Accurate chromosome segregation is dependent upon the integrity of mitotic spindles, which pull each pair of sister chromatids towards opposite poles. In this study, we have characterised fission yeast pop3-5235, a diploidising mutant that is impaired in genome stability. Pop3 is the same as Wat1, a conserved protein containing 7 WD repeats. Pop3/Wat1 has also been isolated from a two-hybrid screen as a binding partner to Prp2, the large subunit of the essential splicing factor U2AF. In wat1 mutants, the cellular amount of α-tubulin is decreased to very low levels, which results in compromised microtubules and spindles, consequently leading to unequal chromosome separation. Further analysis shows that, in spite of the binding between Wat1 and Prp2, Wat1 may not be involved directly in splicing reactions per se. Instead, we find that Wat1 is required for the maintenance of α-tubulin mRNA levels; moreover, transcript levels of genes other than the α-tubulin gene are also equally decreased in this mutant. Wild-type Wat1, but not the mutant protein, forms a large complex in the cell with several other proteins, suggesting that Wat1 functions as a structural linker in the complex. The results suggest that Wat1 plays a role in mRNA maturation as a coupling protein between splicing and synthesis and/or stabilisation.
Maintenance of genome stability is one of the fundamental principals for all eukaryotic organisms. The cell has developed several regulatory networks to ensure genome stability. These include proper ordering of S and M phases (Nurse, 1994; Orr-Weaver, 1994), and the rule that restricts these events to occur only once during each cell cycle (Diffley, 1996). On top of these controls, there are also mechanism to ensure faithful sister chromatid segregation. Duplicated chromosomes are kept together via sister chromatid cohesion after chromosome duplication until metaphase, and upon loss of cohesion at anaphase, each chromatid is pulled apart towards the opposite poles by the mitotic spindle apparatus (Nasmyth et al., 2000). This process has to be highly accurate, otherwise unequal segregation of sister chromatids would result in aneuploid offspring, which can lead to cell death or tumourigenesis (Lengauer et al., 1998).
Schizosaccharomyces pombe is an ideal organism in which to study genetic control of sister chromatid segregation (Umesono et al., 1983; Yanagida, 2000). The cell usually proliferates as a haploid with ploidy being maintained faithfully. In fission yeast, aneuploidy is usually deleterious (Niwa and Yanagida, 1985; Niwa et al., 1989); however, genome instability can be detected by chromosome loss of nonessential artificial mini-chromosomes or diploidisation phenotypes (Broek et al., 1991; Takahashi et al., 1994; Kominami and Toda, 1997). Using these phenotypic markers, previous studies have identified molecules and the genetic network, which ensure genome stability. These include periodic activation and inactivation of Cdc2/cyclins (Broek et al., 1991; Hayles et al., 1994; Moreno and Nurse, 1994; Kominami and Toda, 1997), temporal expression of S-phase regulators (Nishitani and Nurse, 1995; Nishitani et al., 2000), centromere and kinetochore integrity (Allshire et al., 1995; Ekwall et al., 1995; Freeman-Cook et al., 1999; Goshima et al., 1999; Saitoh et al., 1997; Takahashi et al., 2000), sister chromatid cohesion (Furuya et al., 1998; Tomonaga et al., 2000), and chromosome architecture (Yanagida, 1998).
Maturation of pre-mRNA not only plays a housekeeping role in cell division, but also plays an important part in cell cycle progression. For example, several fission yeast mutants defective in splicing reactions show cell-cycle specific phenotypes (Lundgren et al., 1996; Burns et al., 1999; McDonald et al., 1999; Ohi et al., 1994; Ohi et al., 1998; Potashkin et al., 1998; Urushiyama et al., 1996; Urushiyama et al., 1997; Beales et al., 2000), although the reason for this remains unknown. The spliceosome consists of a large complex comprising some 80 components. One of the first splicing factors to interact with the pre-mRNA is U2 auxiliary factor (U2AF). U2AF consists of a large (Prp2 in fission yeast) and a small subunit, and it is the large subunit that is required for binding of the U2 snRNP to the branch point sequence (Parker et al., 1987; Ruskin et al., 1988; Wu and Manley, 1989; Zhuang and Weiner, 1989). Unlike other splicing mutations that produce cell-cycle arrest at G2 phase, temperature-sensitive prp2 mutants show different phenotypes, such as mitotic defects and failure in cell elongation (Takahashi et al., 1994; Beales et al., 2000), suggesting that Prp2 might play a role in progression of the cell cycle that is distinct from other splicing components.
We have performed a large-scale screen to identify mutants with defects in the maintenance of genome ploidy (Kominami and Toda, 1997). We have previously identified Pop1 and Pop2, two F-box proteins consisting of the SCF ubiquitin ligase (Skp1-Cullin-1-F-box) (Kominami and Toda, 1997; Kominami et al., 1998a; Patton et al., 1998; Deshaies, 1999). Loss of either Pop1 or Pop2 results in diploidisation, which is attributable to the failure in the ubiquitin-proteasome-dependent degradation of the CDK inhibitor Rum1. High levels of Rum1 in these mutants lead to the bypass of M phase and successive occurrence of S phase, which results in doubling of genome ploidy.
In this study, we have characterised pop3-5235, a diploidising mutant that was isolated in the same screen as the pop1 mutant. Pop3 is a conserved protein composed of WD repeats (Neer et al., 1994) and is the same as Wat1, which has been identified as the protein required for proper cell morphology and F-actin localisation (Kemp et al., 1997). Pop3/Wat1 has also been isolated as a partner of Prp2 by two-hybrid screening. We will show that, unlike pop1 or pop2 mutants, the reason wat1 mutants diploidise is marked reduction of α-tubulin levels, which results in unequal chromosome separation. Our study highlights a novel mechanism for the maintenance of genome stability involving microtubule integrity via transcriptional regulation.
MATERIALS AND METHODS
Strains, media and genetic techniques
The S. pombe strains used in this study are listed in Table 1. Standard procedurs for S. pombe genetics were followed (Moreno et al., 1991). Flow cytometry analysis (FACS) was performed as described previously (Kumada et al., 1995).
Cloning of the pop3+ gene
An S. pombe genomic library (Barbet et al., 1992) was used for the isolation of genes which complemented the ts pop3-5235 mutant. Three ts+ transformants showed haploid-sized cells, and became fertile. This suppression was plasmid dependent. Plasmid DNAs were recovered from these transformants and restriction enzyme mapping showed that they contained overlapping inserts. Identity of the cloned gene as pop3+/wat1+ was confirmed by genetic crosses between a tagged strain and the original mutant. Random spore analysis between a Pop3/Wat1-HA strain (kanamycin-resistance as a marker) and a ts pop3/wat1-5235 strain showed that the gene that was isolated was derived from the pop3+/wat11+ locus (no recombinants were obtained from more than 103 haploid segregants).
Yeast two-hybrid screening
The entire coding region of prp2+ was cloned into the pGBT9 vector and used as bait. An S. pombe cDNA library constructed in pGADGH (Clonetech, Palo Alto, CA) was used for screening and 30 positive clones were obtained (Wentz-Hunter and Potashkin, 1996; McKinney et al., 1997). Restriction mapping analysis and nucleotide sequencing of recovered cDNA showed that one of these clones (clone G) encoded the wat1+ gene (previously designated uap3+, U2AF59-associated-protein).
Nucleic acid preparation and manipulation
Enzymes were used as recommended by the suppliers (New England Biolabs and Roche Diagnostics). Total RNA was prepared and Northern analysis was performed as described previously (Suda et al., 2000). Nucleotide sequence data reported in this paper are in the DDBJ/EMBL/GenBank databases under Accession Number AB016895 (pop3+/wat1+).
The entire ORF of the wat1+ gene was deleted using PCR-generated fragments (Bähler et al., 1998). Tetrad dissection was performed using a standard method (Moreno et al., 1991).
C-terminal epitope tagging of Wat1
Epitope tagging of Wat1 with HA or Myc peptide was performed using PCR-generated fragments (Bähler et al., 1998). Wild-type Wat1 was tagged with 3HA, while mutant Wat1 protein derived from wat1-5235 was tagged with 13Myc.
Total cell extracts were prepared after disruption of cells with glass beads in the lysis buffer as described previously (Kominami and Toda, 1997). For immunoprecipitation, cell extracts were prepared without boiling. Mouse monoclonal anti-Cig2 (3A11), anti-Cdc2 (Y100), anti-α-tubulin (Sigma), anti-β-tubulin (KMX-1), anti-actin (N350, Amersham), anti-HA (16B12, BAbCO), anti-Myc antibodies (9E10, BAbCO), and rabbit polyclonal anti-Cdc13, anti-Cig2, anti-Rum1 and anti-Cdc18 antibodies were used. Antibodies to Prp2 were raised in rabbits against a synthetic peptide corresponding to the amino acid sequence LSSGSSRIPKRHRDYRDEE (amino acids 7-25 of the protein product; Multiple Peptide Systems, San Diego, CA). The peptide was coupled to keyhole limpet hemocyanin (KLH) via a C-terminal cysteine residue not present in Prp2. Each inoculation contained 0.5 mg of the synthetic peptide. Horseradish peroxidase-conjugated goat anti-rabbit IgG, goat anti-mouse IgG (BioRad Laboratories) and a chemiluminescence system (ECL, Amersham) were used to detect bound antibody.
Indirect immunofluorescence microscopy
Cells were fixed with methanol and primary antibodies (TAT-1, 1/50) applied, followed by Cy3-conjugated goat anti-mouse IgG (Sigma). Immunofluorescence images were viewed with a chilled video-rated CCD camera (model C5985, Hamamatsu) connected to a computer (Apple Power Macintosh G3/400) and processed by use of Adobe® Photoshop (version 5). A confocal microscope LSM510 (Zeiss Co.) was also used.
Gel filtration chromatography
Soluble protein extracts were prepared in buffer A (20 mM Tris-HCl pH 7.5, 20% glycerol, 0.1 mM EDTA, 1 mM mercaptoethanol, 5 mM ATP plus a cocktail of inhibitors, Sigma) as described previously (Vardy and Toda, 2000). Gel filtration chromatography was performed on a Superose-6 column by FPLC (Pharmacia Biotech). To determine molecular weight, a parallel column was run with standards consisting of dextran (2000 kDa), thyroglobulin (669 kDa) and ovalbulin (43 kDa).
Purification of Wat1-interacting proteins
The Wat1-containing complex was immunopurified according to the method that was used for purification of a Cdc5-containing multiprotein complex (McDonald et al., 1999). Wat1-HA tagged strains were metabolically labelled with Tran[35S]-label and used for extract preparation.
Identification of Pop3 as Wat1, a highly conserved WD repeat protein
pop3-5235 was isolated as one of a group of sterile mutants which failed to maintain genome ploidy and became diploidised (Fig. 1A). In addition, pop3-5235 was temperature sensitive (ts) as well as cold sensitive (cs) for growth (Fig. 1B). We cloned the pop3+ gene from a fission yeast genomic library by complementation of ts phenotypes. Nucleotide sequencing showed that Pop3 consists of 304 amino acid residues and is composed almost entirely of seven units of WD repeats (Fig. 2; Neer et al., 1994). pop3+ was previously identified as wat1+, mutations of which result in dislocalisation of F-actin (Kemp et al., 1997; hereafter Wat1 will be used instead of Pop3). Wat1 is an evolutionarily conserved protein and potential homologues are present in budding yeast, plant, fly and human (Fig. 2).
The wat1+ gene is not essential, but is required for genome stability
Gene disruption showed that the wat1+ gene is non-essential for cell growth (Fig. 1C). Dissection of asci from heterozygous diploid cells showed that four viable spores were obtained from 20 tetrads dissected and uracil auxotrophic phenotype segregated 2:2. wat1-deleted (Δwat1) cells, however, grew very poorly at 27°C (doubling times of Δwat1 are 180% and 120% longer than wild type and wat1-5235 mutants, respectively) and showed, like wat1-5235, cs and ts growth defects. Δwat1 is sterile, which is caused by defects in G1 arrest under nitrogen starvation conditions (Kominami and Toda, 1997). We realise that sterile phenotypes are not stable in the wat1 mutant and that phenotypic reversion occurs spontaneously, which could explain why sterility was not described in the previous study (Kemp et al., 1997). Furthermore, Δwat1 haploid cells tended to diploidise (Fig. 1D). These results indicate that Wat1 is a conserved WD-repeat protein that is required for genome stability.
Isolation of the wat1+ gene as a U2AF-Prp2 interacting partner
In a separate set of experiments, we have performed yeast two-hybrid screening using the prp2+ gene as bait (Fields and Song, 1989; McKinney et al., 1997). Prp2 is the large subunit of U2AF and forms a complex with the small subunit p23 (Potashkin et al., 1993; Wentz-Hunter and Potashkin, 1996). In order to identify additional proteins that interact with Prp2, we screened a fission yeast cDNA library. One of the plasmids we identified from this screen, clone G, encodes Wat1. Interaction between Prp2 and Wat1 was specific, as either protein alone without a partner failed to activate the GAL promoters (Table 2). Moreover Wat1 interacted with neither p23 nor Uap2, which was identified as a Prp2-binding protein from the same screening (McKinney et al., 1997).
To confirm an in vivo interaction between Prp2 and Wat1, immunoprecipitation was performed. For this purpose, a strain containing C-terminally tagged Wat1-HA was constructed. Cell extracts were prepared from this strain and immunoprecipitation was performed using anti-HA antibody; the filter was probed with an anti-Prp2 antibody. As shown in Fig. 3A, Prp2 co-immunoprecipitated with Wat1-HA (lane 2). The interaction was specific, as a non-tagged control did not precipitate Prp2 (lane 1). Reciprocal immunoprecipitation was also performed using an anti-Prp2 antibody and preimmune serum, and the interaction was examined with an anti-HA antibody. As shown in Fig. 3B, Wat1-HA was precipitated with this antibody, although a small amount of Wat1-HA was also precipitated with preimmune serum (lanes 1 and 2). Thus, Wat1 forms a complex with the large subunit of U2AF, possibly independently of its small subunit.
α-tubulin levels are reduced in the wat1 mutant
We addressed the mechanism(s) that underlie diploidisation in the wat1 mutant. Unlike pop1 or pop2 mutants, which accumulate the CDK inhibitor Rum1 to high levels, Rum1 levels were not increased in wat1 mutants (data not shown). Consistent with this, again unlike pop1 or pop2 (in which polyploid phenotypes are dependent upon the presence of Rum1; Kominami and Toda, 1997; Kominami et al., 1998a), the rum1wat1-5235 double mutant still diploidised (data not shown). This result indicated that, although Wat1 was identified from the same screen as Pop1, the reason that wat1 mutants diploidise is distinct from that for pop1 and pop2.
During a series of immunoblotting experiments to find proteins abnormally produced in wat1 mutants, we noticed that α-tubulin levels were substantially decreased in this mutant, even under the permissive condition (Fig. 4A). This decrease ofα -tubulin levels appeared to be specific, as the amount of other proteins examined, such as β-tubulin, actin, Cdc13 (B-type cyclin) and Cdc2, was not altered significantly (Fig. 4A). Temperature-shift experiments of wat1 mutants showed that α-tubulin levels became further reduced upon an upwards shift to 36°C (Fig. 4B). In wat1-deleted cells, more drastic reduction of α-tubulin levels was observed (lanes 3, 6 and 9). This result suggests that Wat1 plays an important role in the homeostasis of α-tubulin levels.
Microtubule structure and function are compromised in the wat1 mutant
Given the inability to maintain α-tubulin levels, immunofluorescence microscopy with anti-α-tubulin antibody was performed in the wat1 mutant using confocal microscopy. As shown in Fig. 5A, compared with wild-type cells (left panel), wat1 mutant cells grown at 36°C (2 hours, right panel) and even at 26°C (middle panel) contained shorter and fewer cytoplasmic microtubules. Nuclear staining with DAPI showed that, in a significant fraction of wat1 mutants, in addition to positional displacement, pairs of chromosomes separating to opposite poles at anaphase were not equal and that these segregation defects became especially obvious at 36°C (20% of mitotic cells Fig. 5B). This indicated that unequal sister chromatid segregation occurred in wat1mutants, which resulted in spontaneous diploidisation. Consistent with abnormal and compromised microtubule structures, wat1 mutants were hypersensitive to microtubule-destabilising drugs (Fig. 5C). Taking these results together, Wat1 regulates microtubule integrity by maintaining α-tubulin levels, and lack of Wat1 results in unequal chromosome separation and failure in the maintenance of genome ploidy.
Wat1 may play a crucial role in the maintenance of mature mRNA levels
We sought to examine the involvement of Wat1 in mRNA metabolism. It is known that ts prp2 mutants are defective in splicing reactions when incubated at 36°C, which could be detected with Northern analysis using intron sequence as probes (Potashkin et al., 1989; Potashkin et al., 1993; Urushiyama et al., 1996). Total RNA was prepared from wild type, Δwat1 or ts prp2 mutants incubated at 26°C or 36°C, and Northern analysis was performed. Two probes were used, which corresponded to intron and exon sequences of the tbp1+ gene (encoding the TATA-binding protein; Hoffman et al., 1990). While unspliced premature RNA species were accumulated in ts prp2 mutants to high levels upon upwards temperature shift (Fig. 6A, lanes 4 to 6 in top panel), no equivalent bands were observed in wat1 mutants (lanes 7 to 9) or in wild type (lanes 1 to 3), suggesting that Wat1 may not be involved in the splicing reaction per se. Instead, what was clearly defective in this mutant was that the total amount of tbp1+ mRNA was significantly reduced, especially when incubated at 36°C (lanes 7 to 9 in bottom panel).
In order to examine the specificity of the reduction of mRNA levels, filters were rehybridised with probes corresponding to other genes. Genes examined included nda2+ (encoding α1-tubulin; Toda et al., 1984), cdc2+, cig2+ (encoding S-phase cyclin; Bueno and Russell, 1993; Connolly and Beach, 1994; Obara-Ishihara and Okayama, 1994) and act1+ (encoding actin) (Mertins and Gallwitz, 1987), among which nda2+ and cdc2+ contain introns, while cig2+ and act1+ do not. As shown in Fig. 6B, irrespective of the presence or absence of introns, in all cases, total levels of mRNA were substantially reduced in wat1 mutants. Therefore, although it binds Prp2 in vivo, unlike Prp2, Wat1 appears to play a role in either synthesis or stability of mRNA, but not splicing.
Wild type Wat1, but not its mutant protein, forms a large complex in the cell
It has been generally believed that WD repeats are hallmarks for protein-protein interactions (Wall et al., 1995; Lambright et al., 1996). As shown before, Wat1 binds at least Prp2 in vivo. In order to examine the native size of the Wat1-containing complex, gel filtration chromatography was performed using a tagged strain in both wild-type and wat1-5235 mutant strains. In wild-type cells, the cellular Wat1 protein was found to be almost exclusively included in a large complex (>2000 kDa, Fig. 7A). In clear contrast, in wat1-5235 mutants, the majority of the Wat1 protein was eluted in much smaller fractions, most likely monomers or dimers (Fig. 7B). This result indicates that Wat1 forms a large complex in the cell, which is crucial for its protein function.
We next attempted to purify directly proteins that interact with Wat1. Large-scale immunoprecipitation using anti-HA antibody was performed with cell extracts from a Wat1-HA strain labelled with [35S]. Fig. 7C shows the [35S]-labelling coprecipitated proteins. At least five protein species (marked by arrows), possibly more, specifically co-purified with Wat1; one protein, p33 (smaller than Wat1), notable appeared to bind Wat1 in a stoichiometric manner. Identification of these co-purified proteins is in progress. This analysis has firmly established that, consistent with the protein-protein-interacting WD-repeat structure, Wat1 forms a stable complex with other proteins in the cell.
In this study we have shown that the reason for diploidisation in wat1 mutants is attributable to reduced α-tubulin levels, which result in unequal chromosome separation at mitosis. We have found that reduced protein levels are due to low α-tubulin mRNA. The analysis highlights the importance of transcriptional regulation and/or mRNA stability ofα -tubulin genes for genome integrity.
Selective reduction of α-tubulin protein levels in the wat1 mutant
Finding low α-tubulin levels in wat1 mutants has led us to explore the molecular function of Wat1 in microtubule integrity. We show that Wat1 is involved in the maintenance of mRNA levels of not only α-tubulin but also other genes, suggesting that the primary function of Wat1 is for general transcription. Our results then raised further questions. For example, while transcript levels of all the genes examined are reduced in the wat1 mutant, why are only α-tubulin protein levels significantly decreased? How is the translation rate of α-tubulin mRNA regulated in fission yeast? It is possible that the efficiency or kinetics of translation from each transcript is varied among individual mRNA species and this may be the reason that only α-tubulin protein levels appear to be low in wat1 mutants. Perhaps in α-tubulin synthesis, efficiency of translation is dependent upon the levels of its transcripts or requires stable mRNA or the half-life of α-tubulin is much shorter than that of other proteins examined in this study. It should be noted that, although we have shown only α-tubulin levels are decreased, circumstantial evidence suggests that α-tubulin is not alone and other protein levels are also dependent upon Wat1.
Previous work identified Wat1 as a protein that is required for proper localisation of F-actin (Kemp et al., 1997). Thus, Wat1 plays a crucial role in establishment of both the actin and microtubule cytoskeleton. Unlike α-tubulin, actin levels are not noticeably reduced in wat1 mutants. This suggests that some factor(s) that is required for F-actin localisation is defective in wat1 mutants. In addition, the wat1 mutant is sterile, which is attributable to a failure in G1 arrest upon nutrient starvation (Kemp et al., 1997; Kominami and Toda, 1997) — a prerequisite for conjugation in fission yeast (Kumada et al., 1995; Kominami et al., 1998b). We predict that the levels of protein(s) required for G1 arrest are also reduced in this mutant.
Wat1 is an evolutionarily conserved WD repeat protein
In budding yeast the Wat1 homologue, Lst8 was identified as a factor that is required for transport of amino acid permeases from the Golgi to the plasma membrane (Roberg et al., 1997). Given the role of Wat1 in mRNA metabolism, it is possible that in this organism, protein levels responsible for transport of amino acid permeases are tightly controlled by Lst8-mediated mRNA maturation. In mouse, expression of this homologue in adipocytes is regulated by insulin, while in human the homologue has been isolated as a gene induced by DNA damage and overexpressed in promyelocytic leukemia cell lines (HL-60; database search from DDBJ and NCBI). This suggests that Wat1 also plays a role in response to internal and external cues. Consistent with this, fission yeast wat1 mutants are hypersensitive to various chemicals and drugs (Kemp et al., 1997; I. O. and T. T., unpublished).
In contrast to the previous report (Kemp et al., 1997), we do not see abnormal cell shape in either wat1/pop3-5235 or Δwat1/Δpop3-deletion strain that we have isolated and characterised in this study. It is possible that this phenotype is allele-specific; in fact, a previously isolated deletion allele of wat1+ (wat1.d) only removed part of the protein, while the entire ORF was deleted in Δwat1.
Link between mRNA splicing and other maturation processes
Wat1 binds Prp2, the large subunit of U2AF. Although Prp2 is required for pre-mRNA splicing (Parker et al., 1987; Ruskin et al., 1988; Wu and Manley, 1989; Zhuang and Weiner, 1989; Zamore and Green, 1991; Potashkin et al., 1993), the defective phenotypes of ts prp2 mutants are not the same as mutations in other components of the spliceosome. Notably prp2 is identical to mis11, which was identified as a mutation that caused high frequency loss of mini-chromosomes (Takahashi et al., 1994), demonstrating that Prp2 also plays an important role in genome stability. Therefore, it appears that, in addition to a physical interaction, these two proteins play a common role in genome integrity.
We propose two possibilities for Wat1 function in mRNA metabolism. One is transcription; Wat1 may be involved in mRNA synthesis per se. In this case Prp2 would be implicated in transcription as well as splicing. We have performed gel filtration chromatography to examine whether a Wat1-containing complex is co-fractionated with Tbp1 or RNA polymerase II. It appears that these two proteins do not form a tight complex with Wat1 (I. O. and T. T., unpublished). The Wat1-containing complex may therefore play a regulatory role in transcription reactions. Recent analysis shows that mRNA processing such as splicing is orchestrated with transcription elongation steps (Proudfoot, 2000). The second possibility is the mRNA stabilisation pathway. It has been shown that mRNA stability is regulated by a series of pathways, such as the exosome degradation (Mitchell et al., 1997; van Hoof and Parker, 1999) and 5′ capping and polyadenylation of mRNA (Jacobson and Peltz, 1996). RNA stability and splicing are intimately connected with each other (Serin et al., 2001). Furthermore, accumulating evidence supports the interrelationships of the pathways of mRNA decay and translation (Olivas and Parker, 2000). The cytoplasmic localisation of Wat1 is consistent with a role of this protein in mRNA turnover (K. G. unpublished). In this scenario, Prp2 may play an additional role in inhibition of mRNA decay. Systematic yeast two-hybrid analysis indicates that components of RNA metabolism interact with each other and also with those involved in chromatin structure (Schwikowski et al., 2000). Further identification and characterisation of Wat1-interacting proteins will give clues about the biochemical pathways, which require Wat1.
We thank Drs Keith Gull, Masami Horikoshi and Takashi Umehara, and Paul Nurse, Tokio Tani and Hiroyuki Yamano for strains, plasmids and antibodies. We thank Jacqueline Hayles for critical reading of the manuscript and useful suggestions. K. K. was supported by JSPS Postdoctoral Fellowships for Research Abroad. This work is supported by the ICRF and the Human Frontier Science Program Research Grant (T. T.), by NIH grants (R01GM47487 to J. P.) and by the Howard Hughes Medical Institute of which K. L. G. is an associate investigator.
- Accepted May 1, 2001.
- © The Company of Biologists Limited 2001