Cdc7p interacts preferentially with GTP–Spg1p

Our previous studies have demonstrated that Spg1p and Cdc7p bind to each other and that mutations in the effector domain of Spg1p interfere with the interaction: It was also proposed that the GTP-bound form of Spg1p was the biologically active form (Schmidt et al. 1997). We therefore tested whether Spg1p binding to GDP or GTP affected its interaction with Cdc7p. A GST–Spg1p fusion protein was expressed in Escherichia coli. The purified protein was stabilized in the GDP- or GTP-bound form and mixed with protein extracts from wild-type S. pombe cells, and proteins binding to Spg1p were recovered by affinity chromatography on glutathione–agarose columns. Western blot analysis showed that Cdc7p was recovered much more efficiently by the GTP-bound form of Spg1p (Fig. ​(Fig.22B).

Spg1p is associated with the spindle pole body throughout the mitotic cellcycle

The data described above suggest that regulation of the onset of septation may be mediated by controlling the location of Cdc7p, the availability of substrate, or both. To pursue this idea further, we examined the localization of Cdc7p and Spg1p through the cell cycle. To localize Spg1p, the influenza hemagglutinin (HA) epitope tag, recognized by the 12CA5 monoclonal antibody, was added to the carboxyl terminus of Spg1p (Spg1–HAp) by modifying the chromosomal copy of the spg1 gene (Fig. ​(Fig.2C;2C; see Material and Methods). Indirect immunofluorescence using mAb 12CA5 showed that Spg1–HAp associated with the spindle pole body throughout the cell cycle (Fig. ​(Fig.3A,C).3A,C). Spg1p staining colocalized with the spindle pole body marker Sad1p during interphase and mitosis (Fig. ​(Fig.4A,4A, Hagan and Yanagida 1995). Staining for Sad1p (data not shown) and microtubules (Schmidt et al. 1997) did not reveal any mitotic defects in the spg1 null allele, indicating that Spg1p is not an essential structural component of the spindle pole body.

Open in a separate window

Figure 3

Localization of Cdc7p and Spg1p to the spindle pole body. Cells from the indicated strains were grown at 25°C in YE, fixed, and stained, using the indicated antibodies. DAPI shows the DNA, and phase contrast shows the cell outline. (A) spg1–HA cells stained with mAb 12CA5. Cell 1 is in interphase; in cell 2, the two spindle pole bodies have started to separate; cell 3 is in early anaphase; cell 4 is in late anaphase; and cell 5 has formed the division septum. (B) Wild-type cells (972), stained with anti-Cdc7p antibodies and TAT-1. Cell-cycle stages are indicated in the text. (C) Mitotic spg1–HA cells, stained with mAb 12CA5 and anti-Cdc7p antibodies. (D) cdc7–GFP fluorescence in mitotic cells. The arrows indicate the position of the birth scars.

Open in a separate window

Figure 4

Cdc7p and Spg1p colocalize with the spindle pole body component Sad1p. (A) spg1–HA cells were stained with anti-Sad1p antibodies and mAb 12CA5. Background fluorescence from secondary antibodies was used to show the cell outline. The lower cell is in interphase; the upper cell is undergoing anaphase. Sad1p was detected with FITC-conjugated secondary antibody, which gives green fluorescence, and Cdc7p was detected with CY3-conjugated secondary antibody, which gives red fluorescence. Colocalization of green and red gives a yellow color, that is seen in the merged image. The individual channels are shown in gray tones, the merged image is in color. (B) Wild-type cells were stained with DAPI, anti-Sad1p (raised in sheep), and anti-Cdc7p antibodies. A mitotic cell is shown. In the combined image, DAPI is shown in violet, Sad1p gives red fluorescence, and Cdc7p gives green fluorescence. Colocalization of Sad1p and Cdc7p on one pole results in yellow (see A) color.

Cdc7p is located on both spindle pole bodies early in mitosis and one spindle pole body late inmitosis

Cdc7p was detected using a polyclonal antibody raised against a region carboxy-terminal to the catalytic domain (Fankhauser and Simanis 1994). In mitotic cells with a short spindle, Cdc7p associated with both spindle pole bodies (Fig. ​(Fig.3B,3B, cells 2 and 3). After initiation of anaphase B, Cdc7p immunofluorescence was observed on only one of the two spindle poles (Fig. ​(Fig.3B,3B, cell 4, and Figs. ​Figs.3C3C and ​and4B),4B), and this asymmetry persisted in cells that were forming a division septum (Fig. ​(Fig.3B,3B, cell 5). Despite the fact that the steady-state level of Cdc7p was constant during interphase, mitosis, and cytokinesis (Fig. ​(Fig.1B),1B), no discrete Cdc7p staining was seen either in cells that had completed the septum (Fig. ​(Fig.3B,3B, cell 6), or in interphase cells (Fig. ​(Fig.3B,3B, cell 1). No discrete Cdc7p staining was detected in cells arrested in late G₂ by the cdc25-22 mutation (data not shown), indicating that association of Cdc7p with the spindle pole body depends on entry into mitosis.

To exclude the possibility that the Cdc7p epitopes recognized by the polyclonal serum were masked on one of the two poles, we used strains in which the chromosomal copy of the cdc7 gene had been replaced with a modified version, tagging Cdc7p at its carboxyl terminus with either the HA epitope (cdc7–HA cells; Schmidt et al. 1997) or green fluorescent protein (cdc7–GFP; this study). Indirect immunofluorescence using mAb 12CA5 (Fig. ​(Fig.7D,7D, below) and observation of the Cdc7p–GFP fluorescence (Fig. ​(Fig.3D)3D) confirmed the asymmetric association of Cdc7p with one spindle pole. Using the birth scar as a morphological marker for asymmetry (Mitchison and Nurse 1985), no significant bias in segregation of Cdc7p to the old or new end of the cell was observed (Fig. ​(Fig.3D).3D). We do not know what percentage of the total amount of Cdc7p is associated with the spindle pole body. Therefore, at present, we cannot say whether the asymmetric pattern of Cdc7p localization means that only one of the two daughter cells inherits Cdc7p. The location of Cdc7p during interphase is unknown, but it is possible that the absence of a discrete staining pattern in interphase cells reflects a diffuse cytoplasmic staining that falls below the limit of detection.

Open in a separate window

Figure 7

Localization of Spg1p with an antibody recognizing the GDP-bound form of Spg1p (SuSu1). Cells were grown at 25°C in YE and prepared for immunofluorescence, using the indicated antibodies. DAPI shows the DNA; phase contrast shows the cell outline. The antibodies used in each case are indicated. cdc16-116 cells were shifted to 32°C for 1.5 hr before fixation. (A) Wild-type cells (972), stained with the anti-Spg1p antiserum (SuSu1). Cells 1 and 2 are in interphase; cells 3 and 4 are undergoing anaphase B; cell 5 is forming a division septum; cell 6 has completed septum formation. (B) Early mitotic wild-type cells (972) stained with SuSu1 and TAT-1. (C) cdc16-116 cells stained with SuSu1. (D) Late mitotic cdc7–HA cell stained with SuSu1 and mAb 12CA5.

We were concerned about the possibility that Cdc7p staining associated with microtubules, the medial ring, or equatorial MTOCs might fall below the limit of detection. We therefore stained cells expressing a level of Cdc7p ~20-fold higher than normal from a multicopy plasmid. Although we observed an increase in cytoplasmic staining, no medial ring, microtubule, or equatorial MTOC staining was seen under these conditions, suggesting that Cdc7p does not normally associate with these structures (data not shown). Formally, we cannot exclude the possibility that the number of binding sites for Cdc7p at these locations is limited, precluding its detection with our current reagents; however, we feel that our combined genetic and cytological observations make this interpretation unlikely.

Localization of Cdc7p to the spindle pole body requires Spg1p function, but not viceversa

GTPase-mediated localization of kinases to their sites of action is an emerging theme in the function of signal transduction systems (e.g., Hall 1994; Mochley-Rosen 1995; Peter et al. 1996; Leberer et al. 1997). Because functional Spg1p is not required for Cdc7p kinase activity, we examined the effects of cdc7 and spg1 mutations upon localization of Spg1p and Cdc7p. No discrete Cdc7p staining was detected during mitosis in cells arising from germination of spg1 null allele spores, or in spg1–B8 cells incubated at restrictive temperature (Fig. ​(Fig.5A).5A). Therefore, Spg1p activity is required to localize Cdc7p to the spindle pole body. In cells deleted for the cdc7 gene (Fig. ​(Fig.5B),5B), and the cdc7-24 mutant (data not shown), Spg1p localized to the spindle pole body in both interphase and mitosis [note that the signal is asymmetric in mitotic cells because of the use of a polyclonal antiserum recognizing preferentially GDP–Spg1p (SuSu1; see below)]. Thus, Cdc7p is not required for correct localization of Spg1p. Depolymerization of microtubules by cold shock (Hagan and Yanagida 1995) and staining of the β-tubulin mutant nda3–KM311 (Hiraoka et al. 1984) showed that spindle pole body localization of Cdc7p and Spg1p is independent of microtubule integrity (data not shown).

Open in a separate window

Figure 5

Spindle pole body localization of Cdc7p requires Spg1p, but correct localization of Spg1p occurs in the absence of Cdc7p. spg1–B8 cells were grown at 25°C and shifted for 4 hr to 36°C before fixation. Spores deleted for spg1 or cdc7 were germinated in minimal medium containing supplements allowing only spores carrying the null allele to germinate (described in Materials and Methods). (A) spg1 null allele (top) and spg1-B8 (bottom) cells were stained with anti-Cdc7p antibodies (only mitotic cells are shown). (B) cdc7 null cells stained with anti-Spg1p antibodies (SuSu1). The top panels show a binucleate interphase cell; the bottom panels show a cell undergoing the second mitosis. The positions of the nuclei and the spindle pole bodies in the mitotic cell are schematically represented. Note that the right nuclei of each segregating pair almost colocalize. Spindle pole body-mediated clustering of nuclei is often observed in cdc7 mutants (for a detailed description of nuclear positioning in fission yeast, see Hagan and Yanagida 1997).

Deregulated septum formation correlates with symmetrical distribution of Cdc7p duringmitosis

Because Cdc7p always segregates to only one pole of the spindle in late anaphase in wild-type cells, we examined whether this was the case in cells where septum formation is deregulated. It has been proposed that the cdc16 gene product encodes the GTPase-activating protein (GAP) of Spg1p (Boguski and McCormick 1993; Schmidt et al. 1997). Consistent with this, the multiseptate phenotype resulting from loss of cdc16 function resembles that produced by increased expression of Spg1p and Cdc7p (Mineti et al. 1979; Fankhauser et al. 1993; Fankhauser and Simanis 1994; Schmidt et al. 1997). cdc16-116 cells grown at restrictive temperature were stained with the anti-Cdc7p antiserum. Cdc7p was seen to associate with both spindle pole bodies throughout anaphase and septum formation (Fig. ​(Fig.6A),6A), in contrast to wild-type cells, where Cdc7p was associated with only one pole of late anaphase spindles. Spg1p was also present on both poles (Fig. ​(Fig.6B).6B). The same results were obtained when the experiment was repeated using germinating spores deleted for the cdc16 gene (data not shown). Cells in which septum formation is deregulated by increased expression of Spg1p (Fig. ​(Fig.6C),6C), and cells overexpressing Cdc7p (data not shown), also showed a bipolar distribution of Cdc7p in mitotic cells, confirming the correlation between the symmetric segregation of Cdc7p at mitosis and deregulation of septum formation. Cdc7p was also located on the spindle pole in septated, mononucleate cells induced by increased expression of Spg1p (Fig. ​(Fig.6C,6C, Schmidt et al. 1997). Thus, in wild-type cells Cdc7p is associated with only one spindle pole in late anaphase, whereas deregulation of septum biosynthesis correlates with association of Cdc7p with both poles of the spindle throughout mitosis and septum formation. The potential significance of this is addressed in the Discussion.

Open in a separate window

Figure 6

Loss of Cdc16p function results in segregation of Cdc7p to both poles of the mitotic spindle. cdc16-116 cells were shifted from 25°C to 32°C for 1.5 hr before fixation. Cells were prepared for immunofluorescence and stained using the indicated antibodies. DAPI shows the DNA; phase contrast shows the cell outline. (A) cdc16-116 cells stained with anti-Cdc7p antibodies. Note that the cell marked with an arrowhead has completed anaphase and that the spindle pole bodies guide the movement of the nuclei toward the center of the future daughter cells (Hagan and Yanagida 1997). (B) cdc16-116 spg1–HA cell, stained with mAb 12CA5. (C) Cells overexpressing Spg1p (induced for 18 hr at 25°C from the nmt1 promotor SP1805; Schmidt et al. 1997), stained with anti-Cdc7p antibodies. Cell 1 is undergoing anaphase and shows symmetrical Cdc7p staining; cell 2 is in interphase and has Cdc7p associated with the spindle pole body; cell 3 has Cdc7p associated with the spindle pole body and initiated septum formation during interphase.

Implications of these data for the regulation of septum formation

Kinase assays in vitro suggest that Cdc7p has the potential to be active at all stages of the cell cycle. Assuming that this is an accurate reflection of the in vivo situation, the finding that Spg1p is in the GTP-bound form and Cdc7p is already associated with the spindle poles early in mitosis, suggests that Cdc7p/Spg1p is prevented from signaling the onset of septum formation precociously during the early stages of mitosis. This could be achieved by making the key substrates unavailable, or unresponsive, to Cdc7p kinase activity until the end of mitosis, or by the presence of an inhibitor of Cdc7p during mitosis. Removal of an inhibitor of Cdc7p/Spg1p (or changing substrate availability) at the end of mitosis would provide an excellent means of coordinating mitotic events with the onset of septum formation.

Increased expression of Spg1p in interphase cells promotes medial ring and septum formation (Schmidt et al. 1997). This indicates that the key substrates of Cdc7p/Spg1p are potentially available throughout the cell cycle and therefore suggests that the activity restraining septum formation during mitosis is absent during interphase. Consistent with this, induction of spg1 expression does not induce septum formation in cells arrested in early mitosis by the nda3–KM311 mutation (L. Cerutti and V. Simanis, unpubl.), though it will do so in G₁-, S-, or G₂-arrested cells (Schmidt et al. 1997). Septation during interphase may therefore be restrained by regulating activation of Spg1p.

It is noteworthy that inactivation of Cdc2p kinase in a mitotically arrested cell will bring about cytokinesis or septum formation in both S. cerevisiae and S. pombe (Ghiara et al. 1991; He et al. 1997). Moreover, mutants that arrest late in mitosis with elevated Cdc2p kinase activity do not undergo septum formation and cytokinesis (Moreno et al. 1989; Surana et al. 1993). It is therefore possible that the onset of septum formation is inhibited by high Cdc2p kinase activity during mitosis. Studies of myosin II regulatory light chain phosphorylation in higher eukaryotes provide a precedent for Cdc2p inactivation in triggering cytokinesis. Myosin II regulatory light chain activity appears to be both negatively and positively regulated by changes in its phosphorylation state. During mitosis, it is inhibited through phosphorylation by Cdc2p/cyclin B (Satterwhite et al. 1992). When Cdc2p/cyclin B is inactivated in the normal cell cycle, coincident with the metaphase-anaphase transition, these phosphorylations no longer occur; and coincident with this, phosphorylation of another site by myosin light chain kinase is markedly increased (Yamakita et al. 1994). The changes at these phosphorylation sites are thought to contribute to signaling the initiation of cytokinesis.

The onset of anaphase and inactivation of Cdc2p/cyclin B kinase is brought about by the proteolysis of key substrates targeted by the anaphase-promoting complex (APC). The finding that some cut mutants encode components of the APC (Yamashita et al. 1996; Yamada et al. 1997) argues that the putative inhibitor of septation may not be removed by APC-mediated proteolysis. However, increased expression of the APC component Nuc2p blocks septation but does not interfere with mitosis (Kumada et al. 1995). Clearly, future studies will address how Cdc7p/Spg1p signaling is activated at the end of mitosis and whether the APC plays a role in this signaling.

What mediates the asymmetrical distribution of Cdc7p on the spindle poles during anaphase?

Electron microscopic observation shows that it is highly likely that two apparently identical daughter spindle pole bodies arise by the splitting of a single precursor in S. pombe (Ding et al. 1997). Our data clearly indicate that the two daughter pole bodies produced by this apparently symmetrical structure are not identical.

In S. cerevisiae, a Kar1p–β-galactosidase fusion protein has been shown to localize predominantly to the cytoplasmic face of the new spindle pole body when expressed at high levels (Vallen et al. 1992). However, localization of native Kar1p demonstrated that it is present on the spindle pole body throughout the cell cycle and is a component of the bridge and half-bridge (Spang et al. 1995). The asymmetry observed in the earlier study was attributable to misdirection of the Kar1p–β-galactosidase fusion protein within the spindle pole body. Thus, our data represent the first example of asymmetric distribution of a native molecule on the spindle poles during mitosis. It is interesting to note that an example of nonequivalence in the centrioles of a higher eukaryote has been described recently (Lange and Gull 1995). The antibody marker called Cenexin is acquired by the immature centriole of the pair only at prophase: For the remainder of the cycle, only the mature centriole is Cenexin positive.

The nature of the signal that promotes the switch from bipolar to monopolar Cdc7p localization during anaphase is unknown: It may be attributable to progression past a mitotic checkpoint, or it might be a function of spindle length. The timing of the switch from bipolar to monopolar Cdc7p staining is similar to the timing of the disappearance of Cdc2p/Cdc13p from the poles (Alfa et al. 1990). However, we do not know at present whether APC activity is required for the transition to monopolar Cdc7p localization. Since the disappearance of Cdc7p from one spindle pole correlates with inactivation of Spg1p, and Cdc7p and GDP–Spg1p never colocalize in late anaphase, we favor the view that Cdc7p is delocalized from one pole by asymmetric inactivation of Spg1p. Because Spg1p associates with both spindle pole bodies throughout mitosis, it seems likely that the factors regulating Spg1p activity will be distributed, or regulated, asymmetrically on the spindle poles. Potential negative regulators of Spg1p are Cdc16p (Fankhauser et al. 1993; Schmidt et al. 1997), Byr4p (Song et al. 1996), and Dma1p (Murone and Simanis 1996). Their localization during mitosis will be of considerable interest.

Why does increased expression of cdc7, plo1, or spg1 deregulate septumformation?

Plo1p kinase is essential not only for formation of a bipolar spindle but also for septum formation: Ectopic expression of the protein can also trigger septum formation in interphase cells (Ohkura et al. 1995). Studies of Plk1, the human counterpart of plo1, have demonstrated that microinjection of neutralizing antibodies prevents centrosome maturation (Lane and Nigg 1996). It is tempting to speculate that an untimely increase of Plo1p kinase activity during interphase in S. pombe causes premature spindle pole body maturation and activation of Spg1p. This could be achieved by negative regulation of Spg1p inhibitors, recruitment of Spg1p activators, or both. Because increases in Spg1p activity can induce all stages of septum formation (Schmidt et al. 1997), this would explain the observed phenotypes of plo1 overexpression. It will be interesting to examine the localization of Spg1p and Cdc7p in plo1 mutants, when they become available.

Previous work has shown that cells are very sensitive to increased spg1 expression but much less so to increased cdc7 expression (Fankhauser and Simanis 1994; Schmidt et al. 1997). The finding that spindle pole body association of Cdc7p requires GTP–Spg1p may offer an explanation for this apparent discrepancy. We have shown that Spg1p is mostly GDP bound in interphase cells, so the tendency for Cdc7p to associate with the spindle pole body, where we presume its key substrates are localized, will be low. Thus, increased expression of Cdc7p will not induce septum formation in interphase-arrested cells. In contrast, during mitosis, Spg1p is GTP-bound, and Cdc7p can associate with the spindle pole. It is possible that the excess of Cdc7p titrates away the putative negative regulator from the spindle pole body, giving deregulated signaling and multiple rounds of septum formation at the end of mitosis. Alternatively, it may hinder the inactivation of Spg1p.

Immunofluorescence

Indirect immunofluorescence was performed as described previously (Sohrmann et al. 1996), except that cells were digested for 10 min in 1 mg/ml of Zymolyase 20T. The anti-Sad1p (Hagan and Yanagida 1995), anti-Cdc7p (Fankhauser and Simanis 1994), and anti-Spg1p (SuSu1, Schmidt et al. 1997) antisera were affinity purified before use, according to standard procedures (Harlow and Lane 1988). The monoclonal anti-HA antibody (12CA5) was purchased from BabCO. The α-tubulin mAb TAT-1 (Woods et al. 1989) was a generous gift from K. Gull (University of Manchester). Specificity of the staining was confirmed by competition experiments with the antigen used to raise the antiserum (according to Sohrmann et al. 1996) and by staining of germinating spores deleted for either cdc7 or spg1. The epitopes recognized by the anti-Cdc7p and anti-Spg1p antisera and by mAb 12CA5 are sensitive to glutaraldehyde. Note that fixation in the absence of glutaraldehyde only preserves the mitotic spindle but not the interphase microtubule array. Microtubules were depolymerized by incubating the cells on ice for 25 min before fixation, as described by Hagan and Yanagida (1995). Cells were observed with a Zeiss Axiophot microscope. Production of color images and confocal microscopy has been described (Hagan and Yanagida 1997; Sohrmann et al. 1996).

Construction of the spg1–HA and the cdc7–GFP strains

To construct spg1–HA, the carboxy-terminal half of the spg1 gene was amplified by PCR using the oligonucleotides VS 135 (ATCAAAGAATGGTACCGTCAAGCTCG) and VS 151 (GAG GATCCTTAATCAGCGGCCGCTGTATTCCAA) and cloned into pDW232 (Weilguny et al. 1991). The triple HA tag (Tyers et al. 1992) was cloned into the NotI site introduced by VS 151 just before the stop codon. The ura4⁺ gene and 2 kb of genomic sequence 3′ of the spg1 coding sequence were cloned behind the tag. The whole linear insert was transformed into a ura4D18 strain. Stable uracil prototrophs were checked by Southern and Western blotting for correct replacement of the spg1 gene by the tagged version.

The cdc7–GFP strain was constructed as the cdc7–HA strain described previously (Schmidt et al. 1997), except that the HA tag in the transformed linear fragment was replaced by an improved version of the Aequorea victoria GFP (S65T, F64L). All strains in which the tagged gene replaced the wild-type copy were indistinguishable from wild type.

Protein extracts, kinase assays, and immunoprecipitations

Total protein extracts were prepared as described previously (Sohrmann et al. 1996). Extracts of soluble proteins and kinase assays were performed according to Fankhauser and Simanis (1994), except that mAb 12CA5 was used for Western blot analysis, after immunoprecipitation with polyclonal anti-Cdc7p antiserum. To precipitate Spg1p or Spg1–HAp, 5 μl of affinity-purified anti-Spg1p antiserum, or 1 μg of mAb 12CA5, respectively, was added to protein extracts (1.5 mg) prepared in the absence or presence of 0.6 mm GTPγS (Sigma) and incubated for 60 min at 4°C. The complexes were then precipitated using protein A–Sepharose (Sigma) and analyzed by Western blotting using the same antibody as for the immunoprecipitation.

We thank Lorenzo Cerutti for the gift of purified GST–Spg1p fusion protein. We are grateful to Bruno Amati, Nicola Beltraminelli, Lorenzo Cerutti, Christian Fankhauser, Xavier LeGoff, Matthias Peter, and Suzan Utzig, for discussions and comments on the paper. We are also grateful to Elena Cano for technical assistance. This work was supported by ISREC, the Swiss National Science Foundation, the Swiss Cancer League, the Cancer Research Campaign (CRC), and a short-term European Molecular Biology Organization (EMBO) fellowship to M.S.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC section 1734 solely to indicate this fact.