The Transcriptional Repressor Activator Protein Rap1p Is a Direct Regulator of TATA-binding Protein*

Essentially all nuclear eukaryotic gene transcription depends upon the function of the transcription factor TATA-binding protein (TBP). Here we show that the abundant, multifunctional DNA binding transcription factor repressor activator protein Rap1p interacts directly with TBP. TBP-Rap1p binding occurs efficiently in vivo at physiological expression levels, and in vitro analyses confirm that this is a direct interaction. The DNA binding domains of the two proteins mediate interaction between TBP and Rap1p. TBP-Rap1p complex formation inhibits TBP binding to TATA promoter DNA. Alterations in either Rap1p or TBP levels modulate mRNA gene transcription in vivo. We propose that Rap1p represents a heretofore unrecognized regulator of TBP.

Essentially all nuclear eukaryotic gene transcription depends upon the function of the transcription factor TATA-binding protein (TBP). Here we show that the abundant, multifunctional DNA binding transcription factor repressor activator protein Rap1p interacts directly with TBP. TBP-Rap1p binding occurs efficiently in vivo at physiological expression levels, and in vitro analyses confirm that this is a direct interaction. The DNA binding domains of the two proteins mediate interaction between TBP and Rap1p. TBP-Rap1p complex formation inhibits TBP binding to TATA promoter DNA. Alterations in either Rap1p or TBP levels modulate mRNA gene transcription in vivo. We propose that Rap1p represents a heretofore unrecognized regulator of TBP.
Eukaryotic gene transcription is controlled through the concerted action of DNA-binding transfactors, proteins that functionally interact with the transcription machinery to turn genes on and off. For mRNA-encoding genes the transcription machinery is composed of the general transcription factors (GTFs) 2 TFIIA, TFIIB, TFIID, TFIIE, TFIIF, TFIIH, and RNA polymerase (RNAP) II. DNA-bound transactivators stimulate the GTFs and RNAP II to form pre-initiation complexes (PIC) on cis-linked promoters (1). TBP recruitment to the TATA box of mRNA gene promoters is a critical, and likely rate-limiting step in nucleating this process (2)(3)(4). TBP is chaperoned to promoters by a number of proteins, and in the case of mRNA-encoding genes, TFIID is the predominant protein complex that serves this function. TFIID has 15 evolutionarily conserved subunits, TBP and 14 TBP-associated factors (TAFs) (5,6). The TATA box promoter element of mRNA-encoding genes is bound via the TBP subunit of TFIID through a process facilitated by TAFs (7)(8)(9)(10).
TBP is required for nearly all nuclear gene transcription as it is an integral subunit of the RNAP I-, II-, and III-specific initiation factors SL1, TFIID, and TFIIIB (11). Consequently, it is not surprising that TBP function is subject to tight regulation by non-GTF transcription factors such as NC2, Mot1p, and SAGA and NOT that all directly bind TBP and modulate its activity (12,13).
The only essential domain of Rap1p is its DNA-binding domain (DBD), which is composed of two Myb-like motifs. Yeast expressing only the Rap1p DBD grow, albeit extremely slowly (26). Deletion of the N-terminal BRCT-DNA bending domain has minimal effects on viability (27,28), whereas removal of sequences C-terminal to the DBD results in slow growth (26). A C-terminal activation domain, fused to a heterologous DBD, weakly activates transcription of a reporter gene (29); by contrast, lexA full-length Rap1p fusions fail to activate (30). The C-terminal silencing domain represses mating-type, telomere-proximal, and RP genes under certain circumstances (31,32). Finally, a 34-amino acid (aa) Tox domain inhibits cell growth when overexpressed fused to the DBD (17,(33)(34)(35).
In this report we describe the results of our continuing investigations into how Rap1p interacts with the transcription machinery. We have found that Rap1p directly and specifically binds TBP, and have molecularly dissected this interaction. We discuss the implications of Rap1p-TBP binding vis à vis the regulation of TBP activity.
TFIIA, TFIIB, His 6 -TBP, His 6 -TAND, and NC2 proteins were produced as described as listed above. His 6 -GST-TBP and His 6 -GST-Rap1p variants were expressed and prepared as for His 6 -TBP. His 6 -GST fusion variants were bound to glutathione-agarose beads, glutathione-eluted, and treated with His 6tagged precission protease on Ni-nitriloacetic acid-agarose beads for 1 h at room temperature and centrifuged to remove the beads. Supernatants were collected, analyzed for protein, and stored at Ϫ80°C.
DNase I Footprinting-DNase I footprinting using Ad2 MLP (major late promoter) (Ϫ250 to ϩ195 relative to the ϩ1 transcription start site) and RPS1A (Ϫ420 to ϩ23 relative to the A ϩ1 TG Rps1p methionine codon) probes was performed as described (5).
GST Pulldown Assay-100-l binding reactions were performed in binding buffer: 20 mM HEPES-KOH (pH 7.6), 10% glycerol, 70 mM potassium acetate, 0.1 mM EDTA, 1 mM dithiothreitol, 5 mM MgAc, 0.01% Nonidet P-40, plus 10 l of beads (300 ng of GST fusion protein/reaction), and the proteins to be tested for binding. Reactions were mixed 30 min at room temperature, then beads were washed 3 times with binding buffer, resuspended in 20 l of SDS-PAGE sample buffer, heat denatured, and analyzed by SDS-PAGE on NuPAGE 4 -12% polyacrylamide gels followed by Sypro Ruby staining or immunoblotting.
TBP Cross-linking-TBP (100 ng/l) in 19 l of 20 mM HEPES-KOH (pH 7.6), 10% glycerol, 150 mM NaAc, 5 mM MgAc, and 0.05% Nonidet P-40 was incubated for 30 min at room temperature, then 1 l of 20% formaldehyde, 20 mM HEPES-KOH (pH 7.6), 150 mM sodium acetate was added and incubation continued for 15 min at room temperature. Crosslinking was quenched by addition of 3 l of 2.5 M glycine and incubation for 5 min at room temperature. 20 l of SDS-PAGE sample buffer was added and samples were loaded, without heating, on 4 -12% NuPAGE gels followed by Sypro Ruby staining or immunoblotting.
␤-Galactosidase Assay-Freshly transformed cells were used. Colonies were used to inoculate 5 ml of selective media cultures, grown to 1 A 600 /ml, and cells tested for ␤-galactosidase activity as described by Blair and Cullen (72).
Primer Extension and Chromatin Immunoprecipitation (ChIP) Assays-Primer extension (67) and ChIP assays were performed essentially as described (55) except ChIP PCR products were fractionated by non-denaturing polyacrylamide gel electrophoresis in 1ϫ TBE buffer and DNA quantified by staining with Sypro Gold (Invitrogen).

RESULTS
While conducting footprinting experiments to examine the binding of yeast TFIID to a battery of Rap1p-dependent and Rap1p-independent genes, control reactions showed that Rap1p addition blocked TBP-TATA DNA binding. Given the central role of TBP in nuclear gene transcription, the potency of inhibition of TBP-DNA binding by nanomolar concentrations of Rap1p, the nuclear abundance, and high concentration of both proteins (Ն100 M), we examined this phenomenon in more detail.
Rap1p Inhibits TATA DNA Binding by TBP-The inhibitory effect of Rap1p upon the binding of TBP (see Fig. 1A for purity of proteins) to the Adenovirus 2 major late promoter (Ad2 MLP), a heterologous DNA devoid of RAP1 sites, is shown in  10 -14, and 15-19). Notably, TFIIB failed to block Rap1p from inhibiting TBP-TATA binding (Fig. 1D, compare lanes 20 -22 with lanes 2-19 and 23-25). (Note that the apparent protection of DNA observed when both TBP and TFIIB are added to the footprinting assays is not reproducible (data not shown), and likely results from enhanced nuclease digestion of the probe when these two proteins are added simultaneously.) Together, these data suggested that Rap1p and TBP directly interacted, and that a specific domain(s) of TBP was targeted.
Identification of TBP Residues That Interact with Rap1p-To more precisely map the Rap1p interaction domain we used a family of truncation and point mutant forms of TBP (Fig. 4A). To document the published activities of these TBP mutants (Fig. 4B), and to demonstrate that all variants had some definable biochemical activity and hence were not inactive in Rap1p binding due to the fact that they were, for example, just unfolded, we measured both dimerization (Fig. 4B, lanes 2-19, X-linked, anti-TBP IgG I-BLOT) and DNA binding, ϮRap1p and ϮTFIIA (Fig. 4C). As reported (69,70), these TBPs variably dimerized (Fig. 4B), and only TBP point mutant variants E186Q, Q68R, F177R, G180R, and N95R bound TATA. In every case when DNA binding was observed TATA binding was sensitive to Rap1p and protected by TFIIA (Fig. 4C). When tested via GST-TBP pulldown, all TBP point mutant variants except R98E and R196E bound Rap1p (Fig. 4D, cf. lanes  21 and 41). TBP amino acids Arg-98 and Arg-196 reside on opposite sides of the cavity forming the concave DBD of TBP (red shaded residues, Figs. 3A and 4A). Consistent with these data, only the GST-TBP truncation mutant composed of aa 71 to 240 was able to bind Rap1p (Fig. 4D, lane 52). These data indicated that when Rap1p bound TBP via (minimally) amino acids Arg-98 and Arg-196, access to the TBP DBD and hence DNA binding was blocked.
Rap1p and TBP Form a Complex in Vivo-To assess whether Rap1p and TBP interacted in vivo we performed co-immunoprecipitation (co-IP) studies. Whole cell extracts were prepared from control and tagged strains (TAP-, Myc-, HA-) that express Rap1p from the chromosomal RAP1 promoter. WCEs were fractionated by SDS-PAGE, blotted, and probed with appropriate antibodies to verify the specificity of tagging, WCE TBP content, and equal protein transfer (Fig. 5A, left panels). The WCEs were also used to form IPs using anti-TAG antibodies. IPs were probed by immunoblotting with anti-Rap1p and anti-TBP IgGs to monitor co-precipitation (Fig. 5A, right panel, top and bottom). The results indicated that under normal physiological expression conditions Rap1p and TBP were indeed associated.
To test the ability of WT and mutant Rap1p variants to bind endogenous and exogenous TBP we used cells that expressed a tetracycline (TET)-regulated chromosomal RAP1 allele (71) into which we introduced multicopy plasmids overexpressing Rap1p-FLAG and/or Myc-TBP under UAS GAL control (see schematic, Fig. 5B). Cells were grown on raffinose, Ϫ/ϩ TET, and expression of the episomal TBP and Rap1p variants induced for 3 h by the addition of galactose. Under these conditions Rap1p-FLAG was able to interact with both endogenous and exogenous Myc-TBP (Fig. 5C, lanes 3, 4, 9 ,10,13,14,17,and 18;left panel,ϪTET,ϩTET), and, consistent with the biochemical analyses presented above, complex formation was Rap1p DBD-dependent. Similar results were observed in the reciprocal experiment (Fig. 5C, right panel). Together these data indicated both that Rap1p and TBP indeed interacted in  vivo, and that Rap1p bound TBP through the Rap1p DBD. These results were consistent with our in vitro analyses.

B T B P N c b 2 p T F II B T A N D N c b 1 p G S T -R a p 1 p T F II
Rap1p DBD Modulates mRNA Gene Transcription in Vivo-We used two approaches to test whether Rap1p regulated TBP transcriptional function. First, we took advantage of the fact that overexpression of TBP can increase activator-independent transcription (AIT) of certain mRNA-encoding genes, although the exact mechanism responsible for this phenomenon is at present not fully understood. We observed that different TBP mutant proteins variably stimulated AIT (Fig. 6A, top) when expressed in cells carrying either of two reporter plasmids, pADH1-lacZ or pCYC1-lacZ. These results are consistent with the observations of others (69,(72)(73)(74) who have shown that certain TBP variants, when overexpressed, stimulate AIT. Immunoblotting showed that all of the TBP proteins tested were equally expressed (Fig. 6A, bottom).
Given the fact that Rap1p specifically binds to and inactivates TBP, we reasoned that introduction of both Rap1p and TBP into cells carrying the AIT reporter plasmid should diminish the observed TBP-driven reporter gene transcription. We tested this hypothesis by co-overexpressing forms of TBP (WT, R98E, T124R, V161R, R196E, N69S, and L205K) and various isoforms of Rap1p (⌬Tox, DBD, and ⌬DBD) in cells carrying pADH1-lacZ; all of these proteins were equally expressed (Fig.  6B, right). As seen in the data presented in Fig. 6A, overexpressed TBP increased transcription of the ADH1 promoterdriven lacZ reporter, and two TBP mutants (V161R and N69S) were particularly active in this regard (Fig. 6B, right, white bars). Importantly, co-overexpression of Rap1p DBD (black bars), but not Rap1⌬DBD (light gray bars), significantly reduced TBPdriven AIT in all instances except in the case of the two TBP variants, R98E and R196E, that we had found to be unable to H2' H2' S1 S1 S1' S1'  bind Rap1p (cf. Fig. 4). The DBD-containing Rap1⌬Tox protein, which can be overexpressed as it lacks the Tox domain, also inhibited TBP-driven AIT, although to a lesser extent than the Rap1p DBD (dark gray bars); the R98E and R196E mutants were also refractory to Rap1p⌬Tox-mediated inhibition of AIT. In summary, this data indicated that the Rap1p DBD could modulate the transcriptional activity of TBP in vivo.
Second, to assess whether the Rap1p DBD also affected the transcription of single-copy, endogenous, chromosomal mRNA-encoding genes, we utilized a Cu 2ϩ -modulated expression shutoff system (Copper-degron) (63) to assess the effects of Rap1p depletion on transcription. A control strain (ZMY60) and Cu 2ϩ -depletable RAP1 strain (LEV391) (64) were treated with CuSO 4 for 3 h (see efficacy and time course of Rap1p depletion (supplemental Fig. S1) and transcription modulation (supplemental Fig. S2)), RNA was then extracted and specific RNA levels scored using equal amounts of total RNA from each strain. When Rap1p was depleted, RNAP II-mediated transcription of TFIID-independent (PGK1, ADH1, ADH3, HIS3, and HHT1) and TFIID-dependent mRNA-encoding genes (RPS2, RPS3, and RPL3) significantly increased (Fig. 7A); this increase in transcription was particularly marked for the Rap1p-target genes (PGK1, ADH1, ADH3, RPS2, RPS3, and RPL3). This effect was RNAP II-restricted because neither RDN1 gene transcription by RNAP I (supplemental Fig. S2), nor SCR1 and tL(CAA)C (tRNA Leu3 ) gene transcription by RNAP III was affected by Rap1p depletion (Fig. 7A). This experiment was then repeated, but upon endogenous Rap1p depletion various forms of Rap1p (WT, ⌬Tox, DBD, and ⌬DBD) were expressed from the RAP1 enhancer/promoter at physiological levels (see supplemental Fig. S1D). RNA was again prepared and analyzed for transcript levels. We found that any form of co-expressed Rap1p that carried the DBD (i.e. WT, ⌬Tox, and DBD) blocked the Cu 2ϩ -induced up-regulation of mRNA gene transcription (Fig. 7B), presumably by sequestering TBP. Furthermore, and importantly, the extent of Rap1p-DBD-mediated modulation of mRNA gene transcription was proportional to the levels of these Rap1p variants (supplemental Fig. S2B). Collectively this series of four experiments argued that in vivo, Rap1p, working through DBD-mediated interactions with TBP, modulated mRNA gene transcription.
Up-regulation of mRNA Levels Seen in LEV391 Cells upon Rap1p Depletion Is Associated with Increased Recruitment of the Transcription Machinery-To demonstrate that the observed increases in mRNA levels seen upon Rap1p depletion

R A P 1 R A P 1 -T A P T A P -R A P 1 R A P 1 M y c -R A P 1 R A P 1 H A -R A P 1 R A P 1 R A P 1 -T A P T A P -R A P 1 R A P 1 M y c -R A P 1 R A P 1 H A -R A P 1 R A P 1 R A P 1 -T A P T A P -R A P 1 R A P 1 M y c -R A P 1 R A P 1 H A -R A P 1 R A P 1 R A P 1 -T A P T A P -R A P 1 R A P 1 M y c -R A P 1 R A P 1 H A -R A P 1 R A P 1 R A P 1 -T A P T A P -R A P 1 R A P 1 M y c -R A P 1 R A P 1 H A -R A P 1 R A P 1 R A P 1 -T A P T A P -R A P 1 R A P 1 M y c -R A P 1 R A P 1 H A -R A P 1
Sypro Ruby WCE Anti-Rap1p IgG Anti-TAP IgG Anti-Myc IgG Anti-HA IgG Anti-TBP IgG  were indeed transcriptional effects, we performed ChIP studies to directly measure PIC formation. We scored occupancy of RNAP II, TBP, and Rap1p on chromosomal PGK1: UAS RAP1 , core promoter and open reading frame, in WT (ZMY60) and the Cu 2ϩ -RAP1 degron (LEV391) cells treated ϮCuSO 4 . The resulting data demonstrated that the increase in mRNA PGK1 observed above was the result of increased PIC formation, and hence a direct effect of Rap1p depletion (Fig. 8A). Furthermore, consistent with the above described in vivo transcriptional effects of Rap1p DBD on TBP-directed mRNA gene transcription, we observed that overexpression of the Rap1p DBD, but not the Rap1p-⌬DBD variant, dominantly inhibited yeast cell growth (Fig. 8B).

DISCUSSION
The evolutionarily conserved TATA-binding protein is essential for most nuclear gene transcription in eukaryotes (75,76). By binding to promoter DNA near the start site of transcription TBP catalyzes PIC formation on mRNA-encoding genes by facilitating the cooperative interaction of RNAP II and the GTFs with the promoter. TBP may be escorted to promoters in the form of the TFIID coregulator, or via other coregulators (i.e. Mediator, SAGA, and Mot1p). On RNAP I-and III-transcribed genes the GTFs SL1, SNAP C , and TFIIIB operate in a similar fashion by chaperoning TBP to target promoters for subsequent PIC assembly and transcription.
Consistent with the central role of TBP in transcription, its activity is subject to exquisite control by a plethora of TBP-interacting proteins. These include NC2, TFIIA, TFIIB (77) (87), and the NOT complex (88). Interestingly, many of these TBP-targeted factors both activate and repress transcription (76). Finally, Pugh and colleagues (70,73) have shown that TBP homodimerization likely autoregulates TBP function. An intriguing feature of these disparate TBP-directed regulators is the fact that essentially all utilize direct protein-TBP interactions that target amino acid residues in and around the concave DBD of TBP to regulate TBP activity. Here we report our discovery of a new member of this family of TBPtargeted transcriptional regulatory proteins, Rap1p.
Addition of Rap1p to TBP-TATA DNA binding reactions inhibited the binding of free TBP, but not TBP-TATA complexes, whereas TFIIA, but not TFIIB, protected TBP from inhibition by Rap1p; TBP-Rap1p interaction was bimolecular (Fig.  1). The basic Rap1p DBD (pI 8.6) bound directly to the basic TBP (pI 10.2) to inhibit DNA binding (Fig. 2) and the TFIIA-, TATA-, TAND-, and NC2-interaction domains of TBP were involved in binding (Fig. 3). Mapping indicated that two residues, Arg-98 and Arg-196, on either side of the concave DBD of TBP specifically interacted with the Rap1p DBD providing a molecular explanation for how Rap1p-TBP complex formation could block TBP-DNA binding (Fig. 4). Not only did TBP and Rap1p interact specifically and efficiently in vitro, we demonstrated by co-IP that under normal expression levels TBP could be found associated with Rap1p in vivo. Both untagged endogenous TBP and ectopically co-expressed tagged-TBP interacted with Rap1p in vivo (Fig. 5). Although these co-IP experiments could not indicate if this interaction was direct, we did show that like the in vitro situation, TBP-Rap1p co-IP depended upon the Rap1p DBD. Furthermore, given our data showing that the Taf1p TAND blocked interaction between TBP and Rap1p (Fig. 3), and, that previously we have shown that TFIID directly binds both Rap1p DBD and C-terminal domains with roughly equal affinity, leads us to believe that TBP-Rap1p co-precipitation does not involve TFIID (cf. Ref. 55). Most importantly, we showed that the Rap1p DBD functionally interacted with TBP in vivo by demonstrating that forms of Rap1p carrying the DBD specifically regulated both TBP-driven activator-independent mRNA gene transcription (Fig. 6) as well as endogenous chromosomal mRNA gene transcription (Figs. 7 and 8). Not surprisingly then, overexpression of the DBD, devoid of any Tox domain sequences, dominantly inhibited yeast cell growth (Fig. 8).
Collectively, our data lead us to entertain a model wherein we hypothesize that Rap1p can act as a "TBP buffer," serving to sequester TBP as an inactive TBP-Rap1p complex under conditions of excess free TBP. Such complex formation could have multiple consequences (see below). It should be noted that like TBP, Rap1p is an abundant protein (ϳ2 ϫ 10 4 molecules/cell). However, the critical question for both proteins, indeed for most interacting nuclear transcription proteins is: what fraction of either TBP or Rap1p is free? It is clear from our data that A, Rap1p depletion leads to increased PIC formation as scored by ChIP on PGK1. Control ZMY60, and RAP1degron LEV391 cells were grown for 3 h in the absence (Ϫ) and presence (ϩ) of 500 M CuSO 4 . Cells were formaldehyde cross-linked and processed for ChIP. IPs were prepared using sonicated chromatin and anti-RNAPII CTD (monoclonal antibody 8WG16, Covance; ␣pol II), or rabbit polyclonal anti-TBP (␣TBP) or anti-Rap1p (␣Rap) IgGs (55). Input (2%) and immunoprecipitated DNAs (100%) were used to template multiplexed PCR using PGK1 region-specific primer pairs; following 20 cycles of PCR 20% of both input-and IP-templated reactions were loaded on a TBE-buffered 5% polyacrylamide gel, fractionated, and product DNAs visualized by staining with Sypro Gold (Invitrogen) and scanning, scan image (shown). B, Rap1p DBD dominantly inhibits yeast cell growth. Yeast strain YTW22 (spt15⌬) containing a URA3-marked plasmid expressing SPT15 from the PGK1 promoter was transformed with plasmids (HIS3-marked for TBP; LEU2-marked for Rap1p) that expressed Myc-TBP and FLAG-Rap1p forms from UAS GAL . Cells were pre-grown for 2 days on liquid selective media plus 2% raffinose or galactose, plus uracil and then serially diluted 5-fold and spotted onto plates that contained raffinose, galactose, or galactose ϩ 5-fluoroorotic acid (5-FOA), incubated at 30°C for 3 days and photographed (left). Expression levels of Myc-TBP and FLAG-Rap1p forms were determined by immunoblotting (right).
TBP-Rap1p complex formation excludes functional TBP interactions such as TATA DNA binding and holo-TFIID complex formation; this is schematically illustrated in Fig. 9A.
One consequence of Rap1p-TBP complex formation could be that Rap1p operates mechanistically in a fashion similar to other TBP-directed transcriptional regulators such as NC2, Mot1p, and SAGA coregulators, that all form stable complexes with TBP (77,79,89). A corollary of this model is that by associating or dissociating from TBP, Rap1p will negatively or positively affect transcription by decreasing or increasing the pool of free TBP. In vivo, TBP can clearly shuttle from TFIIIB to TFIID, but apparently fails to shuttle efficiently from TFIID to TFIIIB, or from TFIIIB to the yeast SL1-like RNAP I GTF (90). Moreover, we have previously shown that TBP dynamically associates with TFIID in vitro (5). These data are all consistent with our observations presented in Figs. 7, 8, and supplemental S1 and S2, which demonstrated an RNAP II-specific effect of Rap1p depletion on nuclear gene transcription. In this case, when Rap1p levels dropped, the amount of TBP available to interact with the 14-subunit TFIID TAF complex could go up, presumably increasing the amount of functional, transcriptionally competent holo-TFIID, leading to stimulation of mRNA gene transcription (cf. Fig. 9B). Similar arguments can be made for TFIID-independent transcription, for example, genes driven by the SAGA-TBP, Mot1p-TBP, or other coregulator-TBP complexes. Finally, one cannot rule out the fact that such non-TFIID coregulator-TBP complexes may (spuriously) drive mRNA gene transcription under the artificial TBP excess conditions we have generated here using the LEV391 Rap1p-shutoff strain. Dramatic changes in essential coregulator and/or activator requirements have been seen before in systems genetically manipulated to deplete specific transcription proteins (91)(92)(93). Thus, it is possible that under physiological conditions Rap1p normally functions to chaperone TBP transfer between TFIIIB and TFIID, thereby fine-tuning mRNA, tRNA, and 5 S rRNA gene transcription. Rap1p has been implicated in the regulation of the most actively transcribed genes in yeast cells, the RP-and glycolytic enzyme-encoding genes (18), hence Rap1p and Rap1p-driven mRNA gene transcription is subjected to growth rate-sensitive signaling pathways (16) and thus potentially ideally situated to modulate transcription via interaction with TBP.
A second possible consequence of Rap1p-TBP complex formation is that this interaction could antagonize the negative action(s) of any or all of the other known TBP-targeted transcriptional regulators. Indeed, we have shown here (cf. Fig. 3) that TFIIA, TATA DNA, Taf1p TAND, and NC2 all potently prevented the binding of Rap1p to TBP. Hence, one function of Rap1p-TBP complex formation may be to affect the interplay of these TBP-targeted regulators with TBP under appropriate conditions. Any molecule that dramatically alters intranuclear distribution of TBP between specific, TBP-linked TBP-TAF complexes will effect global nuclear gene expression (94). Experiments are in progress to directly test the ability of Rap1p to influence the interactions between TBP and Mediator, SAGA, Mot1p, NC2, and other TBP-binding proteins.
Finally, third, Rap1p-TBP complex formation could control active/free TBP levels by affecting TBP dimerization. TBP dimerization has been hypothesized to importantly contribute to the control of mRNA gene transcription in vivo, and one of the molecules that blocks TBP-Rap1p complex formation, TFIIA, has been implicated in facilitating TBP-TBP dimer dissociation (95). If TBP dimerization contributes to regulation of free TBP levels, then Rap1p may help pull dimeric TBP, via mass action, to the monomer, the form that is competent to interact with TAFs to form functional GTFs and/or other functional TBP-containing coregulators. If this is the case, our model would need to include one additional TBP state, TBP-TBP dimers, in equilibrium with the Rap1p-TBP complex and monomeric (i.e. functional, non-dimerized) TBP. Our future work will dissect the molecular mechanisms by which the novel TBP regulator, Rap1p, serves to modulate TBP function within  Taf1p TAND; and III, TBP-Rap1p. All three high affinity interactions involve the concave DNA binding surface of TBP, hence binding of TBP to any one of these ligands blocks binding by the others (cf. Fig. 3). Listed are the K d values and complex half-lives we have measured for complexes I and II (58,59,65,68,69); complex III, N.D., not yet determined. B, model illustrating the linkage between the Rap1p-TBP complex and TFIID-driven mRNA gene transcription (1). If the in vivo levels of Rap1p are decreased (X), either naturally or by genetic depletion (as with the copper degron in strain LEV391; Fig. 7 and supplemental Fig. S2) there will be a corresponding increase in free TBP (2); this drop in Rap1p causes an increase in the intranuclear [TBP] (3). Because TBP dynamically interacts with the 14-subunit TAF II complex (65), an increase in the concentration of TBP could, by mass action (4), increase the formation and concentration of holo-TFIID (TFIID). This increase in holo-TFIID could then drive mRNA gene transcription; Txn (5). Note that an alternative to increased holo-TFIID (or in addition to), increased concentrations of TBP could also stimulate formation of SAGA-TBP, mediator-TBP, Mot1p-TBP etc., complexes that themselves could also drive an increase in mRNA gene transcription in a physiological or non-physiological fashion. the nucleus, with a particular eye toward placing Rap1p within the broader context of other TBP-targeted transcriptional coregulators.