A conserved site on Ndc80 complex facilitates dynamic recruitment of Mps1 to yeast kinetochores to promote accurate chromosome segregation

SUMMARY

Accurate chromosome segregation relies on kinetochores carrying out multiple functions, including establishing and maintaining microtubule attachments, forming precise bioriented attachments between sister chromatids, and activating the spindle assembly checkpoint. Central to these processes is the highly conserved Ndc80 complex. This kinetochore subcomplex interacts directly with microtubules, but also serves as a critical platform for recruiting kinetochore-associated factors and as a key substrate for error correction kinases. The precise manner in which these kinetochore factors interact, and regulate each other’s function, remains unknown – considerably hindering our understanding of how Ndc80 complex-dependent processes function together to orchestrate accurate chromosome segregation. Here, we aimed to uncover the role of Nuf2’s CH domain, a component of the Ndc80 complex, in ensuring these processes. Through extensive mutational analysis, we identified a conserved interaction domain comprised of two segments in Nuf2’s CH domain that form the binding site for Mps1 within the yeast Ndc80 complex. Interestingly, this site also associates with the Dam1 complex, suggesting Mps1 recruitment may be subject to regulation by competitive binding with other factors. Mutants disrupting this “interaction hub” exhibit defects in spindle assembly checkpoint function and severe chromosome segregation errors. Significantly, specifically restoring Mps1-Ndc80 complex association rescues these defects. Our findings shed light on the intricate regulation of Ndc80 complex-dependent functions and highlight the essential role of Mps1 in kinetochore biorientation and accurate chromosome segregation.

Graphical Abstract

eTOC Blurb

Parnell et al. identify an ‘interaction hub’ in the Ndc80 complex via mutational analysis in yeast, which associates with both Mps1 and Dam1 complex. Mutants at this site show inviability due to severe defects in biorientation and chromosome segregation; however, restoring Mps1-Ndc80c association rescues viability, highlighting its critical role.

INTRODUCTION

The precise segregation of duplicated chromosomes into daughter cells is a fundamental process during cell division. This segregation relies on the interactions between microtubules and kinetochores, which are large protein complexes that assemble on the centromeres of each chromosome (reviewed in ¹). Accurate chromosome segregation relies on several critical functions performed by kinetochores. First, they must establish and maintain attachments to the tips of microtubules, whose dynamic growing and shrinking generates the forces to physically move chromosomes within the cell. Second, kinetochores on each pair of replicated sister chromatids must form bioriented attachments to microtubules originating from opposite spindle poles. Since initial kinetochore-microtubule attachments are randomly established, these attachments are often erroneous and must be detected and corrected to prevent mis-segregation. Third, kinetochores must activate the spindle assembly checkpoint (SAC) when microtubule attachments are absent, thereby halting the cell cycle until these errors are corrected (reviewed in ^2,3). The precise mechanisms by which kinetochores coordinate and regulate these diverse functions remain important and open questions.

At the core of these essential cellular functions lies the Ndc80 complex (Ndc80c), a highly conserved subcomplex of the kinetochore. This complex is a heterotetrameric, rod-like assembly, comprised of four distinct proteins: Ndc80, Nuf2, Spc24, and Spc25^4–9. The complex engages with microtubules through multiple domains, including the globular CH-domains located at the N-terminal end of the Ndc80:Nuf2 complex, which provides a positively charged interaction surface for microtubules^7,10,11. The unstructured, flexible N-terminal tail of the Ndc80 protein harbors an additional microtubule-binding element^7,11–14. The role of Ndc80’s CH domains in microtubule binding is well-established both in vitro and in vivo^7,10,14,15. In contrast, the function of Nuf2’s CH domain has remained somewhat enigmatic^7,15. Structural studies, for instance, indicate that this domain does not make direct contact with the microtubule^6,9,16. Nevertheless, its high degree of conservation suggests that it likely serves vital functions.

Besides its direct microtubule binding capability, the Ndc80c serves as a critical “landing pad” for numerous microtubule-associated proteins, as well as a substrate for various kinases. To enhance the ability to form load-bearing attachments and to track with dynamic microtubule tips, the Ndc80c associates with the Dam1 complex (Dam1c) in yeast^10,17–22 and the functionally analogous Ska complex in metazoans^23–26. The Ndc80 protein itself serves as a critical phosphorylation target for error correction kinases, namely Ipl1 and Mps1^{7,11–13,27–38}. Additionally, the Ndc80c directly interacts with other factors implicated in correcting erroneous attachments, including Stu2/chTOG^39–41. Lastly, the Ndc80c assumes a pivotal role in triggering the SAC by recruiting its most upstream kinase, Mps1^{33–35,37,42–48}. A significant barrier to understanding how these diverse Ndc80c-dependent functions are coordinated lies in our limited understanding of how these different factors interact with the Ndc80c and the degree to which they influence each other’s functions.

In this study, our aim was to elucidate the role of Nuf2’s CH domain in facilitating chromosome segregation fidelity. Through extensive mutational analysis, we identified a conserved “interaction hub” formed by two segments of Nuf2’s CH domain: its N-terminal loop and G-helix. We demonstrate that this hub constitutes both a key portion of the Mps1 binding interface within the yeast Ndc80c as well as a site of kinetochore association for additional factors such as the Dam1c. Mutants disrupting this region of Nuf2 exhibit severe chromosome segregation errors and viability defects. Surprisingly, restoration of Mps1-Ndc80c association rescues these cellular defects, suggesting Mps1 is the most upstream and/or critical factor that binds to this region. Importantly, Mps1 bound to this interaction hub carries out crucial functions, including promoting sister chromatid biorientation and Dam1c recruitment, in addition to its well-studied role in SAC signaling. The association of Mps1 and the Dam1c with this same region of Nuf2, along with Mps1’s function in recruiting the Dam1c, implies that Mps1’s kinetochore activity may recruit a competitive binder. Our work sheds light on the manner in which multiple Ndc80c-dependent functions are regulated and demonstrates the essential role of Mps1 in kinetochore biorientation and chromosome segregation.

RESULTS

Identification of a portion of Nuf2 that is essential for cell viability

For this study, we wished to determine the role that Nuf2’s CH domain plays in the establishment of correctly bioriented sister chromatids. To identify residues that are important for its function, we looked for spatial patterns of amino acid conservation. Here, we aligned Nuf2 sequences from various fungal species and mapped the resulting conservation scores onto a previously determined Ndc80c structure⁴⁹ (PDB: 5TCS). This analysis revealed a region of conserved residues in close spatial proximity, which is centered on Nuf2’s “G-helix” as well as its amino-terminal “loop” (Figure 1A & Figure S1A). These residues, especially F8, P9, R118, S124, N128 and R131, appear highly conserved in fungal and metazoan Nuf2 sequences (Figure 1A & Figure S1A–B). To examine the functional importance of these residues, we generated strains containing an auxin-inducible degron of the endogenous NUF2, and ectopically expressed mutant nuf2 alleles to examine their phenotypes in the presence of auxin. We focused on mutating residues that showed strong conservation, and assayed effects on cell viability using a spot dilution assay. Many of the mutations resulted in severe cell viability defects, and the most penetrant phenotypes were associated with mutations in the Nuf2 G-helix (nuf2^S124D, nuf2^A125D, nuf2^N128A and nuf2^R131E) or N-terminal loop (nuf2^{F8A P9A}; Figure 1B & Figure S1C). Importantly, these mutations do not appear to affect Nuf2 expression levels or predicted structure (Figure S1D–E). A summary of all mutants and associated viability phenotypes can be found in Figure 1C and Figure S1C. Based on the severity of viability defects, amount of conservation, and degree of spatial proximity, we chose to focus on nuf2^N128A, nuf2^{F8A P9A}, and nuf2^S124D for further analysis.

Figure 1. Mutation of conserved Nuf2 residues affects cell viability.

(A) Cartoon of the heterotetrameric Ndc80 complex (Ndc80, Nuf2, Spc24, Spc25). CH domains are represented as ovals; coiled-coil domains and the N-terminal tail of Ndc80 are shown as wavy lines. Structure shows S. cerevisiae Nuf2 from ⁴⁹ (5TCS), with other members of the Ndc80c not shown. Left: Conservation of Nuf2 among 12 fungal species (listed in Figure S1), viewed using ChimeraX⁸⁰, on a scale of −2.5 to +2.5, with the most conserved residues in red and least conserved in blue. Right: Conserved residues used throughout this work are highlighted (F8 and P9 in the N-terminal loop; S124 and N128 in the G-helix). See Figure S1A–B for multiple sequence alignment among species and structural comparison between S. cerevisiae and human Nuf2.

(B) Yeast cell viability assay with a subset of nuf2 mutant alleles. Strains carry NUF2-AID and an ectopic copy of NUF2-3HA (No covering allele, “None”, M1889; NUF2^WT, M2038; nuf2^{F8A P9A}, M2042; nuf2^s124A, M2040; nuf2^S124D, M2041; nuf2^N128A, M2414). Cells were serially diluted five-fold and spotted onto plates containing DMSO (control), 250 μM auxin (to degrade endogenous Nuf2-AID protein), or 250 μM auxin + 6.5 μg/mL benomyl. See Figure S1C for additional nuf2 mutants.

(C) Structure of Nuf2 (5TCS), illustrating the range of cell viability phenotypes of mutant alleles observed on plates containing benomyl (data shown in Figure S1C). Nuf2 mutants are listed in groups in decreasing shades of purple, from the strongest phenotype (death on auxin only) to the weakest (slight sickness on auxin + benomyl).

See also Figure S1A–E.

Mutants within Nuf2’s patch result in spindle assembly checkpoint defects and substantial chromosome segregation errors

To begin understanding how these nuf2 CH domain mutants result in cell viability defects, we first compared the growth phenotypes to previously described mutations in the CH domain of Nuf2’s binding partner, Ndc80. We found that nuf2 mutants, including nuf2^N128A and nuf2^{F8A P9A}, display a considerably more severe viability defect than ndc80 mutants which perturb microtubule binding^7,10 (i.e. ndc80^K204E and ndc80^{K122E K204E}; Figure S2A–B). Furthermore, while ndc80^{K122E K204E} mutants arrest in mitosis¹⁰, likely by activating the SAC, the nuf2 mutants did not induce a specific cell cycle arrest (Figure 2A). These observations suggest that nuf2 CH domain mutations affect a kinetochore function beyond just microtubule binding. Given the lack of cell cycle arrest, we next assessed SAC function by examining the localization of the SAC component, Bub1, which is recruited to kinetochores upon SAC activation^50–53. We found that approximately 72 ± 6% of NUF2^WT cells showed Bub1-GFP recruitment to the kinetochore upon SAC activation by treatment with the microtubule poison nocodazole, consistent with prior observations^51,53. In contrast, the nuf2 mutants examined showed a dramatic decrease in the number of cells with Bub1 at kinetochores (Figure S2C), consistent with defects in SAC function. Unlike ndc80-1 loss of function mutants⁵⁴, the clustering of Mtw1-mCherry appeared normal in the nuf2 mutants, indicating that these mutations do not cause gross defects in kinetochore organization (Figure S2C).

Figure 2. Inviable nuf2 mutants display a loss of sister chromatid biorientation and mis-segregation of chromosomes.

(A) Comparison of yeast cells expressing nuf2 mutant alleles to those expressing an ndc80 mutant allele that induces a metaphase arrest. Exponentially growing strains with NDC80-AID or NUF2-AID and ectopic copies of either ndc80^{K122E K204E}-3HA (M694) or NUF2-3HA (NUF2^WT, M2933; nuf2^N128A, M2945; nuf2^{F8A P9A}, M2935; nuf2^S124D, M2936) were treated with 500 μM auxin for 2.5 hours to degrade endogenous Ndc80-AID or Nuf2-AID protein prior to fixation. Left: Representative micrographs. Right: Quantitation of cells showing the percentage of large-budded, mononucleate cells in ndc80 or nuf2 mutants, with error bars indicating the standard deviation among three replicates (n ≥ 100 cells for each replicate). Significance was determined by a two-tailed unpaired t test (**; P ≤ 0.01).

(B) Assay for sister chromatid biorientation in nuf2 mutants. Exponentially growing cells carrying NUF2-AID and ectopic copies of NUF2-3HA (NUF2^WT, M4245; nuf^2N128A, M4253; nuf2^{F8A P9A}, M4251; nuf2^S124D, M4249), pMET-CDC20, and CEN III marked with GFP (CEN III:lacO LacI-GFP) were arrested with 1 μg/mL α-factor for 3 hours, followed by release into fresh media containing methionine (to induce metaphase arrest via pMET-CDC20) and 500 μM auxin (to degrade endogenous Nuf2-AID) for 1.5 hours. Left: Representative micrographs. The brightness of images was adjusted individually for GFP and DAPI channels, but acquisition settings for each channel were consistent across samples. Right: Quantitation of percent of large-budded, mononucleate cells displaying biorientation of the chromosome-marked GFP signal, with error bars indicating the standard deviation among three replicates (n= 71-137 cells for each replicate). Note: it is common for the percent of biorientation to be ~50% in such assays^18,35. Significance was determined by a two-tailed unpaired t test (** P ≤ 0.01).

(C) Chromosome segregation assay in nuf2 mutants. Exponentially growing cells carrying NUF2-AID and ectopic copies of NUF2-3HA (NUF2^WT, M4469; nuf2^N128A, M4479; nuf2^{F8A P9A}, M4475; nuf2^S124D, M4473), as well as CEN IV marked with GFP (CEN IV:lacO LacI-GFP) and SPC110-mCherry (spindle pole marker) were arrested with 1 μg/mL α-factor for 3 hours, followed by release into fresh media containing 500 μM auxin for 2 hours. Left: Representative micrographs. The brightness of images was adjusted individually for GFP, mCherry and DAPI channels, but acquisition settings for each channel were consistent across samples. Right: Quantitation of percent of binucleate cells displaying correct segregation of the chromosome-marked GFP signal, with error bars indicating the standard deviation among three replicates (n = 75-151 cells for each replicate). Significance was determined by a two-tailed unpaired t test (****; P ≤ 0.0001).

See also Figure S2A–I.

While these nuf2 CH domain mutants display SAC defects, the lethal phenotypes observed cannot be explained by this limitation alone since the SAC is not essential for cell viability in yeast (Figure S2D). This discrepancy led us to examine sister kinetochore biorientation and chromosome segregation defects as a potential cause of the lethality. To monitor biorientation, we induced a metaphase arrest by depleting Cdc20 (using a methionine-repressible CDC20 allele) in cells that also carry a fluorescently marked centromere of chromosome III⁵⁵. Under these conditions, opposing spindle pulling forces cause bioriented sister chromatids to separate, appearing as two distinct GFP puncta⁴⁹. The frequency of biorientation in NUF2^WT cells was similar to those typically observed in this assay^18,35 (44 ± 6%, Figure 2B). However, in nuf2 mutant cells, we observed a marked reduction in biorientation of CEN III and a spectrum of phenotypic penetrance (nuf2^N128A, 18 ± 1.5%; nuf2^{F8A P9A}, 10 ± 1.5%; nuf2^S124D, 6.6 ± 1.5%). We also noted a significant portion of nuf2 mutant cells undergoing spindle elongation, bypassing the arrest induced by Cdc20 depletion (Figure S2E–F). This seems to result from substantial biorientation deficiencies in the nuf2 mutants, where many or all pairs of sister chromatids form syntelic attachments to the mitotic spindle, consequently lacking the necessary forces to restrict spindle elongation (Figure S2G–H).

To examine chromosome segregation fidelity in these nuf2 mutants, cells again carrying a fluorescently marked centromere (now of chromosome IV) were released from a G1 arrest, and chromosome segregation was examined upon anaphase onset. Cells expressing nuf2 mutants displayed surprisingly high errors in segregating chromosome IV (NUF2^WT 96 ± 4.6% correct segregation compared to nuf2^N128A 14 ± 3.5%, nuf2^{F8A P9A} 4 ± 3.1%, nuf2^S124D 3 ± 1.2%; Figure 2C and Figure S2I). Together, these observations show that this region of Nuf2’s CH domain, consisting of its G-helix and N-terminal loop, is required for SAC function and essential for biorientation and accurate chromosome segregation.

Mps1-kinetochore localization is perturbed in nuf2 patch mutants

The defects in both SAC function and biorientation led us to consider the idea that decreased recruitment of the Mps1 kinase to the kinetochore may be the cause of the observed phenotypes. Mps1 lies at the apex of SAC signaling. Beyond this, its activity is important in the establishment of correctly bioriented kinetochore attachments^{32–37,44,45}, and Nuf2 and Ndc80 have been implicated as the kinetochore binding site for Mps1^30,42,43,56. We therefore used an established kinetochore co-IP assay to examine the levels of Mps1 associated with the kinetochore in NUF2^WT relative to nuf2 mutants^31,57. Kinetochores were immunoprecipitated via α-Flag pulldown of the kinetochore component Dsn1-Flag, and co-purification of Mps1-3V5 was determined by western blot. As expected, NUF2-AID cells expressing NUF2^WT showed Mps1 co-purification with kinetochores. In contrast, the nuf2 mutants displayed nearly undetectable Mps1 binding but preserved Ndc80 association (Figure 3A). Our data demonstrate that this region of Nuf2 forms a key part of the Mps1 binding interface to the yeast Ndc80c, which is consistent with Yu and colleagues who made a similar observation for human Mps1 binding to Ndc80c⁴³, suggesting this is a remarkably conserved interface.

Figure 3. Mps1 kinetochore recruitment is lost in nuf2 mutants, and an mps1 mutant partially phenocopies a nuf2 mutant for chromosome segregation defects.

(A) Detection of Mps1-3V5 association with the kinetochore via immunoprecipitation. Exponentially growing cultures expressing DSN1-6His-3Flag alone (No tag, M2630) or also with MPS1-3V5, NUF2-AID and ectopic copies of NUF2-3HA (NUF2^WT, M3494; nuf2^{F8A P9A}, M3500; nuf2^S124D, M3498; nuf2^N128A, M3504), were treated with auxin 2 hours prior to harvest. Kinetochores were purified from lysates by α-Flag immunoprecipitation (IP) and analyzed by immunoblotting.

(B) Cartoon of the Mps1 protein and a conservation histogram for amino acids 127 to 188. The region from 148 to 173 is shown in detail, with amino acids mutated in this work (R155, L165, R170) each marked by an *. Numbering is according to the S. cerevisiae sequence.

(C) Complementation of the temperature sensitive mps1-1 allele (M56) by ectopically expressed MPS1 (MPS1^WT, M4407; mps1^Δ151-200, M4412; mps1^Δ201-300, M4413; mps1^RLR>EEE, M4949). Note: we used a temperature sensitive allele, mps1-1, to disrupt the endogenous function as an MPS1-AID allele did not fully disrupt cell viability in our strain background (data not shown). Cells were serially diluted five-fold, spotted onto plates, and grown at mps1-1 permissive (23°C) and non-permissive (37°C) temperatures. See Figure S3A for additional mps1 alleles.

(D) MPS1-mNeonGreen assay for recruitment of Mps1 to the kinetochore. Exponentially growing cells carrying the mps1-1 temperature sensitive allele and ectopic copies of MPS1 (MPS1^WT, M5446, M5458; mps1^RLR>EEE, M5447, M5459), as well as SPC110-mCherry (spindle pole marker) were shifted to 37°C and treated with nocodazole for 2 hours. Left: Representative micrographs. The brightness of images was adjusted individually for mNeonGreen, mCherry and DAPI channels, but acquisition settings for each channel were consistent across samples. Right: Quantitation of percent of large-budded cells showing strong, weak or no Mps1-mNeonGreen signal. Error bars indicate the standard deviation among three replicates, representing two independent strains, for each genotype (n = 102 to 126 cells per replicate). See Figure S3C for localization of Mps1-mNeonGreen in asynchronously growing cells.

(E) Chromosome segregation assay in mps1 mutants. Exponentially growing cells carrying the mps1-1 temperature sensitive allele and ectopic copies of MPS1 (MPS1^WT, M4714; mps1^RLR>EEE, M4710, M4947, M4948), as well as CEN IV marked with GFP (CEN IV:lacO LacI-GFP) and SPC110-mCherry (spindle pole marker), were arrested with 1 μg/mL α-factor for 3 hours, followed by release into fresh media at 37°C for 2 hours. Due to the difference in phenotypes between nuf2 mutants and mps1^RLR>EEE, three independent strains were analyzed to ensure validity of the result. Left: Representative micrographs. The brightness of images was adjusted individually for GFP, mCherry and DAPI channels, but acquisition settings for each channel were consistent across samples. Right: Quantitation of percent of binucleate cells displaying correct segregation of the chromosome-marked GFP signal, with error bars indicating the standard deviation among three to four replicates (n = 61-154 cells for each replicate). Significance was determined by a two-tailed unpaired t test (****; P ≤ 0.0001).

See also Figure S3A–F.

If the nuf2 mutant defects stem from reduced Mps1 binding, we hypothesized that an mps1 mutant at this interface would exhibit a similar phenotype. To identify potential interaction regions of Mps1 with the Ndc80c, we sought conserved regions within Mps1 outside its C-terminal kinase domain. Earlier studies associated residues 151-200 with Mps1’s SAC and biorientation functions, while finding residues 201-300 are crucial for spindle pole body duplication⁵⁸. Within residues 151-200 we identified two highly conserved patches encompassing residues 151-157 and 165-171 (Figure 3B). To assess the functional significance of these conserved Mps1 segments, we examined cell viability. Here, we used a temperature sensitive allele, mps1-1, to disrupt the endogenous function. Consistent with previous observations, mps1^Δ151-200 and mps1^Δ201-300 mutants were inviable in this system⁵⁸ (Figure 3C). While neither mps1^Δ151-157 nor mps1^Δ165-171 fully disrupted cell viability, combining these deletions or introducing point mutations to highly conserved residues that span both regions resulted in complete inviability (Figure 3C and Figure S3A). Based on the above observations, we decided to further characterize the mps1^{R155E L165E R170E} mutant (hereafter designated as “mps1^RLR>EEE”).

To evaluate the potential defect in the kinetochore association of mps1^RLR>EEE, we employed tagging of the ectopically expressed alleles to examine their localization at the kinetochore. Initially, we observed that MPS1^WT, tagged with 3V5 and expressed from an ectopic locus, failed to complement mps1-1 cells at non-permissive temperatures. However, we found that an Mps1-mNeonGreen fusion appeared functional under comparable conditions (Figure S3B). Consequently, we opted to use fluorescent microscopy for this analysis. As expected, mps1-1 cells expressing MPS1^WT displayed an increase in the number and intensity of Mps1 foci following nocodazole treatment. In contrast, significantly fewer cells expressing mps1^RLR>EEE exhibited a discernible Mps1 signal (MPS1^WT strong signal 77 ± 6%, mps1^RLR>EEE strong signal 5 ± 3%; Figure 3D). Furthermore, robust spindle-proximal Mps1 signal was observed in asynchronously growing mitotic cells; however, cells expressing mps1^RLR>EEE lacked similar localization (Figure S3C), despite the mps1^RLR>EEE mutant displaying comparable protein expression levels (Figure S3D). Together, these results indicate that Mps1 residues R155, L165, and R170 indeed constitute a portion of the kinetochore binding motif. The notion that this portion of Mps1 facilitates binding to the Ndc80c is well-supported by concurrent studies from Pleuger et al.⁵⁹ and Zahm et al.60, employing biochemical and structural approaches to determine the Mps1-Ndc80c interface.

Next, we examined which specific aspects of Mps1 function are disrupted in mps1^RLR>EEE mutants, with a focus on spindle pole body separation and chromosome segregation. Cells carrying a fluorescently marked centromere of chromosome IV were arrested in G1 at a permissive temperature. Upon release from this G1 arrest at a non-permissive temperature, we examined spindle pole body separation and chromosome segregation during anaphase. mps1^RLR>EEE cells showed spindle pole body separation similar to that of MPS1^WT, indicating this role of Mps1⁶¹ remains functional in these mutants (Figure S3E). In contrast, mps1^RLR>EEE mutants displayed a large degree of chromosome segregation errors (MPS1^WT 99 ± 1.4% correct segregation, compared to mps1^RLR>EEE 56 ± 4.3%; Figure 3E and Figure S3F). We note, however, that the chromosome segregation defects observed in the mps1^RLR>EEE mutant were not as pronounced as those observed in the nuf2 CH domain mutants (14% correct segregation for nuf2^N128A compared to 56% for mps1^RLR>EEE; Figure 2C and Figure 3E). This discrepancy may be due to the mps1^RLR>EEE mutant not completely disrupting Mps1’s ability to bind the Ndc80c. Alternatively, it is plausible that additional factors might bind this region of Nuf2’s CH domain, and the observed phenotypes could arise from the combination of these defects.

Dam1 complex recruitment to kinetochores is dependent on the interaction hub and kinetochore-associated Mps1

The inconsistency in observed chromosome segregation defects between the nuf2 and mps1 mutants prompted us to consider that the nuf2 mutants might disrupt binding to additional factors. Recent structural studies have suggested this region of the Ndc80c as a potential interaction site for the Dam1c^19,22. We therefore investigated the interaction of Dam1-myc with kinetochores in nuf2 mutants. In cells expressing NUF2^WT, Dam1 co-purified with kinetochores. Conversely, cells expressing the nuf2 mutants exhibited a significant reduction in Dam1 binding (Figure 4A). We did observe slight differences in the limited residual association of the Dam1c. This may reflect that the nuf2 mutants differentially affect the Dam1c interface, which is likely distinct from the precise Mps1 binding interface. Regardless, the extent of disruption to Dam1c association was surprising, considering the Dam1c putatively makes numerous contacts with the Ndc80c^20–22. Therefore, to corrobate this observation, we conducted two further experiments. First, we examined synthetic lethality of nuf2 mutants with hypomorphic alleles of Dam1c components. Second, we assessed the localization of the Dam1c along the mitotic spindle using fluorescent microscopy. Consistent with compromised Dam1c localization in nuf2 mutants, we found that hypomorphic nuf2^F8A, nuf2^P9A, and nuf2^{I10A L11A} alleles display synthetic lethality and/or sickness with dad1-1 mutants (Figure 4B and Figure S4A–B). Furthermore, Dam1c localization appeared altered in nuf2 mutants by microscopy. Metaphase-arrested NUF2^WT cells exhibited a characteristic bilobed distribution of the Dam1c component, Ask1-YFP, proximal to the spindle pole (marked by Spc110-mCherry), consistent with expected kinetochore localization. In contrast, nuf2^{F8A P9A} mutant cells displayed a change in Ask1 localization, showing a decrease proximal to the spindle pole, and a noticeable increase in localization along the mitotic spindle (Figure 4C). These observations suggest that Dam1c association with the kinetochore is reduced and mis-localized along the mitotic spindle in the nuf2 mutants, but not eliminated. Together, our findings support the notion that multiple factors, including Mps1 and the Dam1c, bind to this “interaction hub” of Nuf2’s CH domain, with the observed phenotypes likely a result of the combined effects of these binding alterations.

Figure 4. Dam1 localization to the kinetochore is disrupted in nuf2 and mps1^RLR>EEE mutants.

(A) Detection of Dam1-9Myc association with the kinetochore via immunoprecipitation. Exponentially growing cultures (No tag, M3) expressing DSN1-6His-3Flag with DAM1-9Myc, NUF2-AID and ectopic copies of NUF2-3HA (NUF2^WT, M5121; nuf2^{F8A P9A}, M5122; nuf2^S124D, M5124; nuf2^N128A, M5123), were treated with auxin 2 hours prior to harvest. Kinetochores were purified from lysates by α-Flag immunoprecipitation (IP) and analyzed by immunoblotting.

(B) Cell viability assay assessing synthetic lethality of nuf2 mutant alleles with the temperature sensitive dad1-1 allele. Strains carry a NUF2-AID allele and ectopic copies of NUF2-3HA (NUF2^WT DAD1, M2038; NUF2^WT dad1-1, M4465; nuf2^P9A DAD1, M2152; nuf2^P9A dad1-1, M4926; nuf2^{I10A L11A} DAD1, M3750; nuf2^{I10A L11A} dad1-1, M4928). Cells were serially diluted five-fold, spotted onto plates containing DMSO or 250 μM auxin and grown at a temperature permissive for dad1-1 (23°C). Results for a dad1-1 semi-permissive temperature (30°C) are shown in Figure S4A. See Figure S4B for assay of an additional nuf2 allele.

(C) Ask1-YFP localization assay in NUF2^WT and nuf2^{F8A P9A}. Strains carry a NUF2-AID allele and ectopic copies of NUF2-3HA (NUF2^WT, M4985; nuf2^{F8A P9A}, M4987), as well as CDC20-AID, ASK1-YFP (Dam1 complex) and SPC110-mCherry (spindle pole marker). Cells were treated with 500 μM auxin 2 hours prior to harvest (to degrade both endogenous Cdc20-AID and Nuf2-AID). Left: Representative micrographs. The brightness of images was adjusted individually for YFP and mCherry channels, but acquisition settings for each channel were consistent across samples. Middle: Profile plots showing intensities of a line scan through the spindles of NUF2^WT and nuf2^{F8A P9A} cells. The Spc110-mCherry signal intensity (red) was plotted on the left y-axis, and the Ask1-YFP signal intensity (green) was plotted on the right y-axis. AU = Arbitrary Units of fluorescent intensity. Right: Graph of YFP intensities, with each dot representing an individual spindle pole-proximal peak (n ≥ 100 cells, ≥ 200 peaks each for NUF2^WT and nuf2^{F8A P9A}). Significance was determined by a two-tailed unpaired t test (****; P < 0.0001).

(D) Detection of Dam1-9Myc association with the kinetochore via immunoprecipitation. Exponentially growing cultures expressing DSN1-6His-3Flag, DAM1-9Myc and NDC80-3HA in wild-type (M457), mps1-1 (M469), or mps1-1 with ectopic copies of MPS1 (MPS1^WT, M5130; mps1^RLR>EEE, M5132) were shifted to 37°C for 2 hours prior to harvest. Kinetochores were purified from lysates by α-Flag immunoprecipitation (IP) and analyzed by immunoblotting.

We next investigated if there was a functional relationship between the localization of Mps1 and the Dam1c. Previous work suggested that Mps1 activity positively regulates Dam1c localization to the kinetochore^31,36,62,63. It remains uncertain whether it is the kinetochore-associated pool of Mps1 that specifically promotes Dam1c recruitment, especially considering that both rely on the same portion of the Ndc80c for kinetochore association. Consistent with Mps1 recruiting the Dam1c, kinetochores isolated from mps1-1 cells, which lack Mps1 at the restrictive temperature⁵³, showed decreased association with the Dam1c (Figure. 4D). This deficiency was restored by introducing an exogenous copy of MPS1^WT, but not the kinetochore-binding-deficient mutant mps1^RLR>EEE (Figure 4D). These observations imply that kinetochore-bound Mps1 actively participates in recruiting the Dam1c and raises an intriguing possibility that competition for binding to the Ndc80c at this site might contribute to regulating the localization and function of these factors.

Mps1 is the critical factor recruited to the Ndc80 complex interaction hub

Given that various factors appear to interact with this portion of the Ndc80c, a challenging question arises: how can we assess their relative importance in facilitating accurate chromosome segregation? Some insight into this question came from the following observations. We reasoned that if loss of Mps1-kinetochore association was the main cause of viability defects in our nuf2 mutants, perhaps restoring Mps1 levels at the kinetochore could suppress the associated defects. We addressed this question in a number of ways. First, we asked if ectopically tethering Mps1 to the kinetochore, using rapamycin induced dimerization of FRB and FKBP12, would rescue the nuf2 mutant phenotypes (Figure 5A). We generated strains containing NUF2-AID, NDC80-FKBP12 and MPS1-FRB at their endogenous loci, and either NUF2^WT or nuf2 mutants expressed from an ectopic locus. Since tethering Mps1-FRB to the C-terminus of Ndc80 constitutively activates the SAC, causing cell inviability⁶⁴, these strains also included a deletion of MAD3. Remarkably, tethering Mps1 to Ndc80 restored the viability of several nuf2 mutants, particularly nuf2^N128A (Figure 5B and Figure S5A). However, we observed a spectrum in the extent to which different nuf2 alleles could be rescued, ranging from complete viability restoration to no growth rescue for the nuf2^S124D mutant (commented on below; Figure S5A).

Figure 5. Increasing Mps1 concentration via kinetochore tethering or overexpression is sufficient to restore nuf2 mutant viability.

(A) Schematic of system to tether Mps1 to the kinetochore. In the presence of rapamycin, Mps1-FRB associates with Ndc80-FKBP12, artificially localizing Mps1 to the Ndc80c.

(B) Yeast cell viability assay using the rapamycin-induced tethering system. TOR1-1 fpr1Δ (alone, M1375), followed by strains that also contain MPS1-FRB and NDC80-FKBP12 tethering alleles (MAD3, M1463; mad3Δ, M4116) and NUF2-AID, with no nuf2 covering allele (“None”, M4841) or with ectopic copies of NUF2-3HA (NUF2^WT, M4908; nuf2^N128A, M4912), and finally a control nuf2^N128A strain lacking the NDC80-FKBP12 allele (M4955). Cells were serially diluted five-fold and spotted onto plates containing DMSO, 50 ng/mL rapamycin, 250 μM auxin or 250 μM auxin + 50 ng/mL rapamycin. See Figure S5A for additional nuf2 alleles.

(C) Yeast cell viability assay to assess the effect of pGAL-driven overexpression of Mps1 in nuf2 mutants. Strains contain NUF2-AID and ectopic copies of NUF2-3HA, without or with an ectopic pGAL-MPS1 single integration vector allele, respectively (NUF2^WT, M2038 and M3586; nuf2^N128A, M2414 and M3588; nuf2^{F8A P9A}, M2042 and M3587; nuf2^S124D, M2041 and M4619). Cells were serially diluted five-fold and spotted onto plates containing either glucose (control) or galactose (to induce over-expression of MPS1), as well as DMSO or 250 μM auxin. See Figure S5C for an additional nuf2 allele.

(D) Yeast cell viability assay, comparing pGAL-MPS1^WT and pGAL-mps1^RLR>EEE suppression of nuf2^N128A. Strains contain NUF2-AID and an ectopic copy of NUF2-3HA (NUF2^WT, M2038; nuf2^N128A, M2414) or also contain a single, ectopic pGAL-MPS1 allele (nuf2^N128A with MPS1^WT, M3588; nuf2^N128A with mps1^RLR>EEE, M4946). Cells were serially diluted five-fold and spotted onto plates containing either glucose or galactose, as well as DMSO or 250 μM auxin. See also Figure 5G and Figure S5G for additional mps1 alleles.

(E) Structure of S. cerevisiae Nuf2 (5TCS) illustrating the residues with mutants that are suppressed by expression of pGAL-MPS1 (Figure 5C, S5C). Darker shades of blue indicate increasing strength of the suppression. Note that F8A and P9A were tested individually for this analysis (data not shown).

(F) Cartoon depicting the experiment and result in (D). Left: In wild-type, Mps1 associates with the interaction hub on Nuf2, sustaining viability. Middle: The nuf2^N128A mutant is not viable, due to a loss of interaction between Mps1 and the hub. Overexpression of Mps1 leads to viability, suggesting that an increased cellular concentration of Mps1 allows enough interaction between Mps1 and Nuf2 at the partially disrupted binding interface. Right: In the mps1^RLR>EEE mutant, overexpression does not restore viability due to mutations on Mps1 that prevent association with the hub. This demonstrates that overexpression of Mps1 alone is not sufficient to suppress nuf2 interaction hub mutants.

(G) Yeast cell viability assay to compare the effect of overexpression of pGAL-MPS1^WT with pGAL-mps1 deletions on the growth phenotype of nuf2^N128A. Strains carry NUF2-AID, an ectopic copy of NUF2^WT-3HA or nuf2^N128A-3HA, and a single, ectopic pGAL-MPS1 allele (NUF2^WT, M2038; nuf2^N128A with no pGAL-MPS1 allele, “None”, M2414; with pGAL-MPS1^WT, M3587; with pGAL-mps1^Δ151-200, M4055; with pGAL-mps1^Δ201-300, M4056). A previous study implicated Mps1 residues 151-200 in biorientation and residues 201-300 in spindle pole body duplication⁵⁸.

See also Figure S5, S6.

The second way we asked if modulating Mps1 levels could suppress the associated nuf2 mutant defects was via overexpression. Compellingly, overexpression of Mps1 rescues the biorientation and cell viability defects of numerous nuf2 mutants (Figure 5C & Figure S5). However, as with the FRB-FKBP12 tethering experiments, there was a range in the extent to which the nuf2 mutant alleles could be suppressed by Mps1 overexpression. Importantly, the observed rescue is not due to restored SAC function^65–67, since both Mps1 overexpression and Mps1 tethering suppress nuf2 mutant viability defects even when the SAC is inactivated by deleting MAD3 (Figure 5B & Figure S5A,E). Furthermore, this suppression appears specific to Mps1, as overexpression of Ipl1 does not rescue the cell viability of these nuf2 mutants (Figure S5F).

Third, to modulate Mps1 levels at the kinetochore, we utilized an ndc80^ΔN-tail mutant that lacks the N-terminal tail of Ndc80. Prior in vitro analysis with human proteins demonstrated that the N-terminal tail of Ndc80 diminishes Mps1’s capacity to bind to the Ndc80c, likely due to competitive binding⁴³. Consistent with this idea, many of the nuf2 mutants at the described Mps1 binding interface that showed prominent growth defects in the context of NDC80^WT were suppressed by ndc80^ΔN-tail (Figure S6), raising the possibility that the N-terminus of Ndc80 associates directly with the interaction hub to regulate Mps1 localization.

Finally, to investigate whether Mps1’s ability to suppress nuf2 mutant growth defects relies on its interaction with the Ndc80c, we examined the effect of overexpressing the kinetochore binding-deficient mps1^RLR>EEE mutant. Compellingly, while overexpression of MPS1^WT effectively suppresses the cell viability defects of the nuf2 mutants, we found that the overexpression of mps1^RLR>EEE or other mutations at the Ndc80c interface fail to achieve the same outcome (Figure 5D–G and Figure S5G). Interestingly, we also observed a form of intragenic complementation with the spindle pole body duplication-deficient mps1^Δ201-300. While unable to maintain cell viability as the sole copy of Mps1 (Figure 3C), its overexpression in this context effectively compensates for the loss of kinetochore function (Figure 5G). These findings indicate that overexpression of Mps1 alone is insufficient to suppress the nuf2 mutants, but instead requires Mps1’s interaction with the Ndc80c. Thus, kinetochore-association of Mps1 is necessary for its role in biorientation, a process crucial for cell viability. These findings significantly advance our understanding of kinetochore-associated Mps1 functions.

In summary, our findings establish a pivotal “interaction hub” within the Ndc80c that orchestrates SAC signaling, microtubule attachment, and biorientation. Furthermore, they provide important mechanistic insights into the establishment of accurate chromosome attachments, and underscore the indispensable role played by the kinetochore-associated pool of Mps1 in ensuring precise chromosome segregation.

DISCUSSION

In this work, we set out to examine the role that Nuf2’s CH domain plays in the fidelity of chromosome segregation. Through extensive mutational analysis, we discovered a conserved “interaction hub” within Nuf2’s CH domain, formed by portions of its N-terminal loop and G-helix. Importantly, this patch serves as the binding site for Mps1 within the yeast Ndc80c, but also associates with other factors, including the Dam1c. Consequently, mutants disrupting this hub exhibit defects in SAC function and display severe errors in chromosome segregation that, surprisingly, are rescued by restoring Mps1-Ndc80c association. This work sheds light on mechanisms cells use to regulate Mps1 at the kinetochore and underscores the indispensable role that Mps1 plays in kinetochore biorientation to ensure precise chromosome segregation.

The previously uncharacterized region of Nuf2’s CH domain described here interacts with the N-terminus of the kinase Mps1, the C-terminus of the Dam1 protein, potentially the N-terminal tail of the Ndc80 protein and the N-terminus of the kinase Ipl1^19,22,42,43 (this work as well as concurrent work by Pleuger et al.59, Zahm et al.60), among other possible yet-to-be-identified factors. We propose that Mps1 is the most upstream and/or critical factor that binds to this region of the Ndc80c, as specific restoration of the Mps1-Ndc80c association, via Mps1 overexpression or tethering, rescues the cell viability defects of interaction hub mutants. Its sufficiency in this role suggests that Mps1 is capable of recruiting enough Dam1c to support accurate chromosome segregation, or that this specific Dam1c-Ndc80c interaction is dispensable for cell viability. Future work will be required to distinguish between these ideas. In any case, our identification of the Mps1 binding site on the Ndc80c highlights the essential role that kinetochore-bound Mps1 plays in error correction, in addition to its well-studied function in SAC activation.

We propose the following working model, built from this work and the significant work of others (reviewed in ^3,68), outlining the steps from initial microtubule attachment to biorientation (Figure 6). The Ndc80c putatively changes from a ‘closed’ to an ‘open’ conformation⁴³ (Figure S6C), likely facilitated by phosphorylation of Ndc80’s N-terminal tail by Ipl1 and Mps1^{7,11–13,27–31}. This may relieve the N-terminal tail of Ndc80 from occluding binding of other factors⁴³. Mps1 subsequently associates with the Ndc80c interaction hub where it activates the SAC^{33–35,37,44,45}. Importantly, this pool of Mps1 also appears essential for achieving kinetochore biorientation and ensuring accurate chromosome segregation, likely through a tension-sensitive error correction mechanism^31–37, the precise nature of which remains unclear (further discussion below). In part, this error correction function of Mps1 may involve recruiting the Dam1c^31,36,62,63. Once attachments are correctly bioriented, the Dam1c establishes stable association with the Ndc80c via the same interaction hub, thereby displacing Mps1 from its binding site^59,60 (personal communication S. Harrison and S. Westermann). In this manner, Mps1 seems to facilitate the recruitment of its own competitive inhibitor. Other mechanisms previously described to regulate Mps1 kinetochore localization include autophosphorylation^59,69 and competitive binding between Mps1 and microtubules for Ndc80c association^42,43. These collective events ultimately deactivate SAC signaling and also likely terminate Mps1’s participation in error correction. While this model provides insights into certain aspects of this process, several crucial questions remain.

Figure 6. Model of the transition from unattached kinetochore to bioriented attachment.

In the unattached kinetochore state, the interaction hub on Nuf2 (yellow) is available for interaction with Mps1 (purple) and/or the Dam1 complex (dark grey). The microtubule is represented in tan. The interaction hub may initially exist in a “closed state” in which the N-terminus of Ndc80 folds over Nuf2, occluding the interaction hub (Figure S6C) and limiting access of other factors. Movement of the Ndc80 N-terminal tail away from the hub (possibly via phosphorylation by Ipl1 or Mps1) could then generate an “open state”, allowing access to Mps1 and the Dam1c⁴³ (Figure S6).

1. Mps1 associates with the Ndc80 complex at the Nuf2 interaction hub, triggering the SAC and facilitating sister chromatid biorientation through an as yet undetermined mechanism that is likely tension-sensitive error correction.
2. Mps1 also promotes association of the Dam1 complex with the Ndc80 complex, which ultimately allows for stable kinetochore-microtubule interaction and the formation of bioriented attachments.
3. Once the bioriented state is achieved, Mps1 is removed from the kinetochore via competition by the Dam1 complex for the interaction hub^59,60, autophosphorylation⁶⁹ and/or competition with microtubule binding^42,43. It is worth noting that Mps1 binding does not seem to be mutually exclusive to microtubule binding with purified yeast kinetochores in vitro⁶⁹. Loss of Mps1 at the kinetochore also turns off the SAC.

One key question is how Mps1’s activity at kinetochores contributes to biorientation. Previously proposed mechanisms involve either direct phosphorylation of key proteins at the kinetochore-microtubule interface or indirect regulation through controlling the localization of downstream factors. Regarding direct mechanisms, it has been well-demonstrated that Mps1 phosphorylates the N-terminal tail of Ndc80 on kinetochores lacking tension, leading to attachment destabilization³¹. However, disrupting the known Mps1 phosphorylation sites, even in conjunction with a SAC mutant, does not substantially impact cell viability^30,31. Furthermore, the N-terminal tail of the Ndc80 protein is dispensable for viability in yeast^30,70,71. Concerning indirect pathways by which Mps1 may facilitate biorientation, Mps1 has been implicated in activation of the kinase Bub1^32,72 and in recruiting Sgo1 to the pericentromere^72–77, both of which play a role in biorientation but are are non-essential proteins. The fact that these previously described mechanisms are dispensible for cell viability contrasts with the nuf2 and mps1 mutants described here, which exhibit complete inviability, suggesting additional mechanisms must be involved in this process.

An important insight into the requirements for Mps1 to promote biorientation may have come from our examination of the interaction hub mutants. We suggest that the observed spectrum of nuf2 mutant phenotypes results from varying levels of Mps1 binding deficiency. Our assays, unfortunately, lack sensitivity to discern these differences. However, we posit that viability can only be restored in mutants retaining some degree of association with Mps1. The nuf2^N128A, nuf2^{F8A P9A}, and nuf2^S124D mutants appear to constitute an allelic series exhibiting varying degrees of penetrance, with nuf2^N128A being the least disruptive and nuf2^S124D being the most disruptive to biorientation, chromosome segregation, and cell viability (Figure 1B, Figure 2B–C , Figure 5B–C, Figure S5A). Importantly, none of these alleles result in a complete loss of function concerning protein folding or microtubule binding, as the Ndc80c remains intact in all mutants (Figure 3A), and non-disjoined sister chromatids were observed in mother and daughter cells, necessitating kinetochore-microtubule attachments (Figure S2G–I). Likewise, the nuf2 alleles exhibit varying degrees of rescue when modulating Mps1 levels at the kinetochore, particularly through tethering Mps1-FRB to Ndc80-FKBP, that could also reflect differences in their capacity to associate with Mps1 (Figure 5B–C & Figure S5A). The full viability rescue in the nuf2^N128A mutant may indicate only a partial disruption of Mps1’s binding site, such that when the local concentration of Mps1 is increased by overexpression or tethering, Mps1 can bind the interaction hub. In contrast, we suggest the lack of rescue observed in the nuf2^S124D mutant is due to a more significant disruption of the Mps1 binding site that cannot be rescued by additional Mps1, likely by the introduction of a negative charge into a hydrophobic interaction surface, partially displacing Nuf2’s N-terminal loop from its G-helix. Alternatively, it is conceivable that Mps1 function is restored when tethered to the kinetochore in all nuf2 mutants, and that the variability in rescue depends on binding differences of another factor, such as the Dam1c. However, this idea is inconsistent with our kinetochore IP data, which indicate that of the nuf2 mutants, nuf2^S124D exhibits the highest level of residual co-purified Dam1 (Figure 4A). If nuf2 mutants do differ in their Mps1 binding capacity, this suggests the ability of Mps1 to perform its biorientation function is not solely determined by its kinetochore-proximal concentration, but rather by its binding directly at the Ndc80c interaction hub. Interestingly, based on the work of Aravamudhan et al⁶⁴, it appears that a less specific localization of Mps1 near Spc105 is sufficient for activation of the SAC. Thus, distinct mechanisms may govern Mps1 SAC activation and biorientation activities.

Why does Mps1 putatively require specific binding to the interaction hub to execute its function? One possibility to explain this requirement is that Mps1’s kinase activity is regulated through simple clustering and auto-transactivation of the kinase at this site. More intriguingly, this particular interface could facilitate activation via allostery or a similar mechanism, a notion that gains support from the observation that the assembly of the outer kinetochore is essential for generating most, if not all, cellular Mps1 activity⁵⁶. Another critical question concerns the mechanisms that regulate these activities in response to tension. Given the proximity of Mps1 to the microtubule binding domains of the kinetochore, it seems unlikely that spatial separation of substrates from Mps1, as proposed for Aurora B⁷⁸, could explain tension sensitivity. Instead, tension-dependent changes in substrate conformation, as recently demonstrated for Ipl1⁷⁹, provide a plausible mechanism by which tension suppresses phosphorylation of critical targets. Alternatively, Mps1’s activity or localization could be influenced by a kinetochore factor, and this regulation may be responsive to tension. For instance, displacement of Nuf2’s N-terminal loop from its G-helix, either through tension or another mechanism, presents an intriguing way that cells could modulate Mps1 activity. Lastly, the interplay between the Aurora B-mediated and Mps1-dependent pathways in biorientation remains an unresolved question that will be crucial to investigate in future studies. While our inclination is toward Mps1 playing a direct role in promoting biorientation via a tension-sensitive error correction pathway, our discovery of its specific binding site on the yeast Ndc80c equips us with critical tools to explore these questions in future research. This includes investigating the direct versus indirect roles of Mps1 and elucidating the mechanism of tension sensitivity.

In summary, our work identifies a critical “interaction hub” that governs kinetochore-localized functions of Mps1. These functions encompass its role in SAC signaling, but, notably, emphasize the significance of Mps1 in establishing kinetochore biorientation. Our findings clarify a site of Mps1 localization at the kinetochore and its contribution to the fidelity of chromosome segregation.

STAR METHODS

Resource Availability

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Matthew Miller (matt.miller@biochem.utah.edu).

Materials Availability

Plasmids and S. cerevisiae strains made and used in this study will be made available by the lead contact, Matthew Miller, upon request.

Data and code availability

All data utilized in this paper will be shared by the lead contact upon request. Raw microscopy data and original western images have been deposited at https://osf.io/de743/ (DOI 10.17605/OSF.IO/DE743) and are publicly available as of the date of publication. The DOI is listed in the key resource table.

Key Resources Table

REAGENT or RESOURCE		SOURCE>	IDENTIFIER
Antibodies
Mouse monoclonal α-FLAG		Sigma Aldrich	Cat#F3165
Mouse monoclonal α-HA		Roche	12CA5; Cat#11583816001
Mouse monoclonal α-V5		Invitrogen	Cat#R960-25; RRID: AB_2556564
Mouse monoclonal α-Pgk1		Abcam	Cat#22C5D8; abID:ab113687
Rabbit polyclonal α-Ndc80		Desai lab
Sheep α-mouse (HRP)		Cytiva/GE Biosciences	Cat#NA931-1ML; Lot#17697705
Donkey α-rabbit (HRP)		Cytiva/GE Biosciences	Cat#NA934-1ML
Chemicals, peptides, and recombinant proteins
Auxin		Sigma Aldrich	Cat#I3750-5G-A; CAS: 87-51-4
Benomyl		Sigma Aldrich	Cat#381586; CAS: 17804-35-2; Lot#MKCG8313
Nocodazole		Calbiochem	Cat#487928; CAS: 31430-18-9; Lot#B35705
α-factor		University of Utah Core Synthesis Facility
Rapamycin		LC Laboratories	Cat#R5000
Formaldehyde		Fisher Chemical	Cat#F79-500
DAPI (4′,6-Diamidino-2-Phenylindole, Dihydrochloride)		Molecular Probes	Cat#D1306
SuperSignal West Dura Chemiluminescent Substrate		Thermo Scientific	Cat#PI34076; Lot#YD366977
Deposited data
Raw data		This study	https://osf.io/de743/ DOI 10.17605/OSF.IO/DE743
Experimental models: Organisms/strains
S. cerevisiae: Strain background: W303; See Table S1		This work
Oligonucleotides
See Table S2
Recombinant DNA
See Table S2
Software and algorithms
softWoRx		Cytiva/GE Biosciences
ChimeraX 1.5	Goddard et al 201880	https://www.cgl.ucsf.edu/chimerax/download.html
FIJI (ImageJ)		Schindelin et al 201281	https://imagej.net/software/fiji/downloads
AlphaFold2 (ColabFold)		Mirdita et al 202282	https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb
PSI-BLAST		Altschul et al 199783	https://blast.ncbi.nlm.nih.gov/Blast.cgi?CMD=Web&PAGE=Proteins&PROGRAM=blastp&RUN_PSIBLAST=on
Clustal Omega		Madeira et al 202284	https://www.ebi.ac.uk/jdispatcher/msa/clustalo
Jalview 2		Waterhouse et al 200985	https://www.jalview.org/download/
Prism v. 10.1.1		GraphPad Prism	https://www.graphpad.com/features
Other
Dynabeads Protein G		Invitrogen	Cat#10009D

This paper does not report original code.

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Experimental Model and Study Participant Details

All yeast strains used are listed in Table S1 and are isogenic in the W303 background⁸⁶. Standard media and microbial techniques⁸⁷ were used, as well as standard yeast genetic methods for strain construction^88,89.

Single integration vectors (SIV) pM93 (HIS3), pM94 (URA3), pM95 (TRP1), and pM96 (LEU2) were provided by Leon Chan and were used in the construction of the NUF2, NDC80, MPS1 and IPL1 plasmids described below. The genomic NUF2-AID allele NUF2-3V5-IAA7:KanMX was constructed using pM66 (3V5-IAA7:KanMX plasmid; Leon Chan) as a template, and subsequent conversion to NUF2-3V5-IAA7:NATMX was performed as in ⁹⁰. A single integration LEU2 vector containing pNUF2-NUF2-3HA (pM604) was constructed by first tagging endogenous NUF2 with 3HA from pM9 (3HA:HIS3MX6)91, then using the pNUF2-NUF2-3HA strain as a template for plasmid construction. Mutants of NUF2 were constructed by mutagenizing pM604 as described in ^92,93. The NDC80-3V5-IAA7:KanMX is described in ³⁹. A TRP1 single integration vector containing pNDC80-NDC80-3HA (pM270) was constructed in a similar manner to NUF2-3HA above. The pNDC80-ndc80^ΔN-tail-3HA, pNDC80-ndc80^K204E-3HA and pNDC80-ndc80^{ΔN-tail K204E} were subsequently constructed using pM270 as a template for mutagenesis as in ^92,93.

The MPS1-3V5:KanMX allele was constructed using pM32 (V5:KanMX tagging vector; Leon Chan) as a template, and was marker converted to MPS1-3V5:HphMX as in ⁹⁰. pGAL10-MPS1 and pGAL10-IPl1 HIS3 single integration vectors were constructed by genomic PCR amplification of the respective ORFs, followed by Gibson assembly into pM93, downstream of the pGAL10 promoter⁹⁴. The pGAL10 promoter was replaced with genomic PCR-amplified native MPS1 promoter to construct pM1595 (pMPS1-MPS1 HIS3 SIV), and subsequent cloning produced pM1718 (pMPS1-MPS1 URA3 SIV). Mutants of MPS1 were constructed in these plasmids by standard cloning or by mutagenizing pM1489, pM1595 and pM1718 as described in ^92,93. The pMPS1-MPS1-3V5 and pMPS1-MPS1-mNeonGreen SIVs were constructed by mutagenesis of MPS1^WT and mps1^RLR>EEE plasmids as in ⁹³, using a 3V5 PCR product amplified from the endogenous MPS1-3V5 strain described above or PCR of a pFA6a-mNeonGreen HIS3MX6 plasmid. The pGAL10-MPS1 pRS402-based ADE2 Yip was constructed using standard cloning techniques. All plasmids and primers are listed in Table S2. Further details regarding plasmid and strain construction are available upon request.

Integration plasmids containing pGPD1-TIR1 (pM74 for integration at LEU2, pM76 for HIS3, and pM78 for TRP1) were provided by Leon Chan; the URA3 version was constructed from pM74 and the pM94 URA3 single integration vector. SPC110-mCherry:HphMX and ASK1-YFP:HIS3 were provided by Trisha Davis, and BUB1-GFP:KanMX, MTW1-mCherry:HphMX by Sue Biggins. Construction of pCUP1-GFP-Lacl is described in ⁹⁵, CEN III::lacO:TRP1 is described in ⁹⁶, and CEN IV::lacO:TRP1 was provided by Andrew Murray. Strains containing the previously described pMET-CDC20 allele were provided by Frank Uhlmann⁹⁷, and the CDC20-AID allele by Adele Marston. mps1-1 and dad1-1 temperature sensitive alleles were provided by Andrew Murray and Sue Biggins, respectively. The pGAL-MPS1-Myc:URA3 and mad2Δ alleles were provided by Andrew Murray, mad3Δ is described in ⁵⁴, and DSN1-6His-3Flag in ⁵⁷. TOR1-1, fpr1Δ, MPS1-FRB:KanMX, and NDC80-FKBP12:HISMX are described in ^64,98. The Dam1-9Myc allele was provided by Sue Biggins.

Method Details

Auxin inducible degron

The auxin inducible degron (AID) system was used as described in ⁹⁹. Cells expressed C-terminal fusions of NUF2, NDC80, or CDC20 to the gene for an auxin responsive protein (IAA7) at each respective endogenous locus. Cells also expressed TIR1, which is required for auxin-induced degradation. For cell viability assays, 250 μM auxin (indole-3-acetic acid; Sigma Aldrich) dissolved in DMSO was top-plated on agar to induce degradation of the AID-tagged protein. Cells for microscopy were treated with 500 μM auxin in liquid media at the specified time prior to fixation. Auxin was added to liquid media 2 hours prior to harvesting cells for immunoprecipitation analysis.

Spot dilution yeast viability assays

Desired strains were grown overnight for two days on YPA + 2% dextrose plates (YPAD). Equal amounts of each strain were then used for a 1:5 dilution series and spotted onto YPAD + DMSO, YPAD + auxin, YPAD + auxin and 6.5 μg/mL benomyl, YPAD + 50 ng/mL rapamycin, YPAD + auxin and rapamycin, YPA Gal (YPA + 2% galactose) + DMSO, or YPA Gal + auxin. Plates were incubated at 23°C for 2 to 4 days, unless otherwise indicated for temperature sensitive alleles (30°C or 37°C, as noted in figure legends).

Cell fixation, imaging conditions and image analysis

Conditions for growing cells for each type of imaging are described in brief in the figure legends. For most experiments, cells were grown to log phase in standard YPAD liquid medium at room temperature (approximately 20°C). Where indicated, cells were treated with auxin (500 μM) or auxin and nocodazole (10 μM). In some cases, an α-factor arrest was used to synchronize cells (1 μg/mL for 3 hours at room temperature; using bar1-1 strains), followed by three washes in YPAD + 2% DMSO and resuspension in fresh YPAD media for the release. For the biorientation assay in Figure 2B, cells were arrested with α-factor in synthetic liquid medium lacking methionine, then released into the pMET-CDC20 arrest in the presence of methionine at 30°C. 8 mM methionine was subsequently added every thirty minutes until harvest. For the biorientation rescue assay in Figure S5B, cells were grown in synthetic liquid supplemented with 2% raffinose rather than dextrose. Galactose was added to a concentration of 2% for 1 hour prior to release of the pMET-CDC20 arrest by addition of methionine at 30°C, as indicated above. For this assay, cells were not previously synchronized in G1 (via α-factor treatment), which led to variation in the time cells were arrested in metaphase, thereby likely reducing the observed percent bioriented.

Fixation was performed in 3.7% formaldehyde in 100 mM phosphate buffer (pH 6.4) for 5 minutes. Cells were washed once with 100 mM phosphate (pH 6.4) and resuspended in 100 mM phosphate, 1.2 M sorbitol buffer (pH 7.5) and permeabilized with 1% Triton X-100 stained with 1 μg/mL DAPI (4’, 6-diamidino2-phenylindole; Molecular Probes). Cells were imaged with a DeltaVision Ultra microscope with a 60X objective (NA = 1.42), equipped with a sCMOS digital camera. Fifteen Z-stacks (0.3 micron apart) were acquired, and frames were deconvolved using standard settings. For each experiment, acquisition settings for each channel were kept consistent between images and samples. Image stacks were maximally projected. softWoRx image processing software was used for image acquisition and processing.

Quantitation of microscopy images was performed as indicated in the figure legends. For the Bub1-GFP assay for SAC activity, only mononucleate cells with a single Mtw1-mCherry signal were counted (indicating collapse of the spindle from nocodazole treatment). Note that nuf2 mutants often increased the number of cells with two Mtw1 signals, despite nocodazole treatment, and also displayed increased bypass of the CDC20-AID arrest (Figure S2E–H). Collectively, this reduced the number of cells in each field of view available for scoring the presence of Bub1-GFP at the kinetochore relative to NUF2^WT. Numbers of cells counted are indicated in the figure legend. For the biorientation assays in Figure 2B and S5B, only mononucleate cells in metaphase were counted. For chromosome segregation assays in Figure 2C and Figure 3E, only binucleate cells that had undergone anaphase (approximate spindle length ≥ 4 μM) were counted.

Spindle length in Figure S2E was calculated by drawing a line in FIJI (Image J, NIH⁸¹) that connected the edges of the Spc110-mCherry signals from the two poles. To categorize Mps1-mNeonGreen signals from Figure 3D into “Strong”, “Weak” and “No signal”, the green signal intensity of the strongest visible Mps1 dot within ~2μM of the Spc110 spindle pole signal from individual cells of MPS1^WT and mps1^RLR>EEE was measured by calculating the raw integrated density (RID) of circles of uniform size in FIJI. An identical region with no visible signal was subtracted as background. Comparison of RID background-subtracted values with visual inspection of signal in >100 cells in the images was used to delineate the three categories of signal, with “strong” represented by signals >100,000 RID, “weak” by signals 25,000-100,000 and “no signal” by those <25,000 and/or not detectable by visual inspection. Quantitation of additional cells and replicates were then performed by RID measurements and spot-checked by visual confirmation. Spindle pole body separation in Figure S3E was calculated by the percentage of cells that displayed two Spc110-mCherry signals at 90 minutes following the temperature shift. Ask1-YFP intensities in Figure 4C were calculated by drawing a line in FIJI through the Spc110 spindle pole signals and performing plot profile analysis of the YFP signal across this line. The maximum peak intensity proximal to each pole was then determined (two signals for each cell, which were sometimes asymmetric, especially in the nuf2 mutant). Data points in the right graph in Figure 4C represent individual Ask1-YFP peak intensities, two for each cell.

Kinetochore immunoprecipitations

Native kinetochores were purified from asynchronously, exponentially growing S. cerevisiae cells containing Dsn1-6His-3Flag by immunoprecipitation with α-Flag, essentially as described in ⁵⁷. Cells were grown in standard YPAD medium. For strains containing NUF2-AID, cells were treated with 500 μM auxin 2 hours prior to harvest. For examining kinetochore enrichment of Dam1-9Myc in WT and mps1-1 strains, cells were shifted to 37°C for 2 hours prior to harvest. Protein lysates were prepared by mechanical disruption in the presence of lysis buffer using glass beads and a beadbeater (Biospec Products). Lysed cells were resuspended in buffer H (BH) (25 mM HEPES pH 8.0, 2 mM MgCl₂, 0.1 mM EDTA, 0.5 mM EGTA, 0.1% NP-40, 15% glycerol with 150 mM KCl) containing protease inhibitors (at 20 μg/mL final concentration for each of leupeptin, pepstatin A, chymostatin and 200 μM phenylmethylsulfonyl fluoride) and phosphatase inhibitors (0.1 mM Na-orthovanadate, 0.2 μM microcystin, 2 mM β-glycerophosphate, 1 mM Na pyrophosphate, 5 mM NaF) followed by centrifugation at 16,100 g for 30 min at 4°C to clarify the lysate. Dynabeads conjugated with α-Flag antibodies were incubated with extract for 3 hours with constant rotation, followed by three washes with BH containing protease inhibitors, phosphatase inhibitors, 2 mM dithiothreitol (DTT) and 150 mM KCl. Beads were further washed twice with BH containing 150 mM KCl and protease inhibitors. Associated proteins were eluted from the beads by boiling in 2x SDS sample buffer.

Immunoblotting

For immunoblot analysis in Figure S1D, S3D, and S5G, cell lysates were prepared by mechanical disruption with glass beads in 2x SDS-PAGE sample buffer (0.4mM EDTA, 4% SDS, 125 mM Tris pH 6.8, 20% glycerol, 2% bromophenol blue). Lysates for immunoprecipitation were described above. Standard procedures for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting were followed as described in ^100,101. A nitrocellulose membrane (Bio-Rad) was used to transfer proteins from polyacrylamide gels. Commercial monoclonal antibodies used for immunoblotting were as follows: α-HA, 12CA5 (Roche, 1:2000), α-Flag, M2 (Sigma Aldrich, 1:3000), α-V5 (Invitrogen; 1:5000), α-Pgk1 (Abcam; 1:5000). The Ndc80 polyclonal antibody was a gift from Arshad Desai and was used at 1:10,000. Secondary antibodies used were a sheep α-mouse and donkey α-rabbit conjugated to horseradish peroxidase (HRP, GE Biosciences; 1:10,000 dilution). Antibodies were detected using the SuperSignal West Dura Chemiluminescent Substrate (Thermo Scientific).

Multiple sequence alignments

Fungal Nuf2 and Mps1 proteins were identified using a PSI-BLAST⁸³ search on NCBI. Multiple sequence alignments of the entire proteins were generated with ClustalOmega⁸⁴ default parameters and displayed in JalView 2⁸⁵. Species are listed in the figures and/or legends. Conservation scores shown on structures were generated in ChimeraX 1.5⁸⁰, using an entropy-based measure from AL2CO¹⁰².

Quantitation and Statistical Analysis

GraphPad Prism version 10.1.1 was used for statistical analysis. Data normality was assumed for all experiments. Student’s t test was used for comparisons between strains, as indicated in the figure legends. In the text, mean ± standard deviation is reported. Statistical significance is indicated on figures with asterisks. All other statistical details are in figure legends.

Supplementary Material

2

Table S1. Strains used in this study. Related to STAR Methods.

All strains are derivatives of W303 (M3).

3

Table S2. Plasmids and primers used for strain construction. Related to STAR Methods.

Highlights.

• Two segments of Nuf2’s CH domain recruit the Mps1 kinase to kinetochores

• Other factors, including the Dam1 complex, also bind this “interaction hub”

• Kinetochore-bound Mps1 promotes recruitment of the Dam1 complex

• Association of Mps1 with the hub is essential for biorientation and cell viability