The yeast protein kinase Mps1p is required for assembly of the integral spindle pole body component Spc42p

In Saccharomyces cerevisiae, the spindle pole body (SPB)^* serves as the centrosome equivalent organelle. SPBs are duplicated once each cell cycle and function to nucleate microtubules that will form the mitotic spindle. Proper SPB duplication is required to form a bipolar spindle, which in turn is essential for cells to accurately segregate their DNA. SPB morphology is distinct from the typical centrosome of animal cells, having a trilaminar disc-like structure instead of centrioles surrounded by pericentriolar material (Byers and Goetsch, 1974; Kochanski and Borisy, 1990). The SPB remains embedded in the nuclear envelope throughout the cell cycle, allowing it to simultaneously nucleate nuclear and cytoplasmic microtubules (Byers and Goetsch, 1974). Despite morphological differences between the SPB and centrosome, there are several conserved components (Adams and Kilmartin, 2000). These include proteins that form the γ-tubulin complex in S. cerevisiae (Spc98p, Spc97p, and Tub4p), the centrin homologue, Cdc31p, and the molecular spacer protein, Spc110p (Kendrin) (Byers, 1981; Kilmartin et al., 1993; Geissler et al., 1996; Spang et al., 1996; Knop et al., 1997; Middendorp et al., 1997; Flory et al., 2000). Although the γ-tubulin complex clearly serves to nucleate microtubules in both the SPB and the centrosome, the conserved function performed by Cdc31p and Spc110p is still unclear (Moritz et al., 1995; Zheng et al., 1995; Knop and Schiebel, 1997; Marshall and Wilson, 1997; Adams and Kilmartin, 2000).

The SPB duplication pathway has been described through EM analysis of wild-type cells and mutant cells that fail at different stages of SPB duplication (Byers and Goetsch, 1975; Adams and Kilmartin, 1999). Duplication of the SPB occurs in G1 of the cell cycle, beginning with the accumulation of SPB components (the satellite) onto the cytoplasmic face of the half-bridge, a modification of the nuclear envelope distal to the SPB (Byers and Goetsch, 1975). The amorphous satellite appears to develop into a larger ordered structure called the duplication plaque (Adams and Kilmartin, 1999; O'Toole et al., 1999). Immuno-EM analysis of the satellite and duplication plaque show both structures are composed of the core SPB components, Spc29p, Spc94p/Nud1p, Spc42p, and Cnm67p (Adams and Kilmartin, 1999). Assembly of the new SPB is completed when the duplication plaque is inserted into the nuclear envelope and associates with additional SPB components that will make up the inner (nuclear) plaque layers (Adams and Kilmartin, 1999).

The terminal phenotype of various SPB duplication mutants has suggested when the gene products might be required in the process (Byers and Goetsch, 1975; Rose and Fink, 1987; Winey et al., 1991; Schutz et al., 1997; Schutz and Winey, 1998). For example, yeast containing a mutant SPC42 gene fail in SPB duplication after satellite formation (Donaldson and Kilmartin, 1996). The SPC42 gene product forms the electron-dense (two-dimensional crystalline) central layer of the SPB, and Spc42p is found in the duplication intermediates, the satellite, and duplication plaque (Donaldson and Kilmartin, 1996; Bullitt et al., 1997; Adams and Kilmartin, 1999; O'Toole et al., 1999). Unlike mutant alleles of SPC42, different mutant alleles of MPS1, which encodes a dual specificity protein kinase required for SPB duplication, fail at two distinct points in SPB duplication (Winey et al., 1991; Lauze et al., 1995; Schutz and Winey, 1998). This suggests that Mps1p is required for multiple events in SPB duplication.

Mps1p is unusual in that it has a role in the spindle checkpoint and in SPB duplication (Hardwick and Murray, 1995; Weiss and Winey, 1996). Kinetochores that are not attached to microtubules activate the spindle checkpoint (Wang and Burke, 1995; Pangilinan and Spencer, 1996). Failed SPB duplication also triggers this checkpoint, possibly because a monopolar spindle does not nucleate a sufficient number of microtubules to capture all of the kinetochores and cannot produce tension via bipolar spindle attachment (Winey and O'Toole, 2001). The conditional MPS1 mutants isolated thus far are defective in both pathways; therefore, under restrictive conditions MPS1 mutant cells proceed through mitosis with a monopolar spindle, aberrantly segregate their DNA, and rapidly lose viability (Winey et al., 1991; Schutz and Winey, 1998). Although the role of Mps1p in this checkpoint is not yet clearly defined, it requires kinase activity and is probably accomplished through phosphorylation of another checkpoint component, Mad1p (Hardwick et al., 1996).

Previously, we used a mps1-1 strain in genetic screens to identify interactions that would enhance our understanding of the role of Mps1p in SPB duplication and the spindle checkpoint (Schutz et al., 1997; Jones et al., 1999). We identified genes involved in the spindle checkpoint, spindle function, and those involved in stabilizing the Mps1p kinase (Hofmann et al., 1998; Jones et al., 1999). However, we did not identify SPB components. Here, we characterize a novel conditional allele, mps1-8, that is specifically defective in SPB duplication and use this allele in a dosage suppressor screen. We identified SPC42, a gene that encodes an integral SPB component, as a dosage suppressor of the mps1-8 conditional growth defect. We use genetic and biochemical techniques to investigate the interaction between Mps1p and Spc42p, taking advantage of an Spc42p in vivo assembly assay (Donaldson and Kilmartin, 1996) to show that Mps1p is required for Spc42p assembly.

We have investigated further the essential role of MPS1 by generating and characterizing a novel conditional MPS1 mutation, defective only in SPB duplication. This MPS1 allele was isolated from a library of mutagenized MPS1 clones prepared by amplifying the noncatalytic (NH₂-terminal) region of MPS1 using error-prone PCR. We targeted this region of MPS1 for mutagenesis because (a) existing conditional MPS1 alleles have mutations in the catalytic domain and are defective in both Mps1p functions and (b) deletion analysis showed the NH₂ terminus is required for function (unpublished data; Schutz and Winey, 1998). We used a strain that allowed us to screen simultaneously for different MPS1-associated phenotypes (mps1Δ::KanMX cin8Δ::HIS3, supported by counter-selectable plasmids) (see Materials and methods). Counter selection of the plasmid containing CIN8 allowed us to identify MPS1 alleles that might be defective in the spindle checkpoint function of MPS1 (see Materials and methods) (Geiser et al., 1997). These alleles will be described elsewhere. Using temperature sensitivity as a screen for potential SPB duplication defects, we isolated mps1-8 (Fig. 1 A) and showed it is a recessive mutation (unpublished data).

Sequencing revealed that mps1-8 contained multiple mutations (Fig. 1 B); none of the 10 mutations that resulted in amino acid changes individually conferred conditional growth (unpublished data). It is likely that some combination of these mutations is responsible for the mps1-8 phenotype, and all further analysis was performed using the original mps1-8 allele.

We have established previously that the essential MPS1 cellular function is in SPB duplication (Weiss and Winey, 1996). To test if the mps1-8 conditional growth defect might reflect this role, we monitored SPB duplication in mps1-8 cells using immunofluorescence microscopy and EM. Asynchronously growing mps1-8 cells were arrested in G1 using the mating pheromone α-factor, released to the permissive (25°C) or restrictive (36°C) temperature for 3h, and processed for indirect immunofluorescence. The majority of mps1-8 cells complete SPB (shown in green) duplication at 25°C as expected, but they fail to duplicate their SPB at 36°C (92%, n = 42; Fig. 2, A and B). Instead, these cells show a typical S. cerevisiae mitotic arrest state with large buds in which the single unduplicated SPB is associated with cellular DNA (blue) and a focus of microtubules (red) (Fig. 2 B). By immunofluorescence, the SPB duplication defect in the mps1-8 strain is identical to the SPB defect observed in previously characterized conditional MPS1 mutants (Schutz and Winey, 1998).

The morphology of the SPB in mps1-8 cells grown at the restrictive temperature was further examined using EM (see Materials and methods). Serial sections of large budded cells were examined, and a single SPB was found in all cases, verifying that SPB duplication fails in mps1-8 cells grown at the restrictive temperature. As in other MPS1 mutants, the unduplicated SPB in mps1-8 cells is associated with half-bridge material, but unlike them this structure is not elongated in mps1-8 cells (Fig. 2, C–E) (Winey et al., 1991; Schutz and Winey, 1998). It is not yet clear if the subtle difference in the terminal SPB phenotype of the mps1-8 mutant and other MPS1 mutants indicates they are failing at different points in SPB duplication.

As mentioned, the previously characterized conditional MPS1 mutants are defective in both SPB duplication and the spindle checkpoint (Winey et al., 1991; Weiss and Winey, 1996; Schutz and Winey, 1998). We show that the mps1-8 mutant fails in SPB duplication at the restrictive temperature. To determine if mps1-8 cells also fail in the spindle checkpoint, we compared the mps1-8 strain with mps1-1, a strain known to be defective in activating the spindle checkpoint, and mps2-1, a strain able to activate the spindle checkpoint with a monopolar spindle (Hardwick et al., 1996; Weiss and Winey, 1996). Asynchronously growing cultures of mps1-8, mps1-1, and mps2-1 cells were synchronized in G1 with α-factor, released at 25 or 36°C, and samples for flow cytometry and budding indices were taken after 2 and 3 h. At the restrictive temperature, mps1-8 cells exhibit a mitotic arrest similar to that observed for mps2-1 cells; the majority of the cells have a large budded cell morphology, a G2 DNA content (Fig. 3, A and B), and accumulate hyperphosphorylated forms of the spindle checkpoint protein Mad1p, a molecular marker for spindle checkpoint activation (unpublished data; Hardwick et al., 1996). In contrast, the mps1-1 strain fails to arrest in mitosis when grown at the restrictive temperature and instead accumulates cells that appear aploid or aneuploid by flow cytometry (Fig. 3 C). We conclude that mps1-8 is a novel mutant allele leading to defects in SPB duplication but maintaining a functional spindle checkpoint.

We were interested in assessing the level of kinase activity associated with Mps1-8p. To determine this, we compared kinase activity in vitro at the mps1-8 permissive (25°C) and restrictive (36°C) temperatures using protein generated from galactose-inducible glutathione S-transferase (GST) fusion constructs, containing GST alone, GST-MPS1, GST-mps1-1, GST-mps1-KD (kinase dead), and GST-mps1-8 (see Materials and methods). The relative specific activity associated with GST–mps1-8p is similar to that observed for the wild-type protein at either the mps1-8 permissive or restrictive temperature (Fig. 4 A). GST–mps1-1p and GST–mps1-KD were used as controls for MPS1 mutant proteins that have negligible catalytic activity at either assay temperature (Fig. 4 B) (Lauze et al., 1995; Schutz and Winey, 1998). By Western analysis, both GST-Mps1p and GST–Mps1-8p migrate as broad bands slower than the predicted molecular weight of 112 kD (Fig. 4 B) (Lauze et al., 1995; Schutz and Winey, 1998). The band shift observed was shown previously for wild-type protein to be due to autophosphorylation (Fig. 4, A and B). The ability of GST–mps1-8p to phosphorylate an exogenous substrate, myelin basic protein, was also comparable to that observed for the wild-type protein (Fig. 4 B). Therefore, our in vitro results suggest that the SPB duplication defect associated with mps1-8 is not due to a loss of kinase activity itself, but may reflect inappropriate localization or other regulation of Mps1p.

A dosage suppressor screen was performed in a mps1-8 strain to identify genes whose products might interact with Mps1p during SPB duplication. We transformed a mps1-8 strain with a 2-μ URA-based yeast genomic library and screened for transformants that were able to restore growth at 36°C (see Materials and methods). In addition to isolating two different MPS1-containing clones (multiple times), five unique clones were identified as suppressors of the mps1-8 conditional growth defect. In one of these clones (S81), we identified SPC42 as the ORF responsible for mps1-8 suppression (Fig. 5 A). SPC42 mutants exhibit defects in SPB duplication, and Spc42p is a phosphoprotein that localizes to the central plaque of the SPB and duplication intermediates in the satellite and duplication plaque (Donaldson and Kilmartin, 1996; Adams and Kilmartin, 1999). Increased dosage of SPC42 only suppressed the mps1-8 phenotype and not mps1-1 (Fig. 5 B), mps1-737, mps1-412, mps1-1237, or mps1-3796 (unpublished data). This suggests that the mps1-8 defect is distinct from the previously characterized conditional MPS1 mutants.

Two well-characterized SPC42 ts mutants, spc42-10 and spc42-11 (Donaldson and Kilmartin, 1996), exhibit a monopolar phenotype similar to MPS1 mutants when grown at their restrictive temperatures. We generated the double mutant strains, mps1-8 spc42-10 and mps1-8 spc42-11 to look for additional genetic interactions between SPC42 and MPS1 (see Materials and methods). Both double mutant strains show enhanced growth defects when compared with either single mutant strain (unpublished data). The genetic interactions between MPS1 and SPC42 suggest that the gene products may interact in vivo.

The genetic interaction of MPS1 with a bona fide integral SPB component suggested that Mps1p might localize to SPBs, which we initially tested using immunofluorescence microscopy on whole cells. Asynchronously growing cells containing myc epitope-tagged MPS1 at the MPS1 locus (Mps1p-myc) and SPC42 tagged with green fluorescent protein (GFP) (Spc42-GFP) were analyzed by immunofluorescense (see Materials and methods). We observe a strong signal of Mps1p-myc (red) that partially overlaps Spc42p-GFP (green), primarily in unbudded cells, and diffuse nuclear staining (Fig. 6 A). Similar localization was observed with a ProA-tagged Mps1 protein (Schutz et al., 1997; Steiner, 1998). This Mps1p signal is suggestive of SPB and/or kinetochore localization, since yeast kinetochores are adjacent to the SPB during G1 of the cell cycle (Wigge et al., 1998; Wigge and Kilmartin, 2001).

To better characterize Mps1p localization, we determined if Mps1p-myc colocalizes with the kinetochore protein Ndc10p using chromosome spread analysis. A strain containing Mps1p-myc and Ndc10p-HA was grown to midlog phase and samples were harvested for chromosome spread analysis. The diffuse nuclear Mps1p-myc signal we observed by whole cell immunofluorescence appears to overlap with signal from Ndc10p-HA (Fig. 6 B). We also observed Mps1p-myc colocalization with Spc42p-GFP using chromosome spread analysis (Fig. 6 B). In examining chromosome spreads where a single SPB was labeled by Spc42-GFP, and the criterion of ≥90% overlap of Mps1p-myc with Spc42-GFP was considered colocalization, we found that 43% of the spreads (n = 54) exhibited colocalization of these proteins. As a control, we did the same analysis in spreads of strains expressing Spc42-GFP and Ndc10p-HA, and we found only 20% of these spreads showed ≥90% overlap of the two proteins (n = 54), indicating that chromosome spreads can be used to resolve SPBs and kinetochores. Using this technique, Mps1p is found at both organelles.

Finally, we performed immuno-EM on asynchronously growing strains containing Mps1p-myc. Colloidal gold signal overlapping the SPB in the plane of the nuclear envelope suggests that Mps1p localizes to the Spc42p central plaque region (5 examples in 24 cells examined) (Fig. 6 C). Mps1p signal is also detected at the end of microtubules by immuno-EM as is seen for other kinetochore proteins (Wigge et al., 1998; Wigge and Kilmartin, 2001) (18 examples in 24 cells examined) (Fig. 6 D). This dual localization is consistent with a role for Mps1p in SPB duplication and the spindle checkpoint.

The genetic interactions between MPS1 and SPC42 and their colocalization at the SPB prompted us to investigate their physical interaction in the cell. Extracts were prepared from cells containing (a) Mps1p-myc, (b) Ndc1p-myc, as a control for nonspecific interaction with the myc epitope, and (c) no tag. These proteins were immunoprecipitated with anti-myc antibody conjugated to agarose beads and resolved using SDS-PAGE. The presence of Mps1p-myc and Ndc1p-myc was detected using an anti-myc antibody (Fig. 7 B, lanes 5 and 6). Polyclonal anti-Spc42p antibody was used in a duplicate Western analysis and detected a band migrating at the expected molecular weight (46–51 kD) for Spc42p in the Mps1p immunoprecipitate but not the Ndc1p immunoprecipitate (Fig. 7 A, lanes 1 and 2). Indeed, this band migrates at approximately the same position as Spc42p-myc, isolated from a strain overexpressing Spc42p-myc (Fig. 7 A, lane 4). Thus, Mps1p and Spc42p both localize at the SPB and physically interact with each other.

Spc42p is a phosphoprotein whose phosphorylation state varies throughout the cell cycle (Donaldson and Kilmartin, 1996). To determine if Spc42p is a substrate for Mps1p, we used Mps1p-myc and recombinant Spc42p in a kinase assay in vitro. A band corresponding to Spc42p was detected by autoradiography (Fig. 8, A, lane 1) and the presence of this band was dependent on the addition of Mps1p-myc (Fig. 8 A, lane 2). Additionally, we observed autophosphorylation of Mps1p-myc (Fig. 8 A), which has been reported previously (Lauze et al., 1995).

We subsequently assessed whether phosphorylation of Spc42p was dependent on Mps1p function in vivo. We were unable to detect differences in the phosphorylation state of Spc42p-myc isolated from MPS1 and mps1-1 strains using one-dimensional SDS-PAGE (unpublished data). However, differences in Spc42p phosphorylation were revealed using two-dimensional gel electrophoresis (Fig. 8, D–F). For this assay, MPS1 and mps1-1 strains containing SPC42-myc under the control of the galactose promoter were arrested in G1 using α-factor and then released into inducing media at the mps1-1 restrictive temperature for 2 h. Extracts were prepared from these samples and resolved by two-dimensional gel electrophoresis and blotted to PVDF membranes. 11 spots corresponding to Spc42p-myc were identified on the two-dimensional blot prepared using the MPS1 extract (Fig. 8 E). Only six of these spots were detected on the blot prepared using the mps1-1 extract (Fig. 8 D). We believe that the absence of these forms of Spc42p-myc in the mps1-1 strain is due to the inability of Mps1-1p to phosphorylate Spc42p-myc at the mps1-1 restrictive temperature. Indeed, when we treated extract from the MSP1 strain with calf alkaline phosphatase before two-dimensional analysis, the same five spots of Spc42p-myc were absent (Fig. 8 F). The phosphatase-treated sample is missing an additional form of Spc42p-myc detected in the untreated MPS1 extract and shows accumulation of two new forms of Spc42p-myc (Fig. 8 F). Loss of this additional form of Spc42p in the phosphatase-treated sample suggests that another kinase may contribute to Spc42p phosphorylation. The new forms of Spc42p migrate faster and in a more basic position than forms present in the untreated MPS1 sample, suggesting that they are less phosphorylated or unphosphorylated forms of Spc42p (Fig. 8 F, arrow). Interestingly, we do not detect these new forms of Spc42p-myc in the mps1-1 strain. Failure of the phosphatase treatment to produce a single form of Spc42p-myc could be due to incomplete dephosphorylation of Spc42p-myc or to additional posttranslational modification on Spc42p. Phosphorylation of Spc42p by Mps1p in vitro and dependence on Mps1p in vivo suggests that Mps1p may regulate Spc42p during SPB duplication.

Overexpression of Spc42p-myc from an inducible promoter results in the lateral expansion of the central layer of the SPB so that organized layers of Spc42p extend from the SPB in all directions and appear to rest on the nuclear envelope (Donaldson and Kilmartin, 1996; Bullitt et al., 1997; O'Toole et al., 1999). This “super plaque” structure appears by EM to be organized into a two-dimensional crystalline lattice identical to what is observed for Spc42p in the central plaque of a normal SPB (Bullitt et al., 1997). Spc42p isolated from both of these structures is phosphorylated; however, it is not yet clear what role phosphorylation of Spc42p might play in their assembly (Donaldson and Kilmartin, 1996). Here, we analyze formation of the “super plaque” upon SPC42 overexpression in MPS1 mutants. The assembly assay was performed in MPS1 and mps1-1 strains at permissive (25°C) and restrictive temperatures (30 and 34°C) for mps1-1. These strains also contained a myc epitope-tagged version of SPC42 under the control of a galactose-inducible promoter (GAL-SPC42-myc).

Proper formation of the “super plaque” was first analyzed using immunofluorescence. The MPS1 GAL-SPC42-myc and mps1-1 GAL-SPC42-myc strains were grown in noninducing media (no galactose), arrested in G1 using α-factor, and then released into galactose-containing media at both 25 and 30°C for 3 h to induce expression of GAL-SPC42-myc. At the permissive temperature, the mps1-1 (Fig. 9 A) and MPS1 (unpublished data) strains form Spc42p-myc–dependent dome-like structures (super plaque in green) that are associated with DNA (blue) and microtubule (red) signal. The “super plaque” formed in the wild-type strain at 30°C is identical to what we observe for this strain at 25°C (Fig. 9 C). By contrast, in the mps1-1 strain at 30°C we observe a decrease in Spc42p-myc signal at the SPB, and it has a much less organized appearance (Fig. 9 E). To characterize these structures at higher resolution, we analyzed them using EM (see Materials and methods).

Overexpression of SPC42-myc in the wild-type and mps1-1 strain at the permissive temperature results in the formation of the organized dome-shaped “super plaque.” These structures extend laterally from the central plaque of the SPB but appear to be constrained vertically to ∼23 nm as reported previously for the “super plaque” (Fig. 9, B and D) (Bullitt et al., 1997). In contrast, it is clear that Spc42p-myc signal in the mps1-1 strain at the restrictive temperature is an unorganized mass of electron-dense material associated with the half-bridge of the unduplicated SPB (Fig. 9 F). Based on the EM section width (80 nm), we estimate that the structure shown is ∼300 nm × 480 nm × 250 nm (Fig. 9 F). This structure is similar to one formed when spc42-S:A, a mutant allele in which 34 serines have been mutated to alanines, is overexpressed at similar levels (Adams and Kilmartin, 1999). This data suggests that Mps1p function, via phosphorylation of Spc42p, is required for the proper assembly of the central layer of the new SPB during SPB duplication.

We have identified a physical interaction between the integral SPB component Spc42p and Mps1p, a regulator of Spc42p assembly during SPB duplication. We show that Mps1p is capable of phosphorylating Spc42p in vitro and its phosphorylation is dependent on Mps1p in vivo. Additionally, in vivo assembly of Spc42p is compromised in a MPS1 mutant (mps1-1). Mps1-1p has essentially no kinase activity in vitro; therefore, we believe that phosphorylation of Spc42p by Mps1p is important for the assembly of Spc42p into the SPB. We investigated interactions between Mps1p and Spc42p based on our isolation of SPC42 as a dosage suppressor of a unique MPS1 mutant, mps1-8. This conditional mutant was identified in a screen designed to separate the role of MPS1 in SPB duplication from its role in the spindle checkpoint. Using immunofluorescence and EM, we show that mps1-8 mutants fail to duplicate their SPBs at restrictive temperatures. Unlike previously characterized conditional MPS1 mutants, mps1-8 cells are competent to arrest in mitosis through activation of the spindle checkpoint after SPB duplication fails (Schutz and Winey, 1998). Therefore, mps1-8 is an important genetic tool for our further investigation of the role of Mps1p in SPB duplication.

Mutations in the noncatalytic region of MPS1 generated a conditional mutant, mps1-8, that is specifically defective in SPB duplication. We targeted the noncatalytic region of MPS1 for mutagenesis, believing that this region might confer spatial or temporal regulation of Mps1p for each of its roles. In fact, our screen also identified a MPS1 allele that appears to be specifically defective in the spindle checkpoint (to be described elsewhere). In this paper, we show that the molecular defect associated with mps1-8 is likely distinct from our previously characterized conditional MPS1 mutants. Mutations are in the noncatalytic region of mps1-8 and do not affect kinase activity of the mutant protein. By contrast, other conditional MPS1 mutants previously characterized in our lab have mutations in the catalytic region of the gene that affect kinase activity measured in vitro (Schutz and Winey, 1998).

Consistent with the notion that the mps1-8 defect is distinct from other conditional MPS1 mutants, mps1-8 shows unique genetic interactions. We show that mps1-8 but not the other conditional MPS1 mutants is suppressed by an increased dosage of the Mps1p in vitro substrate, Spc42p. By contrast, several of the MPS1 kinase mutants but not mps1-8 are suppressed by an increased dosage of the molecular chaperone CDC37 (Schutz et al., 1997). Although it is likely that CDC37 suppresses through stabilization of the jeopardized Mps1p kinase, it is less clear how SPC42 suppression occurs. We propose that suppression by SPC42 may be through stabilizing an interaction between Spc42p and Mps1-8p. Alternatively, extra Spc42p may help localize Mps1-8p to the SPB or another site of action required for SPB duplication. Future localization studies and coimmunoprecipitation experiments using Mps1-8p should address these questions.

This and previous analyses suggest that Mps1p is required for multiple steps in SPB duplication (Winey et al., 1991; Schutz and Winey, 1998). Here, we show that mps1-8 mutant cells fail very early in SPB duplication when grown at their restrictive temperature. By EM analysis, the unduplicated SPB in mps1-8 cells is similar to those observed in cdc31 and kar1 mutants (Baum et al., 1986; Rose et al., 1986). The unduplicated SPB of mps1-8 does not have an extended half-bridge that is characteristic of several other MPS1 mutants (Winey et al., 1991; Schutz and Winey, 1998). This suggests that there are two distinct requirements for Mps1p early in SPB duplication. Alternatively, the functions defined by mps1-8 and the other conditional MPS1 mutants are not mutually exclusive; proper spatial regulation of Mps1p and Mps1p kinase activity may both be required for its early SPB function. Presumably, the MPS1 kinase domain mutants proceed further in SPB duplication because the mutant proteins localize properly or make appropriate physical interactions, though they are unable to function; the extended half-bridge might be a result of failed attempts to initiate SPB duplication.

MPS1 also functions late in SPB duplication. mps1-737 mutants assemble the central and outer layers of the new SPB but like ndc1, mps2, and bbp1 mutants fail to insert this structure into the nuclear envelope (Winey et al., 1991, 1993; Schramm et al., 2000). Whereas these proteins may function to insert the duplication plaque into the nuclear envelope by providing an opening in the nuclear envelope, it is unclear how Mps1p is involved in this process (Chial et al., 1998; Munoz-Centeno et al., 1999; Schramm et al., 2000). One possibility is that Mps1p facilitates insertion of the duplication plaque into the nuclear envelope by controlling assembly of the inner plaque, a structure not formed in the mps1-737 mutant. Two of the six proteins that make up the inner plaque, Spc110p and Spc98p, are substrates for Mps1p in vitro, and their phosphorylation is dependent on Mps1p in vivo (Pereira et al., 1998; Friedman et al., 2001). The mps1-737 mutant phenotype might reflect a disrupted interaction with Spc110p and Spc98p or the inability of Mps1-737p to phosphorylate Spc110p and Spc98p. The multiple MPS1 mutant phenotypes are consistent with Mps1p, playing a regulatory role in SPB duplication.

We investigated the role of Mps1p as a regulator of SPB assembly using a Spc42p in vivo assembly assay. Overexpression of Spc42p at very high levels causes the central plaque of the SPB (normal site of Spc42p localization) to extend laterally and form a structure we call the “super plaque.” We believe that Mps1p might function in assembly of the central plaque during SPB duplication based on MPS1 mutant phenotypes and because Mps1p physically interacts with and colocalizes with Spc42p in vivo and regulates Spc42p phosphorylation. When Spc42p was overexpressed in a mps1-1 mutant at restrictive temperatures, we observed a defect in assembly of the “super plaque.” This structure is not vertically confined to the two-dimensional crystalline lattice typically seen in MPS1 cells overexpressing SPC42 (Donaldson and Kilmartin, 1996; Adams and Kilmartin, 1999). A similar structure is detected when spc42-S:A, a mutant allele in which 34 serines have been mutated to alanines, is overexpressed at similar levels (Adams and Kilmartin, 1999). Since spc42-S:A is functional at endogenous levels, we believe this common overexpression phenotype reveals the importance of phosphorylation for proper plaque assembly that apparently results in different phenotypes when some phosphorylation is lost (the spc42-S:A allele with threonines intact) versus when most phosphorylation is lost (mutations in the kinase) (Adams and Kilmartin, 1999). In fact, Mps1-1p is severely compromised for kinase activity, and the mps1-1 mutant accumulates fewer phosphorylated forms of Spc42p. We suspect that the phenotype we observe results from the inability of Mps1-1p to properly phosphorylate Spc42p. Failure of Mps1-1p to phosphorylate Spc42p during SPB duplication should then prevent assembly of the central plaque of the new SPB; consistent with this prediction, mps1-1 cells do not show accumulation of satellite material at the half-bridge (Winey et al., 1991).

Is Mps1p function required for assembly of other satellite components? We suspect that Mps1p is required indirectly for assembly of Nud1p, Cnm67p, and Spc29p; Mps1p is required for assembly of Spc42p that may serve as a scaffold upon which the other satellite components assemble. This is supported by the unique suppression of mps1-8 by increased dosage of SPC42. NUD1, CNM67, and SPC29 were not identified in our screen as dosage suppressors of mps1-8 or earlier as dosage suppressors of the mps1-1 (Schutz et al., 1997). We tested SPC29 directly and verified that increased dosage of SPC29 does not suppress mps1-8 (unpublished data). This does not preclude that the function of Mps1p is required for assembly of the other satellite components; the development of assembly assays for Nud1p, Cnm67p, and Spc29p will allow for investigation of the requirement for Mps1p in assembling these components. Regardless of other requirements for Mps1p in SPB duplication, the specific interaction established between Mps1p and Spc42p in this analysis indicates that their association is essential for assembly of the central plaque, an event that may be critical for launching SPB duplication.

Yeast strains used in this study are listed in Table I and are derived from S288c and W303 backgrounds. Yeast and E. coli genetic techniques were as described by Ausubel et al. (1997). Yeast synchronization and 5-fluoroorotic acid screening were done as in Jones et al. (1999). Galactose (GAL) induction was done as in Hardwick et al. (1996).

GST fusion proteins were induced and isolated and kinase assays were performed as in Schutz and Winey (1998). Mps1p-13Xmyc (Ndc1p-3Xmyc isolated similarly, HC12–2b) was isolated from an asynchronously growing strain (SBY650) using techniques similar to those described in Schutz and Winey (1998). Mps1p-13Xmyc and recombinant Spc42p expressed in baculovirus-infected insect cells (a gift from Danni Vinh, Trisha Davis Lab, University of Washington, WA) were used in kinase assays as described in Lauze et al. (1995). Samples for all kinase assays were resolved on 15% Anderson SDS-PAGE gels (Anderson et al., 1973). Gels were subjected to electrophoretic transfer onto nitrocellulose membrane (Schleicher and Schuell Inc.) and membranes were used to expose a phosphorimager screen and for Western analysis to quantitate protein levels. ³²P incorporation was quantitated using a Storm 860 phosphorImager with the ImageQuaNT analysis package (Molecular Dynamics). For Western analysis, nitrocellulose membranes were blocked using Odyssey blocker. Detection was performed using secondary antibodies labeled with the near-infrared fluorescent dyes IRDye800 and Cy5.5, and blots were scanned with an Odyssey Infrared Imager (Odyssey reagents, instruments, and software from LI-COR Biosciences). GST-Mps1 proteins were detected with goat antiglutathione primary antibody (1:1,000; Amersham Pharmacia Biotech) and an anti–goat antibody labeled with the IRDye800 fluorophore (1:10,000). Mps1p-13Xmyc was detected using an anti-myc primary antibody (1:1,000; Santa Cruz Biotechnology, Inc.) and an anti–mouse antibody labeled with IRDye800 (1:10,000). Spc42p was detected using a polyclonal anti-Spc42p primary antibody (a gift from John Kilmartin, Medical Research Council Laboratory of Molecular Biology, London, UK) and an anti–rabbit antibody labeled with Cy5.5 (1:10,000). Protein was visualized and quantitated with the Odyssey Imaging System.

We determined relative specific activity for each fusion protein by using the Storm860 phosphorImager and ImageQuaNT analysis package to measure ³²P incorporation and by using the Odyssey Infrared Imager and analysis software to measure protein amounts. The number representing ³²P incorporation was divided by the number representing protein amount to yield relative specific activity.

In the coimmunoprecipitate experiment, the immunoprecipitated material was resolved on an 8.5% laemmli SDS-PAGE gel (Ausubel et al., 1997). Gels were subjected to electrophoretic transfer onto a polyvinylidene difluoride membrane (Millipore). Polyvinylidene difluoride membranes were blocked as in Chial et al. (1998). Mps1p-13Xmyc and Ndc1p-3Xmyc were detected using an anti-myc primary antibody (1:1,000; Santa Cruz Biotechnology, Inc.) and a sheep anti–mouse antibody conjugated to HRP (1:10,000; Sigma-Aldrich). Spc42p was detected using the anti-Spc42p polyclonal primary antibody (a gift from John Kilmartin, Medical Research Council Laboratory of Molecular Biology, London, UK) and a donkey anti–rabbit antibody conjugated to HRP (1:10,000; Sigma-Aldrich).

Alkaline phosphatase treatment of samples and two-dimensional gel electrophoresis were performed by Kendrick Labs, Inc. as described by Donaldson and Kilmartin (1996) and O'Farrell (1975), respectively. Isoelectric focusing was performed in glass tubes using 2% pH 4–8 ampholines (BDH ampholines Gallard Schlesinger) for 9,600 V/h. The tube gel was sealed to the top of a stacking gel on top of a 10% acrylamide slab gel. Samples were treated with 10 U of calf intestinal alkaline phosphatase (New England Biolabs, Inc.) at 30°C for 30 min.

Flow cytometric analysis of cells was performed as described using the DNA stain propidium iodide (Sigma-Aldrich) (Hutter and Eipel, 1979). Samples were analyzed on a Becton Dickinson FACScan^® flow cytometer using CELL QUEST software (Becton Dickinson).

Indirect immunofluorescence was performed as described in Chial et al. (1998), and chromosome spreads were performed as in Biggins et al. (1999). Tubulin was visualized using rat anti–α-tubulin primary antibody, YOL1/34 (1:150; Accurate Chemical), and anti–rat antibody conjugated to Texas red (1:400; Scientific). An affinity purified primary polyclonal anti-myc antibody (1:1,000; a gift from the Don W. Cleveland lab, University of California, San Diego, La Jolla, CA) and a donkey anti–rabbit Cy3-conjugated antibody (1:1,000; Scientific) were used to detect Mps1p-myc. In some experiments, a mouse anti-GFP primary antibody (1:40; a gift from the Pat O'Farrell lab, University of California, San Francisco, CA) and an anti–mouse FITC-conjugated antibody (1:800; Scientific) were used to detect Spc42-GFP. A monoclonal mouse anti-myc primary antibody (1:450) (9E10; Santa Cruz Biotechnology, Inc.) and anti–mouse FITC-conjugated antibody (1:800) (Jackson ImmunoResearch Laboratories) were used to detect Spc42p-myc. In chromosome spreads, a mouse anti-HA primary antibody (1:500) was used with an anti–mouse FITC-conjugated antibody (1:800). Fluorescent microscopy was performed using a Leica DMRXA/RF4/V automated microscope with a Cooke SensiCam digital camera and Slidebook software (Intelligent Imaging Innovations).

Cells for EM were prepared for thin sectioning by high pressure freezing and freeze substitution (Winey et al., 1995) or by chemical fixation (Byers and Goetsch, 1975). Serial thin sections were viewed on a Philips CM10 electron microscope (Philips Electronic Instruments), and images were captured on film or with a Gatan digital camera and viewed with the Digital Micrograph Software package (Gatan Inc.). Immuno-EM was performed using high pressure frozen and freeze-substituted cells as described by Giddings et al. (2001). Myc-tagged Mps1p was detected with polyclonal anti-myc antibody described earlier and 10 nM colloidal gold-conjugated secondary antibodies.

Primers MPS1AC3 and MPS1AC4 were used to amplify the NH₂-terminus of MPS1 (∼2.2kb) under mutagenic PCR conditions: dATP = 0.1 mM, MnCl₂ = 0.5 mM, and MgCl₂ = 1.5 mM. 25, 20-μl PCR reactions were pooled, cut with EcoRI and BamHI, and ligated to EcoRI- and BamHI-digested pRS314-MPS1 to replace the wild-type NH₂ terminus. This ligation was used to transform E. coli, and transformants (805) were collected, grown for 2 h in Luria broth at 37°C, and prepared for DNA to make the MPS1 mutagenized library.

We screened the MPS1 mutagenized library in a strain (ACY17-16A) (Table I), mps1Δ::KanMX cin8Δ::HIS3 pRS316-pac8-1, pLEU2-CIN8-cyh^s-CEN. pRS316-pac8-1 (pac8-1 is an allele of MPS1) was maintained because MPS1 is essential, and pLEU-CIN8-cyh^s-CEN was maintained because of the lethal interaction between pac8-1 and cin8Δ::HIS3 (Geiser et al., 1997). mps1Δ::KanMX cin8Δ::HIS3 pRS316-pac8-1 plus pLEU-CIN8-cyh^s-CEN was transformed with the CEN-TR–based MPS1 mutagenized library. 56% (17,920) of the TRP⁺ transformants replica plated to 5-fluoroorotic acid were viable. These mps1 alleles were tested for conditional growth at 36°C, benomyl (10 μg/ml; DuPont) sensitivity, and cycloheximide (5 μg/ml; Sigma-Aldrich) sensitivity. We isolated mps1-8 as a conditional mutation (benomyl^R and cycloheximide^R). We sequenced the mutagenized region of mps1-8 using primers MPS1AC2, MPS1X, MPS1Y, and MPS1Z. Sequencing was done by the MCD Biology departmental sequencing facility (ABI automated sequencer).

The mps1-8 allele was integrated at either the URA3 or the LEU2 locus. pRS306–mps1-8 was linearized with NcoI to direct integration at the URA3 gene, and pRS305–mps1-8 was linearized with HpaI to direct integration at the LEU2 gene. Proper integration at the URA3 and LEU2 loci was verified by PCR, using a primer internal to MPS1 (MPS1AC7), and one within the URA3 gene (ACURA3) or the LEU2 (ACLEU2B).

We transformed a mps1-8 strain with a 2-μ URA-based yeast genomic library (Connelly and Hieter, 1996) and screened 40,000 URA⁺ transformants for growth at 36°C. 86 transformants grew at 36°C. We determined that 60 transformants exhibited plasmid-dependent growth at 36°C and isolated clones as done in Schutz et al. (1997). 22 clones conferred growth to mps1-8 at 36°C upon retransformation. We used restriction digests to determine the number of unique, nonMPS1, clones. Seven clones (S14, S36, S45, S47, S79, S81, S89) with unique genomic inserts and two different clones each containing MPS1 were isolated multiple times. ORFs contained in these clones were identified as described in Jones et al. (1999). S81 contained three ORFs from chromosome XI, YKL044W, PHD1, and SPC42. SPC42 was subcloned into pRS202, and we transformed the mps1-8 strain with this clone and found that it conferred growth to mps1-8 at 36°C. The other clones identified in this screen will be reported elsewhere.

We thank John Kilmartin for SPC42 strains and antibody, Sue Biggins for the myc epitope-tagged MPS1 strain and for her support, the Pat O'Farrell lab for anti-GFP antibody, Danni Vinh and Trisha Davis for Spc42p, the Don Cleveland laboratory for affinity purified anti-myc antibody, and LiCor Inc. for reagents and technical assistance. We are grateful to Corine Lau, Susan McBratney, Michele Jones, and Sue Jaspersen for critical reading of this article, Tom Giddings for advice on EM, and all Winey lab members for helpful discussions. A.R. Castillo was supported by an Individual National Research Service Award (National Institutes of Health grant GM19566 to A.R. Castillo). This work was also supported by the National Institute of Health (GM51312 to M. Winey).