RNA polymerase II-associated protein (RAP) 74 binds transcription factor (TF) IIB and blocks TFIIB-RAP30 binding.

A set of deletion mutants of human RNA polymerase II-associated protein (RAP) 30, the small subunit of transcription factor IIF (TFIIF; RAP30/74), was constructed to map functional domains. Mutants were tested for accurate transcriptional activity, RAP74 binding, and TFIIB binding. Transcription assays indicate the importance of both N- and C-terminal sequences for RAP30 function. RAP74 binds to the N-terminal region of RAP30 between amino acids 1 and 98. TFIIB binds to an overlapping region of RAP30, localized to amino acids 1-176 (amino acids 27-152 comprise a minimal binding region). The C-terminal region of RAP74 (amino acids 358-517) binds directly and independently to TFIIB. Interestingly, RAP74 blocks TFIIB-RAP30 binding, both by binding TFIIB and by binding RAP30. When the TFIIF complex is intact, therefore, TFIIB-TFIIF contact is maintained through RAP74. If the TFIIB-RAP30 interaction is physiologically important, the TFIIF complex must dissociate within some transcription complexes.

A minimal pathway for assembly of pre-initiation complexes on RNA polymerase II-dependent promoters has been defined in vitro (1)(2)(3)(4)(5). On promoters that include a TATA box, TBP 1 (TATA-binding protein) first binds to this recognition sequence. TFIIA facilitates this interaction. TFIIB can enter the complex either by binding to these template-associated factors or by binding to RNA polymerase II. For a tight association with the complex, polymerase must first associate with TFIIF, so TFIIB and TFIIF cooperate to bring polymerase into the pre-initiation complex (4 -7). TFIIE and TFIIH are additionally required for accurate initiation from linear DNA templates.
TFIIB is a single polypeptide of 33 kDa (8,9) that can be divided by sequence and mutational analysis into N-and Cterminal regions. Near the N terminus is a probable Zn 2ϩbinding sequence that is important for pre-initiation complex assembly (10). Immediately adjacent to the Zn 2ϩ -binding motif is a highly conserved sequence that is important for selection of transcriptional start sites (11). This N-terminal region of TFIIB binds the RAP30 subunit of TFIIF and also RNA polymerase II (12,13), and mutants in this region inhibit assembly of tran-scription intermediates, probably because they fail to interact with either TFIIF or polymerase (10,(12)(13)(14). The C-terminal region of TFIIB contains two imperfect direct repeats. Near the end of the first repeat is a very basic sequence, and the region surrounding this structure includes separate binding sites for TBP and the acidic activation domain of herpes simplex virus VP16 (15,16). In the absence of other factors, the N-and C-terminal regions of TFIIB fold on one another, masking binding sites for other transcription factors. Binding to VP16 opens up TFIIB to expose the C-terminal region, which binds TBP, and the N-terminal region, which binds RNA polymerase II and TFIIF (16). Thus, TFIIB forms a bridge between TBP and RNA polymerase II-TFIIF, in assembly of the pre-initiation complex, and the assembly steps in which TFIIB is involved are influenced by regulators. Recent reports of an x-ray crystal structure of TATA box DNA with C-terminal fragments of TBP and TFIIB (17) and a nuclear magnetic resonance structure of a C-terminal fragment of TFIIB (18) support this model for TFIIB function in assembly. The C-terminal region of TFIIB binds TBP, activators, and template, and the N-terminal region is presented as a scaffold for assembly of RNA polymerase II-TFIIF into the complex (17,18).
TFIIF is a heteromeric factor of 28-(RAP30) and 58-kDa (RAP74) subunits (19 -22), but there is some indication that in certain contexts these subunits may enter complexes as separate factors (7,23,24). Accurate initiation has been demonstrated from highly supercoiled templates using a system consisting of RNA polymerase II, TFIIB, and either TBP (25) or YY1, which is an initiator binding protein (26). One implication of these observations is that RNA polymerase II and TFIIB might minimally suffice to select transcriptional start sites. Consistent with this view, swapping Schizosaccharomyces pombe for Saccharomyces cerevisiae TFIIB and RNA polymerase II, in a system otherwise comprised of S. cerevisiae factors, shifts the position of the transcriptional start to that characteristic of S. pombe (27). Also, some sua7 mutants in the S. cerevisiae gene encoding TFIIB are altered for selection of initiation sites (28). sua8 mutants, in the gene encoding the largest subunit of RNA polymerase II, affect transcriptional starts in a very similar way to these sua7 mutants (29). Interestingly, mutations in the S. cerevisiae SSU71/TFG1 gene, which encodes the homologue of the RAP74 subunit of human TFIIF, suppress abnormal start site selection in a sua7ssu71 double mutant (30). By itself the ssu71 mutant does not affect transcriptional starts, so TFIIF may suppress the sua7 mutant indirectly through another general factor. The RAP30 subunit of human TFIIF has been shown to interact physically with the N-terminal region of TFIIB (12), so the large subunit of TFIIF interacts genetically, and the small subunit interacts physically, with TFIIB.
The largest subunit of RNA polymerase II has an interesting C-terminal domain (CTD) that consists of 52 repeats of the consensus sequence YSPTSPS (31,32). RNA polymerase II enters the pre-initiation complex most efficiently in the dephosphorylated IIa state. Within the complex the CTD is multiphosphorylated on the SP serines, by a subunit of TFIIH and possibly other CTD kinases. Elongating RNA polymerase II molecules are primarily in the highly phosphorylated IIo state. So the level of CTD phosphorylation may regulate elongation, and a CTD phosphatase may be important for recycling dephosphorylated polymerase IIa to a promoter after termination.
A CTD phosphatase has recently been identified that binds to RNA polymerase II (33,34). This phosphatase interacts with a region of polymerase distinct from the CTD and requires this contact for its dephosphorylation activity. The RAP74 subunit of TFIIF stimulates CTD phosphatase activity by binding RNA polymerase II (34). Consistent with this view, the C-terminal region of RAP74, which binds RNA polymerase II, is required for phosphatase stimulation. The C-terminal region of RAP74 is masked for phosphatase stimulation (34) just as it is for RNA polymerase II binding (35) and for TFIIB binding (this paper). Consistent with functional interaction between TFIIB and TFIIF, TFIIB suppresses stimulation of CTD phosphatase activity by RAP74 (34).
In this report, the region of RAP30 that binds RAP74 and the regions of RAP30 and RAP74 that bind TFIIB were mapped. The region of RAP30 that binds RAP74 overlaps with the region of RAP30 that binds TFIIB. RAP74 binding to TFIIB and RAP74 binding to RAP30 both block TFIIB-RAP30 binding.

EXPERIMENTAL PROCEDURES
Protein Reagents-Fragments of a human RAP30 cDNA (36) were subcloned into pET23d (Novagen) and transformed into Escherichia coli BL21(DE3) for expression (37). To construct RAP30-(1-227), an NcoI to BamHI (blunt: end-filled with Klenow DNA polymerase I and deoxynucleoside triphosphates) fragment was subcloned between the NcoI and HincII sites of the vector. To construct RAP30-(1-176) an NcoI to XhoI (blunt) fragment was cloned between the NcoI and HincII sites of the vector. These mutants have a DKLAAALEHHHHHH C-terminal extension. To construct RAP30-(1-118), an NcoI to PvuII fragment was subcloned between the NcoI and XhoI (blunt) sites of the vector. This mutant has an HHHHHH C-terminal extension. To construct RAP30-(1-98) an NcoI to SstI (blunt) fragment was subcloned between the NcoI and HincII sites of the vector. This mutant has a DKLAAALEHHH-HHH C-terminal extension. chromatography. Protein concentrations were determined using molar extinction coefficients calculated on the basis of aromatic amino acid composition (38).
RAP74 and RAP74 mutants were prepared as described (35). A production vector for human TFIIB and instructions for its use were the kind gifts of D. Reinberg (8). RNA polymerase II was isolated from calf thymus (39).
In Vitro Transcription-The transcriptional activity of RAP30 mutants was tested as described previously (37). A TFIIF-depleted extract was prepared by immunodepletion using anti-RAP30 antibodies. This extract was combined with adenovirus major late promoter template DNA (pML digested with SmaI, 50 g/ml), RAP74, and RAP30 wild type or mutants and preincubated for 1 h in buffer containing 12 mM Hepes, pH 7.9, 12% glycerol, 60 mM KCl, 12 mM MgCl 2 , 0.2 mM EDTA, 3.2 mM EGTA, and 1.2 mM dithiothreitol. 600 M ATP, CTP, UTP, and 25 M [␣ -32 P]GTP were added to the reaction, and transcription was allowed to proceed for 30 min. Transcripts were isolated and analyzed by polyacrylamide gel electrophoresis in 50% (w/v) urea and 1 ϫ TBE. The accurately initiated transcript (217 nucleotides) was quantitated using a Molecular Dynamics phosphorimager.
Protein-Protein Interaction Assays, TFIIB Binding to RAP30 and RAP74 Mutants-TFIIB was immobilized on Affi-Gel 10 (Bio-Rad) at a density of about 1 mg of TFIIB per ml of resin. 20 l of TFIIB beads and 300 pmol of RAP30 or RAP74 mutant were incubated at 4°C for 1 h in 0.5-ml storage buffer containing 0.1 M KCl (SB 0.1) and 0.2% BSA. SB contains 20 mM Hepes pH 7.9, 20% glycerol (w/v), 1 mM EDTA, 1 mM EGTA, and variable KCl concentration (SB 0.1 contains 0.1 M KCl). Beads were washed with 1 ml of SB 0.25, and bound proteins were eluted in 50 l of SB 0.5. 30 l of this eluate was analyzed on a 15% SDS-PAGE gel developed with silver nitrate. 20 -30% of input RAP30 and RAP74 were retained on TFIIB affinity beads in these experiments. Affi-Gel beads without bound protein ligand were used as a negative control.
ELISA assays were done as described (40). Microtiter wells (Becton Dickinson; Pro-Bind) were coated overnight with the protein to be immobilized (5 g/ml) in 100 l of 50 mM sodium borate, pH 9.0, at 4°C. Wells were washed three times with 200 l of phosphate-buffered saline (PBS) containing 0.2% BSA and 0.05% Tween 20 (PBSBT) and blocked with 200 l of PBSBT for an hour at room temperature. Mobile proteins were added in 50 l of SB 0.1 containing 0.2% BSA and incubated for 15 min at room temperature. Glutaraldehyde was added to a final concentration of 1% for 30 min to cross-link protein-protein interactions. Wells were washed three times with 200 l of PBSBT. 100 l of rabbit antiserum (1:1000 diluted) directed against the mobile phase protein was added to each well and incubated 2 h at room temperature. Wells were washed three times with PBSBT, and 100 l of horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (Bio-Rad; 1:3000 diluted) was added and incubated for 1 h at room temperature. Wells were washed 3 times with PBSBT. Color was developed with 100 l of 50 mM sodium citrate, pH 4.0, 0.03% H 2 O 2 , and 0.4 mg/ml ophenylenediamine. The development was stopped by addition of 100 l of 4 N H 2 SO 4 . The reaction was measured for absorbance at 490 nm using a plate reader (Bio Tek Instruments; EL310). All determinations were done in duplicate and reported as average values.

RESULTS
A set of RAP30 and RAP74 deletion mutants was constructed with C-terminal histidine tags to aid in purification and binding reactions ( Fig. 1; Ref. 35). Each of the RAP30 mutants was tested for the ability to stimulate accurate runoff transcription from the adenovirus major late promoter (Fig. 2). Full-length RAP30-(1-249) was much more active in this assay than any of the deletion mutants. RAP30-(1-227) and -(1-176) supported a much lower level of activity. RAP30-(1-152) had barely detectable activity, and RAP30-(1-118) was inactive (data not shown). These results are consistent with previously published data from the Conaway laboratory (41,42) that indicates that the C terminus of RAP30 is involved in DNA binding and is important for accurate initiation. Activity was also very sensitive to deletion of N-terminal sequences. RAP30-(27-249) supported very weak activity (Fig. 2), but RAP30-(51-249) was inactive (data not shown). Using a set of RAP30 mutants with short internal deletions, Conaway and co-workers (43) have recently shown that amino acid sequences between 16 -30 and 136 -210 are critical for initiation of transcription. The extract system used by our laboratory is more tolerant of C-terminal region mutations than the reconstituted system, perhaps because of proteins present in the extract that are not counted among the known general transcription factors.
Yonaha et al. (44) demonstrated that RAP30 amino acids 1-110 were sufficient for binding to RAP74, and Tan et al. (43) demonstrated that amino acids between 16 and 20 were critical for this interaction. Using an affinity bead procedure, we have shown that RAP30 amino acids from 1 to 98 are sufficient for RAP74 binding (data not shown; results tabulated in Fig. 1). Taken together with the results from accurate transcription assays (Fig. 2), these data indicate that most of the RAP30 deletion mutants used in these studies have preserved at least some of the known functions of RAP30 for supporting accurate transcription and/or RAP74 binding.
Interactions between TFIIF subunits and TFIIB-Because of genetic and physical interactions between TFIIF and TFIIB (12,30), RAP30 and RAP74 were tested for binding to TFIIB immobilized on agarose beads (Fig. 3). In Fig. 3A, TFIIB is shown to bind to both RAP30 (lane 2) and RAP74 (lane 4). This result confirms the previous report of a direct interaction between TFIIB and RAP30 (12) and demonstrates for the first time that TFIIB interacts directly with RAP74.
To demonstrate the specificity of TFIIB-TFIIF interactions and to indicate their functional importance, the regions of RAP30 and RAP74 required to bind TFIIB were mapped. In Fig. 3, B and C, RAP30 deletion mutants were tested for TFIIB binding. Interaction was assayed using two methods that may differ in sensitivity. For the experiment shown in Fig. 3B, TFIIB was covalently immobilized on agarose beads and used as an affinity adsorbent for RAP30 mutants. RAP30 and RAP30-(1-176) bound most tightly to TFIIB (lanes 1 and 6). RAP30-(1-152) also bound but much less efficiently (lane 7). None of the other deletion mutants was observed to bind. Apparently, primary sequence distributed over a significant portion of RAP30 contributes to TFIIB binding. In Fig. 3C, an ELISA 10 plate test was used to measure TFIIB-RAP30 binding. Microtiter plate wells were coated with RAP30 or a RAP30 mutant. A control ELISA experiment developed with anti-RAP30 antiserum demonstrated that immobilization of RAP30 mutants was of comparable efficiency (data not shown). After blocking, TFIIB was incubated with RAP30 mutants bound to the wells. After washing, bound TFIIB was detected with anti-TFIIB antiserum, using a horseradish peroxidase-linked second antibody and a colorimetric enzyme assay. ELISA appears to provide a reliable measure of interaction and a somewhat more sensitive detection of RAP30-TFIIB binding than the affinity bead procedure. By ELISA, RAP30, with or without a C-terminal histidine tag, and RAP30- was not observed to bind. Since RAP30-(27-152) binds tightly to TFIIB by this analysis, most TFIIB-RAP30 contacts appear to involve sequence within this region. Since RAP30-(1-118) and -(110 -249) bind TFIIB weakly, amino acids between 27-118 and 110 -176 (or 110 -152) appear to contribute to this interaction. The ELISA binding assay was done in SB 0.1, and the affinity bead assay involved a washing step with SB 0.25, so a difference in salt concentration may account for slightly different but consistent results using these two procedures.
Sequences within the N-terminal region and central region of RAP74 affect masking of the C-terminal domain for RNA polymerase II binding (35) and CTD phosphatase stimulation (34). Since RAP74-(358 -517) appears to bind more tightly to TFIIB than full-length RAP74, or RAP74-(87-517), -(207-517), and -(⌬137-356), the C-terminal domain also appears to be masked for TFIIB binding. RAP74-(207-517) shows the most dramatic masking effect for polymerase binding (35) and phosphatase stimulation (34) just as it does for TFIIB binding (Fig.  3D, lane 3). In summary, the central region of RAP74 (amino acids 207-356) appears to mask the C-terminal region, and the N-terminal region (amino acids 1-205) makes the C-terminal  1. RAP30 and RAP74 deletion mutants. RAP30-(1-249) is the histidine-tagged version of RAP30. Accurate transcription (tx.) was determined from the adenovirus major late promoter (Fig. 2). RAP74 binding to RAP30 was determined using a Ni 2ϩ -affinity bead procedure (data not shown). TFIIB binding to RAP30 and RAP74 was determined by affinity bead and ELISA procedures (Figs. 3 and 4). ϩ, high activity; Ϯ, low activity; ?, barely detectable activity; and -, no detectable activity; n.d., no determination was made for a particular mutant; *, data are not shown in this report. region more accessible for protein-protein interactions.
RAP74 Blocks TFIIB-RAP30 Binding-Since both TFIIB and RAP74 interact independently with overlapping regions of RAP30, RAP74 might stimulate or antagonize TFIIB-RAP30 binding. In Fig. 4A, an ELISA test for interactions between TFIIB, RAP30, and RAP74 was done in which one protein was immobilized in the well of the microtiter plate, another was added to the well as a binding partner, to be detected with FIG. 3. Interactions between TFIIF subunits and TFIIB. A, both RAP30 and RAP74 bind directly and independently to TFIIB. 300 pmol of RAP30 (30) (lanes 1 and 2) or RAP74 (74) (lanes 3 and 4) was incubated with affinity beads containing covalently immobilized TFIIB (IIB) or no protein ligand (CT for control). After washing with SB 0.25, bound protein was eluted and analyzed on an SDS-PAGE gel developed with silver nitrate. 20 -30% of input RAP30 and RAP74 was recovered from the beads. B, RAP30 sequences between amino acids 1-176 are required for tight binding to TFIIB. RAP30-(1-152) binds weakly to TFIIB. The binding assay was done with TFIIB affinity beads and the indicated RAP30 mutants. C, a minimal TFIIB-binding region of RAP30 is localized between amino acids 27-152. RAP30 or RAP30 mutants (500 ng) were immobilized in the wells of a microtiter dish. After blocking, increasing amounts of TFIIB were incubated in the wells. TFIIB binding to RAP30 mutants was detected with anti-TFIIB antiserum. The A 490 nm differs between the two experiments shown in the left and right panels because of the time of color development. RAP30 wild type has no histidine tag. RAP30-(1-249) has a C-terminal histidine tag. D, RAP74 has a masked binding site for TFIIB located within a C-terminal region between amino acids 358 and 517. RAP74 and RAP74 mutants were bound to TFIIB affinity beads. Bound proteins were analyzed in a polyacrylamide gel developed with silver nitrate. None of the RAP74 mutants bound to negative control beads with no bound protein ligand (data not shown). antibody, and the third was added as a potential competitor or facilitator of the binding interaction. Using this protocol, RAP74 was shown to inhibit TFIIB binding to RAP30 (filled circles). This same conclusion was obtained using a Ni 2ϩ -affinity bead procedure in which histidine-tagged RAP30 was used to retain TFIIB in the presence or absence of RAP74 (Fig. 4B). So RAP74 can block the TFIIB-RAP30 interaction, but in a similar test, RAP30 was not observed to block the interaction between TFIIB and RAP74 ( Fig. 4A; (ϫ-ϫ)). This was perhaps the expected result because the N-terminal region of RAP74 binds to RAP30 (35,44), and the C-terminal region of RAP74 binds to TFIIB (Fig. 3D), so both contacts should be maintained simultaneously. TFIIF subunit-subunit interactions are also stable in the presence of TFIIB (Fig. 4, A (open circles) and C).
In the presence of RAP74, therefore, TFIIB does not bind RAP30, but contact between RAP74 and RAP30 is maintained. TFIIB-TFIIF interactions, therefore, are most likely maintained through contact between TFIIB and the C-terminal . RAP30 does not block formation of a TFIIB-RAP74 complex (ϫ-ϫ). TFIIB or RAP74 (500 ng) was immobilized in the wells of a microtiter plate. A mixture of a mobile protein (3.5 pmol) and increasing quantities of a potential binding competitor (or facilitator) were added. Detection of bound protein was with anti-RAP30 or anti-RAP74 antiserum. B, RAP74 blocks formation of the TFIIB-RAP30 complex. A Ni 2ϩ -affinity bead procedure was used. Histidine-tagged RAP30 or RAP30-(1-176) (300 pmol) was bound to TFIIB (600 pmol) in the presence or absence of RAP74 (600 pmol). In the presence of RAP74, TFIIB did not bind to the Ni 2ϩ beads. A Western blot is shown developed with anti-TFIIB antiserum. C, TFIIB does not block formation of a TFIIF complex. Histidine-tagged RAP30 or RAP30-(1-176) retained RAP74 on Ni 2ϩ beads in both the presence and absence of TFIIB. A Western blot is shown developed with anti-RAP74 antiserum. D, RAP74 blocks formation of the TFIIB-RAP30 complex both by binding to TFIIB (left panel) and by binding to RAP30 (right panel). RAP74 or RAP74 mutants (competitor) were tested in the ELISA competition assay as in A. The region of RAP74 that binds TFIIB maps between amino acids 358 and 517 (Fig. 3D) and the region of RAP74 that binds RAP30 maps between amino acids 1 and 172 (35,44). E, calf thymus RNA polymerase II (RNAP II) does not appear to block TFIIB-RAP74 interactions. The ELISA competition assay was done as indicated in the key. region of RAP74. For the TFIIB-RAP30 interaction to be maintained within a transcription complex, TFIIF would have to dissociate into RAP30 and RAP74 subunits.
RAP74 could block TFIIB-RAP30 binding by two mechanisms. 1) RAP74 could bind TFIIB to block the RAP30 binding site on TFIIB. 2) RAP74 could bind RAP30 to block the TFIIB binding site on RAP30. Both mechanisms contribute to blocking the TFIIB-RAP30 interaction (Fig. 4D). RAP74 mutants containing the TFIIB binding region (RAP74 amino acids 358 -517) but missing the RAP30 binding region (RAP74 amino acids 1-172) block TFIIB-RAP30 binding (left panel). There is a close correlation between the RAP74 mutants that bind TFIIB most tightly and the ability of these mutants to compete the TFIIB-RAP30 interaction (compare Figs. 3D and 4D (left panel)). Also, RAP74 mutants containing the RAP30 binding region but missing the TFIIB binding region block TFIIB-RAP30 binding (right panel).
Since the TFIIB-RAP74 interaction takes precedence over the TFIIB-RAP30 interaction, one question that arises is whether the TFIIB-RAP74 interaction is maintained in higher order complexes containing RNA polymerase II. Although this question is difficult to address directly, we have tested the effect of RNA polymerase II on binding of TFIIB and RAP74 (Fig. 4E). Since both TFIIB and RNA polymerase II bind to overlapping regions within the C-terminal domain of RAP74 (amino acids 358 -517 for TFIIB binding versus 363-444 for polymerase binding), these proteins might compete or cooperate for binding. RNA polymerase II does not appear to strongly compete or facilitate the TFIIB-RAP74 interaction. When RAP74 was immobilized and TFIIB was added as the binding partner, addition of RNA polymerase II was not observed to affect formation of complexes containing TFIIB and RAP74 (filled circles). When TFIIB was immobilized and RAP74 added as the binding partner, addition of RNA polymerase II caused a moderate reduction in the retention signal (open circles). Of course, RNA polymerase II is expected to interact with both TFIIB and RAP74, and polymerase might interfere with binding between anti-RAP74 antibody and RAP74. The slight apparent inhibition of binding that is observed, in the experiment in which TFIIB was immobilized (open circles), could be attributable to antibody interference or distribution of TFIIB-polymerase and RAP74-polymerase complexes. It may be difficult to exchange TFIIB and RAP74 bound to separate polymerase molecules into a RNA polymerase II-TFIIB-RAP74 complex, since this requires dissociation of either TFIIB or RAP74 from polymerase. Consistent with this idea, when RNA polymerase II was immobilized and competition for binding was between RAP74 and TFIIB, a smaller reduction in the retention signal was observed. This was true whether detection was for TFIIB (ϫ-ϫ) or RAP74 (filled squares). Of course TFIIB and RAP74 could bind independently to polymerase without maintaining the TFIIB-RAP74 interaction. Additional experiments will be required to demonstrate TFIIB-RAP74 contacts within the preinitiation complex.

DISCUSSION
There is mounting evidence for functional interaction between TFIIF and TFIIB. These factors cooperate to bring RNA polymerase II into the pre-initiation complex (3)(4)(5)(6). Transcriptional start sites are selected by a complex that includes TFIIF, TFIIB, and RNA polymerase II, and genetic experiments implicate each of these factors in this process (27)(28)(29)(30). TFIIB and TFIIF also can cooperate to control dephosphorylation of the CTD (34).
RAP30 appears to consist of three functional regions. The N-terminal region binds RAP74 (43,44), the central region binds RNA polymerase II (45), 2 and the C-terminal region binds DNA (41,42). In the current work, we have mapped the functional domains of human RAP30 that are required for accurate transcription, that bind RAP74, and that bind TFIIB. Tight TFIIB-RAP30 binding requires primary sequence distributed over an extensive region of RAP30 between amino acids 1 and 176 (Fig. 3, B and C).
Analysis of sequence (46,47) and deletion mutants (35) indicates that RAP74 is also divided into three functional regions. The RAP74 primary sequence can be divided into an N-terminal basic region, a highly charged central region with overall negative charge, and a C-terminal basic region. The N-terminal region binds to RAP30 (35,44). The central region appears to be a largely unstructured hinge that controls accessibility of the C-terminal region. The C-terminal region of RAP74 binds directly to TFIIB (Fig. 3D) and to RNA polymerase II (35). Polymerase-RAP74 contact stimulates a CTD phosphatase that must bind to the "body" of polymerase to access the CTD "tail" (34). Deletion from the N terminus of RAP74 initially increases masking of TFIIB binding (Fig. 3D), RNA polymerase II binding (35), and stimulation of CTD phosphatase activity (34). Further deletion from the N terminus and within the central region unmasks TFIIB binding, polymerase binding, and stimulation of CTD phosphatase activity. The central region is also the site of phosphorylation by casein kinase II, and phosphorylation by this and/or other kinases appears to stimulate polymerase binding by TFIIF (48).
Although TFIIB can bind directly to RAP30, RAP74 blocks TFIIB-RAP30 binding through two mechanisms. By binding the N-terminal region of RAP30, the N-terminal region of RAP74 blocks TFIIB-RAP30 binding (Fig. 4D, right panel). Also, by binding TFIIB, the C-terminal region of RAP74 blocks TFIIB-RAP30 binding (Fig. 4D, left panel). The structure of the TATA box-TBP-TFIIB complex strongly indicates that interaction between TFIIB and TFIIF might be through the N-terminal region of TFIIB (17). TFIIB-TFIIF contact, therefore, is most likely maintained through the N-terminal region of TFIIB and the C-terminal region of RAP74.
The strongest contacts between TFIIB and RAP74 map between RAP74 amino acids 358 -517 (Fig. 3D). Although RAP74 blocks TFIIB-RAP30 binding, RAP74 can bind to RAP30 and TFIIB simultaneously (Fig. 4A). This result is consistent with the mapping of RAP74 functional domains, since the N-terminal region binds RAP30 and the C-terminal region binds TFIIB. Although RNA polymerase II and TFIIB bind to overlapping regions within the C-terminal domain of RAP74, RNA polymerase II does not appear to strongly compete TFIIB-RAP74 binding (Fig. 4E). That TFIIB and RAP74 might interact in complexes containing RNA polymerase II is also indicated by the genetic interaction between yeast SUA7 and SSU71 (30) and by TFIIB-mediated suppression of RAP74 stimulation of CTD phosphatase activity (34). Interaction between RAP74 and TFIIB, therefore, may be maintained in complexes with RNA polymerase II.
The SUA7 gene in yeast encodes TFIIB, and the RAP74 homologue in yeast is encoded by SSU71/TFG1. sua7 mutants that confer cold sensitivity and altered transcriptional start site selection, and are suppressed by ssu71 mutants, map to a particular region of TFIIB near the N terminus (11). From our work, we would have predicted that the ssu71 mutations that suppress sua7 might map within the C-terminal region of SSU71 where human TFIIB and RAP74 physically interact. The ssu71 G363D and G363R alleles, 3 however, correspond to the Lys-111 position of human RAP74 (30,35), within the RAP30 binding region. So both TFIIF subunits may cooperate with TFIIB in the process of start site selection.
Although the experiments shown here demonstrate specific protein-protein contacts between general transcription factors TFIIB and TFIIF, the importance of RAP74 blocking TFIIB-RAP30 interaction is not clear. TFIIB is expected to interact with TFIIF through the RAP74 subunit, although in the absence of RAP74, a TFIIB-RAP30 complex could be maintained. Such alternate structures are most easily understood if TFIIF does not invariably function as an intact unit, but rather, under some circumstances, as two independent factors, RAP30 and RAP74. Chang et al. (23) demonstrated that accurate initiation could be reconstituted in a TFIIF-depleted extract by addition of RAP30 alone. In those experiments, elongation to the run-off position required further addition of RAP74. Thus, RAP30 and RAP74 functions appear to be partially separable in initiation and elongation, at least in an extract transcription system. TFIIB-RAP30 interactions, therefore, might be maintained within some promoter-bound complexes to support initiation but not elongation. Addition of RAP74 would displace the TFIIB-RAP30 interaction, and elongation would begin. Regulated interaction between TFIIB, RAP30, and RAP74, therefore, might allow for distinct pathways of pre-initiation complex assembly on different promoters and for separate timing of RAP30 and RAP74 function in initiation and elongation.