DA-Complex Assembly Activity Required for VP16C Transcriptional ...

1 downloads 19 Views 1MB Size Report
silver staining of a sodium dodecyl sulfate (SDS)-polyacrylamide gel of purified ...... Sun, X., D. Ma, M. Sheldon, K. Yeung, and D. Reinberg. 1994. Reconstitu-.
MOLECULAR AND CELLULAR BIOLOGY, July 1998, p. 4023–4031 0270-7306/98/$04.0010 Copyright © 1998, American Society for Microbiology. All Rights Reserved.

Vol. 18, No. 7

DA-Complex Assembly Activity Required for VP16C Transcriptional Activation NAOKO KOBAYASHI,1 PETER J. HORN,2 SUSAN M. SULLIVAN,2 STEVEN J. TRIEZENBERG,2 THOMAS G. BOYER,1 AND ARNOLD J. BERK1* Department of Microbiology and Molecular Genetics, Molecular Biology Institute, University of California, Los Angeles, Los Angeles, California 90095-1570,1 and Department of Biochemistry, Michigan State University, East Lansing, Michigan 48824-13192 Received 6 March 1998/Returned for modification 31 March 1998/Accepted 15 April 1998

One class of transcriptional activation domains stimulates the concerted binding of TFIIA and TFIID to promoter DNA. To test whether this DA-complex assembly activity contributes significantly to the overall mechanism of activation in vivo, we analyzed mutants of the 38-amino-acid residue VP16C activation subdomain from herpes simplex virus. An excellent correlation was observed between the in vivo activation function of these mutants and their in vitro DA-complex assembly activity. Mutants severely defective for in vivo activation also showed reduced in vitro binding to native TFIIA. No significant correlation between in vivo activation function and in vitro binding to human TATA binding protein, human TFIIB, or Drosophila melanogaster TAFII40 was observed for this set of VP16C mutants. These results argue that the ability of VP16C to increase the rate and extent of DA-complex assembly makes a significant contribution to the overall mechanism of transcriptional activation in vivo. icantly to the overall mechanism of activation in vivo. In fact, we observed an excellent correlation between DA-complex assembly activity in vitro and activation function in vivo for a set of eight VP16C mutants with minor to severe defects in activation function. Mutants with severe defects also showed reduced binding to TFIIA in vitro. No correlation was observed between in vivo activation function and in vitro binding to several other polypeptides in the initiation complex that bind in vitro to VP16C to various extents. These results strongly support the model that DA-complex assembly activity is an important component of the VP16C activation mechanism. We discuss these results in light of evidence for other VP16C functions that contribute to transcriptional activation, and we suggest that the 38-residue sequence has evolved to make several distinct interactions with different host proteins, with each type of interaction contributing to transcriptional activation.

Considerable experimental evidence indicates that transcriptional activators bound to enhancer and promoter proximal sequences stimulate transcription through interactions between their activation domains and components of the initiation complex assembled at the associated promoter (33, 36, 39, 42, 46). These interactions have been postulated to stimulate the assembly of initiation complexes at promoters or to affect the activities of general transcription factors in the assembled initiation complex so as to increase the rate of initiation by polymerase II (Pol II). But with few exceptions, relatively little is understood about the detailed mechanism by which specific activators stimulate transcription. Some activation domains have been found to increase both the rate and the final extent of an early step in initiation complex assembly, the concerted binding of TFIIA and TFIID to the TATA box and initiation site region of the promoter (8, 9, 24, 28, 49). This DA-complex assembly activity can be conveniently assayed in vitro by an electrophoretic mobility shift assay (EMSA) using agarose gels to separate the large DNAprotein complexes involved (28). Activators that stimulate DA-complex assembly also interact directly with TFIIA, as observed by coimmunoprecipitation and affinity column chromatography with recombinant proteins (24, 32). An extensive mutational analysis of a well-studied activation domain with DA-complex assembly activity, the 38-residue Cterminal activation subdomain of herpes simplex virus type 1 VP16, called VP16C, has recently been completed (43a). A number of single-, double-, and triple-point mutants of this sequence with various levels of activation function in vivo in Saccharomyces cerevisiae were identified. Since wild-type (wt) VP16C exhibits DA-complex assembly activity (24), the isolation of these VP16C mutants allowed us to test whether this in vitro activity correlated with their in vivo activation function, as would be expected if DA-complex assembly contributes signif-

MATERIALS AND METHODS Plasmid constructions. The EcoRI-BamHI fragment of pTMGAL4-VP16C (24), encoding VP16 amino acid residues 452 to 490, was inserted between the EcoRI and BamHI sites of pBlueScript (SK1) (Stratagene), yielding pBSVP16C. Oligonucleotide-directed mutagenesis (Amersham Sculptor) was used to introduce mutations of interest. EcoRI-SstI fragments of pBSVP16C encoding the wt or mutant VP16C sequences were inserted between the EcoRI and SstI sites of pSG422 (38) after confirmation of the sequence. The resulting plasmids, pSG4VP16C and its mutant derivatives, were used in transient transfection assays in COS7 cells. The same EcoRI-SstI fragments of wt and mutant pBSVP16C constructs were cloned between EcoRI and SstI sites of pETH7GAL4. The resultant constructs, pETH7GAL4VP16Cwt and mutants, were used for in vitro transcription and translation to generate 35S-labeled Gal4-wt VP16C (Gal4-VP16Cwt) and mutants used in binding experiments with TFIIA. pETH7GAL4 was constructed by inserting the PCR product encoding the NcoI-BamHI fragment of Gal4(1-94), which is also flanked by seven consecutive histidine residues at the N terminus, into pET21d (Novagen) digested with NcoI and BamHI. Transfection assays. One microgram of pSG4VP16Cwt or mutant, 5 mg of reporter plasmid pGAL4-M2-Luc (19) or pG5E1bCAT (29), 1 mg of pCH110 (18) (transfection efficiency control expressing b-galactosidase from the simian virus 40 early promoter and enhancer), and 3 mg of salmon sperm DNA were transfected into a 60-mm-diameter plate of COS7 cells at 80% confluence as a calcium phosphate precipitate in 0.5 ml. At 48 h posttransfection, the cells were washed with phosphate-buffered saline (PBS) twice and harvested with 200 ml of lysis buffer (Promega luciferase assay system) for cells transfected with pGAL4-

* Corresponding author. Mailing address: Molecular Biology Institute, UCLA, Los Angeles, CA 90095-1570. Phone: (310) 206-6298. Fax: (310) 206-7286. E-mail: [email protected]. 4023

4024

KOBAYASHI ET AL.

M2-Luc or with buffer containing 0.04 M Tris (pH 7.4), 1 mM EDTA, and 0.15 M NaCl for cells transfected with pG5E1bCAT. Luciferase was assayed with the Promega luciferase assay substrate for 10-s measurements using a Monolight 2010 luminometer (Analytical Luminescence Laboratory). Chloramphenicol acetyltransferase (CAT) was assayed as described previously (13), with quantitation performed by using a Molecular Dynamics PhosphorImager. Values within one transfection experiment were normalized for transfection efficiency according to b-galactosidase activity assayed by o-nitrophenylgalactopyranoside hydrolysis. To examine the level of expression of Gal4-VP16C mutants, 100-mm-diameter plates of COS7 cells were transfected with 5 mg of pCH110, 15 mg of pSG4VP16C, and 5 mg of pGAL4-M2-Luc in 1 ml as described above. At 48 h posttransfection, cells were harvested and two-thirds of the cells were subjected to nuclear extract preparation as described previously (26). The rest of the cells were resuspended in lysis buffer as described above, and luciferase and b-galactosidase activities were determined. Two microliters of nuclear extract was incubated with the 23-bp DNA probe which contains a single Gal4 binding site (59TAGCGGAGTACTGTCCTCCTGAG39) for 20 min at 30°C and subjected to electrophoresis in a 5% polyacrylamide gel run in 45 mM Tris base–45 mM boric acid–1 mM EDTA (pH 8.3). The binding reaction mixture (12.5 ml) contained 12.5 mM HEPES (pH 7.8), 60 mM KCl, 12.5% glycerol, 5 mM MgCl2, 10 mM b-mercaptoethanol, bovine serum albumin (1 mg/ml), poly(dI-dC) z poly(dIdC) (80 mg/ml), and the DNA probe (104 cpm). The gel was dried onto DEAEpaper (Whatman) and exposed to a Molecular Dynamics PhosphorImager screen. DNA binding to a Gal4 site for each Gal4-VP16C mutant in transiently transfected cells was determined by quantitating the specific binding with a PhosphorImager and normalizing by transfection efficiency determined by b-galactosidase activity. DA-complex assembly assays. Binding reactions (15 ml) were performed in 12.5 mM HEPES (pH 7.8)–60 mM KCl–12.5% glycerol, 5 mM MgCl2, 10 mM b-mercaptoethanol–bovine serum albumin (1 mg/ml)–poly(dI-dC) z poly(dI-dC) (40 mg/ml), and the DNA probe (;104 cpm, ;5 fmol). The G5E4T probe used in the binding reaction was the 240-bp HindIII-EcoRI fragment of pG5E4CAT end labeled with Klenow fragments and [a-32P]dATP. Where indicated, an amount of TFIID was added that, in the absence of other proteins, retarded ;5 to 10% of the probe in an agarose gel EMSA. This amounted to 4 ng of TATA binding protein (TBP) polypeptide in endogenous TFIID (eTFIID), assayed by silver staining of a sodium dodecyl sulfate (SDS)-polyacrylamide gel of purified eTFIID compared to a dilution series of a known concentration of recombinant TBP (rTBP). Where indicated, 7 ng of recombinant rTFIIA was added. Where indicated, Gal4-VP16Cwt or mutant was added, by using 1.5 times the amount of purified, recombinant protein required to saturate the Gal4 sites on the probe (;10 ng). Binding reaction mixtures were incubated for 20 min or the indicated time at 30°C. Following incubation, the binding reaction mixture was loaded onto an agarose EMSA gel, prepared and run as described previously (28). After electrophoresis, the gel was dried onto DEAE-paper and analyzed with a Molecular Dynamics PhosphorImager. DA-complex assembly activity was defined as [(fraction of cpm in Gal4-VP16C mutant–TFIID–TFIIA complex) 2 (fraction of cpm in TFIID-TFIIA complex)]/[(fraction of cpm in complex of Gal4-VP16Cwt– TFIID–TFIIA complex) 2 (fraction of cpm in TFIID-TFIIA complex)] 3 100. TFIID was affinity purified from HeLa cells expressing epitope-tagged TBP as described previously (53). TFIIA was prepared from the three isolated, Histagged subunits of human TFIIA (hTFIIA) by renaturation as described previously (32). Gal4-VP16Cwt and mutants were expressed from plasmids pVP16Cwt, -C1, -C2, -C3, etc., in Escherichia coli XA90 as described previously (30), with the following modifications. Polyethylenimine was added to the cell extract to a final concentration of 0.3%. The polyethylenimine pellet was resuspended in buffer A750 (20 mM HEPES [pH 7.5], 750 mM NaCl, 10 mM zinc acetate, 20 mM b-mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride, pepstatin [1 mg/ml], leupeptin [1 mg/ml]), and the undissolved precipitate was removed by centrifugation at 10,000 3 g for 15 min. Ammonium sulfate was added to the supernatant to a final concentration of 40%. After 60 min on ice, the precipitate was collected by centrifugation at 10,000 3 g for 20 min and resuspended in buffer A0 (same as buffer A750 without NaCl). The conductivity was adjusted to that of buffer A200 by addition of buffer A0, and protein was loaded onto a heparin-Sepharose CL-6B column (5 mg of protein/ml) preequilibrated in buffer A200. The column was washed with buffer A400 and eluted with buffer A600. The fractions containing Gal4-VP16C, as assayed by SDS-polyacrylamide gel electrophoresis (PAGE) and Coomassie blue staining, were pooled and dialyzed in buffer D100 (20% glycerol, 20 mM HEPES [pH 7.9], 0.3 mM EDTA, 100 mM KCl, 10 mM b-mercaptoethanol). The proteins were at least 90% pure as judged by Coomassie blue staining of SDS-polyacrylamide gels. Coimmunoprecipitation assays. To assay Gal4-VP16Cwt and mutant binding to TFIIA (Fig. 6), 5 3 104 cpm of 35S-labeled Gal4-VP16Cwt and mutants [with the Gal4(1-94) DNA binding domain] were in vitro translated from the pETH7GAL4VP16C plasmids and incubated for 30 min at 30°C with 0.8 mg of Myc-tagged rTFIIA in 40 ml of D100 buffer containing 0.1% Nonidet P-40 (NP-40). Myc-tagged TFIIA was prepared as described previously (24) after expression of N-terminally tagged TFIIAab from pET-mycTFIIAab and TFIIAg from pQIIA-g (32). Following the binding reaction, immunoprecipitation was performed with protein A-Sepharose beads with prebound monoclonal antibody 9E10 for 75 min at room temperature. The protein A-Sepharose beads

MOL. CELL. BIOL. were washed three times with 10S buffer (250 mM NaCl, 50 mM HEPES [pH 7.2], 0.3% Nonidet P-40, 0.1% Triton X-100, 0.5 mM dithiothreitol, 10 mM sodium phosphate [pH 7.0], 1 mM NaF, 1 mM Na4P2O7; 5-min room temperature incubation in between the two washes), once with 0.8 M LiCl–100 mM Tris-HCl (pH 7.2), and once with PBS. Bound protein was eluted with SDSloading buffer and analyzed by SDS-PAGE (12% gel) followed by counting with a Molecular Dynamics PhosphorImager. Binding to HA1 epitope-tagged recombinant human TBP (hTBP) (Table 2) was assayed similarly except that immunoprecipitation was with monoclonal antibody 12CA5. Binding to HA1-tagged recombinant hTFIIB (Table 2) was assayed similarly except that the binding reaction and subsequent washing of the protein A-Sepharose beads (three times) was in 50 mM KCl–20% glycerol–20 mM HEPES (pH 7.9)–0.3 mM EDTA–10 mM b-mercaptoethanol. To assay binding of Gal4-VP16Cwt and mutants to Drosophila melanogaster TBP-associated factor (TAF) TAFII40 (dTAFII40), the p62 subunit of hTFIIH, and, as an alternative assay for binding to hTBP and hTFIIB (Fig. 7 and Table 3), 5 3 104 cpm of the 35S-labeled polypeptides was individually incubated with 1 mg of Gal4(1-147)-VP16Cwt or mutants at 30°C for 30 min in D100 buffer containing 0.1% NP-40. Three microliters of anti-Gal4(1-147) monoclonal antibody (RK5C1; Santa Cruz Biotechnology) was bound to 25 ml of protein ASepharose in 100 ml of PBS prior to addition to the protein-protein binding assay incubations. The samples were subjected to the same washing steps, SDS-PAGE, and PhosphorImager analysis described above for the TFIIA binding assays. For binding assays of Gal4-VP16Cwt and mutants to human TFIIB, binding and washing were in D50 buffer (20% glycerol, 20 mM HEPES [pH 7.9], 0.3 mM EDTA, 50 mM KCl, 10 mM b-mercaptoethanol) containing 0.1% NP-40. In vitro transcription. The 0.03-ml in vitro transcription reactions (60 min at 30°C) and subsequent CAT-specific primer extension analyses were performed as described previously (5), using 72.5 mg of unfractionated HeLa cell nuclear extract and 100 ng of template DNA pG5E1bCAT (29). The g-32P 59-endlabeled primer used (59CTCAAAATGTTCTTTACGATG CCATTGGGA39) is complementary to the CAT coding region.

RESULTS Effects of VP16C mutations on transcriptional activation in mammalian cells. VP16C (amino acids 453 to 490) mutants used in this study are listed in Table 1. These mutants were chosen from a larger set identified by using two complementary strategies (43a). Some of the mutations (here designated C1, C2, C3, and C7) were deliberately constructed as part of a thorough and systematic scan of the residues of this subdomain. Other mutations (C5, C6, C8, and C9) were isolated by using a genetic screen in yeast based on the toxicity caused by overexpression of strong transcriptional activators. An earlier study (3) had shown that mutations in an activation domain which diminish transcriptional activation also relieve the toxicity caused by overexpression. The VP16C activation domain was randomly mutagenized (by low-fidelity PCR amplification) in the context of a Gal4(1-147) fusion. Mutants that permitted the formation of large colonies when expresed at high level from a multicopy yeast plasmid were isolated. The mutant VP16C domains were then expressed as Gal4 fusions from a low-copy-number yeast vector (to avoid the toxicity) and assayed for the ability to stimulate expression of a lacZ reporter gene expressed from the CYC1 promoter fused to the GAL1-10 upstream activation sequence (pLGSD5 [16]). As indicated in Table 1, the mutants used in this study have transcriptional activities in yeast ranging from fully wt to less than 1% of the wt activity (43a). Our method for assaying DA-complex assembly activity (28) uses hTFIIA and hTFIID. Moreover, the VP16C activation domain normally functions in human cells during the initial phases of an acute herpes simplex virus type 1 infection (37, 50). Consequently, we assayed the effects of these VP16C mutations on in vivo activation in mammalian cells. Since the detailed mechanism of activation may vary for different promoters, we assayed activation from two reporter genes with different promoters to test the generality of defects observed with the various VP16C mutants. COS cells were transfected with an expression vector for Gal4(1-147) fused to wt or mutant VP16C and one of two reporter genes: either the adeno-

VOL. 18, 1998

DA-COMPLEX ASSEMBLY ACTIVITY AND VP16C ACTIVATION

4025

TABLE 1. Phenotypes of Gal4-VP16C mutants Avg % of wt activity 6 SDa Construct

Construct C8 C2 C1 C5 C3 C9 C7 C6

Mutation(s)

F479L Q477A D472A F473S, E474G F479A F479S, D486G F473A, F475A F475S, M478T, F479S

Activation (in vivo) S. cerevisiae

RARb2 (3c)

E1B (2)

90 100 100 12 10 10 ND ,1

94 6 8 58 6 5 27 6 4 3.7 6 1.1 1.3 6 0.7 0.7 6 0.2 0.3 6 0.1 0.1 6 0.1

89 6 24 103 6 3 68 6 17 4.6 6 0.6 1.7 6 2.1 2.2 6 3.1 0.4 6 0.4 0.7 6 0.9

DA assemblyb (4–6)

TFIIA bindingb (3)

50 6 14 63 6 8 35 6 6 1.6 6 6.5 26.1 6 7.2 26.6 6 7.0 216 6 4 218 6 9

102 6 37 94 6 44 110 6 36 52 6 11 45 6 19 42 6 9 20 6 8 12 6 1

a Activity relative to Gal4-VP16Cwt activity for in vivo activation in S. cerevisiae (43a) and in transient transfection assays in COS cells from the RARb2 and Ad2 E1B promoters. ND, not determined. b Determined as described in Materials and Methods. c Number of independent assays performed for each type of analysis.

virus type 2 (Ad2) E1B promoter region with five upstream Gal4 binding sites fused to CAT (pG5E1bCAT [29]) or the retinoic acid receptor b (RARb) promoter region with four upstream Gal4 sites fused to luciferase (19). Reporter gene expression was measured and normalized to the expression observed following activation with Gal4-VP16Cwt (Fig. 1; Table 1). The mutations had similar effects in yeast and COS cells, although the effects were more severe in COS cells. Mutants C1 and C2 were somewhat more defective when assayed with the RARb compared to the E1B promoter, but otherwise the mutations had similar effects in assays with either promoter. In general, mutants C8, C2, and C1 were only partially defective, while mutants C5, C3, C9, C7, and C6 had ,5% of the activity of Gal4-VP16Cwt. To determine if the reduced activity of these mutants in COS cells was due to reduced expression of the mutant activators, nuclear extracts were prepared from transfected cells, and the sequence-specific DNA binding activities of the Gal4

fusion proteins were assayed by EMSA (Fig. 2). None of the mutants was expressed at a significantly lower level than Gal4VP16Cwt. Mutants C3, C6, and C7 accumulated to considerably higher levels than Gal4-VP16Cwt (six- to ninefold higher by quantitation of three EMSAs such as shown in Fig. 2). However, the low in vivo activities of these mutants in COS cells were not due to this high level of expression, because transfection of lower concentrations of the expression vectors for these mutants did not result in increased reporter gene expression (data not shown). The activation function of the VP16C mutants was tested further in in vitro transcription assays, where the concentrations of the Gal4-VP16C mutants could be accurately controlled. In vitro transcription was assayed by using recombinant Gal4-VP16C fusion proteins added at four concentrations to HeLa nuclear extract with pG5E1bCAT as the template (Fig. 3). The VP16C mutants exhibited the same rank order of activity in the in vitro transcription assays as in the in vivo transfection assays, further arguing that the low activities of

FIG. 1. In vivo activation function of the Gal4-VP16C mutants as assayed by transient transfection into COS cells. (A) Activation of CAT expression from the Ad2 E1B promoter in reporter plasmid pG5E1bCAT (29). (B) Activation of luciferase expression from the RARb2 promoter in reporter plasmid pGAL4M2-Luc (19). Standard deviations are shown.

FIG. 2. DNA binding activities of Gal4-VP16Cwt and mutants expressed during COS cell transient transfection assays. Nuclear extract prepared from COS cells following transient transfection with pSGVP16Cwt or mutant expression vectors (as indicated above lanes 1 to 9) was incubated with a 23-bp probe containing a high-affinity Gal4 binding site and subjected to polyacrylamide gel EMSA. Lane 12 contains sample from cells transfected with pCH110 only. Lanes 13 to 16 show binding to 0.025, 0.05, 0.1, and 0.2 ng of purified recombinant Gal4-VP16Cwt; a 100-fold molar excess of unlabeled Gal4 site probe (lane 10) or a 100-fold molar excess of an unrelated oligonucleotide (oligo) containing a binding site for the Zebra activator (Zta) (lane 11) was added to the reaction with nuclear extract from COS cells expressing Gal4-VP16Cwt.

4026

KOBAYASHI ET AL.

MOL. CELL. BIOL.

FIG. 3. In vitro transcription activation by Gal4-VP16Cwt and mutants. In vitro transcription was performed in HeLa cell nuclear extract with template pG5E1bCAT. Reactions contained either no additional protein (lane 1) or 40 ng (lanes 2, 6, 10, 14, 18, and 22), 80 ng (lanes 3, 7, 11, 15, 19, 23, and 26), 160 ng (lanes 4, 8, 12, 16, 20, 24, and 27), or 320 ng (5, 9, 13, 17, 21, 25, and 28) of Gal4-VP16Cwt or mutant or Gal4(1-147) alone, as indicated above the lanes. Specific transcription initiation was assayed by primer extension. The increase in transcription between 160 and 320 ng of Gal4-VP16 mutant C6 (lanes 24 and 25) was not observed in other experiments.

mutants C3, C6, and C7 in the in vivo assays (Fig. 1) were not due to the high expression levels of these mutants. We conclude that the relative reporter gene activities shown in Fig. 1 are accurate measures of the intrinsic in vivo activation functions of these mutants. Effects of VP16C mutations on DA-complex assembly activity. DA-complex assembly was assayed by agarose gel EMSA using a 240-bp probe bearing five Gal4 binding sites upstream of a high-affinity TATA box. For each Gal4-VP16C mutant protein purified from E. coli, the specific DNA binding activity was determined by both DNase I footprinting and polyacrylamide gel EMSA. An equal amount of Gal4 site binding activity (1.5 times the amount required to saturate all five Gal4 sites of the G5E4T probe) was incubated with probe plus rTFIIA and affinity-purified TFIID complex (eTFIID) isolated from HeLa cells expressing epitope-tagged TBP (53). As observed earlier, Gal4-VP16Cwt greatly increased the fraction of probe incorporated into a complex with TFIID and TFIIA (Fig. 4A; compare lanes 3 and 4 and lanes 13 and 16). Also, as observed earlier, TFIID binding was not stimulated by Gal4VP16Cwt in the absence of TFIIA (lane 14). In general, there was a correlation between the in vivo ac-

FIG. 4. DA-complex assembly activity of Gal4-VP16Cwt and mutants. (A) Probe DNA was incubated with the indicated purified proteins for 20 min and subjected to agarose gel EMSA. Positions of free probe and of complexes of probe with protein are indicated to the right. (B) Quantitation of VP16C mutant DA-complex assembly activity. Means and standard deviations of four to six independent assays are shown.

tivation function of the mutants (Fig. 1) and their DA-complex assembly activities (Fig. 4 and Table 1). Mutants C8, C2, and C1, which were only slightly reduced for in vivo activation function, had significant DA-complex assembly activities, although they were less than that of Gal4-VP16Cwt. In contrast, mutants C5, C3, C9, C7, and C6, all defective in in vivo activation function, had negligible DA-complex assembly activity. Mutants C3, C9, C7, and C6 had negative values because the amount of probe in complex with TFIID and TFIIA was slightly reduced in incubations with these mutants compared to that observed following incubation with TFIID and TFIIA alone (lanes 3 and 13). Similar results were observed for these eight VP16C mutants when fused to Gal4(1-94) [as opposed to Gal4(1-147) as described above] also expressed in and purified from E. coli (data not shown). Both the Epstein-Barr virus Zebra activator (also called Zta) and Gal4-VP16C greatly increase the kinetics of TFIID and TFIIA binding as well as the final amount of complex formed (24, 28). To analyze this kinetic aspect of VP16C DA-complex assembly activity, DNA-protein complexes were analyzed after 2-, 10-, and 20-min binding reactions with mutant C8, which is only slightly decreased in in vivo activation function, and mutant C6, which is severely defective, by using fusions to Gal4(194) (Fig. 5). Mutant C8 stimulated the rate of factor binding similarly to VP16Cwt, while mutant C6 had no effect on the kinetics of factor binding and reduced the amount of complex formed compared to TFIID and TFIIA alone, as observed in Fig. 4. Consequently, stimulation of the rate of DA-complex formation, as well as the final equilibrium amount of DNA-

FIG. 5. Kinetics of DA-complex assembly activated by VP16C mutants. Probe was incubated with the indicated proteins (eTFIID [D], rTFIIA [A], and Gal4-VP16Cwt and mutants) for 2, 10, or 20 min as indicated above each lane. Binding reaction mixtures were loaded onto a running agarose gel at the end of the incubation period.

VOL. 18, 1998

DA-COMPLEX ASSEMBLY ACTIVITY AND VP16C ACTIVATION

4027

TABLE 2. Binding of Gal4-VP16Cwt and mutants to hTBP and hTFIIB Construct

C8 C2 C1 C5 C3 C9 C7 C6

% of cpm coprecipitated by wta hTBP binding

hTFIIB binding

78 240 180 150 110 87 130 220

94 ND ND 80 59 90 92 99

a Binding assays were performed as described for Fig. 6 for the binding assay with TFIIA. HA1-tagged recombinant hTBP (0.5 mg) and hTFIIB (1.0 mg) were incubated individually with in vitro-translated, 35S-labeled Gal4-VP16Cwt or mutant and subjected to immunoprecipitation with monoclonal antibody 12CA5, directed against the HA1 epitope, as described in Materials and Methods. For Gal4-VP16Cwt, 11% of the input counts per minute coimmunoprecipitated with HA1-hTBP, compared to a background of 1.4% when no HA1-hTBP was added to the binding reaction; 6.6% of input Gal4-VP16Cwt coimmunoprecipitated with HA1-hTFIIB, compared to a background of 1.7% when no HA1-hTFIIB was added to the binding reaction. ND, not determined.

FIG. 6. Interaction of Gal4-VP16C mutants with TFIIA. (A) Myc-tagged TFIIA (1; lanes 3, 6, 9, 12, 15, 18, 21, 24, and 27) was incubated with 35S-labeled, in vitro-translated Gal4-VP16Cwt or mutants and subjected to immunoprecipitation with monoclonal antibody 9E10 directed against the Myc epitope. Immunoprecipitates were analyzed by SDS-PAGE and PhosphorImager analysis. Input lanes (I; lanes 1, 4, 7, 10, 13, 16, 19, 22, and 25) contained 5% of the input labeled protein used in the binding reactions. Lanes 2, 5, 8, 11, 14, 17, 20, 23, and 26 are control immunoprecipitations with 9E10 in the absence of Myc-TFIIA. Specifically immunoprecipitated 35S-labeled Gal4-VP16C was quantitated with a PhosphorImager. (B) Quantitation of Gal4-VP16C mutant–TFIIA interactions. Means and standard deviations of three independent assays are shown.

protein complex assembled, correlated with in vivo activation function for VP16C. Interaction between VP16C mutants and TFIIA. Earlier, we found that a direct interaction between Gal4-VP16C and native, three-subunit TFIIA could be observed by their coprecipitation following coincubation in vitro or by the binding of Gal4-VP16C to a glutathione S-transferase–TFIIA affinity column. This interaction required the VP16C activation subdomain sequence (24). To test the significance of the VP16CTFIIA interaction for VP16C in vivo activation function and in vitro DA-complex assembly activity, we analyzed the interaction of in vitro-translated wt and mutant Gal4-VP16C proteins with Myc epitope-tagged rTFIIA by coprecipitation (Fig. 6). Typically, 20 to 30% of the input Gal4-VP16Cwt was coimmunoprecipitated with Myc-TFIIA under our experimental conditions. Results from three or more assays are summarized in Table 1. Mutants C8, C2, and C1, which were only modestly reduced for in vivo activation function, coimmunoprecipitated with TFIIA to the same extent as Gal4-VP16Cwt. Mutants C5, C3, C9, and C6, which have severely reduced activation function and in vitro DA-complex assembly activity, were reproducibly reduced for in vitro binding to TFIIA, although mutants C5, C3, and C9 retained 40 to 50% of wt activity in this assay. The mutants that were most defective for in vivo activation and in vitro DA-complex assembly activity, C7 and C6, were the most defective in this TFIIA binding assay. These results suggest that VP16C binding to TFIIA is not sufficient for transcriptional activation because mutant C1 binds to TFIIA as well as Gal4-VP16Cwt yet has reduced activation function. VP16C DA-complex assembly activity

(Fig. 4) correlates better with activation function than does VP16C binding to TFIIA. However, since the mutants that are most defective for activation and for DA-complex assembly activity show reduced TFIIA binding as detected by the coimmunoprecipitation assay, the interaction between VP16C and TFIIA is probably required for DA-complex assembly activity. Interactions between Gal4-VP16C mutants and other initiation complex polypeptides. VP16C consists of only 38 amino acid residues, yet it has been reported to bind in vitro to several different isolated polypeptides which function as components of the Pol II initiation complex. These include TBP (43), TFIIB (35), hTFIIH subunit p62 (51), and dTAFII40 (11) and its human homolog TAFII31 (hTAFII31) (23). To test the significance of these interactions for the mechanism of in vivo activation, we analyzed the activities of the same set of VP16C mutants for in vitro binding to several of these initiation complex polypeptides: hTBP, hTFIIB, the p62 subunit of hTFIIH, and dTAFII40. Binding of Gal4-VP16Cwt and mutants to TBP and TFIIB was assayed similarly to TFIIA in Fig. 6. A high concentration of epitope-tagged TBP or TFIIB was incubated with in vitro-translated, labeled Gal4-VP16Cwt or mutant and then immunoprecipitated with a monoclonal antibody directed against the epitope tag. Significant binding over background was observed, as previously reported, but no correlation between the binding activity and the transcriptional activity of the mutants was apparent (Table 2). In a second set of binding assays, TBP, TFIIB, the p62 subunit of TFIIH, and dTAFII40 were each separately in vitro translated, incubated with high concentrations of purified, recombinant Gal4-VP16Cwt or mutants and subjected to immunoprecipitation with a monoclonal antibody directed against a determinant in Gal4(1-147) (Fig. 7). Significantly more of each of these polypeptides was immunoprecipitated when Gal4-VP16C was added to the incubation mixtures compared to incubation mixtures with equal amounts of Gal4(1147) alone (compare lanes 3 and 12). However, unlike the interaction with TFIIA (Fig. 6), binding of these polypeptides to mutant VP16C sequences did not vary according to the in vivo activation function of the mutants, except possibly for TFIIH subunit p62 (Table 3).

4028

KOBAYASHI ET AL.

MOL. CELL. BIOL.

FIG. 7. Binding of Gal4-VP16Cwt and mutants to hTBP, hTFIIB, dTAFII40, and the p62 subunit of hTFIIH. In vitro-translated proteins were incubated with Gal4-VP16wt or mutants, as indicated, and subjected to immunoprecipitation using a monoclonal antibody against the Gal4 DNA binding domain. Immunoprecipitated proteins were analyzed by SDS-PAGE followed by PhosphorImager analysis. Quantitation of the fraction of 35S-labeled polypeptide bound relative to the fraction bound to Gal4-VP16Cwt is shown in Table 3.

DISCUSSION Naturally occurring activation domains are often composed of multiple subdomains, each of which is independently able to activate transcription when multimerized and fused to a DNA binding domain (47). Each such activation subdomain in a naturally occurring activator might work through a different mechanism, or activation subdomains in a single activator might have redundant functions. Consequently, to simplify mechanistic studies of activation we focused our studies on a single, small activation subdomain, the 38-residue VP16C. In studying a set of VP16C mutants, we found an excellent correlation between the ability of these mutants to activate transcription in vivo and their in vitro DA-complex assembly activity as assayed by agarose gel EMSA. There is now considerable compelling evidence that DAcomplex assembly activity is an important aspect of in vivo activation by one functional class of activation domains that includes VP16 and the Zebra activator of Epstein-Barr virus. (i) When incubated with template and partially purified TFIID and TFIIA, such activators decrease the duration of a ratelimiting step in in vitro transcription (49). (ii) TFIIA must be added to a reconstituted in vitro transcription reaction mixture to observe activation by Zebra or VP16 (32, 44, 52). (iii) For TABLE 3. Binding of Gal4-VP16C to hTBP, hTFIIB, dTAFII40, and hTFIIH p62 Construct

C8 C2 C1 C5 C3 C9 C7 C6

Relative activitya hTBP

hTFIIB

dTAFII40

hTFIIH p62

100 110 160 110 63 130 60 130

100 100 94 110 130 110 110 110

110 200 170 110 120 150 120 120

110 350 190 54 52 57 37 43

a Binding reactions were performed as described for Fig. 7. Each value shown is the fraction of coimmunoprecipitating counts per minute minus background binding to the Gal4 DNA binding domain alone (Fig. 7, lane 12), relative to the value of 100 for Gal4-VP16Cwt. Values are the averages of at least two independent assays.

Gal4-VP16 and Zebra, there is a requirement for multiple copies of upstream activator binding sites for both significant activation in vivo in mammalian cells and in vitro DA-complex assembly activity (9). (iv) Mutations affecting the surface of TBP that interacts with TFIIA reduce TFIIA binding in vitro and also interfere with transcriptional activation in vivo (6, 41). (v) As reported here, a correlation exists between DA-complex assembly activity and in vivo activation function for mutants of VP16C. The ability of an activator to increase DA-complex assembly could stimulate in vivo transcription by providing a platform for the assembly of an initiation complex from Pol II and the remaining general transcription factors. This assembly could be sequential (7) or could result from binding a single preassembled Pol II holoenzyme that includes most of the remaining general transcription factors, Pol II, and other regulatory polypeptides (15, 22, 25). By either mechanism of initiation complex assembly, increased DA-complex could lead to increased transcription for all promoters where TFIID binding in the absence of activators is rate limiting. The kinetic aspect of Gal4-VP16C DA-complex assembly activity may be especially significant for in vivo activation. VP16C mutants that were severely defective for in vivo activation were also severely defective in the ability to increase the rate of complex assembly in vitro (Fig. 5). In vivo, activities such as Mot1 (1) and NC2 (10, 12, 20) inhibit transcription. In vitro, these proteins either inhibit TFIIA and TFIIB binding to a TBP-TATA box complex (21, 31) or remove TBP from TATA box DNA by an active process requiring ATP hydrolysis (1). TBP-TFIIA-TATA box complexes are resistant to the actions of both NC2 and Mot1. Consequently, the ability of Zebra and VP16 activation domains to increase the rate of DA-complex assembly may function in vivo to kinetically oppose inhibitory reactions involving Mot1, NC2, and potentially other general transcriptional inhibitors. For activators like VP16 that increase the rate of DA-complex assembly, the net effect of competing Mot1 and NC2 inhibitory reactions would be to lower the level of basal transcription in the absence of activator, thereby increasing the magnitude of activation. In a recent related study, Lieberman et al. (28a) analyzed mutants of the Epstein-Barr virus Zebra activator for DAcomplex assembly activity and for activation function in vivo and in vitro on several different promoters. They found that two Zebra mutants with significantly reduced DA-complex assembly activity were reduced for activation on some promoters but not others. Studies of additional mutants of the complex Zebra activation domain (8a) indicated that it functions by at least two distinct mechanisms and that the relative importance of the two different mechanisms varies depending on the promoter. DA-complex assembly activity was particularly important on promoters having a TATA box with low affinity for TFIID and under in vitro transcription conditions which limit TFIID or TFIIA binding to the TATA element. The Zebra activation domain is similar to the VP16 activation domain in that it contains distinct activation subdomains, each capable of activating transcription (8a). As discussed by Lieberman et al. (28a), the contradictory effects of Zebra mutations on the activation of different promoters probably result from inhibition of one of the two or more Zebra activation mechanisms but not the others. The effect of the mutations on the activation of different promoters would then depend on the extent to which the mutant function normally stimulates transcription from that promoter. We believe that our studies with the small VP16C activation subdomain revealed a clearer correlation between DA-complex assembly activity and activation function than the Lieberman et al. study (28a) because VP16C has a

VOL. 18, 1998

DA-COMPLEX ASSEMBLY ACTIVITY AND VP16C ACTIVATION

more limited set of distinct functions than the complex Zebra activation domain. We did not observe a good correlation between the in vivo activation function of this set of VP16C mutants and their binding to TBP, TFIIB, or dTAFII40 in vitro under the conditions of our assays. These results do not exclude the possibility that VP16C interactions with these polypeptides contribute to the mechanism of activation in vivo. However, the data indicate that the interactions we detect with these isolated polypeptides in the in vitro coimmunoprecipitation assay are not sufficient for activation in vivo. Mutational analysis of the TBP surface also failed to reveal evidence for the functional significance of a TBP-VP16 interaction (6, 45), although other studies suggest that some activators may make functionally important contacts directly with the surface of TBP (50a). While studies with some activators suggest that interactions with TFIIB may contribute to the activation mechanism (e.g., reference 31a), studies with TBP mutants also argue that VP16 and many other activation domains are unlikely to function by stimulating the binding of TFIIB. Some mutations in TBP residues that form the interface with TFIIB in the initiation complex significantly reduce the affinity of TFIIB for the TBPTATA box complex but have relatively little effect on activation in vivo (6, 27). Also, mutations in TFIIB that reduce the stability of a TBP-TFIIB-TATA box complex do not influence activation at most yeast promoters (9a). These results suggest that TFIIB binding is not a limiting step in transcription, as might be expected for a regulated step. In contrast, mutations in TBP residues in the interface between TBP and TFIIA that reduce the affinity of TFIIA for the TBP-TATA box complex severely reduce the ability of TBP to participate in activated transcription in vivo but do not impair basal transcription (6, 41). The binding of TFIIH subunit p62 to Gal4-VP16C was 10fold higher than a background level of binding to Gal4(1-147) alone (as quantitated by PhosphorImager from experiments such as that shown in Fig. 7). In contrast to TBP, TFIIB, and dTAFII40, the specific binding of p62 to Gal4-VP16C did diminish for the most defective VP16C mutants (Fig. 7). The effect was not very dramatic in that the most defective mutants still bound ;40% as much p62 as wild-type VP16C. However, the results might be explained if ;60% of total binding were due to a specific p62-VP16C interaction required for in vivo activation and ;40% of the total p62 binding were due to a nonspecific interaction. Also, binding of VP16C mutants to TFIIA (Fig. 6) appeared to be a less sensitive assay of mutant function than the agarose EMSA of DA-complex assembly activity (Fig. 4 and 5). A more stringent assay for the incorporation of holo-TFIIH into a functional initiation complex might show a more significant defect for the most deleterious VP16C mutations than the p62 coimmunoprecipitation assay. We were surprised that we did not observe a correlation between dTAFII40 binding and VP16C activation function. DA-complex assembly activity is not observed in binding reactions where TBP is substituted for TFIID (24, 28). Consequently, it seems likely that VP16C interacts with one or more TAFs, as well as TFIIA, in order to activate DA-complex assembly. dTAFII40 seemed like a good candidate for an interacting TAF because of its demonstrated ability to bind to VP16C (11), and the ability of a partial TFIID complex composed of only hTBP, hTAFII250, hTAFII70, and hTAFII31 (the human homolog of dTAFII40) to support transcriptional activation by Gal4-VP16C in vitro (23). Also, a recent two-dimensional nuclear magnetic resonance study (48) showed that at high concentrations, the region of hTAFII31 that is most homologous to dTAFII40 (residues 1 to 140) causes the unstruc-

4029

tured VP16C peptide (40, 48) to undergo a structural transition to an alpha helix. Moreover, the VP16C amino acid residues that exhibited changes in chemical environment when hTAFII31(1-140) was added included D472, F479, L483, and D486. The VP16C mutagenesis study implicated two of these residues, D472 and F479, as being particularly important to VP16C activation function (Table 1) (43a). Substitution of D472 with alanine (mutant C1) decreased VP16C activation in COS cells. Substitution of F479 with a leucine, another bulky hydrophobic residue, had little effect on in vivo activation (mutant C8). However, substitution of F479 with alanine had a very significant effect (mutant C3). Analyses of the in vitro activation function of VP16C deletion mutants were also consistent with the significance of D472 and F479 for the activation mechanism (48). These results tend to support the significance of the VP16C-TAFII31(1-140) interaction in the mechanism of in vivo activation, despite our failure to observe a significant effect of these mutations in the coimmunoprecipitation assay with dTAFII40. The coimmunoprecipitation assay that we used may not measure the functionally important VP16C-TAF interaction required for activation. Other studies support the significance of VP16C interactions with other transcriptional regulatory proteins. The evidence for a functional interaction with histone acetylase (HAT) complexes containing Gcn5 (14) is very strong. High-level expression of Gal4-VP16 (4) and Gal4-VP16C (43a) interferes with the replication of yeast cells. This toxicity is suppressed by inactivating mutations in any one of several subunits of the two Gcn5-containing HAT complexes. Each of these suppressing mutations prevents the formation of these HAT complexes (4, 14). These genetic interactions between the VP16 activation domain and the Gcn5-containing HAT complexes are probably consequences of the physical interaction between the activation domain and the Ada2 subunit of the Gcn5 HAT complexes (2). Finally, the Srb/Mediator complex can support VP16 activation in in vitro transcription reactions with yeast factors in the absence of TFIIA and Pol II TAFs (20, 25). This activity probably involves a direct interaction between the VP16 activation domain and Srb polypeptides in the Pol II holoenzyme complex (17). As discussed above, the correlation reported here between VP16C DA-complex assembly activity and the in vivo activation functions of VP16C mutants adds to the considerable weight of evidence that DA-complex assembly activity is a significant component of the in vivo activation mechanism. Other experimental results cited above suggest that VP16C interactions with other proteins are important in the activation mechanism. Is it reasonable to hypothesize that the simple, 38-residue VP16C peptide makes functionally significant interactions with several apparently unrelated proteins? Perhaps it is, if one considers the biological function and evolutionary history of VP16. The VP16 protein is packaged into herpes simplex virion particles between the outer viral membrane and the inner nucleocapsid. This results in the release of VP16 into the cytoplasm of a host cell during the process of viral infection and its subsequent transport to the nucleus (37). The kinetics of VP16 appearance in the nucleus suggests that it may be transported there in advance of the viral nucleocapsid so that it can bind to host cell transcription factors Oct1 and HCF in preparation for activating high levels of transcription from the viral immediate-early genes as soon as the nucleocapsid arrives in the nucleus (37). This elaborate mechanism, not observed for other classes of DNA viruses, suggests that there has been strong selective pressure for the rapid, high-level expression of viral immediate-early genes. Over the vast number of herpesvirus generations, this selective pressure has likely yielded a

4030

KOBAYASHI ET AL.

VP16C sequence that is optimized for making all possible interactions that synergize with the activation functions of the VP16N activation subdomain (34) to stimulate immediateearly gene transcription. The extensive mutational analysis of VP16C activation function performed by Sullivan et al. (43a) indicates that its interactions with functionally significant targets are largely through bulky hydrophobic residues. The structural studies of Uesugi et al. (48) also revealed that VP16C interacts with TAF31 principally through the hydrophobic face of an induced amphipathic alpha helix. This is reminiscent of the interaction of the CREB activation domain with its target domain in CBP (33a), which, once again, is through the hydrophobic face of an amphipathic alpha helix. Hydrophobic interactions may have evolved because the geometric constraints of such interactions are more flexible than for hydrogen bonds, allowing the evolution of a protein surface capable of interacting with hydrophobic patches on the surfaces of multiple different targets. The DA-complex assembly activity of VP16C probably requires interactions with TFIIA and one or more TAFs on the surface of TFIID. The excellent correlation between DA-complex assembly activity and activation function for VP16C reported here indicates that these interactions make particularly important contributions to the overall mechanism of activation by VP16C. ACKNOWLEDGMENTS This research was supported by grants CA25235 from the NCI to A.J.B. and AI27323 from the NIAID and K04 AI01824 to S.J.T. We thank Carol Eng for excellent technical assistance. REFERENCES 1. Auble, D. T., K. E. Hansen, C. G. Mueller, W. S. Lane, J. Thorner, and S. Hahn. 1994. Mot1, a global repressor of RNA polymerase II transcription, inhibits TBP binding to DNA by an ATP-dependent mechanism. Genes Dev. 8:1920–1934. 2. Barlev, N. A., R. Candau, L. Wang, P. Darpino, N. Silverman, and S. L. Berger. 1995. Characterization of physical interactions of the putative transcriptional adaptor, ADA2, with acidic activation domains and TATA-binding protein. J. Biol. Chem. 270:19337–19344. 3. Berger, S. L., W. D. Cress, A. Cress, S. J. Triezenberg, and L. Guarente. 1990. Selective inhibition of activated but not basal transcription by the acidic activation domain of VP16: evidence for transcriptional adaptors. Cell 61:1199–1208. 4. Berger, S. L., B. Pina, N. Silverman, G. A. Marcus, J. Agapite, J. L. Regier, S. J. Triezenberg, and L. Guarente. 1992. Genetic isolation of ADA2: a potential transcriptional adaptor required for function of certain acidic activation domains. Cell 70:251–265. 5. Boyer, T. G., and A. J. Berk. 1993. Functional interaction of adenovirus E1A with holo-TFIID. Genes Dev. 7:1810–1823. 6. Bryant, G. O., L. S. Martel, S. K. Burley, and A. J. Berk. 1996. Radical mutations reveal TATA-box binding protein surfaces required for activated transcription in vivo. Genes Dev. 10:2491–2504. 7. Buratowski, S., S. Hahn, L. Guarente, and P. A. Sharp. 1989. Five intermediate complexes in transcription initiation by RNA polymerase II. Cell 56: 549–561. 8. Chi, T., and M. Carey. 1996. Assembly of the isomerized TFIIA-TFIIDTATA ternary complex is necessary and sufficient for gene activation. Genes Dev. 10:2540–2550. 8a.Chi, T., and M. Carey. 1993. The ZEBRA activation domain: modular organization and mechanism of action. Mol. Cell. Biol. 13:7045–7055. 9. Chi, T., P. Lieberman, K. Ellwood, and M. Carey. 1995. A general mechanism for transcriptional synergy by eukaryotic activators. Nature 377:254– 257. 9a.Chou, S., and K. Struhl. 1997. Transcriptional activation by TFIIB mutants that are severely impaired in interaction with promoter DNA and acidic activation domains. Mol. Cell. Biol. 17:6794–6802. 10. Gadbois, E. L., D. M. Chao, J. C. Reese, M. R. Green, and R. A. Young. 1997. Functional antagonism between RNA polymerase II holoenzyme and global negative regulator NC2 in vivo. Proc. Natl. Acad. Sci. USA 94:3145–3150. 11. Goodrich, J. A., T. Hoey, C. Thut, A. Admon, and R. Tjian. 1993. Drosophila TAFII40 interacts with both a VP16 activation domain and basal transcription factor TFIIB. Cell 75:519–530. 12. Goppelt, A., and M. Meisterernst. 1996. Characterization of the basal inhib-

MOL. CELL. BIOL. itor of class II transcription NC2 from Saccharomyces cerevisiae. Nucleic Acids Res. 24:4450–4455. 13. Gorman, C. M., L. F. Moffat, and B. H. Howard. 1982. Recombinant genomes which express chloramphenicol acetyltransferase in mammalian cells. Mol. Cell. Biol. 2:1044–1051. 14. Grant, P. A., L. Duggan, J. Cote, S. M. Roberts, J. E. Brownell, R. Candau, R. Ohba, T. Owen-Hughes, C. D. Allis, F. Winston, et al. 1997. Yeast Gcn5 functions in two multisubunit complexes to acetylate nucleosomal histones: characterization of an Ada complex and the SAGA (Spt/Ada) complex. Genes Dev. 11:1640–1650. 15. Greenblatt, J. 1997. RNA Polymerase II holoenzyme and transcriptional regulation. Curr. Opin. Cell Biol. 9:310–319. 16. Guarente, L., R. R. Yocum, and P. Gifford. 1982. A GAL10-CYC1 hybrid yeast promoter identifies the GAL4 regulatory region as an upstream site. Proc. Natl. Acad. Sci. USA 79:7410–7414. 17. Hengartner, C. J., C. M. Thompson, J. Zhang, D. M. Chao, S. M. Liao, A. J. Koleske, S. Okamura, and R. A. Young. 1995. Association of an activator with an RNA polymerase II holoenzyme. Genes Dev. 9:897–910. 18. Herbomel, P., B. Bourachot, and M. Yaniv. 1984. Two distinct enhancers with different cell specificities coexist in the regulatory region of polyoma. Cell 39:653–662. 19. Keveaney, M., A. Berkenstam, M. Feigenbutz, G. Vriend, and H. G. Stunnenberg. 1993. Residues in the TATA-binding protein required to mediate a transcriptional response to retinoic acid in EC cells. Nature 365:562–566. 20. Kim, S., J. G. Na, M. Hampsey, and D. Reinberg. 1997. The Dr1/DRAP1 heterodimer is a global repressor of transcription in vivo. Proc. Natl. Acad. Sci. USA 94:820–825. 21. Kim, T. K., Y. Zhao, H. Ge, R. Bernstein, and R. G. Roeder. 1995. TATAbinding protein residues implicated in a functional interplay between negative cofactor NC2 (Dr1) and general factors TFIIA and TFIIB. J. Biol. Chem. 270:10976–10981. 22. Kim, Y. J., S. Bjorklund, Y. Li, M. H. Sayre, and R. D. Kornberg. 1994. A multiprotein mediator of transcriptional activation and its interaction with the C-terminal repeat domain of RNA polymerase II. Cell 77:599–608. 23. Klemm, R. D., J. A. Goodrich, S. Zhou, and R. Tijian. 1995. Molecular cloning and expression of the 32-kDa subunit of human TFIID reveals interactions with VP16 and TFIIB that mediate transcriptional activation. Proc. Natl. Acad. Sci. USA 92:5788–5792. 24. Kobayashi, N., T. G. Boyer, and A. J. Berk. 1995. A class of activation domains interacts directly with TFIIA and stimulates TFIIA-TFIID-promoter complex assembly. Mol. Cell. Biol. 15:6465–6473. 25. Koleske, A. J., and R. A. Young. 1994. An RNA polymerase II holoenzyme responsive to activators. Nature (London) 368:466–469. 26. Lee, K. A. W., A. Bindereif, and M. R. Green. 1988. A small-scale procedure for the preparation of nuclear extracts that support efficient transcription and pre-mRNA splicing. Gene Anal. Tech. 5:22–33. 27. Lee, M., and K. Struhl. 1997. A severely defective TATA-binding proteinTFIIB interaction does not preclude transcriptional activation in vivo. Mol. Cell. Biol. 17:1336–1345. 28. Lieberman, P. M., and A. J. Berk. 1994. A mechanism for TAFs in transcriptional activation: activation domain enhancement of TFIID-TFIIA-promoter DNA complex formation. Genes Dev. 8:995–1006. 28a.Lieberman, P. M., J. Ozer, and D. B. Gu ¨rsel. 1997. Requirement for transcription factor IIA (TFIIA)-TFIID recruitment by an activator depends on promoter structure and template competition. Mol. Cell. Biol. 17:6624–6632. 29. Lillie, J. W., and M. R. Green. 1989. Transcription activation by the adenovirus E1A protein. Nature 338:39–44. 30. Lin, Y. S., M. F. Carey, M. Ptashne, and M. R. Green. 1988. GAL4 derivatives function alone and synergistically with mammalian activators in vitro. Cell 54:659–664. 31. Mermelstein, F., K. Yeung, J. Cao, J. A. Inostroza, H. Erdjument-Bromage, K. Eagelson, D. Landsman, P. Levitt, P. Tempst, and D. Reinberg. 1996. Requirement of a corepressor for Dr1-mediated repression of transcription. Genes Dev. 10:1033–1048. 31a.Nerlov, C., and E. B. Ziff. 1995. CCAAT/enhancer binding protein-alpha amino acid motifs with dual TBP and TFIIB binding ability co-operate to activate transcription in both yeast and mammalian cells. EMBO J. 14:4318– 4328. 32. Ozer, J., P. A. Moore, A. H. Bolde, A. Lee, C. A. Rosen, and P. M. Lieberman. 1994. Molecular cloning of the small (g) subunit of human TFIIA reveals functions critical for activated transcription. Genes Dev. 8:2324–2335. 33. Ptashne, M., and A. Gann. 1997. Transcriptional activation by recruitment. Nature 386:569–577. 33a.Radhakrishnan, I., G. C. Paerez-Alvarado, D. Parker, H. J. Dyson, M. R. Montminy, and P. E. Wright. 1997. Solution structure of the KIX domain of CBP bound to the transactivation domain of CREB: a model for activator: coactivator interactions. Cell 6:741–752. 34. Regier, J. L., F. Shen, and S. J. Triezenberg. 1993. Pattern of aromatic and hydrophobic amino acids critical for one of two subdomains of the VP16 transcriptional activator. Proc. Natl. Acad. Sci. USA 90:883–887. 35. Roberts, S. G. E., I. Ha, E. Maldonado, D. Reinberg, and M. R. Green. 1993. Interaction between an acidic activator and transcription factor TFIIB is

VOL. 18, 1998

DA-COMPLEX ASSEMBLY ACTIVITY AND VP16C ACTIVATION

required for transcriptional activation. Nature 363:741–744. 36. Roeder, R. G. 1996. The role of general initiation factors in transcription by RNA polymerase II. Trends Biochem. Sci. 21:327–335. 37. Roizman, B., and A. E. Sears. 1996. Virology, 3rd ed., p. 2231–2296. Lippincott-Raven, Philadelphia, Pa. 38. Sadowski, I., and M. Ptashne. 1989. A vector for expressing GAL4(1-147) fusions in mammalian cells. Nucleic Acids Res. 17:7539. 39. Sheldon, M., and D. Reinberg. 1995. Transcriptional activation. Tuning-up transcription. Curr. Biol. 5:43–46. 40. Shen, F., S. J. Triezenberg, P. Hensley, D. Porter, and J. R. Knutson. 1996. Critical amino acids in the activation domain of the herpes virus protein VP16 are solvent-exposed in highly mobile protein segments. J. Biol. Chem. 271:4819–4826. 41. Stargell, L. A., and K. Struhl. 1995. The TBP-TFIIA interaction in the response to acidic activators in vivo. Science 269:75–78. 42. Stargell, L. A., and K. Struhl. 1996. Mechanisms of transcriptional activation in vivo: two steps forward. Trends Genet. 12:311–315. 43. Stringer, K. F., C. J. Ingles, and J. Greenblatt. 1990. An acidic transcriptional activation domain binds directly and selectively to the TATA-box factor. Nature (London) 345:783–786. 43a.Sullivan, S. M., P. J. Horn, V. A. Olson, A. H. Koop, W. Niu, R. H. Ebright, and S. J. Triezenberg. Unpublished data. 44. Sun, X., D. Ma, M. Sheldon, K. Yeung, and D. Reinberg. 1994. Reconstitution of human TFIIA activity from recombinant polypeptides: a role in TFIID-mediated transcription. Genes Dev. 8:2336–2348. 45. Tansey, W. P., and W. Herr. 1995. The ability to associate with activation domains in vitro is not required for the TATA-box binding protein to sup-

4031

port activated transcription in vivo. Proc. Natl. Acad. Sci. USA 93:1119– 1124. 46. Tjian, R., and T. Maniatis. 1994. Transcriptional activation: a complex puzzle with few easy pieces. Cell 77:5–8. 47. Triezenberg, S. J. 1995. Structure and function of transcriptional activation domains. Curr. Opin. Gen. Dev. 5:190–196. 48. Uesugi, M., O. Nyanguile, H. Lu, A. J. Levine, and G. L. Verdine. 1997. Induced a helix in the VP16 activation domain upon binding to a human TAF. Science 277:1310–1313. 49. Wang, W., J. D. Gralla, and M. Carey. 1992. The acidic activator GAL4-AH can stimulate polymerase II transcription by promoting assembly of a closed complex requiring TFIID and TFIIA. Genes Dev. 6:1716–1727. 50. Ward, P. L., and B. Roizman. 1994. Herpes simplex genes: the blueprint of a successful human pathogen. Trends Genet. 10:267–274. 50a.Wu, Y., R. J. Reece, and M. Ptashne. 1996. Quantitation of putative activator-target affinities predicts transcriptional activating potentials. EMBO J. 15:3951–3963. 51. Xiao, H., A. Pearson, B. Coulombe, R. Truant, S. Zhang, J. L. Regier, S. J. Triezenberg, D. Reinberg, O. Flores, C. J. Ingles, and J. Greenblatt. 1994. Binding of basal transcription factor TFIIH to the acidic activation domains of VP16 and p53. Mol. Cell. Biol. 14:7013–7024. 52. Yokomori, K., M. P. Zeidler, J.-L. Chen, C. P. Verrijzer, M. Mlodzik, and R. Tjian. 1994. Drosophila TFIIA directs cooperative DNA binding with TBP and mediates transcriptional activation. Genes Dev. 8:2312–2323. 53. Zhou, Q., P. M. Lieberman, T. G. Boyer, and A. J. Berk. 1992. Holo-TFIID supports transcriptional stimulation by diverse activators and from a TATAless promoter. Genes Dev. 6:1964–1974.