INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Cells express only a subset of genes at a given time. The remaining genes are quiescent. How eukaryotic genes are maintained in a repressed state is a fundamental issue that remains poorly understood (reviewed in reference 21). Specific DNA binding proteins are known to bind promoters of some genes and inhibit gene expression. In addition, for many genes, the arrangement of nucleosomes across the promoter is thought to block transcription. Other genes are repressed due to their location in heterochromatic regions of the chromosome.

The yeast Saccharomyces cerevisiae is a valuable experimental organism for identifying components that maintain the quiescent state of genes. Numerous screens have been done in S. cerevisiae for mutations that restore or enhance transcription of inactive genes. For example, a screen for mutants that restore transcription of a HIS4 locus that has been inactivated by a transposon insertion into its promoter has identified a group of genes known as the SPT (suppressor of Ty) genes (62). The SPT gene products include basal transcription components such as Spt15p, the TATA binding protein (17), chromatin components such as histones (11), and other proteins of undetermined function such as Spt6p, which has been shown to interact with histones (8). Nonet and Young have identified mutations which restore viability to strains whose RNA polymerase II enzyme is compromised by truncation of the carboxyl-terminal domain of the Rpb1p subunit (41). Many of the suppressors of Rpb1p truncations (SRB) encode proteins that copurify with the RNA polymerase II holoenzyme (20, 29, 59). For example, SRB10 and SRB11 encode a cyclin-dependent kinase and cyclin which copurify with the holoenzyme (26, 29).

Studies of the regulation of specific genes have also been invaluable in the search for proteins that negatively regulate transcription. For example, the SUC2 gene is not expressed when cells are grown in glucose-containing medium. Mutations that allow expression of the SUC2 gene in the presence of glucose have been found in the genes SSN6, SRB8, SRB9, SRB10, SRB11, SIN4, ROX3, and RGR1 (54). Sin4p, Rgr1p, and Rox3p are components of a mediator activity that allows yeast RNA polymerase II to be stimulated by activators in vitro (18, 28). Biochemical studies suggest that Srb10p, Srb11p, Sin4p, Rgr1p, and Rox3p are intimately involved with RNA polymerase II function. Independently, genetic studies implicate these proteins in the negative regulation of gene expression (10, 13, 19, 46, 49, 54, 56, 61).

We have been characterizing the negative regulation of transcription of the HO gene in S. cerevisiae. HO is transcribed only in haploid mother cells at Start, the G₁-to-S phase transition (22, 35). Thus, HO is potentially repressed in haploids in daughter cells and in cells that are outside of Start. Indeed, a daughter-specific repressor of HO, Ash1p, has been identified (7, 53). It is not known, however, if there exist cell cycle-specific repressors of transcription (38). The 0.7-kb region, URS2 (Fig. 1, line 1), of the HO upstream region is required for Start-specific expression of HO and for dependence on the transcription factors Swi4p and Swi6p (36-38). Swi4p binds specifically to 10 sites in URS2 (3), called Swi4p cell cycle boxes (SCBs) (9, 37). Swi4p contains a specific DNA binding domain, four ankyrin repeats, and a domain for interaction with Swi6p (4, 5, 43, 51). Together Swi4p and Swi6p activate the transcription of Start-specific genes such as CLN1, CLN2, and PCL1 as well as artificial reporters containing multimerized SCBs (3, 9, 39, 42). Both Swi4p and Swi6p are absolutely required for HO transcription.

View larger version (19K): [in this window] [in a new window]   FIG. 1.   Structures of the intact HO gene (line 1) and reporter constructs UASGAL-URS2R-lacZ reporter (line 2), URS1-URS2R-ho allele (line 3), and URS1-URS2R-lacZ (line 4) are depicted. Numbering indicates the nucleotide positions of the endpoints of each region relative to the HO ATG codon. Nucleotide positions for lines 2 to 4 are as for line 1.

The URS2 region is responsible for dependence of HO transcription on Swi4p and Swi6p, as deletion of URS2 allows transcription of HO in the absence of these proteins. Conversely, insertion of the URS2 region between the GAL1 upstream activation sequence (UAS_GAL) and a TATA box confers Swi4p dependence on UAS_GAL-driven expression (38). Thus, in the absence of Swi4p, URS2 prevents expression of the HO gene despite the presence of potent activation sequences upstream, either UAS_GAL or URS1, the far upstream region of the HO promoter (Fig. 1, line 1).

To characterize regulation by URS2, we screened for mutants, named nut (negative regulation of URS2) mutants, which are defective in the Swi4p dependence of UAS_GAL-URS2R-lacZ, a synthetic reporter gene containing part of URS2. We describe two novel genes, NUT1 and NUT2, that are required for the Swi4p dependence of UAS_GAL-URS2R-lacZ.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains. Standard techniques for strain construction and mutant analysis were used (45). Key yeast strains are listed in Table 1. Except where indicated, all strains are isogenic to RT228, a strain derived from W303 and K1107 as follows. The SWI4 locus of strain K1107 (gift of Kim Nasmyth, Institute of Molecular Pathology, Vienna, Austria) was deleted by using pJO98 (42) digested with EcoRI and SalI to create RT211. This strain, RT211, was crossed to W303 to produce the segregant RT228. To create strains isogenic to RT228, the mating-type locus of RT228 was converted to MAT by two-step gene replacement (47) using pSC9 (gift of S. Chu, University of California, San Francisco). An isogenic SWI4 strain was created by two-step gene replacement using pRKT427.

(i) Deletion of SIN4 and SRB genes. SIN4 was deleted in a haploid strain by using plasmid M1381 (gift of David Stillman, University of Utah). The SRB genes were deleted in strain RT689 by using the following plasmids: to disrupt SRB8, pSL315 (20); to disrupt SRB9, pWS44-11 (gift of Marian Carlson, Columbia University) (54); to disrupt SRB10, pMW14 (gift of Madhu Wahi, University of California, San Francisco) (61); to disrupt SRB11, RY7036 (gift of Richard Young, Massachusetts Institute of Technology).

(ii) Deletion of NUT1. NUT1 was deleted by using plasmid pRKT365 after digestion with NotI and SfiI, which release the disruption cassette. Deletion of the genomic NUT1 locus was verified by PCR.

(iii) Deletion of NUT2. One NUT2 allele of RT575, an a/ W303 diploid, was deleted by transformation with pRKT432 after digestion with SalI and NotI. Strains with deleted alleles were identified by PCR.

(iv) URS1-URS2R-HO. Plasmid pRKT619 was targeted to the ho locus by digestion with NruI and transformation into yeast strains RT690 and RT848 followed by selection for uracil prototrophy. Loop-outs were selected on plates containing 5-fluoro-orotic acid (5-FOA) (47) and verified by PCR. Alleles generated by this procedure, which deletes URS2L, are referred to as urs2L in Table 1 and as URS1-URS2R in Results.

Mutant isolation. Strain RT243 was mutagenized with UV irradiation to 95% inviability and plated on YEP dextrose so that all mutants were independently derived. Colonies were replica plated onto YEP galactose plates covered with Whatman no. 50 filters. After overnight growth, filters were removed and subjected to a filter -galactosidase assay (55). Blue colonies were recovered from the original YEP dextrose plate. The 14 original mutant isolates are strains RT259 through RT271. Mutations were assigned to complementation groups by mating mat derivatives of mutant strains to MAT derivatives obtained by a standard backcross. The Nut phenotype of mat/MAT diploids was assayed by -galactosidase filter assays on YEP galactose plates.

-Galactosidase assays. Plate and liquid -galactosidase assays were performed as described elsewhere (55). For assays of liquid cultures, cells were grown to mid-log phase (optical density at 600 nm of 0.1 to 1.2). Assays were performed in triplicate. Standard deviations of all reported values were less than 10% of experimental values.

Complementation and cloning of NUT genes. (i) SRB gene complementation. Plasmid pRKT439 was isolated from the Rose genomic library (44) as a plasmid that complements the mutant phenotype of RT387, a nut7-2 mutant. To determine if strains RT259 (nut9-1), RT381 (nut6-2), and RT379 (nut8-1) are mutant in other SRB genes, these strains were transformed with plasmids containing SRB8 (pMW18; gift of M. Wahi), SRB9 (pWS8; gift of M. Carlson) (54), or SRB11 (pSK5; gift of S. Kuchin and M. Carlson) (26). Complementation of the Nut phenotype was scored by -galactosidase filter assays on YEP galactose plates.

(ii) ROX3 complementation. To determine if nut3-1 is defective in ROX3, RGR1, or SIN4, RT269 was transformed with URA3 plasmids containing ROX3 (pIT218; gift of M. Carlson) (54), RGR1 (pM2597), and SIN4 (pM1305; gift of D. Stillman) (24). Allelism with ROX3 was determined by integration of HindIII-cut pIT225 (gift of M. Carlson) (54) into strain RT689. This strain was crossed to nut3 mutant strain RT362.

(iii) NUT1 complementation. Plasmid pRKT353 was isolated from the Rose genomic library in YCp50 (44). pRKT354, a derivative of this plasmid obtained by removing the SalI-SphI fragment of pRKT353, complements the Nut phenotype of RT271. YGL151w is the only complete reading frame on this plasmid. pRKT355, which contains a SalI-XbaI deletion that disrupts YGL151w, lacks the complementing activity. To determine if the insert on pRKT353 is allelic to NUT1, the 2.3-kb HindIII fragment from pRKT353 was subcloned into the HindIII site of pRS306 (52) to generate pRKT356. pRKT356 was targeted to the genomic locus of the insert after digestion with BstEII and transformation to uracil prototrophy in the NUT1 NUT2 strain RT243. When the progeny from crosses of this strain to nut1 nut2 RT364 were analyzed, only 1 of 49 URA3 spores was phenotypically Nut, indicating that URA3 is tightly linked to the NUT1 locus.

(iv) NUT2 complementation. Plasmid pRKT404 was isolated from the Rose genomic library. The SalI-HindIII fragment of the genomic insert from pRKT404 was subcloned into pRS306 to generate pRKT417. Integration of AflII-digested pRKT417 marks the locus of the insert with URA3. Analysis of 16 tetrads from a cross between nut2-1 and a strain bearing an integration of AflII-digested plasmid pRKT417 demonstrated that the insert on pRKT404 was allelic to nut2-1 since none of 16 Nut segregants carried the integrated pRKT417 allele. To determine which open reading frame complemented the nut2-1 defect, derivatives of pRKT404 were made by deleting the EcoRI fragment (pRKT447), the SphI-AflII fragment (pRKT448), and the XbaI-AflII fragment (pRKT450) and by filling in the AflII site, which inactivates the reading frame YPR169w (pRKT449). This analysis demonstrated that YPR168w is the open reading frame that complemented the nut2-1 defect.

(v) NUT21/CCR4 complementation. Plasmid pRKT562 was isolated from the Rose genomic library by complementation of the mutant phenotype of RT609. The SalI-NotI fragment was excised from pRKT562 to generate plasmid pRKT564, which still complemented the mutant phenotype. The SpeI-NotI fragment from pRKT562 was subcloned into the SpeI and NotI sites of pRS306 to generate plasmid pRKT566. When integrated at the URA3 or CCR4 locus, pRKT566 complemented the mutant phenotype. CCR4 is the only gene contained on the SpeI-NotI fragment. Finally, this construct was used to mark the CCR4 locus by integration of the SfiI-digested plasmid. Crosses established that CCR4 is tightly linked to nut21-1, since in 10 tetrads, no recombination was observed between nut21-1 and the marked allele of CCR4.

Sequence analysis. Database searches with the Nut1p and Nut2p protein sequences were performed by XREF (6), with additional support from the Wisconsin package of the Genetics Computer Group. Nucleotide positions in the HO promoter are numbered relative to the HO ATG codon such that the A immediately preceding the ATG is 1.

Plasmids. Standard methods for DNA manipulations were as described previously (50). Plasmids generated for this study are listed in Table 2.

(i) lacZ reporters. Two UAS_GAL-URS2R-lacZ reporters, pRKT229 and pRKT188, were used. To create the UAS_GAL-URS2R-lacZ reporter pRKT229, the XhoI-HindIII fragment of pRKT225, which contains the 354-bp SspI-HindIII fragment of URS2 filled in and ligated into the SalI site of pBluescript, was cloned into the XhoI and HindIII sites of pRKT208. To create pRKT188, the 406-bp SspI fragment of URS2 was ligated into the HindIII site of a plasmid closely related to pRKT208. This construct generates a small duplication of the region between the HindIII site at 171 and the SspI site at 122. pRKT208 was generated in two steps. First, the HindIII-BglII fragment of pJO11, a derivative of the ho-lacZ fusion from Russell et al. (48), was ligated into pRS305 (52) cut with HindIII and BamHI. This plasmid was digested with ApaI, filled in with T4 polymerase, and ligated with the Sau3A-AvaI fragment containing UAS_GAL which had been filled in with T4 polymerase. To create pRKT533, URS1 with an ApaI site and an engineered XhoI site (at position 871 of HO) was digested with these enzymes and ligated into the ApaI and XhoI sites upstream of URS2 in the reporter construct. To create pRKT619, the ApaI-to-HindIII fragment from pRKT533 was cloned into the ApaI and HindIII sites of pRS306.

(ii) UAS-less reporter. A fragment containing the URA3 gene was ligated into a HindIII- and SacI-digested pBR322 plasmid containing the HO upstream region and gene to generate plasmid pRKT186. pRKT186 has the HO upstream region replaced with URA3. Strain RT233 was generated by transformation of strain K1107 with SalI- and PstI-digested pRKT186. The resulting allele at the HO locus, ho::uasURA3::ho-lacZ, lacks a UAS and is similar in construction to a UAS-less reporter plasmid used by others (24).

(iii) NUT1 deletion plasmid. The 5.2-kb XhoI fragment containing YGL151w was subcloned into the SalI site of pGEM11(f⁺) to generate pRKT363. Plasmid CY253, containing TRP1, was digested with BglII and HincII and ligated to pRKT363 cut with BglII and EcoRV. This plasmid was then digested with BglII and NcoI, filled in with Klenow DNA polymerase, and religated to remove the entire 5' coding region of YGL151w. This construct, pRKT365, deleted all of the NUT1 open reading frame except the region coding for the carboxyl-terminal 53 amino acids.

(iv) NUT2 deletion plasmid. The 9-kb SalI-HindIII fragment was subcloned into pBluescript KS⁺. The EcoRI-BamHI fragment was then removed by digestion and recircularization of the vector. Divergent primers containing BamHI sites that anneal at the ATG and stop codon of YPR168w were used in the PCR to amplify the flanking regions of YPR168w. The PCR product was digested with BamHI and recircularized. The resulting plasmid was digested with BamHI and ligated to the BamHI-BglII fragment containing the hisG-URA3-hisG cassette from pNKY51 (1) to generate pRKT432.

(v) NUT1 tags. The SacI-XhoI fragment of pRKT363 containing the entire NUT1 locus was cloned into the SacI and XhoI sites of pRS306 (52). Uracil-substituted single-stranded DNA was prepared from this plasmid by using the phage VCSM13 (Stratagene) and Escherichia coli CJ236. By using site-directed mutagenesis (27), a BamHI site was inserted at the amino-terminal end of the NUT1 open reading frame. The resulting plasmid was digested with BamHI, and annealed oligonucleotides encoding the hemagglutinin (HA) tag were ligated in frame. This plasmid, pRKT542, complemented a nut1 mutation.

(vi) NUT2 tags. The XbaI-ClaI fragment containing the NUT2 locus was subcloned into the XbaI and ClaI sites of pRS306. A BamHI site at the carboxyl terminus was inserted by the same method as for NUT1. The HA tag was ligated into this site in frame to create plasmid pRKT535. For immunofluorescence, this construct was digested with BssHII and integrated into the genomic NUT2 locus. This plasmid, which complemented the inviability of a nut2, is otherwise wild type in sequence.

Gap repair of nut2-1. The nut2-1 allele was gap repaired from the strain RT271. Plasmid pRKT447 was digested with XbaI and BssHII, gel purified, and transformed into RT271. The gap-repaired plasmid pRKT519 was recovered by electroporation into XL1-Blue cells (Stratagene). The sequence of the allele was determined by dideoxynucleotide sequencing using primers complementary to NUT2. The XbaI-ClaI fragment containing the nut2-1 allele was subcloned into pRS306 to generate pRKT515. This plasmid was used to replace the genomic NUT2 locus with the nut2-1 allele, using two-step gene replacement (47).

Immunofluorescence. Indirect immunofluorescence was performed as described by Sil and Herskowitz (53). The HA-11 antibody (Babco) was used as the primary antibody at a 1:1,000 dilution.

Northern analysis. Northern analysis was performed as described by Sil and Herskowitz (53).

Temperature-sensitive alleles of NUT2. Temperature-sensitive alleles of NUT2 were generated by PCR mutagenesis as described by Muhlrad et al. (34). One allele, nut2-ts70, was cloned into pRS306 to generate plasmid pRKT559. This plasmid was used to introduce nut2-ts70 into the genomic NUT2 locus by two-step gene replacement (47).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

The UAS_GAL-URS2R-lacZ reporter is Swi4p dependent. To characterize the ability of URS2 to confer Swi4p dependence for transcription, we generated a reporter gene based on the UAS_GAL-URS2 allele of HO created by Nasmyth (38). To assay Swi4p dependence, the UAS_GAL from the GAL1-10 intervening region was fused to the ho TATA box which drives the expression of an ho-lacZ reporter gene. The 360 bp URS2R (right) from 528 to 171 was inserted between the UAS_GAL and the TATA element to generate the UAS_GAL-URS2R-lacZ reporter (Fig. 1, line 2). An identical reporter lacking URS2R, UAS_GAL-lacZ, was used as a control.

View larger version (69K): [in this window] [in a new window]   FIG. 2.   The UASGAL-URS2R-lacZ reporter is dependent on Swi4p and galactose. Expression of the UASGAL-lacZ reporter (pRKT229) was assayed in a wild-type (RT692; lanes 1 and 2) and an isogenic swi4 (RT693; lane 3) strain. Expression of the UASGAL-URS2R-lacZ reporter was assayed in a wild-type (RT688; lanes 4 and 5) and an isogenic swi4 (RT689; lane 6) strain. RNA was isolated from these strains grown in YEP galactose (G; lanes 2, 3, 5, and 6) or YEP dextrose (D; lanes 1 and 4) liquid medium at 30°C. RNA was analyzed by Northern blotting and hybridized with probes for lacZ and TCM1 as a control for RNA content.

Isolation of mutants that exhibit Swi4p-independent expression. To identify gene products that might function at URS2R to confer Swi4p dependence, we screened for mutations that relieve the Swi4p dependence of UAS_GAL-URS2R-lacZ in galactose medium. swi4 deletion strains grown on galactose plates formed white colonies because the lacZ reporter was not expressed, whereas mutant strains formed blue colonies due to lacZ expression. From 80,000 mutagenized swi4 cells, 14 mutants were isolated. Mutants were categorized as either weak suppressors or strong suppressors of the SWI4 deletion (Table 3). For example, strain RT267, which produced 10 U of -galactosidase activity, is representative of the weak class of suppressors. In contrast, strain RT271, which produced 80 U of -galactosidase activity, is representative of the strong class of suppressors. The wild-type parental strain expressed less than 1 U of activity. We also noted that all mutants isolated exhibited similar secondary phenotypes. In particular, they invaded the agar of YEP dextrose plates to much greater extent than the parental strain and were more flocculent than the parental strain when grown in liquid culture.

View this table: [in this window] [in a new window]   TABLE 3.   nut and sin4 mutations alleviate Swi4p-dependent reporter gene expression

MAT

Weak Nut mutants carry mutations in the genes SRB8, SRB9, SRB10, and SRB11. To determine the nature of the defect in strains of the weak phenotype class, we cloned NUT7 by complementation (see Materials and Methods). Sequencing of the insert of the complementing plasmid indicated that it contained SRB10. Mutations in this gene were previously identified as a suppressor of truncations in the carboxyl-terminal domain of RNA polymerase II, as a modifier of the glucose-repressed state of SUC2 (54), and as a modifier of 2-mediated repression of a-specific genes (61). Plasmid pMW11, which contains only SRB10 and none of the adjacent genes from the original complementing plasmid, also complemented the defect in nut7-2. Further, as for the original nut7 alleles, deletion of SRB10 had a weak defect in reporter gene regulation (Table 5).

Mutation of ROX3 or SIN4 causes a Nut phenotype. To determine the identities of the remaining genes, we compared the phenotype of the nut mutants with that of a sin4 deletion in our strain background. Loss of SIN4 function has been previously shown to bypass the Swi4p requirement for transcription of ho-lacZ reporters (24, 32). Likewise, deletion of SIN4 allowed transcription of the UAS_GAL-URS2R-lacZ reporter in swi4 strains (Table 3). The nut3 mutant complemented a sin4 strain, suggesting that nut3 is not an allele of SIN4. Since mutations in RGR1 and ROX3 have phenotypes like mutations in SIN4 (13, 23, 54), we tested whether centromere plasmids bearing these genes could complement nut3-1. A plasmid that contained the ROX3 open reading frame alone complemented the mutant phenotype, whereas plasmids containing RGR1 or SIN4 did not. Further, we found that nut3 was allelic to the ROX3 gene since nut3 segregated away from a ROX3 locus that was marked with the URA3 gene in eight tetrads. Therefore, mutation of either ROX3 or SIN4 causes inappropriate expression of the UAS_GAL-URS2R-lacZ reporter.

Mutant RT271 is defective in two genes, NUT1 and NUT2. The mutant phenotype in RT271, which exhibits a strong defect in the Swi4p dependence of UAS_GAL-URS2R-lacZ, segregated as if it were due to mutations in two unlinked loci that we designated nut1 and nut2. In 32 tetrads, 7 parental ditype tetrads, 4 nonparental ditype tetrads, and 21 tetratype tetrads were observed. This ratio is consistent with the expected 1:1:4 ratio for segregation of two unlinked genes, when at least one gene is far from its centromere. Further, when one mutant locus, NUT1, was homozygous whereas the other locus, NUT2, was heterozygous, single-gene segregation (2:2 segregation) of the phenotype was observed in 14 tetrads.

NUT1 encodes a novel nonessential protein. To clone NUT1 and NUT2, we transformed the double mutant strain with a genomic library, reasoning that a centromere plasmid containing either gene should restore the Swi4p dependence of UAS_GAL-URS2R-lacZ. We obtained two plasmids that differed in their restriction maps. The insert from one plasmid was tightly linked to nut1, whereas the insert of the other plasmid was tightly linked to nut2 (see Materials and Methods).

NUT2 encodes a novel essential protein. By subcloning the insert from the other complementing plasmid, we determined that the NUT2 gene is YPR168w. This open reading frame encodes a protein of 157 amino acids which has sequence homologs of unknown function including the human expressed sequence tag yx99c06.r1 and the Caenorhabditis elegans open reading frame T09A5.6 (Fig. 3).

View larger version (25K): [in this window] [in a new window]   FIG. 3.   Alignment of Nut2p with human and C. elegans homologs. The human expressed sequence tag yx99c06.r1 (GenBank accession no. N40234) and the C. elegans open reading frame T09A5.6 (SwissProt P45966) are aligned with Nut2p. Black highlighting indicates positions conserved among these proteins. The amino acid positions of the Nut2p sequence are indicated above the alignment.  denotes the position mutated to lysine in the nut2-1 allele.

View larger version (45K): [in this window] [in a new window]   FIG. 4.   NUT2 is an essential gene that exhibits synthetic lethality with NUT1. (A) NUT2 is an essential gene. Isogenic strains that are NUT2 (RT781) or nut2 (RT783) were maintained with a centromere plasmid containing URA3 and NUT2. Strains were streaked on synthetic complete medium (left) or on synthetic complete medium containing 5-FOA (right) and incubated for 3 days. (B) A temperature-sensitive allele of NUT2, nut2-ts70, is synthetically lethal with a deletion of NUT1. Isogenic strains derived from RT784 that are NUT1 NUT2 (RT784-6b), nut1 NUT2 (RT784-1d), NUT1 nut2-ts70 (RT784-1a), and nut1 nut2-ts70 (RT784-2c) were constructed bearing pRKT353, a centromere plasmid containing the URA3 and NUT1 genes. Strains were streaked on plates with (center) or without (left and right) 5-FOA and incubated at 25°C (left and center) or 37°C (right) for 3 days.

Nut1p and Nut2p localize to the cell nucleus. To determine the subcellular localization of Nut1p and Nut2p, we introduced epitope tags into the coding sequence of each gene. An amino-terminal fusion of the Nut1p open reading frame joined to two copies of the HA epitope complemented the Nut phenotype of nut1 nut2-1 strains. Likewise, the Nut2p open reading frame was fused in frame at its carboxyl terminus to two copies of the HA epitope. The Nut2p-HA protein complemented the inviability of a NUT2 deletion. By indirect immunofluorescence using a monoclonal antibody directed against the HA epitope, both the HA-Nut1p fusion protein (Fig. 5, columns 1 and 2) and the Nut2p-HA protein (Fig. 5, columns 3 and 4) were localized to the nuclei of yeast cells. Little if any background signal was observed in untagged strains analyzed in parallel with the same antibodies and under the same conditions (Fig. 5, columns 5 and 6). These data suggest that Nut1p and Nut2p are predominantly localized in cell nuclei, where they might affect gene expression.

View larger version (31K): [in this window] [in a new window]   FIG. 5.   Nut1p and Nut2p are localized to nuclei. Cells bearing integrated copies of the HA-NUT1 fusion (pRKT542; columns 1 and 2), the NUT2-HA fusion (pRKT535; columns 3 and 4), or no fusion (columns 5 and 6) were harvested in mid-log phase, fixed, and stained with antibodies against the HA epitope (-HA) as described in Materials and Methods. Localization of the tagged epitope was revealed by rhodamine-conjugated anti-mouse antibodies in the first row. For the same field of cells, DNA was visualized by 4',6-diamidino-2-phenylindole (DAPI) staining in the second row. Cell outlines were visualized by phase-contrast microscopy in the third row.

Loss of function of NUT2 alone causes Swi4p-independent reporter transcription. To determine the phenotype of complete loss of NUT2 function, we generated alleles of NUT2 that are temperature sensitive for viability using PCR mutagenesis. One such allele, nut2-ts70, did not support cell growth at the restrictive temperature of 37°C (Fig. 4B, top right sector). nut2-ts70 cells died with abnormal morphology but no specific cell cycle arrest (58a). We generated swi4 strains carrying the UAS_GAL-URS2R-lacZ reporter that were NUT1 NUT2, nut1 nut2-1, or NUT1 nut2-ts70. Following growth in YEP galactose medium at 25 or 37°C, we assayed reporter gene expression by Northern analysis (Fig. 6A). Consistent with data from analysis of -galactosidase activity, the lacZ transcript was apparent in SWI4 strains but absent from swi4 strains regardless of temperature (Fig. 6A; compare lanes 1 and 2 with lanes 3 and 4). Deletion of NUT1 in combination with the nut2-1 allele restored transcription of the reporter in the absence of Swi4p (Fig. 6A, lanes 5 and 6). The nut2-ts70 allele did not perturb reporter gene expression at the permissive temperature (Fig. 6A, lane 7) but did cause expression of reporter gene transcription at the restrictive temperature of 37°C (Fig. 6A, lane 8), despite NUT1 activity. We conclude that inactivation of NUT2 alone is sufficient to cause Swi4p-independent expression of the UAS_GAL-URS2R-lacZ reporter.

View larger version (47K): [in this window] [in a new window]   FIG. 6.   Inactivation of a temperature-sensitive allele of NUT2 allows Swi4p-independent expression. (A) Expression of the UASGAL-URS2R-lacZ reporter was assayed in wild-type (RT688; lanes 1 and 2), swi4 (RT689; lanes 3 and 4), swi4 nut1 nut2-1 (RT718; lanes 5 and 6), and swi4 NUT1 nut2-ts70 (RT748; lanes 7 and 8) strains. RNA was isolated from these strains after growth in YEP galactose liquid medium at 25°C (odd-numbered lanes) or 37°C (even-numbered lanes) for 3 h. RNA was analyzed by Northern blotting and hybridized with probes for lacZ, CLN2, PCL1, and TCM1. The latter serves as an RNA loading control. The TCM1 signal indicates that lanes 2, 3, and 4 contain slightly more RNA than other lanes. All strains are isogenic. (B) Expression of the SUC2 gene was assayed in wild-type (RT688; lanes 1 and 5), swi4 (RT689; lane 2), swi4 nut1 nut2-1 (RT718; lane 3), and SWI4 nut1 nut2-1 (RT720; lane 4) strains by Northern analysis. Strains were grown in dextrose (D; lanes 1 to 4) or in galactose (G; lane 5) medium. (C) STE2 expression was assayed in MAT cells (lanes 1 and 3) and MATa cells (lanes 2 and 4) that were either wild type (lanes 1 and 2) or nut1 nut2-1 mutants (lanes 3 and 4).

NUT1 is synthetically lethal with a temperature-sensitive allele of NUT2. After replacement of nut2-ts70 into the genomic NUT2 locus, we sporulated a nut1/+ nut2-ts70/+ diploid to determine the phenotype of nut1 nut2-ts70 double mutants. Of 19 tetrads germinated at the permissive temperature, 25°C, no nut1 nut2-ts70 double mutants were found although 19 NUT1 nut2-ts70 spores grew up normally. To determine if nut1 is indeed synthetically lethal with nut2-ts70, we obtained a double mutant that was rescued by a URA3 NUT1 plasmid. Such double mutants were unable to lose the URA3 NUT1 plasmids, as they failed to grow on 5-FOA whereas their isogenic NUT1 nut2-ts70 siblings did (Fig. 4B). Thus, deletion of NUT1 exacerbates the growth phenotype of the nut2-ts70 allele just as it exacerbates the reporter gene phenotype due to the nut2-1 allele.

Mutation of NUT1 and NUT2 affects expression of the UAS_GAL-URS2R-lacZ reporter but not the endogenous HO gene. To determine if NUT1 and NUT2 are physiological regulators of HO transcription, we assayed HO transcription in the nut1 nut2-1 double mutant. We observed that the nut1 nut2-1 double mutant did not affect the Swi4p dependence of the intact ho gene containing URS1-URS2-ho (data not shown). Because the lacZ reporter only contained part of URS2, we constructed an allele of the endogenous ho gene in which the wild-type URS2 region was truncated, leaving only URS2R, the minimal region of URS2 required for Swi4p dependence (Fig. 1, line 3). This allele, URS1-URS2R-ho, required Swi4p (Fig. 7, lane 2) and Swi5p (data not shown) for ho transcription as assayed by Northern hybridization. However, in contrast to the UAS_GAL-URS2R-lacZ reporter, transcription of ho from the URS1-URS2R-ho gene was not restored in the absence of Swi4p in the nut1 nut2-1 mutant (Fig. 7, lane 3).

View larger version (39K): [in this window] [in a new window]   FIG. 7.   The phenotype of nut1 nut2 mutants is manifest only in the presence of lacZ sequences. (A) HO expression was analyzed for strains containing the URS1-URS2R-ho allele (lanes 1 to 4) or the URS1-URS2R-lacZ allele (lanes 5 to 8). Wild-type strains (RT890 [lane 1] and RT973 [lane 5]) were compared with swi4 strains (RT952 [lane 2] and RT976 [lane 6]), swi4 nut1 nut2-1 strains (RT953 [lane 3] and RT953 [lane 7]), and swi4 sin4 strains (RT1118 [lane 4] and RT1120 [lane 8]). (B) PHO5 expression was measured in strains grown in the presence (lanes 1 to 5) or absence (lane 6) of inorganic phosphate. Wild-type (lanes 1, 5, and 6) was compared with swi4 (RT952; lane 2), swi4 nut1 nut2-1 (RT953; lane 3), and swi4 sin4 (RT1118; lane 4) strains.

Mutation of NUT1 and NUT2 affects the expression of another lacZ-containing reporter but not other Swi4p-dependent genes. To determine if the nut1 nut2-1 strain is generally defective in the regulation of gene expression or defective in the regulation of lacZ-containing reporters other than UAS_GAL-URS2R-lacZ, we assayed the expression of another reporter and other genes in this strain. Transcription of the PCL1 gene is dependent on Swi4p (42), whereas transcription of the CLN2 gene is largely but not completely dependent on Swi4p (14). To determine if mutation of NUT1 or NUT2 enhances the Swi4p-independent transcription of PCL1 or CLN2, we analyzed transcription of these genes in wild-type, swi4, swi4 nut1 nut2-1, and swi4 nut2-ts70 strains (Fig. 6A). We were unable to detect any increase in the Swi4p-independent transcription of these two genes in the nut mutant strains.

Mutation of CCR4 in combination with deletion of NUT1 has a Nut phenotype. Because NUT2 is essential for cell viability and is a novel gene, we aimed to identify other genes that function like NUT2. Therefore, we sought other mutants with a Nut2p-like phenotype by screening for constitutive expression of the UAS_GAL-URS2R-lacZ reporter in cells deleted for SWI4 and NUT1. From this screen, we identified genes that define mutations in at least four different complementation groups (21a). One mutant strain, RT609, was characterized further. This strain contained mutation, designated nut21-1, which exhibits a temperature-sensitive growth defect that cosegregated in seven tetrads with the defect in reporter gene regulation. nut21-1 was not allelic to NUT2 but, like the nut2-1 allele, exhibited a Nut phenotype in combination with nut1 (Table 8). Further, nut21-1 exhibited no phenotype alone or in combination with nut2-1. These results suggest that NUT21 encodes an essential protein that, like Nut2p, functions with Nut1p to regulate expression of the UAS_GAL-URS2R-lacZ reporter.

View this table: [in this window] [in a new window]   TABLE 8.   Mutation of NUT21 synergizes with nut1 to cause Swi4p-independent reporter gene expression

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

We have identified a novel class of genes which includes NUT1, NUT2, and CCR4 in screening for mutations that relieve the Swi4p dependence of an artificial reporter regulated by the URS2R region of the HO promoter. Nut1p, Nut2p, and Ccr4p formally behave as negative regulators of gene expression. Their function is distinctive in two respects. First, these proteins contribute substantially to the regulation of artificial reporters but not detectably to their endogenous counterparts. Second, these proteins function cooperatively with each other.

Nut1p and Nut2p behave as a negative regulators of transcription. We found that the 360-bp URS2R segment from the URS2 region is sufficient to confer Swi4p dependence on the UAS_GAL activation sequences. We propose that in the absence of Swi4p, regulatory mechanisms inhibit activation of transcription by Gal4p in the case of UAS_GAL-URS2R-lacZ and by Swi5p in the case of HO. In nut1 nut2 double mutants, however, URS2R is unable to inhibit transcription of UAS_GAL-URS2R-lacZ and URS1-URS2R-lacZ in the absence of Swi4p. Thus, for these reporters, Nut1p and Nut2p act as negative regulators of transcription.

The phenotypes of nut1 nut2-1 mutants are context dependent. We found that the nut1 nut2 double mutation relieved the Swi4p dependence of UAS_GAL-URS2R-lacZ and URS1-URS2R-lacZ but did not affect the Swi4p dependence of URS1-URS2R-ho or URS1-URS2-ho. These data establish that Nut1p and Nut2p are not absolutely required to regulate transcription of the endogenous HO gene. Possibly, unknown components operate redundantly with Nut1p and Nut2p to regulate HO transcription. The insertion of lacZ sequences might disable these components and thus reveal a requirement for Nut1p and Nut2p. Another possibility is that the presence of lacZ sequences in yeast may simply generate a new situation which requires additional proteins for regulation. We cannot exclude the possibility that mutations in NUT1 and NUT2 would increase the expression of any lacZ-containing reporter. Further experimentation will be necessary to resolve this issue.

Nut1p and Nut2p function cooperatively. The other unusual property of the nut1 nut2 double mutation is the strong synergy between loss of NUT1 function and the nut2-1 allele (Table 6). An analogous relationship was observed for genes required for glucose repression of SUC2 (60). The finding of two-gene traits often implies functional redundancy between the products of the two genes. Because the original nut1-1 and nut2-1 alleles exhibit no phenotype alone, Nut1p and Nut2p may function in redundant, parallel pathways.

Nut1p, Nut2p, and Ccr4p may comprise a distinctive class of proteins. Our finding that Nut1p and Nut2p function cooperatively suggests that other proteins might cooperate with Nut1p. Therefore, we asked if deletion of NUT1 could synergize with mutations in genes other than NUT2 to cause a Nut phenotype. From analysis of such mutations, we estimate that at least four such genes exist. An allele of CCR4 caused a Nut phenotype in a nut1 background, indicating that Ccr4p may function with Nut1p. Ccr4p was previously identified as a regulator of ADH2 gene expression (15). Its role in gene regulation appears subtle, as the strongest phenotypes were observed on ADH2 genes that were compromised by SPT/CRE mutations or by delta insertions. Ccr4p is a component of a large protein complex that includes Caf1p, Dbf2p, Not1p, Not2p, Not3p, and Not4p (16, 30, 31). The Not proteins appear to be involved in transcriptional repression (12). Our finding that Ccr4p functions with Nut1p to negatively regulate UAS_GAL-URS2R-lacZ is consistent with these studies. Although we have yet to determine the identity of the remaining genes that function like NUT2, some may encode other proteins in the Ccr4p complex, and likewise Nut1p and Nut2p might physically interact with the Ccr4p complex. Remarkably, Harashima et al. identified two different two-gene traits, one due to mutations in BEL3 and BEL7 and the other due to mutations in BEL5 and BEL6, that affect the repression of PHO5 at a heterologous genomic locus but not at its native locus (19). Not only is the nut1 nut2 mutant similarly defective in two genes, but it shares with these two bel mutants rough colony morphology, flocculent growth in liquid culture, and elevated expression of artificial, albeit different reporter genes. It remains to be determined if bel3, bel5, bel6, and bel7 are alleles of NUT1, NUT2, or CCR4.