Targeting the Ets Binding Site of the HER2/neuPromoter with Pyrrole-Imidazole Polyamides*

Three DNA binding polyamides (1-3) were synthesized that bind with high affinity (K a = 8.7 · 109 m −1 to 1.4 · 1010 m −1) to two 7-base pair sequences overlapping the Ets DNA binding site (EBS; GAGGAA) within the regulatory region of the HER2/neu proximal promoter. As measured by electrophoretic mobility shift assay, polyamides binding to flanking elements upstream (1) or downstream (2 and 3) of the EBS were one to two orders of magnitude more effective than the natural product distamycin at inhibiting formation of complexes between the purified EBS protein, epithelial restricted with serinebox (ESX), and the HER2/neu promoter probe. One polyamide, 2, completely blocked Ets-DNA complex formation at 10 nm ligand concentration, whereas formation of activator protein-2-DNA complexes was unaffected at the activator protein-2 binding site immediately upstream of the HER2/neuEBS, even at 100 nm ligand concentration. At equilibrium, polyamide 1 was equally effective at inhibiting Ets/DNA binding when added before or after in vitro formation of protein-promoter complexes, demonstrating its utility to disrupt endogenous Ets-mediated HER2/neu preinitiation complexes. Polyamide 2, the most potent inhibitor of Ets-DNA complex formation by electrophoretic mobility shift assay, was also the most effective inhibitor of HER2/neu promoter-driven transcription measured in a cell-free system using nuclear extract from an ESX- and HER2/neu-overexpressing human breast cancer cell line, SKBR-3.

expression requires transcription factor (TF) 1 binding to its cognate DNA response element in the gene promoter region (3)(4)(5). The ability to preferentially block gene expression by interfering with TF-DNA complexes could be a powerful tool for elucidating how aberrant gene expression contributes to neoplastic phenotypes.
One strategy for developing gene-specific transcriptional inhibitors is to target DNA binding ligands to the cognate DNA response element of a crucial, promoter-regulating TF (6 -8). A number of DNA binding natural products and their analogs, which interfere with the binding of TFs to their promoter response elements, are potent inhibitors of gene expression (9 -14). Mithramycin, a G,C-specific DNA minor groove binder, inhibits c-Myc expression driven by its G,C-rich P1 promoter (9,10). Similarly, small molecules such as the DNA intercalator mitoxantrone and the minor groove binding distamycin (Dist), both of which can inhibit the binding of E2F1 to the dihydrofolate reductase promoter, are strong inhibitors of dihydrofolate reductase gene expression (11). However, most DNA binding ligands are not promoter-or TF-specific inhibitors; Dist, for example, is also known to inhibit gene transcription by interfering with the association of TATA box-binding protein to its A,T-rich response element (TATA box) found in the proximal promoter of many genes (12).
We have been investigating a new class of DNA minor groove binding ligands, hairpin pyrrole-imidazole polyamides, as potential promoter-and TF-specific inhibitors of gene expression. In this study we have designed several new polyamides specifically targeted to the Ets binding site (EBS) within the proximal promoter of the HER2/neu oncogene. Hairpin polyamides represent a significant advancement in ligand design in that they can achieve a remarkable degree of sequence specificity and high affinity for predetermined DNA sequences (13)(14)(15). Polyamides that contain the aromatic rings N-methylimidazole (Im) and N-methylpyrrole (Py) bind as pairs in an antiparallel fashion to specifically distinguish G⅐C (Im/Py) from C⅐G (Py/ Im). Py/Py pairs are partially degenerate and bind both A⅐T and T⅐A pairs. More than five aromatic rings are overwound relative to the DNA helix, and a ␤-alanine unit has proven to be a conformationally flexible analog of a pyrrole carboxamide unit (15). A ␤/␤ pair can replace a Py/Py pair and allow for recognition of longer sequences while maintaining the specificity for AT/TA base pairs (15). Recently, polyamides designed to interfere with TFIIIA binding to its promoter-response ele-ment, were shown to be potent and specific inhibitors of 5 S RNA gene transcription (16,17). Such designed polyamides have also been shown to specifically inhibit the replication of human immunodeficiency virus, type I virus within the genome of human blood cells (18).
The HER2/neu oncogene is amplified and transcriptionally up-regulated in 25-30% of human breast cancers (19). The dramatic loss of ErbB2/HER2 promoter activity in overexpressing (MDA-453) and normal expressing (MCF-7) cells when a mutation of the ErbB2/HER2 promoter's EBS is introduced (GAGGAA to GAGAGA) into a transfected ErbB2 promoterchloramphenicol acetyltransferase reporter construct demonstrates that the transcriptional up-regulation of HER2/neu depends on a highly conserved EBS and its GAGGAA core recognition sequence within the key regulatory region of the HER2/neu proximal promoter (20). Recent studies have confirmed, both in vitro and in vivo, that ErbB2-mediated tumorigenesis could be inhibited by transfecting an Ets repressor that binds specifically and uniquely to the same ErbB2/HER2 promoter's EBS being targeted by our polyamide ligands (21). A number of polyamides were designed to target the EBS and adjacent upstream or downstream flanking sequences unique to this promoter. Three different hairpin polyamides ImPy-␤-PyIm-(R) H2N ␥-PyPy-␤-ImPy-␤-Dp (1), ImPyPyPyPy-(R) H2N ␥-PyPyPyPy-␤-Dp (2), and ImPy-␤-PyPy-(R) H2N ␥-PyPy-␤-PyPy-␤-Dp (3) were synthesized to target either the upstream or downstream EBS flanking sequences 5Ј-TGCTTGA-3Ј or 5Ј-AGTATAA-3Ј, respectively (Fig. 1A). Quantitative footprint titration analysis confirmed their high affinity binding and sequence specificity for the HER2/neu EBS. Comparisons were made between these polyamides and the classical three ring DNA minor groove binder, Dist, in their abilities to inhibit binding to the HER2/neu EBS by the mammary gland Ets factor, ESX, thought to endogenously regulate this promoter in HER2/neu overexpressing human breast cancers (26). Lastly, a cell-free transcription assay was used to evaluate the specific and differential ability of these three polyamides to interfere with HER2/neu promoter-driven transcription.
ImPy-␤-PyPy-(R) H2N ␥-PyPy-␤-PyPy-␤-Dp (3) were synthesized from ␤-alanine-PAM resin using solid phase as described (22) and was characterized by a combination of analytical high pressure liquid chromatography, UV spectroscopy, and matrix-assisted laser desorption ionization/time of flight mass spectroscopy. MS, m / z observed for Quantitative DNase I Footprint Titrations-A 188-base pair (bp) DNA fragment was obtained by polymerase chain reaction using the plasmid RO6 as a template and the primers P1 (5Ј-GAGAAAGT-GAAGCTGGGAGTT-3Ј) and P2 (5Ј-CCTGGTTTCTCCGGTCCCAAT-3Ј). The primer P2 was 5Ј-radiolabeled with [␥-32 P]ATP using T4polynucleotide kinase (Roche Molecular Biochemicals). Polymerase chain reaction amplification in the presence of P1, P2 (labeled), plasmid RO6, and Taq polymerase (Roche Molecular Biochemicals) gave a 188-bp DNA fragment that was purified on a 6% nondenaturating polyacrylamide gel. All DNase I footprint reactions were performed in a total volume of 400 l containing a 5Ј-32 P-radiolabeled DNA fragment (17,000 cpm) and final concentrations of 10 mM Tris⅐HCl, 10 mM KCl, 10 mM MgCl 2 , 5 mM CaCl 2 , pH 7.0, and either 0.1-10 nM polyamide or no polyamide (control lanes) (23). The solutions were allowed to equilibrate for 12-14 h at 22°C. The footprinting reactions were initiated by the addition of 10 l of a stock solution of DNase I containing 1 mM dithiothreitol and incubated for 7 min at 22°C. The reactions were stopped by adding 50 l of a solution of 2.25 M NaCl, 150 mM EDTA, 28 mM base pair calf thymus DNA, and 0.6 mg/ml glycogen and ethanol precipitation. The precipitates were resuspended in 1ϫ Tris borate-EDTA, 80% formamide loading buffer, denaturated by heating at 85°C for 10 min, and cooled on ice. The reaction products were separated by electrophoresis on a 6% polyacrylamide gel in 1ϫ Tris borate-EDTA at 2000 V for 1 h. Gels were dried on a slab dryer and exposed to a photostimulatable storage phosphorimaging plate (Kodak Storage Phosphor Screen SO230 obtained from Molecular Dynamics) in the dark at 22°C for 12-24 h. The data from the storage screens were obtained using a Molecular Dynamics 400S PhosphorImager and analyzed by volume integration of the target sites and reference blocks using the ImageQuant version 3.3. software. Equilibrium association constants were determined as described previously (24). Each compound was tested three times; the values for the equilibrium association constants (K a ) correspond to average values from three independent gels.
Cell Culture and Nuclear Extract Preparation-SKBR-3 cells were purchased from ATCC (Rockville, MD) and maintained at 37°C with 5% CO 2 and in McCoy's 5a medium (Life Technologies, Inc.) with 10% fetal bovine serum. SKBR-3 cells grown to subconfluence were rinsed twice with phosphate-buffered saline, scraped, and collected by centrifugation at 1200 rpm for 5 min, 4°C (Sorvall RT6000, Newtown, CT). The following steps were performed at 4°C. Cell pellets were suspended in five times the packed cell volume in buffer A (containing 10 mM Hepes-KOH, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 0.75 mM spermidine, 0.15 mM spermine, and 1 mM dithiothreitol), followed by centrifugation at 1200 rpm for 5 min. The pellet was then resuspended in five times the pellet volume in buffer A, kept on ice for 8 min, and homogenized with 10 strokes using a Dounce homogenizer (tight pestle). The homogenate (ϳ95% lysed cells) was centrifuged at 15,000 rpm for ϳ1 min, (JA-17 rotor, JA-21 centrifuge; Beckman, Palo Alto, CA). The pellet was resuspended in buffer B with 20 mM Hepes-KOH (pH 7.9), 20% glycerol, 0.2 mM EDTA, 2.0 mM EGTA, 0.75 mM spermidine, 0.15 mM spermine, 2 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride followed by drop addition of an equal volume of buffer B that included 0.75 M NaCl. After rocking for 20 min, the supernatant was collected by centrifugation at 47,500 rpm for 45 min (SW-55 rotor, Beckman), and dialyzed against Ͼ100-fold buffer C (20 mM Hepes-KOH, pH 7.9, 20% glycerol, 100 mM KCl, 0.2 mM EDTA, 0.2 mM EGTA, 12.5 mM MgCl 2 , 2 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride) for 3 h. Precipitated debris was removed by centrifugation at 15,000 rpm (JA-21 centrifuge, JA-17 rotor, Beckman), and the protein content of the nuclear extract was quantitated using the Bio-Rad protein assay.
Proteins, Antibodies, and Oligonucleotides-Recombinant ESX protein was prepared as described (25). Briefly, full-length ESX cDNA was cloned into a pRSET His tag expression plasmid (NheI-HindIII; Invitrogen) and expressed in isopropyl tagged ESX protein was purified by Ni 2ϩ -chelate affinity chromatography, as recommended by the manufacturer (Qiagen Inc., Valencia, CA). Recombinant AP-2 protein was purchased from Promega Co. (Madison, WI). Monoclonal antibody against AP-2 was purchased from Santa Cruz Biochemical (Carpenteria, CA), and the anti-ESX affinity-purified rabbit polyclonal was prepared as described previously (26). A 34-mer DNA oligonucleotide (oligo) containing the EBS and derived from the HER2/ neu proximal promoter (TA5 sequence shown in Fig. 1A) and its complementary strand, were synthesized by the Biopolymers facility (Roswell Park Cancer Institute, Buffalo, NY). Oligos were gel-purified, annealed, and 5Ј-end-labeled with [␥-32 P]ATP using T4-polynucleotide kinase (New England BioLabs, Beverly, MA) as described previously (27).
Mobility Shift Assay-Demonstration of TFs binding to their DNA response elements in the proximal HER2/neu promoter was performed by EMSA using recombinant TFs (ESX, AP-2), duplexed and 5Ј-endlabeled TA5 promoter probe, with/without anti-TF antibody. In general, recombinant protein at the indicated concentrations and 1 nM 32 Plabeled DNA probe were incubated in a reaction buffer containing 25 mM Tris (pH 7.5), 30 mM KCl, 5% glycerol, 0.1% Nonidet P-40, bovine serum albumin (100 g/ml), and 1 mM dithiothreitol. After incubation at room temperature for 30 min, samples were loaded onto 5% native polyacrylamide gels running with Tris borate-EDTA buffer (44.5 mM Tris base, 44.5 mM boric acid, 1 mM EDTA, pH 8.3). The dried gel was exposed to Kodak film and the protein-DNA complexes were quantitated by computing laser densitometry (Molecular Dynamics, Sunnyvale, CA).
Identification of specific protein-DNA complexes was confirmed by the addition of specific antibodies in the EMSA reaction conditions, as indicated. The ability of polyamides to interfere with the formation of a protein-DNA complex was determined by EMSA. For polyamide effects  on monomeric ESX binding to TA5 (ESX-TA5 complex formation), assays were performed in two ways: (i) polyamides were incubated with a 32 P-labeled probe at room temperature for 30 min prior to the addition of ESX protein, and (ii) ESX protein was complexed with the probe prior to polyamide treatment. The inhibition of ESX-TA5 complex formation was measured by comparing polyamide-treated with nontreated samples. Polyamide ability to inhibit dimeric AP-2 binding to TA5 (AP-2-TA5 complex formation) was measured in a similar manner. IC 50 values (concentration of compound required for 50% inhibition of protein-DNA complex formation) were determined to express the inhibitory activity of each agent; these IC 50 polyamide concentrations were also expressed as r values, the molar ratio of ligand to DNA base pairs.
Cell-free Transcription Assay-In vitro transcription was performed in a cell-free system composed of DNA template, SKBR-3 nuclear extract, and buffer containing 12 mM Hepes-KOH (pH 7.9), 60 mM KCl, 7.5 mM MgCl 2 , 12% glycerol, 0.12 mM EDTA, 0.12 mM EGTA, 1.2 mM dithiothreitol, and 0.6 mM phenylmethylsulfonyl fluoride. The transcription DNA template consisted of CsCl-purified plasmid DNA containing the ϳ500-bp RO6 HER2/neu promoter fragment (20) . Transcript formation proceeded with incubation at 30°C for 60 min, and the reaction was stopped by adding 325 l of 10 mM Tris base (pH 8.0), 7 M urea, 350 mM NaCl, 1% SDS, and 100 g of tRNA, followed by phenolchloroform-isoamyl alcohol extraction and ethanol precipitation. Samples were resuspended in formamide-loading dye and heated at 90 -95°C for Ն1 min before loading onto a 4%, 7 M urea-polyacrylamide gel. The 32 P signal from a dried gel was visualized using a PhosphorImager screen and quantitated by computing laser densitometry (Molecular Dynamics, Sunnyvale, CA).
As with EMSA assessment of polyamide activity, ligand ability to inhibit transcript formation driven off the HER2/neu promoter was analyzed in two ways: (i) DNA template was incubated with polyamide at the indicated concentration in a total volume of 10 l for 30 min prior to the addition of nuclear extract and radiolabeled nucleotide pool, and (ii) preincubation of nuclear extract and DNA template for 15 min was followed by the addition of ligand for another 30 min in the total reaction volume to which radiolabeled nucleotide pool was then added. The degree of transcription was measured by quantitating transcript formation in ligand-treated versus untreated (control) samples and calculating IC 50 and r values. T3 transcript (250 bases; Promega Co., Madison, WI) was used as an internal control. In addition, a time course assay was used to compare transcriptional inhibition off the HER2/neu promoter in the presence of Dist versus polyamide 2 using our previously described procedure (11). For these time course assays, following the addition of ligands and nucleotides to the premixed template and nuclear extract volume, the reaction was stopped at different time points (0 -60 min), and the newly formed transcripts were quantitated as described above.

Design of HER2/neu Promoter Binding Polyamides-We
used the simple pairing code (15) to design polyamides that bind the 5Ј-and 3Ј-flanking sequences overlapping the EBS (GAGGAA) within the endogenous HER2/neu proximal promoter, the RO6 HER2/neu promoter-driven transcript template, and the EMSA TA5 probe. The proximal HER2/neu promoter sequence containing this EBS is shown in Fig. 1A, and this HER2/neu EBS has previously been shown to bind with high affinity to the potent and epithelial-restricted Ets transactivator, ESX (25,26). All Ets family members bind to the major groove of DNA and have additional critical phosphate contacts along flanking minor groove sequences (28,29). Whereas the GGAA core of the EBS is the same for virtually all Ets proteins, these 5Ј-and 3Ј-sequences immediately adjacent to the EBS core are often promoter-and Ets factor-specific. Thus, we synthesized three polyamides to target 7 bp of the HER2/neu promoter sequence 5Ј and 3Ј of the EBS core; polyamide 1 was designed to bind immediately upstream of the ESX core binding site at 5Ј-TGCTTGA-3Ј (site 1), whereas polyamides 2 and 3 were designed to bind immediately downstream and across the adjacent TATA box at 5Ј-AGTATAA-3Ј (site 2).
Selectivity and High Affinity Binding of Polyamides to the HER2/neu Promoter EBS-Quantitative DNase I footprint titration analysis showed that polyamides 1, 2, and 3 bind with high affinity to their target sites (Fig. 2). Polyamide 1 binds with an equilibrium association constant K a ϭ 9.6 ⅐ 10 9 M Ϫ1 to its match site (5Ј-TGCTTGA-3Ј). It binds by a factor of ϳ2 less strongly to a single base pair mismatch site (5Ј-AGAATGA-3Ј) located downstream with respect to the ESX binding site (K a ϭ 4.5 ⅐ 10 9 M Ϫ1 ). Polyamides 2 and 3 both bind with high affinity to their target site (5Ј-AGTATAA-3Ј, K a ϭ 1.4 ⅐ 10 10 M Ϫ1 and K a ϭ 8.7 ⅐ 10 9 M Ϫ1 , respectively).
Polyamide 2 Inhibition of Ets Binding to the HER2/neu Promoter-Because Dist can also bind to the TATA box contained in the 3Ј-EBS element targeted by two of the polyamides (12), polyamide 2 and Dist were compared by EMSA for their abilities to inhibit ESX binding to the HER2/neu promoter probe, TA5. Incubation of 2 with TA5 followed by the addition of ESX resulted in a concentration-dependent inhibition of ESX-TA5 complex formation; 10 nM 2 inhibited complex formation up to 95%, whereas as little as 1 nM resulted in a detectable decrease in complex formation (Fig. 3A, lanes 2-4). The pattern of inhibition of ESX-TA5 complex formation by Dist was similar, but significantly higher Dist concentration was required to achieve the same degree of inhibition observed by 2, because Dist at 2000 nM diminished complex formation by ϳ95% (Fig.  3B, lane 4). Whereas 100 nM of 2 inhibited ESX-TA5 complex formation almost entirely, 100 nM Dist had no effect on ESX-TA5 complex formation (Fig. 3A, lane 2 and Fig. 3B, lane 6). Quantitation of the data in Fig. 3C indicated that 2.2 nM 2 and 500 nM Dist are needed to inhibit complex formation by 50% (IC 50 ); Table I also shows the activity of individual polyamides at inhibiting protein-DNA complex formation expressed as r values, the molar ratio of ligand to DNA base pairs. Inhibitory Effects of Polyamides (1, 2, and 3) on Ets Binding to the HER2/neu Promoter-Because 1 and 2 recognize DNA elements upstream and downstream of the core EBS while 3 also recognizes the same downstream flanking element as 2 (Fig. 1), EMSA was used to test the relative ability of each polyamide to inhibit ESX binding to the HER2/neu promoter probe TA5. As shown in Fig. 4, 1 and 2 appeared similar in their ability to inhibit complex formation with respective IC 50 values of 5 and 2.2 nM. In contrast, 3 required a 9-fold higher concentration (18 nM) to prevent Ets-DNA complex formation by 50% as compared with 2 ( Fig. 4 and Table I).
For certain TF/DNA inhibitory drugs, equilibrium conditions demand greater drug concentrations to inhibit preformed DNA-bound complexes as opposed to preventing the initial formation of such complexes (6). For polyamide 1, however, similar experimental conditions were observed for inhibition of Ets-DNA complexes whether ligand was added before or after ESX binding to the HER2/neu promoter probe, because at 10 nM polyamide 1 concentration nearly the same level of ESX-TA5 complexes were formed within 30 min no matter which order the reagents were added (Fig. 5A). In contrast, 2 required 10-fold more ligand to obtain equal inhibition when added after ESX-TA5 complex formation as compared with the addition of 2 before complex formation (Fig. 5B). A time course assay using 10 nM of 2 indicated that the percentage of ESX-TA5 complexes inhibited by 2 increased with longer incubation time such that a Ն4 h polyamide 2 incubation was needed to achieve equilibrium conditions and maximal inhibition when 2 was added after the initial formation of ESX-TA5 complexes (Fig. 5C).
Inhibitory Effects of Polyamides on AP-2 Versus Ets Binding to the HER2/neu Promoter-Previous studies have suggested that AP-2 contributes to the overexpression of HER2/neu, and footprinting analysis has revealed that there are several AP-2 sites in the proximal HER2/neu promoter (30). We also had observed that both endogenous and recombinant AP-2 binds to a G-rich sequence just upstream of the EBS on the HER2/neu promoter. Dimeric binding of AP-2 to this G-rich element in the TA5 probe (Fig. 1) is demonstrable by EMSA and confirmed by the supershifting effect of an AP-2 monoclonal; in contrast, antibodies nonreactive to AP-2 had no effect on this AP-2-TA5 complex (Fig. 6A). Because AP-2 interacts with this G-rich element adjacent to the EBS in TA5, it was of interest to know whether the EBS-targeted polyamides would affect the binding of AP-2 to this HER2/neu promoter probe. Fig. 6B shows that 1 was capable of inhibiting EMSA-detected AP-2-TA5 complexes in a concentration-dependent manner and with an IC 50 ϭ 48 nM. In contrast, 2 was unable to block complex formation even at the highest concentration (100 nM) tested (Fig. 6C). 2 Likewise, the pattern of inhibition of AP-2-TA5 complexes by Dist was similar to that of 2 in that micromolar concentrations were required to significantly inhibit complex formation (Fig. 6D and Table I). All these compounds were more efficient at inhibiting formation of ESX (versus AP-2) complexes on the HER2/neu promoter probe, with 2 being the most specific and most potent inhibitor.
Transcription Inhibiting Effects of Polyamides on the HER2/ neu Promoter-To determine whether the effects of polyamides on Ets-DNA complex formation resulted in an ability to influence biological function, in vitro transcription assays were performed using a HER2/neu promoter-driven DNA template. With the ϳ500 bp of HER2/neu promoter-inserted plasmid (RO6) linearized with SphI as DNA template and SKBR-3 nuclear extracts (endogenously enriched in ESX, AP-2, TATA box-binding protein, etc.) to provide the transcriptional machinery, a ϳ760-base transcript is produced in this cell-free system. Compounds were first incubated with the DNA template prior to the addition of nuclear extracts and radiolabeled pool of nucleotides. A representative gel shown in Fig. 7A demonstrates the ability of 2 to block synthesis of the 760-base transcript in a concentration-dependent manner. Compared with the untreated control, 5 M of 2 inhibited transcript synthesis by 95%; whereas 1 M produced less than 50% inhibition of transcript formation, at higher polyamide concentrations there was some evidence of the partial transcript production (Fig. 7A, lanes 3 and 4). Comparative inhibition of HER2/neu promoter-driven transcription by 1, 2, 3, and Dist is shown in Fig. 7B. The order of transcription inhibiting potency (2 Ͼ 3 Ͼ 1 Ͼ Dist) is somewhat different from the EMSA ESX-TA5 complex inhibiting potency for these compounds.  Table II. Because EMSA results demonstrated differences between the ability of 1 and 2 to inhibit Ets-DNA complexes when ligand was given before or after initial formation of the ESX-TA5 complex (Fig. 5), to determine if the order of addition of polyamides influenced their transcription inhibitory activity, nuclear extracts were allowed to interact with the promoter and DNA template prior to ligand exposure. Polyamide effectiveness appeared to be reduced when tested in this fashion (Fig. 8). For example, concentrations of 2 at 1.4 or 4.2 M were required to inhibit transcription by 50% when ligand was added before or after nuclear extract binding with the R06 template (Fig. 8A). In the case of 1, a 2-fold higher ligand concentration was needed to inhibit transcription (6.4 M) when extract was prebound to template (Fig. 8B).
Previous time course studies with DNA binding and transcription-inhibiting drugs have shown that the degree of transcription inhibition can change in relation to the in vitro reaction time in the presence of moderately inhibiting drug concentrations (11). Conducting similar time course experiments with the most potent inhibitor 2 revealed a plateau level of transcription inhibition at all time points from 10 -60 min (Fig. 9). In contrast, the level of transcription inhibition by Dist declined somewhat in relation to incubation time. These time course differences between 2 and Dist might be accounted for by the higher DNA binding affinity of the polyamide, making it less likely that ligand is released from the template and transcription is allowed to resume during the longer exposure times. FIG. 7. Polyamide and Dist pretreatment of the HER2/neu promoter inhibits its transcriptional activity. HER2/neu promoterdriven transcription was measured in a cell-free assay as described under "Materials and Methods"; briefly, compound at the indicated concentration was incubated with a SphI-restricted HER2/neu promoter-driven transcription template (RO6) at 30°C for 30 min, followed by the addition of SKBR-3 nuclear extract and a radiolabeled pool of nucleotide precursors. Transcription was allowed to proceed for 60 min at 30°C with the expected ϳ760-base transcript identified by gel electrophoretic separation and phosphorimaging of the scanned gel. Scan from a representative experiment performed in the presence of polyamide 2 is shown in A. Lane   In this study, we examined the ability of sequence-specific polyamides to inhibit Ets-DNA complex formation and EBSregulated transcription off the HER2/neu promoter. Polyamides were synthesized that recognize different elements overlapping and flanking the GAGGAA EBS, located adjacent to and 5Ј of the TATA-box in the regulatory portion of the proximal HER2/neu promoter (20). As compared with the TATA box binding natural product Dist, three designed sequence-specific polyamides were more effective at inhibiting EBS complex formation with the mammary gland Ets transactivator, ESX, as well as HER2/neu driven transcription from a ϳ500 bp HER2/neu promoter sequence known to be regulated at the EBS as well as other endogenous response elements (e.g. AP-2, Sp1, CAAT, and TATA boxes). Of the three polyamides, 2 was the most strongly binding and effective HER2/neu promoter inhibitor, binding with a K a ϭ 1.4 ⅐ 10 10 M Ϫ1 to the 3Ј-flanking EBS element that includes the promoter's TATA box.
Comparison of the three polyamides (1-3) with Dist for inhibition of protein-DNA complex formation on the HER2/neu promoter probe, TA5, revealed the vastly enhanced potency and specificity of the high affinity hairpins as opposed to the latter natural product. Because both Dist and polyamides 2 and 3 bind the same TATA box containing the 3Ј-EBS element (Fig.  1), the higher binding affinity of 2 for this element likely contributed to its greater inhibitory activity over both Dist and polyamide 3. However, because Ets family members also make minor groove phosphate contacts in addition to their major groove base contacts, some of the enhanced inhibitory effects of both these polyamides over Dist may be attributed to steric effects restricting Ets (ESX) access to the HER2/neu EBS (28,31). With a similar comparison in the present study, HER2/neu promoter-targeted polyamides were shown to differentially affect ESX and AP-2 binding to adjacent DNA response elements. The binding of polyamide 1 to its 5Ј-EBS element partially impinges on the G-rich AP-2 binding site present in the TA5 probe ( Fig. 1), probably accounting for the observed ϳ10-fold less effective inhibitory activity of 1 at blocking formation of AP-2-TA5 versus ESX-TA5 complexes (Fig. 6B). Comparing Dist and polyamide 2 (Fig. 6, C and D), both of which bind the same 3Ј-EBS element located more remote from the AP-2 binding element in TA5, demonstrated the vastly improved promoter specificity of a designed polyamide over a less specific natural product like Dist, because the latter showed some AP-2 inhibitory activity whereas the former showed none despite its potent ESX-TA5 inhibitory activity over the same concentration range.
Small molecules that bind DNA near or at a TF response element typically require more time (or higher concentrations) to achieve steady-state inhibition of protein-DNA complexes when added after rather than before the formation of these complexes (6,12,27). Differences in this regard were noted between polyamides 1 and 2 when EMSA was carried out with ligands added before or 30 min after formation of ESX-TA5 complexes; polyamide 1 showed no significant impact by delayed administration but 2 showed a near 50% increase in its IC 50 (Fig. 5, A and B). However, by increasing its post-treatment incubation time from 30 min to 240 min, a 10 nM dose of polyamide 2 regained its full inhibitory activity as seen with a 30-min pretreatment at this same dose (Fig. 5C), demonstrating that 2 required longer post-treatment exposure than 1 to achieve its steady-state inhibitory potential. The difference in FIG. 8. Comparative inhibition of HER2/neu promoter-driven transcription when polyamides are administered before or after promoter binding to endogenous transcription factors contained in SKBR-3 nuclear extract. The experimental procedure was similar to that described in Fig. 7A except that the promoter-driven DNA template was incubated with SKBR-3 nuclear extracts for 15 min before the addition of a labeled pool of nucleotide precursors and polyamides 2 (A) or 1 (B). The percent inhibition of transcript formation () produced by the individual ligands was compared with that produced when ligand exposure to template preceded the addition of nuclear extract (q). Results represent the mean (Ϯ S.D.) of replicate experiments.
FIG. 9. Inhibition of HER2/neu promoter-driven transcription when compound is administered after template binding to nuclear extract and as a function of exposure time. The experimental procedure was similar to that described in Fig. 8 except that the cell-free transcription reaction was stopped after a 5-, 10-, 30-, or 60-min exposure to 2 (q) and Dist (). Results represent the mean (Ϯ S.D.) of replicate experiments. this regard between polyamides 1 and 2 likely reflects 3Јversus 5Ј-asymmetry in the TA5-bound ESX complex, resulting in greater structural interference and reduced access to the TA5 element recognized by 2 in the presence of preformed ESX-TA5 complexes (Fig. 9).
Polyamide 2, which most effectively inhibited ESX-TA5 complex formation at equilibrium also most effectively inhibited HER2/neu promoter-driven transcription, assayed in a cell-free system utilizing endogenous ESX, AP-2, TATA box-binding protein, and other transcriptional components endogenously present in a nuclear extract of the HER2/neu overexpressing breast cancer cell line, SKBR-3. Interestingly, polyamide 3, which was 3-fold less inhibitory than 1 at inhibiting formation of ESX-TA5 complexes on the 34-bp TA5 promoter probe (IC 50 of 18 nM versus 5 nM), was at least as effective as 1 at inhibiting cell-free transcription off the ϳ500-bp (R06) HER2/neu promoter-driven template. Moreover, the inhibitory activity of polyamide 1, which was unaffected in EMSA by prebinding of ESX to TA5, was moderately reduced in the cell-free transcription assay by prebinding of nuclear extract to the HER2/neu promoter-driven template, as was the transcription inhibitory activity of polyamide 2. Similar discordances were observed in comparisons of mitoxantrone and Dist as inhibitors of both protein-DNA complex formation and cell-free transcription with the DHFR promoter (11). Possible variables accounting for these discordances in the present study include the multiplicity of endogenous HER2/neu promoter binding factors present in the nuclear extract fueling the transcription assay (versus the single protein component in the EMSA assay) and potentially different numbers of lower affinity binding sites for each polyamide on the linearized R06 plasmid-containing HER2/neu promoter-driven template (versus the 34-bp TA5 EMSA probe). To address the potential impact of DNA content (bp) as a discordance-inducing variable between the EMSA and cell-free transcription assay, r values were calculated to compare the molar ratios of polyamide to DNA content (Tables I,  and II). The degree of difference between the EMSA and transcription assay r values for Dist is most notable and without obvious explanation. However, the lower overall r values among polyamides tested by transcription assay versus their EMSA determined values suggest that differences in total DNA content or polyamide binding sites on the HER2/neu promotercontaining plasmid template did not substantially contribute to the discordances noted above.
In summary, polyamides designed to selectively target critical 7-bp elements flanking and overlapping on a singular EBS in the regulatory region of the proximal HER2/neu promoter were shown to exhibit high affinity binding to their respective elements and to specifically disrupt binding of a HER2/neu promoter EBS candidate, ESX. These Ets-DNA complex inhibiting hairpin polyamides were significantly more potent inhibitors of HER2/neu promoter-driven transcription than the natural product Dist, a TATA box minor groove binder, and less effective Ets-DNA complex inhibitor. The differences noted in the HER2/neu promoter inhibiting activities of these polyamides is thought to be because of both their respective binding affinities and the choice of EBS flanking elements targeted for polyamide binding. These differences may implicate vulnerable promoter elements for future attempts to repress transcription of the overexpressing HER2/neu oncogene. Studies are now underway to evaluate the effectiveness of polyamides as HER2/ neu transcription inhibitors in whole cell systems.