An Activator Target in the RNA Polymerase II Holoenzyme

Molecular Cell, Vol. 1, 895–904, May, 1998, Copyright 1998 by Cell Press

An Activator Target in the RNA Polymerase II Holoenzyme Sang Seok Koh,* Aseem Z. Ansari,† Mark Ptashne, † and Richard A. Young*‡ * Whitehead Institute for Biomedical Research Nine Cambridge Center Cambridge, Massachusetts 02142 Department of Biology Massachusetts Institute of Technology Cambridge, Massachusetts 02139 † Department of Molecular and Cellular Biology Harvard University Cambridge, Massachusetts 02138

Summary Expression of protein-coding genes in eukaryotes involves the recruitment, by transcriptional activator proteins, of a transcription initiation apparatus consisting of greater than 50 polypeptides. Recent genetic and biochemical evidence in yeast suggests that a subset of these proteins, called SRB proteins, are likely targets for transcriptional activators. We demonstrate here, through affinity chromatography, photo-crosslinking, and surface plasmon resonance experiments, that the GAL4 activator interacts directly with the SRB4 subunit of the RNA polymerase II holoenzyme. The GAL4 activation domain binds to two essential segments of SRB4. The physiological relevance of this interaction is confirmed by mutations in SRB4, which occur within its GAL4-binding domain and which restore activation in vivo by a GAL4 derivative bearing a mutant activation domain. Introduction Transcriptional activators bind to specific DNA sequences and through their activation domains contact the transcriptional machinery. To understand the mechanism of transcriptional activation, it is important to identify the targets of activators in the transcriptional machinery as well as the nature of the interactions that lead to gene activation. According to the recruitment model, the activator–target interaction recruits the transcriptional machinery to the promoter and thereby activates gene expression (Ptashne and Gann, 1997). Interactions between the activator and the transcriptional machinery may also stimulate transcriptional events subsequent to the binding of the machinery to the promoter (Ansari et al., 1995; Triezenberg, 1995; Orphanides et al., 1996; Pugh, 1996; Ranish and Hahn, 1996; Hoffmann et al., 1997). In prokaryotes, specific interactions between activation domains and the transcriptional machinery have been identified using a combination of genetic and biochemical methods (Ptashne and Gann, 1997). For example, a specific interaction between the activation domain of CAP (catabolite gene activator protein) and the a subunit of Escherichia coli RNA polymerase at the lac ‡ To whom correspondence should be addressed.

promoter was identified by: (1) truncations and point mutations in the a subunit that do not respond to CAPmediated activation (Igarashi and Ishihama, 1991; Zou et al., 1992; Tang et al., 1994), (2) positive control (PC) mutants of CAP that do not alter CAP’s ability to bind DNA but that impair its ability to activate transcription (Zhou et al., 1993), (3) DNA footprinting and structural studies, which are consistent with the steric proximity of the regions of a and CAP that are genetically implicated in activator–target interactions (Ren et al., 1988; Kolb et al., 1993), and (4) label transfer photo-crosslinking experiments, which demonstrate that the activation domain of CAP is within 15 A˚ of the a subunit when both proteins are bound to the lac promoter (Chen et al., 1994). Similar experiments at the gal promoter show that a different activation domain of CAP can interact with a different surface of bacterial RNA polymerase to mediate gene expression (Kolb et al., 1993; Kumar et al., 1994; Niu et al., 1996). Studies with the l cI repressor indicate that it activates by contacting the s70 subunit of the RNA polymerase (Li et al., 1994). Interactions between these activators and the polymerase are evidently weak and have not been measured in the absence of DNA. Therefore, both genetic and biochemical lines of evidence must converge for a physiologically relevant contact to be inferred. In eukaryotes, the problem of identifying activator– target interactions is compounded by the fact that the transcriptional machinery consists of more than 50 proteins. Moreover, from several studies it appears that a given activator can contact more than one target in the machinery. For example, the yeast activator GAL4 interacts with TBP and TFIIB, two essential components of the eukaryotic transcriptional machinery, and there is a strong correlation between GAL4’s affinities for TBP and TFIIB measured in vitro and the level of activation measured in vivo for a series of mutations in the activation domain of GAL4 (Melcher and Johnston, 1995; Wu et al., 1996). The activation domain of the herpes simplex virus activator VP16 also binds to TBP, TFIIB, certain TAFs, and other components of the transcriptional machinery (Stringer et al., 1990; Ingles et al., 1991; Lin et al., 1991; Goodrich et al., 1993; Roberts et al., 1993; Xiao et al., 1994; Zhu et al., 1994; Kobayashi et al., 1995). The yeast transcription initiation machinery includes TBP, TFIIA, TFIIB, TFIIE, TFIIF, TFIIH, RNA polymerase II with its 12 subunits, and an SRB/mediator subcomplex. A complex containing all of these components except TBP and TFIIE has been purified from yeast. This complex, called the RNA polymerase II holoenzyme, supplemented with TFIIE and TBP, is capable of responding to activators in a transcription assay performed in vitro (Koleske and Young, 1994). Thus, there may be a few complexes, possibly just one, that upon recruitment to a promoter lead to high levels of gene expression (Koleske and Young, 1995; Stargell and Struhl, 1996; Ptashne and Gann, 1997). Three observations are consistent with the idea that the SRB proteins in the yeast holoenzyme are targets of transcriptional activators. First, the SRB proteins can be purified as

Molecular Cell 896

Figure 1. RNA Polymerase II CTD and SRB Proteins Influence Galactose Activation RNA polymerase II CTD truncation mutants containing wild-type SRB genes (SRB WT) or dominant suppressing SRB alleles (SRB D) were assayed for their ability to grow on SC medium containing galactose as the sole carbon source. The length of the RNA polymerase II CTD in each strain is diagrammed on the right.

a complex separate from RNA polymerase II, and this complex, when added back to the purified core RNA polymerase II and the general transcriptional factors, confers upon them the ability to respond to transcriptional activators in vitro (Kim et al., 1994; Hengartner et al., 1995). Second, the SRB complex binds directly to the activation domain of VP16, a potent viral activator (Hengartner et al., 1995). Third, the SRB complex interacts with the carboxy-terminal domain (CTD) of RNA polymerase II, and the CTD has been implicated in the response to activators in vivo (Scafe et al., 1990; Liao et al., 1991; Gerber et al., 1995). Gain-of-function mutations in SRB2, SRB4, SRB5, and SRB6 make them attractive candidates for activator targets in the RNA polymerase II holoenzyme. We show here, using a combination of biochemical and genetic methods, that the yeast transcriptional activator GAL4 interacts with the SRB4 subunit of the RNA polymerase II holoenzyme. Results Gene Expression Is Dependent upon the SRB Proteins Progressive truncation of the CTD of RNA polymerase II causes a corresponding loss in viability of yeast cells grown in glucose media (Nonet et al., 1987). As shown in Figure 1, truncation of the CTD also leads to a loss of viability of yeast cells that utilize galactose as their sole carbon source. We previously showed that RNA polymerase mutants bearing only 11 intact CTD repeats are cold sensitive for growth on glucose, and that suppressing mutations in SRB2, SRB4, SRB5, and SRB6 can restore viability at that temperature (Nonet and Young, 1989; Koleske et al., 1992; Thompson et al., 1993). These SRB mutations can also suppress the inviability of certain CTD truncations when cells are grown on galactose (Figure 1). Because galactose utilization requires the activation of genes under the control of the GAL4 activator, these results raise the possibility that the SRB2, SRB4, SRB5, and SRB6 proteins play a role in GAL4 responsive activation in vivo.

SRB2, SRB4, SRB5, and SRB6 Form a Subcomplex Recombinant SRB2, SRB4, SRB5, and SRB6 proteins were produced using a baculovirus expression system. Interactions between pairs of SRB proteins were studied by incubating an extract containing an influenza hemagglutinin (HA) epitope-tagged SRB protein with an extract containing an equimolar amount of a second, untagged SRB protein. The epitope-tagged SRB was immunopurified, and the eluate was examined by Western blot analysis probing for the untagged protein (Figures 2A–2F). Pairs of SRB proteins lacking HA epitopes were used in parallel reactions to control for specific immunopurification, and b-galactosidase was added to each reaction to control for nonspecific aggregation. The results identified interactions between SRB2 and SRB5 (Figure 2A) and between SRB4 and SRB6 (Figure 2B), consistent with genetic evidence linking the functions of SRB2 and SRB5 and the functions of SRB4 and SRB6 (Thompson et al., 1993; Thompson and Young, 1995). The results also indicated that SRB2 and SRB4 interact (Figure 2C). In contrast, there were no detectable interactions between SRB2 and SRB6 (Figure 2D), SRB4 and SRB5 (Figure 2E), or between SRB5 and SRB6 (Figure 2F). Figure 2G shows that the four SRBs can form a single complex in vitro. Extracts containing all four SRB proteins were incubated and immunopurified using antibodies against epitope-tagged SRB4 or SRB6. In both cases, all four SRB proteins were coimmunopurified (Figure 2G, lanes 2 and 3). The interaction studies suggest a model for an SRB subcomplex containing all four proteins as shown in Figure 2H. GAL4 Binds to the SRB Subcomplex The activation domain of the yeast activator GAL4, residues 841–875, was expressed fused to glutathione S-transferase (GST) and was incubated with the SRB2, 4, 5, and 6 subcomplex. The GST fusion was affinity purified using glutathione-agarose beads and probed for the presence of SRB proteins by Western blot analysis (Figure 3A). The results show that the purified SRB2, 4, 5, 6 subcomplex bound to the activation domain (GAL4wt) excised from GAL4. The same GAL4 activation domain with two point mutations (GAL4mut, Val to Glu at the residue 864 and Leu to Val at the residue 868), which is impaired for activation in vivo (Melcher and Johnston, 1995), failed to interact with the SRB subcomplex. GAL4 Binds to SRB4 Three independent approaches were used to identify the component of the SRB subcomplex that interacts with the GAL4 activation domain. Affinity chromatography, label transfer affinity–photo-cross-linking, and surface plasmon resonance all demonstrate that the SRB4 subunit is bound by GAL4. Affinity Chromatography GST-GAL4, GST-GAL4 containing activation domain mutations, and GST control protein were loaded separately on glutathione-agarose columns, and each of the four purified SRB proteins were passed over each of the three columns (Figure 3B). Only SRB4 was specifically

Activator Target in RNA Polymerase II Holoenzyme 897

Figure 2. SRB2, SRB4, SRB5, and SRB6 Form a Subcomplex (A–F) Pairwise Interactions of SRB2, SRB4, SRB5, and SRB6 proteins. An insect cell extract containing one recombinant SRB protein was incubated with another extract containing equimolar amounts of a second recombinant SRB, which lacked (lanes 1 and 3) or contained (lanes 2 and 4) an HA-epitope tag. Insect cell extract containing b-galactosidase protein was added to each reaction to serve as a control for specific immunoaffinity. Agarose beads with immobilized 12CA5 anti-HA antibody were added to the reactions, unbound material was removed, and the bound proteins were eluted with HA peptide. Fractions (1/300) of the input (IN; lanes 1 and 2) and 1/40 of the eluate (OUT; lanes 3 and 4) were subjected to SDS-PAGE and analyzed by Western blotting using specific antibodies. The pairs tested were SRB2 and SRB5 (A), SRB4 and SRB6 (B), SRB2 and SRB4 (C), SRB2 and SRB6 (D), SRB4 and SRB5 (E), and SRB5 and SRB6 (F). A schematic interpretation of the binary interactions is presented at the top of each panel. (G) SRB2, 4, 5, 6 subcomplex. Insect cell extracts containing all four SRB proteins were incubated together and immunoaffinity chromatography was performed as in (A)–(F). In the control reaction containing all four SRB proteins (2141516; lane 1), no tagged SRB was included. In the other reactions, either SRB4-HA (214HA1516, lane 2) or SRB6-HA (2141516HA, lane 3) was included. Eluates were subjected to SDS-PAGE and analyzed by Western blotting using specific antibodies. (H) A model depicting interactions between SRB2, SRB4, SRB5, and SRB6.

retained on the column bearing the wild-type GAL4 activation domain. None of the SRB proteins were retained on the column with the mutant activation domain or the GST control column. These experiments show that GAL4 interacts directly with SRB4. Label Transfer Affinity Photo-Cross-Linking To confirm that GAL4 interacts with SRB4, we performed a photo-cross-linking experiment with the SRB subcomplex and the RNA polymerase II holoenzyme. A radioiodinated form of the cross-linker APDP (Chen et al., 1994; Weissman et al., 1994) was attached to a GAL4 derivative comprising the DNA-binding domain (residues 1–100) fused to an activation domain (residues 840–881) (Figure 4). This bifunctional cross-linker forms a disulfide bond with a Cys residue in the dark, and when activated by ultraviolet (UV) light, the nitrene reactive group at the distal end of the cross-linker will react with any molecule in its proximity. The cross-linker was reacted in the dark with a GAL4 derivative bearing a Phe to Cys substitution at the residue 856 in its activation domain. The derivatized GAL4 was altered neither in its ability to bind DNA specifically, nor in its ability to activate transcription (A. Z. A. and M. P., unpublished data). The chemically modified GAL4 was incubated with the SRB subcomplex, a cross-link was induced by UV irradiation, and

the label transfer reaction was completed by reducing the disulfide linkage. The label was transferred to SRB4 (Figure 4, lane 2). The experiment does not eliminate the possibility that label was transferred to SRB2 or SRB6 because these proteins are similar in size to the GAL4 derivative, and the background signal produced by self-labeling of GAL4 obscures the signal in the corresponding size range, but no label was transferred to SRB5. The addition of GAL80, a GAL4 inhibitor that binds to the GAL4 activation domain (Ma and Ptashne, 1987), reduced the amount of label transfer to SRB4 (Figure 4, lane 3) presumably by competing with SRB4 for binding to the GAL4 activation domain. The label transfer to SRB4 was also detected in a reaction with RNA polymerase II holoenzyme (Figure 4, lane 4). This photo-crosslinking result confirms that GAL4 interacts with the SRB subcomplex by binding to SRB4. Surface Plasmon Resonance (SPR) The SPR sensor system can detect interactions between molecules immobilized on a sensor chip and molecules in the solution phase flowing over the chip (Chaiken et al., 1992). A DNA fragment bearing two GAL4 binding sites was immobilized in one channel on the surface of the sensor chip (as in Wu et al., 1996). GAL4(1–1001840– 881) and, in a separate experiment, GAL4(1–100) was

Molecular Cell 898

Figure 3. The SRB Subcomplex Interacts with GAL4 Activation Domain through SRB4 Subunit (A) The SRB subcomplex was incubated with GST fusions of the wild-type (GAL4wt) or mutant (GAL4mut) GAL4 activation domain (AD). A GST control (GST) was included. The GST fusions and bound proteins were isolated using glutathione-agarose beads. Fractions (1/10) of the load and all of the pellets were subjected to SDS-PAGE and analyzed by Western blotting using specific antibodies. (B) SRB4 directly interacts with GAL4. Recombinant SRB2, SRB4, SRB5, or SRB6 proteins were loaded independently onto GSTGAL4wt (lanes 1–4), GST-GAL4mut (lanes 5–8), and GST (lanes 9–12) glutathione-agarose columns. After washing, the columns were eluted with glutathione. Fractions (1/100) of load (L) and the flowthrough (F), and 1/10 of the last wash (W) and the eluate (E) were subjected to SDS-PAGE and analyzed by Western blotting.

bound to the DNA fragment (first injection in Figure 5). The interactions of the GAL4–DNA complex with each of the SRB proteins were studied by flowing sample proteins over the GAL4–DNA bound chips (second injection in Figure 5) and measuring net changes in resonance response units (DRU). In a parallel channel, the same proteins were passed over a surface that contained nonspecific DNA attached to the matrix to measure the nonspecific binding effects and the effects of protein buffers on the RU (dotted lines in each sensogram). The results revealed that SRB4 (Figure 5A) but not SRB6 (Figure 5B) bound to the GAL4–DNA complex containing the activation domain. In agreement with previous results (Wu et al., 1996), neither SRB2 nor SRB5 bound to the GAL4 activation domain (data not shown). None of the four SRB proteins bound to the GAL4 derivative lacking an activation domain (Figure 5, right) nor showed detectable affinity for DNA (data not shown). Therefore, we infer that the increase in RUs in the presence of a functional activation domain in the sensogram in Figure 5A is due to the interaction of the GAL4 activation domain with SRB4. Technical limitations prevent an accurate measurement of the off-rate and consequently the quantitation of the dissociation constant defining the interactions between the two proteins. However, the estimated dissociation constant for GAL4–TBP interactions is 1027 M, and based on the observation that the on-rate for GAL4 binding to SRB4 is slower than that for TBP, we estimate that SRB4 binds the GAL4 activation domain with lower affinity than that for TBP. The experiment of Figure 5D shows that introduction of a mutation (Phe to Ala at the residue 869) into the GAL4 activation domain, a change that reduces activation in vivo and interaction with TBP in vitro (Wu et al., 1996), abolished binding to SRB4 even when tested at SRB4 concentrations 20-fold above that required to saturate binding to the wild-type activation domain. Figure 5D also shows that the mutant GAL4 bound GAL80 as efficiently as did wild type, indicating that the activator was not proteolyzed or otherwise damaged during isolation.

SRB4 Gain-of-Function Mutants Compensate for a GAL4 Activation Mutant Fourteen dominant, gain-of-function SRB4 mutations were previously identified as suppressors of the conditional lethal phenotype of a CTD truncation mutant (Thompson et al., 1993). We investigated whether any of these SRB4 mutants could compensate for an activation-defective GAL4 mutant in an otherwise wild-type cell. The 14 dominant alleles of the SRB4 gene were introduced (on plasmids) independently into yeast cells containing either the wild-type or the mutant (F869A) GAL4 derivatives. To monitor galactose induction, these cells also contained a b-galactosidase reporter gene with five upstream GAL4 binding sites. Two of the mutants tested, SRB4-18 and SRB4-20, partially restored galactose activation in the presence of the mutant GAL4 activator (data not shown). An identical experiment was carried out in cells in which the chromosomal SRB4 gene was replaced with the SRB4-18 and SRB4-20 alleles, with the result that these two alleles substantially restored galactose activation in the presence of the mutant GAL4 activator (Figure 6A). SPR analysis revealed that recombinant SRB4-18 and SRB4-20 proteins bound to the mutant GAL4 activation domain significantly better than did wild-type SRB4 (Figure 6B). Neither SRB4 mutant bound detectably to naked DNA. In no case was the binding of the mutant SRB4 to the mutant activator as robust as the binding of their wild-type counterparts with each other. These results demonstrate a correlation between GAL4-SRB4 binding in vitro and the ability of GAL4 to activate transcription in vivo. GAL4 Binding Sites Are Essential Segments of SRB4 The SRB4 gene is essential for yeast cell viability (Thompson et al., 1993). To obtain clues to the regions of the SRB4 protein that are critical for its function, a series of internal deletion mutations, each affecting 50 amino acids, was constructed in the SRB4 gene and used to replace a wild-type SRB4 gene in yeast cells

Activator Target in RNA Polymerase II Holoenzyme 899

Figure 4. Label-Transfer Affinity Photo-Cross-Linking Analysis Shows GAL4-SRB4 Interaction A radiolabeled cross-linking agent, 125I-APDP (N-[4-(p-azidosalicylamido)butyl]-39[29-pyridyldithio]propionamide), was coupled to the activation domain of a GAL4 derivative by a disulfide bond: the radiolabeled cross-linker is depicted as a star; DNA-binding domain (1–100) and activation domain (840–881) of GAL4 are depicted as a circle and a wave, respectively. This labeled GAL4 derivative was incubated with the SRB subcomplex. Cross-linking was induced by irradiating the samples with UV light. The reaction was terminated by addition of DTT, which reduces both the reactive azide and the disulfide bond between GAL4 and 125I-APDP. The 125I-APDP is transfered and covalently attached to protein bound to the GAL4 activation domain. Proteins were subjected to SDS-PAGE and labeled proteins were visualized with a phosphorimager. Reactions were GAL4 alone control (lane 1), GAL4 plus the SRB subcomplex (lane 2), GAL4 plus the SRB subcomplex with GAL80 (lane 3), and GAL4 plus RNA polymerase II holoenzyme (lane 4). A control experiment with a mutant GAL4 did not show any cross-linking to SRB4.

by plasmid shuffle. The series of SRB4 internal deletion mutations was tested for viability on glucose media at 368C (Figure 7A). The deletion analysis indicates that two segments of SRB4, one consisting of residues 302–351, and another of residues 452–687, are essential for cell viability. Independently, the ability of GAL4 to bind to various regions of SRB4 was studied in vitro. Five overlapping SRB4 protein fragments, each FLAG epitope-tagged at their N termini, were produced and purified. Each recombinant SRB4 fragment was incubated with GST fusion proteins consisting of the wild-type or mutant GAL4 activation domain. Fragments SRB4C and SRB4E bound

Figure 5. SRB4 Specifically Interacts with DNA-Bound GAL4 Activator The SRB4–GAL4 interaction was detected by surface plasmon resonance analysis. In each sensogram, the first change in resonance units (DRU) is due to binding of a GAL4 derivative (depicted as a circle, DNA-binding domain; and a wave, activation domain) to the immobilized DNA sites (solid line) on the surface of the chip (closed rectangle). The net change in RU (DRU) after injection of a sample protein (oval shape) is displayed in the sensogram and attributed to the binding of the protein to the immobilized GAL4 derivative. The time points at which resonance signals were measured are indicated (X). The solid trace shows the binding of proteins in a channel containing the cognate GAL4 DNA sites, and the dotted trace shows the profile of the proteins flowing in a channel containing nonspecific DNA of the same length and immobilized at the same density as the cognate DNA. Total specific binding is presented as DRU after baseline corrections. (A–C) Sensograms on the left and right panels show interactions of sample proteins with GAL4(1–1001840–881) and GAL4(1–100), respectively. Initially, 230–266 RUs of GAL4 derivatives were bound to the cognate DNA on the sensor chips. Interactions between a given GAL4 derivative and the sample proteins were measured by injecting SRB4 (A), SRB6 (B), or TBP (C) into the GAL4-immobilized sensor chips. (D) SRB4 does not bind to an activation defective GAL4 activation domain. RUs (163) of a mutant GAL4 derivative, GAL4(1–1001840– 881 F869A), were first bound to the immobilized DNA of sensor chips. Sequential injections of SRB4 (oval shape) and GAL80 (shaded oval shape) were performed to test binding of the proteins to the GAL4 derivative.

Molecular Cell 900

Figure 6. SRB4 Gain-of-Function Mutants Compensate for a GAL4 Activation Mutant In Vivo (A) Either wild-type (wt) (1–1001840–881) or activation-defective mutant (mut) (1–1001840–881 F869A) GAL4 derivatives were transformed into yeast cells containing wild-type (wt) or gain-of-function SRB4 (4-18 and 4-20) alleles. The yeast cells harbor a chromosomally integrated b-galactosidase gene with upstream DNA containing five consensus 17-mer GAL4 binding sites and GAL1 TATA. The b-galactosidase activity produced by cells with mutant GAL4 is presented as percentage of that produced by isogenic wild-type GAL4 cells. (B) Interactions between various SRB4 and GAL4 recombinant proteins in vitro. SPR analysis was performed as in Figure 5A with purified recombinant SRB4, SRB4-18, and SRB4-20. The DRU values for the various SRB4 proteins are presented.

specifically to the wild-type GAL4 activation domain (Figure 7B). Fragments SRB4A and SRB4B did not bind to the GAL4 activation domain, whereas the SRB4D polypeptide appeared to bind weakly. Thus, the GAL4 activation domain appears to have two relatively strong binding sites in the SRB4 polypeptide, one in the region of residues 331–360 and another in the region of residues 561–687. It is striking that these are the regions that are essential for cell viability (Figure 7A). If the GAL4 activation domain interacts with these two essential segments of SRB4 protein in vivo, then the SRB4 mutations that restore activation in vivo by a GAL4 activation domain mutant (Figure 6) should affect amino acid residues in the vicinity of these two segments. Sequence analysis of SRB4-18 and SRB4-20 alleles revealed that this is indeed the case (Figure 7C). Thus, the segments of SRB4 that are bound by the GAL4 activation domain in vitro contain the suppressing mutations that restore activation by a GAL4 activation domain mutant in vivo.

Figure 7. Essential Segments of SRB4 Are GAL4 Binding Sites and Harbor Suppressing Mutations (A) Regions of SRB4 required for cell viability. A series of internal deletion mutations, each affecting 50 amino acids, was constructed in the SRB4 gene and used to replace a wild-type SRB4 gene in yeast cells by plasmid shuffle. The series of SRB4 internal deletion mutations was tested for viability on glucose media at 368C. (B) Regions of SRB4 that bind to the GAL4 activation domain. Five overlapping SRB4 protein fragments, each FLAG epitope-tagged at their N termini, were produced in a baculovirus system and purified. Each recombinant SRB4 fragment was incubated with GST fusion proteins consisting of the wild-type or mutant GAL4 activation domain. The GST fusions and bound proteins were isolated using glutathione-agarose beads. Fractions (1/10) of the load and all of the pellets were subjected to SDS-PAGE and analyzed by Western blotting using anti-FLAG antibody. (C) Point mutations in SRB4-18 and SRB4-20 proteins.

Discussion The multiple lines of genetic and biochemical evidence that have allowed investigators of prokaryotic transcription activation to identify physiologically relevant activator targets are not as well established for eukaryotic activators. We have attempted to provide the combination of genetic and biochemical evidence necessary to identify a physiologically relevant target for a yeast transcriptional activator. Genetic studies have implicated

Activator Target in RNA Polymerase II Holoenzyme 901

the CTD, and the associated SRB/mediator complex, in the process of transcription activation (Allison and Ingles, 1989; Nonet and Young, 1989; Scafe et al., 1990; Liao et al., 1991; Gerber et al., 1995; Myers et al., 1998; reviewed in Koleske and Young, 1995). Biochemical studies have demonstrated that reconstitution of activated transcription by RNA polymerase II in a yeast system requires the SRB/mediator complex (Kim et al., 1994; Hengartner et al., 1995). RNA polymerase II holoenzymes containing the SRB/mediator complex are responsive to activators in vitro, whereas RNA polymerase II and general transcription factors alone are not (Kim et al., 1994; Koleske and Young, 1994; Hengartner et al., 1995). This genetic and biochemical evidence makes the SRB/mediator complex an attractive candidate for a target of transcriptional activators, but prior to this study a specific target protein had not been identified. We provide here five lines of evidence for the identification of SRB4 as a specific target of the activation domain of GAL4. First, affinity chromatography experiments with individual SRB proteins show that GAL4 binds to SRB4 only when the activator contains a functional activation domain. Second, photo-cross-linking studies confirm that SRB4, when part of the holoenzyme or an SRB subcomplex, becomes closely associated with the activation domain of GAL4. Third, SPR experiments demonstrate that the activation domain of GAL4 binds to SRB4, but not to the other three SRBs, even at submillimolar concentrations. Fourth, mutations have been identified in SRB4 that restore transcription elicited by a weakened GAL4 activator in vivo; unlike the wildtype SRB4 protein, SRB proteins containing the compensatory mutations bind to the weakened GAL4 activation domain in vitro. And finally, we have found that the GAL4 activation domain interacts with the portions of SRB4 in which the compensatory mutations occur. SRB4 joins a substantial list of proposed targets for activators. The most thoroughly studied activator, VP16, has been shown to bind to TBP, TAF IIs, TFIIB, TFIIF, TFIIH, and TFIIA (Stringer et al., 1990; Ingles et al., 1991; Lin et al., 1991; Goodrich et al., 1993; Roberts et al., 1993; Xiao et al., 1994; Zhu et al., 1994; Kobayashi et al., 1995). While it is possible that VP16 does interact with all these factors during activation, it is not yet clear that the binding observed with the isolated general transcription factors also occurs when these factors are components of a holoenzyme. Furthermore, the ability to bind to most of these factors is not sufficient for activation, since transcription systems reconstituted with highly purified TBP, TFIIB, TFIIF, TFIIH, TFIIE, TFIIA, and Pol II are not activated by VP16 (reviewed in Orphanides et al., 1996; Hoffmann et al., 1997). The addition of TAFIIs to some purified systems can reconsitute activation (reviewed in Burley and Roeder, 1996; Verrijzer and Tjian, 1996), but genetic studies reveal that TAFIIs are not essential for expression of all genes in yeast and mammalian cells (Wang and Tjian, 1994; Apone et al., 1996; Moqtaderi et al., 1996; Walker et al., 1996). Further characterization of mammalian SRB/ mediator-containing complexes (Ossipow et al., 1995; Chao et al., 1996; Maldonado et al., 1996; Pan et al., 1997; Scully et al., 1997) and other complexes that confer activation (Burley and Roeder, 1996; Verrijzer and

Tjian, 1996) should improve our understanding of gene activation in higher eukaryotes.

GAL4 Interacts with Multiple Components of the Transcriptional Machinery Several investigators have proposed that TBP and the RNA polymerase II holoenzyme can be recruited independently to promoters, in part because holoenzymes purified thus far from yeast lack TBP (Koleske and Young, 1995; Stargell and Struhl, 1996; Struhl, 1996). The ability of the GAL4 activation domain to interact with TBP (Melcher and Johnston, 1995; Wu et al., 1996) and SRB4 therefore suggests how GAL4 may recruit the entire transcription apparatus to promoters. GAL4 could act by recruiting TBP and then the holoenzyme to a promoter. However, at the GAL1–10 promoter, GAL4 has multiple binding sites, and multiple GAL4 proteins associated with this promoter could bind to TBP and the holoenzyme independently and without a requirement for a particular order of assembly. It seems likely that multiple activator–target interactions play a synergistic role in eliciting high levels of gene expression in vivo. In prokaryotes, catabolite activator protein (CAP) interacting with two different components of E. coli RNA polymerase elicits a much higher level of activation than either single interaction does alone (Busby et al., 1994). Furthermore, multiple interactions by a single activator may affect different steps in activation. Bacterial CAP, for example, interacts with two different regions of the a subunit of E. coli RNA polymerase, and each interaction is responsible for a specific mechanistic consequence (e.g., formation of closed complex and isomerization of closed complex to open complex) (Niu et al., 1996).

SRB/Mediator Complex Is a Likely Target for Multiple Transcriptional Regulators Although the precise subunit composition of the SRB/ mediator complex is not yet resolved, the components that have been identified thus far indicate that this complex is involved in many important regulatory processes in yeast cells. In addition to the nine SRB proteins, GAL11, SIN4, RGR1, ROX3, and seven MED proteins have been identified as components of this complex and of the holoenzyme (Kim et al., 1994; Koleske and Young, 1994; Barberis et al., 1995; Hengartner et al., 1995; Li et al., 1995; Liao et al., 1995; Gustafsson et al., 1997; Lee et al., 1997; Myers et al., 1998). Genes required for glucose regulation (Kuchin et al., 1995; Song et al., 1996), repression of early meiotic-specific genes during mitotic growth (Strich et al., 1989), and repression of a-specific genes (Wahi and Johnson, 1995) have been identified as SRB genes. GAL11 is required for full activity at many promoters, including those encoding enzymes metabolizing galactose (Suzuki et al., 1988). SIN4, RGR1, and ROX3 have been implicated in both positive and negative transcriptional regulation at a variety of genes (Jiang and Stillman, 1992). All these observations suggest that the SRB/mediator complex is a target of many transcriptional regulators. Precise identification of the targets of such regulators in the SRB/mediator

Molecular Cell 902

Table 1. Yeast Strains Strain

Alias

Genotype

Z551 Z552 Z553 Z554 Z555 Z556 Z557 Z572 Z777 Z812 Z813 Z814 Z815

N400 CTY3 CTY8 CTY9 CTY15 CTY20 CTY21 CTY182 SLY84 SKY239 SKY243 S211 JPY9::RJR227

Z859 Z860 Z861

SKY259 SKY260 SKY261

Mata ura3-52 his3D200 leu2-3,112 rpb1D187::HIS3 [RY2204 (rpb1D104 LEU2)] Mata ura3-52 his3D200 leu2-3,112 rpb1D187::HIS3 SRB4-1 [RY2204 (rpb1D104 LEU2)] Mata ura3-52 his3D200 leu2-3,112 rpb1D187::HIS3 SRB5-1 [RY2204 (rpb1D104 LEU2)] Mata ura3-52 his3D200 leu2-3,112 rpb1D187::HIS3 SRB6-1 [RY2204 (rpb1D104 LEU2)] Mata ura3-52 his3D200 leu2-3,112 rpb1D187::HIS3 SRB4-1 [RY2112 (RPB1 URA3)] Mata ura3-52 his3D200 leu2-3,112 rpb1D187::HIS3 SRB5-1 [RY2112 (RPB1 URA3)] Mata ura3-52 his3D200 leu2-3,112 rpb1D187::HIS3 SRB6-1 [RY2112 (RPB1 URA3)] Mata ura3-52 his3D200 leu2-3,112 srb4D2::HIS3 [RY2825 (SRB4 URA3)] Mata ura3-52 his3D200 leu2-3,112 rpb1D187::HIS3 [RY2112 (RPB1 URA3)] Mata ura3-52 his3D200 leu2-3,112 rpb1D187::HIS3 SRB2-2 [RY2112 (RPB1 URA3)] Mata ura3-52 his3D200 leu2-3,112 rpb1D187::HIS3 [RY2128 (RPB1 LEU2)] Mata ura3-52 his3D200 leu2-3,112 rpb1D187::HIS3 SRB2-2 [RY2204 (rpb1D104 LEU2)] Mata his3D200 leu2D1 trp1D63 lys2D385 gal4D11, chromosomally integrated b-galactosidase gene with 5317mer GAL4 binding sites–GAL1 TATA Z815 with srb4D3::TRP1 [RY2832 (SRB4 LEU2)] Z815 with srb4D3::TRP1 [pSK344 (SRB4-18 LEU2)] Z815 with srb4D3::TRP1 [pSK346 (SRB4-20 LEU2)]

complex will improve our understanding of the molecular mechanisms that regulate transcription initiation.

Experimental Procedures Genetic Analysis Yeast strains used in this study are listed in Table 1. Yeast medium was prepared as described (Thompson et al., 1993). Yeast transformations were done using a lithium acetate procedure (Gietz et al., 1992). Plasmid shuffle techniques were performed as described (Boeke et al., 1987). The growth phenotypes of cells containing CTD truncations and dominant suppressing alleles were assayed by plasmid shuffle with strains Z555, Z556, Z557, Z777, and Z812. Plasmids containing various CTD truncations have been described (Nonet et al., 1987). Selection of dominant SRB4 alleles that alleviate decreased galactose inducibility by mutant GAL4 was done with strain Z815. Z815 harboring either wild-type (pRJR182) or mutant (pRJR329) GAL4 derivatives (Wu et al., 1996) were transformed with LEU2 CEN plasmids containing dominant SRB4 alleles. The transformants were assayed for b-galactosidase activity as described (Scafe et al., 1990). Strains Z859, Z860, and Z861, which have a chromosomal SRB4 deletion covered by SRB4, SRB4-18, and SRB4-20 containing CEN plasmids, respectively, were generated by transforming a PCR fragment bearing TRP1 gene flanked by SRB4 sequences. Wildtype or mutant GAL4 was transformed into these strains independently and assayed for b-galactosidase activity as above. Plasmids containing SRB4 deletion mutations were transformed into the strain Z572 and grown on SC medium containing 5-FOA at 368C.

Recombinant Proteins Recombinant SRB2 and SRB5 proteins were purified from bacteria as described (Koleske et al., 1992; Thompson et al., 1993). Recombinant FLAG-SRB proteins were expressed in Sf21 cells and purified using a FLAG M2 immunoaffinity column (Kodak) as described (Koh et al., 1997). For purification of the subcomplex containing SRB2, SRB4, SRB5, and SRB6, Sf21 cells (109 cells) were coinfected with viruses expressing SRB2, FLAG-SRB4, SRB5, and SRB6 at an moi of 3–5 for each virus. Cells were collected 60 hr postinfection. The cell extract (30 ml) was purified using the FLAG M2 immunoaffinity column and a Mono Q 5/5 column (Pharmacia). GST fusions of GAL4 activation domains (GST-34 and mutant 12) were purified as described (Smith and Johnson, 1988). GAL4 derivatives were expressed in the bacterial strain BL21(DE3) pLysS and purified on an SP-Sepharose column (Pharmacia) (Wu et al., 1996) and a Superdex G-75 column (Pharmacia).

Affinity Chromatography Immunoaffinity chromatography was used to monitor interactions between SRB subunits. Monoclonal antibody 12CA5 was coupled to protein A-agarose, and an insect cell extract containing HA-tagged recombinant SRB (10–15 mg) was incubated with a second extract containing an equimolar amount of untagged recombinant SRB for 3 hr on ice. The 12CA5 coupled agarose beads (100 ml), equilibrated in MTB (Hengartner et al., 1995), were added to the reactions and incubated for 3 hr at 48C with constant agitation and washed extensively with MTB, and bound proteins were eluted with 100 ml of 1 mg/ml HA peptide. Interactions between SRB proteins and the GAL4 activation domain were tested by affinity chromatography and GST pull-down experiments. GST affinity columns were prepared by immobilizing GST-GAL4 activation domain fusion proteins on glutathione-agarose beads as described (Smith and Johnson, 1988). The columns contained 200 ml matrix and 300 mg of GST fusion proteins. Purified recombinant SRB proteins (approximately 10 mg) were loaded onto the columns, and proteins interacting with GST fusions were eluted as described (Hengartner et al., 1995). GST pull-down experiments were performed with the SRB subcomplex. The SRB subcomplex (5 mg) was incubated with 20 mg of GST-GAL4 fusion proteins for 3 hr on ice. Glutathione-agarose beads (20 ml) were added and incubated for 1 hr at 48C with constant agitation. Beads were spun down, washed extensively with MTB, and proteins eluted by boiling in 20 ml of sample buffer. Photo-Cross-Linking A GAL4 derivative, GAL4 (1–1001840–881) with a Phe to Cys substitution at the residue 856, was purified and reduced with 10 mM DTT for 1 hr in buffer A (20 mM HEPES [pH 8.2], 150 mM NaCl, 0.1 mM EDTA, 20 mM ZnSO4). Excess DTT was removed by using a gel filtration spin column (Pierce) equilibrated in buffer A. The reduced protein was incubated for 90 min at room temperature with 125 I-APDP (N-[4-(p-azidosalicylamido)butyl]-39[29-pyridyldithio]propionamide). APDP (Pierce) was labeled with NaI125 as suggested by the manufacturer. Unreacted 125I-APDP was also removed with a gel filtration spin column equilibrated in MTB. The radiolabeled GAL4 derivative protein (0.27 mM) was incubated, in the dark, with the SRB subcomplex or RNA polymerase II holoenzyme (0.34 mM) in 30 ml of MTB on ice for 1 hr. For competition experiments, GAL80 was added to the reaction at the final concentration of 1 mM. Photocross-linking was initiated with long-range UV (320 nm) for 5 min. The reaction was stopped and quenched with 20 mM DTT. The proteins were then denatured in sample buffer and subjected to a 4%–20% Tricine-SDS-PAGE. Radiolabeled proteins were visualized using Fuji BAS2000 phosphorimager. Surface Plasmon Resonance (SPR) Analysis Sensor chips with an immobilized double-stranded DNA carrying two consensus GAL4 binding sites were prepared as described

Activator Target in RNA Polymerase II Holoenzyme 903

(Wu et al., 1996). Protein–protein interactions were measured using BIAcore 2000 (Pharmacia) in the buffer MTB at 258C at a flow rate of 15 ml/min. Typically 120 ml of 40 nM GAL4 derivative was passed over the immobilized DNA, which resulted in saturated binding to the DNA-binding sites and gave the first increase in resonance response units (RU). Interactions between this GAL4–DNA complex and other proteins were determined by flowing 120 ml of 50 nM solution of a sample protein over the preformed GAL4–DNA complex. Control experiments with nonspecific DNA and with immobilized streptavidin alone were simultaneously conducted to determine the background level of surface binding of proteins and the increase in RUs due to the difference in the refractive index of the protein storage buffer. Acknowledgments We thank M. Chun, K. Kosik, and R. Knowles for help in setting up a baculovirus expression system; K. Cha and D. Chao for assisting with biochemical analysis; G. Bryant, H. Causton, D. Chao, E. Gadbois, C. Hengartner, F. Holstege, S.-M. Liao, C. Thompson, Z. Zaman, and J. Zhang for advice and discussion; and A. Lee, D. Haile, X. Lu, and J. Ruan for reagents. We are grateful to K. Melcher and S. Johnston for the GST-GAL4 fusion plasmids. A. Z. A. was supported by a Helen Hay Whitney postdoctoral fellowship. This work was supported by National Institutes of Health grants to M. P. and R. A. Y. Received September 23, 1997; revised March 6, 1998. References Allison, L.A., and Ingles, C.J. (1989). Mutations in RNA polymerase II enhance or suppress mutations in GAL4. Proc. Natl. Acad. Sci. USA 86, 2794–2798. Ansari, A.Z., Bradner, J.E., and O’Halloran, T.V. (1995). DNA-bend modulation in a repressor-to-activator switching mechanism. Nature 374, 371–375. Apone, L.M., Virbasius, C.-m.A., Reese, J.C., and Green, M.R. (1996). Yeast TAFII90 is required for cell-cycle progression through G 2/M but not for general transcription activation. Genes Dev. 10, 2368–2380. Barberis, A., Pearlberg, J., Simkovich, N., Farrell, S., Reinagel, P., Bamdad, C., Sigal, G., and Ptashne, M. (1995). Contact with a component of the polymerase II holoenzyme suffices for gene activation. Cell 81, 359–368. Boeke, J., Truehart, J., Natsoulis, B., and Fink, G.R. (1987). 5-Fluoroorotic acid as a selective agent in yeast molecular genetics. Methods Enzymol. 154, 164–175. Burley, S.K., and Roeder, R.G. (1996). Biochemistry and structural biology of transcription factor IID (TFIID). Annu. Rev. Biochem. 65, 769–799. Busby, S., West, D., Lawes, M., Webster, C., Ishihama, A., and Kolb, A. (1994). Transcription activation by the Escherichia coli cyclic AMP receptor protein: receptors bound in tandem at promoters can interact synergistically. J. Mol. Biol. 241, 341–352. Chaiken, I., Rose, S., and Karlsson, R. (1992). Analysis of macromolecular interactions using immobilized ligands. Anal. Biochem. 201, 197–210. Chao, D.M., Gadbois, E.L., Murray, P.J., Anderson, S.F., Sonu, M.S., Parvin, J.D., and Young, R.A. (1996). A mammalian SRB protein associated with an RNA polymerase II holoenzyme. Nature 380, 82–85. Chen, Y., Ebright, Y.W., and Ebright, R.H. (1994). Identification of the target of a transcription activator protein by protein–protein photocrosslinking. Science 265, 90–92. Gerber, H.P., Hagmann, M., Seipel, K., Georgiev, O., West, M.A., Litingtung, Y., Schaffner, W., Corden, J.L. (1995). RNA polymerase II C-terminal domain required for enhancer-driven transcription. Nature 374, 660–662. Gietz, D., St. Jean, A., Woods, R.A., and Schiestl, R.H. (1992). Improved method for high efficiency transformation of intact yeast cells. Nucleic Acids Res. 20, 1425.

Goodrich, J.A., Hoey, T., Thut, C.J., Admon, A., and Tjian, R. (1993). Drosophila TAFII 40 interacts with both a VP16 activation domain and the basal transcriptional factor TFIIB. Cell 75, 519–530. Gustafsson, C.M., Myers, L.C., Li, Y., Redd, M.J., Lui, M., ErdjumentBromage, H., Tempst, P., and Kornberg, R.D. (1997). Identification of Rox3 as a component of mediator and RNA polymerase II holoenzyme. J. Biol. Chem. 272, 48–50. Hengartner, C.J., Thompson, C.M., Zhang, J., Chao, D.M., Liao, S.-M., Koleske, A.J., Okamura, S., and Young, R.A. (1995). Association of an activator with an RNA polymerase II holoenzyme. Genes Dev. 9, 897–910. Hoffmann, A., Oelgeschla¨ger, T., and Roeder, R.G. (1997). Considerations of transcriptional control mechanism: do TFIID-core promoter complexes recapitulate nucleosome-like functions? Proc. Natl. Acad. Sci. USA 94, 8928–8935. Igarashi, K., and Ishihama, A. (1991). Bipartite functional map of the E. coli RNA polymerase a subunit: involvement of the C-terminal region in transcription activation by cAMP-CRP. Cell 65, 1015–1022. Ingles, C.J., Shales, M., Cress, W.D., Triezenberg, S.J., and Greenblatt, J. (1991). Reduced binding of TFIID to transcriptionally compromised mutants of VP16. Nature 351, 588–590. Jiang, Y.W., and Stillman, D.J. (1992). Involvement of the SIN4 global transcriptional regulator in the chromatin structure of Saccharomyces cerevisiae. Mol. Cell. Biol. 12, 4503–4514. Kim, Y.-J., Bjo¨ rklund, S., Li, Y., Sayre, M.H., and Kornberg, R.D. (1994). A multiprotein mediator of transcriptional activation and its interaction with the C-terminal repeat domain of RNA polymerase II. Cell 77, 599–608. Kobayashi, N., Boyer, T.G., and Berk, A.J. (1995). A class of activation domains interacts directly with TFIIA and stimulates TFIIA– TFIID-promoter complex assembly. Mol. Cell. Biol. 15, 6465–6473. Koh, S.S., Hengartner, C.J., and Young, R.A. (1997). Baculoviral transfer vectors for expression of FLAG fusion proteins in insect cells. Biotechniques 23, 622–627. Kolb, A., Igarashi, K., Ishihama, A., Lavigne, M., Buckle, M., and Buc, H. (1993). E. coli RNA polymerase, deleted in the C-terminal part of its a-subunit, interacts with the cAMP–CRP complex at the lacP1 and the galP1 promoter. Nucleic Acids Res. 21, 319–326. Koleske, A.J., and Young, R.A. (1994). An RNA polymerase II holoenzyme responsive to activators. Nature 368, 466–469. Koleske, A.J., and Young, R.A. (1995). The RNA polymerase II holoenzyme and its implication for gene regulation. Trends Biochem. Sci. 20, 113–116. Koleske, A.J., Buratowski, S., Nonet, M., and Young, R.A. (1992). A novel transcription factor reveals a functional link between the RNA polymerase II CTD and TFIID. Cell 69, 883–894. Kuchin, S., Yeghiayan, P., and Carlson, M. (1995). Cyclin-dependent protein kinase and cyclin homologs SSN3 and SSN8 contribute to transcriptional control in yeast. Proc. Natl. Acad. Sci. USA 92, 4006– 4010. Kumar, A., Grimes, B., Fujita, N., Makino, K., Malloch, R.A., Hayward, R.S., and Ishihama, A. (1994). Role of the s 70 subunit of Escherichia coli RNA polymerase in transcription activation. J. Mol. Biol. 235, 404–413. Lee, Y.C., Min, S., Gim, B.S., and Kim, Y.-J. (1997). A transcriptional mediator protein that is required for activation of many RNA polymerase II promoters and is conserved from yeast to humans. Mol. Cell. Biol. 17, 4622–4632. Li, M., Moyle, H., and Susskind, M.M. (1994). Target of the transcriptional activation function of phage l cI protein. Science 263, 75–77. Li, Y., Bjo¨rklund, S., Jiang, Y.W., Kim, Y.-J., Lane, W.S., Stillman, D.J., and Kornberg, R.D. (1995). Yeast global transcriptional regulators Sin4 and Rgr1 are components of mediator complex/RNA polymerase II holoenzyme. Proc. Natl. Acad. Sci. USA 92, 10864–10868. Liao, S.-M., Taylor, I.C.A., Kingston, R.E., and Young, R.A. (1991). RNA polymerase II C-terminal domain contributes to the response to multiple acidic activators in vitro. Genes Dev. 5, 2431–2440. Liao, S.-M., Zhang, J., Jeffery, D.A., Koleske, A.J., Thompson, C.M., Chao, D.M., Viljoen, M., van Vuuren, H.J.J., and Young, R.A. (1995).

Molecular Cell 904

A kinase-cyclin pair in the RNA polymerase II holoenzyme. Nature 374, 193–196. Lin, Y.-S., Ha, I., Maldonado, E., Reinberg, D., and Green, M. (1991). Binding of general transcription factor TFIIB to an acidic activating region. Nature 353, 569–571. Ma, J., and Ptashne, M. (1987). The carboxy-terminal 30 amino acids of GAL4 are recognized by GAL80. Cell 50, 137–142. Maldonado, E., Shiekhattar, R., Sheldon, M., Cho, H., Drapkin, R., Rickert, P., Lees, E., Anderson, C.W., Linn, S., and Reinberg, D. (1996). A human RNA polymerase II complex associated with SRB and DNA-repair proteins. Nature 381, 86–89. Melcher, K., and Johnston, S.A. (1995). GAL4 interacts with TATAbinding protein and coactivators. Mol. Cell. Biol. 15, 2839–2848. Moqtaderi, Z., Bai, Y., Poon, D., Weil, P.A., and Struhl, K. (1996). TBP-associated factors are not generally required for transcriptional activation in yeast. Nature 383, 188–191. Myers, L.C., Gustafsson, C.M., Bushnell, D.A., Lui, M., ErdjumentBromage, H., Tempst, P., and Kornberg, R.D. (1998). The Med proteins of yeast and their function through the RNA polymerase II carboxy-terminal domain. Genes Dev. 12, 45–54. Niu, W., Kim, Y., Tau, G., Heyduk, T., and Ebright, R.H. (1996). Transcriptional activation at class II CAP-dependent promoters: two interactions between CAP and RNA polymerase. Cell 87, 1123–1134. Nonet, M.L., and Young, R.A. (1989). Intragenic and extragenic suppressors of mutations in the heptapeptide repeat domain of Saccharomyces cerevisiae RNA polymerase II. Genetics 123, 715–724. Nonet, M., Sweetser, D., and Young, R.A. (1987). Functional redundancy and structural polymorphism in the large subunit of RNA polymerase II. Cell 50, 909–915. Orphanides, G., Lagrange, T., and Reinberg, D. (1996). The general transcriptional factors of RNA polymerase II. Genes Dev. 10, 2657– 2683. Ossipow, V., Tassan, J.-P., Nigg, E.A., and Schibler, U. (1995). A mammalian RNA polymerase II holoenzyme containing all components required for promoter-specific transcription initiation. Cell 83, 137–146. Pan, G., Aso, T., and Greenblatt, J. (1997). Interaction of elongation factors TFIIS and elogin A with a human RNA polymerase II holoenzyme capable of promoter-specific initiation and responsive to transcriptional activators. J. Biol. Chem. 272, 24563–24571. Ptashne, M., and Gann, A. (1997). Transcriptional activation by recruitment. Nature 386, 569–577. Pugh, B.F. (1996). Mechanisms of transcription complex assembly. Curr. Opin. Cell Biol. 8, 303–311. Ranish, J.A., and Hahn, S. (1996). Transcription: basal factors and activation. Curr. Opin. Genet. Dev. 6, 151–158. Ren, Y.L., Garges, S., Adhya, S., and Krakow, J.S. (1988). Cooperative DNA binding of heterologous proteins: evidence for contact between the cyclic AMP receptor protein and RNA polymerase. Proc. Natl. Acad. Sci. USA 85, 4138–4142. Roberts, S.G.E., Ha, I., Maldonado, E., Reinberg, D., and Green, M.R. (1993). Interaction between an acidic activator and transcription factor TFIIB is required for transcriptional activation. Nature 363, 741–744. Scafe, C., Chao, D., Lopes, J., Hirsch, J.P., Henry, S., and Young, R.A. (1990). RNA polymerase II C-terminal repeat influences response to transcriptional enhancer signals. Nature 347, 491–494. Scully, R., Anderson, S.F., Chao, D.M., Wei, W., Ye, L., Young, R.A., Livingston, D.M., Parvin, J.D. (1997). BRAC1 is a component of the RNA polymerase II holoenzyme. Proc. Natl. Acad. Sci. USA 94, 5605–5610. Smith, D.B., and Johnson, K.S. (1988). Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Gene 67, 31–40. Song, W., Treich, I., Qian, N., Kuchin, S., and Carlson, M. (1996). SSN genes that affect transcriptional repression in Saccharomyces cerevisiae encode SIN4, ROX3, and SRB proteins associated with RNA polymerase II. Mol. Cell. Biol. 16, 115–120. Stargell, L.A., and Struhl, K. (1996). Mechanisms of transcriptional activation in vivo: two steps forward. Trends Genet. 12, 311–315.

Strich, R., Slater, M.R., and Esposito, R.E. (1989). Identification of negative regulatory genes that govern the expression of early meiotic genes in yeast. Proc. Natl. Acad. Sci. USA 86, 10018–10022. Stringer, K.F., Ingles, C.J., and Greenblatt, J. (1990). Direct and selective binding of an acidic transcriptional activation domain to the TATA-box factor TFIID. Nature 345, 783–786. Struhl, K. (1996). Chromatin structure and RNA polymerase II connection: implication for transcription. Cell 84, 179–182. Suzuki, Y., Nogi, Y., Abe, A., and Fukasawa, T. (1988). GAL11 protein, an auxillary transcription activator for genes encoding galactosemetabolizing enzymes in Saccharomyces cerevisiae. Mol. Cell. Biol. 8, 4991–4999. Tang, H., Severinov, K., Goldfarb, A., Fenyo, D., Chait, B., and Ebright, R.H. (1994). Location, structure, and function of the target of a transcriptional activator protein. Genes Dev. 8, 3058–3067. Thompson, C.M., and Young, R.A. (1995). General requirement for RNA polymerase II holoenzymes in vivo. Proc. Natl. Acad. Sci. USA 92, 4587–4590. Thompson, C.M., Koleske, A.J., Chao, C.M., and Young, R.A. (1993). A multisubunit complex associated with the RNA polymerase II CTD and TATA-binding protein in yeast. Cell 73, 1367–1375. Triezenberg, S.J. (1995). Structure and function of transcriptional activation domains. Curr. Opin. Genet. Dev. 5, 190–196. Verrijzer, C.P., and Tjian, R. (1996). TAFs mediate transcriptional activation and promoter selectivity. Trends Biochem. Sci. 21, 338–342. Wahi, M., and Johnson, A.D. (1995). Identification of genes for a2 repression in Saccharomyces cerevisiae. Genetics 140, 79–90. Walker, S.S., Reese, J.C., Apone, L.M., and Green, M.R. (1996). Transcription activation in cells lacking TAFIIs. Nature 383, 185–188. Wang, E.H., and Tjian, R. (1994). Promoter-selective transcriptional defect in cell cycle mutant ts13 rescued by hTAFII 250. Science 263, 811–814. Weissman, J.S., Kashi, Y., Fenton, W.A., and Horwitz, A. (1994). GroEL-mediated protein folding proceeds by multiple rounds of binding and release of nonnative forms. Cell 78, 693–702. Wu, Y., Reece, R.J., and Ptashne, M. (1996). Quantitation of putative activator-target affinities predicts transcriptional activating potentials. EMBO J. 15, 3951–3963. Xiao, H., Pearson, A., Coulombe, B., Truant, R., Zhang, S., Regier, J.L., Triezenberg, S.J., Reinberg, D., Flores, O., Ingles, C.J., and Greenblatt, J. (1994). Binding of basal transcriptional factor TFIIH to the acidic activation domain of VP16 and p53. Mol. Cell. Biol. 14, 7013–7024. Zhou, Y., Zhang, X., and Ebright, R.H. (1993). Identification of the activating region of catabolite gene activator protein (CAP): isolation and characterization of mutants of CAP specifically defective in transcription activation. Proc. Natl. Acad. Sci. USA 90, 6081–6085. Zhu, H., Joliot, V., and Prywes, R. (1994). Role of transcription factor TFIIF in serum response factor-activated transcription. J. Biol. Chem. 269, 3489–3497. Zou, C., Fujita, N., Igarashi, K., and Ishihama, A. (1992). Mapping the cAMP receptor protein contact site on the a subunit of E. coli RNA polymerase. Mol. Microbiol. 6, 2599–2605.