The Sad1-UNC-84 homology domain in Mps3 interacts with Mps2 to connect the spindle pole body with the nuclear envelope

Centrosomes are the primary microtubule-organizing center in most eukaryotic cells. Centrosomes are considered to be soluble cytoplasmic organelles that lack membrane-associated structures. However, centrosomes are closely associated with the nuclear envelope in most interphase cells. A variety of data suggests that a biochemical link exists between the centrosome and nuclear membrane (Bornens, 1977; Nadezhdina et al., 1979; Kuriyama and Borisy, 1981; Malone et al., 2003), and tethering of the centrosome to the nuclear envelope might be important for rapid formation of the mitotic spindle, nuclear positioning within the cell, and/or centrosome inheritance.

SUN (Sad1-UNC-84 homology) domain–containing proteins are an emerging family of inner nuclear envelope proteins that are excellent candidates to play a role in connecting centrosomes with the nuclear envelope. Originally identified based on an ∼150-amino-acid region of homology between the C terminus of the Schizosaccharomyces pombe Sad1 protein and the Caenorhabditis elegans UNC-84 protein (Hagan and Yanagida, 1995; Malone et al., 1999), SUN proteins are present in the proteomes of most eukaryotes (Fig. S2; Mans et al., 2004). In addition to the SUN domain, these proteins contain a transmembrane sequence and at least one coiled-coil domain and localize to the inner nuclear envelope. SUN proteins are anchored in the inner nuclear envelope by their transmembrane segment and oriented in the membrane such that the C-terminal SUN domain is located in the space between the inner and outer nuclear membranes. Here, the SUN domain can interact with the C-terminal tail of an outer nuclear envelope protein that binds to the cytoskeleton, including the centrosome (for review see Starr and Fischer, 2005).

The spindle pole body (SPB) is the Saccharomyces cerevisiae centrosome-equivalent organelle and is the sole site of microtubule nucleation in budding yeast. The SPB is a multilayered cylindrical structure that is embedded in the nuclear envelope throughout the yeast life cycle (for review see Jaspersen and Winey, 2004). The soluble SPB core consists of three primary layers called the outer, inner, and central plaques. Cytoplasmic and nuclear microtubules are nucleated from the outer and inner plaques, respectively, whereas the central layer of the SPB plays an important role in tethering the organelle to the nuclear envelope. Associated with one side of the core SPB is an electron-dense region of the nuclear envelope termed the half-bridge, which is important for SPB duplication as well as for microtubule nucleation during G1 (Byers and Goetsch, 1974, 1975). SPBs in fission yeast have a similar but not identical structure. Importantly, the S. pombe SPB is also embedded in the nuclear envelope, possibly by its SUN protein, Sad1 (Hagan and Yanagida, 1995). Until now, the budding yeast orthologue of Sad1 had not been identified.

The half-bridge is critical for SPB duplication; yet, details of its structure and function at a molecular level are only beginning to emerge. Four proteins are found at the half-bridge: Cdc31, Kar1, Mps3, and Sfi1. Kar1 and Mps3 are integral membrane proteins that localize to the cytoplasmic and nuclear sides of the half-bridge, respectively (Spang et al., 1995; Jaspersen et al., 2002; Nishikawa et al., 2003), whereas Cdc31 and Sfi1 are soluble half-bridge components (Spang et al., 1993; Kilmartin, 2003). Recently, multiple Cdc31 proteins were shown to associate with Sfi1 to form a soluble cytoplasmic filament that spans the length of the half-bridge, and a model in which duplication of the Cdc31-Sfi1 filament generated a non-SPB associated end to initiate assembly of a new SPB was proposed (Li et al., 2006). However, it is unclear how Mps3 and Kar1 associate with this filament and what roles both membrane proteins play during SPB duplication and assembly. One possibility is that Mps3 and Kar1 form the physical half-bridge because mutations in KAR1 and MPS3 cause cells to arrest with unduplicated SPBs that lack any recognizable half-bridge structure (Vallen et al., 1994; Jaspersen et al., 2002). Interestingly, Cdc31 and Sfi1 may play a role in the insertion step later in SPB duplication, although this function is poorly understood (Vallen et al., 1994; Kilmartin, 2003). The insertion step of SPB duplication also requires the membrane proteins Mps2 and Ndc1 and their respective binding partners Bbp1 and Nbp1 (Winey et al., 1991, 1993; Schramm et al., 2000; Araki et al., 2006). These proteins may be recruited to the half-bridge to facilitate insertion of the newly formed SPB into the nuclear envelope and/or to tether the half-bridge to the core SPB. Direct binding between the half-bridge and membrane components of the SPB has not been demonstrated, but genetic and two-hybrid interactions between Bbp1 and Kar1 suggest that the half-bridge is connected to the core SPB through these two proteins (Schramm et al., 2000).

Not only is the half-bridge important for SPB duplication, but it is also essential for nuclear migration and fusion after mating (karyogamy; for review see Rose, 1996). During G1, cytoplasmic microtubules are nucleated from the half-bridge instead of from the outer plaque (Byers and Goetsch, 1975). After mating, half-bridge microtubules interdigitate, allowing the two nuclei to congress, and the juxtaposed half-bridges form the site where SPB and nuclear membrane fusion originates (Byers and Goetsch, 1974, 1975). The importance of the SPB half-bridge for nuclear migration and fusion is illustrated by the fact that mutations in KAR1, MPS3, and CDC31 cause defects in both steps of karyogamy (Rose and Fink, 1987; Vallen et al., 1994; Nishikawa et al., 2003).

In the present work, we show that budding yeast half-bridge protein Mps3 is homologous to the S. pombe SPB component and SUN protein Sad1 (Hagan and Yanagida, 1995). Mps3 also contains a SUN domain, and mutational analysis of this region of Mps3 demonstrated that it is critical for Mps3 function during SPB duplication and karyogamy. The SUN domain of Mps3 binds to the C terminus of another integral membrane component of the SPB, Mps2, and the Mps2–Mps3 interaction is required for formation of an intact SPB. Our results demonstrate at a molecular level how the half-bridge is tethered to the core SPB and describe the consequences of disrupting this interaction. They also support the novel function of SUN proteins in bridging the inner and outer nuclear envelope and provide further evidence that SUN proteins function to tether centrosomes to the nuclear envelope.

MPS3 encodes a 682-amino-acid protein with an acidic N terminus, a transmembrane segment, two coiled-coil domains, and a poly-glutamine region (Jaspersen et al., 2002; Nishikawa et al., 2003). Analysis of secondary structure and local compositional complexity also suggests that the C-terminal region of Mps3 (amino acids 436–682) folds into a discrete globular domain, which has been shown by deletion analysis to be important for Mps3 function (Tables S2 and S3; Nishikawa et al., 2003).

To learn more about the function of the C-terminal domain, we searched for similar sequences in protein databases using PSI-BLAST with the model inclusion cutoff set at 0.05 (Altschul et al., 1997). After the first round of search, only Mps3 orthologues from other budding yeasts were detected; however, after the second iteration, we identified numerous proteins from various eukaryotes that showed homology to the Mps3 C terminus. The best matches were to the C-terminal SUN domain of the S. pombe Sad1 and to other SUN domain–containing proteins. Alignment of the C terminus of Mps3 with the SUN domain from Sad1 and other SUN proteins indicates a good fit between Mps3 and the sequence consensus of the SUN domain family (Fig. 1 A).

The 514-amino-acid Sad1 protein shares a common overall architecture with Mps3 (Fig. 1 B) and localizes to the SPB (Hagan and Yanagida, 1995), raising the possibility that Mps3 and Sad1 perform similar cellular roles. We overexpressed SAD1 in various budding yeast temperature-sensitive (ts) mutants in mps3 and found that growth at the nonpermissive temperature was partially restored in mps3-ΔSUN2 cells (see the following section; Fig. 1 C), indicating that Sad1 and Mps3 are functionally related.

Mutation of the highly conserved asparagine to lysine at position 597 of MPS3 (Fig. 1 A; N597K) suggests that the SUN domain is important for Mps3 function during mitosis and mating (Nishikawa et al., 2003). To more thoroughly analyze the role of the Mps3 SUN domain, we constructed an allelic series within the Mps3 SUN domain by mutating residues that are conserved or similar in most vertebrate SUN domains (Fig. 1 A) or are conserved in at least four out of the six sequenced Saccharomyces species (Cliften et al., 2003). We found that many of the conserved residues are essential for Mps3 function because their mutation results in a nonviable allele (Tables S2 and S3). However, mutation of some conserved residues, such as serine 516, proline 518, and tyrosine 563, had no effect on cell viability. In our analysis, we isolated a series of five new ts MPS3 alleles (mps3-W487A, -Y502H, -A540D, -Q572LQ573L, and -F592S; Fig. 1 A and Tables S2 and S3). We also found that loss of the second half of the SUN domain (mps3-ΔSUN2; amino acids 524–645) or mutation of the conserved tryptophan at position 477 (mps3-W477A) resulted in ts alleles if multiple copies of the mutant gene were present in the cell (Table S3).

We suspected that the SUN domain was critical for Mps3 function during SPB duplication because several of our mutants spontaneously diploidize (Fig. 2, A and B; and Fig. S1), a phenotype shared by mps3-1 and some kar1 and cdc31 mutants that are defective in SPB duplication. To test whether our Mps3 SUN domain mutants show defects in SPB duplication, asynchronously growing wild-type and mutant cells were analyzed after a 4-h shift to the nonpermissive temperature (37°C). Flow cytometric analysis of DNA content and budding index confirmed that all of the SUN domain mutants arrest in mitosis at 37°C. Analysis of mitotic spindle morphology by indirect immunofluorescence microscopy and EM revealed that >75% of large-budded mps3-W487A, -Y502H, -A540D, and -Q572L Q573L cells contain a monopolar spindle at 37°C (Fig. 2, A and B; and Fig. S1). These results suggest that the SUN domain in Mps3 is required for SPB duplication.

In mps3-W477A, -F592S, and -ΔSUN2 mutants, we also observed monopolar spindles in roughly half of the large-budded cells by both indirect immunofluorescence microscopy and EM (Fig. 2, A and B; and Fig. S1). The remaining cells contained short, bipolar spindles with a single DNA mass indicative of a metaphase arrest. This heterogeneous arrest phenotype supports the possibility that Mps3 has a second function in mitotic progression separate from its role in SPB duplication that is defective in these mutants (see Discussion).

Like mps3-1 mutants, we observed that the growth defect of some of our mps3 SUN mutants was partially rescued by overexpression of components of the half-bridge, including CDC31 (Fig. 2 C; Jaspersen et al., 2002). The fact that these mutants also displayed defects in Cdc31 localization at 37°C suggests that their failure in SPB duplication is due to an inability to bind to Cdc31 at the half-bridge (unpublished data). Interestingly, the growth defect of all of the SUN domain mutants at 37°C was suppressed at least in part by overexpression of NBP1 or MPS2 (Fig. 2 C). One interpretation of this genetic interaction is that the SUN region of Mps3 might interact directly with Nbp1 or Mps2; therefore, having more of either Mps3 binding partner present in the cell could stabilize a weak interaction with the mps3 SUN mutant protein and allow the mps3 cells to proliferate at the restrictive temperature. Defects in Cdc31, Mps2, and/or Nbp1 binding could also explain why localization of most of mps3 SUN domain mutant proteins to the SPB is reduced or eliminated in cells grown at 37°C (Fig. 2 E), even though total protein levels determined by Western blotting are not dramatically altered (Fig. 2 D).

Mps3 is important not only for SPB duplication but also for karyogamy. Analysis of the nep98-7 allele, which contains a mutation in one of the most highly conserved residues in the SUN domain, indicated that Mps3 is required for both steps of karyogamy: nuclear congression and fusion (Nishikawa et al., 2003). To test whether our mutants displayed karyogamy defects, we analyzed the position of nuclei in mating mixtures of wild-type and mps3 SUN domain mutants after 5 h at the semipermissive temperature of 30°C; at higher temperatures, karyogamy itself is affected and our mutants begin to exhibit a significant delay in mitosis, a cell cycle stage that is nonpermissive for mating. After 5 h, 95% of zygotes contained a single nucleus in the wild-type mating, whereas only 60–70% of mps3 SUN domain mutant zygotes had successfully completed both steps of karyogamy and contained a single nucleus (Fig. 3). The SUN domain mutants displayed a unilateral defect in karyogamy (unpublished data) and were defective in both nuclear congression and fusion, although the congression defect was more pronounced. We suspect that allele and strain-background differences are responsible for the less severe karyogamy phenotypes that we observed compared with others (Nishikawa et al., 2003). The simplest interpretation of the karyogamy defect, like the SPB duplication defect, is that mps3 SUN domain mutants fail to form an intact half-bridge, which is critical for both nuclear congression and fusion (Byers and Goetsch, 1975).

Because Mps3 is located at the SPB and SUN proteins are thought to interact with other integral membrane proteins, we hypothesized that the Mps3 SUN domain might bind Kar1 or Mps2, the other single pass membrane proteins at the SPB (Vallen et al., 1992; Spang et al., 1995; Munoz-Centeno et al., 1999). Full-length Kar1 and Mps2 expressed as maltose binding protein (MBP) fusions in bacteria were transferred to nitrocellulose membranes after SDS-PAGE of bacterial extracts, and the membrane was probed with the Mps3 SUN domain (amino acids 457–617) labeled with the infrared dye Alexa 680 in a gel overlay assay. We found that the SUN domain bound to full-length Mps2 (MBP-Mps2) and to the Mps2 C terminus (MBP-mps3-Ct) but did not interact with MBP, Kar1 (MBP-Kar1), or other proteins in the bacterial extracts (Fig. 4 A). Binding of the Mps3 SUN domain to Mps2 required a functional SUN domain because most mps3 SUN domain mutant proteins were unable to compete with the wild-type protein for Mps2 binding in the gel overlay assay, even when present in at least 10-fold excess (Fig. 4 B and not depicted).

To test whether Mps2 and Mps3 interact in yeast, we created strains expressing MPS2-13xMYC and/or MPS3-3xFLAG from their endogenous promoters and immunoprecipitated the tagged proteins from the membrane fraction of lysates from spheroplasted cells treated with the short 12-Å membrane-permeable cross-linker dithiobis(succinimidyl)propionate. Mps2-13xMYC coimmunoprecipitated with Mps3-3xFLAG, as did the Mps3 binding protein Cdc31 (Fig. 4 C; Jaspersen et al., 2002). The fact that Mps2-13xMYC did not precipitate in cells lacking Mps3-3xFLAG and that other SPB components such as Spc29 did not coprecipitate indicates that binding is specific and likely occurs through Mps3. Mps2 and Mps3 also coimmunoprecipitated in the absence of a cross-linker when cell lysates were prepared by grinding in liquid nitrogen (Fig. 4 D), providing further evidence of a direct interaction between Mps2 and Mps3 in yeast. However, as SPBs remain largely intact by this method of lysis (Niepel et al., 2005), we cannot totally exclude the possibility that the Mps2–Mps3 interaction is mediated by another SPB component. From the binding data in vitro and our ability to coimmunoprecipitate Mps2 and Mps3 from yeast extracts using two different techniques, we conclude that Mps2 binds to Mps3 in vivo and that this interaction is likely mediated by the Mps3 SUN domain.

Several proteins predicted to bind to SUN domains contain a conserved C-terminal transmembrane segment followed by ∼30 amino acids known as the Klarsicht/ANC-1/Syne homology (KASH) domain (Starr and Fischer, 2005). Protease protection and protein binding experiments demonstrated that the N terminus of Mps3 is exposed to the cytoplasm/nucleoplasm, whereas the C terminus, including the SUN domain, is located in the space between the inner and outer nuclear membranes (Nishikawa et al., 2003). In contrast, the N terminus of Mps2 is exposed to the cytoplasm/nucleoplasm, whereas the C terminus of Mps2 is in the intermembrane space (Schramm et al., 2000). Therefore, we would predict that the Mps3 SUN domain binds to the C terminus of Mps2. However, neither the Mps2 C terminus nor any other protein sequence in S. cerevisiae contain regions of significant similarity to a hidden Markov model of the KASH domain (unpublished data); instead, Mps2 and its orthologues from other budding yeasts have a unique C-terminal region that includes seven conserved aromatic residues (Fig. 5 A).

Despite its lack of a KASH domain, the Mps2 C terminus clearly binds to the Mps3 SUN domain in both the gel overlay assay and pull-down assays in vitro (Fig. 4 A and not depicted). In addition, a version of Mps2 lacking the C-terminal amino acids (mps2ΔC; amino acids 1–328) does not bind to the SUN domain in vitro (Fig. 4 A and Fig. 5 D) and mps2ΔC is unable to rescue the growth defect of mps2Δ and serve as the sole copy of MPS2 in cells (Fig. 5 B). Thus, the C terminus of Mps2, like the Mps3 SUN domain, is critical for cell viability.

Analysis of a series of point mutations (not depicted) and truncations (Fig. 5 C) in the Mps2 C terminus indicates that residues throughout the region are essential for Mps2 function and for Mps3 binding. Insertion of a nonsense codon after residue 381 in Mps2 resulted in an allele of MPS2 (mps2-381) that encodes a protein containing all but the last six amino acids. The fact that mps2-381 cells display a conditional growth defect (Fig. 5, C and E), combined with the fact that the mps2-381 protein only weakly bound to the SUN domain in vitro (Fig. 5 D), is evidence that the C terminus of Mps2 is important for the interaction with Mps3 in vivo. Additional support for the notion that the growth defect in mps2-381 cells is due to loss of Mps3 binding comes from our observations that overexpression of wild-type MPS3, but not SUN domain mutants, partially rescues the growth defect of mps2-381 mutants at 37°C (Fig. 5, E and F). Interestingly, MPS3 overexpression is unable to suppress mps2 mutants that contain lesions in the N terminus (not depicted), whereas BBP1 overexpression, which suppresses some N-terminal mps2 mutants (Schramm et al., 2000), only weakly rescues growth of mps2-381 (Fig. 5 F). These results strongly suggest that the SUN domain of Mps3 binds to the C terminus of Mps2 and supports the possibility that loss of Mps3 binding is responsible for the growth defect of mps2-381 and other C-terminal Mps2-deletion mutants.

Asynchronously growing wild-type and mps2-381 mutant cells were analyzed after a 4-h shift to the nonpermissive temperature (37°C). Flow cytometric analysis of DNA content and budding index confirmed that the mps2-381 mutants spontaneously diploidize at 23°C and arrest in mitosis at 37°C (Fig. 6 A). Analysis of mitotic spindle morphology by indirect immunofluorescence microscopy and EM revealed that >75% of large-budded mps2-381 cells contain a monopolar spindle at 37°C (Fig. 6, B and C). Other mps2 mutants also have monopolar spindles (Winey et al., 1991); however, mps2-381 mutants have a distinct terminal SPB mutant phenotype: they contain a single monopolar spindle on a deep nuclear invagination, and the SPB appears to have failed to initiate SPB duplication (Fig. 6 C). The morphology of the mps2-381 SPB is highly reminiscent of many of our mps3 SUN mutants (Fig. 2 A and Fig. S1), and it appears that mps2-381 SPBs lack a normal half-bridge, although visualization of the half-bridge by EM alone on such a constricted invagination is difficult. We found no evidence by morphology or by immunolabeling of a second partially assembled SPB like that observed in mps2-1 mutants in spite of good overall fixation (unpublished data; Winey et al., 1991). This novel terminal SPB phenotype uncovered by the mps2-381 allele suggests that Mps2 has an earlier function in SPB duplication, perhaps during half-bridge formation or tethering of the half-bridge to the SPB, which was not previously indicated through analysis of other mps2 mutants.

Like the Mps3 SUN domain, the Mps2 C terminus is important not only for SPB duplication but also for both steps of karyogamy. Although other mps2-1 alleles also have a mild karyogamy defect, >30% of mps2-381 cells fail to complete both nuclear congression and fusion, indicating that Mps2–Mps3 binding is important for karyogamy (Fig. 3).

The fact that mps2-381 mutants have phenotypes very similar to our mps3 SUN mutants and to other mutants that lack a functional half-bridge led us to examine the localization and levels of Mps3 and the half-bridge component Cdc31 in mps2-381 cells. Asynchronously growing wild-type and mps2-381 cells that have the endogenous copy of MPS3 fused to GFP were shifted to 37°C for 3 h. Western blot analysis revealed that Mps3-GFP, as well as other SPB components such as Cdc31, Spc110, and Tub4, were expressed at similar levels in MPS2 and mps2-381 cells (Fig. 7 A). However, we consistently saw a decrease in the number of mps2-381 cells grown at 37°C containing Mps3-GFP, as well as a decrease in the overall intensity of Mps3-GFP epifluorescence at the SPB (Fig. 7, B and C). Cdc31 staining at the SPB by indirect immunofluorescence also decreased in mps2-381 mutants at 37°C (Fig. 7, B and C). Instead of a distinct SPB signal, Mps3-GFP was diffusely localized and Cdc31 appeared as punctate spots in mps2-381 cells at 37°C. No change in expression or localization of core SPB components such as Spc110 and Tub4 was observed in mps2-381 mutants at either temperature (Fig. 6 B and Fig. 7 C). The failure of mps2-381 mutants to localize Mps3-GFP and Cdc31 to the SPB is not due to defects in 9xMYC-mps2-381 expression or its localization to the SPB; 9xMYC-Mps2 and 9xMYC-mps2-381 both localized to the SPB at 23 and 37°C (Fig. 7 D). Together with our data showing that mps2-381 binding to the Mps3 SUN domain is reduced in vitro, these data strongly suggest that mps2-381 mutants fail to bind and localize components of the half-bridge to the SPB. We propose that the Mps3 SUN domain and the Mps2 C terminus form an intermolecular bridge spanning the nuclear membranes to tether the half-bridge to the core SPB. This interaction is critical for assembly and/or maintenance of an intact half-bridge (Fig. 8).

Association between the core SPB and the half-bridge is critical for SPB duplication and function, yet the molecular details of this link have not been characterized at a biochemical level. In this work, we show that SPBs in Mps3 SUN domain and Mps2 C-terminal mutants lack an associated half-bridge and find that the Mps2 C terminus directly binds to Mps3 in a SUN domain–dependent manner. This strongly suggests that the Mps2–Mps3 interaction is essential to anchoring the half-bridge to the core SPB (Fig. 8). When combined with previous studies showing that the N terminus of Mps2 binds to Bbp1, which in turn interacts with the central plaque component Spc29 (Schramm et al., 2000), our results provide a missing physical connection between the half-bridge and the core SPB.

Previous analysis of mps2-1 mutants showed that Mps2 function is required for insertion of the newly duplicated SPB into the nuclear envelope (Winey et al., 1991). However, our findings that mps2-381 mutants lack any recognizable half-bridge structure and fail to localize Mps3 and Cdc31 to the SPB suggests that Mps2 has an additional function earlier in SPB duplication. Based on analysis of Mps2–Mps3 binding, we propose that Mps3 recruits and binds to Mps2 at the distal tip of the newly assembled bridge during SPB duplication. Once this connection between the C terminus of Mps2 and the Mps3 SUN domain is established, Mps2 can interact through its N terminus with additional proteins such as Bbp1, Nbp1, and Ndc1 to facilitate insertion of the newly duplicated SPB into the nuclear envelope. This model would explain why previously characterized mps2 alleles, which all contain N-terminal mutations, fail in the late step of SPB duplication like bbp1-1, nbp1-td, and ndc1-1 mutants (Winey et al., 1991, 1993; Schramm et al., 2000; Araki et al., 2006). We would also predict that binding between Mps2 and Mps3 is required to tether the half-bridge to the core SPB from the point of SPB duplication throughout the rest of the cell cycle. Consistent with this hypothesis, mps2-381 cells fail to maintain an intact half-bridge at the restrictive temperature, resulting in defects in the earliest step of SPB duplication as well as in karyogamy.

Is the Mps2–Mps3 interaction the only binding event required to tether the half-bridge to the core SPB? Two-hybrid interactions have been observed between Bbp1 and Kar1 (Schramm et al., 2000) as well as between Ndc1 and Kar1 (unpublished data), and numerous genetic interactions have been reported among Mps2, Ndc1, Bbp1, Nbp1, and components of the half-bridge (Schramm et al., 2000; Jaspersen et al., 2002; Araki et al., 2006), so we suspect that a complex set of protein–protein interactions likely anchors the half-bridge to the core SPB. Multiple overlapping interactions would help ensure the integrity of the entire SPB and allow it to withstand microtubule forces that push and pull on the SPB and nucleus during the yeast life cycle. Redundancy in protein–protein interactions holding the half-bridge to the core SPB could also explain why we observe only ∼50% reduction in Mps3 and Cdc31 localization to the SPB in mps2-381 mutants at 37°C. Further analysis of the molecular interactions between the membrane and half-bridge components of the SPB will allow us to better understand how the SPB is assembled and duplicated.

Analysis of Mps3 SUN mutants in yeast suggest that SUN proteins have multiple evolutionarily conserved functions. First, the role of the Mps3 SUN domain in maintaining SPB integrity by tethering the soluble core proteins to the nuclear envelope components suggests the interesting possibility that this protein family is involved in connecting centrosomes to the nuclear envelope. Although the idea that centrosomes are physically linked to the nuclear envelope is controversial, old and new evidence suggests that such an association may exist (Bornens, 1977; Nadezhdina et al., 1979; Kuriyama and Borisy, 1981; Malone et al., 2003). Low sequence conservation between SUN domains makes it impossible to predict the phylogenetic relationships of most proteins within the SUN family with statistical significance. However, our functional data suggests that Mps3 is most related to fission yeast Sad1, C. elegans SUN-1, and mammalian Spag4, which have been shown to be essential for tethering SPBs, centrosomes, and axonemes, respectively, with the nuclear envelope (Hagan and Yanagida, 1995; Shao et al., 1999; Malone et al., 2003). Interestingly, these Mps3-like SUN proteins not only interact with various types of microtubule-organizing centers but also appear to bind to proteins, such as Kms1, ZYG-12, and now Mps2, that show very little sequence similarity to the KASH domain of other SUN binding proteins that interact with the actin cytoskeleton (Starr and Fischer, 2005). One idea is that microtubule SUN proteins may recognize a distinct or divergent motif that is currently poorly defined because of the low number of known binding proteins.

The requirement of an intact Mps3 SUN domain for nuclear migration during karyogamy is consistent with the role of C. elegans UNC-84 and SUN-1 in nuclear positioning (Malone et al., 1999, 2003). Mps3 SUN mutants also affect the second step of karyogamy. This may be due to a defect in SPB integrity; however, given that hSun1 is required for homotypic membrane fusion in vitro (Hoffenberg et al., 2000), the requirement for Mps3 during nuclear fusion may reflect a more general role for SUN proteins as mediators of membrane fusion. Interestingly, the SUN domain region of Mps3 has been shown to interact with the DnaJ-like ER membrane chaperone Jem1, which is required for membrane fusion after mating (Nishikawa et al., 2003).

SUN proteins also appear to have functions within the nucleus. Sad1 is required for formation of the meiotic bouquet during fission yeast meiosis (Shimanuki et al., 1997; Chikashige et al., 2006), and mammalian Sun1 proteins have Zn²⁺ finger domains that may interact directly with chromatin (Haque et al., 2006). In addition, Mps3 is required for the establishment of sister chromatid cohesion during S phase in budding yeast (Antoniacci et al., 2004), and one of our SUN domain mutants (mps3-Y502H) is found in the mps3-5 allele that has defects in both centromere and telomere cohesion (Antoniacci and Skibbens, 2006). Additional nuclear functions could be one reason for the spindle checkpoint–dependent mitotic delay that we observed in mps3-W477A, -F592S, and -ΔSUN2 cells (unpublished data). How mutations in the SUN domain might affect nuclear functions of Mps3 is unclear at present. However, our identification of Mps3 as the budding yeast SUN protein should facilitate rapid identification and analysis of additional SUN binding partners and greatly enhance our understanding of SUN protein functions.

All strains are derivatives of W303 (ade2-1 trp1-1 leu2-3,112 ura3-1 his3-11,15 can1-100) and are listed in Table S1 . The entire ORF of MPS2 or MPS3 was deleted in a diploid with HIS3MX6 or KANMX6 using PCR-based methods, and the strain was transformed with a centromeric URA3-marked plasmid containing the corresponding wild-type ORF, sporulated, and dissected to generate mps2Δ∷KANMX6 pURA3-MPS2 and mps3Δ∷HIS3MX6 pURA3-MPS3 strains in both mating types. mps2 and mps3 alleles in pRS305 were created by oligonucleotide-directed mutagenesis and integrated into the LEU2 locus in single copy in the appropriate deletion. MPS3 was fused to 3xFLAG and MPS2 was fused to 13xMYC using PCR. The 9xMYC-mps2-381 was created by mutagenesis of pRS304-9xMYC-MPS2 (Munoz-Centeno et al., 1999), and vectors containing both wild-type and mutant forms of MPS2 were cut with BstZ17I to direct integration into the TRP1 locus of SLJ1901 followed by loss of the covering MPS2 plasmid on plates containing 5-fluorootic acid (5-FOA).

SAD1 and MPS3 were amplified by PCR from pREP42-SAD1 (a gift from I. Hagan, Paterson Institute for Cancer Research, Manchester, UK) and pSJ140 (Jaspersen et al., 2002), respectively, and cloned into a pRS306-based vector containing the GAL1/10 promoter to create pURA3-GAL-SAD1 (pSJ379) and pURA3-GAL-MPS3 (pSJ146). MPS2 and mutants, the SUN domain of MPS3 and mutants and SPC29 were amplified by PCR and cloned into pMAL-c2 (New England Biolabs, Inc.) or pQE10/pQE11 (QIAGEN) for expression in bacteria as MBP or 6xHis fusion proteins, respectively.

6xHis-Spc29 was purified from BL21(DE3) bacteria transformed with pQE11-SPC29 (pSJ260) by metal affinity followed by anion exchange chromatography, and the C terminus of Mps3 was purified from BL21(DE3) pLysS bacteria transformed with pMALc2-MPS3-Ct (pSJ151) using amylose resin. Antibodies against purified 6xHis-Spc29 or MBP-Mps3-Ct were generated in rabbits (Animal Pharm) and purified according to the manufacturer's instructions on a Spc29 or Mps3 column created using the Sulfo-link kit (Pierce Chemical Co.). Mps2 antibodies were generated by injecting a KLH-coupled peptide corresponding to residues 212–233 of Mps2 into guinea pigs (Pocono) followed by affinity purification on MBP-Mps2 bound to nitrocellulose filters.

The Mps3 SUN domain was also purified from BL21(DE3) pLysS bacteria transformed with pQE10-mps3SUN (pSJ413) by metal affinity chromatography, and 0.5 mg purified 6xHis-mps3SUN labeled with Alexa 680 as described previously (Jaspersen et al., 2002). Expression of MBP, MBP-Mps2 (pSJ426), and MBP-Kar1 (pSJ425) in BL21(DE3) pLysS cells was induced by the addition of 0.1 mM IPTG for 2 h at 23°C. Cells from 1 ml of culture were resuspended in 200 μl 2× SDS sample buffer and heated to 100°C for 5 min, 5 μl of MBP-Mps2 and MBP-Kar1 and 2 μl of MBP-only samples were separated by 8% SDS-PAGE, and proteins were electrophoretically transferred to nitrocellulose. Membranes were probed with 10 μg Alexa 680–labeled 6xHis-mps3SUN, and binding was assessed using the Odyssey imaging system (LiCor). Expression of bacterial proteins was confirmed by Western blotting with a 1:1,000 dilution of anti-MBP antibodies (New England Biolabs, Inc.). For bacterial pull-down assays, MBP, MBP-Mps2, and MBP-mps2 mutants were expressed in BL21(DE3) pLysS cells by the addition of 0.1 mM IPTG for 2 h at 23°C and cell pellets from 2 ml of culture were resuspended in 0.5 ml PBS containing 0.1% Triton X-100 and 1 mM PMSF, vortexed for 5 min at 4°C, and centrifuged at 14,000 g for 10 min at 4°C. 100 μl of each extract was mixed with 1 μg of purified 6xHis-mps3SUN in a 500-μl immunoprecipitation that included 2 μl protein A–coated magnetic beads (Dynal) and 1 μl of affinity-purified anti-Mps2 antibodies. After a 1-h incubation a 4°C, immunoprecipitates were washed five times in PBS containing 0.1% Triton X-100 and an additional 250 mM NaCl.

Spheroplast lysates were prepared from 100 OD₆₀₀ midlog phase cells. Harvested cells were resuspended in 5 ml spheroplast buffer (50 mM Tris-HCl, pH 7.5, 1.2 M sorbitol, and 10 mM NaN₃) containing 40 mM β-mercaptoethanol and treated with 300 μg/ml zymolase 100T for 30 min at 30°C. Cells were then washed three times in spheroplast buffer, resuspended in 2.5 ml spheroplast lysis buffer (20 mM Hepes-KOH, pH 7.4, 100 mM K-acetate, 5 mM Mg-acetate, 1 mM EDTA, 1 mM PMSF, and 1 μg/ml each of pepstatin A, aprotinin, and leupeptin), and lysed by dounce homogenization. Lysates were transferred to ice and treated with 0.2 mg/ml dithiobis(succinimidyl)propionate for 10 min. Cross-linking was quenched by addition of 50 mM Tris-HCl, pH 8.0, and lysates were centrifuged at 13,000 g for 15 min at 4°C. The pellet was resuspended in 2.5 ml solubilization buffer (spheroplast lysis buffer plus 1 M NaCl and 1% Triton X-100) and recentrifuged at 13,000 g for 15 min at 4°C. The supernatant was added to 20 ml spheroplast lysis buffer immediately before immunoprecipitation.

Ground lysates were prepared from 500 OD₆₀₀ of midlog phase cells as described previously (Niepel et al., 2005). In brief, cells were frozen in liquid nitrogen and ground with a Retsch ball mill. Ground cell powder was allowed to thaw on ice and resuspended in 10 ml extraction buffer (25 mM Hepes-NaOH, pH 7.5, 300 mM NaCl, 0.1 mM EDTA, 0.5 mM EGTA, 2 mM MgCl₂, 0.5% Triton X-100, 1 mM DTT, 1 mM PMSF, and 1 μg/ml each of pepstatin A, aprotinin, and leupeptin). After homogenization with a Polytron 10/35 for 30 s, lysates were centrifuged at 3,000 g for 10 min at 4°C, and the resulting supernatant was used for immunoprecipitations.

100 μl FLAG M2 (Sigma-Aldrich) or 9E10 conjugated (Santa Cruz Biotechnology, Inc.) beads were added to lysates to immunoprecipitate Mps3-FLAG or Mps2-13xMYC, respectively. After a 2-h incubation at 4°C, beads were washed five times in extraction buffer, and 1/10 of the bound protein and 1/1,000 of the input protein was analyzed by SDS-PAGE followed by Western blotting. The following primary antibody dilutions were used: 1:5,000 anti-MYC A14 (Santa Cruz Biotechnology, Inc.), 1:1,000 anti-FLAG (Sigma-Aldrich), 1:2,000 anti-Cdc31 (Jaspersen et al., 2002), 1:5,000 anti-Spc29, 1:2,000 anti-Mps3, 1:2,000 anti-Spc110 (a gift of T. Davis, University of Washington, Seattle, WA), 1:1,000 anti-GFP B34 (Covance), 1:2,500 anti-MBP (New England Biolabs, Inc.), and 1:10,000 anti–glucose-6-phosphate dehydrogenase (G6PDH; Sigma-Aldrich). Alkaline phosphatase–conjugated secondary antibodies were used at 1:10,000 (Promega).

Approximately 5 × 10⁶ midlog phase cells of opposite mating types grown at 23°C were mixed onto a 45-μm nitrocellulose filter, and filters were incubated on YPD plates for 5 h at 30°C. Cells were then stained with DAPI, and nuclear position was analyzed by fluorescence microscopy.

Analysis of DNA content by flow cytometry, EM, and protein localization by indirect immunofluorescence and epifluorescence microscopy were performed as previously described (Jaspersen et al., 2002). Spc110 polyclonal antibodies were a gift from T. Davis and were used at a 1:2,000 dilution. Mps3 was localized using affinity-purified anti-Mps3 antibodies diluted 1:500 in PBS containing 3% BSA and 0.1% NP-40 followed by detection with Alexa 555–conjugated goat anti–rabbit secondary antibodies (Invitrogen) diluted 1:20,000 in the same buffer. Cells were examined with an Axioimager (Carl Zeiss MicroImaging, Inc.) using a 100× Plan-Fluar lens (NA = 1.45; Carl Zeiss MicroImaging, Inc.), and images were captured with a digital camera (Orca ER; Hamamatsu) and processed using Axiovision 4.0 (Carl Zeiss MicroImaging, Inc.).

Homologues of Mps3 were detected through iterative searches of the nonredundant protein sequence database (National Center for Biotechnology Information) using the PSI-BLAST program (Altschul et al., 1997). Sequences corresponding to the SUN domain of Mps3 homologues and C termini of putative SUN binding proteins were aligned using ClustalW, and phylogenetic analysis was performed using the MEGA3 software.

Fig. S1 shows that ts SUN domain mutants arrest in mitosis because of a defect in SPB duplication. Fig. S2 shows a phylogenetic analysis of SUN domains. Table S1 lists yeast strains. Table S2 shows SUN domain mutants. Table S3 gives copy number–dependent SUN domain mutants.

We thank I. Hagan and T. Davis for reagents; K. Delventhal, J. Wunderlich, and J. Schwartz for technical assistance; and D. Starr, J. Gerton, K. Lee, C. Pearson, and S. Hawley for valuable discussions and comments on the manuscript. We are especially grateful to S. Shelton, J. Bupp, and E. Hoffman for their contributions to this project.

This work was supported by funding from the National Institutes of General Medical Sciences (grant GM51312 to M. Winey) and the Leukemia and Lymphoma Society (to S.L. Jaspersen). S.L. Jaspersen and A. Mushegian are also funded by the Stowers Institute for Medical Research.