Functional Dissection of the Global Repressor Tup1 in Yeast: Dominant Role of the C-Terminal Repression Domain

Corresponding author: Robert J. Trumbly, Medical College of Ohio, 3035 Arlington Ave., Toledo, OH 43614-5804., rtrumbly{at}mco.edu (E-mail)

Communicating editor: M. HAMPSEY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In the yeast Saccharomyces cerevisiae, Tup1, in association with Cyc8 (Ssn6), functions as a general repressor of transcription. Tup1 and Cyc8 are required for repression of diverse families of genes coordinately controlled by glucose repression, mating type, and other mechanisms. This repression is mediated by recruitment of the Cyc8-Tup1 complex to target promoters by sequence-specific DNA-binding proteins. We created a library of XhoI linker insertions and internal in-frame deletion mutations within the TUP1 coding region. Insertion mutations outside of the WD domains were wild type, while insertions within the WD domains induced mutant phenotypes with differential effects on the target genes SUC2, MFA2, RNR2, and HEM13. Deletion mutations confirmed previous findings of two separate repression domains in the N and C termini. The cumulative data suggest that the C-terminal repression domain, located near the first WD repeat, plays the dominant role in repression. Although the N-terminal repression domain is sufficient for partial repression, deletion of this region does not compromise repression. Surprisingly, deletion of the majority of the histone-binding domain of Tup1 also does not significantly reduce repression. The N-terminal region containing potential -helical coiled coils is required for Tup1 oligomerization and association with Cyc8. Association with Cyc8 is required for repression of SUC2, HEM13, and RNR2 but not MFA2 and STE2.

AWD repeat containing phosphoprotein (REDD et al. 1997 ), Tup1 is required for repression of transcription of many gene families in the yeast Saccharomyces cerevisiae. Tup1 associates with Cyc8 (= Ssn6) to form the Cyc8-Tup1 corepressor complex consisting of one Cyc8 and four Tup1 subunits (WILLIAMS et al. 1991 ; VARANASI et al. 1996 ). About 3% of yeast genes are derepressed in tup1 mutants, as determined by microarray analysis (DERISI et al. 1997 ). The target genes can be classified into at least nine groups, including glucose-repressible genes (TRUMBLY 1992 ), mating-type-regulated genes (MUKAI et al. 1991 ; KELEHER et al. 1992 ), DNA-damage-inducible genes (ZHOU and ELLEDGE 1992 ), oxygen-regulated genes (ZITOMER and LOWRY 1992 ), osmostress-inducible genes (MARQUEZ et al. 1998 ), a fatty-acid-regulated gene (FUJIMORI et al. 1997 ), flocculation-related genes (TEUNISSEN et al. 1995 ), sporulation-related genes (FRIESEN et al. 1997 ), and a meiosis-related gene (MIZUNO et al. 1998 ). According to the corepressor model, the Cyc8-Tup1 complex is recruited by pathway-specific DNA-binding proteins to mediate repression of these different gene families (KELEHER et al. 1992 ; TZAMARIAS and STRUHL 1995 ).

The Cyc8-Tup1 complex appears to repress transcription by two general mechanisms: (i) alteration of chromatin structure so as to mask DNA targets for activators or transcriptional factors and (ii) inhibition of activation of transcription. There are several lines of evidence for the first explanation. Simpson and co-workers (ROTH et al. 1990 ) found that the 2 operator was capable of controlling nucleosome positioning, which was directly implicated in transcriptional repression. This nucleosome positioning was dependent on both Cyc8 and Tup1, but tup1 mutations had more severe effects on chromatin structure than did cyc8 mutations (COOPER et al. 1994 ). Roth and co-workers showed that Tup1 interacts directly with the amino-terminal ends of histones H3 and H4 (EDMONDSON et al. 1996 ), and mutations in the N termini of H3 and H4 compromise the repression mediated by Tup1 in vivo (HUANG et al. 1997 ). Tup1 interacted more strongly with these histone tails in their deacetylated state (EDMONDSON et al. 1996 ), corresponding to the inverse relation between histone acetylation and repression observed in many systems. Furthermore, repression by Tup1 is accompanied by local histone deacetylation (BONE and ROTH 2001 ), and repression is eliminated by combined deletion of three specific genes encoding deacetylases (WATSON et al. 2000 ). Mutations in histone tails or histone deacetylases strongly decrease the association of Tup1 with its target genes, as assayed by chromatin immunoprecipitation (DAVIE et al. 2002 ).

In favor of the second model of repression, the Mat2 protein repressed transcription in vitro from 2 operators, and this repression was dependent upon Cyc8 and Tup1 (HERSCHBACH et al. 1994 ). Isolation of TATA-binding protein mutants that overcame repression by Cyc8-Tup1 showed that repression may work, at least in part, by inhibiting the function of transcriptional activators (GEISBERG and STRUHL 2000 ). As a specific example, Tup1 was shown to inhibit the ability of Mcm1 to activate expression of a-specific genes (GAVIN et al. 2000 ). As further support for this model, mutations affecting repression have been identified in genes encoding subunits of the RNA polymerase II Srb/Med complex (BALCIUNAS and RONNE 1995 ; KUCHIN and CARLSON 1998 ; LEE et al. 2000 ). Recent work has shown that Tup1 can interact directly with several subunits of the Srb/Med complex: Srb7 (GROMOLLER and LEHMING 2000 ), Srb10 (ZAMAN et al. 2001 ), and Hrs1/Med3 (PAPAMICHOS-CHRONAKIS et al. 2000 ). Disruption of the Tup1-Srb7 interaction produced substantial derepression of several Tup1 targets, suggesting that this interaction is an important pathway for repression by Tup1. Therefore it seems likely that the Cyc8-Tup1 complex can repress transcription by two different mechanisms: modification of chromatin structure and direct inhibition of transcriptional activation.

Several studies have defined the functional domains in Tup1. Tup1 can be divided into three regions: the first 72 amino acids (aa), the C-terminal seven WD repeats (340–713 aa), and the middle region in between. The first 72 amino acids are involved in formation of the Cyc8-Tup1 complex (TZAMARIAS and STRUHL 1994 ; CARRICO and ZITOMER 1998 ). Amino acids 73–385 associate with histones H3 and H4, and the region of high affinity to histone H3/H4 was located at 120–316 aa (EDMONDSON et al. 1996 ), providing a molecular basis for direct connection between the Cyc8-Tup1 complex and chromatin organization. It was claimed that the histone-binding domain corresponded to the repression domain in the middle region (EDMONDSON et al. 1996 ). In fact, the results of TZAMARIAS and STRUHL 1994 suggest that there are two independent repression domains, one in 73–200 aa and the other within 288–389 aa. Johnson and co-workers also pointed out that Tup1 has two repression domains, one in the amino terminus that also contains the Cyc8-association domain and one in the carboxyl terminus in a region overlapping with the first WD repeats (KOMACHI et al. 1994 ). So the total histone-binding domain overlaps with both the N- and C-terminal repression domains.

To identify the functional domains of Tup1, we created a library of XhoI linker insertions and internal in-frame deletion mutants within the TUP1 coding region. The Tup1 repression function was assayed by expression of five representative genes: SUC2, RNR2, MFA2, HEM13, and OLE1. We found that the main histone-binding domain in Tup1 can be deleted with little derepression of SUC2, RNR2, MFA2, and HEM13. Insertion mutations outside the WD repeats were wild type, while insertions within the WD repeats induced mutant phenotypes with differential effects on the target genes. The N-terminal 200 amino acids of Tup1, when overexpressed, nearly completely repress SUC2, RNR2, and OLE1, but act as a dominant negative with regard to MFA2, resulting in expression levels greater than those in a tup1 strain. However, deletion of the C-terminal repression domain but not the N-terminal repression domain alleviates repression, reinforcing the model from the linker insertion results that the C-terminal repression domain plays a dominant role for transcriptional repression in the context of Tup1.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains and plasmids:
Yeast strains BJ2168 (MATa leu2 pep4-3 prb1-1122 prc1-407 trp1 ura3-52), RTY418 (MAT his4-519 leu2-3 leu2-112 trp1-289 tup1-1::TRP1 ura3-52), and RTY535 (MATa leu2 pep4-3 prb1-1122 prc1-407 trp1 tup1-1::TRP1 ura3-52) were described earlier (WILLIAMS et al. 1991 ). RTY545 (MATa cyc8-1::LEU2 leu2 pep4-3 prb1-1122 prc1-407 trp1 tup1-1::TRP1 ura3-52) was derived from RTY535 by transformation with pDSB (TRUMBLY 1988 ). RTY510 was derived from the BJ2168 strain by transformation with pTXL63, which contains both TUP1 and CYC8 (WILLIAMS et al. 1991 ).

Yeast transformation was performed by an improved lithium acetate method (GIETZ et al. 1992 ). For Northern blots, yeast cells were grown in SD-uracil medium plus 5% glucose and harvested in log phase. The Escherichia coli strain XL1-Blue was used for plasmid manipulations.

Construction of linker insertion mutants:
The starting plasmid for mutagenesis was pUV35, which consists of the entire TUP1 gene on a 3.5-kb HindIII fragment subcloned into pRS306, a yeast-integrating plasmid carrying the URA3 gene (SIKORSKI and HIETER 1989 ). An NdeI site (CATATG) was created by site-directed mutagenesis at the start codon of TUP1 in pUV35 to facilitate subcloning to expression vectors, as described previously (VARANASI et al. 1996 ). Restriction fragments containing different regions of the TUP1 coding sequence were subcloned into pUC19 (YANISCH-PERRON et al. 1985 ), mutagenized by linker-tailing (LATHE et al. 1984 ), and then returned to pUV35 for expression in yeast. To create linker insertions the plasmid DNA was linearized by partial digestion with PvuII, FspI, RsaI, DpnI, or HaeIII in the presence of 20 µg ethidium bromide/ml, purified by agarose gel electrophoresis, and ligated with a 12-bp nonphosphorylated XhoI linker (5'-CTCGAGCTCGAG-3'). Excess linker was removed by gel filtration over a Sephadex G-50 column. The plasmids were religated after removing the unligated linker strands by denaturing at 75° for 5 min and hybridizing cohesive strands. The insertion positions were confirmed by digesting with appropriate restriction enzymes and by DNA sequencing to resolve ambiguous cases.

Construction of deletion mutants from insertion mutants:
Nucleotide positions for TUP1 refer to the published sequence, GenBank accesssion no. M31733 (WILLIAMS and TRUMBLY 1990 ). Appropriate pairs of plasmids carrying the linker insertion mutants were digested with SmaI and XhoI, and complementary restriction fragments were ligated to create internal deletions in the TUP1 coding sequence. The reading frame was adjusted for a potential +2 frameshift by completely filling in with Klenow DNA polymerase in the presence of all four dNTPs, and for a +1 frameshift by partial filling in by Klenow with only TTP and ligating with a 10-bp CGAGGTWACC adaptor, where W = A or T. Insertion of this adaptor was verified by digestion with BstEII. The plasmid pMB3 encoding a truncated Tup1 protein missing the N-terminal 51 aa was constructed by subcloning a 2251-bp PvuII-SacI fragment containing most of the TUP1 coding sequence into pVT100-U, a yeast expression vector with the ADH1 promoter (VERNET et al. 1987 ), so that the third Met codon of the Tup1 coding sequence is used as the start codon. The tup1673-713 mutation, which deletes the C-terminal WD repeat, was created by digesting at the EcoRI site at position 2525 in the TUP1 sequence (WILLIAMS and TRUMBLY 1990 ) and then filling in with Klenow DNA polymerase, introducing a stop codon at aa position 673.

Mutant N201 (Tup1-201stop) was created by site-directed mutagenesis using the oligonucleotide 5'-CCACCTCCCTAGGTTTCCGTCGCAC-3', which changes the codon at position 202 from CAG (glutamine) to TAG (stop). Mutant N254 (Tup1-254stop) was made by deleting the sequence between the BamHI site at codon 253 (made blunt by fill-in with Klenow polymerase) and the MscI site at codon 637. This deletion results in a frameshift encoding the N-terminal 254 amino acids of Tup1 followed by the tetrapeptide His-His-Thr-Lys at the C terminus.

For expression at the single-copy level, plasmids containing the mutant genes in pRS306 were linearized by digestion with StuI to direct integration at the URA3 locus and then transformed into RTY418. For expression as multicopy plasmids, the mutant genes were recloned as HindIII fragments into pEMBLYe24 (BALDARI and CESARINI 1985 ) and then transformed into RTY418.

Construction of D1, D2, and middle-deletion mutants (M2–M5):
Primers PstR (5'-TATAATCCATGGCAGCAGAATCGTCAGATAAACGGG-3'; NcoI site underlined, 3' position 1643 in TUP1 sequence) and PstL (5'-ATATATCCATGGCCCCATCATCCGACTTGTATATCCGTTCAG-3'; NcoI site underlined, 3' position 1854 in TUP1 sequence) were synthesized and a long PCR was performed using the TaqPlus precision PCR system (Stratagene, La Jolla, CA) and pEMBLYe24-TUP1 as template. The product was digested with NcoI and ligated to produce tup1ST (385-431; = D2). With pEMBLYe24-tup173-279 as template and with the same primers, tup173-279ST (= D1) was created. Both mutants were transformed into RTY418. The middle-deletion mutants M2–M5 were constructed based on two plasmids, pEMBLYe24-TUP1 and pEMBLYe24-TUP1(IN129) (insertion mutant with XhoI at codon 129). Primers Pa-30 (5'-GCCACGCGTTGCACCAGGATCACTACTTAG-3'; MluI site underlined, 3' position 1364 in TUP1 sequence), Pb-33 (5'-ACCACGCGTCGACTTTGGATCATACTTCAGTTG-3'; MluI site underlined, 3' position 1547 in TUP1 sequence), and P10-24 (5'-GTAACAGATCTATCTAATGAGCCG-3'; BglII site underlined, 3' position 2287 in TUP1 sequence) were synthesized and two PCR reactions using pEMBLYe24-TUP1 as template were performed with the primer pairs Pa-30–P10-24 and Pb-33–P10-24, respectively. The PCR products were cut with MluI and BglII and both were ligated into pEMBLYe24-TUP1 cut by the same two enzymes, thus creating tup173-279 (M2) and tup173-340 (M3). pEMBLYe24-TUP1(IN129) was cut by MluI and XhoI, filled in by the Klenow fragment and ligated to make tup173-129 (M5). To make tup1130-340 (M4), PCR was performed with primers Pb-33 and P10-24 using pEMBLYe24-TUP1 as template, and the PCR product after digestion with SalI and BglII was ligated into pEMBLYe24-TUP1(IN129) cut by XhoI and BglII. The mutations were confirmed by restriction digests and dideoxy sequencing. The D1, D2, and M2–M5 middle-deletion mutants were transformed into RTY418 to test for complementation of TUP1 function.

Protein extractions and Western blots:
Protein extracts from yeast cells were prepared as described previously (VARANASI et al. 1996 ). For Western blots, 20 µg of crude protein was resolved on either a 10% acrylamide SDS gel or nondenaturing 6% acrylamide gel, electroblotted onto nitrocellulose, and immunostained with either Cyc8 or Tup1 affinity-purified polyclonal antibodies as described by WILLIAMS et al. 1991 .

Yeast RNA isolation and Northern blotting:
RNA for Northern blots was isolated from early log phase cultures by the method of CHOMCZYNSKI and MACKEY 1995 with some modifications. Briefly, yeast cells were collected from log phase cultures by centrifugation. For each 0.2 g wet weight cells we added 2.4 ml denaturation solution (25 g guanidine thiocyanate dissolved in 33 ml 42 mM sodium citrate/0.83% w/v N-lauryl sarcosine/0.2 mM ß-mercaptoethanol) and 3.6 ml glass beads. The suspension was vortexed seven times for 1 min, followed by addition of 0.3 ml 2 M sodium acetate (pH 4.0) buffer and 2.88 ml phenol:chloroform:isoamyl alcohol (ratio of 25:24:1, buffered with 42 mM sodium citrate, pH 4.0). The mixture was vortexed 4 min, stored on ice 20 min, then centrifuged at 10,000 x g for 20 min at 4°. To the supernatant was added 1/2 volume high salt solution (1.2 M NaCl plus 0.8 M sodium citrate; CHOMCZYNSKI and MACKEY 1995 ) and 1/2 volume isopropanol. After incubation at -70° for 1 hr, the solution was centrifuged at 10,000 x g for 20 min. The RNA pellet was dissolved in 400 µl denaturation solution at 65° and cooled to room temperature. A total of 400 µl isopropanol was added to reprecipitate the RNA. The RNA was vacuum dried, dissolved in 400 µl Formazol (CHOMCZYNSKI 1992 ), and stored at -20°. The final RNA concentration was 0.5–1.0 µg/µl.

Northern blots were performed as described (TRUMBLY 1986 ) except that formaldehyde was included in the running buffer to a final concentration of 0.22 M. Approximately 10 µg total RNA was loaded per lane on a 1% agarose gel containing 0.22 M formaldehyde. The RNA was transferred to nitrocellulose membrane, prehybridized at 42° for 1 hr, and hybridized at 42° for 16 hr in the presence of 50% formamide with each probe labeled with ³²P by the random primer method. The SUC2 probe was a 4.0-kb EcoRI fragment from pRT1 (WILLIAMS et al. 1985 ), and the MFA2 probe was a 1.7-kb HindIII fragment from pSM29 (MICHAELIS and HERSKOWITZ 1988 ). The following probes were amplified by PCR of yeast genomic DNA: a 1.5-kb ACT1 fragment, primers 5'-CAAGAAGAAAAAGAAAAGGTC-3' and 5'-AAGCAGAGATTAGAAACACT-3'; a 1.0-kb RNR2 fragment, primers 5'-AGGAAATGGAAAAGGAGGAAC-3' and 5'-TTGAAGAGA CTGCGTAAAAAG-3'; a 2.0-kb STE2 fragment, primers 5'-ACACCAAAGATTCAAGATAAG-3' and 5'-CTCGTAAAAGCAAAGGTGGTT-3'; and a 1.2-kb HEM13 fragment, primers 5'-AACATAGCAAACCGAACTCTT-3' and 5'-TCATAAATAGGACGCAATAAT-3'. The films from the Northern blots were scanned on a flat bed scanner, and the band densities were quantified using Kodak 1D image analysis software. The band densities for each sample were normalized to the actin band and then expressed as a percentage of the tup1 null mutant value.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Construction of insertion and deletion mutants:
Several previous studies have addressed the structure-function relationships of Tup1 by analyzing mutants (KOMACHI et al. 1994 ; TZAMARIAS and STRUHL 1994 ; CARRICO and ZITOMER 1998 ). These studies have provided a general picture of the functional organization of Tup1. The N terminus is required for oligomerization of Tup1 to a tetramer, as well as association with Cyc8 to form the Cyc8-Tup1 corepressor complex. Deletion analysis has shown that there are two major repression domains, one N-terminal and the other C-terminal. The precise location, mechanism, and relative importance of these two domains have not been established. Here we present characterization of a large collection of insertion and deletion mutants designed to clarify some of these questions.

A collection of 12-bp XhoI linker insertion and deletion mutants distributed throughout the TUP1 coding sequence were constructed to identify the functional domains of the Tup1 protein. It was reasonable to believe that distinct functional domains would be identified, since the Tup1 protein must perform several roles: association with Cyc8 to form the Cyc8-Tup1 complex; interaction with the 2, Mig1, and other repressors; and repression of transcription by one or more mechanisms. The structural features of the Tup1 protein are shown in Fig 1A. There are seven WD repeat domains in the C-terminal region of the protein, two polyglutamine domains in the N-terminal region, and a serine-threonine-rich region between WD1 and WD2.

View larger version (80K): In this window In a new window Download PPT slide   Figure 1. Construction and expression of TUP1 insertion mutants. (A) The coding region and the domains of Tup1 are shown. Numbers under each domain correspond to the position of the first amino acid within the domain. The diagrammed domains are the mixed charge cluster (MCC: aa 66–89), glutamine domains 1 and 2 (Q1: aa 97–118 and Q2: aa 181–198), a serine-threonine-rich domain (ST: aa 393–419), and WD repeats 1 (340–385) and 2–7 (aa 431–713). The positions of the XhoI linker insertions are designated by the codon where the restriction enzyme cuts or, if the enzyme cuts between codons, the codon before the restriction site. Details for construction of the insertion mutants are given in MATERIALS AND METHODS. (B). Expression of the mutant proteins. Extracts (10 µg protein per lane) from cells expressing the mutant proteins were loaded on 6% acrylamide native gels and electroblotted and were probed with either Cyc8 or Tup1 antibodies. The strains were RTY535 (tup1) transformed with the mutated TUP1 derivatives carried on plasmid pEMBLYe24. The negative control strain (tup1) had the vector alone, and the Tup1 + Cyc8 strain was RTY510, overexpressing both Cyc8 and Tup1.

The insertion mutants were transformed into the tup1 deletion strain RTY535 on a multicopy episomal plasmid. We chose to express the mutant Tup1 proteins from multicopy plasmids in most experiments since we observed that many Tup1 deletion derivatives were expressed at low levels, and lack of function might be due in some cases to this low expression level rather than to deletion of a functional domain. It is possible that some partially defective mutants would behave as wild type due to their overexpression. Expression of the Tup1 insertion mutant proteins, tested by Western blots of SDS-polyacrylamide gels, was similar to that of the wild-type protein expressed from the same vector (data not shown). The ability of the mutant proteins to form a stable complex with Cyc8 was tested by electrophoresis of yeast cell extracts on 6% acrylamide native gels, followed by Western blotting using both Cyc8 and Tup1 antibodies as probes (Fig 1B). In the control strain overexpressing both Cyc8 and Tup1, the Cyc8-Tup1 complex was the major band, and the faster migrating Cyc8 monomer and Tup1 oligomer (Fig 1B, lane 1) were also evident. None of the insertions affected complex formation except IN48. In the other insertion mutants, only the intact complex was detected by the Cyc8 antibodies (Fig 1B). On the other hand, two bands were recognized by the Tup1 antibodies in most of the insertion mutants, the upper band representing the intact complex and the lower band representing the Tup1 oligomer. The latter species was present because Tup1 was overexpressed 5- to 10-fold in these strains, while Cyc8 was expressed as a single copy. In the mutant IN48 the band corresponding to the complete complex was not present. The Cyc8 antibodies recognize a smear extending from the position of the Tup1 oligomer to the Cyc8 monomer (Fig 1B, lane 3). The Tup1 antibodies recognize primarily a band corresponding to the Tup1 oligomer, but also a smear from the Tup1 oligomer to monomer positions (Fig 1B, lane 3). It appears that the IN48 mutation destabilizes both the oligomerization of Tup1 and the association of Cyc8 with Tup1. The intact complex is presumably present in vivo and disassociates during electrophoresis, since the IN48 mutation does not induce a mutant phenotype (see below).

The effect of the linker insertion mutations on expression of four target genes is shown in Fig 2. Repression of these four target genes requires Cyc8-Tup1, but is mediated by different repressor proteins: MFA2 by Mat2, SUC2 by Mig1, RNR2 by Crt1, and HEM13 by Rox1. In general, only linker insertions within the WD repeat domain showed significant derepression of any of the target genes. The mutation with the greatest effect was at position 379 in WD1, with near complete derepression of all targets with the exception of MFA2. IN342 had complete derepression of HEM13, partial derepression of RNR2, and minor derepression of SUC2 and MFA2. The other three insertion mutations in the WD region, IN515, IN572, and IN644, displayed significant expression of RNR2 and lesser effects on the other gene targets, although MFA2 and HEM13 were partially derepressed in IN644. These results suggest that the structure of the WD repeats is very sensitive to the perturbations introduced by the linker insertions. In contrast, insertions in other regions of the protein did not disrupt function, suggesting that the other domains of the protein are less sensitive to these perturbations.

View larger version (70K): In this window In a new window Download PPT slide   Figure 2. Effect of Tup1 insertion mutations on expression of target genes. Northern blots were performed with total RNA from yeast cells harboring different insertion mutations in TUP1. The blots were hybridized with probes for the targets MFA2, SUC2, RNR2, and HEM13 and ACT1 as a loading control. The yeast strains were RTY535 (tup1) transformed with the wild-type TUP1 or the insertion mutations carried on plasmid vector pEMBLYe24, or the vector itself. The values below the bands represent the band densities normalized to actin, with the tup1 null mutant as 100%.

A striking feature of the insertions in the WD region is their differential effects on the different gene targets. With the exception of IN644, these mutations have very little effect on repression of MFA2. RNR2, on the other hand, is significantly derepressed in all of the insertions in the WD domains. These results are surprising, since one might have expected greater effects of WD mutations on MFA2 expression, since Mat2 interacts directly with the WD domain of Tup1, while the other repressors interact primarily with Cyc8.

Interestingly, IN48, which destabilizes the Cyc8-Tup1 complex in vitro, has no apparent effect on repression of the target genes. Presumably, the complex is more stable under intracellular conditions than during electrophoresis.

Effect of tup1 deletion mutations on protein expression and complex formation:
Deletion mutants were constructed by combining complementary segments of plasmids containing the insertion mutants. The deletion mutants were expressed from multicopy plasmids as explained above (Fig 3A, not all mutants are shown). The mutant 1-51 was expressed at higher levels since it was expressed from the strong ADH1 promoter, while the other mutants were expressed from the native TUP1 promoter.

View larger version (64K): In this window In a new window Download PPT slide   Figure 3. Expression and complex formation of tup1 deletion mutants. (A) SDS gels. Mutant proteins (10 µg crude protein per lane) were resolved on 10% polyacrylamide gels containing SDS, electroblotted, and immunostained with Tup1 or Cyc8 antibodies. (B) Effect of tup1 mutations on complex formation. Mutant proteins were resolved on 6% acrylamide native gels, electroblotted, and stained with either Cyc8 or Tup1 antibodies.

Compared to the insertion mutations, the tup1 deletion mutations had much more dramatic effects on complex formation. Mutations that removed segments of the N terminus, 1-51, 48-129, and 48-431, completely eliminated complex formation as assayed by gel electrophoresis, since the complex band was absent, and all Cyc8 protein migrated as a monomer (Fig 3B, lanes 3, 4, and 7). However, any conclusions regarding the mutant 48-431 are weakened by the very low amounts of this mutant protein that were expressed (Fig 3A). The mutant 129-282 exhibited a complex of increased mobility compared to the wild-type complex, as expected (Fig 3B, lane 5). In the N-terminal deletion mutants the Tup1 protein migrated primarily free from Cyc8 and with greater electrophoretic mobility than the wild-type Tup1 oligomer (Fig 3B, lanes 3, 4, and 5). As shown previously, the 1-51 and 48-129 mutants behaved as monomers, and the 129-282 mutant and wild-type Tup1 behaved as a trimer or tetramer (VARANASI et al. 1996 ). The finding that deletions in the N terminus that disrupted Tup1 oligomerization also prevented association with Cyc8 reinforces the correlation seen with the IN48 mutation and suggests that Tup1 oligomerization may be a prerequisite for association with Cyc8.

Deletions at the C terminus of Tup1 that removed various combinations of WD repeats had similar effects: the complex was present, but appeared faint and diffuse. Although these diffuse bands might not by themselves be considered convincing evidence for complex formation, the absence of free Cyc8 protein on the native gels argues strongly for complex formation in these mutants. The Cyc8 protein was present in these extracts at normal levels, without apparent proteolysis (Fig 3A, bottom). Therefore the WD repeats do not appear to play a necessary role in the association of Tup1 with Cyc8.

Effect of tup1 deletions on repression of target genes:
The TUP1 and CYC8 genes are required for repression of transcription mediated by several diverse repressor proteins, including Mig1, which represses glucose-repressible genes; Mat2, a repressor of a-specific genes; and Mata1/Mat2, repressors of haploid-specific genes. Fig 4B shows the effects of the N-terminal and C-terminal tup1 deletions on mRNA levels of target genes representative of these four classes: SUC2 (glucose repression), MFA2 (a-specific gene), RNR2 (DNA-damage-inducible gene), and HEM13 (oxygen-regulated gene). The expression levels were normalized to the value of the tup1 deletion strain, except for MFA2, which was normalized to the maximum value, because of the anomalously low MFA2 value of the tup1 deletion strain in this experiment.

View larger version (158K): In this window In a new window Download PPT slide   Figure 4. Effect of tup1 C- and N-terminal deletions on repression of representative target genes. (A) Diagram showing tup1 C- and N-terminal deletions. (B) Northern blots. Approximately 10 µg total RNA prepared from yeast strains harboring the wild-type TUP1, the vector alone, and the deletion mutants was analyzed by Northern blots. (C) Northern blots showing the effect of selected tup1 mutations on regulation of MFA2 and STE2.

The tup11-51 mutation shows partial derepression of all target genes with the exception of MFA2. The 48-129 mutation induces high-level expression of all the target genes, again with the exception of MFA2. Both of these N-terminal deletion mutations eliminate both Tup1 oligomerization and association with Cyc8 as assayed on native gels (Fig 3B). Since 1-51 only partially reduces repression, it probably does not completely prevent association of Cyc8 and Tup1 in vivo, allowing sufficient complex formation for repression of its target genes. On the other hand, 48-129 must have a more drastic effect on complex formation in vitro and in vivo. TZAMARIAS and STRUHL 1994 showed that the first 72 aa of Tup1 were sufficient to mediate association with Cyc8. Our results indicate that sequences both N- and C-terminal to aa 48 are involved in association with Cyc8, but that sequences C-terminal to aa 48 are more critical.

All internal deletions that remove portions of the WD domains (mutants D3, D4, C2, C3, C6, and C8) produced 50% derepression of SUC2, HEM13, and RNR2 and complete derepression of MFA2 (Fig 4B). Deletion of the region between WD1 and WD2 (the ST region, residues 385–431) in mutant D2 did not diminish repression, suggesting that this region, which is unique to Tup1 from S. cerevisiae and the closely related Kluy-veromyces lactis, is not important for repression. But generally, mutations that delete portions of the WD domains cause dramatic loss of repression for most targets. However, since these mutations probably disrupt the compact toroidal WD structure, they cannot precisely localize the repression domains, but point to the general importance of the C-terminal region in repression.

The effect of the tup1 deletions on regulation of MFA2 was markedly different from the effect on the other targets. To generalize this pattern, we compared the expression of two genes repressed by Mat2, MFA2 and STE2. The expression levels of MFA2 and STE2 were similar in three representative tup1 deletions: N2, D3, and C4 (Fig 4C). The deletion mutant N2 (48-129) produced only slight derepression of both MFA2 and STE2, as opposed to the much greater derepression observed for the other targets. This difference may reflect the fact that Tup1 may be recruited by Mata2 in the absence of Cyc8, whereas Cyc8 is required for recruitment of Cyc8-Tup1 by the other repressors. The D3 and C4 deletions induced complete derepression of both MFA2 and STE2, while producing only partial derepression of the other targets.

The N-terminal region is sufficient for partial repression of most target genes:
Deletions that remove C-terminal sequences generally result in substantial derepression of all tested target genes. Deletions from the C terminus ending at positions 673, 431, 394, and 304 have roughly similar effects. However, deletions that leave only the first 201 or 254 aa, when expressed in multiple copies on a 2-µm vector, produce nearly complete repression of the target genes SUC2, RNR2, and OLE1 (Fig 5). The N201 deletion is somewhat more effective than N254. The repression of these targets is much less efficient when the mutants are expressed from single-copy chromosomal integrants. These results agree in general with those of TZAMARIAS and STRUHL 1994 , TZAMARIAS and STRUHL 1995 . However, in their experiments N200 functioned almost as well as full-length Tup1, but N253 had almost completely lost function. Since the tup1 mutants in their study were expressed from the strong ADH1 promoter, it seems the N-terminal region must be overexpressed to attain repression levels near the wild type.

View larger version (26K): In this window In a new window Download PPT slide   Figure 5. Effect of copy number on the function of the N-terminal repression domain. Derivatives of Tup1 containing only the first 201 or 254 N-terminal amino acids were expressed as an integrated single copy or on a multicopy plasmid in a tup1 deletion strain. Their repression activity was assayed by Northern blots of target genes. Lane 1, Tup1 N201 in pEMBLYe24; lane 2, Tup1 N254 in pEMBLYe24; lane 3, Tup1 N201 in pRS306; lane 4, Tup1 N254 in pRS306; lane 5, tup1, pEMBLYe24 alone; lane 6, TUP1 in pEMBLYe24.

The pattern of expression of MFA2 in the strains carrying the N201 and N254 mutants was in sharp contrast to the other targets. These Tup1 mutants in single copy mediated partial repression of MFA2, but in multiple copies behaved as dominant negative mutants, with expression of MFA2 much higher than in tup1 cells. Similarly, most of the deletions that removed part of the WD domains resulted in expression of MFA2 at levels greater than the tup1 null mutants. However, none of the insertion mutations within the WD domains behaved as dominant negatives. The dominant negative phenotype of tup1 deletions lacking the WD repeat domains was noted previously (KOMACHI and JOHNSON 1997 ).

Our results confirm the general pattern that deleting part of the C terminus results in almost complete loss of function, while deletion up to aa 200 or 254 restores function. A possible explanation would be that partial deletion of the WD region, which has been proposed to form a compact barrel shape composed of the seven WD "propellers," would result in gross misfolding of that region to render the entire protein nonfunctional. One problem with this model is that 304-713, which completely lacks the WD region, is also largely nonfunctional. Therefore, the region between residues 254 and 304 may be capable of inhibiting repression by the N terminus in the C-truncated proteins. In addition, insertion and point mutations in the WD domains that do not cause major disruption of the WD structure also result in almost complete derepression of most targets, suggesting that the N-terminal repression domain is either normally masked or much less important than the C-terminal repression domain.

Major deletions of the histone-binding domain of Tup1 maintain transcriptional repression:
A series of mutations that deleted the middle region of Tup1 was constructed and assayed for their effect on gene expression. This region was chosen for attention since terminal deletion mutations suggested that the major repression domains were located in this region (TZAMARIAS and STRUHL 1994 ; EDMONDSON et al. 1996 ), and region 120–316 was shown to bind to histones in vitro (EDMONDSON et al. 1996 ). Furthermore, this histone-binding region was thought to coincide with the repression domains (EDMONDSON et al. 1996 ). The mutant Tup1 proteins were expressed on multicopy plasmids (Fig 6A) or integrated as a single copy (Fig 6B). When the mutant proteins were expressed from multicopy plasmids, it was evident that deletions that remove a major part of the histone-binding domain, namely, 129-282 (M1) and 73-279 (M2), have little effect on expression of any of the target genes. Moderate derepression was seen in 130-340 (M4) and almost complete derepression in 73-340 (M3), the only mutant completely lacking the entire histone-binding domain determined in vitro (EDMONDSON et al. 1996 ). This latter result may reflect a requirement for histone binding for repression. However, this large deletion also removes the tetramerization domain and is expressed at low levels as noted below. Further work is needed to identify the histone-binding domain more precisely and determine its role in repression.

View larger version (192K): In this window In a new window Download PPT slide   Figure 6. Effect of tup1 deletions affecting the middle region on repression of different target genes. (A). Northern blots for Tup1 middle-deletion mutants expressed on multicopy plasmids. (B) Northern blots for Tup1 middle-deletion mutants expressed in single copy on pRS306 integrated at the URA3 locus. (C) The sequences affected in the middle-deletion mutants are compared with the histone-binding domains defined by EDMONDSON et al. 1996 . The region between 72 and 385 had histone-binding activity, with the strongest activity between 120 and 316. The degrees of repression of the target genes shown on the Northern blot in A are classified as near complete repression (+++), partial (>50%) repression (++), slight (<50%) repression (+), and no repression (-).

With single-copy expression, repression is mostly maintained by 129-282 (M1), but repression by the other mutants is relaxed, especially in the case of SUC2 by 73-279 (M2). The interpretation of the effect of copy number on repression is complicated by the different levels of expression of the mutant proteins. The mutants 129-282 (M1; Fig 3A) and 73-128 (M5) are expressed at levels near wild type, while M2, M3, and M4 are expressed at significantly lower levels (data not shown). In the case of the latter mutants, expression from multicopy plasmids may represent more faithfully the effect of these deletions on protein function.

These results support the interpretation of the Tup1 terminal deletions of TZAMARIAS and STRUHL 1994 that there are two major repression domains, an N-terminal domain containing aa 73–129 and a C-terminal repression domain containing aa 277–340. Our results suggest that deletion of the N-terminal domain does not significantly reduce repression, deletion of the C-terminal domain moderately reduces repression, and simultaneous deletion in both domains (73-340) almost eliminates repression (Fig 6C). An exception to this is MFA2, which is still mostly repressed in all of these deletions and apparently needs the WD region only for repression.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Location of repression domains:
The first structure/function analysis of Tup1, using N- and C-terminal deletions, identified two repression domains, localized between residues 72–200 and 288–389 (TZAMARIAS and STRUHL 1994 ). We will refer to these as the N-terminal and C-terminal repression domains, respectively. Our results are in general agreement with theirs and support the existence of two repression domains. However, our results with both internal deletion and linker insertion mutations suggest that the C-terminal repression domain is dominant in the context of the Tup1 protein. Of the linker insertion mutants, only those localized in the WD repeats induced a mutant phenotype. The insertions at aa 379 in WD1 severely reduced repression of all targets tested except for MFA2. IN342 severely derepressed RNR2 and HEM13. Insertions in WD repeats other than WD1 resulted in differential derepression of RNR2, possibly reflecting the requirement for interaction of these WD repeats with the Crt1 repressor (HUANG et al. 1998 ) that regulates RNR2. The linker insertion mutants have normal protein expression and complex formation with Cyc8, so their effects suggest local perturbation of the WD structure that weakens interaction with other proteins, presumably the downstream elements of the repression pathway. In a previous study, point mutations in Tup1 that derepressed hypoxic genes were isolated in two regions, the N terminus and the WD repeats (CARRICO and ZITOMER 1998 ). The N-terminal mutation disrupted Tup1 interaction with Cyc8, which is necessary for repression of the hypoxic genes. The mutations in the WD repeats had strong but differential effects on different target genes. The location of these point mutations supports the dominant role of the C-terminal repression domain.

The deletion mutants also suggest a dominant role for the C-terminal repression domain. Previous results with only terminal deletion mutants may have been misleading, since deletion of the N terminus prevents oligomerization of Tup1 (TZAMARIAS and STRUHL 1994 ). The repression activity of the Groucho repressor, which is homologous to Tup1, requires tetramerization via its native N terminus or a heterologous tetramerization domain (CHEN et al. 1998 ). Our results—that deletion of the N-terminal repression domain, but maintenance of the tetramerization and C-terminal repression domains, does not compromise repression—suggest that the Tup1 C-terminal domain is also dependent upon oligomerization for function. Mutations that eliminate the N-terminal repression domain, for example, M2 (73-279), preserve almost complete repression of all target genes tested. The analysis of deletions affecting the C-terminal domain is more problematical. The C-terminal repression domain defined by Tzamarias and Struhl was large, from residues 288–389 (TZAMARIAS and STRUHL 1994 ). A problem with defining this domain further is that it overlaps the WD domain and deletions into the WD domain perturb the structure, making precise localization difficult. All deletions affecting the WD region, with the exception of the ST region, gave significant derepression. Middle-deletion mutants that extended to position 340 produced partial derepression. This may indicate that the sequences just prior to the WD domain are important for repression. Alternatively, deleting up to residue 340 may result in perturbation of the WD domain, which begins at residue 333 (SPRAGUE et al. 2000 ), indirectly weakening repression that is directly mediated by the WD repeats.

The effects of deletions that extend into both the N- and C-terminal repression domains suggest that they may sometimes play redundant roles in repression. For example, M1 (129-282) has a near wild-type phenotype, but M4 (130-340), which extends to the boundary of the WD domain, has a partial mutant phenotype. Similarly, M3 (73-340), which removes the N-terminal repression domain and part of the C-terminal repression domain, has a much stronger mutant phenotype than M4 (130-340). The cumulative results suggest that the C-terminal repression domain is located near the beginning of the WD repeat domain, with the predominant role played by WD1 and a lesser role by the sequences 327–340 preceding WD1.

The repression domains presumably function by interacting with other proteins more directly responsible for repression. We have already discussed in the RESULTS that our deletion analysis suggests that the repression domains do not coincide with the region with the highest affinity for histones in vitro (EDMONDSON et al. 1996 ). Recently, several other proteins that interact with Tup1 and that may at least in part mediate repression have been identified. The Srb7 protein, a subunit of the Srb/Med complex, can bind to the two regions of Tup1 defined as repression domains (GROMOLLER and LEHMING 2000 ). Disruption of the Srb7-Tup1 interaction produced significant derepression of several Tup1 target genes (GROMOLLER and LEHMING 2000 ). Srb7 could mediate repression by Tup1 by blocking the signal for activation of transcription transduced by other Med subunits (GROMOLLER and LEHMING 2000 ). Tup1 also interacts directly with Srb10, another subunit of the mediator complex (ZAMAN et al. 2001 ), establishing a mechanism for attenuation of Tup1 repression in srb10 mutants (BALCIUNAS and RONNE 1995 ; KUCHIN and CARLSON 1998 ). The Hrs1/Med3 protein, another subunit of the Srb/Med complex, also interacts directly with the Tup1 N-terminal repression domain (PAPAMICHOS-CHRONAKIS et al. 2000 ). However, Hrs1 does not seem to mediate repression, since deletion of HRS1 does not reduce repression. In fact, overexpression of Hrs1 alleviates repression by Cyc8-Tup1.

In addition to histones, Tup1 interacts with two other types of proteins that affect chromatin structure, histone deacetylases (HDACs) and the Nhp6A/Nhp6B proteins. WATSON et al. 2000 found that simultaneous deletion of three specific genes encoding HDACs, hos1, hos2, and rpd3, severely reduced repression of Tup1 target genes. Furthermore, the Hos2 and Rpd3 HDACs interact directly with the Cyc8-Tup1 complex. The functional significance of these interactions was demonstrated by the finding that the promoter regions of genes repressed by Cyc8-Tup1 show decreased histone acetylation (BONE and ROTH 2001 ). In other cases, repression by Tup1 is mediated in part by recruitment of the Hda1 histone deacetylase (WU et al. 2001 ). The Hda1 protein interacts with GST-Tup1(73-386) in vitro (WU et al. 2001 ). By use of the split-ubiquitin assay, Tup1 was shown to interact with the very similar Nhp6A and Nhp6B proteins, members of the HMG1 family of DNA-binding proteins (LASER et al. 2000 ). Deletion of the NHP6 genes reduced repression of some, but not all, Tup1 targets. Clearly, many proteins involved in the repression pathway interact directly with Tup1, suggesting that the Tup1 repression domains are complex and in need of further characterization.

Analysis of the Tup1 N terminus:
The N-terminal region of Tup1 containing residues 1–72 is required for both oligomerization of Tup1 and association with Cyc8 (TZAMARIAS and STRUHL 1994 ). Predictions of the secondary structure of the Tup1 N terminus were made by the computer program PHD (LUPAS et al. 1991 ; ROST et al. 1994 ; Fig 7). Three regions in the N terminus of Tup1 had a high (>82%) propensity to form -helices: aa 9–36 (termed H-1), aa 42–82 (H-2), and aa 90–121 (H-3), the latter composed of three small helices rich in glutamine. Since coiled-coil regions of proteins are common stabilizing motifs among oligomers (COHEN and PARRY 1994 ), the propensity of these helices to form coiled-coil regions is also presented in Fig 7. Both H-1 and H-2 contain helical domains scored >70% probability as potential coiled coils. H-1, H-2, and H-3 had 3, 3, and 4 predicted heptads, respectively. Sequences adopting a coiled-coil structure normally consist of heptad repeats (a b c d e f g) with hydrophobic residues at positions a and d. Recently, a variety of approaches including circular dichroism spectroscopy and nuclear magnetic resonance imaging showed that the N terminus of Tup1 has a coiled-coil structure (JABET et al. 2000 ) and confirmed our finding that Tup1 normally forms a tetramer (VARANASI et al. 1996 ).

View larger version (28K): In this window In a new window Download PPT slide   Figure 7. Predicted secondary structure of the Tup1 N terminus. The Tup1 protein sequence was analyzed using the PHD program (LUPAS et al. 1991 ; ROST et al. 1994 ) via e-mail (PredictProtein). The line HELIX shows residues likely to adopt -helical structures as H. The COILS line shows the probability on a scale of 0 (low) to 9 (high) that residues will adopt a coiled-coil structure. The Phase line refers to the phasing of the heptad repeats predicted by the COILS program.

In the mutant 1-51, H-1 is missing and the other coiled-coil domains are present, while in the mutant 48-129 H-1 is present and both H-2 and H-3 are absent. Neither of these deletion mutants formed a Tup1 oligomer or associated with Cyc8, suggesting that the helical domains present in both deleted regions are required for these functions. The oligomerization of Tup1 and association with Cyc8 are both destabilized in the mutant IN48 (Fig 1B and Fig C), which results in an insertion of four amino acids, LELE, after aa 48. This insertion is within H-2 and the predicted coiled-coil region. Analysis of the secondary structure by the PHD program predicts that IN48 does not interrupt H-2 but just changes H-2 from 42–82 to 43–90 aa (data not shown). The insertion may destabilize the protein-protein association by changing the spacing between different helical domains directly involved in protein-protein associations.

Several point mutations in the Tup1 N terminus that compromise Tup1 association with Cyc8 have been reported: L62R (CARRICO and ZITOMER 1998 ), I55K, and V59D (CAO 1997 ). These mutations are located in H-2, the second helical domain, and are all hydrophobic to charged amino acid substitutions in the a or d (hydrophobic) positions of the proposed coiled coil (Fig 6). These mutations resulted in derepression of SUC2, but did not prevent oligomerization of Tup1. No point mutations have yet been identified that prevent Tup1 oligomerization and still maintain association with Cyc8.

Groucho and related proteins are possible Tup1 homologs from metazoans (FISHER and CAUDY 1998 ; PARKHURST 1998 ; COUREY and JIA 2001 ). Groucho and Tup1 share the same general domain structures, with an N-terminal domain required for tetramerization, a poorly conserved middle region, and a C-terminal domain of seven WD repeats. TLE1 and TLE2, two related Groucho homologs from mammals, have two repression domains located in the N terminus and at the beginning of the WD domains, as does Tup1 (GRBAVEC et al. 1998 ). The evidence for homology between Groucho and Tup1 was recently strengthened by the finding that TLE1 and TLE2 interact with the yeast Cyc8(Ssn6) and Cyc8- related proteins from mammals (GRBAVEC et al. 1999 ). Therefore, Groucho may participate in a corepressor complex homologous to the yeast Cyc8-Tup1 complex. However, Groucho may sometimes act independently of Cyc8 homologs and in most cases interacts directly with repressors via its WD region, as Tup1 interacts with the Mat2 repressor.

Groucho, like Tup1, can interact directly with histone H3 in vitro, and this interaction is believed to play a role in repression (PALAPARTI et al. 1997 ). Furthermore, Groucho binds more strongly to hypoacetylated histones, as does Tup1 (FLORES-SAAIB and COUREY 2000 ). Deletion mutations of Groucho that decrease histone binding reduce repression activity. Recently, Groucho was found to interact directly with the histone deacetylase Rpd3 and thus may directly modify chromatin structure (CHEN et al. 1999 ). Tup1 can interact with the yeast histone deacetylases Rpd3 and Hos2, and deletion of three yeast histone deacetylase genes was shown to partially relieve repression by Tup1 (WATSON et al. 2000 ). The accumulated evidence suggests that Groucho and Tup1 are true homologs, but have diverged sufficiently to interact with different repressor proteins and to repress by related but somewhat different mechanisms.

	FOOTNOTES

¹ These authors contributed equally to this work.
² Present address: School of Life Science and Biotechnology, Shanghai Jiao Tong University, Shanghai 200030, PR China.
³ Present address: Department of Pathology, Wayne State University, Detroit, MI 48201.

	ACKNOWLEDGMENTS

The authors thank Dr. Alexander D. Johnson of the University of California at San Francisco for providing the MFA2 plasmid and Dr. Charles E. Martin of Rutgers University for providing the OLE1 plasmid. We thank Frederick Williams for the construction of the mutants N201 and N254 and Andrea Shawaker for technical assistance. This research was supported by grants from the American Cancer Society (CMV-450) and the National Institutes of Health (GM-48148).

Manuscript received February 28, 2002; Accepted for publication March 19, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BALCIUNAS, D. and H. RONNE, 1995  Three subunits of the RNA polymerase II mediator complex are involved in glucose repression. Nucleic Acids Res. 23:4421-4425.[Abstract/Free Full Text]

BALDARI, C. and G. CESARINI, 1985  Plasmids pEMBLY: new single-stranded shuttle vectors for the recovery and analysis of yeast DNA sequences. Gene 35:27-32.[Medline]

BONE, J. R. and S. Y. ROTH, 2001  Recruitment of the yeast Tup1p-Ssn6p repressor is associated with localized decreases in histone acetylation. J. Biol. Chem. 276:1808-1813.[Abstract/Free Full Text]

CAO, X., 1997 Mutagenesis of Tup1 N-terminus to identify the amino acid residue(s) essential for Tup1 function. M. S. Thesis, Medical College of Ohio, Toledo, OH.

CARRICO, P. M. and R. S. ZITOMER, 1998  Mutational analysis of the Tup1 general repressor of yeast. Genetics 148:637-644.[Abstract/Free Full Text]

CHEN, G., P. H. NGUYEN, and A. J. COUREY, 1998  A role for Groucho tetramerization in transcriptional repression. Mol. Cell. Biol. 18:7259-7268.[Abstract/Free Full Text]

CHEN, G., J. FERNANDEZ, S. MISCHE, and A. J. COUREY, 1999  A functional interaction between the histone deacetylase Rpd3 and the corepressor groucho in Drosophila development. Genes Dev. 13:2218-2230.[Abstract/Free Full Text]

CHOMCZYNSKI, P., 1992  Solubilization in formamide protects RNA from degradation. Nucleic Acids Res. 20:3791-3792.[Free Full Text]

CHOMCZYNSKI, P. and K. MACKEY, 1995  Short technical reports. Modification of the TRI reagent procedure for isolation of RNA from polysaccharide- and proteoglycan-rich sources. Biotechniques 19:942-945.[Medline]

COHEN, C. and D. A. PARRY, 1994  Alpha-helical coiled coils: more facts and better predictions. Science 263:488-489. (Erratum. Science 264: 1068).[Free Full Text]

COOPER, J. P., S. Y. ROTH, and R. T. SIMPSON, 1994  The global transcriptional regulators, SSN6 and TUP1, play distinct roles in the establishment of a repressive chromatin structure. Genes Dev. 8:1400-1410.[Abstract/Free Full Text]

COUREY, A. J. and S. JIA, 2001  Transcriptional repression: the long and short of it. Genes Dev. 15:2786-2796.[Free Full Text]

DAVIE, J. K., R. J. TRUMBLY, and S. Y. DENT, 2002  Histone-dependent association of Tup1-Ssn6 with repressed genes in vivo. Mol. Cell. Biol. 22:693-703.[Abstract/Free Full Text]

DERISI, J. L., V. R. IYER, and P. O. BROWN, 1997  Exploring the metabolic and genetic control of gene expression on a genomic scale. Science 278:680-686.[Abstract/Free Full Text]

EDMONDSON, D. G., M. M. SMITH, and S. Y. ROTH, 1996  Repression domain of the yeast global repressor Tup1 interacts directly with histones H3 and H4. Genes Dev. 10:1247-1259.[Abstract/Free Full Text]

FISHER, A. L. and M. CAUDY, 1998  Groucho proteins: transcriptional corepressors for specific subsets of DNA-binding transcription factors in vertebrates and invertebrates. Genes Dev. 12:1931-1940.[Free Full Text]

FLORES-SAAIB, R. D. and A. J. COUREY, 2000  Analysis of groucho-histone interactions suggests mechanistic similarities between groucho- and Tup1-mediated repression. Nucleic Acids Res. 28:4189-4196.[Abstract/Free Full Text]

FRIESEN, H., S. R. HEPWORTH, and J. SEGALL, 1997  An Ssn6-Tup1-dependent negative regulatory element controls sporulation-specific expression of DIT1 and DIT2 in Saccharomyces cerevisiae. Mol. Cell. Biol. 17:123-134.[Abstract]

FUJIMORI, K., S. ANAMNART, Y. NAKAGAWA, S. SUGIOKA, and D. OHTA et al., 1997  Isolation and characterization of mutations affecting expression of the delta9- fatty acid desaturase gene, OLE1, in Saccharomyces cerevisiae. FEBS Lett. 413:226-230.[Medline]

GAVIN, I. M., M. P. KLADDE, and R. T. SIMPSON, 2000  Tup1p represses mcm1p transcriptional activation and chromatin remodeling of an a-cell-specific gene. EMBO J. 19:5875-5883.[Medline]

GEISBERG, J. V. and K. STRUHL, 2000  TATA-binding protein mutants that increase transcription from enhancerless and repressed promoters in vivo. Mol. Cell. Biol. 20:1478-1488.[Abstract/Free Full Text]

GIETZ, D., A. ST. JEAN, R. A. WOODS, and R. H. SCHIESTL, 1992  Improved method for high efficiency transformation of intact yeast cells. Nucleic Acids Res. 20:1425.[Free Full Text]

GRBAVEC, D., R. LO, Y. LIU, and S. STIFANI, 1998  Transducin-like Enhancer of split 2, a mammalian homologue of Drosophila Groucho, acts as a transcriptional repressor, interacts with Hairy/Enhancer of split proteins, and is expressed during neuronal development. Eur. J. Biochem. 258:339-349.[Medline]

GRBAVEC, D., R. LO, Y. LIU, A. GREENFIELD, and S. STIFANI, 1999  Groucho/transducin-like enhancer of split (TLE) family members interact with the yeast transcriptional co-repressor SSN6 and mammalian SSN6-related proteins: implications for evolutionary conservation of transcription repression mechanisms. Biochem. J. 337:13-17.

GROMOLLER, A. and N. LEHMING, 2000  Srb7p is a physical and physiological target of Tup1p. EMBO J. 19:6845-6852.[Medline]

HERSCHBACH, B. M., M. B. ARNAUD, and A. D. JOHNSON, 1994  Transcriptional repression directed by the yeast alpha 2 protein in vitro. Nature 370:309-311.[Medline]

HUANG, L., W. ZHANG, and S. Y. ROTH, 1997  Amino termini of histones H3 and H4 are required for a1-alpha2 repression in yeast. Mol. Cell. Biol. 17:6555-6562.[Abstract]

HUANG, M., Z. ZHOU, and S. J. ELLEDGE, 1998  The DNA replication and damage checkpoint pathways induce transcription by inhibition of the Crt1 repressor. Cell 94:595-605.[Medline]

JABET, C., E. R. SPRAGUE, A. P. VANDEMARK, and C. WOLBERGER, 2000  Characterization of the N-terminal domain of the yeast transcriptional repressor Tup1. Proposal for an association model of the repressor complex Tup1 x Ssn6. J. Biol. Chem. 275:9011-9018.[Abstract/Free Full Text]

KELEHER, C. A., M. J. REDD, J. SCHULTZ, M. CARLSON, and A. D. JOHNSON, 1992  Ssn6-Tup1 is a general repressor of transcription in yeast. Cell 68:709-719.