The Ndc80 Loop Region Facilitates Formation of Kinetochore Attachment to the Dynamic Microtubule Plus End

Summary Proper chromosome segregation in mitosis relies on correct kinetochore-microtubule (KT-MT) interactions. The KT initially interacts with the lateral surface of a single MT (lateral attachment) extending from a spindle pole and is subsequently anchored at the plus end of the MT (end-on attachment) [1]. The conversion from lateral to end-on attachment is crucial because end-on attachment is more robust [2–4] and thought to be necessary to sustain KT-MT attachment when tension is applied across sister KTs upon their biorientation [1]. The mechanism for this conversion is still elusive. The Ndc80 complex is an essential component of the KT-MT interface [1, 5], and here we studied a role of the Ndc80 loop region, a distinct motif looping out from the coiled-coil shaft of the complex [6], in Saccharomyces cerevisiae. With deletions or mutations of the loop region, the lateral KT-MT attachment occurred normally; however, subsequent conversion to end-on attachment was defective, leading to failure in sister KT biorientation. The Ndc80 loop region was required for Ndc80-Dam1 interaction and KT loading of the Dam1 complex, which in turn supported KT tethering to the dynamic MT plus end [3, 7]. The Ndc80 loop region, therefore, has an important role in the conversion from lateral to end-on attachment, a crucial maturation step of KT-MT interaction.


References cited in supplemental information
Results: the defect in sister KT bi-orientation led to failure to satisfy the spindleassembly checkpoint (SAC) in ndc80Δ490-510 and ndc80-7A mutants of the Ndc80 loop region. These mutants showed a delay in anaphase onset, which was dependent on MAD2, a key component of SAC [1]. A defect in sister KT bi-orientation was also found in ipl1 (Ipl1 is an orthologue of metazoan Aurora B) and mps1 mutants; however, in contrast to the ndc80 loop region mutants, they did not engage the SAC [2][3][4]. We also found that, in ndc80Δ490-510 and ndc80-7A cells with mad2 deletion, sister CEN5s showed frequent missegregation during anaphase (56 % at 110 min and 48 % at 90 min, respectively), in contrast to wild-type NDC80 with mad2 deletion (0 %).

D. Mono-oriented sister
CEN5s often change their associated spindle poles in the Ndc80 loop mutants, in contrast to ipl1 mutants. ndc80Δ490-510 (T6690) and ipl1-321 (T2863) cells with P GAL -CEN3-tetOs TetR-GFP Venus(or YFP)-TUB1 P MET3 -CDC20 were treated with α factor in methionine drop-out medium with 2 % glucose and released to YP medium containing 2 % glucose and 2 mM methionine at 35 °C to arrest cells in metaphase. After 2.5 hrs, cells were immobilized and GFP and Venus (YFP) images were acquired every 20 sec for 30 min. The number of events, in which mono-oriented sister CEN3s changed their associated spindle poles on the spindle, were counted and shown in the graph.
Results: a defect in sister KT bi-orientation was also found in mutants of Ipl1 and Mps1, which facilitate turnover of KT-MT attachment with aberrant orientation [3,5]. We compared behaviour of mono-oriented CEN5 in the Ndc80 loop mutants, ipl1 and mps1 mutants. When sister CEN5s failed to bi-orient in both the Ndc80 loop region and ipl1 mutants, they remained unseparated and located in the vicinity of a spindle pole and this was also the case for mps1 mutants [5]. Intriguingly, whereas the associated sister CEN5s often changed their associated spindle poles in the ndc80 loop region mutants (red bars), such changes were infrequent in ipl1 and mps1 mutants (blue bars for ipl1) [3,5]. Thus, Ndc80 loop region mutants may still be able to facilitate turnover of KT-MT attachment, in contrast to ipl1 and mps1 mutants. The results in Figs S2C and S2D are consistent with the possibility that the Ndc80 loop region facilitates sister KT bi-orientation with a different mechanism, at least in part, from that of Ipl1 and Mps1 kinases.  B. In a MT co-sedimentation assay in vitro, the Dam1 complex enhances MT binding of the Ndc80 complexes similarly between NDC80 wild-type and ndc80Δ490-510.
(i) To obtain the S. cerevisiae Ndc80 complex from recombinant proteins in high yield, Ndc80-Spc25 and Nuf2-Spc24 fusion proteins were expressed in E. coli. (ii) Ndc80-Spc25 and Nuf2-Spc24 were co-purified from the bacteria. Two versions of Ndc80-Spc25 were used; one with wild-type Ndc80 and the other with ndc80Δ490-510. Meanwhile, 10 components of the S. cerevisiae Dam1 complex were also expressed in E. coli and co-purified. These proteins were applied and separated on SDS-PAGE gels. (iii) Microtubule co-sedimentation of the Ndc80 complex (with wild-type Ndc80 and the ndc80Δ490-510 mutant) in the presence and absence of the Dam1 complex in a buffer containing 100 mM NaCl. Representative images of gels are shown here. Co-sedimentation of the Ndc80 complex with MTs was quantified and plotted (means and standard errors), using the data obtained from three independent experiments.
Results: Using a MT co-sedimentation assay, it was recently demonstrated that the Dam1 complex is able to enhance MT binding of the Ndc80 complex in a buffer containing 100 mM NaCl [6]. Using the same condition, we evaluated MT cosedimentation of the purified Ndc80 complex with the loop mutant ndc80Δ490-510; its enhancement by the Dam1 complex was similar to that of the wild-type Ndc80 complex. Note that, for MT-binding reactions in this experiment, samples were incubated at room temperature for 20 min. Similar results were obtained when they were incubated at 35 ˚C for 20 min (data not shown). These results are consistent with the result that the isolated Ndc80-Nuf2 head was sufficient for enhancement of MT association by the Dam1 complex, under these conditions [6].

C. Stu2 and Ndc80 do not show interaction in a two-hybrid assay.
A two-hybrid assay was carried out as in Fig 4A. Bik1 was included in this assay as a control to confirm that our Stu2 construct fused with the activation domain produces a functional fusion protein [7,8]. We also tried to detect possible Ndc80-Stu2 interaction using co-immunoprecipitation but could not detect it (data not shown).

D. Stu2 shows localization at CEN3 (which is reactivated, but not yet interacted with MTs or the spindle) and on the metaphase spindle, similarly between wildtype and ndc80Δ490-510 cells.
Ndc80 wild-type (T7305) and ndc80Δ490-510 (T7269) cells with P GAL -CEN3-tetOs TetR-3×CFP STU2-4×mCherry Venus-TUB1 P MET3 -CDC20 were treated as in Fig 3B. Images were acquired 5-10 min after cells were suspended in medium with glucose (CEN3 reactivation). Representative images and the quantification (means and standard errors) of Stu2 signals at CEN3 are shown here. n.s.: difference is not significant.
Results: Using the assay shown in Fig 3A, we previously found that Stu2 localized at CEN after its reactivation and before its interaction with a MT extending from a spindle pole [9,10]. This was also the case during lateral KT-MT attachment, albeit with a reduced amount of Stu2 at CEN [10]. The intensity of Stu2 signals at CEN in these conditions was similar between Ndc80 wild-type and ndc80Δ490-510 (this figure and data not shown). The intensity of Stu2 signals on the metaphase spindle was also similar between the two strains (this figure, quantification not shown). Note that tubulin signals are associated with uncaptured CEN3 and this is dependent on Stu2 [10]; the intensity of the tubulin signals was similar between Ndc80 wild-type and ndc80Δ490-510 (this figure, quantification not shown).
Supplemental notes (associated with Figure 4): 1) If proper Ndc80-Dam1 interaction is necessary for correct KT-MT attachment, dam1 mutants may also show defects in KT-MT attachment similarly to ndc80 loop region mutants. We previously found that, in dam1-1 mutant cells, lateral attachment happened normally, but then CEN was often tethered to the end of a MT that could not subsequently depolymerize; i.e. end-on pulling was defective [11]. Thus the phenotypes of the dam1-1 and ndc80 loop region mutants were similar, except that CEN was tethered at the MT end in dam1-1 while a MT was rescued in ndc80 loop mutants. We speculate that, in dam1-1, residual Dam1 function may still allow interaction with Ndc80; however the resultant end-on attachment does not lead to end-on pulling due to a defect in dam1.
2) It is intriguing that, when Ndc80 loop region mutants failed to establish end-on attachment, MT rescue happened rather than KT detachment from the MT. Ndc80 loop region mutants show KT-MT interaction during this process, presumably because their intact CH domain and N-terminal region can still maintain MT association [12][13][14]. However, we reason that such attachment would not be able to complete end-on configuration in the absence of proper Ndc80-Dam1 interaction, and MT rescue occurs, facilitated by Stu2 associated with the KT ( [9]; our unpublished result).
3) Is the role of the Ndc80 loop region in the end-on KT-MT attachment conserved in evolution? The Dam1 complex is found in various yeast species, but its orthologues have not been identified in metazoa [15]. Nonetheless, several amino acid residues are well conserved within the Ndc80 loop region between yeasts and metazoa [16] (see Fig 1D). This raises an interesting possibility that the Ndc80 loop region may facilitate interaction with functional counterparts of the Dam1 complex to establish end-on KT-MT attachment in metazoan cells. For example, in these organisms, the Ska1 complex is proposed to be a functional equivalent to the Dam1 complex [17,18].

Constructing ndc80 mutants
Mutations and deletions were introduced by site-directed mutagenesis on the NDC80 gene using the Quick Change kit (Stratagene) following manufacturer's protocol, and verified by DNA sequencing. NDC80 wild-type (as a control) and mutants with its original promoter and terminator were cloned into pRS303. A single copy of this plasmid was integrated at his3 locus in a diploid strain where one of the original NDC80 genes had been deleted at its original locus. Through tetrad dissection of spores from this diploid strain, we obtained a haploid cell whose only NDC80 is wildtype or a mutant inserted at his3 locus. To make a version of Ndc80 tagged with HA, six copies of tandem HA tags were inserted at the C-terminus of NDC80 wild-type and mutants, which had been cloned in pRS303, and their single copy was subsequently integrated at his3 locus similarly to a non-tagged version.

Computer program and statistical analyses
The coiled-coil probability was calculated using COILS [28]. Multiple sequences of amino acid residues were aligned using Jalview [29]. Statistical analyses were carried out using Prism (Graph Pad) software, by choosing the unpaired t-test (Figs 2A, S2D, S4A, S4D), Fisher's exact test for contingency table (Fig 3C i) and the chi-square test for trend (Figs S2B, 3C ii).

Live-cell imaging
The procedures for time-lapse fluorescence microscopy were described previously [30]. Scientific) CCD camera. We acquired 7-9 (0.5-0.7 mm apart) z-sections, which were subsequently deconvoluted, projected to two-dimensional images and analyzed with SoftWoRx and Volocity (Improvision) software. CFP, Venus (or GFP), and mCherry (or RFP) signals were discriminated using the 89006 multi-band filter set (Chroma).
GFP and Venus signals were discriminated using the JP3 filter set (Chroma). Images of fixed cells were acquired similarly except that cells had been treated with paraformaldehyde for fixation beforehand.

Protein immunoprecipitation
Protein extracts were prepared from cells by 2-min bead beating at 4˚C in buffer containing 50mM HEPES pH 7.5, 150mM NaCl, 0.05% Tween 20, 1mM βmercaptoethanol, protease inhibitor cocktail IV (Calbiochem) and PhosSTOP phosphatase inhibitor cocktail (Roche). Extracts were cleared by centrifugation at 10,000 g for 10 min at 4˚C and equal amounts (500 µg) of total protein were used for immunoprecipitations. Myc-tagged proteins were immunoprecipitated using anti-myc agarose beads (Sigma). Beads were rolled with extracts at 4˚C for 1 hour, washed 3 times by re-suspension in the above buffer containing 0.5 M NaCl with a final wash in extraction buffer. Proteins were eluted by heating at 100˚C in Laemmli sample buffer for 3 min, separated by PAGE on 4-12% Novex Bis-Tris gels and HA-tagged proteins were detected on Western blots using 12CA5 anti-HA antibody.

Two-hybrid assay
Two-hybrid assay was carried out as described previously [31]. The constructs of NDC80 (wild-type and mutants) and BIK1 were amplified by PCR, cloned into pBTM116 plasmid [32] and fused with the DNA binding domain. The constructs of DAM1, NUF2 and STU2 were amplified by PCR, cloned into pGAD-C2 plasmid [33] and fused with the activation domain for transcription. These plasmids were transformed into the L40 strain of S. cerevisiae. Two-hybrid interaction was evaluated by expression of HIS3 reporter gene on medium lacking histidine. Constructs with RAS and RAF were used as controls.

Chromatin immunoprecipitation
Chromatin immunoprecipitation was carried out as described previously [34]. The percentage of DNA recovery in immunoprecipitation was quantified by making a serial dilution of immunoprecipitated DNA and total DNA from whole cell extracts, followed by their use as PCR templates, as described previously [34]. The PCR amplified a 300-bp DNA fragment spanning CEN3 and, as an example of a noncentromere locus, a 230-bp DNA fragment within MPS1 gene (45 kb from CEN4).

Purification of recombinant protein
To obtain the purified S. mM NaCl, 1 mM EDTA using a PD10 desalting column (GE Healthcare) and the desalted fraction was loaded on 1 ml Q-Sepharose and protein was eluted with a stepwise gradient of NaCl. Fractions were dialyzed against 20 mM sodium phosphate (pH 6.8), 150 mM NaCl, 1 mM EDTA.

Microtubule co-sedimentation assays
Assays were conducted using using tubulin polymerized with taxol (both from Cytoskeleton, Inc) in BRB80 (80mM PIPES-NaOH pH6.8, 1 mM MgCl 2 , 1mM EGTA) containing 1mM GTP and 20 μM taxol, as described in [6,37]. Ndc80 sequoia complex and Dam1 complex were cleared by centrifugation and 0.3 µM sequoia, in the presence or absence of 0.3 µM Dam1 complex, was mixed with microtubules in BRB80 containing 1 µM BSA and 100 mM NaCl and incubated at room temperature for 20 min, after which the reaction was centrifuged at 90,000 rpm for 10 min at 25 ˚C in TLA100 rotor. Equivalent volumes of the supernatant and pellet were run on 4-12 % Novex Bis-Tris SDS-PAGE gels (Invitrogen), which were stained with Simply Blue Safe Stain (Invitrogen) and scanned to quantify the amounts