Functional characterization of the interacting domains of the positive coactivator PC4 with the transcription factor AP-2α

Gene 320 (2003) 155 – 164 www.elsevier.com/locate/gene

Functional characterization of the interacting domains of the positive coactivator PC4 with the transcription factor AP-2a Li Zhong a, Yingqun Wang b, Perry Kannan b, Michael A. Tainsky a,* a

Program in Molecular Biology and Genetics, Barbara Ann Karmanos Cancer Institute, Wayne State University School of Medicine, 110 East Warren Avenue, Detroit, MI 48201, USA b MetroHealth Medical Center, Case Western Reserve University, 2500 MetroHealth Drive, Cleveland, OH 44109, USA Received 14 April 2003; received in revised form 26 June 2003; accepted 11 July 2003 Received by A.J. van Wijnen

Abstract The transcriptional positive cofactor 4 (PC4) physically interacts with the transcription factor, activator protein-2 (AP-2) a, and overexpression of PC4 results in a relief of the AP-2 transcriptional self-interference, which is induced by high levels of AP-2a expression. PC4 was initially described as a DNA-binding protein that enhances the activator-dependent transcription of class II genes in vitro, but it was later shown that PC4 could also act as a potent repressor of transcription on specific DNA structures such as single-stranded (ss) DNA, DNA ends and heteroduplex DNA. To further explore the functional domains of PC4 and its ssDNA-binding effect in the interaction with AP-2a and on AP-2 transcriptional activity, we investigated the C-terminal domain of PC4 (PC4-CTD) and several PC4 mutants in which the ssDNA binding function was interrupted. We found that the C-terminal domain of PC4 physically interacts with AP-2a and retains the function of full-length protein in relieving transcription self-interference of AP-2. A point-mutated form of PC4 within the C-terminal domain h-ridge, PC4 W89A, or a triple mutant in the h2 – h3 loop of PC4, F77A/K78G/K80G, inactivate the ability of PC4 to bind AP-2a and to relieve the transcription self-interference of AP-2a. In addition, point-mutated forms of AP-2a within the activation domain (AD) that inactivate AP-2 transcription activity also lose their self-interference function. Our data suggest that the C-terminal domain of the transcription cofactor PC4 is critical for AP-2a transcriptional interference that is mediated by the activation domain of AP-2a. D 2003 Elsevier B.V. All rights reserved. Keywords: AP-2; Transcription self-interference; Transcriptional cofactors

1. Introduction Activator protein-2 (AP-2), a 52-kDa protein, is a eukaryotic transcriptional regulator that is required for normal growth and morphogenesis during mammalian development (Nottoli et al., 1998; Mitchell et al., 1987). There are four highly homologous AP-2 family genes, AP-2a, AP-2h, AP2g and AP-2y, which have homologues in most mammalian cells (Bosher et al., 1996; McPherson et al., 1997; Zhao et

Abbreviations: AD, activation domain; CTD, C-terminal domain; CAT, chloramphenicol acetyl transferase; GST, glutathione S-transferase; TNT, transcription/translation systems. * Corresponding author. Tel.: +1-313-833-0715x2641; fax: +1-313832-7294. E-mail address: [email protected] (M.A. Tainsky). 0378-1119/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0378-1119(03)00823-0

al., 2001). These isotypes share a conserved C-terminal DNA-binding motif with an integral helix – span – helix homodimerization motif and a less-conserved proline-rich trans-activation domain (AD) near the N-terminus (Williams et al., 1988). AP-2a regulates the genes involved in a spectrum of biological functions including p21WAF1/CIP1 (Zeng et al., 1997), transforming growth factor-a (Wang et al., 1997), estrogen receptor (McPherson et al., 1997), keratinocyte-specific genes (Leask et al., 1991), tyrosine kinase receptor gene c-KIT (Huang et al., 1998), type IV collagenase (Frisch and Morisaki, 1990), HER-2/neu (Bosher et al., 1996), insulin-like growth factor-binding protein-5 (Duan and Clemmons, 1995) and the dopamine h-hydroxylase gene (Greco et al., 1995). AP-2a also negatively regulates a number of genes, including: MCAM/MUC18 (Jean et al., 1998); c/EBP-a, during adipogenesis; and c-myc (Gaubatz et al., 1995).

156

L. Zhong et al. / Gene 320 (2003) 155–164

Previous experiments have shown that an activated ras oncogene induced an increase in the expression of AP-2a in the human teratocarcinoma cell line, PA-1, and that abundant AP-2a, in turn, resulted in lower AP-2 transcription activity in these cells (Kannan et al., 1994). These results suggested that the ras oncogene might exploit the mechanism of AP-2 transcriptional self-interference to transform PA-1 cells. Further studies have demonstrated that a positive coactivator of transcription, positive cofactor 4 (PC4), physically interacted with AP-2a and reversed the AP-2 transcriptional self-interference (Kannan and Tainsky, 1999). PC4 was initially described as a DNA-binding protein that enhances the activator-dependent transcription of class II genes in vitro. It stimulates transcription in vitro with diverse kinds of activators, including VP16 (Kretzschmar et al., 1994; Ge and Roeder, 1994), thyroid hormone receptor (Fondell et al., 1999), octamer transcription factor-1 (Luo et al., 1998), and BRCA-1 (Haile and Parvin, 1999), presumably by facilitating assembly of the preinitiation complex through bridging between activators and the general transcriptional machinery (Ge and Roeder, 1994; Kaiser et al., 1995). It was later shown that PC4 could also act as a repressor of basal transcription (Malik et al., 1998; Werten et al., 1998a). The N-terminal region of PC4 contains a serine-rich portion termed the SEAC domain (Fig. 1A), which exhibits similarity to viral immediate-early proteins (Kretzschmar et al., 1994). Although stimulation of RNA polymerase II transcription by PC4 is probably largely effected by protein –protein interactions depending on the N-terminal half of the protein, the C-terminal half of the protein in addition comprises a powerful single-stranded (ss) DNA-binding domain (Kretzschmar et al., 1994; Werten et al., 1998b). The C-terminal domain of PC4 (PC4-CTD), spanning amino acids 63 –127, was identified to form a dimeric fold that provides an intriguing binding surface for two antiparallel ssDNA strands (Brandsen et al., 1997; Werten et al., 1999). The physiological role of this domain has long remained enigmatic. Recent reports have shown that PC4 is a very potent repressor of transcription on specific DNA structures such as ssDNA, DNA ends and heteroduplex DNA. These related DNA structures serve as effective initiation sites for RNA polymerase II (Malik et al., 1998; Schang et al., 2000). To explore the functional domains of PC4 in their interactions with AP-2a, we selected the PC4-CTD and several mutant forms of PC4 in which the ssDNA binding function was interrupted to test with the wild-type (WT) AP-2a and several AP-2a activation domain mutants (Fig. 1). Although we expected the N-terminal serine-rich domain (SEAC in Fig. 1), our results suggest that the CTD of PC4 is the main functional domain in the interaction with AP-2a to relieve its transcription self-interference. Disruption of the DNA-binding domain of PC4 leads to a loss of the physical interaction with AP-2a, and also to a loss of the ability of relieving AP-2 transcription self-interference

Fig. 1. Schematic representation of the wild-type and mutant forms of PC4 and AP-2a used in the experiments. (A) SEAC (serine – acidic) regions containing major CKII phosphorylation sites; lysine-rich (K-rich) region; single-stranded DNA-binding sites located in the C-terminal domain (PC4CTD); dimerization sites located in the CTD. Positions of mutations in the CTD were indicated by underlined letters. In h2 – h3 loop interruption triple mutation, both of the Lys residues and Phe residues were replaced by Gly and Ala, respectively (F77A/K78G/K80G). The single point mutation W89 was replaced by Ala (W89A). (B) Schematic structure of AP-2a indicates activation domain, dimerization domain and DNA-binding domain. The locations of the mutations in the activation domain of AP-2a that inhibit its activity are also shown.

under conditions that the protein expression levels for both AP-2a and PC4 unchanged.

2. Materials and methods 2.1. Cell culture 9117 cells, a subline of PA-1 human teratocarcinoma (Kannan et al., 1994), were grown in modified Eagle’s medium with Earl’s salt and L-glutamine (GIBCO BRL, Gaithersburg, MD) supplemented with 5% fetal bovine serum (Hazelton Biologics, Lenexa, KS) and 1% Penicillin –Streptomycin (GIBCO BRL) at 37 jC in 5% CO2 and 95% air. 2.2. Expression vectors AP-2a cDNA isolated from 6928 PA-1 cells was cloned in proper orientation in the EcoRI site of plasmid pSG5 (Stratagene, La Jolla, CA) to generate pSAP2. SV40 early promoter and h-globin intron sequences that enable efficient

L. Zhong et al. / Gene 320 (2003) 155–164

expression of the cloned genes are located before the EcoRI site in pSG5. The wild-type PC4 and five PC4 mutants were obtained as a generous gift from Dr. Michael Meisterernst’s laboratory in a bacterial expression plasmid pET-11 (Stratagene). The mutants that were designed for interrupting ssDNA binding were W89A PC4, F77A/K78G/K80G (h2 – h3 loop), the PC4-CTD, W89A PC4-CTD and h2 – h3 loop PC4 (Werten et al., 1998b, 1999). In order to generate the mammalian expression vector pSG5 for these mutants, PCR was used to generate a 5VEcoRI site plus a Kozak box and 3V BglII site, which was then cloned into the pSG5 plasmid EcoRI and BglII enzyme digestion sites. The forward PCR primer for the full-length PC4 was: 5V-GGAATTCACCATGCCTAAATCAAAGGAACT; the forward primer for the CTD of PC4 was: 5V-GGAATTCACCATGTTTCAGATTGGGAAAATG; the reverse primer for both full length and the CTD was: 5V-GATAGATCTTCACAGCTTTCTTACTGCGTC. The PCR products were purified, enzyme digested with EcoRI and BglII, and then ligated into pSG5. After cloning, all changes were verified by DNA sequencing. The GAL4 DBD and AP-2a activation domain containing fusion construct GAL4 – AP-2/11 – 226 was made by inserting the sequences encoding amino acids 11 – 226 of AP-2a into EcoRI-cut and mung bean nuclease-blunted pSG5. The mutations in full-length AP-2a at D52A and L107A/L108A were described in detail by Wankhade et al. (2000). 2.3. Transient transfections of PA-1 cells and chloramphenicol acetyl transferase (CAT) assays Transient transfections using calcium phosphate precipitation as described previously (Kannan et al., 1994) were performed to introduce the expression plasmid DNA into PA-1 cells. The amount of DNA used in all transfections was equalized by the addition of empty vector DNA, pSG5. The trans-activation activity of various activators was determined by measuring the CAT activity using respective expression plasmids and reporter constructs as follows. For the detection of AP-2 transcription activity, three AP-2 response element sequences were constructed (3  AP-2CAT) adjacent to the HSV tk promoter in the vector pBLCAT2 (Kannan et al., 1994). To detect GAL4 –AP-2a expression, we used the G5E1bCAT reporter, a generous gift from Dr. Ptashne. Forty-eight hours after transfection, cells were harvested and suspended in 1  phosphate-buffered saline, pH 7.2. Cells were lysed by three cycles of freezing ( 80 jC) and thawing (37 jC). CAT activity, normalized to 30 –50 Ag of protein, was measured by the conversion of [14C]chloramphenicol to monoacetyl- and diacetyl-chloramphenicol, essentially as described earlier (Werten et al., 1999). After partitioning the acetylated forms of chloramphenicol by thin-layer chromatography (TLC), the TLC plate was exposed to a phosphorimager plate, and the percentage conversion was calculated by measuring radioactivity on a Storm Analyzer (Molecular Dynamics, Sunnyvale, CA). The experiments were repeated two to three times

157

for each assay, and each transfection was performed in duplicate. The transfection efficiencies were monitored by Western blots of cell extracts from replicate plates. 2.4. Western blotting Forty-eight hours after transfection, cells were harvested and suspended in lysis buffer (1.5% Triton X-100, 2 mM phenylmethylsulfonyl fluoride, 0.5 mM EDTA, pH 8.0, protease inhibitor cocktail in phosphate-buffered saline, pH 7.2). Cells were lysed by three cycles of freezing ( 80 jC) and thawing (37 jC). The supernatant containing crude protein extract, 50 Ag, was separated on a 10% SDS – PAGE for AP-2 or GAL4– AP-2 protein and 15% SDS – PAGE for PC4 protein. Proteins were then transferred to a polyvinylidene difluoride membrane (NEN Life Science Products) and probed with either anti-AP-2a, anti-GAL4 DBD rabbit polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA) or anti-PC4 (a gift from Dr. Hui Ge). The signals were detected using anti-rabbit IgG conjugated with horseradish peroxidase and SuperSignal Chemiluminescent substrate (PIERCE). Images of the gels were developed by exposure to X-ray film. As an internal control, the same membranes were stripped and reprobed with an anti-hactin monoclonal antibody (Santa Cruz Biotechnology). 2.5. Analysis of glutathione S-transferase (GST)– AP-2a interactions with PC4 proteins The GST bacterial expression vector pGEX-AP-2a (Promega) was transformed into BL21 Escherichia coli competent cells via heat shock. The GST –AP-2a protein expression was induced by IPTG. Cells were pelleted then resuspended and lysed in PBS in the presence of PMSF and protease inhibitors cocktail using sonication. To purify the protein, the supernatant containing 500 Ag of crude cell extracts was applied onto glutathione – Sepharose beads (Promega) according to the manufacture’s instructions. The WT and mutant PC4 proteins were synthesized using the transcription/translation systems (TNT) kit (Promega) with 35S-methionine labeling. An aliquot of 30 Al of different 35S-labeled PC4 proteins was mixed with 30 Ag of bacterial GST – AP-2a purified protein that was bound to glutathione –Sepharose beads. After 4 h of incubation at 4 jC, the mixture was washed three times with PBS. PC4 proteins associated with GST – AP-2a were released by using reduced glutathione and 0.1% Triton. All elutes were boiled in SDS Laemmli loading buffer and resolved on a 15% SDS – PAGE. The gel was dried and exposed to a phosphorimager plate for imaging on a Storm Imager (Molecular Dynamics). Five microliters of each TNT-synthesized PC4 protein was also analyzed on a 15% gel to monitor binding affinity with AP-2a. To test the interaction of WT PC4 with AP-2a AD mutants, the D52A and L107A/ L108A mutants of AP-2a were PCR-amplified, cloned and fused with GST as described for WT AP-2a above. WT PC4

158

L. Zhong et al. / Gene 320 (2003) 155–164

protein was purified from pET11-PC4 plasmid (a generous gift from Dr. Hui Ge) containing bacteria as described earlier. The mutant AP-2a proteins fused with GST were purified, linked to glutathione – Sepharose beads and tested for the binding of bacterially purified WT PC4 as above. 2.6. Co-immunoprecipitation The WT and two mutant forms of AP-2a, D52A and L107A/L108A, were synthesized using the TNT system. Thirty microliters of each product was then mixed with 20 Al of 35S-methionine labeled WT PC4 protein generated also by the TNT kit (Promega) in 500 Al of 1  TBST for 4 h at 4 jC. Five microliters of anti-AP-2a antibody was then added to each reaction and incubated overnight at 4 jC. The mixture was bound to protein A agarose (20 Al) for 4 h at 4 jC. The protein A agarose pellets were washed three times with TBST, and then resuspended in SDS Laemmli loading buffer. After 5 min of boiling, the agarose was directly loaded onto a 15% SDS – PAGE gel. The gel was dried and exposed to a phosphorimager plate to digitize the image.

3. Results 3.1. The CTD of PC4 relieves AP-2 transcription self-inhibition The PA-1 human teratocarcinoma cell lines with differing propensities to form tumors offer an opportunity to study the physiological effects of a ras oncogene on transcriptional control mechanisms associated with differentiation and tumorigenicity (Kannan and Tainsky, 1999). N-ras-transformed PA-1 cells overexpress AP-2a but have low AP-2 transcriptional activity. In contrast, non-ras-transformed PA1 cells have low AP-2a protein level but higher endogenous AP-2 transcriptional activity. When transient transfection of AP-2a is performed into these two kinds of cells, both the N-ras-transformed and non-ras-transformed PA-1 cells show a dose-dependent decrease in AP-2 transcriptional reporter activity that we term transcriptional self-interference or self-inhibition. Previous results have shown that this AP-2 transcriptional self-interference can be reversed by cotransfection of transcription positive cofactor, PC4,

(Kannan and Tainsky, 1999) or poly-ADP-ribose polymerase (PARP) (Kannan et al., 1999). To investigate this effect of PC4 on AP-2 transcription activity in PA-1 cells, we selected several mutants of PC4 that were designed to affect the ssDNA binding property. Among these mutants, the CTD of PC4, spanning amino acids 63– 127, has a strong affinity to ssDNA binding (Brandsen et al., 1997). Other mutations, W89A and F77A/K78G/K80G (Fig. 1) either in full length or in the CTD of PC4, have shown a significant loss in the ssDNA binding property (Werten et al., 1998a). The tryptophan residue at amino acid position 89, W89, is located in the h-ridge separating the two antiparallel channels in PC4. Replacing this W89 with alanine (W89A) alters the h-ridge structure and interrupts the ssDNA binding affinity of PC4. The triple mutation F77A/K78G/K80G of PC4, referred to as the h2 –h3 loop, is a typical ssDNAbinding loop found in several other ssDNA-binding proteins, suggestive of the L45 loop of oligonucleotide/oligosaccharide binding-fold proteins (Murzin, 1993). It contains two positive charges (Lys78 and Lys80) and an aromatic residue (Phe77). Altering these two Lys and Phe to Gly and Ala, respectively, leads to an interruption of this special structure and results in a loss of ssDNA binding function of the protein (Werten et al., 1998a,b). Transient transfection experiments were performed to measure the effects of the WT and mutant PC4 on AP-2 transactivation properties. The PC4 constructs WT and mutants were cotransfected with the full-length AP-2a in PA-1 9117 cells (Fig. 2A). The endogenous transcriptional activity of AP-2 was inhibited by overexpressed AP-2a, and cotransfection of the PC4-CTD reversed this AP-2 selfinterference as efficiently as the WT PC4. Both the WT PC4 and PC4-CTD completely reversed the effects of exogenous, transfected AP-2a and restored the same or even higher transcriptional activity of AP-2. However, a single point mutation in full-length PC4 protein (W89A) or the h2– h3 triple mutants of CTD PC4 lost this activity. These results implied that the ssDNA binding structure of PC4 also has a strong influence on AP-2 transcription activity. The transfection efficacies for AP-2a or PC4 were monitored by Western blots in parallel with transfection assays. Furthermore, the expression of the mutant forms of PC4 produced similar levels of abundance as compared to the WT PC4 (Fig. 2B,C). An antibody to h-actin was used as an internal control for Western blotting (Fig. 2B). Little difference of

Fig. 2. WT and CTD PC4 relieves AP-2 transcriptional self-interference in PA-1 cells. Transient transfections and CAT assays were preformed as described in Section 2. After 48 h of transfection, cell lysates were prepared via freeze – thaw cycles, and aliquots were used for CAT assay and Western blotting to determine the AP-2 transcription activity and transfection efficiency. The amounts of expression plasmids for AP-2a and PC4 were shown at the bottom in the presence of 4 Ag AP-2 reporter plasmid 3  AP-2-CAT. The total amounts of plasmid DNA were equalized with the addition of the empty expression vector pSG5 and each treatment was in duplicate. The lowest AP-2 activity was taken as 1 to calculate the fold changes of the CAT activity affected by transfection of GAL4AD-2 and pSPC4. (A) Cotransfection of the WT and different mutant forms of PC4 have different effects on the AP-2 activity. The maximum AP-2 selfinhibition that is in the presence of 20 Ag pSAP-2 was taken as 1 to calculate the fold changes of AP-2 CAT activity. (B) The AP-2 transfection efficiency was tested via a Western blot using an anti-AP-2a antibody. Bacterially expressed GST – AP-2a was used as a positive control. The nontransfected PA-1 cells showed a very faint band as endogenous AP-2a protein level. The loading control was monitored using the same membrane and reprobed with anti-h-actin antibody. (C) Likewise, the PC4 (WT and mutants) transfection efficiency was detected using anti-PC4 antibody. The nontransfected PA-1 cells also showed a faint band of the endogenous PC4 level (Lane 1). HeLa cell nuclear extract was used to be a positive control (Lane 2).

L. Zhong et al. / Gene 320 (2003) 155–164

the h-actin control was observed among the different cotransfection conditions. To demonstrate that the interaction of the PC4-CTD was limited to the activation domain of AP-2a, we tested

159

whether the PC4-CTD affects the transactivation of a fusion protein, GAL4 – AP-2a-AD (activation domain), using a different transcriptional reporter, G5E1B-CAT, which detects the activation of transcription through the GAL4 DNA-

160

L. Zhong et al. / Gene 320 (2003) 155–164

binding domain in CAT assays. Because GAL4 is not normally present in mammalian cells, we do not expect any G5E1B-CAT activity without transfection of the GAL4 –AP2a-AD construct into PA-1 cells. Fig. 3A shows that after

transfection of 3 Ag of GAL4– AP-2a-AD, there is a moderate amount of GAL4-CAT activity, but as the amount of GAL4 –AP-2a-AD DNA was increased to 25 Ag, this CAT activity was suppressed 2.5-fold. Similar to the effect ob-

L. Zhong et al. / Gene 320 (2003) 155–164

161

Fig. 5. Activation domain mutant forms of AP-2a lost self-inhibition function. Transient transfections and CAT assays were preformed as described in Section 2 to determine the effect of the AP-2a activation domain mutants on the AP-2 transcriptional activity. The value of the WT AP-2 self-inhibition was taken as 1 to calculate the fold changes of the CAT activity. The transfection efficacies were monitored by Western blot (data not shown).

AP-2a is essential for PC4 modulation of transcriptional selfinterference activity in PA-1 cells, and that the CTD of PC4 is sufficient for this activity. Western blots were used to demonstrate equal expression of GAL4– AP-2a under the conditions of the transfection assays. Because there is no endogenous GAL4 level in the mammalian cells, no GAL4 protein was observed in nontransfected cells (Fig. 3B). Equal protein loading was monitored using Western blotting with a h-actin antibody. Fig. 4. Physical interactions between AP-2a and different forms of PC4. The GST – AP-2a was expressed and purified as described in Section 2. The WT and all mutant forms of PC4 proteins were synthesized using TNT kit (Promega), and all the proteins were labeled with 35S-methionine. (A) Thirty microliters of each in vitro synthesized PC4 proteins was incubated with purified GST – AP-2a and passed through GST columns. The elutions were electrophoresed on a 15% SDS – PAGE, and results were visualized by phosphorimager plate exposure. (B) Five microliters of the synthesized PC4 samples was subjected onto 15% SDS – PAGE gel, and the dried gel was visualized by phosphorimager plate exposure. Binding affinity between AP2a and PC4 can be relatively reflected by the loading volume and band intensities.

served on endogenous AP-2 activity, cotransfection of the WT and CTD of PC4 showed a sharp (five- to sixfold) increase in GAL4 –AP-2a-activated transcription. Likewise, the presence of the PC4 W89A and loop interruption mutants did not have any effect on the activity of the GAL4-CAT reporter. These results suggested that the activation domain of

3.2. Analysis of PC4 proteins and their interaction with AP-2a We next examined the physical interaction of various mutants of PC4 with a bacterially expressed AP-2a protein using a GST pull-down method. We previously reported that the WT PC4 interacts physically with AP-2a, and this interaction leads to a relief of AP-2 transcriptional selfinterference (Kannan and Tainsky, 1999). The bacterially expressed GST – AP-2a fusion protein was adsorbed to glutathione –Sepharose beads, washed and incubated with the PC4 proteins. The WT and mutant forms of PC4 were in vitro synthesized and labeled with [35S]-methionine, and the products were detected using a phosphorimager. The fulllength PC4 proteins migrated as f 17 kDa, and the CTD PC4 migrated at f 9 kDa (Fig. 4B). We observed an interaction of AP-2a with the WT and CTD forms of PC4 (Fig. 4A) as retention on the GST – AP-2 bound to glutathi-

Fig. 3. Effects of WT and mutant PC4 on GAL4 – AP-2a transcriptional activity. Transient transfections and CAT assays were preformed as described in Section 2 to determine how the exogenous transfected GAL4 – AP-2a activity was affected by AP-2 self-inhibition and the WT and mutants of PC4. (A) The amount of expression plasmids for AP-2 activation domain GAL4 – AP-2a and different forms of PC4 were shown at the bottom in the presence of 5 Ag GAL4 reporter plasmid G5E1bCAT. The value of the maximum GAL4 – AP-2a self-inhibition was taken as 1 to calculate the fold changes of GAL4 activity. Note that PA-1 cells do not have endogenous GAL4 activity. (B) The GAL4 – AP-2a transfection efficiency was tested by Western blot using anti-GAL4 antibody in the condition of transfection assays. (C) The PC4 transfection was also monitored by Western blot using anti-PC4 antibody, and the same membrane was washed and reprobed with anti-h-actin antibody as the internal control.

162

L. Zhong et al. / Gene 320 (2003) 155–164

one beads. By comparing the bands’ intensities between Fig. 4’s panels A and B, we could demonstrate that the relative binding efficiency of AP-2a for the WT PC4 is f 8%, and for PC4-CTD, it is f 5%, under the conditions of our assay. Interestingly, W89A and loop interruption mutants within the CTD of PC4 failed to display this physical interaction with AP-2a in vitro.

Fig. 6. Physical interactions between the activation domain mutant forms of AP-2a and WT-PC4. Co-immunoprecipitation was preformed as described in Section 2 to examine the physical interactions between the mutant forms of AP-2a and PC4. All forms of AP-2a and PC4 proteins were also synthesized using the TNT kit. The WT PC4 protein was labeled with 35Smethionine. (A) The products of the co-immunoprecipitation were subjected onto a 15% SDS – PAGE, and the gel was dried and exposed to a phosphorimager plate. Lane 1 is the 35S-methionine-labeled PC4; Lane 2 is the precipitation with the WT AP-2a; Lane 3 is the precipitation with the D52A mutant AP-2a; Lane 4 is the precipitation with the L107A/L108A mutant AP-2a; Lane 5 is a negative control, in which the in vitro synthesis reaction contained only the empty pSG5 vector. A Western blot with an anti-AP-2a antibody was used to demonstrate that similar amounts of protein were produced of the in vitro synthesized preparations of AP-2a (data not shown). The binding percentage between PC4 and different AP-2 can be calculated by comparing the bands densities to Lane 1, which is the unbound total PC4. (B) GST – AP-2a binding assays were performed as described in Section 2 using bacterially purified WT PC4. Detection of PC4 was performed using Western blotting. Lane 1, bacterially purified PC4; Lane 2, GST control; Lane 3, GST – AP-2a WT; Lane 4, GST – AP-2a/ D52A; Lane 5, GST – AP-2a/P56A; Lane 6, GST – AP-2a/P59A; Lane 7, GST – AP-2a/L107A/L108A.

3.3. AP-2 activation domain mutations lose transcriptional self-interference The activation domain of AP-2a has been shown to reside between amino acids 52 and 108, which contains many residues and motifs conserved among all AP-2 isotypes (Wankhade et al., 2000). The amino acids D52 (the beginning of AD) and L107/L108 (the end of AD) are among the most conserved residues, and a mutation at any of these positions (D52A or L107A/L108A) results in a >90% loss of the AP-2 transcription activity in CAT reporter assays (Wankhade et al., 2000). To test the self-interference of these AP-2a mutants, we used non-ras-transformed PA-1 9117 cells that have a significant level of endogenous AP-2 activity. Although 15 Ag of the WT AP-2a inhibited 85% of the endogenous activity in the PA-1 cells, the D52A or L107A/L108A AP-2a mutants transfected at similar DNA levels had little effect on the endogenous AP-2 activity (Fig. 5). These results indicated that the activation domain of AP2a is the region that possesses its self-interference activity. To further understand this function of the activation domain of AP-2a and the WT PC4, we performed a co-immunoprecipitation assay. The D52A mutant form of AP-2a showed a low level of binding to PC4, while the L107A/ L108A mutant bound even more weakly to PC4 (Fig. 6A). Because the input amounts of 35S-PC4 were the same in all lanes, by comparing the bands’ intensities to Lane 1, which is the total input PC4, we can calculate the relative binding efficiency. WT AP-2 bound to f 25% of the input PC4, whereas the mutant AP-2 proteins, D52A and L107A/ L108A, bound 5% and >1% of the input PC4, respectively. To confirm these results, the WT and mutant AP-2a proteins were cloned and fused with the bacterial glutathione Stransferase gene, and GST – AP-2-binding assay was performed to test their interaction with bacterially purified WT PC4. Similar to co-immunoprecipitation studies, PC4 bound to WT GST –AP-2a fusion protein, but did not bind to GST – AP-2a/D52A and bound only weakly to GST –AP2a/L107A/L108A mutants as well as other activation-defective mutants of AP-2a (Fig. 6B). These results revealed that the activation domain of AP-2 is involved in sequestration of common coactivators, which are the targets for transcriptional interference.

4. Discussion We have found that the CTD of transcription cofactor PC4 is responsible for the activity of PC4 to relieve the transcription self-interference of AP-2. In addition, this CTD of PC4 is also the functional domain physically interacting with the activation domain of AP-2a. All mutants of PC4 that interrupted ssDNA binding interruption were unable to modulate AP-2 transcription activity or demonstrate a physical interaction with AP-2a. As a transcriptional cofactor, PC4 can act as an activator or a

L. Zhong et al. / Gene 320 (2003) 155–164

repressor on basal transcription. Recent reports have shown that PC4 represses transcription on specific DNA structures including ssDNA, DNA ends and heteroduplex DNA. These related DNA structures serve as effective initiation sites for RNA polymerase II (Malik et al., 1998; Schang et al., 2000). The interaction of PC4 with AP-2a in PA-1 cells more likely functions as a transcriptional repressor to inhibit the basal transcription of AP-2a. The AP-2a promoter contains a number of AP-2 binding sites (Bauer et al., 1994). We hypothesize that AP-2 family transcription factors elicit transcriptional self-interference property as a feedback mechanism so that the increased AP-2 protein level will repress expression and AP-2 transcriptional activity, and, in turn, the decreased AP-2 protein level will stimulate the AP2 transcriptional activity to restore the proper level of AP-2 protein in the cells. After overexpression of AP-2, the elevated level of AP-2 signals the transcriptional machinery to lower its transcriptional activity, and as a result, we observe the transcription self-interference. When PC4 is cotransfected into the cells at this moment, the overexpressed PC4 will bind to AP-2 transcriptional initiation site (presumably ssDNA) to repress AP-2 production, and as a result, AP-2 protein level decreases to its regular level, and its transitional activity resumes to its normal level. We observe a reversed AP-2 self-interference by PC4. Our pervious results (Kannan and Tainsky, 1999) showed that transfection of PC4 alone into PA-1 cells increases endogenous transcriptional activity of AP-2, which further indicates PC4 as a physiological repressor of AP-2. The PC4 transcription repressing property can be alleviated by the presence of general transcription factor TFIIH (Muchardt et al., 1992). However, the high concentration of PC4 dominantly represses transcription regardless of the presence of TFIIH. In fact, PC4 is a very abundant protein in mammalian cells that might well repress transcription via competition with RNA polymerase II (RNA Pol II) in the presence of TFIIH (Malik et al., 1998). The ras oncogene induces high levels of AP-2a mRNA in PA-1 cells leading to low AP-2 activity due to AP-2 transcription self-inhibition. Likewise, transfection of exogenous AP-2a into PA-1 cells also causes significant transcription inhibition of AP-2 (Kannan and Tainsky, 1999). Our previous results showed that the tumorigenic properties in ras-transformed PA-1 cells could be reverted to nontumorigenic using constitutive expression of PC4 in this cell line (Kannan and Tainsky, 1999). In the N-terminal region, PC4 consists of serine-rich and acidic residues (SEAC) that are responsible for phosphorylation in vitro. PC4 shows little transcription repression when phosphorylated by Casein Kinase II (CKII) in vitro (Ge and Roeder, 1994). Our results did not show any difference between the WT PC4 and the CTD of PC4 on AP-2 transcription activity. All the CKII phosphorylation sites are in the N-terminal SEAC region missing from the CTD (Fig. 1A). Therefore, we conclude that the ras oncogene does not decrease AP-2 activity through the CKII-mediated phosphorylation of PC4.

163

As a transcription cofactor for AP-2a, PC4 physically interacts with the AP-2a activation domain. In vitro synthesized PC4 mutants that have lost their coactivator function on AP-2 activity assays were used to study the interactions of these forms of PC4 with GST – AP-2a. Only the WT PC4 and CTD of PC4 have a clear interaction with the bacterially expressed GST – AP-2a. All the ssDNAbinding PC4 missense mutants have lost the ability to physically interact with the AP-2a protein. These results indicate that these mutations, which were designed to interrupt ssDNA binding, also affected binding to AP-2a, suggesting that the h-ridge and h2 – h3 loop structures are essential for PC4 binding to AP-2a. Because these PC4 proteins were synthesized in vitro using a wheat-germ translation system, they should not be phosphorylated, therefore indicating that phosphorylation is not required for the PC4 – AP-2a interaction. Point mutations in the AP-2a activation domain that inhibit transactivation activity also block its self-interference and disrupt the interaction with PC4. In conclusion, transcription cofactor PC4 regulates AP-2 transcription through its CTD by physically interacting with the N-terminal activation domain of AP2a, releasing AP-2 from transcriptional self-interference.

Acknowledgements We thank Dr. M. Meisterernst for the gift of bacterial expression vectors for PC4 mutants. We also thank Dr. H. Ge for the PC4 polyclonal antibody. This work was supported by NIH grant R01-CA53475 to MAT and the Cancer Center Core Sequencing Facility P30-CA22453.

References Bauer, R., Imhof, A., Pscherer, A., Moser, M., Seegers, S., Kerscher, M., Tainsky, M.A., Hofstaedter, F., Buettner, R., 1994. The genomic structure of the AP-2 transcription factor. Nucleic Acids Res. 22, 1413 – 1420. Bosher, J.M., Totty, N.F., Hsuan, J.J., Williams, T., Hurst, H.C., 1996. A family of AP-2 proteins regulates c-erbB-2 expression in mammary carcinoma. Oncogene 13, 1701 – 1707. Brandsen, J., Werten, S., van der Vliet, P.C., Meisterernst, M., Kroon, J., Gros, P., 1997. C-terminal domain of transcription cofactor PC4 reveals dimeric ssDNA binding site. Nat. Struct. Biol. 4, 900 – 903. Duan, C., Clemmons, D.R., 1995. Transcription factor AP-2 regulates human insulin-like growth factor binding protein-5 gene expression. J. Biol. Chem. 270, 24844 – 24851. Fondell, J.D., Guermah, M., Malik, S., Roeder, R.G., 1999. Thyroid hormone receptor-associated proteins and general positive cofactors mediate thyroid hormone receptor function in the absence of the TATA box-binding protein-associated factors of TFIID. Proc. Natl. Acad. Sci. U. S. A. 96, 1959 – 1964. Frisch, S.M., Morisaki, J.H., 1990. Positive and negative transcriptional elements of the human type IV collagenase gene. Mol. Cell. Biol. 10, 6524 – 6532. Gaubatz, S., Imhof, A., Dosch, R., Werner, O., Mitchell, P., Buettner, R.,

164

L. Zhong et al. / Gene 320 (2003) 155–164

Eilers, M., 1995. Transcriptional activation by Myc is under negative control by the transcription factor AP-2. EMBO J. 14, 1508 – 1519. Ge, H., Roeder, R.G., 1994. Purification, cloning, and characterization of a human coactivator, PC4, that mediates transcriptional activation of class II genes. Cell 78, 513 – 523. Greco, D., Zellmer, E., Zhang, Z., Lewis, E., 1995. Transcription factor AP-2 regulates expression of the dopamine beta-hydroxylase gene. J. Neurochem. 65, 510 – 516. Haile, D.T., Parvin, J.D., 1999. Activation of transcription in vitro by the BRCA1 carboxyl-terminal domain. J. Biol. Chem. 274, 2113 – 2117. Huang, S., Jean, D., Luca, M., Tainsky, M.A., Bar-Eli, M., 1998. Loss of AP-2 results in downregulation of c-KIT and enhancement of melanoma tumorigenicity and metastasis. EMBO J. 17, 4358 – 4369. Jean, D., Gershenwald, J.E., Huang, S., Luca, M., Hudson, M.J., Tainsky, M.A., Bar-Eli, M., 1998. Loss of AP-2 results in up-regulation of MCAM/MUC18 and an increase in tumor growth and metastasis of human melanoma cells. J. Biol. Chem. 273, 16501 – 16508. Kaiser, K., Stelzer, G., Meisterernst, M., 1995. The coactivator p15 (PC4) initiates transcriptional activation during TFIIA – TFIID – promoter complex formation. EMBO J. 14, 3520 – 3527. Kannan, P., Tainsky, M.A., 1999. Coactivator PC4 mediates AP-2 transcriptional activity and suppresses ras-induced transformation dependent on AP-2 transcriptional interference. Mol. Cell. Biol. 19, 899 – 908. Kannan, P., Buettner, R., Chiao, P.J., Yim, S.O., Sarkiss, M., Tainsky, M.A., 1994. N-ras oncogene causes AP-2 transcriptional self-interference, which leads to transformation. Genes Dev. 8, 1258 – 1269. Kannan, P., Yu, Y., Wankhade, S., Tainsky, M.A., 1999. PolyADP-ribose polymerase is a coactivator for AP-2-mediated transcriptional activation. Nucleic Acids Res. 27, 866 – 874. Kretzschmar, M., Kaiser, K., Lottspeich, F., Meisterernst, M., 1994. A novel mediator of class II gene transcription with homology to viral immediate-early transcriptional regulators. Cell 78, 525 – 534. Leask, A., Byrne, C., Fuchs, E., 1991. Transcription factor AP-2 and its role in epidermal-specific gene expression. Proc. Natl. Acad. Sci. U. S. A. 88, 7948 – 7952. Luo, Y., Ge, H., Stevens, S., Xiao, H., Roeder, R.G., 1998. Coactivation by OCA-B: definition of critical regions and synergism with general cofactors. Mol. Cell. Biol. 18, 3803 – 3810. Malik, S., Guermah, M., Roeder, R.G., 1998. A dynamic model for PC4 coactivator function in RNA polymerase II transcription. Proc. Natl. Acad. Sci. U. S. A. 95, 2192 – 2197. McPherson, L.A., Baichwal, V.R., Weigel, R.J., 1997. Identification of ERF-1 as a member of the AP-2 transcription factor family. Proc. Natl. Acad. Sci. U. S. A. 94, 4342 – 4347.

Mitchell, P.J., Wang, C., Tjian, R., 1987. Positive and negative regulation of transcription in vitro: enhancer-binding protein AP-2 is inhibited by SV40 T antigen. Cell 50, 847 – 861. Muchardt, C., Seeler, J.S., Nirula, A., Gong, S., Gaynor, R., 1992. Transcription factor AP-2 activates gene expression of HTLV-I. EMBO J. 11, 2573 – 2581. Murzin, A.G., 1993. OB(oligonucleotide/oligosaccharide binding)-fold: common structural and functional solution for non-homologous sequences. EMBO J. 12, 861 – 867. Nottoli, T., Hagopian-Donaldson, S., Zhang, A., Perkins, A., Williams, T., 1998. AP-2-null cells disrupt morphogenesis of the eye, face, and limbs in chimeric mice. Proc. Natl. Acad. Sci. U. S. A. 95, 13714 – 13719. Schang, L.M., Hwang, G.J., Dynlacht, B.D., Speicher, D.W., Bantly, A., Schaffer, P.A., Shilatifard, A., Ge, H., Shiekhattar, R., 2000. Human PC4 is a substrate-specific inhibitor of RNA polymerase II phosphorylation. J. Biol. Chem. 275, 6071 – 6074. Wang, D., Shin, T.H., Kudlow, J.E., 1997. Transcription factor AP-2 controls transcription of the human transforming growth factor-a gene. J. Biol. Chem. 272, 14244 – 14250. Wankhade, S., Yu, Y., Weinberg, J., Tainsky, M.A., Kannan, P., 2000. Characterization of the activation domains of AP-2 family transcription factors. J. Biol. Chem. 275, 29701 – 29708. Werten, S., Langen, F.W., van Schaik, R., Timmers, H.T., Meisterernst, M., van der Vliet, P.C., 1998a. The acetyltransferase activity of CBP stimulates transcription. J. Mol. Biol. 276, 367 – 377. Werten, S., Stelzer, G., Goppelt, A., Langen, F.M., Gros, P., Timmers, H.T., van der Vliet, P.C., Meisterernst, M., 1998b. Interaction of PC4 with melted DNA inhibits transcription. EMBO J. 17, 5103 – 5111. Werten, S., Wechselberger, R., Boelens, R., van der Vliet, P.C., Kaptein, R., 1999. Identification of the single-stranded DNA binding surface of the transcriptional coactivator PC4 by NMR. J. Biol. Chem. 274, 3693 – 3699. Williams, T., Admon, A., Luscher, B., Tjian, R., 1988. Cloning and expression of AP-2, a cell-type-specific transcription factor that activates inducible enhancer elements. Genes Dev. 2, 1557 – 1569. Zeng, Y.X., Somasundaram, K., el-Deiry, W.S., 1997. AP-2 inhibits cancer cell growth and activates p21WAF1/CIP1 expression. Nat. Genet. 15, 78 – 82. Zhao, F., Satoda, M., Licht, J.D., Hayashizaki, Y., Gelb, B.D., 2001. Cloning and characterization of a novel mouse AP-2 transcription factor, AP-2delta, with unique DNA binding and transactivation properties. J. Biol. Chem. 276, 40755 – 40760.