Nucleic Acids Research

Strains and plasmids

Yeast strains (Table 1) were all derivatives of strain SK-1. RME1 alleles have been described in detail (7-9). rme1::P_GAL1-S53-RME1::TRP1 replaces the RME1 promoter with the GAL1 promoter and modifies the N-terminus of Rme1p to include an S53 epitope. This allele, referred to in the text as P_GAL1-RME1, is functional. rme1::P_GAL1-S53-rme1-213::TRP1 is a missense allele of P_GAL1-RME1 causing a substitution in the second Rme1p zinc finger motif. This allele, referred to as P_GAL1-rme1-213, abolishes DNA binding and Rme1p-dependent repression, but not polypeptide accumulation (8).

Table 1. Yeast strains used in this study^aAll SK-1 derivatives carried the mutations (SK1*) leu2::hisG trp1::hisG lys2 ura3 ho::LYS2 gal80::LEU2, except as indicated.

The IME1-81, IME1-82 and IME1-81,82 alleles were created by oligonucleotide mutagenesis of plasmid pAC82, which carries the IME1 gene with 2.2 kbp of upstream and 1.5 kbp of downstream flanking nucleotides in the BamHI site of pRS314. IME1-81 is a deletion of nt -2044 to -2025 and insertion of one C residue to create a SalI site (8). IME1-82 is a substitution of the sequence 5[prime]-CCGGGA-3[prime] for nt -1954 to -1949 to create a SmaI site. The IME1-81,82 allele contains both the IME1-81 and IME1-82 mutations. IME1 mutations were introduced by two-step integration from inserts in the URA3 vector pRS306 and were confirmed by PCR and restriction digestion.

Plasmid pAC153-4, which carries the RC (IME1 nt -2146 to -1743) upstream of a CYC1-lacZ reporter, has been described previously (8).

In vivo footprinting

In vivo dimethyl sulfate (DMS) and UV-photo footprinting were performed as described previously (12,14-16). Early exponential phase cells from a 1 l culture in YPAc medium were collected and resuspended in 15 ml fresh YPAc medium. For DMS footprinting a 1.5 ml portion of the resuspended cells was treated with 0.12 and 0.06% DMS for 2 min at room temperature and DNA was isolated. As a control, purified DNA was dissolved in 0.3 M Tris-HCl, pH 7.5, and methylated with 0.1% DMS for 3 min. For UV-photo footprinting 1.5 ml portions of the resuspended cells were irradiated at 254 nm with a UV crosslinker (Stratagene or Funakoshi), using doses of 250, 500 or 1000 mJ/cm², and DNA was isolated. As a control, purified DNA in PBS buffer (0.2 M NaCl, 2.7 mM KCl, 15.3 mM Na₂HPO₄, 1.5 mM KH₂PO₄, pH 7.4) was irradiated with a dose of 120 mJ/cm². The sites of UV photoproducts and methylation were analyzed by multi-cycle primer extension mapping using Taq polymerase (Boehringer) as described previously (12,14,17,18). Primers used included: IM1 (-2173 to -2139), 5[prime]-TGC TGT TCT TTC CGC CAC GGC CCG TAT GGT GTT GG-3[prime]; IM2 (-2084 to -2050), 5[prime]-AGA AGG GTG GGG GTG TAT TTG CTG CAA GAA TGC GG-3[prime]; IM3 (-1880 to -1916), 5[prime]-TCA CTT TTT TCC TAC CCA CGT CTA TAT ATG CAA CGC C-3[prime]; IM4 (-1816 to -1850), 5[prime]-AAT GCA GGA CCC TAT TTC TTC ACG AGG GTG TAT CC-3[prime]. The numbers are relative to the ATG translation start site in IME1 as defined previously (8; GenBank accession no. M37188).

Gel mobility retardation assay

In vitro DNA binding assays employed a purified maltose binding protein-Rme1p fusion (MBP-Rme1p), constructed through cloning of a full-length RME1 PCR product between the XmnI and HindIII sites of plasmid pMAL-c2 (New England Biolabs). MBP-Rme1p was overexpressed in Escherichia coli strain TB1 and purified according to the manufacturer's instructions. Purity of MBP-Rme1p was checked by SDS-PAGE. Protein concentrations were determined by the method of bicinchonine-Cu⁺ chelation (Pierce) using bovine serum albumin as the standard.

The 404 bp RC fragment was generated by PCR using pAC153-4 as template and subcloned into the EcoRI site of a plasmid pBR322 derivative. The RC fragment was excised by digestion with EcoRI, purified by agarose gel electrophoresis and radiolabeled by filling-in with Klenow fragment in the presence of [[alpha]-³²P]dATP. An aliquot of 0.15 pmol ³²P-labeled RCfragment was mixed with 6 µg MBP-Rme1p in 50 mM Tris-HCl, pH 8.0, 50 mM KCl, 0.5 mM MgCl₂, 1.5 mM dithiothreitol, 8% glycerol, 0.25 mg/ml bovine serum albumin (total 30 µl) containing 2 µg poly(dI-dC)·poly (dI-dC) and 7.5 µg bovine serum albumin. Binding reactions were incubated on ice for 30 min, then subjected to electrophoresis on a 6% polyacrylamide/Tris-borate-EDTA buffer gel at 5°C. DNA bands were visualized by autoradiography. The sequences of competitor oligo duplex DNA were as follows: the -2030 oligo DNA, 5[prime]-GGA TGT CAA AAG AAC CTC AAG AAG TCC ACT AAA TGG-3[prime] (corresponding to IME1 -2051 to -2016); the -1990 oligo DNA, 5[prime]-TGC CGC GCT TAC ATT TTT AGC GAC TGC CGA AAA CG-3[prime] (corresponding to IME1 -2008 to -1974); the -1950 oligo DNA, 5[prime]-GGC TAA CTG CTG TAC CTC AAA AGC ATA AAA TTG TG-3[prime] (corresponding to IME1 -1970 to -1936).

Miscellaneous methods

Procedures for preparation of RNA, electrophoresis and Northern blot analysis have been described (5,6). [beta]-Galactosidase assays were performed on cells grown in YPAc or synthetic galactose medium to early exponential phase (8,19). Sporulation was measured by counting the number of asci per 200 cells after 24 h in sporulation medium (8). [beta]-Galactosidase and sporulation assay results are the average of three independent cultures of each strain. The yeast genome was searched with Patscan (http://www.mcs.anl.gov/home/overbeek/PatScan/HTML/patscan.html).

Identification of Rme1p binding sites in the chromosomal IME1 locus by in vivo footprinting

We have used DMS and UV-photo footprinting to analyze protein binding to genomic DNA in vivo. To identify Rme1p-dependent footprints we compared strains that overexpress functional and non-functional Rme1p derivatives (P_GAL1-RME1 and P_GAL1-rme1-213 strains, respectively). Prior studies indicate that the P_GAL1-rme1-213 product is defective in binding to DNA and repression (8).

DMS footprints of the IME1 non-coding strand near nt -2000 are shown in Figure 1A. Sites protected against modification in vivo were identified with the P_GAL1-RME1 strain by comparison of DMS-treated whole cells (lanes 5 and 7) and purified DNA (lanes 6 and 8). In vivo protections were observed at -2037, -2036, -2034, -1956, -1955, -1953 and -1947. These residues are all guanines, as expected from the fact that DMS modifies primarily N-7 of guanine (e.g. 15,16). None of the guanine residues were protected in vivo in the P_GAL1-rme1-213 strain (lanes 9 and 11 compared with lanes 10 and 12) or in an rme1 deletion mutant (data not shown). Similar analysis of the IME1 coding strand (Fig. 1B) reveals Rme1p-dependent protection at -2040 and -1959. These results indicate that there are two protein binding sites in this region of the IME1 promoter and detectable binding to both sites depends upon Rme1p. We refer to these sites as `the -2030 site' and `the -1950 site'.

Figure 1. In vivo dimethyl sulfate (DMS) footprints of the IME1 genomic locus upstream region. (A) Footprints of the non-coding strand. Lanes 1-4, sequencing ladders for C, T, A and G, whose reactions were terminated with the complementary dideoxynucleotides ddG, ddA, ddT and ddC respectively; lanes 5-8, strain AMP1122, in which functional Rme1p (PGAL1-RME1) was expressed; lanes 9-12, strain AMP1124, in which non-functional Rme1-213p (PGAL1-rme1-213) was expressed. The IM1 primer was used in analysis of both the -2030 and -1950 sites (lanes 5, 6, 9 and 10). The IM2 primer was used in analysis of the -1950 sites (lanes 7, 8, 11 and 12). (B) Footprints of the coding strand. Lanes 1-4, strain AMP1122, in which functional Rme1p (PGAL1-RME1) was expressed; lanes 5-8 strain AMP1124, in which non-functional Rme1-213p (PGAL1-rme1-213) was expressed. The IM4 primer was used in analysis of both the -2030 and -1950 sites (lanes 1, 2, 5 and 6). The IM3 primer was used in analysis of the -2030 site (lanes 3, 4, 7 and 8). The sequences of Rme1p binding sites are shown to the left of the gels. Nucleotides are numbered with respect to the IME1 translation start. Dots indicate sites protected from DMS modification in cells. * indicates a DMS hypersensitive site induced by Rme1p binding. Lanes marked C are samples from intact cells treated with DMS; lanes marked D are samples from purified DNA treated with DMS.

We noticed an additional difference between P_GAL1-RME1 and P_GAL1-rme1-213 strains. On the non-coding strand nt -2000 was hyper-reactive to DMS in the presence of functional Rme1p (Fig. 1A, lanes 5 and 7 compared with 9 and 11, nucleotide indicated by an asterisk). This nucleotide is an adenine residue, which is modified primarily at N-3 by DMS. This Rme1p-dependent modification may represent a third protein binding site in this region or a more general structural change in DNA between the -2030 and -1950 sites.

UV-photo footprints of the IME1 non-coding strand are shown in Figure 2. In vivo protection was observed for two sets of four consecutive thymines at -1952 to -1949 and -1944 to -1941 in the P_GAL1-RME1 strain (lanes 1 and 3 compared with 2 and 4) and not in the P_GAL1-rme1-213 strain (lanes 5 and 7 compared with 6 and 8). Protection of thymines is consistent with the idea that protein-DNA binding may alter rates of UV-stimulated dimerization of adjacent pyrimidines (e.g. 14,16). These results complement the conclusion from DMS footprinting that an Rme1p-dependent complex forms at the -1950 site. Failure to detect an Rme1p-dependent complex at the -2030 site may reflect reduced reactivity of the pairs of thymine residues at the -2030 site.

Figure 2. In vivo UV-photo footprints of the non-coding strand of the IME1 genomic locus upstream region. Lanes 1-4, strain AMP1122, in which functional Rme1p (P_GAL1-RME1) was expressed; lanes 5-8, strain AMP1124, in which non-functional Rme1-213p (P_GAL1-rme1-213) was expressed. The IM1 primer was used in analysis of both the -2030 and -1950 sites (lanes 1, 2, 5 and 6). The IM2 primer was used in analysis of the -1950 site (lanes 3, 4, 7 and 8). The sequences of Rme1p binding sites are shown to the left of the gels. + indicates sites protected from UV irradiation in cells. Other symbols are as described in the legend to Figure 1.

Binding of MBP-Rme1p to IME1 DNA in vitro

The -2030 and -1950 sites have very similar nucleotide sequences and Rme1p binds to the -2030 site in vitro (8). To determine whether Rme1p binds directly to both sites in vitro, we carried out gel retardation assays with ³²P-labeled RC DNA (IME1 nt -2146 to -1743) as a probe and Rme1p fused with maltose binding protein (MBP-Rme1p) purified from E.coli. MBP-Rme1p formed two electrophoretically distinguishable complexes with RC DNA (Fig. 3, lane 2 compared with lane 1) and the minor band with lower mobility may result from binding to both sites. The presence of 10 mM EDTA abolished formation of the Rme1p-DNA complex (data not shown), as expected for a zinc finger protein (20). Specificity of complex formation was determined through competition assays (Fig. 3). Binding reactions included a labeled RC DNA probe and an unlabeled competitor oligo duplex DNA with sequences of the -2030 site, the -1950 site or the -1990 region (to serve as a non-specific control). As seen in Figure 3, the -2030 and -1950 site oligo DNA inhibited RC DNA complex formation with similar efficiencies (lanes 3-6 and 7-10 respectively); the -1990 region oligo DNA did not (lanes 11-14). Competition assays thus indicate that Rme1p binds directly to both the -2030 site and the -1950 site in vitro.

Figure 3. Gel mobility retardation assay for MBP-Rme1p with the ³²P-labeled RC (repression cassette) fragment (IME1 -2146 to -1743). The labeled RC fragment was mixed with MBP-Rme1p with unlabeled oligo DNA as a competitor and then electrophoresed on a 6% polyacrylamide gel. Binding reactions contained: no MBP-Rme1p (lane 1); MBP-Rme1p (lanes 2-14); an excess molar ratio of unlabeled competitor, at either 1- (lanes 3, 7 and 11), 5- (lanes 4, 8 and 12), 25- (lanes 5, 9 and 13) or 125-fold (lanes 6, 10 and 14). Unlabeled double-stranded oligo DNA was used as competitor: the -2030 oligo DNA (IME1 -2051 to -2016) (lanes 3-6); the -1950 oligo DNA (IME1 -1970 to -1936) (lanes 6-10); the -1990 oligo DNA (IME1 -2008 to -1974) (lanes 11-14).

Effect of Rme1p site mutations on Rme1p binding and IME1 expression

We have used mutational analysis to evaluate the functional importance of each Rme1p binding site. Mutations were introduced into the chromosomal IME1 locus that disrupt the -2030 site (IME1-81 allele), the -1950 site (IME1-82 allele) or both sites (IME1-81,82 allele). We first determined effects of the mutations on binding of Rme1p through in vivo DMS footprinting in P_GAL1-RME1 strains (Fig. 4). Mutational disruption of the -2030 site permitted protection of the -1950 site (lanes 1 and 3 compared with 2 and 4) and disruption of the -1950 site permitted protection of the -2030 site (lane 5 compared with 6). We observed no protection at either mutational site in the single (lanes 1-6) or double mutants (lanes 7 and 8). These results indicate that Rme1p can occupy each site in vivo in the absence of the other site.

Figure 4. Effect of mutations in the Rme1p binding site in the IME1 genomic locus on binding of Rme1p. In vivo DMS footprinting of the non-coding strand of IME1 was performed as in Figure 1A. Lanes 1-4, strain AMP1423 (IME1-81 allele); lanes 5 and 6, strain AMP1424 (IME1-82 allele); lanes 7 and 8, strain AMP1425 (IME1-81,82 allele). The IM1 primer was used in analysis of both the -2030 and -1950 sites (lanes 1, 2 and 5-8). The IM2 primer was used in analysis of the -1950 sites (lanes 3 and 4). The sequences of Rme1p binding sites are shown to the left of the gels. Other symbols are described in the legend to Figure 1.

It should be noted that single (IME1-81 or IME1-82 alleles) or double (IME1-81,82 allele) mutations of Rme1p caused loss of the hyper-reactive adenine residue at -2000 (compare Figs 4 with 1A). Thus this hyper-reactive site is observed only when Rme1p binds to both sites. These results suggest that interaction occurs between two Rme1p-DNA complexes.

We determined the effect of Rme1p site mutations on IME1 expression through Northern blot analysis (Fig. 5). Diploids homozygous for P_GAL1-RME1 or P_GAL1-rme1-213 and different IME1 alleles were subjected to nitrogen starvation, which promotes IME1 transcript accumulation and sporulation. As seen in Figure 5, IME1 mRNA levels were undetectable in the P_GAL1-RME1 strain (lanes 1-4). In contrast, IME1 transcript increased after starvation of the P_GAL1-rme1-213 strain (lanes 5-8) and accumulation was maximal after starvation for 2-4 h (lanes 6 and 7). This result is consistent with previous findings (3,5,6). In the presence of functional Rme1p, mutation of the -2030 site permitted slight IME1 expression after starvation (lanes 9-12); mutation of the -1950 site permitted greater IME1 expression (lanes 13-16); mutation of both sites permitted still greater IME1 expression (lanes 17-20). We also examined the effect of Rme1p site mutations on sporulation in these strains. As shown in Figure 5, bottom, in the wild-type IME1 strains cells sporulated (99% sporulation) in the absence of Rme1p, whereas they did not sporulate (<0.1% sporulation) in the presence of Rme1p. In the presence of Rme1p mutation of the -2030 site permitted 12% sporulation in the cells; mutation of the -1950 site permitted 35% sporulation; mutation of both sites permitted 93% sporulation. Thus levels of IME1 mRNA in these strains were reflected in their relative sporulation abilities. These results indicate that both Rme1p binding sites are required for full repression of the chromosomal IME1 gene. In addition, the -1950 site alone confers greater repression than the -2030 site alone in the context of the IME1 upstream region.

Figure 5. Effect of genomic Rme1p binding site mutations on expression of IME1 mRNA and sporulation. RME1 was expressed from the GAL1 promoter. RNA was prepared 0, 2, 4 or 6 h after starvation of a/a strains. Northern blots were probed with the IME1 probe (5,6,9) and then stripped and reprobed with the control probe pC4/2, which hybidizes to an RNA unaffected by starvation or cell type (40). Sporulation was measured 24 h after cells were shifted to sporulation culture. The strains used were AMP1428, 1429, 1430, 1497 and 1431.

Insertion of the RC upstream of the CYC1-lacZ gene permits Rme1p to repress CYC1-lacZ expression (8,9). We used CYC1-lacZ expression assays to quantitate the effects of Rme1p site mutations on repression (Fig. 6). Without the RC, CYC1-lacZ expression was insensitive to Rme1p (plasmid pKB112); insertion of the RC permitted >50-fold repression by Rme1p (plasmid pAC153-4), consistent with a previous report (8). Mutation of the -2030 site permitted 17-fold repression (plasmid pBS8); mutation of the -1950 site permitted 3-fold repression (plasmid pBS7); mutation of both sites permitted <2-fold repression (plasmid pBS9). Because the -2030 site mutation caused a slight reduction in expression of CYC1-lacZ, we used deletion analysis to confirm the contribution of the -1950 site to repression. Two deletions that remove the -2030 site permitted 19- and 12.5-fold repression (plasmids pWL121 and 122); a deletion that does not remove the -2030 site did not affect repression (plasmid pWL120). These results also show that both Rme1p binding sites are essential for repression. They also confirm that the -1950 site has a greater role in repression than the -2030 site.

Figure 6. Effect of Rme1p binding site mutations and 5[prime] deletions of the RC on Rme1p-dependent repression of RC-CYC1-lacZ. The open box indicates the Rme1p binding sites, whereas the black box with X indicates mutation of the binding site. End-points are numbered with respect to the IME1 translation start. Strains that expressed P_GAL1-RME1 or P_GAL1-rme1-213 and contained these plasmids were assayed for [beta]-galactosidase activity. The strains used were AMP1122 and 1124. The cells were grown in synthetic galactose medium to early log phase. Numbers are averages of three determinations from three independent transformants. The range for all values is within ±20% of the means.

Table 2. Effect of mutations in the Rme1p binding sites in the IME1 upstream genomic locus on ime2-lacZ expression

Relevant genotype	[beta]-Galactosidase activity (Miller unitsa)		Fold repression
	RME1	rme1[Delta]5
Wild-type	26	393	15.1
IME1-81	64	289	4.5
IME1-82	80	378	4.7
IME1-81,82	175	248	1.4

^a[beta]-Galactosidase activities were determined using cells grown in YPAc medium. Numbers are averages of three determinations from three independent colonies. The range for all values is within ±20% of the means. The strains used were WL481, 482, 483, 484, 485, 486, 487 and 488.

Our studies thus far have examined repression by Rme1p that is overexpressed. To determine whether both sites are required for repression when Rme1p is expressed at natural levels, we compared RME1 and rme1Δ5 haploid strains that carry different IME1 alleles (Table ). IME1 expression was monitored through expression of an ime2-lacZ reporter gene (5-8), which is activated by Ime1p. As seen in Table 2, in a wild-type IME1 background ime2-lacZ expression was reduced 15-fold by Rme1p. Mutation of either the -2030 site or the -1950 site caused ime2-lacZ expression to be reduced only 5-fold by Rme1p; mutation of both sites relieved the effect of Rme1p almost entirely. Therefore, both Rme1p binding sites contribute to repression of IME1 at natural Rme1p levels. However, these results yield no distinction between the sites in repression.

Effect of rgr1 and sin4 mutations on Rme1p binding and repression

Rgr1p and Sin4p are required for Rme1p-dependent repression (9). Failure of Rme1p to repress in rgr1 or sin4 mutants may result from failure of Rme1p to bind to DNA or from failure of Rme1p to exert repression once bound. We have examined Rme1p binding through in vivo DMS footprinting to distinguish between these models. We found the same DMS protection pattern in rgr1-100, sin4[Delta] and rgr1-100 sin4[Delta] (double mutant) strains (Fig. 7) that we had observed in the wild-type (Fig. 1A, lanes 5-8). In addition, as observed in the wild-type strain (Fig. 1A), the adenine residue at -2000 was hyper-reactive to DMS in rgr1-100 and sin4[Delta] strains (nucleotide indicated by an asterisk in Fig. 7). Also, Rme1p-dependent UV-photo footprints were observed at the -1950 site in rgr1-100 and sin4[Delta] mutant strains (data not shown). These results indicate that Rgr1p and Sin4p are not simply required for Rme1p to bind to DNA.

Figure 7. Effect of rgr1 and sin4 mutations on binding of Rme1p. In vivo DMS footprinting of the non-coding strand of IME1 was performed as in Figure 1A. Lanes 1-4, strain AMP1420 (rgr1-100); lanes 5-8, strain AMP1396 (sin4[Delta]); lanes 9-12, strain AMP1426 (rgr1-100 sin4[Delta]). The IM1 primer was used in analysis of both the -2030 and -1950 sites (lanes 1, 2, 5, 6, 9 and 10). The IM2 primer was used in analysis of the -1950 site (lanes 3, 4, 7, 8, 11 and 12). Symbols are as described in the legend to Figure 1.

We checked CYC1-lacZ expression in P_GAL1-RME1 strains with rgr1-100 or sin4[Delta] single mutations or with rgr1-100 sin4[Delta] double mutations under the same growth conditions as employed for in vivo DMS footprinting. Either mutation permitted expression of CYC1-lacZ in the presence of Rme1p (1.4- and 1.8-fold repression respectively), in agreement with previous results (9). The rgr1-100 sin4[Delta] double mutant had a similar phenotype (1.4-fold repression). This finding suggests that Rgr1p and Sin4p act in the same genetic pathway for repression.

DISCUSSION

Previous studies argued that Rme1p acts directly as a repressor of IME1 (8), based upon identification of the -2030 Rme1p binding site upstream of IME1. However, mutational disruption of the -2030 site had little effect on repression of IME1. This finding, combined with the observation that Rme1p can activate transcription (8,11) and the unusually large distance between the Rme1p site and the IME1 gene, raised the possibility that Rme1p might repress IME1 through both direct and indirect effects. Studies reported here now explain and extend those earlier observations. We find that the IME1 upstream region has two Rme1p binding sites and that the presence of either site is sufficient for partial repression of IME1. Our results also establish that Rme1p can bind to DNA, but not repress, in rgr1 and sin4 mutants.

Rme1p-DNA interaction

The -2030 Rme1p site had been defined from -2044 to -2024 (5[prime]-AAA AGA ACC TCA AGA AGT CCA-3[prime]) through chemical nuclease footprinting of an electrophoretically isolated Rme1p-DNA complex (8). We have now narrowed the -2030 Rme1p site to nt -2040 to -2030 and identified a second binding site from -1959 to -1949, based upon in vivo footprinting. The -2040 to -2030 segment is in agreement with previous mutational analysis, which showed that mutations that disrupted binding all lay in the -2035 to -2032 segment (8). It also permits a simple alignment with the second binding site in IME1 (Fig. 8). Gel retardation assays reported by Toone et al. (11) indicated that Rme1p binds to a fragment containing nt -515 to -614 from the CLN2 gene. As shown in Figure 8, the -561 to -551 CLN2 promoter segment strongly resembles the sites we identified in the IME1 promoter. Four guanine contacts revealed by DMS footprinting are present in all three Rme1p sites. In addition, a tract of four thymines that are protected from UV are common to the IME1 -1950 and CLN2 sites. Failure to detect UV protection for this segment of the IME1 -2030 site may result from reduced reactivity of the interrupted thymine tract, because the RRE4 mutant -2030 site (GAACCTCAAAAA) (8) has increased UV reactivity in the absence of Rme1p and shows UV protection by Rme1p that is comparable with the -1950 site (unpublished results).

Figure 8. Comparison of Rme1p binding sites. Data are shown for the -2030 and -1950 sites, with filled circles representing sites protected from methylation by DMS (see Fig. 1) and plus signs denoting sites protected from UV-induced thymine dimer formation (see Fig. 2). The Rme1p recognition site in the CLN2 promoter is defined by similarity to the -2030 and -1950 sites (11). A consensus sequence for Rme1p binding is shown at the bottom, corresponding to the underlined IME1 and CLN2 sequences. W, A or T; R, G or A.

Comparison of the three Rme1p binding sites yields the consensus sequence 5[prime]-GWACCTCAARA-3[prime] (Fig. 8; W, A or T; R, G or A). A search of the yeast genome reveals 36 such sites. Aside from the IME1 -2030 and -1950 sites, the only additional site on chromosome X lies 95 kbp from IME1. A search that permits a mismatch reveals 1143 such sites, but again the -2030 and -1950 sites are the closest to IME1. These results confirm that Rme1p represses IME1 from the two identified binding sites.

In vivo footprinting indicates that Rme1p shares some features with other zinc finger proteins. Protection of N-7 of guanine residues from DMS indicates that Rme1p binds in the major groove of DNA, as do crystallized Cys₂-His₂ zinc finger proteins (21-23). In these cases arginine or lysine residues in the zinc finger [alpha]-helix play important roles in contact with guanine (21-23). Rme1p contains three zinc finger segments and may make such contacts through zinc fingers 2 and 3, which contain basic residues in predicted [alpha]-helical regions (7). However, another aspect of Rme1p-DNA interaction may reflect more novel features: for several other zinc finger proteins each zinc finger makes primary contacts in a 3 bp subsite (21-23). However, Rme1p-dependent DMS protection at the -1950 site extends over 13 bp. The extended footprint may result from binding of an accessory protein along with Rme1p or from DNA contacts with both the Rme1p zinc finger domains and an additional Rme1p domain.

Role of Rgr1p and Sin4p in repression

Rgr1p and Sin4p are required for Rme1p-dependent repression (9). Herein we have shown that single and double mutations in RGR1 and SIN4 do not affect binding of Rme1p in vivo. Thus, binding of Rme1p to DNA is not sufficient for repression. This observation argues that repression does not result from a passive mechanism, such as competition between Rme1p and a hypothetical transcriptional activator for the same site. Our study of effects of binding of Rme1p to the RC further substantiates the view that Rme1p is an active repressor: Rme1p prevents activator binding to sites over 250 bp from the Rme1p binding site (12). Rgr1p and Sin4p may play a direct role in this repression mechanism; for example they may be recruited to the promoter region by interaction with Rme1p, much as Ssn6p and Tup1p act to repress transcription when recruited to target promoters by [alpha]2-Mcm1p, a1-[alpha]2 and Mig1p (24-27). Alternatively, they may act more indirectly, for example through their effects on global chromatin structure (28-32), integrity of which may be required for Rme1p to exclude activators.

The rgr1 sin4 double mutant has a repression defect similar to either single mutant. This finding suggests that Rgr1p and Sin4p act in the same genetic pathway for repression. Recent studies indicate that Rgr1p and Sin4p lie in an RNA polymerase II-associated subcomplex, called the mediator, with Gal11p and a protein called p50 (33). Our results are consistent with the idea that the mediator is required for repression by Rme1p.

Classes of yeast repressors

The general properties of yeast transcriptional repressors fall into two classes, which we call `proximal' and `distal'. Proximal repressors, such as [alpha]2-Mcm1p (17,24,25), a1-[alpha]2 (24), Mig1p (26,34,35), Rox1p (36) and Ume6p (19,37,38), recognize DNA sites that lie within 300 bp of the 5[prime]-ends of target genes. Introduction of the binding sites for these repressors into heterologous promoters is sufficient to cause repression. For [alpha]2-Mcm1p, it was reported that repression does not result from inability of transcriptional activators to bind to nearby sites (13). In contrast, distal repressors recognize DNA sites that may lie several kilobase pairs from the 5[prime]-ends of target genes. This class includes Rme1p and Rap1p (reviewed in 39), the repressor of silent mating-type loci and one component of the telomeric silencer. Introduction of the binding sites for these repressors into heterologous promoters causes transcriptional activation; repression depends upon additional sequences or DNA binding proteins. Repression by both Rme1p and Rap1p results in inability of DNA binding proteins to interact with their cognate sites. Therefore, it is possible that distal repressors achieve repression through analogous mechanisms.

ACKNOWLEDGEMENTS

We thank the members of the Mitchell laboratory for helpful advice and discussions and Dr Y.Shibusawa for his support in the course of this work. This work was supported by a Grant-in Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan to M.S., by PHS grant GM39531 and an American Cancer Society Faculty Research Award to A.P.M. and by a research grant from the Human Frontier Science Program Organization.

REFERENCES

1. Atcheson,C.L. and Esposito,R.E. (1993) Curr. Opin. Genet. Dev., 3,736-744. MEDLINE Abstract

2. Kupeic,M., Bryers,B., Esposito,R.E. and Mitchell,A.P. (1997) In Pringle,J.R., Broach,J.R. and Jones,E.W. (eds), The Molecular and Cellular Biology of the Yeast Saccharomyces. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, Vol. 3, pp. 889-1036.

3. Mitchell,A.P. (1994) Microbiol. Rev., 58, 56-70. MEDLINE Abstract

4. Kassir,Y., Granot,D. and Simchen,G. (1988) Cell, 52, 853-862. MEDLINE Abstract

5. Smith,H.E. and Mitchell,A.P. (1989) Mol. Cell. Biol., 9, 2142-2152. MEDLINE Abstract

6. Smith,H.E., Su,S.S.Y., Neigeborn,L., Driscoll,S.E. and Mitchell,A.P. (1990) Mol. Cell. Biol., 10, 6103-6113. MEDLINE Abstract

7. Covitz,P.A., Herskowitz,I. and Mitchell,A.P. (1991) Genes Dev., 5,1982-1989. MEDLINE Abstract

8. Covitz,P.A. and Mitchell,A.P. (1993) Genes Dev., 7, 1598-1608. MEDLINE Abstract

9. Covitz, P. A., Song,W. and Mitchell,A.P. (1994) Genetics, 138, 577-586. MEDLINE Abstract

10. Mitchell,A.P. and Herskowitz,I. (1986) Nature, 319, 738-742. MEDLINE Abstract

11. Toone,W.M., Johnson,A.L., Banks,G.R., Toyn,J. H., Stuart,D., Wittenberg,C. and Johnston,L.H. (1995) EMBO J., 14, 5824-5832. MEDLINE Abstract

12. Shimizu,M., Li,W., Shindo,H. and Mitchell,A.P. (1997) Proc. Natl. Acad. Sci. USA, 94, 790-795. MEDLINE Abstract

13. Redd,M.J., Stark,M.R. and Johnson,A.D. (1996) Mol. Cell. Biol., 16, 2865-2869. MEDLINE Abstract

14. Axelrod,J.D. and Majors, J. (1989) Nucleic Acids Res., 17,171-183. MEDLINE Abstract

15. Giniger,E., Varnum,S.M. and Ptashne,M. (1985) Cell, 65, 767-774.

16. Murphy,M.R., Shimizu,M., Roth,S.Y., Dranginis,A.M. and Simpson,R.T. (1993) Nucleic Acids Res., 21, 3295-3300. MEDLINE Abstract

17. Roth,S.Y., Shimizu,M., Johnson,L., Grunstein,M. and Simpson,R.T. (1992) Genes Dev., 6, 411-425. MEDLINE Abstract

18. Shimizu,M., Roth,S.Y., Szent-Gyorgyi,C. and Simpson,R.T. (1991) EMBO J., 10, 3033-3041. MEDLINE Abstract

19. Bowdish,K.S., Yuan,H.E. and Mitchell,A.P. (1995) Mol. Cell. Biol., 15, 2955-2961. MEDLINE Abstract

20. Pedon,P.V., Ghirlando,R., Clore,G.M., Gronenborn,A.M., Felsenfeld,G. and Omichinski,J.G. (1996) Proc. Natl. Acad. Sci. USA, 93, 2822-2826.

21. Fairall, L., Schwabe,J.W.R., Chapman,L., Finch,J.T. and Rhodes,D. (1993) Nature, 366, 483-487. MEDLINE Abstract

22. Pavletich,N.P. and Pabo,C.O. (1991) Science, 252, 809-817. MEDLINE Abstract

23. Pavleitch,N.P. and Pabo,C.O. (1993) Science, 261, 1701-1707.

24. Keleher,C.A., Redd,M.J., Schultz,J., Carlson,M. and Johnson,A.D. (1992) Cell, 68, 709-719. MEDLINE Abstract

25. Komachi,K., Redd,M.J. and Johnson,A.D. (1994) Genes Dev., 8, 2857-2867. MEDLINE Abstract

26. Treitel,M.A. and Carlson,M. (1995) Proc. Natl. Acad. Sci. USA, 92, 3132-3136 MEDLINE Abstract

27. Tzamarias,D. and Struhl,K. (1995) Genes Dev., 9, 821-831. MEDLINE Abstract

28. Jiang,Y.W. and Stillman,D.J. (1992) Mol. Cell. Biol., 12, 4503-4514. MEDLINE Abstract

29. Jiang,Y.W. and Stillman,D.J. (1995) Genetics, 140, 103-114. MEDLINE Abstract

30. Jiang,Y.W., Dohrman,P.R. and Stillman,D.J. (1995) Genetics, 140, 47-54. MEDLINE Abstract

31. Harashima,S., Mizuno,T., Mabuchi,H., Yoshimitsu,S., Ramesh,R., Hasebe,M., Tanaka,A. and Oshima,Y. (1995) Mol. Gen. Genet., 247, 716-725. MEDLINE Abstract

32. Macatee,T., Jiang,Y.W., Stillman,D.J. and Roth,S.Y. (1997) Nucleic Acids Res., 25, 1240-1247. MEDLINE Abstract

33. Li,Y., Bjorklund,S., Jiang,Y.W., Kim,Y.-J., Lane,W.S., Stillman,D.J. and Kornberg,R.D. (1995) Proc. Natl. Acad. Sci. USA, 92, 10864-10868. MEDLINE Abstract

34. Flick,J.S. and Johnston,M.(1992) Genetics, 130, 295-304. MEDLINE Abstract

35. Nehlin,J.O., Carlberg,M. and Ronne,H. (1991) EMBO J., 10, 3373-3377. MEDLINE Abstract

36. Balasubramanian,B., Lowry,C.V. and Zitomer,R.S. (1993) Mol. Cell. Biol., 13,6071-6078. MEDLINE Abstract

37. Park,H.D., Luche,R.M. and Cooper,T.G. (1992) Nucleic Acids Res., 20, 1909-1915. MEDLINE Abstract

38. Strich, R., Surosky,R.T., Steber,C., Dubois,E., Messenguy,F. and Esposito,R.E. (1994) Genes Dev., 8, 796-810. MEDLINE Abstract

39. Shore,D. (1994) Trends Genet., 10, 408-412. MEDLINE Abstract

40. Law, D.T.S. and Segall,J. (1988) Mol. Cell. Biol., 8, 912-922. MEDLINE Abstract

---

*To whom correspondence should be addressed at School of Pharmacy, Tokyo University of Pharmacy and Life Science, 1432-1 Horinouchi, Hachioji, Tokyo 192-0392, Japan. Tel: +81 426 76 4546; Fax: +81 426 76 4542; Email: shimizum@ps.toyaku.ac.jp

---

This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 13 May 1998
Copyright©Oxford University Press, 1998.
CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				D. van Dyk, G. Hansson, I. S. Pretorius, and F. F. Bauer Cellular Differentiation in Response to Nutrient Availability: The Repressor of Meiosis, Rme1p, Positively Regulates Invasive Growth in Saccharomyces cerevisiae Genetics, November 1, 2003; 165(3): 1045 - 1058. [Abstract] [Full Text] [PDF]

				M. Shimizu, K. Takahashi, T. M. Lamb, H. Shindo, and A. P. Mitchell Yeast Ume6p repressor permits activator binding but restricts TBP binding at the HOP1 promoter Nucleic Acids Res., June 15, 2003; 31(12): 3033 - 3037. [Abstract] [Full Text] [PDF]

				P. D. Rogers and K. S. Barker Genome-Wide Expression Profile Analysis Reveals Coordinately Regulated Genes Associated with Stepwise Acquisition of Azole Resistance in Candida albicans Clinical Isolates Antimicrob. Agents Chemother., April 1, 2003; 47(4): 1220 - 1227. [Abstract] [Full Text] [PDF]

				A. Blumental-Perry, W. Li, G. Simchen, and A. P. Mitchell Repression and Activation Domains of Rme1p Structurally Overlap, but Differ in Genetic Requirements Mol. Biol. Cell, May 1, 2002; 13(5): 1709 - 1721. [Abstract] [Full Text] [PDF]

				M. Shimizu, A. Murase, M. Hara, H. Shindo, and A. P. Mitchell A C-terminal Segment with Properties of alpha -Helix Is Essential for DNA Binding and in Vivo Function of Zinc Finger Protein Rme1p J. Biol. Chem., September 28, 2001; 276(40): 37680 - 37685. [Abstract] [Full Text] [PDF]

This Article Abstract Print PDF (551K) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (14) Request Permissions Commercial Re-use Guidelines for Open Access NAR Content Google Scholar Articles by Shimizu, M. Articles by Mitchell, A. P. Search for Related Content PubMed PubMed Citation Articles by Shimizu, M. Articles by Mitchell, A. P. Social Bookmarking          What's this?