An interaction hub 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 accurate chromosome segregation. Through extensive mutational analysis, we identified a conserved “interaction hub” comprising two segments in Nuf2’s CH domain, forming the binding site for Mps1 within the yeast Ndc80 complex. Intriguingly, the interaction between Mps1 and the Ndc80 complex seems to 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.

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 in¹). 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 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 in the field.

At the core of these essential cellular functions lies the Ndc80 complex, 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. Furthermore, 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 Ndc80 complex 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 Ndc80 complex associates with the Dam1 complex 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 the error correction kinases, namely Ipl1 and Mps1^{7,11–13,27–38}. Additionally, the Ndc80 complex directly interacts with additional factors implicated in correcting erroneous attachments, including Stu2/chTOG^39–41. Lastly, the Ndc80 complex assumes a pivotal role in triggering the spindle assembly checkpoint by recruiting the most upstream component, the kinase Mps1^{33–35,37,42–48}. A significant barrier to understanding how these diverse Ndc80 complex-dependent functions are coordinated lies in our limited understanding of how these different factors interact with the Ndc80 complex 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 a key portion of the Mps1 binding interface within the yeast Ndc80 complex. This same region of Nuf2 associates with additional factors, including the Dam1 complex, suggesting that Mps1’s kinetochore localization may be regulated by competitive binding. Mutants that disrupt this interaction hub exhibit defects in spindle assembly checkpoint function and severe chromosome segregation errors. Remarkably, Mps1 appears to be the critical factor that binds to this region, since the cellular defects can be rescued by specifically restoring Mps1-Ndc80 complex association. Our work sheds light on the manner in which numerous Ndc80 complex-dependent functions are regulated and demonstrates the essential role of Mps1 in kinetochore biorientation and chromosome segregation.

RESULTS

Identification of a Nuf2 patch 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 first looked for spatial patterns of amino acid conservation within Nuf2’s CH domain. Here, we aligned Nuf2 sequences from various fungal species and then mapped the resulting conservation scores onto a previously determined Ndc80 complex structure⁴⁹ (PDB: 5TCS). This analysis revealed a region of conserved residues in close spatial proximity, which is centered on Nuf2’s so called “G-helix” as well as its amino-terminal “loop” (Fig 1A & Fig S1A). These residues, especially F8, P9, R118, S124, N128 and R131, appear highly conserved in fungal and metazoan Nuf2 sequences (Fig 1A & Fig 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 phenotype 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. We found that many of the mutations resulted in severe cell viability defects, and the most penetrant phenotypes were associated with mutations within the Nuf2 G-helix (nuf2^S124D, nuf2^A125D, nuf2^N128A and nuf2^R131E) as well as N-terminal loop (nuf2^{F8A P9A}; Fig 1B & Fig S1D). Importantly, these mutations do not appear to affect Nuf2 expression levels (Fig S1C). A summary of all mutants and associated viability phenotypes can be found in Fig 1C and Fig S1D. Based on the severity of viability defects, amount of conservation, and degree of spatial proximity, we chose to focus on mutations nuf2^{F8A P9A}, nuf2^S124D, and nuf2^N128A 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 from⁴⁹ (5TCS), with other members of the Ndc80 complex not shown. Left: Conservation of Nuf2 among 12 fungal species (listed in Fig S1), viewed using ChimeraX, 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).

(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.

(C) Structure of Nuf2 (5TCS), illustrating the range of cell viability phenotypes of mutant alleles observed on plates containing benomyl (data shown in Fig S1D). 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).

Spindle assembly checkpoint function is perturbed in nuf2 patch mutants

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^{F8A P9A} and nuf2^N128A, display a considerably more severe viability defect than ndc80 mutants which perturb microtubule binding^7,10 (i.e. ndc80^K204E and ndc80^{K122E K204E}; Fig 2A and Fig S2). Furthermore, while ndc80^{K122E K204E} mutants arrest in mitosis¹⁰, likely by activating the spindle assembly checkpoint, the nuf2 mutants did not induce a specific cell cycle arrest (Fig 2B). These observations suggest that these nuf2 CH domain mutations affect a kinetochore function beyond just microtubule binding. Given the lack of cell cycle arrest, we next assessed spindle assembly checkpoint (SAC) function by examining the localization of the SAC component, Bub1, which is recruited to kinetochores upon SAC activation^50–53. For this assay, CDC20-AID cells were arrested in metaphase by the addition of auxin. Once arrested in metaphase, cells were treated with nocodazole, and co-localization of Bub1-GFP with the kinetochore marker Mtw1-mCherry was monitored. Approximately 72 ± 6% of NUF2^WT cells showed Bub1-GFP recruitment to the kinetochore upon nocodazole treatment, consistent with prior observations^51,53. In contrast, nuf2^{F8A P9A} and nuf2^N128A showed a dramatic decrease in the number of cells with Bub1 at kinetochores (14 ± 5.5% and 14 ± 3.5%, respectively; Fig 2C), consistent with defects in spindle assembly checkpoint function. We also note that Mtw1-mCherry distribution appeared normal in the nuf2 mutants, suggesting that these mutations do not cause gross defects in kinetochore organization (Fig 2C).

Figure 2. Lethal nuf2 mutant alleles fail to trigger the spindle assembly checkpoint.

(A) Yeast cell viability assay comparing nuf2 mutants with previously characterized ndc80 alleles. Strains carry either NUF2-AID or NDC80-AID at their endogenous loci, and ectopic copies of NUF2–3HA (None, M1889; NUF2^WT, M2038; nuf2^{F8A P9A}, M2042; nuf2^N128A, M2414) or NDC80–3HA (None, M4182; NDC80^WT, M692; ndc80^K204E, M693; ndc80^{K122E K204E}, M694). Cells were serially diluted five-fold and spotted onto plates containing DMSO, 250 μM auxin or 250 μM auxin + 6.5 μg/mL benomyl.

(B) 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^{F8A P9A}, M2935; nuf2^S124D, M2936; nuf2^N128A, M2945) 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).

(C) Bub1-GFP assay for SAC activity. Exponentially growing NUF2-AID CDC20-AID strains (NUF2^WT, M2792; nuf2^{F8A P9A}, M2794; nuf2^S124D, M2795; nuf2^N128A, M2801) were treated with 500 μM auxin for 2 hours to arrest cells in metaphase, followed by 10 μM nocodazole for 1 hour to activate the SAC. Strains carry BUB1-GFP (SAC) and MTW1-mCherry (kinetochore) alleles. Left: Representative micrographs. Right: Quantitation of percent of large-budded, mononucleate cells with Bub1-GFP localized at the kinetochore, with error bars indicating the standard deviation among three replicates (n=66–112 cells for each replicate). Significance was determined by a two-tailed unpaired t test (***; P ≤ 0.001). Only cells with a single Mtw1 signal, indicating collapse of the spindle, were analyzed.

Mutants within Nuf2’s patch result in substantial chromosome segregation defects

While these nuf2 CH domain mutants display spindle assembly checkpoint defects, the lethal phenotypes observed cannot be explained by this limitation alone since the SAC is not essential for cell viability in yeast (Fig 3A). 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 III⁵⁴. 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 (44 ± 6%, Fig 3B). However, in nuf2 mutant cells, we observed a marked reduction in biorientation of CEN III (nuf2^{F8A P9A}, 10 ± 1.5%; nuf2^S124D, 6.6 ± 1.5%; and nuf2^N128A, 18 ± 1.5%). Furthermore, we noted a significant portion of nuf2 mutant cells bypassing the arrest induced by Cdc20 depletion (Fig S3A). In these nuf2 mutants, spindle elongation led exclusively to instances of sister chromatid non-disjunction and unequal DNA masses (Fig S3B–C). These findings can be attributed to substantial deficiencies in biorientation, 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 3. Inviable nuf2 mutants display a loss of sister chromatid biorientation and mis-segregation of chromosomes.

(A) Comparison of nuf2 mutant cell viability relative to the mad2 SAC mutant. Strains carrying a NUF2-AID and ectopic copies of NUF2–3HA (NUF2^WT, M2038; nuf2^{F8A P9A}, M2042; nuf2^N128A, M2414) are shown in comparison to mad2Δ (M35). Cells were serially diluted five-fold and spotted onto plates containing DMSO, 250 μM auxin, 6.5 μg/mL benomyl or both auxin and benomyl.

(B) Assay for sister chromatid biorientation in nuf2 mutants. Exponentially growing cells carrying NUF2-AID and ectopic copies of NUF2–3HA (NUF2^WT, M4245; nuf2^{F8A P9A}, M4251; nuf2^S124D, M4249; nuf2^N128A, M4253), pMET-CDC20, and CEN III marked with GFP (CEN III:lacO LacI-GFP) were arrested with 1 μg/mL alpha 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. 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). 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^{F8A P9A}, M4475; nuf2^N128A, M4479), 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 alpha factor for 3 hours, followed by release into fresh media containing 500 μM auxin for 2 hours. Left: Representative micrographs. 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).

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^{F8A P9A} and nuf2^N128A displayed surprisingly high errors in segregating chromosome IV (NUF2^WT 96 ± 4.6% correct segregation compared to nuf2^{F8A P9A} 4 ± 3.1%, nuf2^N128A 14 ± 3.5%, Fig 3C and Fig S3D). Together, these observations show that the region of Nuf2’s CH domain, consisting of its G-helix and N-terminal loop, are required for spindle assembly checkpoint function and essential for biorientation and accurate chromosome segregation.

Mps1-kinetochore localization perturbed in nuf2 patch mutants

The observed 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 spindle assembly checkpoint 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 binding site for Mps1^30,42,43,55. 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^31,56. Kinetochores were immunoprecipitated via anti-Flag pulldown of the kinetochore component Dsn1-Flag, and co-purification of Mps1-3V5 was determined by western blot. As expected, NUF2-AID cells (treated with auxin 2 hours prior to harvesting) expressing NUF2^WT showed Mps1 co-purification with kinetochores. In contrast, nuf2^{F8A P9A}, nuf2^S124D, and nuf2^N128A mutants displayed nearly undetectable Mps1 binding but preserved Ndc80 association (Fig 4A). We propose this region of Nuf2 forms a key part of the Mps1 binding interface to the yeast Ndc80 complex. These results are also consistent with Yu and colleagues who made a similar observation for human Mps1 binding to Ndc80 complex⁴³, suggesting this is a remarkably conserved interface.

Figure 4. 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 (No tag, M2630) as well as 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 anti-Flag immunoprecipitation (IP) and analyzed by immunoblotting.

(B) Cartoon of the Mps1 protein and a conservation histogram for amino acids 126 to 190. The region from 151 to 171 is shown in detail, with amino acids mutated in this work 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). Cells were serially diluted five-fold, spotted onto plates, and grown at mps1-1 permissive (23°C) and non-permissive (37°C) temperatures.

(D) 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 alpha 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. 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).

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 Ndc80 complex, we sought conserved regions within Mps1 outside its C-terminal kinase domain and identified two highly conserved patches encompassing residues 151–157 and 165–171 (Fig 4B). Earlier studies associated residues 151–200 with Mps1’s SAC and biorientation functions, while finding residues 201–300 are crucial for spindle pole body duplication⁵⁷. To assess the functional significance of these conserved Mps1 segments, we initially examined cell viability. Here, 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). Consistent with previous observations, mps1^Δ151−200 and mps1^Δ201−300 mutants were inviable in this system⁵⁷ (Fig 4C). While neither mps1^Δ151−157 nor mps1^Δ165−171 fully disrupted cell viability, combining these deletions or introducing point mutations spanning one or both regions resulted in complete inviability (Fig 4C and Fig S4A). The notion that this portion of Mps1 facilitates binding to the Ndc80 complex is well-supported by concurrent studies from Pleuger et al.⁵⁸ and Zahm et al., employing biochemical and structural approaches to determine the Mps1-Ndc80 complex interface (personal communication S. Westermann and S. Harrison).

Based on the above observations we decided to further characterize the mps1^{R155E L165E R170E} mutant (hereafter designated as “mps1^RLR>EEE”), focusing 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 Mps1⁵⁹ remains functional in these mutants (Fig S4B). 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%, Fig 4D and Fig S4C). It is worth noting 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, Fig 3C and Fig 4D). This discrepancy may be due to the mps1^RLR>EEE mutant not completely disrupting Mps1’s ability to bind the Ndc80 complex. 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 altered in nuf2 CH domain mutants

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 Ndc80 complex as a potential interaction site for the Dam1 complex^19,20. To examine if the nuf2 CH domain mutants disrupt Dam1 recruitment to the kinetochore, we assessed the localization of a component of the Dam1 complex, Ask1-YFP, along the mitotic spindle in metaphase arrested cells. NUF2^WT cells exhibited a characteristic bilobed distribution of Ask1 signals proximal to the spindle pole (marked by Spc110-mCherry), consistent with expected kinetochore localization. On the other hand, nuf2^{F8A P9A} mutant cells displayed a substantial change in Ask1 localization, showing a dramatic decrease proximal to the spindle pole, and a noticeable increase in localization along the mitotic spindle (Fig 5A). These observations align with the nuf2^{F8A P9A} mutant disrupting normal localization of the Dam1 complex to the kinetochore. If Dam1 complex localization is compromised in nuf2 mutants, we might anticipate synthetic lethality with hypomorphic alleles of Dam1 complex components. In line with this idea, we found that nuf2^F8A, nuf2^P9A, and nuf2^{I10A L11A} display synthetic lethality and/or sickness with dad1-1 mutants (Fig 5B and Fig S5). Collectively, these findings support the notion that multiple factors, including Mps1 and the Dam1 complex, interact with this “interaction hub” of Nuf2’s CH domain, and the observed phenotypes are likely a result of the combined effects of these binding alterations. This also raises an intriguing possibility that competition for binding to the Ndc80 complex might contribute to regulating the localization and function of these factors.

Figure 5. Dam1 localization to the kinetochore is disrupted in nuf2 mutants.

(A) 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). Left: Representative micrographs. 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. 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}, respectively). Significance was determined by a two-tailed unpaired t test (****; P < 0.0001).

(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 dad1-1 permissive (23°C) and semi-permissive (30°C) temperatures.

Ndc80 complex interaction hub mutants can be suppressed by modulating Mps1 levels

Considering that multiple factors seem to interact with this interaction hub, such as Mps1, the Dam1 complex, and potentially other factors, including the Aurora B kinase, Ipl1 (see Zahm et al., personal communication with S. Harrison), a challenging question arises: how do we determine the relative importance of each of these factors in promoting accurate chromosome segregation? Some insight into this question came from the following observations. We reasoned that if defects in Mps1-kinetochore association was the main cause of viability defects in our nuf2 mutants, perhaps modulating Mps1 levels at the kinetochore could suppress the associated defects. Prior research with human proteins demonstrated that the presence of the N-terminal tail of Ndc80 diminishes Mps1’s capacity to bind to the Ndc80 complex in vitro, likely due to competitive binding. Furthermore, this inhibition appears to be relieved by phosphorylation of the N-terminal tail⁴³. If this idea is correct, and Ndc80’s N-terminal tail forms a ‘closed’ conformation to compete with Mps1’s ability to bind in cells, we reasoned that deletion of this tail should increase kinetochore-associated Mps1 (Fig 6A) and may suppress the nuf2 mutant growth defects. We constructed a NUF2-AID NDC80-AID ‘double AID’ strain to conduct these experiments, and found that this genetic background was far more sensitized, i.e. nuf2 mutants which show no detectable growth defect in a NUF2-AID ‘single AID’ background, now result in dramatic cell viability defects (Fig S6A). Consistent with the idea above, many of the nuf2 mutants at the described Mps1 binding interface showed prominent growth defects in the context of NDC80^WT, however, many of these mutants were suppressed by ndc80^ΔN-tail (Fig 6B and Fig S6B–C). These findings also raise the possibility that the N-terminus of Ndc80 associates directly with the interaction hub to regulate Mps1 localization.

Figure 6. Increasing the level of Mps1 suppresses the lethality of nuf2 mutants.

(A) Model illustrating two states of the Ndc80 N-terminal tail: “closed”, in which the N-terminal tail of Ndc80 associates with Nuf2, precluding Mps1 binding, and “open”, in which the tail is altered, likely via phosphorylation, to allow Mps1-Nuf2 association⁴³.

(B) Yeast cell viability assay comparing nuf2 mutant phenotypes in NDC80^WT and ndc80^ΔN-tail strains. Each is a “double AID” strain, containing both NUF2-AID and NDC80-AID alleles, as well as ectopic copies of NUF2–3HA and NDC80–3HA (NUF2^WT NDC80^WT, M2998; NUF2^WT ndc80^ΔN-tail, M2999; nuf2^I10A NDC80^WT, M2654; nuf2^I10A ndc80^ΔN-tail, M2755; nuf2^S124A NDC80^WT, M2641; nuf2^S124A ndc80^ΔN-tail, M2742; nuf2^F132A NDC80^WT, M2841; nuf2^F132A ndc80^ΔN-tail, M2774). Cells were serially diluted five-fold and spotted onto plates containing DMSO or 250 μM auxin.

(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^{F8A P9A}, M2042 and M3587; nuf2^S124D, M2041 and M4619; nuf2^N128A, M2414 and M3588). Cells were serially diluted five-fold and spotted onto plates containing either glucose (control) or galactose (induce over-expression of MPS1), as well as DMSO or 250 μM auxin.

(D) Structure of S. cerevisiae Nuf2 (5TCS) illustrating the residues with mutants that are suppressed by expression of pGAL-MPS1. Darker shades of blue indicate increasing strength of the suppression.

(E) 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.

(F) Detection of Dam1-9Myc association with the kinetochore via immunoprecipitation. Exponentially growing wild-type (M457) and mps1-1 (M469) cultures expressing DSN1–6His-3Flag, DAM1–9Myc and NDC80–3HA were shifted to 37°C for 2 hours prior to harvest. Kinetochores were purified from lysates by anti-Flag immunoprecipitation (IP) and analyzed by immunoblotting.

Another way to modulate Mps1 levels at the kinetochore is through overexpression. Compellingly, overexpression of Mps1 rescues the cell viability defects of both nuf2^{F8A P9A} and nuf2^N128A, as well as another nearby mutation (Fig 6C & Fig S6D–E). In fact, when we examine the nuf2 mutant alleles whose cell viability defects are suppressed by Mps1 overexpression, we find that the mutated residues spatially cluster on a surface of Nuf2 (Fig 6D). However, viability of nuf2^S124D, the most penetrant single substitution nuf2 mutant we tested, was not suppressed by Mps1 overexpression (Fig 6C) even though it is close in spatial proximity to F8, P9 and N128 (Fig 1A & Fig 6D). We suggest the increased severity of nuf2^S124D may be due to physical displacement of the N-term loop of Nuf2 from its G-helix, leading to a complete disruption of the Mps1 binding site. The finding that the overexpression of Mps1 can suppress the detrimental effects of mutations in this portion of the Ndc80 complex implies that these mutations result in lethality due to reduced association between Mps1 and the Ndc80 complex. However, it was plausible that the rescue of these mutants by Mps1 overexpression was solely due to the restoration of spindle assembly checkpoint function^60–62. This does not appear to be the case, as Mps1 overexpression still suppresses nuf2 mutants even when the SAC is inactivated by deleting MAD3 (Fig S6F). Furthermore, this suppression appears specific to Mps1 as overexpression of Ipl1 does not rescue the cell viability of these nuf2 mutants (Fig S6G). Finally, the ability of Mps1 overexpression to suppress nuf2 mutant growth defects appeared to be dependent on the ability of Mps1 to interact with the Ndc80 complex. While the 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 Ndc80 complex interface fail to achieve the same outcome (Fig 6E and Fig S6H–I). These findings indicate that overexpression of Mps1 alone is insufficient to suppress the nuf2 mutants, but instead requires Mps1’s interaction with the Ndc80 complex. Moreover, the observation that certain nuf2 mutants (e.g. nuf2^{F8A P9A} and nuf2^N128A) are susceptible to suppression by Mps1 overexpression suggests that these mutants disrupt, but do not entirely eliminate, the ability of Mps1 to bind the Ndc80 complex.

While modulating the levels of Mps1 at the kinetochore appears to suppress the nuf2 CH domain mutants, how do we reconcile the observation that these mutants also display reduced Dam1 complex recruitment (as described in Fig 5A)? A potential explanation is that the localization of the Dam1 complex to the kinetochore is regulated by Mps1 activity, also suggested by previous work^31,36,63,64. To investigate this hypothesis, we used a kinetochore co-IP assay⁵⁶ in which kinetochores were immunoprecipitated, and co-purification of the Dam1 complex was determined by western blot. In line with the notion above, kinetochores purified from mps1-1 cells lacking Mps1⁵³ exhibited reduced association with the Dam1 complex (Fig 6F). These findings potentially elucidate how modulating Mps1 levels at the kinetochore can mitigate the effects caused by nuf2 CH domain mutants, which disrupt the interaction hub and presumably lead to the loss of binding of multiple factors. Additionally, these results suggest an active role of Mps1 in recruiting a competing factor, the Dam1 complex, vying for the same kinetochore binding site.

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

Finally, while modulating Mps1 levels appears to suppress the cell viability defects of the nuf2 CH domain mutants, we wanted to more directly ask if the phenotypes of nuf2 mutants are due specifically to diminished Mps1-kinetochore association. Here, we asked if restoring Mps1 localization, using rapamycin induced dimerization of FRB and FKBP12, would rescue the nuf2 mutant phenotypes (Fig 7A). We generated strains containing NUF2-AID, NDC80-FKBP12 and MPS1-FRB at their endogenous loci, and either NUF2^WT or nuf2^N128A expressed from an ectopic locus. Since tethering Mps1-FRB to the C-terminus of Ndc80 constitutively activates the SAC, causing cell inviability⁶⁵, these strains also included a deletion of MAD3. Compellingly, the tethering of Mps1 to Ndc80 effectively restored the viability of nuf2^N128A mutants (Fig 7B). These data argue that Mps1 needs to be bound to the Ndc80 complex to carry out its kinetochore biorientation function. In summary, our findings establish a pivotal “interaction hub” within the Ndc80 complex that orchestrates SAC signaling and microtubule attachment. Moreover, it underscores the indispensable role played by the kinetochore-associated pool of Mps1 in ensuring precise chromosome segregation.

Figure 7. Tethering Mps1 to the kinetochore 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 Ndc80 complex.

(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 plus 50 ng/mL rapamycin.

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 pinpointed 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 Ndc80 complex, but also associates with numerous other factors. Consequently, mutants disrupting this hub exhibit defects in spindle assembly checkpoint function and display severe errors in chromosome segregation. The array of interactions occurring at this specific site on the Ndc80 complex appears critical for the functional regulation of Mps1’s association with the kinetochore. In fact, restoring Mps1-Ndc80 complex association completely rescues the cellular defects associated with mutations at this site. Overall, this work sheds light on mechanisms cells use to regulate Mps1 at the kinetochore, and also underscores the essential role of Mps1 in kinetochore biorientation and precise chromosome segregation.

Mps1 is the critical factor that associates with the Ndc80 complex’s interaction hub

Our work describes a previously uncharacterized region of Nuf2’s CH domain that is responsible for interacting with segments of various proteins. This list encompasses 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 Ipl1^19,20,42,43 (this work as well as concurrent work by Zahm et al, Pleuger et al; personal communication S. Harrison and S. Westermann), among potentially other yet-to-be-identified factors. Notably, mutations affecting this interaction hub lead to cell inviability and substantial errors in kinetochore biorientation and chromosome segregation. Interestingly, our findings suggest that these severe phenotypes do not result from disrupting the binding of multiple factors to this site. Instead, our observation that specifically restoring Mps1-Ndc80 complex association, whether through overexpression or tethering via FRB-FKBP, implies that Mps1 might be the sole essential factor binding to this interaction hub. It remains possible that Mps1’s role may involve, in part, recruiting and/or regulating these additional factors when bound to the Ndc80 complex. Our observation that Mps1 appears to recruit the Dam1 complex aligns with this latter concept, and is consistent with previous work^31,36,63,64. In any case, our identification of the Mps1 binding site on the Ndc80 complex highlights the crucial role that kinetochore-bound Mps1 plays in error correction, in addition to its well-studied function in spindle assembly checkpoint activation. This error correction function of Mps1 is essential for cell viability, revealing a more significant role than we previously recognized.

The activity of Mps1 at the kinetochore is subject to regulation at multiple levels

We propose the following working model, building upon the excellent work of others (reviewed in^3,66), outlining the steps from the initial microtubule attachment to biorientation (Fig 8). Prior to initial attachment to spindle microtubules, the Ndc80 complex is in a ‘closed’ conformation with the N-terminal tail of Ndc80 folded back and occluding the association of other factors to this interaction hub⁴³. Both error correction kinases, Ipl1 and Mps1, can phosphorylate the N-terminal tail^{7,11–13,27–31}, which changes Ndc80 to an ‘open’ conformation⁴³, allowing binding of Mps1. Kinetochore-associated Mps1 can now carry out its well-described function in spindle assembly checkpoint activation^{33–35,37,44,45}. Importantly, this pool of Mps1 also appears essential for achieving kinetochore biorientation and ensuring accurate chromosome segregation. This is likely achieved through a tension-sensitive error correction mechanism^31–37. However, the precise working mechanism of this process remains unclear (further discussion below). In part, this error correction function of Mps1 may involve recruiting the Dam1 complex^31,36,63,64. Regardless, once attachments are correctly bioriented and no longer susceptible to error correction-mediated destabilization, the Dam1 complex establishes stable association with the Ndc80 complex, thereby displacing Mps1 from its binding site (Zahm et al., Pleuger et al.; personal communication S. Harrison and S. Westermann). In this manner, Mps1 seems to facilitate the recruitment of its own competitive inhibitor. Another mechanism through which Mps1 regulates its kinetochore localization is autophosphorylation^58,67. Furthermore, competition between Mps1 and microtubules for binding to the Ndc80 complex is proposed as an additional means to displace Mps1^42,43, although Mps1 binding does not appear to be mutually exclusive to microtubule binding⁶⁷. These cumulative events subsequently deactivate spindle assembly checkpoint 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 8.

Model of the transition from unattached kinetochore to bioriented attachment, incorporating known protein-protein associations and events that may occur at the interaction hub on Nuf2.

The interaction hub on Nuf2 (yellow) interacts with Mps1 (purple) and the Dam1 complex (dark grey), and potentially with the N-terminal tail of Ndc80 (light blue). The microtubule is shown in tan.

1. In the unattached kinetochore state, the N-terminus of Ndc80 folds over Nuf2 to occlude the interaction hub (“closed state”), precluding association of Mps1 with the kinetochore.
2. Movement of the Ndc80 N-terminal tail away from the Nuf2 interaction hub (possibly via phosphorylation by Ipl1 or Mps1) leads to the “open state”, allowing access to other factors.
3. Mps1 associates with the Ndc80 complex at the Nuf2 hub, triggering the SAC and participating in sister chromatid biorientation through an as yet undetermined mechanism that is likely tension sensitive.
4. Mps1 also promotes association of the Dam1 complex with Ndc80 complex, which ultimately allows for stable kinetochore-microtubule interaction and the formation of bioriented attachments.
5. Once the bioriented state is achieved, Mps1 is removed from the kinetochore via competition by the Dam1 complex for the interaction hub, autophosphorylation⁶⁷ and/or competition with microtubule binding^42,43. Loss of Mps1 at the kinetochore also turns off the SAC.

One key question is how does Mps1’s activity at kinetochores contribute to biorientation? Additionally, what mechanisms regulate these activities in response to tension, a critical signal that cells use to discern correct from incorrect attachments? It has been well-demonstrated that Mps1 phosphorylates sites on the N-terminal tail of Ndc80, leading to attachment destabilization. Notably, Mps1’s phosphorylation of Ndc80 is stimulated on kinetochores lacking tension³¹. However, disrupting the known Mps1 phosphorylation sites, even in conjunction with a spindle assembly checkpoint mutant, does not substantially impact cell viability^30,31. Furthermore, the N-terminal tail of the Ndc80 protein is dispensable for viability in yeast^30,68,69. These observations are in contrast to the nuf2 and mps1 mutants described here, which exhibit complete inviability. This discrepancy suggests that additional mechanisms must be involved in this process. Mps1 has been implicated in two other pathways that promote biorientation: the activation of the kinase Bub1^32,70 and the recruitment of pericentromeric localization of Sgo1^70–75. The interplay between these pathways, particularly in the context of our new mutants disrupting kinetochore-associated Mps1, requires further investigation. It’s worth noting that Bub1 and Sgo1 are non-essential proteins, and furthermore, the role of tension in these processes remains unclear.

Several key questions surround the regulation of the kinetochore-bound Mps1 pathway by tension. Given the proximity of Mps1 to the microtubule binding domains of the kinetochore, spatial separation of substrates from Mps1, as proposed for Aurora B⁷⁶, seems unlikely. Instead, tension-dependent changes in substrate conformation, as recently demonstrated for Ipl1⁷⁷, provide a plausible mechanism by which tension suppresses phosphorylation of critical targets. Alternatively, Mps1’s kinase activity or localization could be influenced by a kinetochore factor, potentially a portion of the Ndc80 complex, and this regulation may be responsive to tension. For instance, the displacement of Nuf2’s N-terminal loop from its G-helix (similar to the nuf2^S124D mutant) either via tension or another mechanism appears to be an intriguing way that cells could modulate Mps1 activity. This notion gains support from the observation that the assembly of the outer kinetochore is essential for generating most, if not all, cellular Mps1 activity⁵⁵. Additionally, the interplay between Aurora B and Mps1 activities remains an open question.

In summary, our work identifies a critical “interaction hub” that governs essential kinetochore functions of the kinase Mps1. These functions encompass its broadly studied role in spindle assembly checkpoint signaling, but, notably, emphasize the significance of Mps1 in establishing kinetochore biorientation. Our findings clarify the mechanisms that govern kinetochore-associated Mps1 and how these mechanisms contribute to the fidelity of chromosome segregation.

METHODS

Yeast strains and media

All yeast strains used are listed in Table S1 and are isogenic in the W303 background⁷⁸. Standard media and microbial techniques⁷⁹ were used, as well as standard yeast genetic methods for strain construction^80,81.

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 in⁸². A single integration LEU2 vector containing pNUF2-NUF2–3HA (pM604) was constructed by first tagging endogenous NUF2 with 3HA from pM9 (3HA:HIS3MX6)⁸³, then using the pNUF2-NUF2–3HA strain as a template for plasmid construction. Mutants of NUF2 were constructed by mutagenizing pM604 as described in^84,85. The NDC80–3V5-IAA7:KanMX is described in³⁹. A TRP1 single integration vector containing pNDC80-NDC80–3HA (pM270) was constructed in a similar manner to NUF2–3HA above, and pNDC80-ndc80^ΔN-tail-3HA (pM1342) was subsequently constructed using pM270 as a template for mutagenesis as in^84,85.

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 in⁸². 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 promoter⁸⁶. 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^84,85. 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-LacI is described in⁸⁷, CEN III::lacO:TRP1 is described in⁸⁸, and CEN IV::lacO:TRP1 was provided by Andrew Murray. Strains containing the previously described pMET-CDC20 allele were provided by Frank Uhlmann⁸⁹, 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 in⁹⁰, and DSN1–6His-3Flag in⁵⁶. TOR1–1, fpr1Δ, MPS1-FRB:KanMX, and NDC80-FKBP12:HISMX are described in^65,91. The Dam1-9Myc allele was provided by Sue Biggins.

Auxin inducible degron

The auxin inducible degron (AID) system was used as described in⁹². 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 plus 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, 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, 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.

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. 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 (Fig S3A–C). 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 assay in Fig 3B, only mononucleate cells in metaphase were counted. For chromosome segregation assays in Fig 3C and Fig 4D, only binucleate cells that had undergone anaphase (approximate spindle length ≥ 4 μM) were counted.

Spindle length in Fig S3A was calculated by drawing a line in FIJI (Image J, NIH) that connected the edges of the Spc110-mCherry signals from the two poles. Spindle pole body separation in Fig S4B was calculated by the percentage of cells that displayed two Spc110-mCherry signals at 90 minutes following the temperature shift. Ask1-YFP intensities in Fig 5A 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 Fig 5 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 in⁵⁶. 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 Fig S1C, cell lysates were prepared by mechanical disruption with glass beads in 2x SDS-PAGE sample buffer (0.4mM EDTA, 4% SDS, 125 mM Tris pH6.8, 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^93,94. 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 anti-mouse and goat anti-rabbit conjugated to horseradish peroxidase (HRP, GE Biosciences; 1:10,000 dilution). Antibodies were detected using the SuperSignal West Dura Chemiluminescent Substrate (Thermo Scientific).

Statistics

GraphPad Prism version 10 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 are reported.

Multiple sequence alignments

Fungal Nuf2 proteins were identified using a PSI-BLAST⁹⁵ search on NCBI. Multiple sequence alignments of the entire proteins were generated with ClustalOmega default parameters and displayed in JalView 2⁹⁶. Species are listed in the figures and/or legends. Conservation scores shown on structures were generated in ChimeraX 1.5, using an entropy-based measure from AL2CO⁹⁷.