The Saccharomyces cerevisiae RNase Mitochondrial RNA Processing Is Critical for Cell Cycle Progression at the End of Mitosis
Ti Cai, Jason Aulds, Tina Gill, Michael Cerio, Mark E. Schmitt

Abstract

We have identified a cell cycle delay in Saccharomyces cerevisiae RNase MRP mutants. Mutants delay with large budded cells, dumbbell-shaped nuclei, and extended spindles characteristic of “exit from mitosis” mutants. In accord with this, a RNase MRP mutation can be suppressed by overexpressing the polo-like kinase CDC5 or by deleting the B-type cyclin CLB1, without restoring the MRP-dependent rRNA-processing step. In addition, we identified a series of genetic interactions between RNase MRP mutations and mutations in CDC5, CDC14, CDC15, CLB2, and CLB5. As in most “exit from mitosis” mutants, levels of the Clb2 cyclin were increased. The buildup of Clb2 protein is not the result of a defect in the release of the Cdc14 phosphatase from the nucleolus, but rather the result of an increase in CLB2 mRNA levels. These results indicate a clear role of RNase MRP in cell cycle progression at the end of mitosis. Conservation of this function in humans may explain many of the pleiotropic phenotypes of cartilage hair hypoplasia.

RIBONUCLEASE mitochondrial RNA processing (RNase MRP) is a ribonucleoprotein endoribonuclease that cleaves an RNA sequence in a site-specific manner (Chang and Clayton 1987). RNase MRP was initially isolated from mammalian mitochondria; however, cellular fractionation and immunolocalization experiments have revealed that the majority of the complex is localized to the nucleus of mammalian cells (Reimeret al. 1988). The enzyme has been remarkably conserved from yeast to humans in both the RNA and protein components (Schmitt and Clayton 1992; Lygerouet al. 1996; van Eenennaamet al. 2000). In the yeast Saccharomyces cerevisiae, a role for RNase MRP in nucleolar processing of rRNA has been found (Schmitt and Clayton 1993; Chuet al. 1994; Henryet al. 1994). Mutations in the RNA component of the human RNase MRP have been shown to be the cause of the genetic disease cartilage hair hypoplasia (CHH; Ridanpääet al. 2001). This disease is characterized by short stature, brittle and sparse hair, and immunodeficiency (Mäkitieet al. 1995; Clayton 2001).

The gene for the S. cerevisiae MRP RNA is called NME1 for nuclear mitochondrial endonuclease 1 (Schmitt and Clayton 1992). In addition, at least nine yeast proteins associated with the MRP RNA in vivo have been identified. Eight of these proteins are shared with the ribonucleoprotein endoribonuclease RNases P (Lygerouet al. 1994; Chuet al. 1997; Dichtl and Tollervey 1997; Stolc and Altman 1997; Chamberlainet al. 1998). One protein encoded by the SNM1 gene encodes an RNA-binding protein that is associated only with the RNase MRP RNA and not the RNase P RNA (Schmitt and Clayton 1994).

All of the components of RNase MRP are essential for the viability of yeast. Mutations in the yeast RNase MRP components lead to a defect in 5.8S-rRNA processing, specifically at the A3 site in the pre-rRNA (Schmitt and Clayton 1993; Chuet al. 1994; Henryet al. 1994). Removal of the A3 processing site and loss of mitochondrial DNA are not lethal in yeast, indicating that there is an unknown essential function for RNase MRP (Henryet al. 1994; Venema and Tollervey 1999). In support of this idea, mutations in the SNM1 gene can lead to defects in plasmid stability (Caiet al. 1999). In Schizosaccharomyces pombe, a mutation in the MRP RNA was found to lead to a defect in septation (Paluh and Clayton 1996). The cause of the phenotypes in these mutants was not determined.

We report here the identification of a specific cell division cycle delay in RNase MRP mutants. Cells accumulate late in the mitotic cycle with large budded cells, dumbbell-shaped nuclei, and extended spindles, identical to that seen with previously described exit from mitosis mutants (EFM; Suranaet al. 1993). A series of genetic interactions were identified between mutations in RNase MRP components and known EFM genes. Western analysis of the mitotic cyclin Clb2 showed accumulation of Clb2 protein as is the case in most EFM mutants. These observations suggest an entirely novel role for RNase MRP in control of the cell cycle and potentially identify its essential function. In addition, the cell cycle delay lends insights as to the cause of the pleotropic phenotypes seen in CHH (Mäkitieet al. 1995).

MATERIALS AND METHODS

Strains and media: Yeast media and genetic manipulations have been described (Caiet al. 1999). Basic molecular biology techniques were performed as described (Sambrooket al. 1989). The genotypes of all yeast strains used are given in Table 1.

Construction of the yeast strain YMC6: YMC6 expresses three tagged subunits of the RNase MRP enzyme simultaneously: Snm1:GFP:6HIS, Pop3:GFP, and Pop4:3×HA. To construct this strain the POP3 gene was deleted in the strain MES101 and marked with the HIS3 gene (Longtineet al. 1998) to create the strain YGB22 (see Table 1). This strain was transformed with the plasmid pTC143 (pRS315 carrying the POP3 gene) to generate the strain YGB24 and was subsequently sporulated to generate the strain YGB24-38A. YL323 was transformed with the plasmid pSC93 (kind gift of L. Lindahl, University of Maryland; Chuet al. 1997) and subsequently sporulated to generate the strain YGB26-14D. YGB26-14D was mated to YGB24-38A to create the diploid strain YGB28 and the two plasmids were allowed to be lost. This strain was transformed with the plasmid pGAB116 (contains both a GFP:POP3 fusion gene and a POP4:3XHA gene; Chuet al. 1997). The resulting strain, YGB29, was sporulated and the double pop3 and pop4 mutant YGB29-4A was chosen for subsequent use. The strain, THR200, was transformed with the plasmid pGAB119 (SNM1:GFP:6His, LEU2; Caiet al. 1999). The resultant strain YGB30 was crossed to YGB29-4A and the diploid, MES250, was sporulated. A strain with chromosomal deletions of snm1, pop3, and pop4 was selected and named YMC6.

Construction of the CDC14-GFP fusion: PCR-targeted gene modification was performed as described (Longtineet al. 1998). Primers were used to amplify the CDC14 gene and the pFA6a-GFP(S65T)-kanMX6 cassette placing the green fluorescent protein (GFP) tag exactly before the CDC14 termination codon. The final PCR product was purified and transformed into the strain MES117. The transformants were plated on YPD-G418 plates to select for the fusion. The CDC14-GFP construct was verified by PCR and fluorescence microscopy. RNase MRP RNA mutants were shuffled into the resulting strain YTC240 as previously described (Shadelet al. 2000).

Immunofluorescence microscopy: Cells were prepared for immunofluorescence as previously described (Pringleet al. 1991). Rat monoclonal antitubulin antibody, YOL1/34, and Cy3 conjugated rabbit anti-rat secondary antibody were obtained from Accurate Chemical (Westbury, NY). Cells were viewed using a Zeiss Axioskop microscope equipped with epifluorescent and Nomarski optics and a Zeiss Plan-Apochromat ×100 objective. Images were captured in real time using a Diagnostics Instruments Spot Camera 2 directly linked to an Apple PowerMac G3 computer.

Plasmid segregation assay: Plasmid segregation assays were carried out as previously described (Caiet al. 1999), using the yeast strains MES111-140 as a wild-type control and MES111-P6 as the experimental.

Identification of high-copy suppressors of the snm1-172 mutation: The yeast strain YTC150-172 carrying the snm1-172 mutation (T189G; G190C [Cys64Ala] in the SNM1 gene; Caiet al. 1999) was transformed with a yeast, URA3, 2μ genomic library (Carlson and Botstein 1982) and plated onto Ura media at 24°. Ten thousand transformants were replicated to synthetic complete dextrose media (SCD; Shadelet al. 2000) at 37°. Colonies growing at the nonpermissive temperature were picked and retested. Those transformants that continued to grow at 37° on SCD plates, but failed to grow at 37° on SCD plates with 5-fluoroorotic acid, were classified as plasmid-linked suppressors. Yeast 2μ plasmids that conferred 37° growth were rescued from yeast into Escherichia coli (Sikorski and Boeke 1991) and then retested for suppression in the original snm1-172 strain.

Cell arrest experiments: Yeast strains were grown to 106 cells/ml at 25° in SCD, arrested in hydroxyurea (15 mg/ml) or nocodazole (20 μg/ml) for 3 hr, and then shifted to 37° for 3 hr. Cells were harvested and whole-cell protein extracts were made.

Preparation of yeast cell extracts: Yeast was grown in 50 ml of liquid SCD at 25° until they reached exponential growth (107 cells/ml). The cultures were then shifted to the nonpermissive temperature (37°) and grown for 4 hr. The cells were collected and yeast whole-cell extracts were prepared as described (Cai and Schmitt 2001).

Western blot analysis: SDS-PAGE was performed as previously described (Caiet al. 1999). A total of 40 μg of yeast protein extract was denatured in 2% (w/v) SDS, 5% (v/v) β-mercaptoethanol for 5 min at 95°, and resolved on a 15% (w/v) polyacrylamide gel. The anti-rabbit POD was detected with a Boehringer Mannheim (Indianapolis) chemiluminescence Western blotting kit and exposed to film for 10 min and then again overnight. Anti-Vma1 antibody, used as a loading control, was a kind gift of P. Kane (SUNY Upstate Medical University).

Analysis of yeast RNA: RNA was extracted as previously described (Schmittet al. 1990). Approximately 10 μg of whole-cell RNA was separated on 1% (w/v) agarose gels or 6% (w/v) acrylamide/7 m urea gels (Sambrooket al. 1989). Gels were stained with ethidium bromide to visualize the 5.8S rRNA or Northern blot analysis was performed (Sambrooket al. 1989; Schmitt and Clayton 1993). Probes used for Northern analysis were CLB2 (653-bp BglII-HindIII fragment of the CLB2 gene; Fitchet al. 1992) and ACT1 (1141-bp XhoI-KpnI fragment of the ACT1 gene; Shortleet al. 1982). Probes were radiolabeled for hybridization with [α-32P]dCTP using the Prime-It Kit (Stratagene, La Jolla, CA). Radioactive blots were analyzed on a Molecular Dynamics (Sunnyvale, CA) PhosphoImager. Northern blots were probed a second time with actin to ensure equal loading.

Construction of EFM and RNase MRP double mutants: Strains carrying the cdc5-1, cdc14-1, and cdc15-1 were the kind gift of D. Botstein, Stanford University. These strains were backcrossed to our strain background three times in the case of cdc5 and four times for cdc14 and cdc15. Haploid strains from the final cross carrying the relevant markers were used to shuffle in the different NME1 and SNM1 alleles as previously described (Caiet al. 1999; Shadelet al. 2000). Deletions for clb1, clb2, and clb5 were obtained from Research Genetics (San Diego, CA). These double-mutant strains were generated in the same way as the cdcs.

Growth tests of yeast mutants: Mutant yeast strains were analyzed for temperature conditional growth on YPD media plates at 24°, 30°, 34°, and 37°. A total of 100 μl of sterile water was aliquoted to each well of a 96-well tissue culture plate. Two fresh colonies from each yeast strain were mixed into the first row wells. Tenfold serial dilutions of these cell mixtures were done in subsequent wells. A 48-pin Frogger (Dankar) was used to transfer diluted cell mixtures onto YPD media plates. The cells were grown at the temperatures specified for 3 days. All growth tests were performed at least three separate times in duplicates to ensure reproducibility.

RESULTS

Mutations in yeast RNase MRP components cause a cell cycle delay at the end of mitosis: While studying RNase MRP from yeast, genes for several of the known protein subunits of the complex were fused to the gene for GFP, GFP-6 histidine (GFP::6HIS), or hemaglutinin (HA) and used to replace the wild-type copies of the respective genes in yeast cells (Figure 1A). The fluorescent tags allow for cellular localization of the RNase MRP components. Each of the tagged components was able to complement a strain deleted for the original gene and grew as well as the original wild-type strain. However, a strain designated YMC6, which expressed all three tagged subunits simultaneously, showed a significantly increased generation time at 30° and marked temperature sensitivity at 37°. The GFP-tagged subunits corresponding to the GFP::POP3 localized to the nucleus, providing bright nuclear fluorescence (Figure 1B). Microscopic observation of YMC6 at 30° (semipermissive) and, in particular, at 37° (nonpermissive) revealed an accumulation of cells at a late stage of the cell cycle. This stage of the cell cycle was exemplified by an hourglass-shaped nucleus (Figure 1B). Cells grown at 30° and shifted to 37° for 4 hr displayed an increased number of large-budded cells with dumbbell-shaped nuclei and what appeared to be an extended, contiguous spindle. This cell cycle arrest phenotype is identical to that displayed by the class of M-to-G1 transition genes involved in the exit from mitosis (Zachariae and Nasmyth 1999). Mutants in these genes have been shown to arrest in telophase with well-divided DNA, an hourglass-shaped nucleus, and an extended mitotic spindle (Suranaet al. 1993). As can be seen in Figure 1C, this is also the case for the telophase-arrested cells in the YMC6 strain.

View this table:
TABLE 1

List of S. cerevisiae strains used in this study

Figure 1.

(A) Schematic of protein fusions present in the YMC6 strain. The genes for three of the known protein subunits of RNase MRP were fused to GFP, GFP:6HIS, or three copies of the HA epitope (Chuet al. 1997; gift of Lasse Lindahl, University of Maryland, Baltimore County) and used to replace the wild-type copies of the respective genes in yeast cells. The yeast strain YMC6 contains replacements of all three subunits shown simultaneously, SNM1::GFP::6HIS, POP3::GFP, and POP4::3×HA. (B) Examples of the cell cycle delay observed in the YMC6 strain. The GFP fluorescence localized to the nucleus. The YMC6 strain was grown to 106 cells/ml at 30° in YPD and cells were examined under a fluorescence microscope with a GFP filter set. Observed fluorescence is on the left and differential interference contrast is on the right. Approximately 40% of the cells display the morphology shown. (C) Simultaneous localization of GFP staining, mitotic spindle, and DNA in telophase-arrested YMC6 (see materials and methods). Clockwise from the top left is GFP fluorescence, Cy3-staining tubulin, visible light, and DNA stained with DAPI.

Since two of the tagged-protein components in the YMC6 strain are shared with RNase P, we examined whether the phenotype was the result of an RNase MRP defect alone. A strain carrying the temperature-conditional MRP RNA mutation, nme1-P6, which has a point mutation (G-to-A transition at position 122 of MRP RNA) in the NME1 gene (Schmitt and Clayton 1993; Shadelet al. 2000), was examined for a cell cycle defect. We used 4′,6-diamidino-2-phenylindole (DAPI) staining to visualize DNA and an anti-tubulin antibody to visualize the mitotic spindle allowing for easy cell cycle staging. To a greater extent than the YMC6 strain, cells built up with large budded cells, a long mitotic spindle, and divided DNA (Figure 2). In addition, a strain carrying the temperature-sensitive mutation in the SNM1 gene, snm1-p18, displayed a similar phenotype (Figure 2 and Caiet al. 1999). Snm1 is a protein component RNase MRP and not a component of RNase P (Schmitt and Clayton 1994; Chamberlainet al. 1998). Random fields of 200 cells were scored for their stage in the cell cycle. In the wild type we found that ~9.5% of cells were in telophase at any one time, with long spindles and divided DNA. The strain with the nme1-P6 mutation and the YMC6 strain both have a much higher percentage of cells in telophase even at the semipermissive temperature of 30°, 71 and 39%, respectively. These results indicate a cell cycle delay as opposed to a 100% arrest.

Plasmid missegregation in RNase MRP RNA mutants: It has been demonstrated that mutations in the Snm1 protein will cause a plasmid segregation problem (Caiet al. 1999). This missegregation may be caused by a prolonged cell cycle arrest at telophase caused by the loss of RNase MRP function. In addition, other EFM mutants also have plasmid segregation problems (Hardy and Pautz 1996; Shou and Deshaies 2002). Because of this we tested plasmid segregation in a strain with the nme1-P6 mutation. The reporter plasmid pTC185, which contains a wild-type ADE2 gene in the CEN/URA3 vector pRS316 (Caiet al. 1999), was transformed into both a wild-type and nme1-P6 strain. Because of a ade2-1 mutation on the chromosome, these strains normally form red colonies on YPD plates. However, after having been transformed with the plasmid pTC185 containing the ADE2 gene, they form white colonies. Loss of the plasmid during colony growth produces red sectors in white colonies. If there is a segregation problem, an increase in the number of red sectors is expected.

Figure 2.

Immunolocalization of mitotic spindles in the strain YMC6 and RNase MRP RNA and protein mutants. Yeast strains were grown to 2 × 106 cells/ml at 30° in YPD. The snm1-P18 strain was shifted to 37° for an additional 3 hr before harvesting (Caiet al. 1999). Cells were fixed and the mitotic spindle and DNA were localized (see materials and methods). Cy3-staining tubulin is shown. NME1(wild type), MES111-140; nme1-P6, MES111-P6; snm1-p18, YTC150-p18.

The results are shown in Figure 3. Most wild-type cells are white, although a small number of red colonies are observed due to plasmid loss during the nonselective growth. A few sectoring colonies are found, most of which have fewer than two small sectors. In the nme1-P6 strain a large increase in plasmid missegregation is observed, as indicated by the increase in the number of colonies with multiple sectors. The sectoring results demonstrate that defects in the MRP RNA can cause plasmid missegregation. On the basis of our cell cycle delay findings, it is predicted that the missegregation is the result of a prolonged period of telophase in RNase MRP mutants even at the permissive temperature.

CDC5 was identified as a suppressor of the snm1-172 mutation: snm1-172 is a temperature-sensitive, site-directed mutation (Cys64 to Ala) of the unique RNase MRP protein component Snm1p. A strain carrying the snm1-172 mutation is temperature sensitive and shows a 5.8S rRNA processing defect characteristic of RNase MRP mutants (Caiet al. 1999). Because this mutant both contained a single missense mutation and displayed strong temperature sensitivity, we used it to perform a high-copy suppressor search for new proteins involved in MRP function. A yeast 2μ library was transformed into the snm1-172 strain, and library transformants were selected on uracil minus medium, replicated, and grown at 37° for 3 days. A single suppressor clone was identified four times independently. The suppressor was retested, sequenced, and found to contain a 9.3-kb genomic DNA fragment of chromosome XIII (Figure 4A). The insert was subcloned and subjected to deletion analysis. A shorter fragment containing the CDC5 gene and CEN13 (Sph1Δ) gave the same strong suppression as the original clone; however, the CDC5 gene alone gave only weak suppression of the snm1-172 mutation. It is known that high-level overexpression of CDC5 will result in growth arrest with a nonuniform terminal phenotype in yeast cells (Charleset al. 1998), and the presence of the CEN13 may reduce the copy number of the 2μ plasmid to a lower level and provide better suppression. To confirm that CDC5 is the suppressor gene, a fragment in the CDC5 open reading frame was removed in the Sph1Δ clone and this construct was tested. It did not suppress the snm1-172 mutation nor did CEN13 alone (Figure 4B). Cdc5, a polo-like kinase, is a high-copy suppressor of many EFM mutants, including cdc15, tem1, and dbf2 (Kitadaet al. 1993; Jaspersenet al. 1998). Overexpression of CDC5 has been shown to promote the destruction of Clb2, the B-type cyclin whose proteolysis allows progression from anaphase to G1 (Charleset al. 1998).

Figure 3.

Plasmid segregation in a RNase MRP RNA mutant. The nme1-P6 mutant yeast strain was transformed with a plasmid carrying a wild-type ADE2 gene so that plasmid stability could be monitored by a simple color assay (see materials and methods; Caiet al. 1999). Pictures were taken after 3 days growth at 30° (a semipermissive temperature for the nme1-P6 mutant). NME1 (wild type), MES111-140 + pTC185; nme1-P6, MES111-P6 + pTC185.

A strain carrying the snm1-172 mutation has previously been shown to have a distinct rRNA processing defect, and the small-to-large 5.8S rRNA ratio was changed from the normal 10:1 ratio (Caiet al. 1999). We examined rRNA processing when the snm1-172 mutation is being suppressed by a high-copy CDC5 gene. Identical to the strain without the suppressor, the 5.8S rRNA ratio was altered, and an aberrant precursor (5.8Sb; Schmitt and Clayton 1993) was observed (Figure 4C). The amounts of the individual bands were quantitated on a BioRad Fluor-S MultiImager using Bio-Rad's Quantity One quantitation software, with the 5S rRNA as a loading standard, and found to have no significant difference between the snm1-172 mutant and the snm1-172 carrying the high-copy CDC5. These results indicate that CDC5 suppresses the mutation through an alternate pathway from the known rRNA processing activity. It also suggests that the temperature-sensitive growth arrest caused by the snm1-172 mutation is not the result of defective rRNA processing.

Figure 4.

CDC5 is a high-copy suppressor of the snm1-172 mutation. (A) Map of the genomic DNA insert found as a high-copy suppressor of the snm1-172 mutation. The snm1-172 mutation is a strong temperature-sensitive point mutation in the SNM1 gene (Caiet al. 1999). A gene map of the original suppressing clone is shown. The lines below the map represent fragments of this region that were tested for suppression of the snm1-172 mutation when in the yeast 2μ/URA3 vector YEP352. The ability of these clones to suppress is indicated by a (−), no suppression; (+), partial suppression; or (+++), suppression at 37°. (B) Suppressor clone with a deletion in CDC5 coding region cannot suppress the snm1-172 mutation. Yeast cells were spread onto SCD plates at 37° and grown for 24 hr before photographing. Yeast strain YTC150-172, which carries the snm1-172 mutation, was tested with an empty YEP352 vector (YEP352), the CDC5 gene (Sph1Δ, in Figure 4A) on YEP352 or YEP13, or a deletion in the CDC5 coding region (YEP13 [cdc5Δ1]). (C) The CDC5 suppressor restores cell growth but has no effect on the defect in ribosomal RNA processing in the snm1-172 mutant. Yeast strains were grown to 2 × 107 cells/ml at 24° in YPD and then shifted to 37° for 4 hr. Total RNA was isolated, and equal amounts were separated on a 6% acrylamide/7 m urea gel and stained with ethidium bromide (Schmittet al. 1990). The locations of the relevant 5.8S, small and large, and 5.8Sb rRNAs are indicated (Schmitt and Clayton 1993). Strains are the same as in B; SNM1, YTC150.

Increase of Clb2 protein level in an RNase MRP RNA mutant: The EFM mutants or M-to-G1 mutant group (Hartwellet al. 1974) all show higher mitotic cyclin levels and a resultant increased cyclin kinase activity (Jaspersenet al. 1998). Clb2, the major mitotic cyclin in yeast at this stage in the cell cycle, has increased protein levels (two- to fourfold) in these mutants. To examine the Clb2 protein level, we replaced the CLB2 gene in our strain background with a 3× HA-tagged CLB2 gene (gift of R. Hallberg, Syracuse University; Shuet al. 1997). This allowed us to examine Clb2 protein using an anti-HA antibody (see Figure 5). In asynchronous cells the Clb2 protein level increased about two- to fourfold (as quantitated by densitometry) in the nme1-P6 mutant as compared to the wild-type control. However, in hydroxyurea and nocodazole (drugs that arrest cells in S phase and M phase, respectively) Clb2 protein levels show no significant difference between wild-type and mutant cells, similar to the other EFM mutants. These changes were reproducible in six separate experiments and are a characteristic of most EFM mutants (Jaspersenet al. 1998). This increase in B-cyclin protein confirms that cells are delaying at the end of mitosis. The cell cycle delay found in EFM mutants is due to accumulation of B-cyclin protein in telophase. It is removal of this B-cyclin that is required to allow cells to exit from mitosis.

Figure 5.

Western analysis of Clb2 protein in a RNase MRP RNA mutant. Wild-type (NME1, YTC221) and nme1-P6 (YTC222) mutant strains were grown to 106 cells/ml at 25° in SCD. Cultures were then either directly shifted to 37° for 3 hr (AS) or first arrested in hydroxyurea (HU) or nocodazole (NZ) for 3 hr followed by the shift to 37° (see materials and methods). Cells were harvested and whole-cell protein extracts were made, and Western analysis was performed. The HA-tagged Clb2 was detected with an anti-HA monoclonal antibody, followed by a mouse anti-Vma1 antibody as a loading control (gift of Dr. P. Kane, SUNY Upstate Medical University). The Clb2p band is indicated.

Figure 6.

Immunolocalization of Cdc14p in a RNase MRP mutant. A yeast strain carrying the nme1-P6 mutation (YTC242) was grown to 2 × 106 cells/ml at 24° in YPD and shifted to 37° for 4 hr. The mitotic spindle was visualized using a TUB3-GFP gene that allowed easy cell cycle staging. Cdc14 was localized using a CDC14-GFP gene. Nucleolar staining of Cdc14p in nontelophase cells is indicated by the white arrows. Cells in telophase (which predominate in the mutant) are identified by the long mitotic spindle and diffuse cdc14p throughout the cell (gray arrows). (Left) Phase contrast; (right) GFP fluorescence.

Cdc14 localization was unaffected in a RNase MRP RNA mutant: Cdc14 is a protein phosphatase that activates the degradation of Clb2 protein by dephosphorylation of the Hct1/Cdh1 protein, allowing it to bind and activate Clb2-directed APC/C degradation (Visintinet al. 1998). Cdc14 is sequestered in the nucleolus for most of the cell cycle and released from nucleolus during anaphase (Shouet al. 1999; Visintinet al. 1999). RNase MRP is also located in the nucleolus where it processes ribosomal RNA precursors (Reimeret al. 1988; Schmitt and Clayton 1993). We examined the possibility that a defect in RNase MRP could affect release of Cdc14 from the nucleolus and in turn affect exit from mitosis. We followed the Cdc14 protein by fusing the coding region for the GFP to the 3′ end of the endogenous CDC14 gene (Longtineet al. 1998). This fusion gene was used to replace the wild-type gene. Cells were grown at permissive temperature to mid-log phase and then shifted to nonpermissive temperature for the nme1-P6 mutation for 4 hr. Cdc14 localization was examined by fluorescence microscopy. To easily stage cells that were in telophase, we visualized the mitotic spindles by transforming cells with a tubulin-GFP plasmid (gift of D. Amberg, SUNY Upstate Medical University). In both wild-type and mutant cells, Cdc14 was localized in the nucleolus during most of the cell cycle (Figure 6, white arrows) and was released from the nucleolus late in telophase (Figure 6, gray arrows). This indicated that the cell cycle delay of a mutation in RNase MRP is not caused by a failure to mobilize the Cdc14 phosphatase.

Increase of CLB2 mRNA levels in an RNase MRP RNA mutant: RNase MRP is a well-characterized endoribonuclease (Chang and Clayton 1987). We predict that the function of RNase MRP in the cell cycle involves the processing of some cellular RNA. We examined whether the CLB2 mRNA level might be altered in a strain carrying the nme1-P6 mutation (MES111-P6). Northern analysis was performed on asynchronous and drug-arrested populations of cells. As shown in Figure 7, in asynchronous cultures the CLB2 mRNA level is three- to fourfold higher (as quantitated on a phosphoimager) in mutant cells than in wild-type cells. CLB2 mRNA levels are not significantly different in either hydroxyurea- or nocodazole-arrested cells. This result is very similar to the changes that were seen at the protein level, suggesting that the increased mRNA level is the direct cause of the increased Clb2 protein levels leading to the block in cell cycle progression. It should be noted that this result is in contrast to a cdc15 mutant (exit from mitosis mutant) that has been shown to have very low or undetectable levels of the CLB2 mRNA when arrested at the nonpermissive temperature (Fitchet al. 1992; Spellmanet al. 1998).

Figure 7.

Northern analysis of CLB2 mRNA in nme1-P6 cells. Wild-type (MES111-140) and nme1-P6 (MES111-P6) mutant strains were prepared as in Figure 5 (see materials and methods). Total RNA was isolated and equal amounts of total RNA were subjected to Northern analysis. Location of the CLB2 transcript is indicated. AS, asynchronous cells; HU, hydroxyurea-synchronized cells; NZ, nocodazole-synchronized cells.

Genetic interactions of RNase MRP mutants with EFM mutants and mitotic cyclins: Because of the phenotypic resemblance of RNase MRP mutants to EFM mutants and the high-copy suppression with CDC5 we investigated various genetic interactions between mutations in EFM genes and mutations in RNase MRP genes. Temperature-sensitive mutations in cdc5, -14, and -15 and deletions of clb1, -2, and -5 were combined individually with RNase MRP mutants. These analyses of RNase MRP and EFM double mutants produced a number of genetic interactions. Double mutants were generated with four separate MRP mutants, nme1-P6, nme12, snm1-P18, and snm1-172 (Caiet al. 1999; Shadelet al. 2000). All four of these mutants are temperature sensitive for growth at 37°. The nme1-P6 and snm1-P18 mutants are strong mutants in terms of slow growth and defects in rRNA processing. The nme12 and the snm1-172 are milder mutants in that they have normal growth rates at permissive temperatures and moderate rRNA processing defects. CLB1 and CLB2 are functionally redundant cyclins; however, CLB2 is the major cyclin of this pair (Clb2p is at twice the level of Clb1p; Crosset al. 2002). CLB5 has partially overlapping functions with CLB1 and CLB2 but is expressed in an earlier pattern in the cell cycle. Alone none of the clbΔs display a growth phenotype except the clb2Δ that is mildly temperature sensitive.

The clb1Δ was found to partially suppress all of the mutants tested except the snm1-P18 that had a worsening of its growth defect. The clb2Δ was found to cause all the mutants to grow much more slowly, with the nme1-P6/clb2Δ and the snm1-P18/clb2Δ strains growing extremely slowly and at only 25° (see Figure 8). The two weaker mutations demonstrated an increase in temperature sensitivity when combined with the clb2Δ. The clb5Δ was able to dramatically suppress the two stronger mutants.

Crosses between RNase MRP mutants and the EFM mutants cdc14 and cdc15 yielded synthetic interactions only with the snm1-p18 mutant. A weak suppression of cdc15-1 temperature sensitivity by mutations in snm1 (Figure 9, right panel) was also seen. On the basis of our previously identified suppression of an RNase MRP mutant by multicopy CDC5, we expected an increased temperature sensitivity in the cdc5/RNase MRP double mutants. This result was borne out in all four RNase MRP mutants tested (see Figure 9). The extreme case was in the snm1-p18 mutation that could not be combined with the cdc5-1 mutation. Taken together, RNase MRP mutants exhibit a high degree of genetic interaction with EFM mutants.

DISCUSSION

RNase MRP mutations and exit from mitosis: The M-to-G1 cell cycle delay in RNase MRP mutants was originally identified in the YMC6 strain that had a slow growth phenotype and displayed cells arrested with large buds, well-divided nuclei, and extended contiguous spindles. We demonstrate that temperature-sensitive mutations in both the MRP RNA component and the unique protein component (Schmitt and Clayton 1993; Caiet al. 1999; Shadelet al. 2000) have an accumulation of cells in telophase. In addition, we identified a plasmid missegregation defect that has also been reported in other EFM mutants (Hardy and Pautz 1996; Shou and Deshaies 2002). The plasmid missegregation is most likely the result of the long delay of cells in telophase (Caiet al. 1999). Together these results suggest that the cell cycle phenotype is due to the loss of RNase MRP function, as opposed to the loss of individual enzyme components, and that RNase MRP enzymatic activity may be an important component for EFM. In our strongest temperature-sensitive RNase MRP RNA mutant, we saw an accumulation of >70% of the cells in telophase at the nonpermissive temperature. Because we did not see all of the cells in the telophase arrest, we consider this a cell cycle delay as opposed to an arrest.

Examination of the RNase MRP RNA mutant revealed that Clb2 protein accumulates in this strain. In wild-type cells, Clb2 protein is expressed in late S phase and degrades rapidly after anaphase. Destruction of Clb2 cyclin plays an important role in cyclin-dependent kinase (Cdk) inactivation and exit from mitosis (Visintinet al. 1998). Elevated Clb2 protein levels are a ubiquitous characteristic of EFM mutants and excess Clb2p promotes accumulation of cells at telophase. This is also true of RNase MRP mutants and indicates that MRP mutants accumulate at a similar cell cycle stage by a similar mechanism as other EFM mutants. Furthermore, we have shown that the accumulation of Clb2p is a result of an increase in steady-state levels of CLB2 mRNA. CLB2 transcripts begin to accumulate late in S phase, remain elevated until late in mitosis, and are degraded rapidly as cells complete mitosis (Fitchet al. 1992). This suggests that CLB2 mRNA levels play an important role in control of Clb2-CDK activity. Transcription plays a major part in the control of CLB2 mRNA levels in the cell cycle. However, little has been reported about how CLB2 mRNA is degraded, even though degradation of the mRNA is essential for tight control of the mRNA levels. It is noteworthy that cdc15 mutants arrested in telophase have been reported to show very low levels of the CLB2 mRNA (Fitchet al. 1992); this is in contrast to what we observe in RNase MRP mutants. On the basis of these observations it is clearly worth investigating if RNase MRP plays a role in degrading B-cyclin mRNAs.

Cdc14, a protein phosphatase and a EFM mutant, is found inactively sequestered in the nucleolus during most of the cell cycle and released in telophase to spread through the cell and dephosphorylate Swi5, Sic1, and Hct1/Cdh1. Dephosphorylation of these substrates is required to degrade the Clb2 protein and inactivate the kinase. Since RNase MRP is located in the nucleolus where it processes rRNAs (Schmitt and Clayton 1993), we examined mobilization of Cdc14. We found Cdc14 release to be normal in MRP mutants, suggesting that the regulatory mechanism of Clb2 proteolysis is also functional. Proper functioning of this pathway is also consistent with the identification of CDC5 as a high-copy suppressor of an RNase MRP mutant.

Cdc5 is a rate-limiting determinant of APC/C activity and B-cyclin destruction. Overexpression of Cdc5 has been shown to lead to increased destruction of Clb2 protein and suppression of other EFM mutants (Shirayamaet al. 1998). CDC5 rescues the snm1-172 mutation growth defect, without restoring the rRNA processing defect. This finding indicates that the cell cycle delay caused by RNase MRP mutations is the result of the accumulated Clb2 cyclin and not defective rRNA processing. This is consistent with the view that rRNA processing is not the essential function of RNase MRP and an additional function is essential (Henryet al. 1994; Venema and Tollervey 1999).

Figure 8.

Growth phenotypes of double RNase MRP and B-cyclin mutants. Consecutive 10-fold dilutions of each indicated strain were made in a microtiter dish, spot plated (using a 48-pin replica-plating device) onto YPD plates, and grown at the indicated temperature. The images are a composite of several plates and the experiment was repeated three times to ensure reproducibility. The full genotypes of strains are provided in Table 1.

We were able to identify a number of genetic interactions between RNase MRP mutations and both EFM mutations and B-cyclin deletions. On the basis of the phenotypes of these double mutants there is a clear role for RNase MRP in the pathway leading to exit from mitosis. The phenotypes of the double mutants indicate that RNase MRP is directly involved in modulating B-cyclin levels but does not place MRP in that pathway.

Cdc14p was found to be released normally from the nucleolus in a RNase MRP mutant. Since this is one of the last steps in the known mitotic exit pathway, RNase MRP must be acting at a stage downstream from this point or in a parallel pathway that effects B-cyclin accumulation. In S. cerevisiae, Clb1 and Clb2 are an important functionally redundant pair of B-type cyclins required for initiation and completion of mitosis. Clb2 is the predominant cyclin of this pair. Increased levels of Clb1/2 proteins correlate with associated CDK activity, which peaks just before and disappears immediately following anaphase. Suppression of RNase MRP mutants by the clb1Δ and clb5Δ indicates that a mild reduction in cyclin levels may be enough to pass through the MRP block. The increase in temperature sensitivity found with the RNase MRP mutants and the clb2Δ or the EFM mutations indicates that a drastic reduction or increase in B-cyclin levels results in an exacerbation of the MRP block.

Figure 9.

Growth phenotypes of double RNase MRP and EFM mutants. Consecutive 10-fold dilutions of each indicated strain were made in a microtiter dish, spot plated (using a 48-pin replica-plating device) onto YPD plates, and grown at the indicated temperature. The images are a composite of several plates and the experiment was repeated three times to ensure reproducibility. The full genotypes of strains are provided in Table 1.

RNase MRP and cartilage hair hypoplasia: Mutations in the RNA component of the human RNase MRP have been shown to cause a pleiotropic disease, cartilage hair hypoplasia (Ridanpääet al. 2001). This disease is manifested by short stature, sparse and brittle hair, and a compromised immune system (Mäkitie et al. 1995, 1998; Clayton 2001). The underlying problem of all of these phenotypes is a common cell proliferation defect (Pierce and Polomar 1982; Juvonenet al. 1995). On the basis of conservation to the yeast RNase MRP system, it is tempting to speculate that human RNase MRP is playing a similar role. We are currently investigating this possibility.

Acknowledgments

We thank M. Hale, P. Kane, and D. Amberg for comments and helpful discussions during the preparation of this manuscript. We are grateful to P. Kane, SUNY Upstate Medical University, for the Vma1 antibody and D. Amberg, SUNY Upstate Medical University, D. Hallberg, Syracuse University, D. Botstein, Stanford University, and L. Lindahl, University of Maryland Baltimore County, for strains and plasmids. This work was supported by grant no. RPG-96-109 from the American Cancer Society.

Footnotes

  • Communicating editor: S. Sandmeyer

  • Received December 12, 2001.
  • Accepted April 25, 2002.

LITERATURE CITED

View Abstract