. Error bars, s.d. (C) Increase in the percentage of cell death after Hec1 or Zwint-1 siRNA transfection. The cell death was observed after Hec1 or Zwint-1 siRNA transfection, and the dead cells were revealed by Trypan blue staining and counted by hemacytometry (n>1000). (D) The duration of M phase in HeLa/H2B-GFP cells after treated with Luc or Hec1 siRNA. Cells were cultured in CO2 independent medium after transfection with the siRNA and photographed using fluorescence and phase-contrast time-lapse imaging at 1-h intervals for 40 h. M phase length was determined from the intervals between chromosome condensation and completion of cytokinesis. (E) The mitotic progression of cells after treated with Luc siRNA. Cells completed M phase within 2 h and entered interphase with one nucleusin both cases( a and b). (F) The mitotic progression of cells after treated with Hec1 siRNA. Cells underwent cytokinesis without proper chromosome segregation: (a) the cell exited M phase with many lagging chromosome (arrows); (b) the cell exited M phase without chromosome congression and segregation. These cellsentered interphasewith multinuclei and eventually died. The numbers indicated the hours passed during the recording.

Oncogene Hec1 recruits Zwint-1 and ZW10 to kinetochores Y-T Lin et al 6909 localized at kinetochoresin Zwint-1-depleted cells,while Mad2 level waspartially reduced at kinetochores (Figure 7a). The average fluorescence intensity of Mad2 in control cells(149.7 U) ismuch higher than that of Zwint-1-depleted cells(42.8 U, Po0.001) (Figure 7b). The overall protein levelsof Mad2 and BubR1 were comparable in Luc and Zwint-1 siRNA- treated cells (Figure 7c). These results suggest that Zwint-1 isrequired for the activation of the Mad2 pathway of the spindle checkpoint control. We speculate that Zwint-1 may function downstream of Hec1 and that the efficient recruitment of Mad2 at kinetochores requiresan intact Hec1/Zwint-1/ZW10 pathway.

Discussion

In thiscommunication, we have shown that Hec1 interactsdirectly with Zwint-1 and co-localizeswith Zwint-1 and ZW10 at kinetochores. Hec1 is required for the sequential recruitment of Zwint-1 and ZW10 to kinetochores during M phase. Disruption of Hec1, Zwint-1 and ZW10 recruitment to kinetochoresresults in chromosome missegregation, failure in full activation of spindle checkpoint, and eventually cell death. These results provide, at least in part, a molecular basis of how Hec1 playsa crucial role in promoting faithful chromo- some segregation as well as spindle checkpoint control.

Hec1 recruits Zwint-1 and ZW10 to kinetochores for chromosome segregation It is known that the kinetochores comprise several layers of structures that are likely assembled in a hierarchical Figure 6 Depletion of Hec1 or Zwint-1 causes the spindle order by multiple copies of distinct protein subcom- checkpoint failure. (a) Top panel showed a schematic for the time course of the experiment. Unsynchronized HeLa/H2B-GFP cells plexes(Brinkley and Stubblefield, 1966; Zinkowski were transfected with the respective siRNA. At 32 h of post- et al., 1991; McEwen et al., 1998; Kniola et al., 2001; transfection, nocodazole (80 ng/ml) was added and maintained for De Wulf et al., 2003). Both Hec1 and Zwint-1 are the 24 h. (b) Western analysis on cells transfected with luciferase, outer kinetochore components, while ZW10 is present at Zwint-1 or Hec1 siRNA and subsequently treated with nocodazole for the designated periods of time. Aliquots of cell lysates were the fibrouscorona (Kitagawa and Hieter, 2001; Chan separated by SDS-PAGE and probed with antibodies against Hec1, et al., 2005). Hec1 and Zwint-1 start to appear at Zwint-1 or p84. p84 served as the internal loading control. kinetochoresat G2/M transitionto prophase,and (c) siRNA treated cells were randomly photographed at the ZW10 joinsthem later at prometaphase(Starr et al., designated time point using fluorescence and phase-contrast 2000). Hec1 is required for the subsequent localization microscopy. (d) The percentagesof cellsin the M phase(mitotic index, top panel, n>1000) and those in the interphase with of Zwint-1 and ZW10 to kinetochores, suggesting a multinuclei (bottom panel, n>1000) were determined. Error bars, temporal mode of kinetochore association for these s.d. moleculesmediated by Zwint-1. Similar conclusions were drawn in two recent studies using different systems (Emanuele et al., 2005; Kops et al., 2005). However, a proper localization at kinetochore iscritical for spindle direct physical interaction between Hec1 and Zwint-1 checkpoint proteinsto execute their function (Amon, was not shown. Intriguingly, in frog system, the 1999; Cleveland et al., 2003). Among them, Mad2 and recruitment of Hec1 to kinetochoresalsorequires BubR1 have been demonstrated to repress anaphase Zwint-1 (Emanuele et al., 2005), while in mammalian entry by inhibiting APC/C and Cdc20 in vitro (Fang system as described herein, the co-dependency of Hec1 et al., 1998; Tang et al., 2001). To assess how Hec1 and and Zwint-1 recruitment at kinetochoreswasnot Zwint-1 are involved in the full activation of the spindle observed. Although the underlying reason remains checkpoint, the kinetochore association of Mad2 and unclear, the biological significance of this recruitment BubR1 were examined. Aspreviouslyreported, the can be at least twofolds. One is to facilitate chromosome kinetochore localization of Mad2, but not BubR1, was motility and stabilize the chromosome attachment at abolished in Hec1-depleted cells (Martin-Lluesma et al., prometaphase and the other is to modulate spindle 2002; Meraldi et al., 2004). Similarly, BubR1 wasstill checkpoint control.

Oncogene Hec1 recruits Zwint-1 and ZW10 to kinetochores Y-T Lin et al 6910

Figure 8 Proposed model for Hec1 function in kinetochores. (A) Hec1 complex (Hec1, Nuf2, Spc24 and Spc25) iscritical for stable kMT attachment at kinetochores. (B) Hec1 recruitsZwint-1, and ZW10, which in turn recruitsdynein/dynactin for proper poleward chromosome movement. (C) Hec1 isinvolved in spindle checkpoint through the recruitment of ZW10 complex and Mad2. The ZW10 complex pathway and Mad2 pathway are converged at Hec1 platform. In the absence of Hec1 and Zwint-1, BubR1 can still localize at kinetochores and transiently arrest cells at M phase. (D) Silencing of mitotic checkpoint upon Mad2 release from kinetochores.

microtubule attachment at kinetochoresleadsto chro- mosome congression failure as observed in Hec1- depleted cells(Figure 5F). If Hec1 and Zwint-1 works together to coordinate thisfunction, a similar attach- ment defect should have occurred in Zwint-1- or ZW10- depleted cells. However, Zwint-1 depletion did not apparently hamper the chromosome alignment. Thus, it suggests that Hec1 plays multiple mitotic roles mediated by distinct downstream effectors, one of which Figure 7 Zwint-1 ispartially required for Mad2 to localize at is Zwint-1. As chromosome missegregation was ob- kinetochores, but not for BubR1. Cells were immunostained with served in both Hec1- and Zwint-1-depleted cells, it is antibodiesagainstZwint-1, Mad2 and BubR1 to determine their localizations after transfection for 44 hrs with the respective possible that the Hec1/Zwint-1/ZW10 association is siRNA. DNA was visualized by DAPI staining. (a) The primarily required for efficient and coordinated chro- kinetochore localization of Mad2 wasdiminished in cellstreated mosome segregation. ZW10 has been shown to recruit with Zwint-1 siRNA, compared to those treated with the control dynactin and dynein, which directly facilitate chromo- Luc siRNA. In contrast, the kinetochore localization of BubR1 some movement during the M phase (Starr et al., 1998; wasnot affected in Zwint-1-depleted cells.( b) Fluorescence intensities of Zwint-1 and Mad2 at individual kinetochores in Savoian et al., 2000; Sharp et al., 2000). The localization control or Zwint-1-depleted cells were quantified and statistically of Hec1, Zwint-1 and ZW10 at kinetochoresmay analysed. The bars represent the fluorescence intensities of Zwint-1 provide a stage to facilitate chromosome poleward and Mad2 in arbitrary unit, respectively. Error bars, s.d. (c) The movement through dynactin and dynein during meta- corresponding cells were harvested and analyzed by western blotting with antibodiesagainstZwint-1, Mad2, BubR1. The phase and anaphase (Figure 8). expression levels of Mad2 and BubR1 proteins were not affected in cellsafter treated with Zwint-1 siRNA. Hec1, Zwint-1 and ZW10 at kinetochores for spindle checkpoint control Inactivation of Hec1 in T24 bladder carcinoma cellsby It wasreported that kinetochoresin Hec1-depleted specific antibody microinjection resulted in disordered cellsfailsto form and/or maintain outer kinetochore chromosome alignment and sister chromatid segrega- structure, thus removing the sites for microtubule end- tion, fragmented nuclei and cell death (Chen et al., on attachment (Deluca et al., 2005). The defect in 1997a). These findings show that cytokinesis still

Oncogene Hec1 recruits Zwint-1 and ZW10 to kinetochores Y-T Lin et al 6911 proceeds despite chromosome segregation errors, which ZW10 and Mad2 at kinetochores, the spindle check- would otherwise activate the spindle checkpoint and point will eventually fail to maintain. Noticeably, in arrest cells during M phase. Similar abnormal pheno- Hec1 and Zwint-1-depleted cellsBubR1 wasrecruited to types were seen in scHec1/NDC80 null yeast rescued by kinetochoresupon activation of the spindle checkpoint. a temperature sensitive mutant of human Hec1 grown at Thus, the transient mitotic arrest in these cells may be the non-permissive temperature (Zheng et al., 1999). A achieved through the BubR1 pathway (Figure 8, C3). recent report demonstrated that cells injected with We have observed a higher incidence of unaligned neutralizing Hec1 antibodiesprematurely exited from chromosomes at metaphase and premature chromosome mitosis and failed to arrest in response to microtubule segregation in Hec1-depleted cells than in Zwint-1- inhibiting drugs, indicating that the complex containing depleted cells, suggesting that Hec1 plays a more crucial Hec1 and Nuf2 isrequired for the spindle checkpoint in role in the chromosome segregation than Zwint-1 does the Xenopus system (McCleland et al., 2003). Consistent (Figure 5A). It is less likely that these phenotypic with thisobservation, othersand we have observeda differencesare due to the difference in the depletion similar phenomenon where Hec1-depletion by siRNA efficiencies of these two proteins. Rather, a plausible leadsto the spindle checkpoint failure (Meraldi et al., explanation isthat Hec1 hasadditional functionssuch 2004). These results, taken together, strongly support as providing an attachment sites for spindle through the previousobservation that inactivation of Hec1 forming separate complex independently of Zwint-1 and causes a spindle checkpoint defect (Chen et al., 1997a; ZW10 (Figure 8A). Similarly, the different strength of Zheng et al., 1999). the spindle checkpoint control observed in Hec1- and How are Hec1 and Zwint-1 involved in the spindle Zwint-1-depleted cells may be resulted from the presence checkpoint control? Recent studies have provided of different checkpoint proteinsat kinetochores.For convincing evidence that Hec1 isrequired for Mad2 example, in Zwint-1-depleted cells, Mad2 is partially recruitment at kinetochoresto allow the efficient spindle localized at kinetochores, while it is completely absent in checkpoint activation (Martin-Lluesma et al., 2002; Hec1-depleted cells. Similarly, Mps1 is abrogated at McCleland et al., 2003). In thisstudy,we argue that kinetochoresin Hec1-depleted cellsbut may not be Zwint-1 might manifest its checkpoint activity, in part, affected in Zwint-1-depleted cells(Martin-Lluesma by mediating the Hec1/ZW10 regulatory pathway to et al., 2002). allow proper Mad2 docking at kinetochores. Mechan- istically, this can be attributed to at least two separate Potential regulation of the Hec1 pathway by mitotic events associated with the defective MT-kinetochore kinases attachment and the lack of tension between the sister Many proteinsthat localize to kinetochoresexistas kinetochores, respectively (Li and Nicklas, 1995; Rieder preassembled complexes. In yeast, four complexes, et al., 1995). First, Hec1 interacts with the inner Ndc80, Ctf19, Dam1 and Ipl1, have been purified asa kinetochore component CEPN-H and Mis12 and forms distinct complex within kinetochores (Cheeseman et al., a stable complex with Nuf2/Spc24/Spc25 at the outer 2002a). In human, the counterpart of Ndc80 complex is kinetochore (Janke et al., 2001; Wigge and Kilmartin, identified and containsHec1, hNuf2, Spc24 and Spc25 2001; Bharadwaj et al., 2004; McCleland et al., 2004; (Wigge and Kilmartin, 2001; Bharadwaj et al., 2004; Obuse et al., 2004; Mikami et al., 2005). Studiesin Ciferri et al., 2005; Wei et al., 2005). Nevertheless, only variousorganismsindicatedthat Hec1 isrequired for two of the components, Hec1 and Spc24, were identified the formation of an integral kinetochore structure for in the human Zwint-1 complex (Kops et al., 2005). proper microtubule attachment (Janke et al., 2001; ZW10 hasalsobeen identified in a complex with Rod Martin-Lluesma et al., 2002; McCleland et al., 2003, and Zwilch (Williams et al., 2003). It appearsthat the 2004; Bharadwaj et al., 2004; Obuse et al., 2004; Deluca interrelationship of the Hec1-Zwint-1-ZW10 proteins is et al., 2005; Mikami et al., 2005). Second, ZW10 formsa rather complicated. One possible mode of regulation is large complex with ROD and Zwilch and thiscomplex is phosphorylation by mitotic kinases, which includes involved in dynein anchoring to power a proper tension Aurora kinases, Polo-like kinases (Plk) and Nek2 kinase between the sister kinetochores in a microtubule (Bigginsand Murray, 2001; He et al., 2001; Nigg, 2001; dependent manner (Williams et al., 2003). Clearly, Chen et al., 2002). Ipl1, the yeast homologue of Aurora, Zwint-1 might bridge ZW10 and Hec1 to promote the isa key protein kinaserequired for regulating kineto- proper MT-kinetochore attachment and the tension chore function in yeast, where scHec1/NDC80 is one of generation. Based on the observations of others and the substrates in vitro (Tanaka et al., 2002; Cheeseman ours, Mad2 fails to localize at kinetochores in Hec1- et al., 2002a). It islikely that Aurora kinases,the depleted cellsand partially failsin Zwint-1-depleted mammalian homologuesof Ipl1, may regulate the cells, suggesting that Hec1 and Zwint-1 may serve as a function of hsHec1 by phosphorylation. Our recent platform for the Mad2 recruitment (Martin-Lluesma results indicated that the phosphorylation of Hec1 at et al., 2002; Meraldi et al., 2004). The failure in serine 165 by Nek2A kinase is essential for its function recruiting Mad2 to kinetochoresin thesecellsmay in faithful chromosome segregation, which is consistent provide an explanation for the defect of the spindle with the finding that Nek2A localizesat kinetochores checkpoint activation (Figure 8). Apparently, both during M phase (Chen et al., 2002; Lou et al., 2004). ZW10 and Mad2 are required for a full activation of However, whether Hec1 issubjected to the regulation by the spindle checkpoint as observed in Figure 6. Without other mitotic kinases remains unclear. Furthermore,

Oncogene Hec1 recruits Zwint-1 and ZW10 to kinetochores Y-T Lin et al 6912 how these mitotic kinases modulate the Hec1, Zwint-1 ml antipain) and clarified by centrifugation at 16 000 g and ZW10 complex formation and functionality will be for 20 min at 41C. The protein concentration of clarified cell an important subject to pursue. lysate was determined by Bio-Rad protein assay (Bio-Rad, Hercules, CA, USA). For co-IP, the cell lysates were first pre- incubated with protein A-Sepharose to remove the non-specific binding proteins. The cleaned lysates were then incubated with Materials and methods anti-Hec1 antibody at 41C for 1 h followed by protein A- Sepharose, as described (Xiao et al., 2001). The immuno- Yeast two-hybrid interaction system and in vitro binding assay precipitates were washed with cold PBS three times. The For yeast two-hybrid interaction, Hec1 deletion mutants lysates and immunoprecipitates were separated by SDS– (Hec1-P, amino acids(aa) 251–618; Hec1-PR, aa 251–431; PAGE, transferred to Immobilon-P membranes (Millipore Hec1-S, aa 361–547; Hec1-HB, aa 547–618) were fused in Corp., Bedford, MA, USA) and subjected to immunoblot frame with GAL4 DNA-binding domain in the pAS vector. analysis using different antibodies as described previously Zwint-1 was fused in frame with GAL4 transactivation (Chen et al., 1997a). domain in the pSE1107 vector and co-transformed with the individual pAS clone into yeast strain Y153. The interaction was assessed by measuring the b-galactosidase activity using Immunostaining and immunofluorescence microscopy the colony lift method and quantified with the chlorophenol The immunostaining procedure was adapted as previously red-b-D-galactopyranoside assay as described previously described (Starr et al., 2000). Briefly, HeLa cellscultured on (Chen et al., 1997a). For the in vitro binding assay, fusion coverslips were washed with PBS and fixed with 4% proteinsincluding GST-Hec1, GST-Zwint-1, GST-ZW10 and paraformaldehyde in 9-(2-phosphonylmethoxyethyl)guanine GST were expressed in Escherichia coli and purified with (PEMG) buffer (80 mM piperazine-1,4-bis(2-ethanesulfonic glutathione-coupled sepharose beads. Comparable amounts of acid), pH6.8, 5 mM ethylenaglycol-bis([beta]-aminoether)- 0 GST and GST fusion proteins on the beads were incubated N,N -tetra-acetic acid (EGTA), 1 mM MgCl2, 4 M glycerol) with the in vitro translated 35S-methionine labeled Zwint-1, containing 0.5% Triton X-100 for 30 min. Cellswere then Hec1 or ZW10 using TNT Quick Coupled Transcription/ permeabilized with 0.05% saponin at room temperature for Translation Systems (Promega Corp., Madison, WI, USA) for 30 min, and washed five times with PBS. After being blocked 40 min at room temperature. The beadswere washedthree with 10% goat serum in PBS for 30 min, cells were incubated times with phosphate-buffered saline (PBS), and the proteins with primary antibodiesat 4 1C overnight, followed by five bound to the beadswere releasedby adding 2 Â sodium washes with PBS/0.5% Nonidet P-40. Affinity-purified pri- dodecyl sulfate (SDS) sample buffer. The samples were then mary antibodiesagainsthuman Hec1, Zwint-1 and ZW10 were separated by SDS–ployacrylamide gel electrophoresis (PAGE) prepared as previously described (Chen et al., 1997a), and used and detected by autoradiography. In the triple binding assay, at the appropriate dilution. Primary antibodiesagainst 35S-methionine labeled Hec1 and ZW10 were incubated with phospho-histon H3 at serine 10 (Upstate, Lake Placid, NY, purified His-Zwint-1 for 40 min before incubated with GST USA), Mad2 (Covance, Berkeley, CA, USA) and human fusion proteins on the GST beads. His-Zwint-1 was purified by CREST antibody (kindly provided by B Brinkley) were used as Ni-NTA agarose beads (Qiagen, Valencia, CA, USA) and recommended. Appropriate goat anti-human, -rabbit, or eluted with 250–500 mM imidazole (Sigma, St Louis, MO, -mouse IgG Alexa Fluor 488 and Alexa Fluor 595 secondary USA). antibodies(Molecular probes,Eugene, OR, USA) were added at a dilution of 1:2000 for 45 min and further stained with 1 mg/ ml DAPI (Fisher, Pittsburgh, PA, USA) for 15 min. After Cell culture and synchronization extensively washing with PBS/0.5% Nonidet P-40, cells were T24, a human bladder carcinoma cell line, and HeLa, a human mounted in PermaFluor (Shandon, Pittsburgh, PA, USA) and cervical carcinoma cell line, were obtained from American photographed by a Zeiss Axioplan 2 microscope using Type Culture Collection (Manassas, VA, USA) and cultured in Axiocam HR software. The fluorescent intensity of kineto- low-glucose Dulbecco modified Eagle medium containing 10% chore dotswasquantified in gray scaleby Metamorph fetal bovine serum and antibiotics. To obtain cell populations software. For chromosome spreads, active growing T24 cells synchronized at different cell cycle stages, T24 cells were first were treated with 1 ng/ml colcemid for 90 min; then, mitotic arrested at G /G by growing at high-density with low-serum 0 1 cells were shaken off and swollen in 60 mM KCl hypotonic medium, and then released from the arrest by seeding at a solution. Cells were then cytospun onto coverslips and lower-density with serum containing medium. Cells were immunostained as described above. collected at various time points after release as described (Chen et al., 1989; Xiao et al., 2001). To obtain cell populations synchronized at different cell cycle stages, HeLa siRNA, fluorescence microscopy and time-lapse imaging cells were first released from thymidine (2.5 mM) block for 4 h Three 21-nucleotide siRNA duplexes were designed according and then treated with nocodazole for 4 h to arrest cells at to the principles as described (Elbashir et al., 2001) to target mitotic phase. Mitotic cells were collected by shake-off method the Hec1 sequence (50-AAC AAG GCC TCT ATA CCC CTC- and washed with PBS before being plated on culture dish to 30), Zwint-1 sequence (50-AAG CTG CTC TGC AGC CAG allow re-entry into the cell cycle. Any cellsthat had not CTT-30), and firefly luciferase sequence (50-AAG ATT CAA attached within 6 h after replating were discarded. The AGT GCG CTG CTG-30). Each siRNA was synthesized by attached cellswere then collected at varioustime points Dharmacon Research (Lafayette, CO, USA) and transfected thereafter. by oligofectamine (Invitrogen, Carlsbad, CA, USA) into HeLa-H2B-GFP cells, which stably express GFP-fused histone Immunoprecipitation and immunoblotting H2B introduced by retrovirusinfection. Chromosomebeha- Cells were lysed in Lysis 250 buffer (50 mM Tris, pH 7.4, viorsin live cellswere monitored directly under a Nikon 250 mM NaCl, 5 mM ethylene diamine tetra acetic acid inverted TE300 microscope. For live cell time-lapse imaging, (EDTA), 0.1% Nonidet P-40, 50 mM NaF, 1 mM PMSF, cellscultured in CO 2-independent medium (Gibco-BRL, 1 mg/ml pepstatin A, 1 mg/ml leupeptin, 1 mg/ml aprotinin, 1 mg/ Grand Island, NY, USA) were photographed at one-hour

Oncogene Hec1 recruits Zwint-1 and ZW10 to kinetochores Y-T Lin et al 6913 intervals for 40 h on a heated stage using a Nikon inverted Abbreviations TE300 microscope 20 Â objective and Metamorph software. Hec1, highly expressed in cancer 1; Zwint-1, human ZW10 Cell death and mitotic index analysis interacting protein; ZW10, Zeste White 10; CIN, Chromosome To determine dead cell number, siRNA transfected HeLa cells instability; M phase, mitotic phase; Rb, retinoblastoma were trypsinized and stained with 0.2% trypan blue (Sigma, St protein; MT, microtubule; kMT, kinetochore microtubule; Louis, MO, USA) in PBS for 5 min and counted with a RNAi, RNA interference; siRNA, Short interfering RNA; Sc, hemacytometer. For mitotic index analysis, random popula- Saccharomyces cerevisiae; GFP, Green fluorescent protein; tionsof HeLa cellswere incubated with 80 ng/ml of GST, glutathione S-tansferase. nocodazole (Sigma, St Louis, MO, USA) for different periods of time after siRNA transfection. Cells were photographed Acknowledgements randomly to collect 25–30 imagesfor a total of more than 1000 cellsat each indicated time point usinga Nikon inverted We thank Dr Michael Goldberg for the human ZW10 cDNA TE300 microscope 20 Â objective and Metamorph software. plasmid, Dr Phang-Lang Chen and Bingnan Gu for technical The mitotic index and the interphase cells with multinuclei support, Dr Daniel J Riley, Ms Saori Furuta and Lily Khidr were counted from the overlapped fluorescent and phase- for critically reading the manuscript. This work was supported contrast images. by a grant from NIH (CA 107568) to WHL.

References

Amon A. (1999). Curr Opin Genet Dev 9: 69–75. Fang G, Yu H, Kirschner MW. (1998). Dev 12: Basto R, Gomes R, Karess RE. (2000). Nat Cell Biol 2: 1871–1883. 939–943. Fodde R, KuipersJ, RosenbergC, SmitsR, Kielman M, Bharadwaj R, Qi W, Yu H. (2004). J Biol Chem 279: Gaspar C et al. (2001). Nat Cell Biol 3: 433–438. 13076–13085. He X, RinesDR, EspelinCW, Sorger PK. (2001). Cell 106: BigginsS, Murray AW. (2001). Genes Dev 15: 3118–3129. 195–206. Brinkley BR, Stubblefield E. (1966). Chromosoma 19: 28–43. Hori T, Haraguchi T, Hiraoka Y, Kimura H, Fukagawa T. Chan GK, Jablonski SA, Starr DA, Goldberg ML, Yen TJ. (2003). J Cell Sci 116: 3347–3362. (2000). Nat Cell Biol 2: 944–947. Hoyt MA, TotisL, RobertsBT. (1991). Cell 66: 507–517. Chan GK, Jablonski SA, Sudakin V, Hittle JC, Yen TJ. Hsu JY, Sun ZW, Li X, Reuben M, Tatchell K, Bishop DK (1999). J Cell Biol 146: 941–954. et al. (2000). Cell 102: 279–291. Chan GK, Liu ST, Yen TJ. (2005). Trends Cell Biol 15: Janke C, Ortiz J, Lechner J, Shevchenko A, Magiera MM, 589–598. Schramm C et al. (2001). EMBO J 20: 777–791. Cheeseman IM, Anderson S, Jwa M, Green EM, Kang J, Kanda T, Sullivan KF, Wahl GM. (1998). Curr Biol 8: YatesIII JR et al. (2002a). Cell 111: 163–172. 377–385. Cheeseman IM, Drubin DG, Barnes G. (2002b). J Cell Biol Kaplan KB, BurdsAA, Swedlow JR, Bekir SS, Sorger PK, 157: 199–203. Nathke IS. (2001). Nat Cell Biol 3: 429–432. Chen PL, Scully P, Shew JY, Wang JY, Lee WH. (1989). Cell Kitagawa K, Hieter P. (2001). Nat Rev Mol Cell Biol 2: 58: 1193–1198. 678–687. Chen Y, Riley DJ, Chen PL, Lee WH. (1997a). Mol Cell Biol Kniola B, O’Toole E, McIntosh JR, Mellone B, 17: 6049–6056. Allshire R, Mengarelli S et al. (2001). Mol Biol Cell 12: Chen Y, Riley DJ, Zheng L, Chen PL, Lee WH. (2002). J Biol 2767–2775. Chem 277: 49408–49416. KopsGJ, Kim Y, Weaver BA, Mao Y, McLeod I, YatesIII Chen Y, Sharp ZD, Lee WH. (1997b). J Biol Chem 272: JR et al. (2005). J Cell Biol 169: 49–60. 24081–24087. Lengauer C, Kinzler KW, Vogelstein B. (1997). Nature 386: Ciferri C, De Luca J, Monzani S, Ferrari KJ, Ristic D, Wyman 623–627. C et al. (2005). J Biol Chem 280: 29088–29095. Li R, Murray AW. (1991). Cell 66: 519–531. Cleveland DW, Mao Y, Sullivan KF. (2003). Cell 112: Li X, NicklasRB. (1995). Nature 373: 630–632. 407–421. Lou Y, Yao J, Zereshki A, Dou Z, Ahmed K, Wang H et al. Craig JM, Earnshaw WC, Vagnarelli P. (1999). Exp Cell Res (2004). J Biol Chem 279: 20049–20057. 246: 249–262. Martin-Lluesma S, Stucke VM, Nigg EA. (2002). Science 297: De Wulf P, McAinsh AD, Sorger PK. (2003). Genes Dev 17: 2267–2270. 2902–2921. McCleland ML, Gardner RD, Kallio MJ, Daum JR, Gorbsky Deluca JG, Dong Y, Hergert P, Strauss J, Hickey JM, Salmon GJ, Burke DJ et al. (2003). Genes Dev 17: 101–114. ED et al. (2005). Mol Biol Cell 16: 519–531. McCleland ML, Kallio MJ, Barrett-Wilt GA, Kestner CA, Desai A, Rybina S, Muller-Reichert T, Shevchenko A, Hyman Shabanowitz J, Hunt DF et al. (2004). Curr Biol 14: A, Oegema K. (2003). Genes Dev 17: 2421–2435. 131–137. DoblesM, Liberal V, Scott ML, Benezra R, Sorger PK. McEwen BF, Hsieh CE, Mattheyses AL, Rieder CL. (1998). (2000). Cell 101: 635–645. Chromosoma 107: 366–375. Earnshaw WC, Rothfield N. (1985). Chromosoma 91: 313–321. Meraldi P, Draviam VM, Sorger PK. (2004). Dev Cell 7: Earnshaw WC, Sullivan KF, Machlin PS, Cooke CA, Kaiser 45–60. DA, Pollard TD et al. (1987). J Cell Biol 104: 817–829. Michel LS, Liberal V, Chatterjee A, Kirchwegger R, Pasche B, Elbashir SM, Harborth J, Lendeckel W, Yalcin A, Weber K, Gerald W et al. (2001). Nature 409: 355–359. Tuschl T. (2001). Nature 411: 494–498. Mikami Y, Hori T, Kimura H, Fukagawa T. (2005). Mol Cell Emanuele MJ, McCleland ML, Satinover DL, Stukenberg PT. Biol 25: 1958–1970. (2005). Mol Biol Cell 16: 4882–4892. Nigg EA. (2001). Nat Rev Mol Cell Biol 2: 21–32.

Oncogene Hec1 recruits Zwint-1 and ZW10 to kinetochores Y-T Lin et al 6914 Obuse C, Iwasaki O, Kiyomitsu T, Goshima G, Toyoda Y, Wang H, Hu X, Ding X, Dou Z, Yang Z, Shaw AW et al. Yanagida M. (2004). Nat Cell Biol 6: 1135–1141. (2004). J Biol Chem 279: 54590–54598. Palmer DK, O’Day K, Wener MH, AndrewsBS, Margolis Wei RR, Sorger PK, Harrison SC. (2005). Proc Natl Acad Sci RL. (1987). J Cell Biol 104: 805–815. USA 102: 5363–5367. Rieder CL, Cole RW, Khodjakov A, Sluder G. (1995). J Cell Wei Y, Yu L, Bowen J, Gorovsky MA, Allis CD. (1999). Cell Biol 130: 941–948. 97: 99–109. Savoian MS, Goldberg ML, Rieder CL. (2000). Nat Cell Biol Weiss E, Winey M. (1996). J Cell Biol 132: 111–123. 2: 948–952. Wigge PA, Jensen ON, Holmes S, Soues S, Mann M, Sharp DJ, RogersGC, Scholey JM. (2000). Nat Cell Biol 2: Kilmartin JV. (1998). J Cell Biol 141: 967–977. 922–930. Wigge PA, Kilmartin JV. (2001). J Cell Biol 152: 349–360. Starr DA, Saffery R, Li Z, Simpson AE, Choo KH, Yen TJ WilliamsBC, Goldberg ML. (1994). J Cell Sci 107: 785–798. et al. (2000). J Cell Sci 113: 1939–1950. WilliamsBC, Karr TL, Montgomery JM, Goldberg ML. Starr DA, WilliamsBC, HaysTS, Goldberg ML. (1998). J Cell (1992). J Cell Biol 118: 759–773. Biol 142: 763–774. WilliamsBC, Li Z, Liu S, WilliamsEV, Leung G, Yen TJ et al. Starr DA, WilliamsBC, Li Z, Etemad-Moghadam B, (2003). Mol Biol Cell 14: 1379–1391. Dawe RK, Goldberg ML. (1997). J Cell Biol 138: Xiao J, Liu CC, Chen PL, Lee WH. (2001). J Biol Chem 276: 1289–1301. 6105–6111. Takahashi T, Haruki N, Nomoto S, Masuda A, Saji S, Osada Zheng L, Chen Y, Lee WH. (1999). Mol Cell Biol 19: H. (1999). Oncogene 18: 4295–4300. 5417–5428. Tanaka TU, Rachidi N, Janke C, Pereira G, Galova M, Zhou H, Kuang J, Zhong L, Kuo WL, Gray JW, Sahin A et al. Schiebel E et al. (2002). Cell 108: 317–329. (1998). Nat Genet 20: 189–193. Tang Z, Bharadwaj R, Li B, Yu H. (2001). Dev Cell 1: Zinkowski RP, Meyne J, Brinkley BR. (1991). J Cell Biol 113: 227–237. 1091–1110.

Oncogene