A conditional lethal mutant in the fission yeast 26 S protease subunit mts3+ is defective in metaphase to anaphase transition.

We have isolated a conditional lethal mutant mts3 in the fission yeast Schizosaccharomyces pombe which at the permissive temperature is resistant to the mitotic poison MBC and at the restrictive temperature is defective in metaphase to anaphase transition. The predicted amino acid sequence of mts3 is 36% identical with the budding yeast gene NIN1. NIN1 cloned into a fission yeast expression vector can rescue both mts3 temperature-sensitive and null alleles demonstrating that NIN1 is the budding yeast homologue of the fission yeast mts3 gene. The phenotype of the mts3 null is identical with the mts3 ts mutant demonstrating that the phenotype of the mts3 ts mutant is due to loss of mts3 function. The deduced amino acid sequences of both mts3 and NIN1 show homology to peptide sequences obtained from subunit 14 of the 26 S protease purified from bovine or human cells.

We have isolated a conditional lethal mutant mts3 in the fission yeast Schizosaccharomyces pombe which at the permissive temperature is resistant to the mitotic poison MBC and at the restrictive temperature is defective in metaphase to anaphase transition. The predicted amino acid sequence of mts3 ؉ is 36% identical with the budding yeast gene NIN1. NIN1 cloned into a fission yeast expression vector can rescue both mts3 temperature-sensitive and null alleles demonstrating that NIN1 is the budding yeast homologue of the fission yeast mts3 ؉ gene. The phenotype of the mts3 null is identical with the mts3 ts mutant demonstrating that the phenotype of the mts3 ts mutant is due to loss of mts3 ؉ function. The deduced amino acid sequences of both mts3 ؉ and NIN1 show homology to peptide sequences obtained from subunit 14 of the 26 S protease purified from bovine or human cells.
It is becoming increasingly recognized that the stability of proteins is an important factor in the regulation of gene expression. The main nonlysosomal intracellular proteolysis pathway in the cell is the ubiquitin pathway. The ubiquitin pathway has been implicated in the instability of a number of important regulatory proteins such as p53, c-Myc, c-Mos, and the mitotic cyclins. Ubiquitin is a 76-amino acid protein, that when covalently attached to a lysine residue of a protein, targets it for destruction (reviewed in Refs. [1][2][3]. Biochemical analysis has determined that a multiprotein complex called the 26 S protease degrades proteins that have been targeted for destruction by the ubiquitin pathway. The 26 S protease is made up of two multiprotein functional components, the catalytic 20 S proteasome and the regulatory complex. The 20 S proteasome is a cylindrical structure composed of 14 different protein subunits and contains all the proteolytic activities of the complex. In vitro analysis on purified 20 S complexes has shown that proteins are degraded in a completely unregulated manner. Regulated proteolysis requires the addition of another multiprotein complex composed of at least another additional 15 different subunits to each end of the 20 S proteasome cylinder to form the 26 S protease. The different subunits of this regulatory complex are named according to their size on an SDS-polyacrylamide gel electrophoresis gel, S1 being the largest and S15 the smallest. With the formation of the 26 S protease, degradation of proteins is now carried out in a highly regulated manner. Certain substrates are degraded only if they have been polyubiquitinated and proteolysis is now ATP-dependent (4,5).
Recently, we isolated a conditional lethal mutant, mts2, in subunit 4 of the 26 S protease regulatory complex in the fission yeast Schizosaccharomyces pombe. Characterization of the phenotype of mts2 strain at the restrictive temperature demonstrated that the cells arrested at a particular point in the cell cycle, the metaphase stage of mitosis (6). In this paper, we describe another conditional mutant, mts3, which was isolated in the same screen. We show that mts3-1 arrests with a phenotype similar to the mts2 mutant and that the S. cerevisiae NIN1 gene is the budding yeast homologue of the mts3 gene and that both the mts3 ϩ and NIN1 gene products are similar to subunit 14 of the regulatory complex of the 26 S protease purified from mammalian cells.

EXPERIMENTAL PROCEDURES
Strains, Media, and Genetic Methods-The S. pombe strains described in this paper were all derived from the 972h Ϫ and 975h ϩ wild type heterothallic strains: (8). Media and general methods for handling the S. pombe strains are as described (7).
Plasmids-Two libraries were used in this work. An S. pombe Sau3a wild type genomic library cloned into the vector pSPARS305 (9) originally constructed in David Beach's laboratory (obtained from Robin Allshire) and an S. pombe wild type cDNA library cloned into the S. pombe expression vector pREP1 (a gift from Chris Norbury). The plasmid pREP1 (10) or pSP1 (6) contains the thiamine-repressible S. pombe nmt1 ϩ promoter. The shuttle plasmid pSPARS(305) was used for routine cloning in S. pombe and the integrating vector pRS305 (9) for integrating cloned S. pombe DNA.
Fluorescence Microscopy-DAPI 1 staining was as described in Ref. 11 and calcofluor staining in Ref. 6. For immunofluoresence microscopy, cells were grown up overnight in YPD (0.5% yeast extract, 0.5% peptone, and 2% glucose per liter of water), and, immediately before fixation, sorbitol was added directly to the culture to give a final concentration of 1.2 M. The culture was then fixed in 38% paraformaldehyde at 32°C, or 36°C for ts mutants, for 30 min. The cells were then prepared as described (12). The anti-tubulin monoclonal antibody TAT1 was used to stain microtubules (13). Stained cells were observed on a Zeiss Axioplan microscope. Images were processed using IPLab software and Digital Scientific Smart capture. Pictures were taken directly from the VDU screen on Fuji daylight 100 film.
Kinase Assay-The kinase assay was carried out as detailed (14), using p13 beads (Oncogenic Science) to precipitate the Cdc2-Cdc13 complex.
Western Analysis-Western analysis was carried out as detailed (6). The Cdc13 antibody used to probe the blot was obtained from Dr. Iain Hagan (School of Biological Sciences, University of Manchester) and used at a 1/200 dilution.
Flow Cytometry-Flow cytometry was carried out using a Becton Dickinson FACScan using the method described (15).
Isolation of the Budding Yeast NIN1 Gene-The NIN1 gene was amplified by PCR from S. cerevisiae genomic DNA as a SalI/PstI fragment using the primers NIN1-5Ј (GCAGTCGACATGCCCTCGTTAGC-CGAATTGACC) and NIN1-3Ј (GCACTGCAGACACAATATTTTCAA-TACTTA). This fragment was then cloned into the S. pombe expression vector pSP1 to give the plasmid pNIN1 Gene Disruption-The mts3 ϩ gene was disrupted by generating two BamHI/XbaI segments of the mts3 ϩ coding sequence by PCR using S. pombe genomic DNA. When the two XbaI sites were ligated together, the resulting BamHI fragment encoded the mts3 gene with 400 base pairs of its coding sequence deleted. The primers used were: GCAG-GATCCGGGAGCTGCAAGTTCTATTC(C/G)CATCTAGACCTCCCA-CTGCTAGCAAGC and GCAGGATCCGTTACAATTGCTTAAGT(G/G)-CGTTCTAGAGAAATAACCC. The two primers to the mts3 ϩ gene are shown in Fig. 5A, and the other two primers contain sequences 400 base pairs 3Ј to the mts3 ϩ stop codon and 164 base pairs 5Ј of the mts3 ϩ gene start codon. The S. pombe ura4 ϩ gene was inserted into the XbaI site.The BamHI fragment was cut out, gel-purified, and used to transform a mts3 ϩ /mts3 ϩ ade6.216/ade6.210leu1.32/leu1.32ura4.D18/ ura4.D18h ϩ /h Ϫ sporulating diploid strain selecting for uracil prototrophy. Stable ura ϩ transformants were selected. PCR analysis was used to demonstrate that the ura4 ϩ gene had disrupted the mts3 ϩ locus. Primers to the ura4 ϩ gene were used in conjunction with an oligonucleotide to the mts3 ϩ genomic sequence just upstream of the original primer used to amplify the mts3 deletion. A fragment of the correct size was generated with DNA prepared from the heterozygous mts3 ϩ / mts3⌬ura4 ϩ but not from the homozygous mts3 ϩ /mts3 ϩ diploid (data not shown).
Spore Germination of the mts3 ϩ /mts3⌬ura4 ϩ Diploid-Spore germination was carried out as described in Ref. 16. Sporulation was carried out in malt extract broth at 25°C. Asci formed after 2-3 days of incubation. These asci were then digested with Glusulase (Sigma) overnight at 25°C. When all the vegetative cells had lysed, the spores were harvested by centrifugation, washed once in water, and resuspended in 5 ml of water. The spores were then layered on top of 25 ml of 40% glucose in a 50 ϫ 50 ml tube. The tube was spun at 1600 rpm. The pelleted spores were then washed once in water and resuspended in water. The spore concentration was determined by counting on a hemocytometer. 4 ϫ 10 6 spores were inoculated into minimal media supplemented with leucine and adenine and incubated at 20°C for 10 h. The spores were transferred to 32°C and sampled at the times shown.

RESULTS
Isolation of the mts3 Mutant-The mts3 mutant was isolated in a screen to isolate mutants that were both resistant to the microtubule destabilizing drug MBC and also temperaturesensitive for growth. One mutant allele mts3-1 was isolated in the screen. Crossing the mts3-1 mutant strain to wild type cells of the opposite mating type showed that the temperature sensitivity segregated 2ts ϩ :2ts Ϫ in tetrads demonstrating that a single gene mutation was responsible for the conditional lethality. When these segregants were tested for MBC R , it was found that MBC R cosegregated with the temperature sensitivity demonstrating that the same mutation was responsible for both the MBC R and the conditional lethality. Construction of heterozygous mts3 ϩ /mts3-1ade6-M210/ade6-M216 leu1.32/leu1.32 ura4-D18/ura4-D18h ϩ /h Ϫ diploids demonstrated that the temperature-sensitive phenotype of the mts3-1 mutation was recessive to wild type.
Characterization of the mts3 Mutant Phenotype-A detailed investigation of the mts3 mutant phenotype was carried out. An early log phase mts3-1leu1.32h Ϫ culture was shifted from the permissive to the restrictive temperature and sampled every hour. Cell number was determined by counting cells on a Coulter counter (Fig. 1). This demonstrates that the mts3-1 strain arrests within the first cell cycle after shift to the restrictive temperature. mts3-1 cells grown at 25°C continued to grow exponentially at a rate similar to wild type cells.
mts3-1 cells were also fixed for cytological examination as detailed under "Experimental Procedures." The cells were stained with the DNA-specific stain DAPI and with the antitubulin antibody TAT1 to investigate the state of the DNA and microtubules. The DNA and microtubule structures seen in wild type cells are shown in Fig. 2a. After shift to the restrictive temperature, an increasing proportion of the population displayed a characteristic mutant phenotype in that the DNA becomes highly condensed (Fig. 2b, B). Associated with this condensed DNA is a short mitotic spindle (Fig. 2b, A). The condensed DNA is found to be located in the middle of the short spindle ( Fig. 2b, C) consistent with these cells being at the metaphase stage of mitosis. After 4 h at the restrictive temperature, up to 75% of cells in the population are displaying this phenotype (Fig. 2c). On cells sampled at later time points, the DNA was never found to separate nor the short spindle elongate to a length typical of anaphase in S. pombe (Fig. 2a, C). The mts3-1 mutant therefore appears to be defective in the metaphase to anaphase transition. This metaphase arrest phenotype was found to be transient in that after further incubation at the restrictive temperature the short metaphase spindle was found to disassemble and the DNA was found to decondense to reform a nucleus of normal appearance. In most cases, the single nucleus became displaced to one end of the cell and a septum formed to divide the cell into an anucleate and nucleate half (Fig. 2b, E). This phenotype is very similar to that shown by the S. pombe mutant mts2-1 isolated in the same screen (6). The only difference between the phenotypes displayed at the restrictive temperature is that a much higher proportion (75%) of the mts3-1 strain showed a metaphase arrest compared to 25% of cells in the mts2-1 culture. The septa that formed appeared aberrant when stained with the septum specific stain calcofluor and did not appear to be functional as cell number was not observed to increase (Fig. 1). Cytoplasmic microtubules were found to reform after the metaphase spindle disassembled (Fig. 2b, D). These microtubules, however, stained far less efficiently with the anti-tubulin antibody than wild type cells, implying that they could be defective. Interestingly, after extended incubation at the restrictive temperature, the culture did not go through another round of chromosome condensation as has been reported for some of the S. pombe cut mitotic mutants (17).
Histone H1 kinase assay on an mts3 asynchronous culture shifted from the permissive to the restrictive temperature was performed (Fig. 3A, a) and indicated that H1 kinase activity increases after shift to the restrictive temperature. Quantita-FIG. 1. First cycle arrest of mts3-1 strain after shift from the permissive to the restrictive temperature. Two early log phase wild type and mts3-1 cultures were grown at the permissive temperature of 25°C. At time 0 min, one culture was kept at 25°C while the remaining culture was shifted to the restrictive temperature of 35°C. The cultures were sampled at the times shown, and the cell number was measured using a Coulter counter and plotted on a logarithmic scale. mts3 ϩ Gene Encodes Fission Yeast 26 S Protease Subunit tion on the PhosphorImager shows that activity peaks after a 2-h incubation at the restrictive temperature displaying approximately 3-fold more activity than cells growing at the permissive temperature. The increase in chromosome condensation in samples taken from the same culture occurs later, at 4 h (Fig. 2c), than the peak in H1 kinase activity. After the peak in activity, the level of H1 kinase activity decreased to a level that was much lower than the original value before shift to the restrictive temperature.
Recently, Schwob et al. (18) showed that the budding yeast SIC1 gene encodes a specific inhibitor of the CLB5/CDC28 kinase and that degradation of SIC1 is required to release the CLB5/CDC28 activity essential for G 1 /S phase transition (18). In addition, the degradation of SIC1 protein was shown to be by the ubiquitin pathway. To investigate if a similar inhibition event was occurring in the mts3 extracts at the restrictive temperature, mixing experiments were carried out. Increasing amounts of crude extracts prepared from cells sampled 8 h (t ϭ 8 h) after shift to the restrictive temperature were mixed with 50 g of protein extract prepared from cells before they were shifted (t ϭ 0 h). No evidence of any inhibiting activity could be found to be present in sample t ϭ 8 h (Fig. 3A, b).  A, a, at time 0 h, an early logarithmic growing culture of mts3-1 cells was shifted from 25°C to 36°C sampled at the times shown and assayed for H1 kinase activity. Lane c shows the H1 kinase activity obtained from an exponentially growing haploid wild type culture. b, mixing experiment to assay for possible inhibitor. To 50 g of extract from t ϭ 0 h, increasing amounts of extract from t ϭ 8 h were added (0, 50, 100, and 150 g) and then assayed for H1 kinase activity as before. B, steady state levels of Cdc13 protein in mts3-1 after shift to the restrictive temperature. The same extracts assayed in A, panel a, were probed by Western analysis for Cdc13 protein. 50 g of each extract was separated on an 10% SDS-polyacrylamide gel electrophoresis gel. The samples were transferred to nitrocellulose and probed with an anti-Cdc13 antibody. The antibody was detected with a rabbit monoclonal antibody conjugated to alkaline phosphatase (Promega). The alkaline phosphatase activity was detected according to the manufacturer's instructions. H1 kinase activity measures the amount of activity of the Cdc2 kinase when it is complexed to the S. pombe cyclin B homologue Cdc13. To investigate if the extracts were losing activity due to a decrease in the levels of Cdc13 protein, the extracts assayed in Fig. 3A, a, were subjected to Western blot analysis with an antibody against the S. pombe Cdc13 protein.
As shown in Fig. 3B, the level of Cdc13 protein remained essentially constant throughout the course of the experiment. The differences in H1 kinase activity that were observed could be due to differences in the level of the Cdc2 protein or alternatively to differences in post-translational modification of either the Cdc2 or Cdc13 protein.
mts3 Mutant Cells Re-replicate Their DNA at the Restrictive Temperature-An mts3-1 mutant strain was shifted from the permissive to the restrictive temperature, and cells were sampled for flow cytometry to assay their DNA content. The mts3 cells replicate their DNA from a 2n to 4n content of DNA at 8 h (Fig. 4A). At least two possibilities can explain this re-replication in the mts3-1 mutant strain at the restrictive temperature. First, mts3 is involved in the control to ensure that replication occurs only once per cell division cycle. Alternatively, the re-replication observed is an indirect effect due to a breakdown in the normal feedback control caused by the mts3 cells entering mitosis arresting at metaphase and exiting mitosis (as assayed by loss of H1 kinase activity) before chromosome segregation had occurred. To differentiate between these two possibilities, we crossed the mts3-1 mutation into a cdc25 background. cdc25 cells are temperature-sensitive for growth and at the restrictive temperature the cells arrest in the G 2 phase of the cell division cycle before mitosis (8). Therefore, the mts3cdc25 double mutant would also arrest in G 2 before mitosis, and, thus, at the restrictive temperature, no cells should FIG. 4. A, FACS analysis of mts3-1 strain shifted to the restrictive temperature. The top three panels show nitrogen (n) starved wild type haploids which have predominantly cells with a 1n DNA content (a(i)), exponentially growing wild type haploid which have a 2n DNA content (a(ii)), and N starved wild type diploid which shows a 2n and 4n peak of DNA (a(iii)).

Panels b(i), (b)(ii), and (b)(iii) shows mts2 cells, panels c(i), (c)(ii), and (c)(iii) show mts3 cells, and panels d(i), (d)(ii), and (d)(iii) show wild type cells shifted from 25°C to 36°C and sampled at 0 h (i), 4 h (ii), and 8 h (iii) after the shift.
For each sample, the 1n, 2n or 4n peaks obtained are indicated. In b, c, and d, the mts2, mts3, and wild type strains were all haploid strains growing exponentially prior to the shift. B, FACS analysis of cdc25mts3 strain shifted to the restrictive temperature. A cdc25mts3 double mutant strain was shifted to 36°C, sampled at the times shown, and subjected to FACS analysis. The single 2n peak of DNA observed is marked. mts3 ϩ Gene Encodes Fission Yeast 26 S Protease Subunit make a mitotic spindle. This would allow the re-replication phenotype to be investigated in the absence of the metaphase arrest phenotype.
At 25°C, the cdc25mts3 double mutant grew very slowly compared to the cdc25 or mts3 single mutants and appeared elongated, being 2 to 3 times longer at division. At 20°C, however, the double mutant was observed to grow normally. Consistent with this finding was the observation that cdc25mts3 cells growing at this temperature divided at the roughly same size as wild type cells. Therefore, for the cdc25mts3 double mutant, 20°C instead of 25°C was used as the permissive temperature.
An asynchronous culture of cells was shifted from the permissive to the restrictive temperature and sampled at the times shown (Fig. 4B). These cells were then fixed, stained with propidium iodide, and subjected to FACS analysis. During the course of the experiment, at the restrictive temperature, the cells elongated to give the characteristic cdc phenotype typical of cdc25 cells. Consistent with this observation, the DNA was not found to condense, implying that the cells had all arrested at the cdc25 G 2 block and no 4n peak was observed (Fig. 4B). In conclusion, the observed re-replication in the mts3-1 strain appears to be due to the indirect effect of the cells entering mitosis and not the result of deregulation of the initiation of DNA replication.
Isolation of the mts3 Gene-The mts3 ϩ gene was isolated by complementation of the temperature-sensitive mutation of the mts3-1 mutant strain using an S. pombe wild type genomic library in the autonomously replicating vector, pSPARS(305). A 3.5-kilobase HindIII fragment was found to rescue the mts3-1 ts defect. This fragment was subcloned into the integrating vector pRS(305) and the 3.5-kilobase fragment used to integrate by homology. The integrant could rescue the temperature-sensitive defect of a mts3-1 strain. The site of integration was mapped relative to the mts3 mutation. Out of 19 full tetrads, no temperature-sensitive segregants were observed, indicating that the site of integration was very closely linked to the mts3 mutation and therefore the cloned genomic fragment contained the authentic mts3 ϩ gene and not an extragenic suppressor.
The genomic clone was used to isolate a corresponding cDNA clone carried in the S. pombe expression vector pREP1. Overexpression of the mts3 ϩ cDNA in pREP1 complemented the ts mts3 mutation and resulted in no obvious phenotype. The nucleotide sequence of the mts3 ϩ cDNA was determined and is shown in Fig. 5A.
Comparison of the mts3 ϩ coding sequence with other DNA sequences in the EMBL data base revealed a substantial homology with NIN1, a known S. cerevisiae gene. The deduced amino acid sequence of mts3 ϩ is 36% identical with the NIN1 amino acid sequence (Fig. 5B). The nin1 mutant is a conditional lethal mutation in which cells at the restrictive temperature arrest in mitosis with a single nucleus and a G 2 DNA content (19). The described phenotype is very similar to that of the mts3-1 strain at the restrictive temperature. In addition, the two proteins are very similar in size.
The S. cerevisiae NIN1 Gene Can Rescue the mts3 ts and Null Alleles-To investigate if the NIN1 gene could be the S. cerevisiae homologue of mts3 ϩ , the NIN1 gene was isolated by PCR from S. cerevisiae genomic DNA and subcloned into the S. pombe expression vector pSP1 to give the plasmid pNIN1 (described under "Experimental Procedures"). This plasmid was then used to transform the mts3-1 strain to investigate if the NIN1 gene can rescue the conditional lethal phenotype of the mts3-1 ts mutation. As shown, the NIN1 gene can rescue, but the vector itself or the mts2 ϩ cDNA cloned into the same vector cannot rescue the S. pombe mts3 ts mutation (Fig. 6A). From the size of colonies obtained with a mts3-1 strain carrying the pNIN1 or pmts3 ϩ plasmid at the restrictive temperature, it can be seen that the pNIN1 gene can rescue the mts3 ts mutation very well.
Although the budding yeast NIN1 gene can rescue an mts3 ts mutant, a much more stringent test would be to ask if the NIN1 gene could rescue an mts3 deletion allele.
The cloned mts3 ϩ gene was used to make a null allele of the mts3 ϩ gene as detailed under "Experimental Procedures." The mts3 ϩ gene was deleted by PCR and replaced with the selectable ura4 ϩ marker. This construct was used to transform an S. pombe diploid strain selecting for uracil prototrophy. PCR analysis was used to confirm that integration had occurred at the mts3 locus (data not shown).
The heterozygous diploid was sporulated and tetrad analysis carried out on the resulting asci. Out of 20 tetrads, only two viable spores were obtained in each ascus. In addition, these viable spores were all uracil auxotrophs demonstrating that the mts3 ϩ gene is essential for growth.
To investigate if the NIN1 gene could also rescue a null allele, the mts3 ϩ /mts3⌬ura4 ϩ leu1.32/leu1.32 ura4D18/ ura4D18 heterozygous diploid was transformed with the pNIN1 plasmid and sporulated. By selecting simultaneously for the ura4 ϩ gene in the disrupted mts3 allele and the leucine marker carried on the pNIN1 plasmid, the ability of the NIN1 gene to rescue the mts3 null could be investigated. The NIN1 gene could rescue the mts3 null allele as well as the S. pombe mts3 ϩ gene (Fig. 6B). As a control, both the mts2 ϩ cDNA and the vector itself could not rescue the mts3 null allele. This functional complementation of a mts3 null allele combined with the similarity between the deduced amino acid sequences confirms that the NIN1 gene is the budding yeast homologue of the fission yeast mts3 ϩ gene.
Phenotype of the mts3 Null Allele-The mts3 ϩ /mts3⌬ura4 ϩ heterozygous diploid was sporulated, the vegetative cells lysed by digestion with Glusulase, and the spores used to innoculate fresh media selecting for the disrupted mts3 allele. Cells were then sampled and fixed for FACS analysis and cytology. The same metaphase arrest phenotype was seen in approximately 40% of cells (Fig. 7, A and B). As observed for the mts3 ts mutant, the metaphase arrest phenotype was transient, as on prolonged incubation, the spindle disassembled, the DNA decondensed, and a septa formed to give essentially the same phenotype as seen with the mts3 ts strain (data not shown). FACS analysis on cells sampled at the same time points showed that a proportion of cells replicated their DNA to give a 4n content of DNA (Fig. 7c). The 1n peak was due to spores containing the undisrupted allele, and this peak disappeared if uracil was added to the media inoculated with the spores (data not shown). In conclusion, the phenotype observed with the mts3 ts allele is identical with that of a mts3 null allele.
Homology of Both the mts3 and NIN1 Gene Products to Subunit 14 of the Human 26 S Protease-The mammalian 26 S protease has been recently purified to homogeneity in a number of laboratories, and the resulting purified subunits were used to obtain peptide fragments which were then sequenced (20,21). When the mts3 ϩ deduced amino acid sequence was compared to these peptide sequences, it was found that mts3 ϩ had a high degree of similarity to some of the peptide fragments obtained from subunit 14 of the 26 S protease (Fig. 8A). The NIN1 deduced amino acid sequence also shows a high degree of similarity to peptide fragments from subunit 14 (Fig. 8B). The size of the Mts3 protein is also similar to that found for the mammalian subunit 14 (20). Recent experiments have shown that the Mts3 protein is present in the 26 S protease affinitypurified from fission yeast. 2

DISCUSSION
In this paper we describe the phenotype of the mts3 mutant of S. pombe. At the restrictive temperature, mts3 cells enter mitosis and arrest at metaphase. Spindle elongation and DNA segregation characteristic of anaphase did not occur, demonstrating that the mutant is defective in the metaphase to anaphase transition. In the same original screen, we isolated another conditional lethal mutant, mts2, which encodes subunit 4 of the same regulatory complex (6). The mts2 ϩ gene product belongs to a highly conserved family of ATPases (22), while the mts3 ϩ amino acid sequence bears no homology to these ATPases nor to any other protein domains found in the EMBL data base. Comparison of the Mts3 peptide sequence to the peptide sequence obtained from tryptic digests of mammalian 26 S subunits indicated a high degree of homology to peptide sequence from subunit 14. Consistent with this observation, we have shown recently that the mts3 ϩ gene product is present in the regulatory complex of the 26 S protease purified from S.pombe. 2 The phenotype of mts3-1 at the restrictive temperature is very similar to that which we have previously described for mts2-1. Therefore, the metaphase arrest phenotype appears to be associated with lack of 26 S protease function. In addition, the demonstration that an mts3 null allele has the same metaphase arrest phenotype as the ts mutant at the restrictive temperature shows that this phenotype is a result of lack of mts3 ϩ (subunit 14) gene product function.
We have previously shown that overexpressing the mouse MSS1 gene, which encodes subunit 7 of the 26 S protease, could rescue the mts2 ts mutation but not an mts2 null mutation (6). 2 M. Seeger, K. Ferrell, C. Gordon, and W. Dubiel, manuscript in preparation.
FIG. 6. The budding yeast NIN1 gene can rescue the conditional lethality of the mts3 ts and null alleles. A, the mts3leu1.32h Ϫ and the mts2leu1.32h Ϫ strains were transformed with the S. pombe expression vector pSP1 or pSP1 carrying the mts2 ϩ (pmts2 ϩ ), mts3 ϩ (pmts3 ϩ ), or NIN1 (pNIN1) cDNAs. The top two plates show the mts3leu1.32h Ϫ strain carrying the plasmids streaked out at the permissive temperature of 25°C (left) or the restrictive temperature of 36°C (right). Only the mts3 cells carrying the plasmids containing the NIN1 or mts3 ϩ genes can grow at the restrictive temperature. The bottom two plates show the mts2leu1.32h Ϫ strain transformed with the same plasmids. In this case, only mts2 cells carrying the plasmid containing the mts2 ϩ gene can grow at the restrictive temperature. B, the mts3 ϩ / mts3⌬ura4 ϩ heterozygous diploid was transformed with pSP1, pmts2 ϩ , pmts3 ϩ , and pNIN1. The diploid was sporulated and by selecting simultaneously for the ura4 ϩ gene in the disrupted mts3 allele and the leucine marker carried on the S. pombe expression vector the ability of the different plasmids to rescue the mts3 null allele could be investigated. Only the pmts3 ϩ and pNIN1 plasmids could rescue the mts3 null allele.
This provided genetic evidence for a possible interaction between subunits 4 and 7 in the 26 S complex itself. We could find no such genetic evidence for an interaction between the mts2 ϩ (subunit 4) and mts3 ϩ (subunit 14) gene products either in the phenotype of the mts2mts3 double mutant strain (data not shown) or by rescue of the temperature sensitivity by overexpression of mts2 ϩ in a mts3-1 strain or vice versa.
When the mts3 ϩ peptide sequence was compared to other sequences in the EMBL data base, it was found to have 36% identity to the S. cerevisiae gene NIN1. The phenotype of the nin1 mutant at the restrictive temperature is similar to that described for the S. cerevisiae ts mutants cim3(MSS1) and cim5(SUG1) which encode subunits 7 and 8 of the 26 S protease (23). We show here that the S. cerevisiae NIN1 gene can rescue both S.pombe mts3 ts and null alleles demonstrating that NIN1 is the budding yeast homologue of the mts3 ϩ gene.
Surprisingly, the 36% identity found between the mts3 ϩ and NIN1 gene products is much lower than that found between the mts2 ϩ and corresponding human S4 homologue, which show 73% identity with each other. Both subunits are present in the same multiprotein complex so what model could account for the difference in divergence between different homologues of one subunit (subunit 4) compared to the other (subunit 14). Perhaps this could be a consequence of where each protein is situated in the complex itself. Subunit 4, for example, could lie in the middle of the complex in intimate association with a number of different proteins allowing little chance of divergence while subunit 14 could be located at one end of the complex interacting with fewer proteins. Whatever the reason, when homologues of subunit 14 are isolated from other organisms, this reduced degree of identity between homologues should aid in the identification of functionally important do-(filled circles). c, FACS analysis of germinated spores sampled at the times (in hours) shown. The 1n, 2n, and 4n peaks are marked. The mts2 and mts3 mutants were both isolated in a screen looking for mutants that were resistant to the mitotic poison MBC. It is not apparent why such a screen should seem to be specific for subunits of the regulatory complex of the 26 S protease. One possible explanation is that the concentrations of MBC used in the screen causes cells to make a slightly defective spindle at the permissive temperature which although it is still able to function relatively normally activates a checkpoint pathway to cause cell cycle arrest. The mutants are defective in this putative checkpoint response and are therefore able to grow at the permissive temperature in the presence of the drug. Whether such a model accounts for the MBC R found in the mutants and whether this is related to mitotic arrest at the restrictive temperature will have to wait until further experiments have been carried out.
The mts2 and mts3 mutants at the restrictive temperature both arrest at the metaphase stage of mitosis. In addition, both genes encode different subunits of the same regulatory complex of the 26 S protease. This implies that some substrate(s) has to be degraded by the ubiquitin pathway to proceed from metaphase to anaphase in mitosis. Recent experiments on frog extracts have essentially come to the same conclusion (24). It was originally thought that cyclin B was one of these substrates and that it had to be degraded for cells to proceed from metaphase to anaphase in mitosis. More recent work has disproved this hypothesis and demonstrated that cyclin B destruction is in fact required for exit from mitosis (24,25). Recent work on the cyclin B destruction pathway has implied that some components involved in this pathway also seem to be required for the destruction of the substrate(s) to allow metaphase to anaphase transition. ts mutants in three genes were isolated in S. cerevisiae that were defective in the destruction of the budding yeast cyclin B homologue, CLB2. Surprisingly, at the restrictive temperature, all three mutants appear to be defective in metaphase to anaphase transition. When the wild type genes were cloned, all three mutations turned out to be in known genes, CDC16, CDC23, and CSE1. The CDC16 and CDC23 genes encode two members of the TPR family of proteins which are characterized by blocks of 34 amino acid tandem repeats repeats known as tetratricopeptide repeats (26,27). Antibodies that recognize these proteins and an additional TPR protein called CDC27 were used to show that they are present in a large multiprotein complex purified from Xenopus extracts. This complex was isolated by its ability to attach ubiquitin to cyclin B molecules in a cell cycle-dependent manner, an event which targets the cyclin B molecule for destruction (28 -31). To explain the metaphase arrest phenotype in the cdc16 and cdc23 mutants at the restrictive temperature, it is postulated that, in addition to being required for the destruction of cyclin B, they are also required for the destruction of the putative substrate that has to be degraded for metaphase to anaphase transition to occur.
The S. pombe cut9 and nuc2 mutants are temperature-sensitive mutants whose genes encode members of this TPR family (17,31). nuc2 ϩ has the highest degree of similarity to CDC27 while cut9 ϩ is most similar to CDC16. At the restrictive tem-perature, like the mts2-1 and mts3-1 ts strains, both nuc2 and cut9 mutants are defective in metaphase to anaphase transition. Genetic evidence suggests that the nuc2 ϩ and cut9 ϩ proteins interact with each other (17). The mts2, mts3, nuc2, and cut9 mutants could therefore all define genes required in a pathway in S. pombe to degrade the putative substrate whose destruction has been postulated to be required for sister chromatid separation at the metaphase to anaphase transition in mitosis.