Abstract
The CCR4-NOT transcriptional regulatory complex affects expression of a number of genes both positively and negatively. We report here that components of the CCR4-NOT complex functionally and physically interact with TBP and TBP-associated factors. First, mutations in CCR4-NOT components suppressed the his4-912δ insertion in a manner similar to that observed for the defective TBP allele spt15-122. Second, using modified HIS3 promoter derivatives containing specific mutations within the TATA sequence, we found that the NOT proteins were general repressors that disrupt TBP function irrespective of the DNA sequence. Third, increasing the dosage of NOT1 specifically inhibited the ability of spt15-122 to suppress the his4-912δ insertion but did not affect the Spt− phenotype of spt3 or spt10 at this locus. Fourth, spt3, spt8, and spt15-21 alleles (all involved in affecting interaction of SPT3 with TBP) suppressed ccr4 and caf1 defects. Finally, we show that NOT2 and NOT5 can be immunoprecipitated by TBP. NOT5 was subsequently shown to associate with TBP and TAFs and this association was dependent on the integrity of TFIID. These genetic and physical interactions indicate that one role of the CCR4-NOT proteins is to inhibit functional TBP-DNA interactions, perhaps by interacting with and modulating the function of TFIID.
THE CCR4-NOT complex is one of several large groups of proteins involved in transcription and consists of two complexes, 1.9 × 106 and 1.0 × 106 D (1 mD) in size (Liuet al. 1998; Baiet al. 1999). Both of these complexes are distinct from other large transcriptionally important groups of proteins, such as the SNF/SWI complex, the SAGA complex, and the RNA polymerase II holoenzyme (Draperet al. 1995; Liuet al. 1998). The 1-mD CCR4-NOT complex consists of CCR4, CAF1, the five NOT proteins (NOT1–5), and three other proteins (Draper et al. 1994, 1995; Liuet al. 1998; Y.-C. Chiang, unpublished observations). The CCR4 gene was initially identified by mutations that suppressed spt10-enhanced ADH2 expression and that affected the derepression of ADH2 and other nonfermentative genes (Denis 1984; Denis and Malvar 1990). However, it has become clear since then that CCR4, CAF1, and NOT proteins affect diverse processes in yeast and do so both positively and negatively (Malvaret al. 1992; Sakaiet al. 1992; Collart and Struhl 1993; McKenzieet al. 1993; Schild 1995; Bensonet al. 1998; Leeet al. 1998; Liuet al. 1998; Tabtiang and Herskowitz 1998).
NOT genes were originally isolated as mutations that cause defects in cell cycle progression and were also isolated as regulatory factors in the pheromone response pathway (Barros Lopeset al. 1990; Cade and Errede 1994; Irieet al. 1994). The first insight into the function of the NOT proteins, however, came from studies examining the regulation of HIS3 expression. The five NOT genes were isolated as mutations that allow increased TC-dependent HIS3 expression in a strain that lacks optimal GCN4 function (Collart and Struhl 1993, 1994; Oberholzer and Collart 1998). The HIS3 promoter contains two important elements, TR, which consists of the canonical TATA sequence, and TC, which is a longer, ill-defined element containing several nonconsensus TATA sequences (Mahadevan and Struhl 1990). Mutations in the NOT genes allowed increased HIS3 expression from the TC element. It is not clear, however, if this effect is specific to the TC element or if it affects other TATA sequences as well.
Since the NOTs can distinguish between TC and TR, the NOT proteins were suggested to restrict TATAA-binding protein (TBP) access to noncanonical TATA sequences. Collart (1996) further showed that defects in MOT1, which is required for TBP removal from DNA (Aubleet al. 1994), and SPT3, which binds TBP (Eisenmannet al. 1992) and is required for its action at certain promoters, genetically interacted with certain mutations in the NOT genes. More recently, not and caf1 alleles have been found to be specific suppressors of srb4, probably due to their ability to increase transcription at critical SRB4-dependent promoters (Leeet al. 1998). Similarly, defects in NCB1 and NCB2, components of the NC2 complex that binds TBP and represses its function (Goppelt and Meisterernst 1996; Goppeltet al. 1996), also suppress srb4 (Leeet al. 1998). The NC2 complex has been shown to directly bind TBP and repress its function by blocking the access of TFIIB and TFIIA. In contrast, the mechanism of action of the CCR4-NOT complex remains unclear.
In this study we have assayed NOT function in vivo using strains with simplified HIS3 promoter derivatives containing specific mutations within the TATA sequence. Specifically, we tested the effect of not mutations on activated HIS3 transcription in the presence of wild-type GCN4 from mutant TATA sequences that vary by a single base pair. From this analysis we found that the NOT proteins are general repressors that reduce TBP activity irrespective of the sequence to which TBP binds. In addition, we also obtained evidence that caf1 and not4 deletions affect his4-912δ expression in the same manner as certain TBP mutations. We also show that TBP and TBP-associated factors (TAFs) display multiple genetic interactions with CCR4-NOT complex components. While it has recently been shown that glutathione S-transferase (GST)-TBP can retain NOT1 from yeast crude extracts (Leeet al. 1998), we show additionally that TBP can immunoprecipitate NOT2 and NOT5. Moreover, GST-NOT5 was found to interact with several TAF components of TFIID and, in vivo, an intact TFIID complex was required for these interactions. These genetic and physical interactions indicate that one role of the CCR4-NOT complex is to inhibit TBP activity, possibly through interactions with TFIID.
MATERIALS AND METHODS
Yeast strains: Yeast strains are listed in Table 1.
Genetic analysis: Nutrient auxotrophies and suppression of δ insertions at the HIS4 locus were scored on His− medium supplemented with the appropriate amino acids, as described in Arndt et al. (1994). YD plates consist of 1% yeast extract, 2% bactopeptone, 2% agar, and 2% glucose. Phenotypic analysis of the modified HIS3 promoters was carried out by testing growth of strains on His media containing 3-aminotriazole (3-AT), as described in Arndt et al. (1994).
Gene disruptions and plasmids: The plasmids and gene disruption constructs used have been described previously (Collart and Struhl 1994; Madison and Winston 1997; Liuet al. 1998; Oberholzer and Collart 1998).
Immunoprecipitation: Immunoprecipitations were carried out as described previously (Draperet al. 1994; Liuet al. 1998) with yeast extracts prepared in Ip/wash buffer containing 50 mm phosphate, pH 7.6, 150 mm potassium acetate, 1 mm sodium pyrophosphate, 1 mm sodium fluoride, 1% NP40, 10% glycerol, 2 mm magnesium chloride, 1 mm EDTA, 4 mg/liter leupeptin, 2 mg/liter pepstatin A, 1 mm benzamidine, 1.25 mg/liter chymastatin, 1 mm phenylmethylsulfonyl fluoride (PMSF). A total of 2 mg of yeast crude extract protein in 250 ml of Ip/wash buffer was incubated with 5 mg of anti-TBP antibody or equivalent amount of anti-TBP preimmune serum for 1 hr at 4° with constant rocking. Twenty-five microliters of packed Protein A agarose beads was added to the above mixture and incubated for 1 hr at 4° with constant shaking. The beads were washed four times with 1 ml of the same buffer to remove unbound proteins from the yeast extract, mixed with SDS loading buffer, boiled for 5 min, and separated on a SDS-PAGE gel prior to Western analysis.
GST pulldowns: GST fusion proteins were expressed and bound to glutathione-agarose beads as described in Chiang et al. (1996). Yeast whole-cell extracts were prepared in the same buffer used for carrying out the immunoprecipitation experiments. Washed gutathione-agarose beads were incubated with 2 mg of yeast whole-cell extract protein in 250 ml of Ip/wash buffer for 2 hr at 4° on a rocking platform. Unbound proteins were removed by four washes with 1 ml of the same buffer and specifically bound proteins were resolved by SDS-PAGE after boiling beads directly in sample buffer. Proteins resolved by SDS-PAGE were transferred and analyzed by Western blotting as described in Draper et al. (1994).
RESULTS
caf1 and not4 suppress his4-912δ expression in a manner similar to that of spt15-122: We initially observed that a caf1 deletion suppressed a his4-912δ− insertion and thus exhibited a Spt− phenotype (Figure 1). Mutations in several genes involved in transcription have been isolated as suppressors of his4-912δ or other Ty insertions at the HIS4 or LYS2 loci, including SPT15, which encodes the yeast TBP protein (Eisenmannet al. 1989). Because it has been previously shown that the CCR4-NOT proteins were functionally linked to TBP (Collart 1996; Leeet al. 1998), we explored further whether other components of the CCR4-NOT complex also elicited an Spt− phenotype. We found that a not5 deletion exhibited a weak Spt− phenotype (not shown) while a not4 deletion caused a moderately strong suppression of his4-912δ (Figure 1). A ccr4 deletion, in turn, did not affect his4-912δ expression (Figure 1). It should be noted, however, that ccr4 did suppress the his4-912δ insertion-containing promoter when it was fused to the ADE2 gene (not shown). These above effects were found to be specific to his4-912δ as none of the above-mentioned defects suppressed a his4-917δ or lys2-128δ− insertion (data not shown). Also, this Spt− phenotype was not observed with defects in DBF2 and DHH1, two proteins known to associate with CCR4 (Liuet al. 1997; Hataet al. 1998) but that are not components of the 1-mD CCR4-NOT complex (Baiet al. 1999; Y.-C. Chiang, unpublished observations).
We next analyzed whether the Spt− phenotypes of caf1 and not4 occurred in a manner similar to that of certain spt15 mutations. One such mutation, spt15-122, containing the L205F alteration in TBP, suppresses the his4-912δ insertion (Arndtet al. 1994). Residue 205 is known to contact the TATA box and this mutated allele of TBP shows increased affinity for binding the nonconsensus TATA sequence CATAAA (Arndtet al. 1994). The ability of spt15-122 to suppress his4-912δ correspondingly depends on the presence of the CATAAA sequence at position III within the delta element (Arndtet al. 1994). We consequently investigated the ability of caf1 and not4 deletions to suppress his4-912δ in a strain wherein the CATAAA sequence at position III within the delta element had been mutated to GAGCTC (designated *CATAA* in Figure 1; Arndtet al. 1994). We found that caf1 and not4 were unable to suppress his4-912δ in this strain (Figure 1). In contrast, mutations in SPT6 and SPT10, whose protein products are not part of the CCR4-NOT complex (Liuet al. 1998) and are not functionally related to TBP (Winston and Carlson 1992; Dollardet al. 1994; Bortvin and Winston 1996), suppressed his4-912δ even when the CATAAA sequence was mutated (spt10, Figure 1; and spt6, F. Winston, personal communication). These data suggest that the Spt− phenotypes of caf1 and not4 occurred in a manner similar to that displayed by spt15-122.
Yeast strains
not4 and not5 deletions increase transcription from nonconsensus TATA sequences irrespective of sequence: On the basis of our above observation that a CATAAA sequence was important to the Spt− phenotype elicited by caf1 and not4 and that NOT proteins are presumed to affect TBP access to noncanonical TA-TAs, we subsequently examined whether the effects of CCR4-NOT proteins were dependent on a particular noncanonical TATA sequence. To investigate this, we utilized a simplified HIS3 promoter in which the TC and TR elements in the HIS3 promoter were replaced by an oligonucleotide positioned 23 bp downstream from the GCN4 binding site (Chen and Struhl 1988). The oligonucleotides were TATAAA or derivatives of this sequence containing single base changes (Arndtet al. 1994). Using resistance to 3-AT, a competitive inhibitor of the HIS3 gene product, as a measure of relative HIS3 expression (Klopotowski and Wiater 1965; Struhl and Davis 1977), we analyzed the effects of deletions in CCR4-NOT components on these HIS3 promoter derivatives. We initially found that ccr4 and caf1 deletions did not have any effect on a his3-202 promoter containing a CATAAA sequence instead of TATAAA and did not confer the ability of the yeast to grow on 5 mm 3-AT (not shown). In contrast, not4 and not5 deletions allowed growth on 20 mm 3-AT in the strain carrying his3-202 (Table 2). These data suggest that not4 and not5 deletions allow increased stability of TBP binding to the CATAAA sequence and thus allow increased transcription from the CATAAA sequence. This effect is similar to that observed with the mutated spt15-122 allele that allows increased transcription from the his3-202 (CATAAA) promoter (Arndtet al. 1994). These data also indicated that the modified HIS3 promoter used in this study was behaving in a manner similar to that observed for HIS3 since not4 and not5 deletions augment TC-dependent expression at the HIS3 locus (Collart and Struhl 1994; Oberholzer and Collart 1998) whereas ccr4 and caf1 do not (Liuet al. 1998).
caf1 and not4 suppress the his4-912δ insertion. Growth was detected on complete medium (his+) or minimal medium lacking histidine (his−). *CATAA* indicates that the CATAA sequence at position III within the δ element has been mutated to GTGCAC (Arndtet al. 1994). The left-most column in each separate panel represents a 1/1000 dilution of a saturated yeast culture, and the two succeeding columns represent 1/10 dilutions, respectively. One-half the volume was spotted for the his+ samples as compared to the his− samples. (Left) The effect on his4-912δ expression for wild type, ccr4, caf1, and not4 was analyzed in DY3462 by making the indicated deletion allele and for spt10 in strain 1370-5b. (Right) The effect on his4-912δ (*CATAA*) expression was analyzed in strain DY3855 (wt), DY3855-1a (ccr4), 1374-6c (caf1), 1361-1-1c (not4), and 1370-1c (spt10).
Although spt15-122 and not4 and not5 have the same effect on the his3-202 (CATAAA) promoter, it should be noted that spt15-122 is a point mutation in TBP that alters its binding specificity and increases its binding affinity for the CATAAA sequence. Therefore, the effect of spt15-122 is specific to his3-202 (CATAAA) and does not affect other HIS3 promoter derivatives containing other substitutions in the TATAAA sequence (Arndtet al. 1994). We subsequently asked whether spt15-122 and not4 and not5 affected his3-202 (CATAAA) by the same mechanism. This question was investigated using two approaches. First, if not4 and not5 do not alter the binding specificity of TBP, then their effects should not be limited to one particular nonconsensus TATAAA element. To test this possibility, we made not4 and not5 deletions in strains containing other HIS3 promoter derivatives. We found that not4 and not5 deletions allowed increased HIS3 transcription from several different nonconsensus TATA elements (CATAAA, TATAAG, TACAAA, and TGTAAA) that were not affected by spt15-122 (Table 2; Arndtet al. 1994).
Second, if mutations in TBP and NOT4 affected nonconsensus TATA elements by the same mechanism, then combining these mutations should not have any significant additive effects. It should be noted that, for reasons unrelated to HIS3 transcription, strains containing spt15-122 mutations are very sensitive to high levels of 3-AT and hence could not be used for this latter analysis (Arndtet al. 1994). Therefore, we used spt15-301(L114F), which is another point mutation in TBP that increases its affinity for the sequence TATAAG and thus allows increased transcription from his3-217 (TATAAG; see Table 3 and Arndtet al. 1992). While the not4 deletion and the spt15-301 mutation each allowed growth on 20 mm and 40 mm 3-AT, respectively, in combination they allowed growth on 80 mm 3-AT (Table 3). Thus, the additive effect of these two mutations suggests that they function by different mechanisms. Taken together, these data indicate that while defects in TBP alter TBP binding specificity and cause increased transcription from certain specific nonconsensus TATA elements, the not4 and not5 deletions allow increased transcription from all nonconsensus TATA elements tested. The not4 and not5 deletions appear to stabilize TBP-TATA box interactions irrespective of the sequence and by a mechanism that is apparently different from augmenting the affinity of TBP binding to a particular noncanonical TATA sequence. In this regard it should be noted that increasing TBP levels in the cell did not augment HIS3 expression from these noncanonical TATA sequences (data not shown).
Effect of not4 and not5 deletions on different his3 promoter alleles
Effect of not4 and spt15-301 (L114F) alleles on his3-217 (TATAAG) expression
Genetic interactions between CCR4-NOTs, TBP, and TBP-associated factors: The above data implicate a balance between components of the CCR4-NOT complex and TBP and its associated factors in regulating gene expression. The effect of increasing the levels of CCR4-NOT protein components in combination with the TBP allele spt15-122 was subsequently investigated. Increasing the dosage of NOT1 was found to specifically inhibit the Spt− phenotype of spt15-122 (Table 4) but did not affect the Spt− phenotype of spt10 and spt3 mutations (data not shown). Elevated expression of NOT2 or NOT4 did not have any effect in the spt15-122 strain while increasing the dosage of NOT3 caused a weak inhibition of the Spt− phenotype (Table 4).
Effect of increased NOT gene dosage on the ability of spt15-122 to suppress the his4-912δ insertion
The above experiments indicate that components of the CCR4-NOT complex can functionally interact with TBP and its associated factors. Although it is clear from these studies that several of the NOTs appear more closely aligned with TBP and its associated factors than CCR4 and CAF1, ccr4 and caf1 do display some genetic interactions with TBP as evidenced by the similar behavior of caf1 to spt15-122 in terms of suppressing his4-912δ. We also investigated the effect of spt3 and spt8 defects on ccr4 and caf1 phenotypes since it has been shown that the temperature-sensitive phenotype associated with the not1-2 allele is specifically suppressed by spt3 and spt8 deletions whereas defects in other NOT genes are not suppressed by these spt alleles (Collart 1996). An spt3 deletion suppressed the caffeine sensitivity and cold sensitivity of a ccr4 deletion or a caf1 deletion but did not suppress the glycerol temperature-sensitive phenotype (Table 5). An spt8 deletion also suppressed ccr4 and caf1 defects (Table 5). Using an spt15-21 allele that results in a TBP that is specifically defective in binding SPT3 (Eisenmannet al. 1992), we examined whether the ability of spt3 to suppress ccr4 was related to the ability of SPT3 to bind TBP. The spt15-21 allele also suppressed defects caused by a ccr4 deletion (Table 5). An spt7 deletion did not have any effect on ccr4 and caf1 phenotypes (data not shown). These results confirm that CCR4 and CAF1 also functionally interact with TBP and SPT3 and indicate that different components of the CCR4-NOT complex display different genetic interactions with TBP and its associated factors.
NOT2 and NOT5 co-immunoprecipitate with TBP: On the basis of all the above data, it is likely that components of the CCR4-NOT complex physically interact with TBP, TAFs, or other TBP-associated factors. To initially address this possibility, we carried out immuno-precipitation experiments using LexA-tagged versions of the NOT proteins. We found that immunoprecipitating TBP with anti-TBP antibody specifically brought down LexA-NOT2 and LexA-NOT5 but did not bring down LexA alone (Figure 2, lanes 7 and 8 as compared to lane 6). As an additional control, we showed that TBP preimmune serum did not immunoprecipitate LexA-NOT2 (Figure 2, lane 5). Immunoprecipitating TBP did not bring down LexA-CCR4, LexA-CAF1, or LexA-NOT4 (data not shown), indicating that the LexA-NOT2 and Lex-NOT5 interaction with TBP was not due to the LexA moiety. These data agree with recent studies indicating that NOT1 can physically associate with GST-TBP (Leeet al. 1998). These results also suggest that NOT2 and NOT5 are more closely associated with TBP and its associated factors than other components of the CCR4-NOT complex in agreement with the observation that within the CCR4-NOT complex NOT2 and NOT5 are very closely associated (Baiet al. 1999).
spt3, spt8, and spt15-21 suppress ccr4 and caf1 defects
TBP immunoprecipitates LexA-NOT2 and LexA-NOT5. Yeast extracts from haploid strain EGY188 containing LexA fusion proteins as indicated were immunoprecipitated with anti-TBP antibody (lanes 6–8) and with preimmune serum (lanes 4 and 5). LexA fusion proteins were detected by Western analysis following SDS-PAGE using LexA antibody. LexA-NOT2 (lanes 2, 5, 7), LexA-NOT5 (lanes 3 and 8), and LexA (lanes 1, 4, 6) contained full-length NOT2, NOT5, and LexA, respectively. Crude extracts are represented in lanes 1–3. Ten microliters of whole-cell extract was loaded as crude extract and 250 ml of extract was used for each immunoprecipitation reaction. Roughly 5–10% of LexA-NOT2 and LexA-NOT5 co-immunoprecipitated with TBP. It should be noted that in lanes 4–7 LexA antibody was not used to probe proteins of molecular weight >49 kD since neither LexA nor LexA-NOT2 migrate at that size.
NOT5 interacts with TFIID: Since TBP can associate with NOT5 and NOT2, we subsequently determined whether TAFs also associated with NOT5 and NOT2. We initially tried immunoprecipitating NOT5 or NOT2 and ascertaining whether TBP and TAFs co-immunoprecipitated. However, the NOT2 antibody was not of sufficient quality to conduct these experiments and NOT5 antibody nonspecifically immunoprecipitated several TAFs even in strains lacking NOT5 (data not shown). It should be noted that CCR4 or CAF1 immunoprecipitations only immunoprecipitated those proteins present in the 1-mD CCR4-NOT complex (Baiet al. 1999).
Therefore, to conduct these studies we examined the ability of TBP and its associated factors to bind the GST-NOT5 protein that was expressed and purified from Escherichia coli. When whole-cell extracts from wild-type yeast strains were passed over GST-NOT5, it specifically retained TAF90, TAF61, and TBP while the control fusions (GST or GST-Vpu) did not retain any of the above-mentioned proteins (Figure 3, lane 3 as compared to lanes 2 and 4). GST-NOT2 in contrast did not specifically retain TBP or TAFs (data not shown). Other TAF-directed antibodies (TAF130, TAF40, and TAF19) were also used in this analysis but were unable to detect the corresponding protein in the crude extracts (data not shown).
The above data suggest that NOT5 is capable of contacting TAFs and TBP. To address whether this contact was through TFIID, we repeated the GST-NOT5 binding experiments using extracts from yeast strain K151L, K156Y, containing a TBP with alterations of K151L and K156Y (Ranalloet al. 1999). TBP (K151L, K156Y) associates with TAFs in a TFIID complex at 30° but at the restrictive temperature of 38° is unable to immunoprecipitate the complete complement of TAFs (Ranalloet al. 1999). When extracts from the wild-type parent of K151L, K156Y grown at the restrictive temperature were passed over the GST-NOT5 column, TBP, TAF60, TAF61, and TAF90 were specifically retained (Figure 4, lane 4). Similar results were obtained at the permissive temperature for the wild-type strain (data not shown). In contrast, when extracts from K151L, K156Y grown at the permissive temperature were incubated with GST-NOT5, diminished levels of TBP, TAF60, and TAF90 were bound although all proteins were present in the crude extracts (compare Figure 4, lane 5 to lane 2). When the K151L, K156Y strain was grown at the restrictive temperature, GST-NOT5 retained little of the TBP, TAF60, TAF61, and TAF90 proteins. These results imply that an intact TFIID complex must be present for NOT5 to associate with TBP and TAFs.
GST-NOT5 retains TBP and TAFs. GST fusions were induced as described (Chianget al. 1996), bound to glutathione agarose beads, and incubated with yeast crude extract from the strain yEK20 as described in materials and methods. Proteins were eluted from glutathione agarose beads by boiling and separated by SDS-PAGE. Western analysis was conducted by probing with appropriate antibodies. From the left: lane 1, crude extract; lane 2, GST; lane 3, GST-NOT5; lane 4, GST-VpU.
GST-NOT5 interaction with TBP and TAFs requires an intact TFIID complex. GST-NOT5 pull-down experiments were performed as described in Figure 3 using strain BYΔ2 (wild type) and BYΔ2(K151L K156Y). Western analysis was conducted using antibodies directed against TBP, TAF60, TAF61, and TAF90. Several other TAFs were probed for, such as TAF130, TAF40, and TAF19, but the antibodies were of insufficient quality to detect the protein in the crude extracts. Lanes 1–3, crude extracts from wild type or K15lL K156Y at 38° or 30°; lanes 4–6, GST-NOT5 pull downs using crude extracts described in lanes 1–3. It should be noted that TBP (K151L K156Y) migrates slightly faster than TBP (wild type).
DISCUSSION
Functional interaction between the CCR4-NOT complex with TBP and associated factors: In these studies we have demonstrated functional and physical interactions between components of the CCR4-NOT complex and that of TBP and its associated factors. The functional interaction between TBP and components of the CCR4-NOT complex is based on several lines of evidence. First, mutations in CAF1 and NOT4 suppress the his4-912δ insertion in a manner similar to that observed for the mutated TBP allele spt15-122. The mechanism of suppression by spt15-122 has been suggested to involve stabilization of TBP binding to a specific nonconsensus TATA sequence, CATAAA, in the his4-912δ element (Arndtet al. 1994). Our data suggest similarly that defects in CCR4-NOT protein components allow increased binding of TBP to this CATAAA sequence, although this has not been established directly. This model is consistent with the previously proposed role of NOT proteins in repressing the nonconsensus TC-dependent transcription from the HIS3 promoter (Collart 1996).
Second, we subsequently showed that mutations in the not genes allow increased transcription from several different “mutant” TATA sequences that could be classified as nonconsensus TATA elements. The unique and distinguishing feature of this analysis is that it was carried out in strains in which the HIS3 promoter was modified such that it contained only one TATA element, expression from which was activated by a wild-type allele of GCN4. The results from these studies indicate that the function of the NOT proteins is not specific to the sequence at the TATAA element. Rather, the NOT proteins are general repressors of TBP function, irrespective of the DNA sequence bound by TBP. This model is also consistent with the role of the NOT proteins as global repressors of transcription, as inferred from the fact that mutations in CAF1 and NOT genes can suppress mutations in SRB4 and RPB2, both of which are essential for global mRNA synthesis (Leeet al. 1998).
However, the mechanism by which not alleles allow increased transcription from these nonconsensus TATA sequences appears different from altering the binding specificity of TBP. This is suggested by the observation that the expression from several different TATA-like sequences was augmented by not alleles. In contrast, TBP alterations such as L205F or L114F tend to affect only expression from one or another TATA-like sequence (Arndtet al. 1994). Therefore, it is unlikely that the not deletions result in altered binding specificity of TBP. Also, a not4 deletion caused increased 3-AT resistance even in a TBP(L114F) background, suggesting that the not4 effect occurred by a mechanism separate from that of L114F effects on the DNA binding specificity of TBP. In addition, increasing the dosage of NOT1, but not any of the other CCR4-NOT components, specifically inhibited the Spt− phenotype of the spt15-122 allele. This genetic interaction suggests that there exists a dosage-dependent interaction between the NOT proteins and TBP. Similarly, we have found that a ccr4 deletion resulted in a synthetic growth defect with the taf130-ts1 allele but not with the taf130-ts2 allele (C. Denis, unpublished observations). This allele-specific interaction further suggests the functional interaction between CCR4-NOT proteins and TBP and its associated factors.
Finally, spt3, spt8, and spt15-21 alleles (all involved in affecting interaction of SPT3 with TBP) suppressed ccr4 and caf1 defects. Previously it had been shown that spt3 can suppress a not1-2 allele but was unable to suppress defects in other NOT genes (Collart 1996). This suppression of ccr4, caf1, and not1-2 by spt3 confirms the previously observed close physical association between the CCR4 and CAF1 proteins and the NOT1 protein in the CCR4-NOT complex (Draperet al. 1994; Baiet al. 1999).
NOT5 physically interacts with TFIID: The above functional interactions predict that the CCR4-NOT protein components should physically interact with TBP and/or its associated factors. We have found that NOT2 and NOT5 do indeed coimmunoprecipitate with TBP. Recently, it has also been shown that NOT1 is retained from crude extracts by GST-TBP (Leeet al. 1998) and that NOT2 can associate indirectly with ADA2, a component of the SAGA complex (Bensonet al. 1998). We failed, however, to detect NOT4, CCR4, and CAF1 in our TBP immunoprecipitates. These differences may derive from the physical associations observed within the CCR4-NOT complex (Baiet al. 1999). NOT2 and NOT5 associate with a different part of NOT1 than do CCR4 and CAF1 and we have found that NOT2 and NOT5 can associate independently of CCR4 and CAF1 (Baiet al. 1999). These observations suggest that NOT2 and NOT5 could function independently of CCR4, CAF1, and perhaps NOT4. The synthetic lethality of not4 and not5 double deletions (Oberholzer and Collart 1998) and that of a not2 and not5 deletion with ccr4 or caf1 (Baiet al. 1999) further suggest that CCR4, CAF1, and NOT4 have different roles than do NOT5 and NOT2 and therefore separate functional and physical interactions. Whether these separate roles derive solely from separate locations and actions within the CCR4-NOT complex or from NOT5-NOT2 being components of a complex separate from that of CCR4-NOT cannot as yet be determined.
In addition to TBP-NOT5 interactions, we additionally showed that TAFs 60, 61, and 90 could be specifically retained by GST-NOT5. An intact TFIID complex was found to be required for TBP and TAFs to bind to GST-NOT5. These results establish that NOT5 at least interacts with TBP through contacts with components of TFIID. However, other potential contacts of CCR4-NOT proteins to factors present in SAGA (Collart 1996; Bensonet al. 1998) cannot be excluded. Based on the diversity of phenotypes observed for CCR4-NOT defects, it is likely that the CCR4-NOT proteins are capable of making multiple contacts with TFIID, SAGA, and other components.
Mechanism of action of CCR4-NOT complex in affecting TBP function: Our genetic results and previous data (Collart and Struhl 1994; Collart 1996; Leeet al. 1998) support the model that the CCR4-NOT proteins exert their repressive effects by destabilizing functional interaction of TBP with DNA. The CCR4-NOT proteins could accomplish this by several possible mechanisms: (1) The CCR4-NOT proteins could directly bind TBP and inhibit its ability to bind DNA. This possibility seems unlikely in that, as described above, the not alleles affected functional TBP binding to a variety of nonconsensus TATA sequences and by a mechanism that is apparently separate from TBP binding to DNA. This model is also inconsistent with the fact that overexpression of TBP did not overcome the repressive effects of the NOTs at the modified HIS3 promoters used in these studies (data not shown). This implies that TBP direct binding to DNA by itself is not being affected. (2) It is possible that the CCR4-NOT proteins bind directly to TBP and prevent the interaction of TBP with other stabilizing factors like TFIIA and SPT3. While we cannot exclude this possibility on the basis of the results presented herein and elsewhere (Leeet al. 1998), it seems unlikely since no direct interaction was detected between TBP and any of the CCR4-NOT proteins (Komarnitskyet al. 1998; K. Reed and C. L. Denis, unpublished observations). (3) The CCR4-NOT proteins might contact certain factors such as TFIIA, SPT3, or TAFs that function to stabilize TBP-DNA interactions. The CCR4-NOT proteins may interact with more than one of these factors and therefore may actually destabilize the TBP-DNA interaction by multiple mechanisms. As indicated above, this latter possibility is consistent with the fact that mutations in different CCR4-NOT genes confer overlapping but not identical phenotypes.
We envisage that the CCR4-NOT proteins could sequester some components of TFIID and/or SAGA and thereby affect TBP function at certain promoters. For example, TFIID components have been shown to be required for transcription from nonconsensus TATA elements at the TRP3 promoter and the TC element of the HIS3 gene promoter (Moqtaderiet al. 1996). Therefore, the binding of NOT5 to TFIID may modulate TAF activity or prevent TFIID access to these promoters. Alternatively, the interaction of CCR4-NOT protein components with TFIID or SAGA could disrupt these complexes and thereby prevent their function. These models should be compared with the known mechanisms of action for other repressors such as NC2 and MOT1. NC2 appears to generally repress transcription by interfering with the association of TFIIB with TBP and by affecting TBP or TBP-DNA interactions so as to reduce TFIIA binding (Maldonadoet al. 1999). MOT1, in contrast, may function as a factor that aids in promoting release of TBP from nonproductive TATAA sites. This would, for instance, allow increased availability of TBP for promoters where TBP displays less affinity (Collart 1996; Maldonadoet al. 1999). Because defects in NC2 and the CCR4-NOT complex components were all able to suppress defects in general transcription due to an srb4 allele (Leeet al. 1998), our results and those of others strongly suggest that one function of the CCR4-NOT complex is to affect TBP (TFIID) activity.
Acknowledgments
We are especially grateful to David Stillman for his valuable comments and initial recommendations for pursuing this research. We are also indebted to K. Arndt, S. Buratowski, F. Winston, K. Struhl, M. Collart, T. Weil, R. Young, and members of their laboratories for providing strains, plasmids, and antibodies used in this study. The technical assistance of J. Farrell is appreciated. This research was supported by National Institutes of Health grant GM-41215, National Science Foundation grant MCB95-13412, and Hatch project H291 to C.L.D. V.B. also wishes to thank UNH for their generous support with a Dissertation Fellowship Award. This is publication number 2017 from the New Hampshire Agricultural Experiment Station.
Footnotes
-
Communicating editor: M. Hampsey
- Received December 21, 1999.
- Accepted March 30, 2000.
- Copyright © 2000 by the Genetics Society of America
