A common molecular mechanism underlies the role of Mps1 in chromosome biorientation and the spindle assembly checkpoint

Abstract

The Mps1 kinase corrects improper kinetochore–microtubule attachments, thereby ensuring chromosome biorientation. Yet, its critical phosphorylation targets in this process remain largely elusive. Mps1 also controls the spindle assembly checkpoint (SAC), which halts chromosome segregation until biorientation is attained. Its role in SAC activation is antagonised by the PP1 phosphatase and involves phosphorylation of the kinetochore scaffold Knl1/Spc105, which in turn recruits the Bub1 kinase to promote assembly of SAC effector complexes. A crucial question is whether error correction and SAC activation are part of a single or separable pathways. Here, we isolate and characterise a new yeast mutant, mps1‐3, that is severely defective in chromosome biorientation and SAC signalling. Through an unbiased screen for extragenic suppressors, we found that mutations lowering PP1 levels at Spc105 or forced association of Bub1 with Spc105 reinstate both chromosome biorientation and SAC signalling in mps1‐3 cells. Our data argue that a common mechanism based on Knl1/Spc105 phosphorylation is critical for Mps1 function in error correction and SAC signalling, thus supporting the idea that a single sensory apparatus simultaneously elicits both pathways.

Through isolation and characterization of the new mps1‐3 mutant, this study proposes that Mps1‐dependent phosphorylation of Spc105 contributes to both proper chromosome segregation and the spindle assembly checkpoint in budding yeast.

Introduction

Accurate chromosome segregation is a fundamental aspect of mitosis and secures the genetic stability of a cell lineage. During the mitotic cell cycle, sister chromatids that have been generated by chromosome replication in S phase must attach through their kinetochores to microtubules emanating from opposite spindle poles in M phase. This occurs through an error correction mechanism where kinetochores dynamically attach and detach from microtubules until bipolar arrangement of sister chromatids is finalised (reviewed in ref. 1). Bipolar attachment, also referred to as amphitelic attachment, leads to tension across kinetochores, which arises from the balance of pulling forces exerted by microtubules and counteracting forces exerted by sister chromatid cohesion (reviewed in ref. 2). This tug of war leads also to chromosome congression to the metaphase plate.

Key to the error correction process is the evolutionary conserved CPC (chromosome passenger complex) that is composed by the Aurora B kinase and three additional centromere‐targeting and activating subunits (INCENP, survivin and borealin). The CPC detects lack of kinetochore tension (e.g. in case of monotelic attachment where only one sister kinetochore is bound to microtubules or syntelic attachment where both sister kinetochores are attached to the same spindle pole) and favours amphitelic attachments by destabilising incorrect kinetochore–microtubule connections through phosphorylation of specific kinetochore substrates (reviewed in ref. 3). The Mps1 kinase is also required for the error correction pathway in many organisms 4, 5, 6, 7, 8. In human cells Mps1 acts in concert with Aurora B in this process 7, 9, 10. Additionally, Mps1‐specific phosphorylation targets, like the kinetochore Ska complex, have been recently involved in the error correction mechanism 11. Importantly, in budding yeast Mps1 seems to operate independently of Aurora B in chromosome biorientation and its critical targets are unknown 6, 8, 12. Although the budding yeast functional orthologue of the SKA complex, the Dam1 complex, is phosphorylated by Mps1, mutations altering these phosphorylations have no impact on chromosome segregation 12, 13.

The error correction process is intimately coupled to a conserved surveillance device, the spindle assembly checkpoint or SAC, that prevents sister chromatid separation and the onset of anaphase until all chromosomes are bioriented on spindle microtubules. SAC signalling fires at unattached kinetochores, where SAC proteins are recruited, and is propagated to the whole cell for inhibition of the E3 ubiquitin ligase anaphase‐promoting complex (APC) bound to its activator Cdc20. A soluble protein complex called MCC (mitotic checkpoint complex) and composed of the SAC components Bub3, BubR1, Mad2 and Cdc20 is essential for APC/Cdc20 inhibition. This, in turn, prevents degradation of APC substrates, such as securin and cyclin B, thereby halting the cell cycle before the onset of anaphase (reviewed in ref. 14).

The Mps1 kinase acts at the apex of the SAC signalling pathway, by phosphorylating the Knl1/Spc105 kinetochore scaffold on its MELT repeats, which in turn recruit the Bub3‐Bub1 SAC complex 15, 16, 17, 18. Subsequent binding of the SAC factor Mad1 to Bub1, together with Mps1‐mediated Mad1 phosphorylation, is a prerequisite for recruitment of Mad2 to kinetochores and MCC assembly 19, 20, 21, 22, 23. Consistently, phosphorylation of Knl1/Spc105 MELT repeats by Mps1 is essential for SAC signalling 15, 17, 18. The Knl1/Spc105 kinetochore protein has emerged as a key factor not only as a platform for SAC signalling, but also to silence the SAC once all pairs of sister chromatids are bipolarly attached. Indeed, Knl1/Spc105 harbours at its N‐terminus an “RVSF” amino acid motif that binds to PP1, which extinguishes SAC signalling at least partly by dephosphorylating Knl1/Spc105 MELT repeats 24, 25, 26, 27, 28, 29. Additionally, PP1 dephosphorylates the catalytic loop of Mps1, thus further silencing SAC 30.

Aurora B has also been implicated in SAC signalling. While an early model proposed that its role in the SAC would be indirect and linked to the generation of unattached kinetochores during the correction of tension‐less kinetochore–microtubule attachments 31, 32, 33, 34, increasing evidence supports a direct role of Aurora B in SAC activation from unattached kinetochores 9, 35, 36, 37, 38, 39. For instance, Aurora B promotes kinetochore recruitment of several SAC factors, including Mps1, and weakens PP1 binding to Knl1 through phosphorylation of Knl1 RVSF motif, suggesting that it acts upstream of SAC signalling 7, 9, 25, 31, 40.

The involvement of Mps1 and Aurora B in both error correction and SAC signalling raises a crucial, yet unsolved, question: are these two pathways concomitantly elicited by a common upstream sensory device or, alternatively, they respond to distinct defects (i.e. lack of kinetochore tension and lack of kinetochore attachment)? Is it possible to activate SAC without engaging the error correction machinery and vice versa? If the molecular machineries detecting these defects were separable, it should be possible to generate hypo‐ or hyper‐morphic mutants in the two apical kinases that selectively affect either pathway without perturbing the other.

Here, we report the characterisation of a novel temperature‐sensitive mps1 mutant of budding yeast (named mps1‐3) that is defective in chromosome biorientation and SAC signalling. An unbiased genetic screen for extragenic suppressors of the lethality of mps1‐3 cells at high temperature revealed that mutations affecting the Spc105‐PP1 interface restore proper chromosome biorientation as well as a proficient SAC response. Furthermore, artificial anchoring of Bub1 near the MELT repeats of Spc105 in mps1‐3 cells results in a similar rescue. Altogether, our data indicate that Mps1 promotes both error correction and SAC signalling through a common molecular mechanism based on Spc105 phosphorylation at the MELT repeats and recruitment of the Bub1 kinase to kinetochores, thereby strengthening the idea that a single sensory apparatus engages both processes.

Results

The novel mps1‐3 mutant is defective in chromosome segregation and proficient in SPB duplication

The targets of budding yeast Mps1 in the error correction pathway are unknown so far. Dissecting the function of Mps1 in kinetochore error correction is hampered by its involvement in the duplication of spindle pole bodies (SPBs) and the assembly of a bipolar spindle 41, the first step towards chromosome biorientation.

We isolated the mps1‐3 mutant through a random mutagenesis screen for temperature‐sensitive mps1 mutants (see Materials and Methods). Haploid mps1‐3 cells grew slowly at 25–30°C and were unable to form colonies at temperatures above 32°C (Fig 1A). Sequencing of the mps1‐3 allele revealed a single mutation replacing the conserved serine 635 in the kinase domain by phenylalanine (Fig 1B). To characterise the primary defects underlying the temperature sensitivity of mps1‐3 cells, we synchronised in G1 mps1‐3 cells, together with a wild‐type control and the well‐characterised mps1‐1 mutant 41, and released cells at the restrictive temperature of 34°C. FACS analysis of DNA contents revealed the expected bimodal distribution of DNA contents in wild‐type cells, while mps1‐3 and mps1‐1 cells showed abroad distribution of DNA contents after cytokinesis (90–105 min), indicative of massive chromosome missegregation (Fig 1C). However, while mps1‐1 cells were unable to assemble bipolar spindles due to their failure to duplicate SPBs and to activate the SAC 41, mps1‐3 cells proficiently formed bipolar spindles and elongated them during anaphase (Fig 1D and Appendix Fig S1A), suggesting that mps1‐3 might be a separation of function mutant specifically defective in chromosome biorientation but proficient at SPB duplication.

Figure 1. The mps1‐3 mutant is defective in chromosome biorientation.

• A
  Serial dilutions of stationary phase cultures of the indicated strains were spotted on YEPD and incubated at the indicated temperatures.
• B
  Mps1 sequence alignment around Saccharomyces cerevisiae Ser635. Sc: Saccharomyces cerevisiae; Sp: Schizosaccharomyces pombe; Mm: Mus musculus: Hs: Homo sapiens; Dm: Drosophila melanogaster.
• C, D
  Wild‐type, mps1‐3 and mps1‐1 mutant cells were synchronised in G1 with α‐factor at permissive temperature (25°C) and then released at restrictive temperature (34°C, t = 0). Cells were collected at the indicated time points for FACS analysis of DNA contents (C) and for immunofluorescence using anti‐tubulin antibodies in order to score metaphase and anaphase spindles (D). Budding and nuclear division were scored on the FACS samples.
• E
  Frequencies of chromosome loss were quantified at permissive temperature (25°C).
• F
  Wild‐type and mps1 mutant cells bearing the TetO/TetR‐GFP system to mark the centromere of chromosome V 43 and expressing the SPBs marker Spc97‐mCherry were synchronised as in (C, D) and arrested in anaphase through the temperature‐sensitive cdc15‐2 allele. At 120 and 150 min, cells were fixed for scoring chromosome V segregation (n ≥ 172). Arrowheads indicate the sister chromatids of chromosome V. Error bars: SD. N = 3. Representative images of cells for each genotype are shown on the right. Scale bar: 5 μm.
• G
  Wild‐type and mps1‐3 mutant cells carrying the TetO/TetR‐GFP markers for CEN5 labelling and expressing mCherry‐Tub1 were grown at 25°C and then shifted to 34°C for 1 h before filming. Cells were filmed at 34°C every 2 min by time‐lapse fluorescence microscopy. DIC: differential interference contrast. Scale bar: 5 μm.

To quantify the mps1‐3 chromosome segregation defects, we used a genetic assay based on the ability of diploid cells to mate with a haploid tester strain upon events leading to loss of one of the two copies of the MAT locus, including loss of one entire chromosome III 42. We used wild‐type and homozygous mps1‐1/mps1‐1, mps1‐3/mps1‐3 diploid cells that were grown for a defined number of generations at 25°C. Already under these permissive conditions, the appearance of mating‐competent cells was 1,000‐ to 2,000‐fold higher in the mps1 mutants relative to the wild type (Fig 1E), suggesting pronounced chromosome III loss even in conditions that allow cell proliferation.

To follow directly chromosome biorientation in mps1‐3 haploid cells, we visualised by fluorescence microscopy the segregation of chromosome V tagged with the TetO/TetR system 1.4 kb from the centromere (CEN5‐GFP) 43. Cells were also expressing the SPB marker Spc97‐mCherry and carried a cdc15‐2 temperature‐sensitive allele to arrest cells in telophase. G1‐synchronised cells were released at the restrictive temperature of 34°C, and samples were collected at 120 and 150 min from the release. As expected, near the totality of wild‐type cells bioriented and properly segregated the two sister chromatids of chromosome V at opposite SPBs, while almost all mps1‐1 cells underwent CEN5 mono‐orientation on the single unduplicated SPB (Fig 1F). Strikingly, the vast majority of mps1‐3 mutant cells displayed duplicated SPBs that separated far apart, but at 120′ the two chromatids had segregated correctly to opposite spindle poles in 42% of the cells, whereas in 41% of the cells the two chromatids had remained at the same spindle pole. Only 4% of the cells did not duplicate the SPBs, while 12% of the cells had duplicated SPBs that remained very close to each other in the mother cell, accompanied by lack of CEN5 segregation (Fig 1F). We conclude that the mps1‐3 mutant, though capable to duplicate the SPBs and form bipolar spindles, is specifically defective in chromosome biorientation at 34°C. It is worth noting, however, that at higher temperatures (37°C) mps1‐3 cells also failed SPB duplication and bipolar spindle assembly, suggesting that mps1‐3 is a hypomorphic mutant.

Centromeric and pericentromeric cohesin is crucial for chromosome biorientation, but while centromeres transiently split in metaphase upon bipolar attachment, pericentromeric cohesion resists to pulling spindle forces 43. We therefore asked if premature loss of pericentromeric cohesion could underlie the chromosome biorientation defect of mps1‐3 cells. To answer this question, we tagged with the TetO/TetR system the pericentromere of chromosome V (TetO array at −13 kb from the centromere ref. 43) and analysed its behaviour in metaphase‐arrested cells by fluorescence microscopy. After a pre‐synchronisation in G1, cells carrying the temperature‐sensitive cdc16‐1 allele (compromising APC/C activity at high temperatures and arresting cells in metaphase 44) were released at 34°C. Under these conditions, sister pericentromeric regions of chromosome V did not disjoin in most mps1‐3 cells, similar to wild‐type and mps1‐1 cells (Appendix Fig S1B). Thus, the chromosome biorientation defect of mps1‐3 cells is not explained by a premature loss of pericentric cohesion.

Biorientation defects could be due to lack of kinetochore–microtubule attachments, or to a failure in the correction of faulty kinetochore–microtubule connections that are not under tension. To discriminate between these two processes, we imaged live cells with GFP‐tagged CEN5 and expressing Tub1‐mCherry to visualise microtubules. Cells were pre‐incubated at 34°C for 1 h before filming at the same temperature. Under these conditions, while 100% of wild‐type cells correctly bioriented the two sister CEN5 and segregated them to opposite spindle poles (n = 201, Figs 1G and 3H), only 51.3% of mps1‐3 mutant cells bioriented CEN5, whereas 48.7% of the cells did not (n = 179, Figs 1G and 3H). During the whole movie, we did not detect lagging chromosomes or CEN5 signals off the spindle in mps1‐3 cells, suggesting that kinetochore–microtubule attachment is not affected. Furthermore, in contrast to ipl1 mutant cells that missegregate chromosomes mostly towards the old SPB 45, we found no such bias in mps1‐3 cells under the above conditions (Fig 3H). Interestingly, however, when we filmed mps1‐3 after a G1 arrest and release at 34°C (i.e. during the first cell cycle at restrictive temperature, when the pre‐existing SPB is likely fully functional), we found a strong missegregation bias towards the bud (Appendix Fig S1C). This bias tends to be lost in the subsequent cell cycles (Appendix Fig S1C). From these data, we conclude that mps1‐3 cells are defective in the correction of monopolar attachments, like ipl1 mutants and in agreement with previous conclusions 6. Additionally, they could alter the normal pattern of SPB asymmetric inheritance (see Discussion).

The mps1‐3 mutant does not engage the spindle assembly checkpoint

To assess if mps1‐3 cells can engage the SAC, pre‐synchronised G1 cells were released at 34°C in presence of the microtubule‐depolymerising drug nocodazole. FACS analysis of DNA contents showed that wild‐type cells arrested as expected in metaphase with 2C DNA contents, indicative of proficient SAC signalling. Contrariwise, mps1‐3 and mps1‐1 cells kept progressing through the cell cycle accumulating 4C DNA contents, indicating that they cannot engage a SAC response (Fig 2A). This was confirmed by Western blot analysis of two APC substrates, the securin Pds1 and the cyclin B Clb2, which are normally degraded in anaphase and stabilised by the SAC. As expected, both proteins remained stable in wild‐type cells treated with nocodazole, while they underwent proteolysis in mps1‐1 and mps1‐3 mutants (Fig 2B). Thus, the mps1‐3 mutant is severely defective in SAC activation at high temperature.

Figure 2. The mps1‐3 mutant is defective in SAC signalling and does not localise Mps1 at kinetochores.

• A
  Wild‐type and mps1 mutant cells were synchronised in G1 with α‐factor at 25°C and then released at 34°C in the presence of nocodazole (t = 0). Cells were collected at the indicated time points for FACS analysis of DNA contents.
• B
  Cells were treated as in (A), and after 2 h from the release, α‐factor was re‐added to prevent cells from undergoing a second cell cycle. Cell samples were collected at the indicated time points for Western blot analysis of the indicated proteins. Pgk1 was used as loading control. Cyc: cycling cells.
• C
  Top: in vitro kinase assays with recombinant Mps1 and Mps1‐3 purified from Escherichia coli and incubated at 34°C for the indicated times in the presence of radioactive ATP and a recombinant Spc105 N‐terminal fragment (aa 1–320) as substrate. Bottom left: kinase activity was quantified on the autoradiographs by ImageJ and normalised to the levels of the full‐length protein on the Coomassie blue‐stained gel (right). Bottom right: 1 μg of the recombinant Mps1 kinases used for kinase assays was loaded on SDS–PAGE and stained with Coomassie blue for normalisation of kinase assays.
• D
  Wild‐type and mps1‐3 cells were synchronised in G1 with α‐factor at 25°C and then released at 34°C in the presence of nocodazole. Cells were collected after 90 min and fixed with formaldehyde for ChIP‐seq analysis. ChIP sequence reads were normalised against sequence reads from corresponding input samples, and relative enrichment is plotted for chromosome III around the centromere (see the centromeric regions of all 16 yeast chromosomes in Appendix Fig S2). Y‐axis shows enrichment values (linear scale, range is 0–10). Values below 1.5 are shown in grey, and values above 1.5 (i.e. sequences enriched in ChIP samples) are red coloured.
• E
  Cells with the indicated genotypes were synchronised in G1 with α‐factor at 25°C and then released in fresh medium at 34°C (t = 0). Cells were collected at the indicated time points for FACS analysis of DNA contents.

The Mps1‐3 mutant protein has enhanced kinase activity but has reduced levels at kinetochores

The catalytic activity of Mps1 is essential for both chromosome biorientation and SAC activation 5, 6, 8, 15. In order to understand if the defects of the mps1‐3 mutant were due to impaired Mps1 kinase activity, we performed in vitro kinase assays using recombinant Mps1 and Mps1‐3 proteins purified from bacteria. We measured the ability of these proteins to auto‐phosphorylate and to phosphorylate the N‐terminal part of Spc105 (aa 1–320) at different times of incubation in the presence of P³³γATP at 34°C. Surprisingly, the Mps1‐3 catalytic activity turned out to be more elevated than that of Mps1 at all time points (Fig 2C). This was confirmed by similar kinase assays in the presence of the exogenous substrate MBP and Mps1/Mps1‐3 proteins affinity‐purified from yeast cells using galactose‐inducible glutathione S‐transferase (GST) fusion constructs 46 (Appendix Fig S2A). Contextually, Mps1‐3 protein levels in yeast cells were also 2‐ to 3‐fold higher than wild‐type Mps1 and had aberrant electrophoretic mobility (Appendix Fig S2A). The elevated Mps1‐3 levels correlated with a decreased proteolysis of the protein in mitosis (Appendix Fig S2B), consistent with the idea that Mps1 turnover inversely correlates to its kinase activity 47.

Since Mps1 levels at kinetochores increase when its kinase activity is inactivated 7, 40, 48, 49, 50, it is conceivable that the elevated Mps1 kinase activity in mps1‐3 cells accelerates Mps1 kinetochore turnover to a threshold that could be incompatible with SAC signalling and chromosome biorientation. To test this possibility, we tagged with three HA epitopes the MPS1 and mps1‐3 genes at their genomic locus and checked the distribution of the corresponding proteins on chromosomes by ChIP‐seq using cells pre‐synchronised in G1 and released in nocodazole for 90 min at 34°C, time at which both MPS1 and mps1‐3 cells were still in mitosis. Strikingly, while wild‐type Mps1 clearly accumulated at the centromeric region of all 16 yeast chromosomes, Mps1‐3 did not (Fig 2D and Appendix Fig S3), suggesting that its residence at kinetochores is severely impaired.

To definitely ascertain if reduced kinetochore levels are the main reason for the Mps1‐3 loss of function, we artificially recruited to kinetochores GFP‐tagged Mps1 and Mps1‐3 in cells that co‐expressed the kinetochore protein Mtw1 tagged with the high affinity GFP‐nanotrap (GFP‐binding domain or GBD) 51. This strategy led to constitutive association of Mps1 to the kinetochore in 91.3% of the cells and of Mps1‐3 in 88.5% of the cells (n = 196 and n = 209, respectively). Cells also carried the deletion of MAD2 to obliterate SAC, which would be otherwise constitutively active following Mps1 tethering to kinetochores 49, 52. Notably, anchoring Mps1‐3 to Mtw1 could partially restore proper chromosome segregation at 34°C, as shown by the bimodal distribution of DNA contents by FACS analysis (Fig 2E), suggesting that the chromosome biorientation defects of mps1‐3 cells are largely accounted for by decreased Mps1 kinetochore levels.

A genetic screen for spontaneous suppressors of the mps1‐3 mutation

Since SAC is not essential in budding yeast 53, 54, we reasoned that chromosome missegregation could be the main cause of the lethality of mps1‐3 cells at high temperatures, which makes this mutant a unique genetic tool to study the role of Mps1 in chromosome biorientation. To gain insights into this process, we performed an unbiased genetic screen for spontaneous mutations suppressing the temperature sensitivity of mps1‐3 cells at 34°C. Classical genetic analyses allowed us to establish if the suppressing mutations were recessive or dominant, distinguish the intragenic from the extragenic suppressors (i.e. suppressing mutation inside or outside the mps1‐3 gene, respectively) and assess if the extragenic suppressors were allelic to one another. Through this screen, we could isolate 27 suppressors. Out of these, 17 turned out to be intragenic while 10 were extragenic suppressors. We focused on the extragenic suppressors, which could provide valuable insights into the molecular mechanism underlying the role of Mps1 in chromosome biorientation. Through genetic analyses and whole genome sequencing, we could establish that the suppressing mutations hit three different genes (Fig 3A and B): SPC105, encoding for the kinetochore protein Spc105/Knl1; GLC7, encoding for the catalytic subunit of the phosphatase PP1; and GRR1, encoding for an F‐box protein of the E3 ubiquitin ligase complex SCF. Three GRR1 suppressors out of four carried also a mutation in RSC30, encoding for a subunit of the RSC chromatin‐remodelling complex 55. The basis for the suppression by the GRR1 mutations, alone or in combination with the mutation in RSC30, was not immediately clear and will be investigated in the future.

Figure 3. Characterisation of the extragenic suppressors of mps1‐3 mutant cells.

• A
  Serial dilutions of stationary phase cultures of the indicated strains were spotted on YEPD and incubated at 25°C and 34°C.
• B
  List of extragenic suppressors found in the unbiased genetic screen.
• C
  Serial dilutions of stationary phase cultures of the indicated strains were spotted on YEPD and incubated at the indicated temperatures.
• D
  Cells with the indicated genotypes were synchronised in G1 with α‐factor at 25°C and then released in fresh medium at 34°C in presence of nocodazole (t = 0). Cells were collected at the indicated time points for FACS analysis of DNA contents.
• E
  Cells were treated as in (D) and collected at the indicated time points for Western blot analysis of Mad1‐3HA. Cyc: cycling cells.
• F
  Cells with the indicated genotypes were synchronised at 25°C and then released in fresh medium at 34°C (t = 0). Cells were collected at the indicated time points for FACS analysis of DNA contents.
• G, H
  Cells with the indicated genotypes carrying the TetO/TetR‐GFP markers for CEN5 labelling and expressing mCherry‐Tub1 were grown at 25°C and then shifted to 34°C for 1 h before filming. Cells were filmed at 34°C every 2 or 4 min by time‐lapse fluorescence microscopy. Chromosome V segregation errors are reported in the table (H). Representative cells are shown as examples in the montages (G). Representative montages for wild‐type and mps1‐3 cells are shown in Fig 1F. DIC: differential interference contrast. Scale bar: 5 μm.
• I
  Wild‐type, mps1‐3, mps1‐3 spc105‐18 and mps1‐3 GLC7‐24 cells were synchronised in G1 with α‐factor at 25°C and then released at 32°C in the presence of nocodazole (note that the presence of 3HA tags at the C‐terminus of Mps1‐3 slightly decreases the maximal temperature of suppression). Cells were collected after 90 min and fixed with formaldehyde for ChIP‐seq analysis. ChIP sequence reads were normalised against sequence reads from corresponding input samples, and relative enrichment is plotted for chromosome III around the centromere (see the centromeric regions of all 16 yeast chromosomes in Appendix Fig S5). Y‐axis shows enrichment values (linear scale, range is 0–10). Values below 1.5 are shown in grey, and values above 1.5 (i.e. sequences enriched in ChIP samples) are red coloured.

Spc105 (Knl1 in metazoans) is a known target of Mps1 in the SAC 15. Mps1‐dependent phosphorylation of its MELT repeats is crucial for SAC signalling by recruiting Bub1 and Bub3 15, 16, 17, 18. The PP1 Glc7 in turn dephosphorylates Spc105 MELT repeats by binding directly the consensus RV/IXF motif at the N‐terminus of Spc105 (RVSF, aa 75–78) to release Bub1 and silence the checkpoint 15, 28. Strikingly, all our suppressing mutations in SPC105 encompassed or were close to the PP1‐binding motif, while both GLC7 suppressors carried a missense mutation of Phe256 (to Ile or Val), which resides in the hydrophobic groove involved in PP1 interaction with partners carrying the RV/IXF consensus motif 56 (Fig 3B). Collectively, these observations suggest that reduced PP1 interaction with Spc105 likely accounts for the rescue of the temperature sensitivity of mps1‐3 cells. Interestingly, the spc105‐V76G mutation was recessive and lethal in otherwise wild‐type cells, consistent with the notion that mutations in the PP1‐binding site of Spc105 are inviable due to constitutive SAC activation 28, while spc105‐A79T and spc105‐V82D were viable, suggesting that they might interfere with PP1 recruitment to kinetochores less severely than spc105‐V76G. It is also worth noting that spc105‐V76G suppressed more robustly than spc105‐A79T and spc105‐V82D the lethality of mps1‐3 cells at high temperatures (Fig 3A).

Consistent with the idea that lack of PP1 recruitment to Spc105 rescues the lethality of mps1‐3 cells, the well‐characterised RASA mutation in the PP1 binding motif of Spc105 28 also suppressed the temperature sensitivity of mps1‐3 cells (Fig 3C; note that the spc105‐RASA mutation is lethal unless SAC activity is abolished, e.g. by MAD2 deletion 28). Additionally, fusing the PP1 catalytic subunit Glc7 to the N‐terminus of the Spc105‐RASA mutant protein 28 almost completely abolished its suppressing properties (Fig 3C), indicating that suppression stems only from loss of PP1 binding to Spc105.

The GLC7 mutations isolated in our screen were dominant and viable in otherwise wild‐type cells. Since dominant mutations are often associated with gain of function, this observation was quite puzzling and prompted us to characterise PP1 activity in GLC7‐24 cells. As a readout, we analysed the phosphorylation status of the well‐characterised PP1 cytoplasmic substrate Snf1, which is hyperphosphorylated on T210 under low glucose conditions (1%) and gets dephosphorylated by Glc7 in high glucose (4%) 57. No significant differences in the phosphorylation of Snf1 T210 were apparent in mps1‐3 GLC7‐24 and GLC7‐24 cells relative to the wild type (Appendix Fig S4A), suggesting that the cytoplasmic activity of PP1 is not affected. To assess PP1 phosphatase activity at kinetochores in GLC7‐24 cells, we monitored phosphorylation of Sli15, a subunit of the Aurora B complex that resides at kinetochores and on spindle microtubules and is involved in chromosome biorientation 45, 58. Sli15 is phosphorylated by Ipl1/Aurora B 59, whose activity is often reversed by PP1 60, 61. Interestingly, Sli15 phosphorylation was more pronounced in cycling cells and slightly advanced during the cell cycle in the GLC7‐24 mutant relative to the wild type (Appendix Fig S4B). Thus, Glc7 phosphatase activity is not globally elevated in GLC7‐24 cells; rather, it might impair dephosphorylation of specific kinetochore substrates.

Suppressing mutations in SPC105 and GLC7 restore SAC activation and normal chromosome segregation in mps1‐3 cells

To investigate further the phenotypes of the suppressors, we analysed their ability to engage the SAC and undergo normal chromosome segregation. Pre‐synchronised G1 cells were released at 34°C in the presence and absence of nocodazole and subjected to FACS analysis of DNA contents at different time points. Importantly, the spc105‐18 and GLC7‐24 mutations restored proficient SAC activation in mps1‐3 cells (Fig 3D). In mps1‐3 GLC7‐24 cells, this was accompanied by Mad1 hyperphosphorylation, which requires Mps1 activity 62, 63 and was abolished in mps1‐3 cells (Fig 3E; note that we could not study Mad1 phosphorylation pattern in mps1‐3 spc105‐18 cells due to their synthetic lethality with HA‐tagged Mad1). Since Mad1 is recruited to unattached kinetochores and is not part of soluble MCC 64, 65, these data further strengthen the idea that in GLC7‐24 cells PP1 has impaired ability to dephosphorylate kinetochore substrates. Additionally, they raise the possibility that PP1 could dephosphorylate Mad1 directly.

Remarkably, the spc105‐18 and GLC7‐24 suppressors also rescued the chromosome segregation defects of mps1‐3, as shown by the bimodal distribution of DNA contents after one complete cell cycle at 34°C (Fig 3F). Moreover, live cell imaging of mps1‐3 spc105‐18 and mps1‐3 GLC7‐24 cells with GFP‐tagged CEN5 and expressing Tub1‐mCherry showed nearly complete suppression of chromosome missegregation (Fig 3G and H). Importantly, the spc105‐18 and GLC7‐24 alleles did not restore normal levels of Mps1‐3 at centromeres, as shown by ChIP‐seq (Fig 3I and Appendix Fig S5). Furthermore, suppression of chromosome missegregation was not a consequence of restored SAC activity, which might impose sufficient time for error correction, because it remained unchanged upon MAD2 deletion (Fig 3H).

Thus, these data strongly argue that a common mechanism involving Mps1 and antagonised by PP1 simultaneously promotes error correction of faulty kinetochore attachments and SAC signalling.

The spc105‐18 and GLC7‐24 suppressors restore Spc105 phosphorylation in mps1‐3 cells

To further explore the possibility that Mps1 could promote SAC signalling and chromosome biorientation through the same molecular mechanism, we analysed the phosphorylation status of Spc105 in wild‐type and mps1‐3 cells at 34°C in presence or absence of nocodazole. As expected, Spc105 tagged with 3HA (Spc105‐3HA) was promptly phosphorylated in wild‐type cells treated with nocodazole, as indicated by its mobility shift in SDS–PAGE (times 60–150 min after release from G1, Fig 4A). Phosphorylation was impaired in mps1‐3 cells under the same conditions, in agreement with the notion that Spc105 is a critical Mps1 substrate during SAC activation 15. Importantly, Spc105 phosphorylation was efficiently restored, if not accentuated, in the mps1‐3 spc105‐18 and mps1‐3 GLC7‐24 mutants (Fig 4A), consistent with defective PP1 activity at kinetochores that counteracts Mps1‐dependent phosphorylation of Spc105 during SAC signalling. Interestingly, we could detect a slight electrophoretic mobility shift of Spc105 also in wild‐type cells undergoing a normal cell cycle in the absence of nocodazole (Fig 4B, times 30′–40′ after release from G1, which corresponds to cells in S phase), suggesting that Mps1 might phosphorylate a low number of MELT repeats in unperturbed conditions. Consistent with this hypothesis, an unphosphorylatable MELT mutant (spc105‐6A) is unable to recruit Bub1 to kinetochores also during unperturbed conditions and displays chromosome segregation defects 15. Spc105 phosphorylation was not apparent in mps1‐3 cells at 34°C and was restored in mps1‐3 spc105‐18 cells (Fig 4B). It is worth noting that the mobility shift of Spc105 was even more conspicuous and present for a larger cell cycle window in mps1‐3 spc105‐18 cells than in the wild type (Fig 4B), in line with the impaired PP1 activity at Spc105 in this mutant. Moreover, it was prominent in mps1‐3 spc105‐18 cells at 25°C, but less so at 34°C (Appendix Fig S6A), confirming that it depends on Mps1 activity at kinetochores.

Figure 4. Spc105 phosphorylation and Bub1 kinetochore recruitment are impaired in mps1‐3 cells.

• A, B
  Wild‐type and mps1‐3 cells were synchronised in G1 with α‐factor at 25°C and then released at 34°C in the presence (A) or absence (B) of nocodazole (t = 0). Cells were collected at the indicated time points for Western blot analysis of the indicated proteins. Equal amounts of protein extracts were loaded on two different gels, for Western blot of Spc105‐3PK and Clb2/Pgk1, respectively. Clb2 was used as mitotic marker and Pgk1 as loading control. Cyc: cycling cells.
• C, D
  Cells with the indicated genotypes expressing Bub1‐GFP and the kinetochore marker Mtw1‐Tomato were grown in SD glu 2% and then shifted to 34°C for 1 h before filming; they were then filmed every 4 min by time‐lapse fluorescence microscopy at 34°C. Note that BUB1‐GFP is synthetic lethal with mps1‐3. Thus, we used mps1‐3 GALs‐MPS1 BUB1‐GFP cells that were grown in ‐His RG medium at 30°C; glucose was added to the culture for 30 min to shut off GALs‐MPS1, followed by shifting cells to SD medium at 34°C for 1 h before imaging in the same medium. Montages show representative cells (C). Arrowheads indicate Bub1‐GFP signals at kinetochores. DIC: differential interference contrast. Scale bar: 5 μm. GFP signals (Bub1) co‐localising with Tomato signals (kinetochores) were scored on the movies (D).
• E–H
  Wild‐type and mps1‐3 cells were synchronised in G1 with α‐factor at 25°C and then released at 34°C in the presence (E, F) or absence (G, H) of nocodazole (t = 0). Cells were collected at the indicated time points for Western blot analysis of the indicated proteins (E, G) and for FACS analysis of DNA contents (F, H). Equal amounts of protein extracts were loaded on two different gels, for Western blot of Bub1‐3HA/Cdc5 and Clb2/Pgk1, respectively. A white asterisk indicates a hyperphosphorylated isoform of Bub1 that in wild‐type cells correlates with lack of chromosome biorientation. Cdc5 and Clb2 were used as mitotic markers and Pgk1 as loading control. Cyc: cycling cells.

Bub1 recruitment to kinetochores is impaired in mps1‐3 cells but is restored by the spc105‐18 and GLC7‐24 suppressing mutations

As mentioned above, Spc105/Knl1 phosphorylation primes Bub1 recruitment to kinetochores for SAC signalling 15, 16, 17, 18. Since Bub1 transiently associates with kinetochores also under unperturbed conditions 65, we visualised by time‐lapse video microscopy Bub1‐GFP along with the kinetochore marker Mtw1‐Tomato in wild‐type and mps1‐3 cells at 34°C. Under these conditions, Bub1 was transiently recruited to kinetochores in wild‐type cells with a small and medium bud, i.e. during S phase and early mitosis, in 100% of the cells (n = 260, Fig 4C and D). In mps1‐3 cells, we detected low levels of Bub1 at kinetochores only in 20% of the cells at 34°C (n = 170, Fig 4C and D), while the protein co‐localised normally with Mtw1 in small‐ and medium‐budded cells at 25°C (Appendix Fig S6B). Therefore, lack of Spc105 phosphorylation in mps1‐3 cells under unperturbed conditions correlates with loss of Bub1 at kinetochores. Remarkably, Bub1 kinetochore localisation was restored to nearly wild‐type levels in mps1‐3 spc105‐18 and mps1‐3 GLC7‐24 cells (Fig 4C and D), further strengthening the correlation between Spc105 phosphorylation and Bub1 residence at kinetochores. It is worth noting that while in wild‐type cells Bub1 dissociated from the kinetochores at the time of chromosome biorientation, it persisted at kinetochores up to anaphase in the majority of mps1‐3 spc105‐18, in good agreement with Spc105 hyperphosphorylation in these cells.

In order to be fully functional for SAC activation, Bub1 also needs to be directly phosphorylated by Mps1, which in turn allows its binding to Mad1 19, 20, 21. Thus, we checked if Bub1 phosphorylation was also impaired in mps1‐3 cells. Pre‐synchronised G1 cells were released at 34°C in the presence or the absence of nocodazole. Although several forms of Bub1 with different electrophoretic mobility were present throughout the cell cycle, one particular isoform with slowest mobility appeared in the wild type at 20′ after G1 release, i.e. when cells entered S phase, thus correlating with the timing of Bub1 kinetochore localisation (Fig 4E–H). While this hyperphosphorylation was transient in the absence of nocodazole, disappearing by 60′ (Fig 4G), it accumulated upon microtubule depolymerisation in wild‐type cells (Fig 4E). Under both conditions, this isoform was not apparent in mps1‐3 cells, suggesting that it is phosphorylated by kinetochore‐bound Mps1 (Fig 4E and G). Furthermore, Bub1 hyperphosphorylation was not restored in the mps1‐3 spc105‐18 and mps1‐3 GLC7‐24 suppressors undergoing cell cycle progression in the absence (Appendix Fig S6C) or in the presence of nocodazole (Appendix Fig S6D), although a modest Bub1 hyperphosphorylation could be observed in the mps1‐3 GLC7‐24 suppressor treated with nocodazole at late time points (Appendix Fig S6D). Thus, the mechanism underlying the suppression of mps1‐3 mutant cells might be independent of the phosphorylation status of Bub1.

Since Bub1 has been involved in kinetochore biorientation in budding yeast 8, 66, we conclude that lack of Bub1 recruitment to kinetochores is likely the main reason underlying the chromosome segregation defects of mps1‐3 cells. Lowering PP1 levels at Spc105 in the suppressors favours phosphorylation of the MELT repeats of Spc105, thereby restoring Bub1 recruitment at kinetochores and chromosome biorientation.

Artificial targeting of Bub1 to Spc105 restores proper chromosome segregation and SAC response in mps1‐3 cells

If our above conclusions are correct, artificial tethering of Bub1 to kinetochores could rescue the chromosome segregation defects of mps1‐3 cells at high temperatures. To test this hypothesis, we generated yeast strains expressing as sole source of Spc105 an Spc105 variant where we had inserted the GFP‐binding domain upstream of the MELT repeats of Spc105 (Spc105‐GBD). In the same strains, we co‐expressed Bub1‐GFP or Bub3‐GFP (note that although Bub3 enhances Bub1 binding to Spc105 by directly reading its phosphorylated MELT repeats, it has no additional role in Bub1 activation 16, 67). This strategy led to constitutive recruitment of Bub1 to kinetochores throughout the cell cycle (Fig 5A and Movie EV1), with no obvious impact on the proliferation rate of otherwise wild‐type cells (Fig 5B), indicating that prolonged Bub1 residence at kinetochores is insufficient to constitutively activate the SAC. Artificial recruitment of Bub1 to Spc105 only modestly, if at all, rescued the lethality of mps1‐3 cells at high temperatures (Fig 5B). However, it did prolong the viability of mps1‐3 cells after transient incubation at restrictive temperature (Fig 5C), suggesting that proper chromosome segregation might be partially restored. In agreement with this hypothesis, FACS analysis of DNA contents in synchronised mps1‐3 cells, either lacking or carrying the SPC105‐GBD and BUB1‐GFP or BUB3‐GFP constructs, showed that the massive chromosome missegregation of mps1‐3 cells at 34°C was partially (BUB3‐GFP) or largely (BUB1‐GFP) rescued, as indicated by the bimodal distribution of DNA contents (Appendix Fig S7A and Fig 5D). Interestingly, artificial anchoring of Bub1 to Spc105 also partially suppressed the SAC defects of mps1‐3 cells (Appendix Fig S7B). Nevertheless, the rescue of chromosome missegregation was independent of restored SAC signalling, as it remained unaffected by MAD2 deletion (Fig 5C and D). Thus, the main reason underlying the chromosome biorientation, and possibly SAC defects, of mps1‐3 cells is the lack of Bub1 at kinetochores.

Figure 5. Artificial recruitment of Bub1 to Spc105 is sufficient to suppress the chromosome biorientation defects of mps1‐3 cells.

• A
  SPC105‐GBD cells expressing Bub1‐GFP were grown at 30°C and filmed every 4 min at 30°C by time‐lapse fluorescence microscopy. DIC: differential interference contrast. Scale bar: 5 μm.
• B
  Serial dilutions of stationary phase cultures of the indicated strains were spotted on YEPD and incubated at the indicated temperatures.
• C
  Cells with the indicated genotypes were grown at 25°C and then shifted at 34°C (t = 0). Every hour after the temperature shift, the same number of cells was plated on YEPD for each strain and incubated at 25°C to determine the number of colony‐forming units. Percentages of viable cells have been calculated for each strain relative to t = 0. Error bars: SD. N = 3.
• D
  Cells with the indicated genotypes were synchronised in G1 with α‐factor at 25°C and then released in fresh medium at 34°C. Cells were collected at the indicated time points for FACS analysis of DNA contents.

The function of Mps1 and Bub1 in chromosome biorientation has been previously linked to pericentromeric localisation of Shugoshin, which in yeast requires Mps1 and Bub1 kinase activity 10, 66, 68, 69. Shugoshin senses the lack of kinetochore tension and engages the error correction machinery through kinetochore recruitment of the CPC complex (chromosome passenger complex) 68, 69, 70, 71, 72, 73, 74. Although the role of budding yeast Mps1 in chromosome biorientation seems independent of the CPC and Aurora B activity 6, 8, we decided to investigate if the chromosome segregation defects of mps1‐3 cells could be ascribed to poor Shugoshin activity at pericentromeres. Although co‐localisation of Sgo1 (the only budding Shugoshin) with kinetochore clusters was partially compromised in mps1‐3 cells at 34°C (Appendix Fig S8A), in agreement with previous observations 8 and with the loss of Bub1 at kinetochores, mps1‐3 cells turned out to be synthetically lethal with SGO1 deletion even at 25°C (Appendix Fig S8B), strongly arguing against a linear pathway where Mps1 acts upstream of Sgo1.

The role of budding yeast Aurora B (Ipl1) in chromosome biorientation relies on its alternative targeting to centromeres or the inner kinetochore 75, 76, 77. While centromeric localisation depends on histone H2A phosphorylation by Bub1 and subsequent recruitment of Shugoshin and the Bir1 subunit of CPC (survivin in metazoans) 71, 77, 78, kinetochore localisation requires binding of Aurora B and the Sli15 subunit of CPC (INCENP in metazoans) to the COMA complex 75, 77. A truncated Sli15 lacking its N‐terminal region (Sli15‐ΔNterm) or alanine mutations in its CDK‐dependent phosphorylation sites (Sli15‐6A) might favour Ipl1 binding to the COMA complex and/or microtubules 75, 77, 79 and, concomitantly, rescue the lethality/growth defects of bir1Δ, bub1Δ and sgo1Δ cells 80. If the chromosome segregation defects of mps1‐3 cells arose exclusively from poor activity of centromeric Sgo1/CPC, then they should be similarly suppressed by the Sli15‐ΔNterm and Sli15‐6A variants. This was definitely not the case, as shown by proliferation assays (Appendix Fig S8C) and FACS analysis of DNA contents on synchronised cells (Appendix Fig S8D).

Finally, since Dam1 is a critical phosphorylation target of Ipl1 in the error correction process 12, 81, we monitored its phosphorylation state in wild‐type, mps1‐3 and ipl1‐321 cells during a synchronous release from G1 at 34°C. While the electrophoretic mobility of Dam1‐3PK was dramatically altered by the ipl1‐321 mutation, it was mainly unaffected in mps1‐3 cells (Appendix Fig S8E), consistent with the notion that phosphorylation of Dam1 by Mps1 is dispensable for chromosome biorientation 12, 13 and suggesting that Ipl1 kinase activity is largely proficient in mps1‐3 cells.

We therefore conclude that lack of Bub1 at kinetochores is mainly responsible for the chromosome biorientation flaws of mps1‐3 mutant cells, at least partly independently of the role of Bub1 in Sgo1 kinetochore recruitment and CPC activation.

Discussion

The mps1‐3 mutant as a novel genetic tool to study the role of Mps1 in chromosome segregation

The precise function of budding yeast Mps1 in chromosome biorientation has been elusive so far. We have shed light on this process through the characterisation of the conditional mps1‐3 mutant that at 34°C is defective in this process but proficient at SPB duplication and bipolar spindle assembly. Our analysis of the segregation pattern of chromosome V in mps1‐3 cells suggests that each pair of sister chromatids has equal chance to segregate in a bipolar or monopolar fashion. This random segregation applied to the 16 yeast chromosomes makes the likelihood for each single mps1‐3 cell to segregate correctly all chromosomes extremely low (1.5 × 10⁻⁵), which is in agreement with the massive chromosome missegregation we observe by FACS analysis. Being mps1‐3 cells also SAC‐deficient makes monopolar attachments undetected by the checkpoint, thus generating highly aneuploid cells.

We could establish through live cell imaging that mps1‐3 cells are not defective in kinetochore–microtubule attachment, but rather in the error correction mechanism that converts monotelic or syntelic linkages into amphitelic attachments, in agreement with previous conclusions 6, 8. This function has been classically attributed to Aurora B and the CPC (reviewed in ref. 82), but in more recent years Mps1 has also been implicated in this process. Mps1 was shown to impact on Aurora B activity through phosphorylation of the CPC subunit borealin 4, as well as to act downstream of Aurora B 7, suggesting a complex interplay between the two kinases. In contrast, in budding yeast Mps1 and Aurora B (Ipl1) were proposed to operate independently 6, 8.

While ipl1 mutants missegregate chromosomes preferentially towards the old SPB, presumably because the new SPB must undergo a maturation time before becoming fully able to nucleate microtubules 45, we find different results in mps1‐3 cells depending on the experimental conditions. In particular, during the first cell cycle at restrictive temperature mps1‐3 cells missegregate chromosomes preferentially towards the bud (i.e. towards the old SPB) similar to ipl1 mutant cells. In contrast, during the following cell cycles this bias seems to be progressively lost. These results can be rationalised by surmising that Mps1 and Ipl1 are both involved in error correction of kinetochore attachments, but Mps1 could be additionally required for SPB maturation. In other words, when Mps1 activity is disabled, young SPBs fail to age properly and SPB asymmetry is obliterated, thus explaining the loss of missegregation bias after the second cell cycle at restrictive temperature. It is worth noting that apparently contradictory data exist in the literature as to whether in mps1 mutants chromosomes missegregate preferentially towards the bud 5, 6, 83. The above hypothesis offers a plausible explanation to reconcile these data.

Although the mps1‐3 Ser635Phe mutation lies in the kinase domain, the in vitro kinase activity of the corresponding protein is surprisingly enhanced, rather than impaired, at high temperature. The crystal structure of the Mps1 kinase domain has been solved 84 and reveals that Ser635 is semi‐buried, thus precluding any straightforward prediction of the possible outcome of its mutation on Mps1 folding or activity. Importantly, the Mps1‐3 protein does not accumulate at kinetochores, suggesting that lack of phosphorylation of specific kinetochore substrates underlies the chromosome biorientation and SAC defects of mps1‐3 cells. For what concerns SAC signalling, although MPS1 overexpression is able to hyperactivate SAC in the absence of functional kinetochores 85, our data imply that endogenous levels of Mps1 must accumulate at kinetochores for a productive SAC response, in agreement with previous reports 9, 86, 87, 88. Mps1 kinetochore recruitment depends on its N‐terminus that interacts with the kinetochore protein Ndc80/Hec1 86, 87. Additionally, Mps1 kinase activity promotes its own turnover at kinetochores 7, 40, 49, 50. It is therefore tempting to speculate that the enhanced kinase activity of Mps1‐3 might accelerate its own turnover and account for its low levels at kinetochores. The characterisation of the intragenic suppressors identified in our genetic screen might help shedding light on this hypothesis.

Mps1 promotes SAC signalling and chromosome biorientation through a common molecular mechanism

Our unbiased screen for extragenic suppressors of mps1‐3 has been particularly revealing and points at the antagonism between Mps1 and PP1 at kinetochores as critical for the correction of faulty kinetochore–microtubule attachments and correct chromosome segregation. Our data strongly argue that the same mechanism implicated in SAC signalling, i.e. Spc105 phosphorylation and subsequent Bub1 recruitment, is a crucial function of Mps1 also in chromosome biorientation and independently of SAC activation. Several lines of evidence are consistent with this conclusion. First, the spc105 and GLC7 suppressors that we have identified can re‐establish both proper chromosome segregation and SAC signalling in mps1‐3 cells without bringing back Mps1 to kinetochores. Second, Spc105 phosphorylation and Bub1 kinetochore localisation, which are severely impaired in mps1‐3 mutant cells, are restored in the suppressors. Third, constitutive binding of Bub1 to Spc105 is sufficient to rescue considerably chromosome missegregation and partly SAC signalling in mps1‐3 cells at restrictive temperature. The incomplete rescue of SAC activation is not surprising, given that Mps1 plays additional roles in the SAC, such as promoting Bub1 and Mad1 phosphorylation 19, 20. Along this line, it is worth noting that the spc105‐18 and GLC7‐24 suppressors, which restore a normal SAC response in mps1‐3 cells, also restore wild‐type Mad1 (but not Bub1) hyperphosphorylation upon nocodazole treatment.

The idea that a common molecular mechanism based on Knl1/Spc105 phosphorylation would underlie the role of Mps1 in both SAC and chromosome biorientation was previously proposed in fission yeast. Indeed, phospho‐mimicking mutations in the MELT repeats of Spc7 (the fission yeast homologue of Knl1/Spc105) were shown to partially suppress the chromosome biorientation defects of mps1Δ cells 18. However, in fission yeast Mps1 is not essential for viability, and mps1Δ cells experience only modest chromosome missegregation 18. Thus, our data gathered in an organism where Mps1 is critical for chromosome biorientation and with the aid of an unbiased genetic screen largely extend and strengthen this idea.

Despite our data strongly argue that Spc105 is a critical Mps1 substrate for the error correction pathway, we cannot rule out that other Mps1 substrates contribute to this process. Further experiments will help refining the precise mechanism underlying the control of chromosome segregation by Mps1.

The N‐terminal region of Knl1 and Bub1 was previously shown to stimulate Aurora B kinase activity at kinetochores and increase the turnover of erroneous kinetochore–microtubule attachments 89, 90, 91. Furthermore, Bub1 has been involved in chromosome biorientation through phosphorylation of histone H2A and Sgo1 accumulation at pericentromeres 66, 78, 92, 93. Sgo1, in turn, enhances CPC activity in the error correction mechanism 71, 74, 94. Consistent with the idea that Mps1 is required for Bub1 recruitment to kinetochores, Sgo1 co‐localisation with kinetochores is compromised in mps1 mutants (ref. 8 and our data). Our data suggest, however, that impaired Sgo1 kinetochore localisation does not fully account for the chromosome biorientation defects of mps1‐3 cells. Indeed, Sgo1 centromeric recruitment is weakened, but not abolished, in mps1‐3 cells and the mps1‐3 mutation and SGO1 deletion are synthetically lethal. Furthermore, mutations in the CPC centromere‐targeting subunit Sli15 that rescue the sickness of sgo1Δ and bub1Δ cells 80 have no ameliorating effects on the lethality and chromosome segregation defects of mps1‐3 cells at high temperature. Finally, phosphorylation of the Aurora B kinetochore target Dam1 is unaffected in mps1‐3 cells, suggesting that the error correction defects of this mutant are not caused by reduced Aurora B kinase activity. Thus, although Mps1 could promote chromosome biorientation partly through the Bub1‐Sgo1‐CPC axis, it must have at least one additional function in this process. Consistent with this hypothesis, Bub1 and Sgo1 are not essential proteins, unlike Mps1, and their deletion leads to milder chromosome segregation defects than mps1‐3 cells at restrictive temperature. One intriguing possibility that is supported by our genetic results is that Mps1 could directly harness PP1 activity at kinetochores. In this scenario, Mps1 would master two opposing activities, acting indirectly on the CPC and perhaps more directly on PP1, thereby allowing the highest accuracy of bipolar attachments. This hypothesis would also explain why the spc105 and GLC7 suppressors isolated in our screen rescue the lethality and chromosome segregation defects of mps1‐3 cells considerably better than artificial recruitment of Bub1 to Spc105. Whether and how Mps1 controls PP1 activity at kinetochores will be an important subject for future investigations.

In sum, our data provide experimental support to the hypothesis that a single sensory device may be used for both error correction of faulty kinetochore–microtubule attachments and SAC signalling 89, 95. Given the evolutionary conservation of the players involved, it is very likely that Mps1 promotes SAC activation and chromosome biorientation through a common mechanism also in other organisms.

Materials and Methods

Yeast strains, plasmids and growth conditions

All yeast strains (Appendix Table S1) are congenic to or at least four times backcrossed to W303 (ade2‐1, trp1‐1, leu2‐3,112, his3‐11 and 15 ura3).

One‐step tagging techniques were used to generate 3HA‐, 3PK‐, GFP‐ and mCherry‐tagged proteins at the C‐terminus. The CEN5::tetO2X224 (12.6 kB) tetR‐GFP strain was a generous gift from T. Tanaka 43; BUB1‐GFP from P. de Wulf; MTW1‐Tomato from A. Marston 68; GFP‐TUB1 from A. Straight 96; mCherry‐TUB1 from A. Amon 97; PDS1‐myc18 from K. Nasmyth 98; spc105‐RASA and GLC7‐spc105‐RASA mutants from H. Funabiki and F. Cross 28; SGO1‐GFP and CEN5::tetO2X112 (1.4 kB) tetR‐GFP from E. Schwob; and ura3::URA3::tetOX224 (35 kB) tetR‐GFP from K. Nasmyth. The sli15‐ΔNterm (sli15Δ2‐228) and sli15‐6A plasmids (pCC253 and pCC297) were a gift of C. Campbell 80 and were integrated at the SLI15 locus by cutting with HindII and SalI, respectively.

The mps1‐3 mutant was obtained by hydroxylamine‐induced mutagenesis of an MPS1‐bearing plasmid and selection of temperature‐sensitive mutants through plasmid shuffling 99. The mps1‐3 allele (entire coding sequence) together with 480 bp of 5′ UTR and 900 bp of 3′ UTR was cloned as an EcoRI/SalI fragment in the integrative plasmid pFL34 to generate pSP1041 and integrated at the MPS1 locus by the Pop‐in/Pop‐out method 100.

The SPC105‐GBD‐bearing plasmid (pSP1530) was generated by cloning in KpnI/HindIII of Yiplac204 a KpnI/PfoI fragment containing SPC105 promoter and part of the coding region from plasmid SB1884 15 together with a PfoI/HindIII SPC105 synthetic gene (aa 61–386 of Spc105) carrying the GBD domain after aa 130 (pSP1516). SPC105‐GBD was integrated at the SPC105 locus by digesting pSP1530 with BglII. This generates a partial duplication of SPC105, with full‐length SPC105‐GBD gene followed by a truncated spc105 allele (aa 1–386).

The GALs‐MPS1 plasmid (pSP149) was constructed by amplifying a DNA fragment comprising 30 bp of MPS1 5′ UTR, its CDS and 105 bp of 3′ UTR by PCR from W303 genomic DNA with primers SP127 and SP129, followed by its cloning into EcoRI/XhoI of a pRS416 harbouring the attenuated GALs promoter 101. Growth of wild‐type cells carrying the GALs‐MPS1 construct is unaffected (data not shown), in contrast to high‐level overexpression of MPS1 that causes a SAC‐dependent metaphase arrest 62.

The GAL1‐GST‐MPS1 plasmid (pSP228) was constructed by amplifying MPS1 by PCR with primers SP238 and SP239 and cloning it into pEG(KT) 102 after cutting SmaI/HindIII. The GAL1‐GST‐mps1‐3 plasmid (pSP1202) was generated by replacing a BspEI/XbaI fragment of pSP228 internal to MPS1 coding region by the same fragment containing the mps1‐3 mutation from pSP1041.

The plasmids for expression of His‐tagged Mps1 and Mps1‐3 in bacteria (pSP1575, pSP1576) were generated by amplifying the coding sequence of MPS1/mps1‐3 by PCR with primers MP1124‐MP1125, followed by cloning into EcoRI/HindIII of pPROEX‐HTa.

The plasmid for expression of His‐tagged Spc105 N‐terminus (aa 1–320) in Escherichia coli (pSP1571) was generated by PCR‐amplifying part of the coding sequence of SPC105 with primers MP1126‐1127, followed by cloning into BamHI/PstI of pPROEX‐HTa.

Yeast cultures were grown at 25–30°C, in either SD medium (6.7 g/l yeast nitrogen base without amino acids) supplemented with the appropriate nutrients or YEP (1% yeast extract, 2% bactopeptone, 50 mg/l adenine) medium. Raffinose was supplemented to 2% (SD‐raffinose or YEPR), glucose to 2% (SD‐glucose or YEPD), and galactose to 1% (SD‐raffinose/galactose or YEPRG). Cells were synchronised in G1 by alpha factor (4 μg/ml) in YEP medium containing the appropriate sugar at 25°C. G1 arrest was monitored under a transmitted light microscope, and cells were released in fresh medium (typically after 120–135 min of alpha factor treatment) at 34°C, unless otherwise specified, after being collected by centrifugation at 2,000 g followed by one wash with YEP containing the appropriate sugar. Nocodazole was used at 15 μg/ml.

Primers used in this study for gene tagging: sequences in bold anneal to the tag‐bearing cassette

SP84 (tagging BUB1 with 3HA::K.l.URA3; fwd)

ACGTAATTCTAAGCATTGAAGAGGAGTTATCACATTTTCAATATAAGGGGAAACCGTCAAGGAGATTTTCCGGTTCTGCTGCTAG

SP85 (tagging BUB1 with 3HA::K.l.URA3; rev)

GTCATTGCTATGGAATCTGGCAGGACACCAAAAAGTCACCTATGCGGGAGATGAAGGCATATTTATTCACCTCGAGGCCAGAAGAC

SP96 (tagging MPS1 with 3HA::K.l.URA3; fwd)

CATATATCACAAGATGATCTCAATGATGTGGTAGACACTGTTTTAAGGAAATTTGCAGATTACAAAATTTCCGGTTCTGCTGCTAG

SP97 (tagging MPS1 with 3HA::K.l.URA3; rev)

CATGCATGGCAAACTCTAAATGTATTTATGTTCATAACTGGCACATGCTTTTCTTCCTTATGCGGCTCTTCCTCGAGGCCAGAAGAC

SP134 (tagging MAD1 with 3HA::K.l.URA3; fwd)

GATAGAGGTCAACTTCCGTGCTTTTTGGCAACAATAACATTGCGTCTGTGGGAACAGCGACAAGCCAAATCCGGTTCTGCTGCTAG

SP135 (tagging MAD1 with 3HA::K.l.URA3; rev)

ATGTCAGCGGATAGGAGTTTATCATATTATAAAACCGATTACTATTATCTATTAGAAATGTATATACACCCTCGAGGCCAGAAGAC

MP560 (tagging SPC97 with mCherry::HPHMX; fwd)

GTTCTTTACTCTATAGTACCTCCTCGCTCAGCATCTGCTTCTTCCCAAAGAGGTCGACGGATCCCCGGG

MP561 (tagging SPC97 with mCherry::HPHMX; rev)

GATTCCACTTTCCGCAAGTTGGTGCACGTCGTTAGTGACATAACGCCGCGTATCGATGAATTCGAGCTCG

MP629 (tagging mps1‐3 with 3HA::K.l.URA3; fwd)

CTCAATGATGTGGTAGACACTGTTTTAAGGAAATTTGCAGATTACAAAATTTCCGGTTCTGCTGCTAG

MP630 (tagging mps1‐3 with 3HA::K.l.URA3; rev)

ATGTATTTATGTTCATAACTGGCACATGCTTTTCTTCCTTATGCGGCTCTTCCTCGAGGCCAGAAGAC

MP761 (tagging MPS1 with 6Gly‐eGFP::kanMX; fwd)

CTCAATGATGTGGTAGACACTGTTTTAAGGAAATTTGCAGATTACAAAATTGGGGGAGGCGGGGGTGGAGGTGACGGTGCTGGTTTA

MP762 (tagging MPS1 with 6Gly‐eGFP::kanMX; rev)

ATGTATTTATGTTCATAACTGGCACATGCTTTTCTTCCTTATGCGGCTCTTTCGATGAATTCGAGCTCG

MP763 (tagging SPC105 with 3PK::K.l.HIS3; fwd)

AGTGGAGTTCTTCCTTCATTTACGAAAAGTAGAATACATTTAGAGTTTACGTCCGGTTCTGCTGCTAG

MP764 (tagging SPC105 with 3PK::K.l.HIS3; rev)

AAAAAAAAGTGATGAGATATTACTAGTCATCGTTGTCCTATTATAAACACTCCTCGAGGCCAGAAGAC

MP894 (tagging DAM1 with 3PK::K.l.HIS3; fwd)

TCTGTTGCAAAGAAAACTGAAAAAAAAATAAATACAAGGCCCCCCTTCAGATCCGGTTCTGCTGCTAG

MP895 (tagging DAM1 with 3PK::K.l.HIS3; rev)

CTAGCGATATATTTTGTGAGGAGGATAATTCTTTGGTTGGGTTGGGCGTAGCCTCGAGGCCAGAAGAC

MP900 (tagging SNF1 with 3PK::K.l.HIS3; fwd)

CATTTAACAACAAAACTAATTATGGAATTAGCCGTTAACAGTCAAAGCAATTCCGGTTCTGCTGCTAG

MP901 (tagging SNF1 with 3PK::K.l.HIS3; rev)

ACGATACATAAAAAAAAGGGAACTTCCATATCATTCTTTTACGTTCCACCACCTCGAGGCCAGAAGAC

MP940 (tagging SLI15 with 3HA::K.l.URA3; fwd)

AATCCTAGGCTAAACAGGTTGAAACCGCGTCAAATTGTGCCCAAAAGGTCTTCCGGTTCTGCTGCTAG

MP941 (tagging SLI15 with 3HA::K.l.URA3; rev)

CTTGATTTAATGTTAACCAGTTTGAATTTTTCTTTTCTGGGGTAATCGAATCCTCGAGGCCAGAAGAC

MP1024 (tagging BUB3 with eGFP::kanMX; fwd)

ACTATTGAACTAAACGCAAGTTCAATATACATAATATTTGACTATGAGAACGGGGGAGGCGGGGGTGGAGGTGACGGTGCTGGTTTA

MP1025 (tagging BUB3 with eGFP::kanMX; rev)

AATTTTTTTTCTGGAATGTTCTATCATACTACACGAATCTTCACGAAGATATCGATGAATTCGAGCTCG

MP1128 (tagging MTW1 with GBD::kanMX; fwd)

GTTAGTATAGATATTGAAGAGCCTCAATTGGATTTACTTGATGATGTGTTACGTACGCTGCAGGTCGAC

MP1129 (tagging MTW1 with GBD::kanMX; rev)

ATTCGTGAATACATACATCATATCATAGCACATACTTTTTCCCACTTTATAATCGATGAATTCGAGCTCG

Primers used in this study for cloning: sequences in bold indicate the restriction sites

SP127 (to clone MPS1, EcoRI)

CGGAATTCCTGCAAGAACTGGATGATCTC

SP129 (to clone MPS1, XhoI)

CCGCTCGAGCGGTGTAAAAAAGAGTTGATC

SP238 (cloning of GAL1‐GST‐MPS1; fwd)

TCCCCCGGGATCAACAAACTCATTCCATGATTAG

SP239 (cloning of GAL1‐GST‐MPS1; rev)

CCCAAGCTTAAGTATTCAGTGTCTGTGTCG

MP1124 (cloning of MPS1, EcoRI; fwd)

ACTTAGAATTCTCAACAAACTCATTCCATGATTATG

MP1125 (cloning of MPS1, HindIII; rev)

ACTTAAAGCTTCTAAATTTTGTAATCTGCAAATTTC

MP1126 (cloning N‐terSPC105, BamHI; fwd)

ACTTAGGATCCAATGTGGATGAAAGAAGCCG

MP1127 (cloning N‐terSPC105, PstI; rev)

ACTTACTGCAGCTAGTCAAAGTTTGTCTGTATTTGCG

Microscopy

Scoring of GFP‐tagged chromatid separation was carried out on cells fixed in 100% ethanol at −20°C. After two washes with PBS, cells were laid between a glass slide and a coverslip and imaged by fluorescence microscopy (Epifluorescent widefield Zeiss Axioimager Z1 controlled by the Metamorph software). Z‐stacks (11 planes, 0.3 μm spaced) were max‐projected before scoring.

To detect spindle formation and elongation, alpha‐tubulin immunostaining was performed on formaldehyde‐fixed cells using the YOL34 monoclonal antibody (1:100; MCA78S AbDSerotec, Raleigh, NC), followed by indirect immunofluorescence using CY2‐conjugated anti‐rat antibody (1:100; 31645 Pierce Chemical Co.).

For time‐lapse video microscopy, cells were mounted on 1% agarose pads in SD medium on Fluorodishes (World Precision Instruments) and filmed at controlled temperature with a 100× 1.45 NA oil immersion objective mounted on a Spinning disk CSU‐X1 Andor Nikon Eclipse Ti microscope coupled to an iXon Ultra camera controlled by the Andor iQ3 software. Z‐stacks of 10 planes were acquired every 2–4 min with a step size of 0.5 μm (Figs 1G, 3G, 4C, and 5A, Movie EV1), 0.3 μm (Appendix Fig S8A) or 0.67 μm (Appendix Fig S1A).

Images of Sgo1‐GFP (Appendix Fig S8A) have been deconvolved using the software Huygens using the following settings: background = 200; ratio signal/noise = 4; and iterations = 5.

Budding index was scored under a transmitted light microscope on cells fixed with 70% ethanol and treated for FACS analysis. Percentage of nuclear division was determined on the same samples after DNA staining with propidium iodide under a fluorescence microscope using a CY3 filter.

FACS analysis of DNA contents

For DNA quantification by fluorescence‐activated cell sorting (FACS), 1–2 × 10⁷ cells were collected, spun at 10,000 g and fixed with 1 ml of 70% ethanol for at least 30 min at RT. After one wash with 50 mM Tris–Cl pH 7.5, cells were resuspended in 0.5 ml of the same buffer containing 0.05 ml of a preboiled 10 mg/ml RNAse solution and incubated overnight at 37°C. The next day cells were spun at 10,000 g and resuspended in 0.5 ml of 5 mg/ml pepsin freshly diluted in 55 mM HCl. After 30‐min incubation at 37°C, cells were washed with FACS buffer (200 mM Tris–Cl pH 7.5, 200 mM NaCl, 78 mM MgCl₂) and resuspended in the same buffer containing 50 μg/ml propidium iodide. After a short sonication, samples were diluted (1:20–1:10) in 1 ml of 50 mM Tris–Cl pH 7.5 and analysed with a FACSCalibur device (BD Biosciences). 10,000 events were scored for each sample and plotted after gating out the debris.

Protein extracts, Western blotting and kinase assays

For TCA protein extracts, 10–15 ml of cell culture in logarithmic phase (OD₆₀₀ = 0.5–1) were collected by centrifugation at 2,000 g, washed with 1 ml of 20% TCA and resuspended in 100 μl of 20% TCA before breakage of cells with glass beads (diameter 0.5–0.75 mm) on a Vibrax VXR (IKA). After addition of 400 μl of 5% TCA, lysates were centrifuged for 10 min at 845 g. Protein precipitates were resuspended in 100 μl of 3× SDS sample buffer (240 mM Tris–Cl pH 6.8, 6% SDS, 30% glycerol, 2.28 M β‐mercaptoethanol, 0.06% bromophenol blue), denatured at 99°C for 3 min and loaded on SDS–PAGE after elimination of cellular debris by centrifugation (5 min at 20,000 g).

Proteins were wet‐transferred on Protran membranes (Schleicher and Schuell) overnight at 0.2 A and probed with monoclonal anti‐HA 12CA5 (1:5,000); anti‐PK (alias anti‐V5; MCA1360 AbDSerotec, 1:3,000); polyclonal anti‐Cdc5 (sc‐6733 Santa Cruz, 1:3,000); anti‐Clb2 (a generous gift from W. Zachariae, 1:2,000); anti‐Pgk1 (Invitrogen Molecular Probes, 1:40,000); monoclonal anti‐myc 9E10 (1:5,000); anti‐phospho‐AMPK (Thr172) (Cell Signalling, 1:1,000); and anti‐GST (1:2,000, 27‐4577‐01 GE Healthcare). Antibodies were diluted in 5% low‐fat milk (Regilait) dissolved in TBST. Secondary antibodies were purchased from GE Healthcare, and proteins were detected by a home‐made enhanced chemiluminescence system.

For kinase assays with recombinant Mps1 proteins, Mps1/Mps1‐3 and Spc105 have been purified as follows: BL21 DE3 cells transformed with pSP1571, pSP1575 or pSP1576 were grown in LB containing 50 μg/ml of ampicillin, 34 μg/ml of chloramphenicol and 1% glucose and induced with 0.2 mM IPTG at 16°C overnight. Cells were lysed in 50 mM Tris–Cl pH 8, 300 mM NaCl, 0.1% Tween 20 containing 1 mg/ml of lysozyme and a cocktail of protease inhibitors (Complete EDTA‐free Roche). Cells were then sonicated 3× with 1′30″ cycles of 8″ pulses/8″ ice and extracts cleared at 30,000 g for 30′ at 4°C. Lysates were incubated 2 h with 1 ml of Ni‐NTA agarose beads (Qiagen), pre‐washed three times with 50 mM Tris–Cl pH 8, 300 mM NaCl, 0.1% Tween 20 and 10 mM imidazole at 4°C. After incubation with the extracts, the slurry was washed three times with 50 mM Tris–Cl pH 8, 300 mM NaCl, 0.1% Tween 20 and 20 mM imidazole and loaded on Polyprep columns (Bio‐Rad). Fractions of 0.5 ml were eluted with 50 mM Tris–Cl pH 8, 300 mM NaCl and 250 mM imidazole and quantified by Nanodrop. The most concentrated fractions were washed and concentrated through the use of Amicon Ultra filter units (10 kDa cut‐off for Spc105 and 50 kDa cut‐off for Mps1) to finally have them in 50 mM Tris–Cl pH 8, 50 mM NaCl. Glycerol was added to the final concentration of 20%. Protein concentration was quantified by loading them on SDS–PAGE along with a BSA standard curve followed by staining with Coomassie blue. Kinase assays were performed at 34°C for different times in 25 μl of a reaction mix containing 50 mM Tris–Cl pH 7.5, 10 mM MgCl₂, 0.5 mM DTT, 10 μM ATP, 1 μl ATP‐γ^P33 3,000 Ci/mmol and 80 μg/ml MBP. 75 ng of Spc105 N‐term fragment and 1 μg Mps1/Mps1‐3 kinases were used for each assay. Reactions were stopped using 7 μl of 3× SDS sample buffer (described in TCA extracts). Samples were then boiled 3 min at 95°C and loaded on SDS–PAGE. The gel was then dried and exposed on autoradiography films.

Native yeast protein extracts for kinase assays were performed by collecting cells in logarithmic phase (OD₆₀₀ = 1) grown in ‐Ura RG at 30°C, centrifuging at 2,000 g and washing with 1 ml of cold 10 mM Tris–Cl pH 7.5. Cells were then broken in 50 mM Tris–Cl pH 7.5, 50 mM NaCl, 0.2% Triton X‐100 containing a cocktail of protease inhibitors (Complete EDTA‐free Roche) and phosphatase inhibitors (PhosSTOP Roche). Lysates were cleared at 20,000 g for 10 min at 4°C. The equivalent of 6 OD₆₀₀/assay was incubated on a nutator for 2 h at 4°C with 50 μl of Glutathione Sepharose 4 Fast Flow (GE Healthcare) beads pre‐washed with the above buffer. After incubation with the extracts, beads were spun down at 4°C at 845 g, washed twice with the same buffer without protease and phosphatase inhibitors, twice with the same buffer containing 0.1% of Triton X‐100, twice with the same buffer without Triton X‐100 and once with 50 mM Tris–Cl pH 7.5, 10 mM MgCl₂ and 0.5 mM DTT. Kinase assays were performed as above using 2 μg of MBP as exogenous substrate.

Chromosome loss assay

This assay is based on the loss of heterozygosity of the MAT locus, carried by chromosome III 42. Briefly, wild‐type diploid cells are heterozygous MATa/MATalpha and cannot mate unless they lose one copy of the MAT locus through mutations, chromosome re‐arrangements or chromosome loss. Cells were grown overnight at 25°C, and their concentration was measured at the time of inoculation and the day after to calculate the number of cell divisions. 2 × 10⁷ diploid Met⁺ cells were incubated 4 h at RT with an excess (2 × 10⁸) of MATalpha met4 tester cells in 400 μl of YEPD. 10 μl of the suspension was then diluted to plate 500 cells on −Met plates (where only the Met⁺ diploids can form colonies) to calculate the percentage of viable (i.e. colony‐forming); the remaining cells were plated on minimal SD plates (MIN), where only triploids, generated from the mating of diploids with the haploid tester cells, can grow. Plates have been incubated at 25°C. Frequency of chromosome loss has been calculated as follows: {[(no. of colonies on MIN/2 × 10⁷)/fraction of viable cells]/no. of cellular divisions}. We note that with this assay we likely underestimate the real frequencies of chromosome loss because our measurements do not take into account the frequency of mating for each strain, which probably is much lower than 100%.

Suppressor screen

2.6 × 10⁷ mps1‐3 cells were plated on YEPD and incubated at 34°C for 3–5 days until colonies’ appearance. The isolated spontaneous suppressors were then characterised by classical genetic tests to distinguish intragenic from extragenic mutations, recessive from dominant mutations, and to group the suppressors in complementation/allelic groups. Genomic DNA was extracted from one representative suppressor for each complementation/allelic group and sent for whole genome sequencing to identify the mutated genes. The identity of the suppressing mutations was confirmed by classical genetics. Targeted gene re‐sequencing approaches were used to identify the suppressing mutations in the other extragenic suppressors.

Viability test

Cells were grown at 25°C to 2 × 10⁶ cells/ml and then serially diluted to plate 200 cells/plate in duplicate (wild‐type, mps1‐3 SPC105‐GBD BUB1‐GFP and mps1‐3 SPC105‐GBD BUB3‐GFP) or 200, 500 and 1,000 cells/plate in duplicate for mps1‐3 cells (t = 0). Cells were then shifted to 34°C, withdrawn every hour, serially diluted to plate 200 cells/plate in duplicate (wild‐type, mps1‐3 SPC105‐GBD BUB1‐GFP and mps1‐3 SPC105‐GBD BUB3‐GFP) or 200, 500 and 1,000 cells/plate in duplicate for mps1‐3 cells. Plates were incubated at 25°C until colony appearance. The number of colonies at each time point (average values of the duplicates) was divided by the number of colonies at t = 0 to obtain percentages of viability.

ChIP‐seq analysis

100 ml of cells (concentration 10⁷ cells/ml) was fixed in 1% formaldehyde for 30′ at RT with shaking and then transferred ON at 4°C. Cells were then spun down 3 min at 2,000 g at 4°C and washed three times with TBS. Cells were resuspended in 0.4 ml of 50 mM Hepes‐KOH pH 7.5, 140 mM NaCl, 1 mM EDTA, 1% Triton X‐100, 0.1% Na‐deoxycholate containing a cocktail of protease inhibitors (Complete EDTA‐free Roche) and 1 mM PMSF, followed by cell breakage through a Multi‐beads shaker (Yasui‐Kikai) at 0°C. Lysates were eluted in 15 ml Falcon tubes (centrifuged twice at 3,000 rpm for 1 min at 4°C), transferred in 1.5 Eppendorf tube, centrifuged at 2,500 g 1 min at 4°C and sonicated through a Sonifier Branson 2508 (5–7 cycles of 15 s, tune 1.5 with centrifugation of 5,000 rpm for 1 min between each cycle). Sonicated lysates were then centrifuged for 5 min at 21,000 g at 4°C. Supernatants were recovered. 10 μl of lysates was withdrawn at this stage (=Inputs). For immunoprecipitations (IPs), 60 μl of Protein G‐dynabeads pre‐washed twice with PBS containing 5 mg/ml BSA and pre‐adsorbed with 3 μg Babco 16B12 anti‐HA antibody was added to each lysate and incubated overnight at 4°C on a nutator. IPs were then washed twice with ice‐cold 50 mM Hepes‐KOH pH 7.5, 140 mM NaCl, 1 mM EDTA, 1% Triton X‐100, and 0.1% Na‐deoxycholate; twice with ice‐cold 50 mM Hepes‐KOH pH 7.5, 500 mM NaCl, 1 mM EDTA, 1% Triton X‐100, and 0.1% Na‐deoxycholate; twice with ice‐cold 10 mM Tris–Cl pH 8, 250 mM LiCl, 0.5% NP‐40, 0.5% Na‐deoxycholate and 1 mM EDTA; and once with ice‐cold TE (50 mM Tris–Cl pH 8, 1 mM EDTA pH 8). IPs were eluted at 65°C for 15 min with 50 μl of 50 mM Tris–Cl pH 8, 10 mM EDTA, 1% SDS, resuspended in 50 μl of TE/1% SDS and incubated overnight at 65°C to reverse crosslinking. IPs were then incubated at 37°C for 1 h with 100 μl of TE containing 10 μg of RNAseA and 2 h at 37°C with 2 μl of ProteinaseK 50 mg/ml. DNA was then purified with the Qiagen PCR purification kit. Libraries for Illumina sequencing have been then prepared following the NEBNext Ultra II DNA Library Prep Kit for Illumina. Sequencing results have been analysed using the DROMPA software 103.

Author contributions

GB characterised the phenotypes of the mps1‐3 mutant, carried out the screen for suppressors and studied the effects of artificial Bub1 recruitment to kinetochores. GB and SP did the Mps1 kinase assays and live cell imaging. AC generated the mps1‐3 mutant. YA, YK and KS analysed the kinetochore localisation of Mps1 by ChIP‐seq. GB and SP designed the work. GB made the figures with data provided by all contributing authors. SP supervised the work and acquired funding. SP and GB wrote the paper with input from all authors.

Conflict of interest

The authors declare that they have no conflict of interest.

Supporting information

Appendix

Movie EV1

Review Process File

EMBO Reports (2020) 21: e50257

Data availability

The ChIP‐seq data from this publication have been deposited to the GEO NCBI database (https://www.ncbi.nlm.nih.gov/geo/) and assigned the identifier GSE144741.