Estrogen Receptor Involves Stabilized Interactions with a. Limiting Target ... synergistically with other activators recognizing the same promoter. ...... Breathnach, R., and B. A. Harris. 1983. .... Lees, J. A., S. E. Fawell, and M. G. Parker. 1989.
MOLECULAR AND CELLULAR BIOLOGY, June 1991, p. 2937-2945
Vol. 11, No. 6
0270-7306/91/062937-09$02.00/0 Copyright © 1991, American Society for Microbiology
Synergistic Transcriptional Activation by CTF/NF-I and the Estrogen Receptor Involves Stabilized Interactions with a Limiting Target Factor ERNEST MARTINEZ,t YVES DUSSERRE, WALTER WAHLI, AND NICOLAS MERMOD* Institut de Biologie Animale, Batiment de Biologie, Universite de Lausanne, CH-1015 Lausanne, Switzerland Received 19 December 1990/Accepted 26 February 1991
Transcription initiation at eukaryotic protein-coding gene promoters is regulated by a complex interplay of site-specific DNA-binding proteins acting synergistically or antagonistically. Here, we have analyzed the mechanisms of synergistic transcriptional activation between members of the CCAAT-binding transcription factor/nuclear factor I (CTF/NF-I) family and the estrogen receptor. By using cotransfection experiments with HeLa cells, we show that the proline-rich transcriptional activation domain of CTF-1, when fused to the GAL4 DNA-binding domain, synergizes with each of the two estrogen receptor-activating regions. Cooperative DNA binding between the GAL4-CTF-1 fusion and the estrogen receptor does not occur in vitro, and in vivo competition experiments demonstrate that both activators can be specifically inhibited by the overexpression of a proline-rich competitor, indicating that a common limiting factor is mediating their transcriptional activation functions. Furthermore, the two activators functioning synergistically are much more resistant to competition than either factor alone, suggesting that synergism between CTF-1 and the estrogen receptor is the result of a stronger tethering of the limiting target factor(s) to the two promoter-bound activators.
Transcription of eukaryotic protein-coding genes is controlled by the combinatorial arrangement of regulatory proteins that bind specific DNA elements located either upstream or downstream of a core promoter. Structural and functional analyses have revealed that sequence-specific activators are modular in structure. They most often contain at least two generally independent functional regions: a DNA-binding domain and a transcriptional activation domain, in addition to oligomerization and nuclear localization determinants. Four main distinct types of activation motifs, tentatively grouped according to their different amino acid compositions, have been identified thus far: the acidic negatively charged domains, the glutamine-rich domains, the proline-rich domains (34, 37), and the metal-binding cysteine-containing activating domain of the adenovirus Ela protein (29). Activation domains are thought to function by interacting directly or indirectly with general components of the transcription initiation machinery (26, 38). This hypothesis is supported by the observation that an activator present at an artificially high intracellular concentration can inhibit its own transcriptional activity as well as that of other activators; this inhibition was previously called squelching or transcriptional interference (12, 29, 33, 46, 48). Squelching is thought to occur by the titration of a limiting soluble cellular component of the transcriptional machinery (the target) through its interaction with the activation domain of the overexpressed activator. Thus, squelching may result in the sequestration of the target, which can no longer interact with promoter-bound activators to mediate their transcriptional activation functions (37, 38). A fundamental property of many sequence-specific DNAbinding activators is their ability to activate transcription
synergistically with other activators recognizing the same promoter. Thus, the increased rate of transcription initiation conferred on a target promoter by two activators functioning simultaneously is higher than the sum of the transcription activities conferred by each activator acting alone. In some cases, cooperative DNA recognition has been implicated in synergistic transcriptional activation in higher eukaryotes (10, 24, 31, 36, 44, 49). However, transcriptional synergism has also been shown to occur in vitro independently of cooperative DNA binding (6, 28), suggesting alternative mechanisms. For example, several activators functioning synergistically may simultaneously, and thus tightly, interact with a common target component of the general transcription machinery, therefore favoring the formation and/or activation of the transcription preinitiation complex. Alternatively, different activators could contact separate targets, and synergism would then be triggered by subsequent steps leading to the formation of an active transcription initiation complex. The analysis of such processes is a prerequisite to understanding the combinatorial regulation of gene transcription by various activators. In this article, we have analyzed the mechanisms involved in the synergistic activation of transcription by two activators: the estrogen receptor and CTF-1, a member of the CCAAT-binding transcription factor/nuclear factor I (CTF/ NF-I) family of proteins. We show, by using GAL4-CTF-1 fusion activators, that the proline-rich transcriptional activation domain of CTF-1 is the main determinant of synergism with either activating region of the estrogen receptor. These two trans-activators do not bind cooperatively to the reporter promoter in vitro, suggesting that another mechanism is involved. Indeed, our in vivo competition (squelching) experiments suggest that the synergistic transcriptional activity conferred by the CTF-1 fusion protein and the estrogen receptor activation domains results from the stronger tethering of a common target transcription factor.
* Corresponding author. t Present address: Laboratory of Biochemistry and Molecular Biology, The Rockefeller University, New York, NY 10021.
2937
2938
MARTINEZ ET AL.
MATERIALS AND METHODS Plasmid constructions. pGlBCAT consists of a synthetic 17-mer binding site for GAL4 protein inserted upstream of the adenovirus Elb TATA box and chloramphenicol acetyltransferase (CAT) gene (cat), while pG5BCAT contains five such GAL4 recognition sites (27) (kindly provided by J. Lillie). The reporter construct pGlBCAT-B-ERE results from the insertion of a BglII oligonucleotide linker into the Klenow-repaired HindIII site of pGlBCAT, followed by the insertion of a synthetic estrogen receptor-binding site consensus sequence (ERE) (30, 51) into the BglII site. The expression vectors pKCR2 (4), HEO (15), HE12, HE15, HE19, and HE21 (22) were kindly provided by P. Chambon. pSG424 (41) and pSGVP (40) were a kind gift from I. Sadowski. pSG424 and pSGVP were used to express the unfused GAL4 N-terminal 147 amino acids (Gal) or the fusion of this GAL4 sequence to the negatively charged activator of VP16 (Gal VP16), respectively. The Gal 399-499, Gal 399-437, and Gal 399-430 expression vectors were constructed by inserting the Sall and Klenow-repaired BglII small fragments of pCTF-1, pCTF-lAC437, and pCTF-2 (32, 42), respectively, into pSG424 cleaved with Sall and BamHI, with the BamHI site filled in with Klenow enzyme. The pT7-CTF-1 (32) small SmaI-XbaI fragment was inserted into pSG424 cut with SmaI and XbaI, resulting in the Gal 438-499 expression vector. The Gal 1-499 and Gal 220-499 expression vectors result from the insertion of the 1.8-kb SalI-Klenow-repaired Spel fragment or the 1-kb SaclEcoRV fragment of pCTF-1 into pSG424 cleaved with Sall and BamHI or with Sacl and XbaI, respectively, with the BamHI and XbaI sites filled in with Klenow enzyme. The conservation of the proper reading frames in fusion constructions was ascertained by DNA sequencing. The p113 expression vector (a kind gift of V. Baichwal) consists of the pBluescriptKS+ (Stratagene, Inc.) vector into which the 578-bp NdeI-HindIII fragment of pRSVCAT (13) was inserted into the EagI and NdeI sites, followed by insertion into the resulting plasmid, cleaved with EcoRI and SalI, of the pRSVCAT 1.3-kb EcoRI-BamHI fragment, with all sites filled in with Klenow enzyme. The orientation of the inserted fragments was such that transcription from the Rous sarcoma virus long terminal repeat promoter is directed toward the pBluescript polylinker, cat gene, and simian virus 40 polyadenylation sequences. The competitor expression vectors p113-CTF-1 and p113-CTF-1A were constructed by inserting the EcoRV-BamHI partially digested 1.8-kb fragment of pCTF-1 or the BamHI-HindIII 1.3-kb fragment of pCTF-1AC399 (32), respectively, into the p113 3.6-kb fragment generated by HpaI and BamHI cleavage. p113-Spl and p113-VP16 were generated by excision of the CTF-1 sequence of p113-CTF-1 by BamHI cleavage and its replacement by either the BamHI partial cleavage 2.1-kb fragment of pPADH-SP1 (9) or the 1.65-kb BamHI fragment of pVP16C+119bp (48) (kindly provided by S. Triezenberg), respectively. Cell culture, transfection, and CAT assays. HeLa cells cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum were plated at 106 cells per 10-cm dish (in 10 ml of medium) the day before transfection. Just before transfection, 0.5 ml of 1 M HEPES (N-2hydroxyethylpiperazine-N'-2-ethanesulfonic acid, pH 7.2) was added to the cells. Transfection was performed by the calcium phosphate technique (52) with, unless otherwise indicated, 4 ,ug of either the pGlBCAT-B-ERE or the pGSBCAT reporter plasmid and 0.2 ,g of activator expres-
MOL. CELL. BIOL.
sion vector per dish. Carrier DNA (sonicated salmon sperm DNA or pBluescriptSK+ plasmid) was added to a total of 20 ,ug per dish. For the competition experiments, 1.4 or 3.5 pmol of competitor expression plasmid was added per dish, which corresponds to 5 or 12.5 ,ug of p113-CTF-1A, 5.6 or 14 ,ug of p113-CTF-1, 7.6 or 15.6 ,ug of p113-SP1, or 7.3 or 14.6 ,ug of p113-VP16. Cells were incubated with the DNA precipitate for 4 to 6 h, the medium was removed, and fresh medium, containing 10-8 M 17p-estradiol or no hormone, was added to the cells. After 40 to 48 h, total cellular extracts were made and CAT assays were performed as described previously (14). CAT activity was normalized to the protein concentration in cellular extracts and quantitated by liquid scintillation counting. The results represent the mean and standard deviation of four independent experiments. HeLa WCE and DNA-binding analyses. HeLa cells were transfected essentially as described above with 10 to 13 ,ug of the activator expression vector (HEO or Gal399-499) or the control vector (pKCR2). Transfected cells were left for a total of 37 h, and 10-8 M 17,B-estradiol was added 13 h before cell extracts were prepared. Whole-cell extracts (WCE) were prepared from transfected cells as previously described (21). The ERE 17-mer probe used in gel mobility shift assays consists of a 95-bp HaeIII-XbaI restriction fragment of the reporter pG1BCAT-B-ERE plasmid that was labeled at the XbaI end with the Klenow enzyme in the presence of [ao-32P]dATP and the four cold deoxynucleoside triphosphates. Binding reaction mixes contained 13 to 15 ptg of WCE protein, 4 ,ug of poly(dI-dC), 10-8 M 17p-estradiol, 10 ,M ZnSO4, 10 mM Tris-HCl (pH 7.5), 100 mM KCl, 10% glycerol, and 1 mM dithiothreitol in a final volume of 10 ,l. All components for the binding reaction, except the probe, were incubated for 15 min on ice before 1 RI of the radiolabeled DNA probe (-15,000 cpm; 10 fmol) was added, and the binding reaction mix was further incubated for 15 min at 20 to 23°C. For the mixing experiment shown in Fig. 4, a first binding reaction with 13 ,ug of HEO-containing WCE proteins was performed as described above. Then increasing amounts of Gal 399-499-containing WCE or a negative control extract, preincubated in parallel for 15 min on ice but without the probe, was added. The final protein amount was maintained at 28 ,ug by addition of the negative control WCE. The resulting 20-RI reaction mixes were further incubated for 15 min at 20 to 23°C. Just before gel loading, the final KCI concentration was adjusted to 200 mM by adding salt-containing gel loading buffer to the reaction mixes. The native 5% polyacrylamide gel was run at 25 mA at 4°C with buffer (6.7 mM Tris-HCl [pH 7.5], 3.3 mM sodium acetate, 1 mM EDTA) circulation. RESULTS Synergistic transcriptional activation by the CTF-1 prolinerich activating domain and the estrogen receptor. CTFINF-I has been shown previously to activate transcription in synergy with the estrogen receptor (7, 8). To approach the molecular mechanisms involved in synergistic activation by these two trans activators, we first assessed whether the transcriptional activation domain of CTF-1 (32, 42) is sufficient to synergize with the estrogen receptor. Since mammalian cells usually already contain endogenous CTF/NF-I activity, we altered the DNA-binding specificities of CTF derivatives. We constructed a chimeric cDNA expression vector that encodes a fusion polypeptide containing the proline-rich activation domain of CTF-1 (amino acids 399 to 499) (32), fused to the DNA-binding and dimerization do-
VOL . 1 l,
TRANSCRIPTIONAL SYNERGISM VIA STABILIZED TARGET
1991
A 0,-
0 0-
u
la
1 Gal 399-499
2
.2r1 5 3 4
0.02 0.2
2
0.2
I
6
transcription activities obtained with each activator acting alone (compare lanes 2 to 4 and 8 to 10 with lanes 5 to 7). *- Estrogen * + Estrogen
7
8
9
10
. 12 13 14
0.2 0.2
0.02 0.2
HEO
2
0.02 0.2
2
0.02 0.2 2
0.2 0.2 0.2
Gal
2
pKCR2
B Reporter
CAT /p
Z ERE
17 MER
2939
EIB TATA box
FIG. 1. CTF-1 proline-rich domain activates traniscription synergistically with the human estrogen receptor. (A) Thie pGlBCAT-BERE reporter plasmid was transfected into HeLa celIls alone (lane 1) or together with the indicated amounts (in microgranns) of the cDNA expression vectors encoding either the chimeric p olypeptide consisting of the GAL4 DNA-binding domain fused proline-rich domain (CTF-1 amino acid positions:399 to 499 [Gal (HEO), the 399-499]; see also Fig. 2B), the human estrogen rece GAL4 DNA-binding domain alone (Gal), or a conitrol expression vector without any cDNA insert (pKCR2). Extract :s of transfected HeLa cells were made and CAT activity assays weire performed as described in the Materials and Methods section. CA]r activities from cells cultured in the absence (open bars) or presencte (hatched bars) of estrogen are indicated as fold stimulation over th e basal level of activity measured in cells transfected with the r only. (B) Structure of the pGlBCAT-B-ERE reportei a synthetic bacterial cat coding sequence is under the control promoter consisting of the adenovirus Elb gene TATA box, a consensus GAL4-binding site (17-mer), and the 1 3-bp consensus estrogen response element (ERE).
3ptor
rconstrluct.The rof
main of the yeast GAL4 activator (amino acid,s 1 to 147) (5). The resulting GAL4-CTF-1 fusion polypepti(de was called Gal 399-499. We then tested the capacity 4of this fusion protein to activate transcription in synergy wit;h the estrogen receptor by transiently cotransfecting variouis amounts of Gal 399-499 and human estrogen receptor (HE_0) expression vectors into HeLa cells together with a repoorter construct (Fig. 1A). The reporter construct contains the cat gene placed under the control of a synthetic promc ter composed of the adenovirus Elb TATA box and two upstream elements: a high-affinity estrogen receptor bindi ng site (ERE) adjacent to the 17-mer GAL4 binding site ( Fig. 1B). Gal 399-499 and the estrogen receptor by themselvtes are moderate activators, and as shown previously (15) the estrogen receptor activated transcription only in the presence of estrogen (Fig. 1A). However, these two act:ivators, when assayed together, yielded a strong estrogen-d4 ependent activation of the reporter promoter (70- to 270-foild activation), representing a 3- to 4-fold higher activity than the sum of the
This indicates that Gal 399-499 and the estrogen receptor activate transcription in a synergistic manner. In contrast, the DNA-binding domain of GAL4 (Gal), by itself, did not
activate transcription (lanes 11 to 13, in the absence of estrogen), nor did it act synergistically with the estrogen receptor (lanes 11 to 13, in the presence of estrogen). This demonstrates that the CTF-1 proline-rich activation domain, when tethered to the GAL4 heterologous DNA-binding domain, can synergize with the estrogen receptor. In order to further characterize the CTF-1 domain involved in the synergism with the estrogen receptor, the N-terminal and C-terminal portions of the CTF-1 prolinerich activating domain were fused to the GAL4 DNAbinding domain and analyzed as before. Transfection of various amounts of the resulting Gal fusion proteins, Gal 399-437 and Gal 438-499, activated transcription less efficiently (about fivefold less) than Gal 399-499, which contains the whole proline-rich activating domain of CTF-1 (Fig. 2A and data not shown). Interestingly, synergism was correspondingly reduced with either Gal 399-437 or Gal 438-499 compared with Gal 399-499, to levels indicating little or no synergism with the estrogen receptor (Fig. 2B). Thus, we conclude that maximal transcriptional activation and optimal synergism with the estrogen receptor require most of the CTF-1 activating domain. To assess the role of other CTF-1 regions in the synergistic activation with the estrogen receptor, we transcriptional tested Gal-CTF-1 fusion proteins containing additional CTF-1 sequences flanking the proline-rich activation domain. The addition of the central part of CTF-1 (Fig. 2B, Gal 220-499) did not increase synergism, although the expression of Gal 220-499 resulted in a stronger activation than expression of Gal 399-499 (Fig. 2A), possibly owing to the presence of additional proline residues flanking the 399 to 499 region. In addition, the expression of another Gal-CTF-1 fusion protein that contains the entire CTF-1 sequence, Gal 1-499, conferred a weaker activation than Gal 399-499 (about threefold weaker; Fig. 2A) and did not synergize with the estrogen receptor (Fig. 2B). The absence of synergism may be the result of a steric hindrance due to the large size of the Gal 1-499 fusion protein, since this particular fusion protein even repressed the transcriptional activation mediated by the receptor (Fig. 2A, compare lanes 2 and 12). Taken together, the above results identify the CTF-1 proline-rich activation domain as the protein region responsible for the synergistic transcriptional activation by CTF-1 and the estrogen receptor and further suggest that other CTF-1 regions do not contribute to this synergy. We further asked whether CTF-2, another member of the human CTF/NF-I family having a distinct proline-rich acti-
vation domain generated by alternative splicing (32, 42), could also work synergistically with the estrogen receptor. Therefore, we tested a GAL4-CTF-2 fusion protein (Gal 399-430, Fig. 2) which contains the proline-rich domain of CTF-2 fused to the GAL4 DNA-binding region. The expression of Gal 399-430 activated transcription about sevenfold less efficiently than that of Gal 399-499 (Fig. 2A). Moreover, in contrast to the proline-rich domain of CTF-1, the activation domain of CTF-2 did not significantly synergize with the estrogen receptor (Fig. 2B). This most likely does not stem from poor expression of Gal 399-430, since similar results were obtained from the transfection of larger amounts of
expression vectors (data not shown). Both activating regions of the estrogen receptor synergize
2940
MARTINEZ ET AL.
MOL. CELL. BIOL.
A iso1 -
a c
U U
c
ioo0
*
19 *-1
;0*-
01
t)
*>
50 :p
t
1 2 3 4
HEO
+
-
Gal fusions:
-
5
+
-
10 11 12 13 14 + - + +
6 7 8 9 +
-
+
20-499 1-499 399-430
399.499 399.437 438-499 21
-
B
ACTIVATION
DNA BINDING DIMERIZATION
CTF-l
1
C TF-2
1
_499 lstvovovt\tvvtvOvvtvvs
E,
430
Zs
SYNERGISM 399
Gal 399-499
N
Gal
499
43
39_437
Gal 399-437
Gal 4
Gal 438-499
Gal 4
Gal 220-499
Gal 4
Gal 1-499
Gal4
Gal 399-430
Gal4
~ 438
220
3x 1.3 x
* 499 1.5 U x 499 3 x
x
_9
0.4 x
: 430 1.4 X
FIG. 2. Integrity of the CTF-1 proline-ricch domain is important for both maximal transcriptional activation aand synergism with the estrogen receptor. (A) Relative CAT activitie s measured from HeLa cells transfected with the reporter construct p)GlBCAT-B-ERE (Fig. 1B) and chimeric expression vectors encodin g various GAL4-CTF-1 or GAL4-CTF-2 fusion proteins (Gal fusion: s), with (+) or without (-) the estrogen receptor expression vectoi HEO. The Gal fusion polypeptides consist of the GAL4 DNA-bindling domain tethered to CTF-1 sequences as indicated by their aminc acid positions (lanes 3 to 12) or fused to the 31 C-terminal amino ac basal and 14), as depicted in part B. CAT activity i: nduction cellsw^&^-^ transfected with level was normalized to that obtained from cel trnfce ith both the Gal 399-499 and HEO expression ve ctors, which was given an arbitrary value of 100 (lane 4). Transfe,cted HeLa cells were cultured in the presence of estrogen and prc cessed as described in the legend to Fig. 1. (B) Structure and fun ctional domains of the human CTF-1 and CTF-2 proteins and of the GAL4-CTF fusion proteins. The locations of the DNA-bindingg, oligomerization, and transcriptional activation domains of the CTF,-1 and CTF-2 polypeptides are shown as shaded boxes at the top of the figure (32). The C-terminal portion of the CTF-2 activation domain, depicted by a black box, indicates the only region whern e the CTF-2 sequence diverges from that of CTF-1, which resul ts from an alternative splicing event of the precursor mRNA enc oding both CTF-1 and CTF-2 (42). The open boxes of the fusiorn proteins labeled Gal
overCthe
--v--T
with the proline-rich domain of CTF-1. The human estrogen receptor contains two distinct transcription activating regions, called TAF-1 and TAF-2; TAF-1 is a constitutive activation domain, whereas TAF-2, which colocalizes with the hormone-binding region, functions in a hormone-dependent manner (25, 47). In order to test whether TAF-1 or TAF-2 or both are involved in synergistic transcriptional activation with the proline-rich domain of CTF-1, estrogen receptor deletion mutants (22, 23) were tested for their ability to synergize with Gal 399-499, as described in the preceding section. The data presented in Fig. 3 show that receptor mutants lacking either TAF-1 (HE19) or TAF-2 (HE21 and HE15) all activated transcription in synergy with Gal 399-499. However, only HE19, which retains TAF-2, synergized in a hormone-dependent manner, in contrast to HE21 and HE15, which retain only TAF-1 and synergized constitutively with Gal 399-499. Since all three deletion mutants (HE19, HE21, and HE15) contain sequences of the D region (Fig. 3B), we tested the importance of these sequences by using an internal deletion mutant which lacks most ofthe Dregion (HE12). As shown in Fig. 3A, HE12 also had the capacity to synergize with Gal 399-499, and as expected, this synergism was hormone dependent, since the hormone-binding TAF-2 regulatory domain was present. Thus, we conclude that region D is not necessary for the observed synergistic action of the estrogen receptor. These experiments do not exclude the possibility that the DNAbinding domain (region C in Fig. 3B), present in all mutants, contributes to the observed synergism. However, this is unlikely since previous experiments have demonstrated that the estrogen receptor DNA-binding domain by itself cannot activate transcription from different promoters (22). Altogether, our results strongly suggest that both TAF-1 and TAF-2 can synergize with the proline-rich transcription activating domain of CTF-1. Estrogen receptor and Gal 399-499 do not cooperate for DNA binding. Previous experiments have indicated that steroid hormone receptors activating transcription synergistically bind cooperatively in vitro to their recognition sites (19, 31, 44, 49). Therefore, we used a similar assay to address the possibility that the estrogen receptor and Gal 399-499 bind cooperatively to the promoter used to demonstrate their functional synergism. WCE from transfected HeLa cells expressing either the estrogen receptor (HEO) or Gal 399-499 were used in combination with a reporter promoter fragment containing the ERE adjacent to the 17-mer GAL4 binding site in a gel mobility shift assay (11). This allowed the detection of specific protein-DNA complexes formed on the ERE-17-mer probe with the estrogen receptor- and Gal 399-499-containing extracts (Fig. 4A) (31). If cooperative binding occurs, one would expect Gal 399-499 to bind preferentially a receptor-bound probe (HEO-bound probe in Fig. 4A) rather than to the protein-free probe, whereas a noncooperative DNA-binding mechanism implies 4 represent the GAL4 DNA-binding domain sequence (amino acids 1 to 147) fused at the C-terminal end with various portions of CTF-1, as indicated by amino acid positions, except for Gal 399-430, which results from the fusion of CTF-2 sequences between amino acid positions 399 and 430. The numbers to the right of the Gal fusion polypeptides indicate the synergism between each Gal fusion and the estrogen receptor as the ratio of CAT activity induction mediated by HEO and a particular Gal fusion polypeptide acting together over the sum of CAT activity induction mediated by each activator acting alone, as calculated from the results presented in part A.
VOL.
TRANSCRIPTIONAL SYNERGISM VIA STABILIZED TARGET
11, 1991
A
10117 Estrogen
Estrogen
*
10
i
-
St
-S 10b
-
I=
1
2
3
Gal399-499-+ Receptor
B
4
-
+
HEO
TAF-1
I
180 A/B
A
263 302
C
I
6
TAF-2 HORMONE
DNA 1
i
7 8 9 10 11 12 -- + - + - + + HE19 HE21 HE1S HE12
5
IDl
SYNERGISM 595
E/F
4x c HEO N iln 8x HE19 _c 4x 341 HE21 N _ 3x 282 HE15 N 4x HE12 N _ 271V300 FIG. 3. Estrogen receptor mutants truncated in either transcriptional activation region synergize with the CTF-1 proline-rich domain. (A) Relative induction of pG1BCAT-B-ERE reporter cat gene expression by various estrogen receptor mutants (depicted in part B) in the absence (-) or presence (+) of Gal 399-499 was tested by the transfection of HeLa cells as described in the legend to Fig. 1. CAT activity inductions were normalized to that obtained from the transfection of estrogen-stimulated HeLa cells with the HEO, Gal 399-499, and reporter construct (lane 4, hatched bar). The control expression vector pKCR2, having no cDNA insert, was cotransfected in the experiments of lanes 1 and 2. (B) Structural and functional features of the receptor mutants tested in part A. The locations of TAF-1 and TAF-2, corresponding to the two transcription-activating regions of the estrogen receptor, and the DNA- and hormone-binding domains are indicated (22, 23, 46). Endpoints of the deletions in each mutant polypeptide are shown by the corresponding amino acid positions. The synergism factors to the right of each of the full-length and truncated receptors were calculated as described in the legend to Fig. 2B from the CAT activity inductions in estrogen-stimulated cells presented in part A. C
that Gal 399-499 should bind with the same affinity to both probes. Therefore, the probe was first incubated with the estrogen receptor-containing extract in conditions that give about 50% of the receptor-DNA complex formation obtained at equilibrium (lane 3), and then increasing amounts of Gal 399-499-containing extract were added (lanes 4 to 6). The addition of the Gal 399-499-containing extract resulted in the formation of two additional specific complexes, Cl and C2, that resulted from the binding of Gal 399-499 to the 17-mer site on the protein-free probe and on the estrogen receptorbound probe, respectively. The quantitation of Cl and C2 complex formation as a function of added Gal 399-499containing extract revealed that both complexes were formed with the same efficiency (Fig. 4B). We conclude that Gal 399-499 has the same affinity for the free probe as for the
2941
estrogen receptor-bound probe, which demonstrates that Gal 399-499 does not bind cooperatively with the estrogen receptor to its target site in vitro. This suggests that the in vivo synergism between Gal 399-499 and the estrogen receptor is not the result of a direct interaction between the two activators leading to cooperative DNA binding. Synergism between Gal 399-499 and the estrogen receptor involves a stabilized interaction with their target factor. We next tested whether the estrogen receptor and Gal 399-499 synergize through their interactions with a target factor(s) in vivo, which would mediate their transcriptional activation functions. To address this issue, we first determined whether a proline-rich activator can be inhibited by homologous or heterologous activation motifs acting as competitors. In these experiments, HeLa cells were transiently cotransfected with the cat reporter construct pG5BCAT containing five GAL4 binding sites upstream of the adenovirus Elb gene TATA box and cat sequences (27), together with either Gal 399-499 or the Gal VP16 activator, a GAL4(1-147) fusion with the negatively charged activation domain of the herpes simplex virus VP16 protein (40). As competitors, we used different constructs expressing either the intact human CTF-1 protein, a truncated CTF-1 derivative (CTF-1A) lacking its proline-rich transcriptional activation domain, or intact VP16 and Spl activator proteins. Previous transfection experiments have indicated that the truncated CTF-1A derivative localizes properly in the nuclei of transfected cells and is stably expressed in amounts similar to full-length CTF-1 (32). Since these proteins have no binding sites on the reporter promoter, they should act as pure competitors. As shown in Fig. 5A, the expression of CTF-1 but not that of the truncated CTF-1A efficiently inhibited Gal 399-499mediated activation, while in contrast the expression of the glutamine-rich Spl heterologous competitor had little effect on Gal 399-499 activity. The inhibition of Gal 399-499 transcriptional activity by CTF-1 did not depend on the five GAL4 recognition sites of the pG5BCAT reporter promoter, because similar results were obtained with a reporter construct containing only one GAL4 site (pGlBCAT-B-ERE; Fig. 6A). Moreover, the CTF-1 inhibition effect was specific for Gal 399-499-mediated activation, since CTF-1 overexpression did not repress either the activation mediated by the heterologous Gal VP16 activator (Fig. 5B) or the basal level of expression in the absence of any activator (data not shown). Thus, the CTF-1 competitor effect does not result from nonspecific inhibition of the activator's expression, such as the repression of the promoter used for both the Gal VP16 and Gal 399-499 expression vectors. Therefore, these results suggest that the CTF-1 repression of Gal 399-499 activity indeed results from the competition for a limiting target factor, but that it does not stem from a nonspecific inhibitory effect, such as cellular toxicity, or the inhibition of a component of the basal transcription machinery. To assess whether the estrogen receptor trans-activation functions are mediated by the same limiting target factor as for CTF-1, we next assayed transcriptional activation by the estrogen receptor (HEO) alone and in the presence of the competitor. As reported in Fig. 6C, the activation domaindeleted mutant CTF-1A did not prevent HEO-dependent transcriptional activation, while in contrast, intact CTF-1 strongly inhibited estrogen receptor-dependent activation. These results therefore indicate that the limiting target transcription factor(s) that mediates the transcriptional activation properties of CTF-1, possibly by directly or indirectly interacting with the CTF-1 proline-rich activating domain,
2942
MOL. CELL. BIOL.
MARTINEZ ET AL.
B
A HEO C,,a] 399-499
0--I 113
13
* Cl versus Free probe o C2 versus IFIEO-bound probe
J15J1 I_ 511L5 -
z 9
.
..
9 ,, -L,,,,
,, -..
.4
C2
iE )meV R ,..W 0 - 4
HEO-bound probe
c-
11
gmi -
.
n z Z)
A ...
.1-1.
(ZN
CN
C. "ps
R99-4CJ
E
Free probe _-*-
ER
6.
l7mer
5
10
15
gg Gal 399-499-containing extract FIG. 4. Gal 399-499 and the estrogen receptor do not cooperate for specific promoter binding in vitro. (A) The ERE-17-mer probe, generated from the pG1BCAT-B-ERE plasmid, and WCE from HeLa cells transfected with 10 ,ug of either the control vector pKCR2 (lane 2), Gal 399-499 (lanes 1 and 4 to 6), or HEO (lanes 3 to 6) were used in a gel mobility shift assay. In lanes 3 to 6, a constant amount of HEO-containing extract was preincubated with the probe to equilibrium, followed by the addition of increasing amounts of Gal 399-499-containing extract (lanes 4 to 6). The amount of HEO- and Gal 399-499-containing extract used in each reaction is shown in micrograms of total protein at the top of the figure. Each reaction mix was adjusted to 28 ,ug of total protein by the addition of a control extract from cells transfected with pKCR2. The specific complexes generated by the binding of the estrogen receptor (HEO-bound probe) and by Gal 399-499 (Cl), as well as the complex resulting from both proteins (C2), are shown by arrows and are represented schematically at the right-hand side. (B) The complexes shown in part A were cut out from the gel, and radioactivity was quantitated by liquid scintillation counting. The graph represents the percentage of Gal 399-499-occupied probe versus free probe (C1/[C1 + free probe]) and the percentage of the probe bound by both Gal 399-499 and HEO relative to the HEO-bound probe (C2/[C2 + HEO-bound probe]) as a function of the amount (micrograms of total protein) of added Gal 399-499-containing extract. 1 2 3 4 5 6
also mediates most of the transcription-activating functions of the estrogen receptor. Next, we compared the extent of transcriptional inhibition by an excess of CTF-1 or CTF-1A competitor in the presence of either one activator or both. The data presented in Fig. 6B show that, as for each activator alone, CTF-1A did not inhibit synergistic transcriptional activation by Gal 399-499 and the estrogen receptor (HEO) acting together. However, in contrast to the strong CTF-1-mediated inhibition of each activator alone (Fig. 6A and C), CTF-1 only slightly reduced activation when both activators were present simultaneously (Fig. 6B). For instance, an amount of CTF-1 expression vector corresponding to a 25-fold molar excess over Gal 399-499 (3.5 pmol of the CTF-1 expression vector) totally abolished activation by Gal 399-499 (Fig. 6A), and activation by HEO alone was reduced by over 80% (Fig. 6C), whereas activation mediated by both Gal 399-499 and HEO together was reduced by less than 50% (Fig. 6B). Note that the open bars in Fig. 6B represent the expected residual transcriptional activities resulting from both activators acting together in the presence of the CTF-1 competitor, as calculated from the inhibitory effect of the competitor on each factor acting alone. This demonstrates that the two activators, when acting in synergy, are much more resistant to competition than either activator alone. Consistently, in the presence of the CTF-1 competitor, a further threefold increase of the apparent synergism occurred.
A
D
B
100
I-
*10
50
0
pmole competitor: competitors:
activators:
-
-
1.4
3.5
1.4
3.5
CTF-IA CTF-1
Gal 399-499
1.4
3.5
Spl
-
1.4 3.5 1.4 3.5 1.4 3.5
-
CTF-IA CTF-1
VP16
Gal VP16
FIG. 5. Specific inhibition of a proline-rich activator by CTF-1 acting as a competitor. Relative CAT activities were measured from extracts of HeLa cells cotransfected with the pG5BCAT reporter and 2 jig (0.7 pmol of plasmid) of the expression vector for either Gal 399-499 (A) or Gal VP16 (B), together with various amounts (picomoles of plasmid) of the p113-CTF-1A, p113-CTF-1, p113-Spl, and p113-VP16 competitor expression vectors, as indicated. The minus signs indicate that no competitor plasmid was added. CAT activities were normalized to that obtained from the expression of Gal 399-499 (part A, 790-fold induction over the basal level of expression in the absence of activator and competitor) or that of Gal VP16 (part B, 5,500-fold induction over the basal level) in the absence of competitor, which were assigned a value of 100.
TRANSCRIPTIONAL SYNERGISM VIA STABILIZED TARGET
VOL . 1 l, 1991
2943
imply that the estrogen receptor and the CTF-1 proline-rich activating domain require at least one similar limiting target transcription factor to activate transcription. Moreover,
150
100
50
pmole competitor:
-
-
1.4 3.5 1.4 3.5
-
-
1.4 3.5 1.4 3.5
competitors:
-
-
CTF-IA CTF-1
-
-
CIF-IA CITF-
activators:
-
Gal 399-499
Gal 399-499
+
HEO
-
1.4 3.5 1.4 3.5
CIT-lA
CTF-i
HEO
FIG. 6. Synergistic transcriptional activation by Gal 399-499 and HEO involves the stronger tethering of a limiting target factor. Relative CAT activities were measured from estrogen-induced HeLa cells cotransfected with the reporter pGlBCAT-B-ERE, with or without 0.4 ,ug of the expression vector for either Gal 399-499 (part A), HEO (part C), or both (part B), together with various amounts of competitor expression vectors coding for full-length CTF-1 or a derivative deleted of the proline-rich activating sequence (CTF-1A), as indicated. The minus sign indicates that no competitor or activator plasmid was added. CAT activities were normalized to those obtained from the expression of Gal 399-499 (part A) or HEO (part C) alone or both Gal 399-499 and HEO (part B) in the absence of competitor, which were assigned an arbitrary value of 100. The open bars (part B) represent the relative CAT activities in the presence of the competitor CTF-1, as calculated from the sum of the activations conferred by each activator alone in the presence of the CTF-1 competitor.
DISCUSSION CTF-1 proline-rich activating domain synergizes with the
estrogen receptor. By using various Gal-CTF fusion proteins transiently expressed in HeLa cells, we demonstrated that the transcriptional activation domain of CTF-1, when fused to the GAL4 DNA-binding domain, is both necessary and sufficient for synergistic activation with the estrogen receptor. Both synergism and transcriptional activation functions colocalize in the human CTF-1 protein, since deletions in the proline-rich domain that affect one function correspondingly alter the other. Interestingly, the proline-rich activating domain of CTF-2 appears to be weaker than that of CTF-1 and does not synergize significantly with the estrogen receptor. This may reflect a functional diversity among distinct members of the CTF/NF-I family of activators. Further studies will be needed to address the possibility that fulllength CTF-1 and CTF-2 differ in their ability to regulate transcription synergistically with other activators on natural promoters. Similar target transcription factor(s) mediates transcriptional activation by CTF-1 and the estrogen receptor. Our in vivo competition experiments indicate that an excess of CTF-1 but not of CTF-1A, which lacks the proline-rich activating domain, specifically inhibits transcriptional activation by Gal 399-499, while neither competitor represses basal-level expression or the activation conferred by the negatively charged Gal VP16 protein. These results demonstrate that the three CTF-1 activities, transcriptional activation, synergism with the estrogen receptor, and competition for a limiting target, colocalize within the CTF-1 proline-rich domain and further suggest that these three activities are mediated by the direct or indirect interaction of this domain with a similar limiting target transcription factor. Furthermore, transcriptional activation by the estrogen receptor is also specifically inhibited by the CTF-1 competitor and is unaffected by overexpression of CTF-1A. These results
estrogen receptor mutants deleted of either the TAF-1 or TAF-2 activation domain are inhibited by the CTF-1 competitor (unpublished results), which is consistent with the recent demonstration that the two estrogen receptor activation domains can compete for each other's transcriptional activation (46). Consistently, our functional analyses suggest that both TAF-1 and TAF-2 synergize with the CTF-1 proline-rich activation domain. This contrasts with the previous finding that the two estrogen receptor activation domains have different properties concerning their synergistic activation with other types of activators (46, 47). These differences could be due to the different activator constructs tested or could reflect the spatial organization of these activators onto specific promoters. Together, our data suggest that both estrogen receptor TAF regions, although
having in some cases distinct synergistic properties (46), have at least one limiting target transcription factor in common with the CTF-1 proline-rich domain. In addition, the distinct specificities of various types of competitor molecules (e.g., CTF-1, Spl, and VP16) in squelching experiments (Fig. 5 and unpublished results) would be consistent with a multiplicity of different target factors in addition to that used by CTF-1 and the estrogen receptor, as suggested previously by in vitro transcription reconstitution experiments (35, 39). Synergism between CTF-1 and the estrogen receptor involves stabilized interactions with the target transcription factor(s). The finding that the CTF-1 activation domain, when fused to the heterologous GAL4 DNA-binding domain, is sufficient for synergistic transcriptional activation with the estrogen receptor, together with the correlation between the transcriptional activities of particular mutants and their capacities to synergize with the estrogen receptor, suggests that it is the CTF-1 activation domain by itself, and not a simple cooperative DNA binding of the two activators, that mediates synergism. Indeed, our DNA-binding analyses indicate that both activators, Gal 399-499 and the estrogen receptor, do not cooperate for specific DNA binding in vitro. This suggests that synergism is not mediated by direct interactions between these two activators during promoter recognition but rather results from a subsequent step in the transcriptional activation pathway. In this respect, our in vivo competition experiments demonstrate that CTF-1 is a much weaker inhibitor of transcriptional activation when Gal 399-499 and the estrogen receptor are acting synergistically than when each activator is functioning alone. Furthermore, an excess of CTF-1 competitor that abolishes Gal 399-499mediated activation does not reduce synergism. In contrast, the apparent synergism is increased even further by the presence of the competitor. These results thus suggest a stronger tethering of the limiting target factor(s) with both activators acting synergistically than with either one alone. Therefore, the stronger interaction of the target(s) with both trans activators than with only one may be responsible for the synergistic transcriptional activation. These findings raise the question of how the target factor(s) may be stabilized by the two activators. One possibility is that the transcriptional domains of both activators simultaneously interact with the same limiting target(s), either directly or through intermediary proteins. This would be consistent with the observation that the CTF-1 and estrogen receptor activation domains stimulate transcription by a similar type of limiting target factor(s). In this model,
2944
MARTINEZ ET AL.
MOL. CELL. BIOL.
A Transcription activation
0
1.
B
TARGETj
TFlI1A.
-B,
-5,-FJ
FIG. 7. Model for competition and synergism through a common a target factor is needed to mediate the activating function of an upstream promoter-bound activator. The limiting target may interact directly or indirectly with the activation domain (labeled A) of the DNA-bound activator (having a DNAbinding domain labeled D). In competition experiments, the target can also interact with the homologous activation domain of a competitor molecule that lacks a promoter-specific DNA-binding domain. Therefore, an excess of the promoter-unassociated competitor would sequester or mask the target, impeding its interaction with the promoter-bound activator, as shown in part A. This would in turn result in the inhibition of activator-dependent transcriptional activation, as indicated by the minus sign on the right-hand side of the figure. Note that in this model the target could be either a mediator (as shown here) between the activator and the general transcription machinery or a component of the latter. Note also that the target could be either a single molecule or a stable homo- or heteromultimer complex of polypeptides (see also the Discussion section). In contrast, when two activators stimulate transcription synergistically, as in part B, the addition of an excess of the specific competitor cannot inhibit transcriptional activation to a great extent (as indicated by the plus sign). This can result from the target's associating simultaneously and thus more stably with both activators acting synergistically than with either one alone, thereby preventing the displacement of the target by the competitor molecules. This implies a stronger tethering of the target in the presence of the two promoter-bound activators than when only one activator is present. The promoter-associated target factor(s) could then interact with the TATA box-binding factor (TFIID), as suggested here, and/or with other components of the general transcription machinery. Pol II, RNA polymerase II.
target. In this model,
distinct proteins (see reference 38 for a discussion). However, our observation that factors functioning synergistically are much more resistant to competition than factors acting alone implies that if the two activators associate with the limiting target through intermediary proteins, the latter factors should form a relatively stable complex that does not readily dissociate during a competition experiment. Previous reports have suggested that the targets of upstream activators may be components of the general transcription machinery, such as TFIID (16, 17, 43, 45, 50) and RNA polymerase II (1, 3). Therefore, TFIID and/or other general transcription factors could interact directly with CTF-1 and the estrogen receptor. However, it has been proposed recently that several upstream activators, including CTF/NF-I, may require specific coactivators or adaptor molecules that would bridge the activators with the general transcription machinery (2, 18, 29, 35, 39). Therefore, CTF-1 and the estrogen receptor may interact with an adaptor/coactivator molecule(s), which in turn would contact one or several components of the general transcription machinery, including perhaps TFIID (as proposed in Fig. 7B). It is quite possible that distinct classes of upstream activators may use one or the other of the above general activation pathways. Our evidence for a stronger tethering of a limiting target factor(s) to the two promoter-bound activators acting synergistically emphasizes the importance of concerted protein interactions arising subsequent to promoter recognition by the site-specific DNA-binding factors, in the context of combinatorial regulation of gene expression. ACKNOWLEDGMENTS We thank V. Baichwal, P. Chambon, J. Lillie, I. Sadowski, S. Triezenberg, and T. von der Weid for the kind gift of plasmids. We are also grateful to S. Catsicas, M.-C. Lebeau, M. Taylor, M. Tsai, and T. Williams for critical reading of the manuscript. This work was supported by the ttat de Vaud and grants from the Swiss National Science Foundation to N.M. and W.W. REFERENCES 1. Allison, L. A., and C. J. Ingles. 1989. Mutations in RNA polymerase II enhance or suppress mutations in GAL4. Proc. Natl. Acad. Sci. USA 86:2794-2798. 2. Berger, S. L., W. D. Cress, A. Cress, S. J. Triezenberg, and L. Guarente. 1990. Selective inhibition of activated but not basal
3. 4.
the interaction of the target with a single DNA-bound activator is relatively weak, and consequently an excess of the specific competitor efficiently titrates it from the promoter-bound activator. This results in a strong inhibition of transcriptional activation (Fig. 7A). In contrast, the competitor cannot efficiently displace a limiting target that would be simultaneously and thus tightly tethered by both DNAbound activators, by either direct or indirect interactions (Fig. 7B). Note that a more stable association of both activators with the promoter could indirectly result from their simultaneous interaction with a common transcription factor in vivo. What could be the nature and function of the proposed target? What we call the target in Fig. 7 could represent a single polypeptide or a complex of several identical or
5. 6.
7.
8.
transcription by the acidic activation domain of VP16: evidence for transcriptional activators. Cell 61:1199-1208. Brandl, C. J., and K. Struhl. 1989. Yeast GCN4 transcriptional activator protein interacts with RNA polymerase II in vitro. Proc. Natl. Acad. Sci. USA 86:2652-2656. Breathnach, R., and B. A. Harris. 1983. Plasmids for the cloning and expression of full-length double-stranded cDNAs under the control of the SV40 early or late gene promoter. Nucleic Acids Res. 11:7119-7136. Carey, M., H. Kakidani, J. Leatherwood, F. Mostashari, and M. Ptashne. 1989. An amino-terminal fragment of GAL4 binds DNA as a dimer. J. Mol. Biol. 209:423-432. Carey, M., Y.-S. Lin, M. R. Green, and M. Ptashne. 1990. A mechanism for synergistic activation of a mammalian gene by GAL4 derivatives. Nature (London) 345:361-364. Corthesy, B., J.-R. Cardinaux, F.-X. Claret, and W. Wahli. 1989. A nuclear factor I-like activity and a liver-specific repressor govern estrogen-regulated in vitro transcription from the Xenopus laevis vitellogenin Bi promoter. Mol. Cell. Biol. 9:5548-5562. Corthesy, B., F.-X. Claret, and W. Wahli. 1990. Estrogen receptor level determines sex-specific in vitro transcription from the Xenopus vitellogenin promoter. Proc. Natl. Acad. Sci. USA 87:7878-7882.
VOL . 1 l, 1991
TRANSCRIPTIONAL SYNERGISM VIA STABILIZED TARGET
9. Courey, A., and R. Tjian. 1988. Analysis of Spl in vivo reveals multiple transcriptional domains, including a novel glutaminerich activation motif. Cell 55:887-898. 10. Davidson, I., J. H. Xiao, R. Rosales, A. Staub, and P. Chambon. 1988. The HeLa cell protein TEF-1 binds specifically and cooperatively to two SV40 enhancer motifs of unrelated sequence. Cell 54:931-942. 11. Fried, M., and D. M. Crothers. 1981. Equilibria and kinetics of lac repressor-operator interactions by polyacrylamide gel electrophoresis. Nucleic Acids Res. 9:6505-6525. 12. Gill, G., and M. Ptashne. 1988. Negative effect of the transcriptional activator GAL4. Nature (London) 334:721-724. 13. Gorman, C. M., G. T. Merlino, M. C. Willingham, I. Pastan, and B. H. Howard. 1982. The Rous sarcoma virus long terminal repeat is a strong promoter when introduced into a variety of eukaryotic cells by DNA-mediated transfection. Proc. Natl. Acad. Sci. USA 79:6777-6781. 14. 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. 15. Green, S., P. Walter, V. Kumar, A. Krust, J. M. Bonert, P. Argos, and P. Chambon. 1986. Human oestrogen receptor cDNA: sequence, expression and homology to v-erb-A. Nature (London) 320:134-139. 16. Horikoshi, M., M. F. Carey, H. Kakidani, and R. G. Roeder. 1988. Mechanism of action of a yeast activator: direct effect of GAL4 derivatives on mammalian TFIID-promoter interactions. Cell 54:665-669. 17. Horikoshi, M., T. Hai, Y.-S. Lin, M. R. Green, and R. G.
Roeder. 1988. Transcription factor ATF interacts with the TATA factor to facilitate establishment of a preinitiation complex. Cell 54:1033-1042. 18. Kelleher, R. J., P. M. Flanagan, and R. D. Kornberg. 1990. A novel mediator between activator proteins and the RNA polymerase II transcription apparatus. Cell 61:1209-1215. 19. Klein-Hitpass, L., M. Kaling, and G. U. Ryffel. 1988. Synergism of closely adjacent estrogen responsive elements increases their regulatory potential. J. Mol. Biol. 201:537-544. 20. Klein-Hitpass, L., S. Tsai, N. L. Weigel, G. F. Allan, D. Riley, R.
Rodriguez, W. T. Schrader, M.-J. Tsai, and B. O'Malley. 1990. The progesterone receptor stimulates cell-free transcription by enhancing the formation of a stable preinitiation complex. Cell 60:247-257. 21. Kumar, V., and P. Chambon. 1988. The estrogen receptor binds tightly to its responsive element as a ligand induced homodimer. Cell 55:145-156. 22. Kumar, V., S. Green, G. Stack, M. Berry, J.-R. Jin, and P.
Chambon. 1987. Functional domains of the human estrogen receptor. Cell 51:941-951. 23. Kumar, V., S. Green, A. Staub, and P. Chambon. 1986. Localization of the oestradiol-binding and putative DNA binding domains of the human oestrogen receptor. EMBO J. 5:22312236. 24. LeBowitz, J. H., R. G. Clerc, M. Brenowitz, and P. Sharp. 1989.
The Oct-2 protein binds cooperatively to adjacent sites. Genes Dev. 3:1625-1638. 25. Lees, J. A., S. E. Fawell, and M. G. Parker. 1989. Identification of two transactivation domains in the mouse oestrogen receptor. Nucleic Acids Res. 17:5477-5488. 26. Lewin, B. 1990. Commitment and activation at Pol II promoters: a tail of protein-protein interactions. Cell 61:1161-1164. 27. Lillie, J. W., and M. R. Green. 1989. Transactivation by the adenovirus Ela protein. Nature (London) 338:39-44. 28. Lin, Y.-S., M. Carey, M. Ptashne, and M. R. Green. 1990. How
different eukaryotic transcriptional activators can cooperate promiscuously. Nature (London) 345:359-361. 29. Martin, K. J., J. W. Lillie, and M. R. Green. 1990. Evidence for
interaction of different eukaryotic transcriptional activators with distinct cellular targets. Nature (London) 346:147-152. 30. Martinez, E., F. Givel, and W. Wahli. 1987. The estrogenresponsive element as an inducible enhancer: DNA sequence
requirements and conversion to a glucocorticoid-responsive element. EMBO J. 6:3719-3727.
2945
31. Martinez, E., and W. Wahli. 1989. Cooperative binding of estrogen receptor to imperfect estrogen-responsive DNA elements correlates with their synergistic hormone-dependent enhancer activity. EMBO J. 8:3781-3791. 32. Mermod, N., E. A. O'Neil, T. J. Kelly, and R. Tjian. 1989. The proline-rich transcriptional activator of CTF/NF-I is distinct from the replication and DNA binding domain. Cell 58:741-753. 33. Meyer, M.-E., H. Gronenmeyer, B. Turcotte, M.-T. Bocquel, D. Tasset, and P. Chambon. 1989. Steroid hormone receptors compete for factors that mediate their enhancer function. Cell 57:433-442. 34. Mitchell, P. J., and R. Tjian. 1989. Transcriptional regulation in mammalian cells by sequence-specific DNA-binding proteins. Science 245:371-378. 35. Peterson, M. G., N. Tanese, F. Pugh, and R. Tjian. 1990. Functional domains and upstream properties of cloned human TATA binding protein. Science 248:1625-1630. 36. Poellinger, L., B. K. Yoza, and R. G. Roeder. 1989. Functional cooperativity between protein molecules bound at two distinct sequence elements of the immunoglobulin heavy-chain promoter. Nature (London) 337:573-576. 37. Ptashne, M. 1988. How eukaryotic transcriptional activators work. Nature (London) 335:683-689. 38. Ptashne, M., and A. A. F. Gann. 1990. Activators and targets. Nature (London) 346:329-331. 39. Pugh, B. F., and R. Tjian. 1990. Mechanism of transcriptional activation by Spl: evidence for coactivators. Cell 61:1187-1197. 40. Sadowski, I., J. Ma, S. Triezenberg, and M. Ptashne. 1988. GAL4-VP16 is an unusually potent transcriptional activator. Nature (London) 335:563-564. 41. Sadowski, I., and M. Ptashne. 1989. A vector for expressing GAL4 (1-147) fusions in mammalian cells. Nucleic Acids Res. 17:7539. 42. Santoro, C., N. Mermod, P. C. Andrews, and R. Tjian. 1988. A family of human CCAAT-box-binding proteins active in transcription and DNA replication: cloning and expression of multiple cDNAs. Nature (London) 334:218-224. 43. Sawadogo, M., and R. G. Roeder. 1985. Interaction of a genespecific transcription factor with the adenovirus major late promoter upstream of the TATA box region. Cell 43:165-175. 44. Schmid, W., U. Striihle, G. Schutz, J. Schmitt, and H. Stunnenberg. 1989. Glucocorticoid receptor binds cooperatively to adjacent recognition sites. EMBO J. 8:2257-2263. 45. Stringer, K. F., J. Ingles, and J. Greenblatt. 1990. Direct and selective binding of an acidic transcriptional activation domain to the TATA-box factor TFIID. Nature (London) 345:783-786. 46. Tasset, D., L. Tora, C. Fromental, E. Scheer, and P. Chambon. 1990. Distinct classes of transcriptional activating domains function by different mechanisms. Cell 62:1177-1187. 47. Tora, L., J. White, C. Brou, D. Tasset, N. Webster, E. Scheer, and P. Chambon. 1989. The human estrogen receptor has two independent nonacidic transcriptional activation functions. Cell 59:477-487. 48. Triezenberg, S. J., R. C. Kingsbury, and S. L. McKnight. 1988. Functional dissection of VP16, the trans-activator of herpes simplex virus immediate early gene expression. Genes Dev. 2:718-729. 49. Tsai, S. Y., M.-J. Tsai, and B. W. O'Malley. 1989. Cooperative binding of steroid hormone receptors contributes to transcriptional synergism at target enhancer elements. Cell 57:443-448. 50. Van Dyke, M., R. G. Roeder, and M. Sawadogo. 1988. Physical analysis of a transcription preinitiation complex assembly on a class II gene promoter. Science 241:1335-1338. 51. Walker, P., J.-E. Germond, M. Brown-Luedi, F. Givel, and W. Wahli. 1984. Sequence homologies in the region preceding the transcription initiation site of the liver estrogen-responsive vitellogenin and apo-VLDLII genes. Nucleic Acids Res. 12: 8611-8626. 52. Wigler, M., R. Sweet, G. K. Sim, B. Wald, A. Pellicer, E. Lacy, T. Maniatis, R. Silvester, and R. Axel. 1979. Transformation of mammalian cells with genes from procaryotes and eucaryotes. Cell 16:777-785.
