Cell, Vol. 61, 1199-1206,

June 29, 1990, Copyright

0 1990 by Cell Press

Selective Inhibition of Activated but Not Basal Transcription by the Acidic Activation Domain of VP16: Evidence for Transcriptional Adaptors Shelley L. Berger,” W. Douglas Cress,7 Andrea Cress,7 Steven J. Triezenbetgt and Leonard Guarente* * Department of Biology Massachusetts Institute of Technology Cambridge, Massachusetts 02139 tDepartment of Biochemistry Michigan State University East Lansing, Michigan 48824

Summary The interaction between the chimeric activator GAL4VP16, consisting of the DNA binding domain of GAL4 and the acidic activation domain of VP16, and its target in the transcriptional machinery was studied in vitro. GAL4-VP16 stimulated transcription from a promoter bearing GAL4 sites, and greatly inhibited transcription from a promoter bearing binding sites for the dA:dT activator and from a basal promoter bearing only a TATA box. Mutations in the acidic domain that reduced activation from the GAL4 site promoter also reduced inhibition from the dA:dT promoter, indicating a similar interaction between VP16 and its target in both processes. Strikingly, if the DNA binding domain of GAL4-VP16 was occupied by a GAL4 site oligonucleotide, the protein inhibited activation by the dA:dT actiwator but did not inhibit basal transcription. We propose that, under these conditions, GAL4VP16 acted to titrate an “adaptor” that bridges an interaction between the upstream activator and the basic transcriptional machinery at the TATA box. Introduction ‘The activation of transcription is of central importance in ,the regulation of genetic programs that range from control Iof cell growth to embryonic development. The dissection of promoters from various eukaryotes has revealed two essential sequence elements: the proximal promoter, including the TATA box (Benoist and Chambon, 1981) and the site of upstream activation, termed the UAS or enhancer (Banerji et al., 1981; Guarente et al., 1982). The TATA box, which may lie hundreds or thousands of base pairs from the UAS, is present in promoters of most genes transcribed by RNA polymerase II and is the binding site for a protein termed TFIID (Matsui et al., 1980; Samuels et al., 1982). Binding of TFIID to the TATA box initiates an ordered assembly of the components of the basic transcriptional machinery (Buratowski et al., 1989; Van Dyke et al., 1988). The component factors of this machinery have been identified by fractionation of the mammalian in vitro transcription system, and include, in addition to TFIID: TFIIA, TFIIB, TFIIE, and RNA polymerase II itself (Matsui et al., 1980; Samuels et al., 1982). The second class of essential elements in the promoter,

the UASs, are gene-specific DNA sequence elements and are binding sites for transcriptional activators (Guarente, 1987; Johnson and McKnight, 1989; Struhl, 1989). Genespecific regulation of transcription occurs largely because different transcriptional activators respond to specific regulatory signals. Transcriptional activators contain distinct domains dedicated to DNA binding at the UAS and to the activation of the transcriptional machinery (Brent and Ptashne, 1985; Hope and Struhl, 1986). The DNA binding domains of activators are of various classes, which have been termed helix-loop-helix, helix-turn-helix, leucine zipper, and zinc finger (Johnson and McKnight, 1989). There are also several types of transcriptional activation domains. A major class, including many yeast activators, has a highly acidic domain enriched in glutamate and aspartate (Hope and Struhl, 1986; Ma and Ptashne, 1987). Alteration of the sequence of acidic residues, without changing the negative charge, strongly affects activation potential (Giniger and Ptashne, 1987; Hope et al., 1988). This result indicates that a specific structure in the acidic domain is required for function. Other activators (so far found only in mammals) have nonacidic domains, including the glutaminerich activation domain of Spl (Courey et al., 1989) and the proline-rich CTFINF-I (Mermod et al., 1989). In some cases, certain residues around the zinc finger of the DNA binding domain may play a more direct role in the activation of transcription (Kim and Guarente, 1989; Schena et al., 1989). In still other cases, the activator is a heteromeric complex containing several protein subunits (Olesen et al., 1987; Gentz et al., 1989; Landshulz et al., 1989; Turner and Tjian, 1989). In some heteromeric complexes the acidic activation domain exists as a discrete polypeptide chain. Examples of these proteins are the yeast HAP4 (Forsburg and Guarente, 1989) and the herpes simplex virus VP16 (Triezenberg et al., 1988b), which associate with DNA binding proteins in the nucleus and are thereby directed to particular promoters. The basic transcription factors are remarkably conserved in eukaryotes. The sequences of the large subunit of RNA polymerase II from human cells and yeast show considerable similarity (Sweetser et al., 1987) including a highly repeated sequence at the carboxyl tail (Allison et al., 1985; Corden et al., 1985). The yeast TATA box binding protein TFIID was actually identified and purified because it can substitute for human TFIID in the mammalian in vitro transcription reaction (Buratowski et al., 1988). Recently, TFIID from yeast and other eukaryotes has been cloned (Hahn et al., 1989; Horikoshi et al., 1989; Cavallini et al., 1989; Schmidt et al., 1989; Fikes et al., 1990; Hoey et al., 1990). Much of the primary structure of this transcription factor is highly conserved in divergent organisms. In addition to this conservation of specific factors, the mechanism of transcriptional activation itself appears to be conserved. For example, the yeast activator GAL4 functions in mammalian (Kakidani and Ptashne, 1988;

Cell 1200

TEMPLATES

TATA I WC, PROMOTER

dA:dTu CadT “AS

LAC I.Z

b

TATA

basal

WC1 PROMOTER

LAC I.Z

ACTIVATORS

GAL4

i/i’ 147

I

413

GAL4-VP16 GAL4 DNA-

Figure 1. DNA Templates and Activator

Proteins

(Top) Structure of the three DNA templates used in the in vitro transcription reactions. The gal.3 template has three tandem copies of an oligomer encoding a consensus binding site for the yeast activator GAL4. The dA:dT template has a binding site for an activator that is present in the yeast nuclear extract. The basal template has no known activator binding sites. All templates have multiple TATA box sequences and multiple transcriptional initiation sites. The CYC7 gene is fused at nucleotide +75 to the bacterial /e&Z genes. (Bottom) Structure of the activator GAL4-VP16 and its mutant derivatives. GAL4VPI6 contains the DNA binding domain of the yeast activator GAL4 (residues 1-147) fused to the acidic activation domain of the herpes activator VP16 (residues 413-490) via a 7 amino acid linker (Chasman et al., 1989). The protein designated GAL4 contains only residues 1-147 of the GAL4 DNA binding domain (Lin et al., 1988). Derivative de1456 is the same as GAL4VP16. except the carboxy terminal 35 residues have been deleted (Triezenberg et al., 1988a). FP442 is the same as de1458 except for a single amino acid substitution from phenylalanine to proline at position 442 (W. D. C. and S. J. T., unpublished data).

Webster et al., 1988) plant (Ma et al., 1988) or Drosophila (Fischer et al., 1988) cells. Similarly, mammalian steroid hormone receptors activate transcription in yeast (Metzger et al., 1988; Schena and Yamamoto, 1988). What is the mechanism by which these transcriptional activators function over a distance to stimulate the basic machinery? One model (reviewed by Ptashne, 1986,1988) is that activation surfaces contact a target factor of the basic machinery, thereby looping out the intervening DNA. This contact would activate transcription by stabilizing the binding of the target factor to the promoter, or by converting the conformation of the bound factor to an active form. Consistent with this notion are in vitro experiments which indicated an effect of upstream activators on the DNAase I footprint around the TATA box (Horikoshi et al., 1988a). Also, in vivo evidence supports this model that activators stimulate a target factor in the transcriptional machinery. High level expression of the herpes simplex virus activator VP16 in animal cells (Triezenberg et al., 1988a) or of GAL4

in yeast (Gill and Ptashne, 1988) inhibits normal transcription by RNA polymerase II. It has been suggested that the target factor has been bound in solution by these high levels of the activator, and sequestered from the promoter. This interaction in solution is thought to reflect the actual association between the activation domain and the target. An alternative proposal is that activators function by antagonizing histones, which otherwise wrap up the promoter DNA into a repressed chromatin state (Han and Grunstein, 1989). In this view, activators would not directly activate a target factor in the transcriptional machinery. A finding consistent with this model is that in vivo depletion or overexpression of histones in yeast can activate transcription, or alter the start site, of certain genes (Han and Grunstein, 1989; Clark-Adams et al., 1988). This activation bypasses the normal requirement for the UAS. Also, histones can repress transcription in vitro by competing with an activator for binding to the promoter region (Workman et al., 1988). In this case the activator would prevent repression, rather than actually stimulating transcription. A directly positive role for activators is suggested by the fact that transcriptional stimulation can be achieved in vitro in mammalian (Horikoshi et al., 1988b) or yeast extracts (Chasman et al., 1989). However, because the extracts are crude fractions, a role of chromatin in the process cannot be excluded. Yeast transcription in vitro responds strongly to activation by the chimeric activator GAL4-VP16 (Chasman et al., 1989). Recent experiments showed that excessive levels of GAL4-VP16 can inhibit transcription in the yeast in vitro system (Kelleher et al., 1990) and, as discussed above, it has been suggested that inhibition results from titration of the target factor in the machinery. In this study, we examined the response to GAL4-VP16 in the yeast in vitro system in greater detail. Our findings show that inhibition and activation use the same surface of VP16, since both processes are affected similarly by mutations in this domain. We believe these results indicate a direct and positive role of the activator in stimulation of transcription. Moreover, we devised a means to use the activation domain of VP16 to specifically block trancriptional activation by another activator at a promoter bearing binding sites for that protein. Strikingly, under these conditions, basal transcription from the TATA box is not affected. We propose that these conditions selectively deplete a novel class of molecules, termed adaptors, that couple transcriptional activators to the basic machinery. Results To study mechanisms involved in transcriptional activation, we used an in vitro transcription assay in yeast nuclear extracts (Lue and Kornberg, 1987). Transcription was analyzed from the three DNA templates depicted in Figure 1. These templates have identical proximal promoters (the yeast CYC7 promoter, which contains multiple TATA consensus sequences and multiple start sites) and identical transcribed sequences (the first 80 bp of the yeast CYCl gene fused to the bacterial lad-Z genes). The difference between the three templates is the UAS located 65 nucleo-

In Vitro Transcriptional 1 2!01

Inhibition

A

6

C SC

0

0.4

1

2

4

10

loo

200

300

365

gal.3 Fogure 2. Activation

and Inhibition of Transcription

0

10

4

2

1

10

LIN ‘0

pmoleGAL4v~l6

dA:dT by GAL4-VP16

GAL4VP16 was added to the in vitro reaction in the amounts (in pmol) indicated at the top of each lane. The DNA templates used were (A) gal.3 (R) dA:dT, and (C) basal. Supercoiled (SC) DNA templates were used except in (C) in the last two lanes, where the basal template was linearized (LIN) using Xhol, which cleaves just upstream of the TATA boxes, The gels were autoradiographed for (A) 6 hr, (B) 18 hr, and (C) 18 hr.

tides 5’ of the transcriptional initiation site. The gal.3 template has a tandem repeat of three consensus binding sites for the yeast GAL4 transcriptional activator protein. The dA:dT template has a binding site for the yeast dA:dT transcriptional activator protein present in the nuclear extract (Lue et al., 1989). Although the dA:dT activator has not yet been purified or cloned, its existence was inferred because adding an oligonucleotide encoding the dA:dT 1JAS to an in vitro reaction reduced transcription from the dA:dT template. The basal template (Figure 1) contains no known binding sites for transcriptional activators upstream of the CyC7 promoter. Following the in vitro transcription reaction, which was performed essentially as described by Lue and Kornberg (1987) RNA was purified and analyzed by primer extension from a radiolabeled oligonucleotide probe complementary to the lad gene adjacent to the CYClllacl-Z fusion junction. The 5’ends of the primer extension products from the in vitro transcription reaction matched those synthesized in vivo (not shown). However, the predominant 5’ start site used in vitro was not the most frequent !Y end in vivo (Hahn et al., 1985); this difference has also been noted by others using the CyCi promoter in the yeast in vitro transcription reaction (Lue and Kornberg, 1987). GAL4-VP16 Strongly Activates and inhibits Transcription In Vitro ‘The chimeric activator protein GAL4VP16 (Sadowski et

Selective inhibition of activated but not basal transcription by the acidic activation domain of VP16: evidence for transcriptional adaptors.

The interaction between the chimeric activator GAL4-VP16, consisting of the DNA binding domain of GAL4 and the acidic activation domain of VP16, and i...
2MB Sizes 0 Downloads 0 Views