DNA Double Helix Unwinding Triggers Transcription Block-Dependent

Chem. Res. Toxicol. 2009, 22, 483–491

483

DNA Double Helix Unwinding Triggers Transcription Block-Dependent Apoptosis: A Semiquantitative Probe of the Response of ATM, RNAPII, and p53 to Two DNA Intercalators Zhichao Zhang,*,† Yuanyuan Wang,‡ Ting Song,‡ Jin Gao,‡ Guiye Wu,† Jing Zhang,‡,§ and Xuhong Qian†,| State Key Laboratory of Fine Chemicals, Dalian UniVersity of Technology, Dalian 116012, People’s Republic of China, School of EnVironmental and Biological Science and Technology, Dalian UniVersity of Technology, Dalian 116024, China, State Key Laboratory of Elemento-organic Chemistry, Nankai UniVersity, Tianjin 300071, China, and Shanghai Key Laboratory of Chemical Biology, School of Pharmacy, East China UniVersity of Science and Technology, Shanghai 200237, China ReceiVed August 2, 2008

We have previously shown the binding modes of two DNA interacting analogues 1a {3-(4-methylpiperazin)-8-oxo-8H-acenaphtho[1,2-b]pyrrole-9-carbonitrile} and 3a {3-(3-dimethylamino-propylamino)8-oxo-8H-acenaphtho[1,2-b]pyrrole-9-carbonitrile} with the DNA double helix. In this study, we have determined the notably different DNA damage signal pathway elicited by 1a and 3a due to the different extents to which they unwind the DNA double helix. First, we have identified that ataxia-telangiectasiamutated (ATM) protein kinase can respond to DNA double helix unwinding caused by both 1a and 3a. In addition, the amount of ATM activation is consistent with the degree to which the DNA double helix was unwound. Consequently, we used 1a and 3a to semiquantitatively probe the response of RNA polymerase II (RNAPII) and p53 toward DNA double helix unwinding in vivo. By means of flow cytometry, immunocytochemistry, ChIP, quantitative real-time polymerase chain reaction, and Western blot analyses, we measured the level of p53 and RNAPII phosphorylation, in addition to the dynamics of the RNAPII distribution along the c-Myc gene. These results provided novel evidence for the impact of subtle DNA structural changes on the activity of RNAPII and p53. Moreover, DNA double helix conformational damage-dependent apoptosis was studied for the first time. These results indicated that 1 a can induce transcriptional blockage following a shift of the unphosphorylated IIa form of RNAPII to the phosphorylated IIo form, while 3a is unable to induce the same effect. Subsequently, p53 accumulation and phosphorylation events occur that lead to apoptosis in the case of 1a exposure. This suggests that the transcriptional blockage is also correlated to the degree of double helix unwinding. Furthermore, we found that the degree of DNA conformational damage determines whether or not apoptosis occurs through transcriptional blockage. Under our experimental conditions, ATM does not participate in the downstream events even when it has been activated. Thus, p53-mediated apoptosis may be independently triggered by transcriptional blockage. Introduction To maintain the integrity of the genome, mammalian cells activate a complex network of DNA damage signaling pathways to eliminate cells that contain a significant amount of DNA damage. In an extraordinary demonstration of the complexity of such systems, a plethora of distinct factors are known to be involved in these intricate pathways. These factors include ataxia-telangiectasia-mutated protein kinase (ATM), RNA polymerase II (RNAPII), and their effector p53, and all play a key role in response to DNA damage (1, 2). ATM protein kinase is the initial factor responsible for detecting DNA damage (3). ATM is known to act at an early stage in the signal transduction pathway in mammalian cells. After decades of research, * To whom correspondence should be addressed. Tel: 86-411-88993871. Fax: 86-411-83673488. E-mail: [email protected]. † State Key Laboratory of Fine Chemicals, Dalian University of Technology. ‡ School of Environmental and Biological Science and Technology, Dalian University of Technology. § Nankai University. | East China University of Science and Technology.

numerous aspects of the mechanisms underlying ATM activation have been elucidated. ATM has been shown to be an important protein involved in sensing DNA strand breaks caused by ionizing radiation. Notably, ATM senses the DNA strand breaks through the loss of chromatin topology (4). In addition, RNAPII has been suggested recently to act as a sensor of DNA damage in the template and to alert the cell by activating the DNA damage signaling pathway (5). Moreover, it has been established that RNAPII may detect DNA photodamage by UV light in addition to bulky DNA adduct lesions (6). As a consequence, the concept of transcription as an antitumor target is certainly worthy of further investigation. However, the detailed mechanisms by which ATM and RNAPII are triggered are still not well understood. In addition to the DNA lesions described above, there are other important types of damage, such as changes in the topology of the DNA double helix. Interestingly, the activity of ATM and RNAPII under conditions where the topology of the DNA double helix has been changed remains to be elucidated (7). DNA topology has long been known to be involved in the regulation of transcription in vivo (8). It is well-established that alterations

10.1021/tx800288v CCC: $40.75  2009 American Chemical Society Published on Web 01/30/2009

484

Chem. Res. Toxicol., Vol. 22, No. 3, 2009

Figure 1. Structures of 1a and 3a.

of the tertiary structure of DNA are involved in a variety of transcription events. While superhelicity is inherent to cellular DNA due to the spooling of DNA around histones, RNAPII itself may also introduce DNA torsional tension, thus creating both positive and negative supercoils in the DNA that are located upstream and downstream of RNAPII, respectively (9). Because transcription is a process that undresses the DNA from its histone coat and separates the two strands of DNA, it is reasonable to speculate that the topology of the double helix is also linked to transcription (10). Several studies have been performed regarding the effects of a DNA binding drug on the process of transcription in vitro, which can be correlated to a change in the double helix conformation (11). With recent advances in cell biology and chemical biology, this longstanding problem can now be reexamined in vivo. The same problem is relevant in the context of ATM activity. It has been demonstrated that ATM monitors DNA strand breaks by sensing a change in the higher order chromatin structure. Previous results indicate that ATM is activated by hypotonic conditions or via exposure to chloroquine (4), and it is concluded that ATM can sense alterations in the DNA tertiary structure. However, the question remains as to whether or not ATM can sense a conformational change in the DNA double helix. The precise nature of the aforementioned chromatin structural alterations that activate ATM, in addition to the impact that these factors may have on ATM activation, remains to be explored (12-14). Interestingly, the network of damage signaling pathways is so complex that the factors involved in these pathways often overlap functionally, rather than being strictly dependent on the activity of a partner protein (15). For example, while p53 serves as one of the substrates that are activated by ATM, p53 can also be triggered by transcriptional blockage (16). So, the question remains as to whether or not ATM and RNAPII can act synergistically to stimulate p53 and thus induce apoptosis. The investigation into the response of ATM and RNAPII to conformational changes in the DNA double helix is important not only for the understanding of the DNA damage signaling pathway but also for the development of antitumor therapeutics. Most of the known DNA intercalators, a very large and everevolving class of noncovalent antitumor agents (17), only change the DNA double helix conformation through winding and unwinding effects (18). Thus, it is still unknown whether or not the RNAPII transcription machinery and ATM activation contribute to the antitumor properties brought about by this class of drugs. Here, we investigated the activation and dynamics of these aforementioned factors in cells after exposure to two noncovalent DNA interacting analogues 1a and 3a (Figure 1). The intercalation geometries of these drugs with respect to DNA were illustrated in detail in our previous report (19). Specifically, these drugs intercalate into the base pair and subsequently transform B-form DNA to an unwound A-like form. However, the intercalation modes are different for the two analogues. For instance, 1a intercalates in a fashion where the long axis of the

Zhang et al.

molecule is positioned in parallel to the long axis of the DNA base pairs, while 3a is found to be positioned roughly pependicular to the base pairs. As a consequence, 1a is able to unwind the double helix to a more significant extent than 3a. Interestingly, we found that only 1a can induce cell apoptosis, while our observations indicate that 3a cannot induce apoptosis. In the current study, we have characterized the involvement of ATM, RNAPII, and p53 in the apoptotic pathway in the presence of 1a and 3a. Moreover, we have made a semiquantitative assessment of the impact of the topology of DNA double helix damage on ATM, p53, and RNAPII by exploiting the inherent ability of 1a and 3a to alter the DNA conformation. Thus, for the first time, the role of ATM and RNAPII during the DNA double-helix-unwinding-triggered p53 response was explored. Our results suggested that although ATM can be activated by the unwinding of the DNA double helix in vivo, the blockage of transcription is the key feature that independently determines whether or not the cell undergoes apoptosis.

Experimental Procedures Chemicals and Chemical Treatment of Cells. Both 1a and 3a were synthesized in our laboratory according to our previous report (19), and the structures are shown in Figure 1. The purity of the compounds was examined using HPLC in two systems, and each compound was observed to be greater than 99% pure. Both 1a and 3 a were dissolved in dimethyl sulfoxide (DMSO) (10 mM) and were subsequently stored as stock solutions in dark-colored bottles at 4 °C. Each stock was diluted to the desired concentration immediately before addition to the growth media. The ATM inhibitor KU-55933 was purchased from Merck. The cells were exposed to either 1a or 3a at different concentrations ranging from 0 to 10 µM for different periods of time (0-36 h). Cells that were grown in media containing an equivalent amount of DMSO without either of the intercalators served as a control. Antibodies. The antibodies used in this study include the A-10 antibody against RNAPII (sc-17798, Santa Cruz), the H5 antibody against RNAPII CTD that is phosphorylated on residue Ser2 (MMS129R, Covance), the DO-1 antibody against p53 (sc-126), the 10H11.E12 antibody against ATM that is phosphorylated on Ser1981 (sc-47739, Santa Cruz), the C4 antibody against β-actin (sc-47778, Santa Cruz), the Phospho-Histone H2AX (Ser 139) antibody (2577, Cell Signaling), and the Phospho-p53 (Ser15, Ser20) antibody (9284, 9287, Cell Signaling). Cell Culture. Human MCF-7 cells were cultured in RPMI1640 medium containing 10% (v/v) fetal calf serum. Exponentially growing cells were exposed to various concentrations of 1a or 3a for the indicated times. Flow Cytometry Analysis. For the apoptosis assays, hypodiploid DNA was measured as described previously (20). Briefly, 106 cells were centrifuged and fixed in 1 mL of ice-cold 70% ethanol for 12 h, washed once with PBS, and resuspended in 1 mL of PBS containing 0.04 mg of RNase A and 0.05 mg of propidium iodide. After incubation at room temperature for 30 min, the cells were analyzed. The fluorescence from 30000 cells was counted for each sample. The results were analyzed using a FACScan flow cytometer in combination with CellQuest software (BD Biosciences, San Jose, CA). For the expression assays of p53 and phospho-p53, cells were harvested, treated with trypsin (0.1% trypsin in PBS for 5 min), fixed in 1 mL of 95% ethanol, and kept at room temperature for 20 min before analysis of antibody binding. Antibody staining was performed as described previously (21). Briefly, 5 × 105 cells were centrifuged and resuspended with p53, p-p53 (Ser15), or p-p53 (Ser20) antibodies. After 1 h, cells were rinsed and resuspended in a medium containing a secondary antibody conjugated with FITC. The cell pellets were finally resuspended in PBS. The fluorescence of 10000 cells was analyzed using a FACScan flow cytometer in combination with CellQuest software (BD Biosciences). All of the experiments were repeated at least three times.

DNA Damage Signaling Pathway Triggered by a DNA Intercalator

Chem. Res. Toxicol., Vol. 22, No. 3, 2009 485

Figure 2. Activation of ATM by 1a and 3a in MCF-7 cells. MCF-7 cells grown in 35 mm glass bottom culture dishes were incubated with 0.1% DMSO alone (a and e) or different concentrations of 1a (b, 2 µM; c, 4 µM; and d, 6 µM) and 3a (f, 6 µM; g, 8 µM; and h, 10 µM) for 12 h. Cells were fixed, and images of immunofluorescence for p-ATM (Ser1981) were taken by confocal microscope. A 60× object lens was used. Excitation wavelength, 488 nm; laser power, 48.50%; iris, 5.40 mm; and offset, 0.

Chromatin-Immunoprecipitation (ChIP) Procedures. Chromatin preparation and ChIP procedures have been described previously (8). Briefly, cells were fixed with 1% (v/v) formaldehyde for 15 min at room temperature, and the chromatin was sonicated to an average DNA fragment size of 300-400 bp. The chromatin was equilibrated in RIPA buffer, and samples were precleared with nonimmune rabbit serum. Equal amounts of precleared chromatin were then incubated with the specific antibody or nonimmune IgG to measure background recoveries. The immunoprecipitated DNA was purified after proteinase K treatment and phenol extraction. The DNA recovery was measured using a SYBR Q-PCR (SYBR Premix Ex Taq kit for Real Time, TaKaRa, Dalian, China) with a Smart Cycler II (TaKaRa, United States). The polymerase chain reaction (PCR) was performed in a final volume of 25 µL containing primers (0.5 µM) and amplified using the following conditions: a step at 95 °C for 10 s followed by a step composed of a 5 s period at 95 °C and a 20 s period at 60 °C for a total of 45 cycles. The primers to amplify the fragment A (P2 promotor) gene were 5′GAGAAGGGCAGGGCTTCTCA-3′ (sense) and 5′-TCTGCCTCTCGCTGGAATTA-3′ (antisense). The primers to amplify the fragment B (exon fragment) gene were 5′-CCTCTGTTGAAATGGGTCTGG-3′ (sense) and 5′-CCTTTGCCTACCTCTCACCTTCT-3′ (antisense). The primers to amplify the C fragment (intron fragment) were 5′-TCCACCTCCAGCTTGTACCTG-3′ (sense) and 5′-TCGTTGAGAGGGTAGGGGAAG-3′ (antisense). These experiments were repeated in triplicate. The specificity of the PCR products was routinely controlled by performing melting curve analyses and analyzing via agarose gel electrophoresis. In addition, several criteria were followed to verify that each measurement and the recovered DNA were normalized relative to the maximal recovery. Immunofluorescence Staining. This procedure was performed as described previously (22). Briefly, after cold ethanol fixation at -20 °C for 10 min, cultures were incubated with normal goat serum for 30 min. The cultures were then stained with the p-ATM, p-p53 (Ser15), and p-p53 (Ser20) primary antibodies for 2 h at 37 °C. After they were washed with PBS, the cells were incubated with FITC-conjugated secondary antibodies for 1 h at 37 °C and analyzed with a Nikon TE 2000 inverted microscope or a Radiance 2000 laser scanning confocal microscope (Bio-Rad, United States) equipped with an objective lens (60×/1.4 oil) and operated using the Laserpix 4.0 software. Western Blot Analysis. After cells were exposed to each type of cell-damaging treatment, the cells were lysed in a lysis buffer (62.5 mM Tris-HCl, pH 6.8, 2% w/v SDS, 10% glycerol, 50 mM DTT, and 1 mM PMSF). Whole cell lysates, corresponding to 6 × 105 cells, or 10-15 µg of nuclear protein, were separated by SDSPAGE. After electrophoresis, the proteins were transferred to a PVDF membrane (Millipore). The membranes were subsequently blocked with a blocking solution containing 5% nonfat dry milk in TBS-T for 1 h at room temperature and were then incubated with each antibody for 1 h at room temperature. The membranes were

also incubated with a peroxidase-conjugated IgG, and the expected proteins were detected using the ECL method (Amersham Pharmacia Biotech). Measurement of RNA Synthesis. Exponentially growing cells were exposed to 1a at various concentrations (1, 2, 4, and 6 µM). After a 12 h incubation period at 37 °C, the medium was replaced with a 1a-free medium containing 1 µCi [5-3H]uridine (Atom HighTech Co., Ltd. Beijing), and the cells were cultured for an additional 12 h. The incorporated radioactivity was measured using Ultimacold liquid scintillation counting fluid (Packard, United States). All of the experiments were repeated at least three times. Statistical Analysis. The data are presented as means ( SEMs with p < 0.01 considered significant. Statistical significance was determined using the Student’s unpaired t test with equal variance.

Results ATM Is Activated by 1a and 3a through Chromatin Conformational Damage in Vivo. Although previous in vitro experiments had shown that 1a and 3a could cause a DNA conformational change without inducing a strand break, we wanted to determine whether or not such an event could occur in vivo without generating a DNA break. Because ATM is the only factor known to initiate the signaling pathways in cells following the alteration of higher order DNA structure, we first performed spatiotemporal analysis of the pS1981-ATM foci using immunofluorescence staining. After 12 h of exposure, the phosphorylation of ATM in the nucleus was detected at concentrations of 4 µM 1a and 8 µM 3a (Figure 2). Subsequent analysis indicates that the foci increased in a concentrationdependent manner for both 1a and 3a. However, it is worth noting that although the working concentration of 1a was lower than that of 3a, the number of foci induced by 1a was greater than that induced by 3a. At a concentration of 6 µM 1a, the average number of ATM foci was 32 ( 1.5, which was significantly higher than that observed at a concentration of 8 µM 3a (15 ( 0.5, p < 0.5). However, ATM can respond to both DNA conformational changes and DNA strand breaks. To determine whether or not strand breaks were induced by the two chemicals in vivo, we examined complex formation with γ-H2AX, which is a substrate of ATM. γ-H2AX complex formation is a marker of DNA double strand breaks since its phosphorylation is confined to chromatin domains surrounding the DNA strand breaks (23, 24). Our results indicated that no obvious γ-H2AX foci were induced as a consequence of these two chemical analogues (1a and 3a), even at concentrations known to induce ATM foci (Figure 3). These findings are consistent with the mechanism whereby

486


Zhang et al.

Figure 3. No activation of γ-H2AX by 1a and 3a in MCF-7 cells. MCF-7 cells were incubated with 0.1% DMSO alone (a) or 6 µM 1a (b) and 10 µM 3a (c) for 12 h, and cells were exposed to 1 Gy IR for 1 h as a positive control (d). (A) Cells were fixed, and images of immunofluorescence for γ-H2AX were taken by confocal microscope. A 60× object lens was used. (B) γ-H2AX was detected by Western blot. a, 0.1% DMSO as control; b, 6 µM 1a; c, 10 µM 3a; and d, 1 Gy IR.

Figure 4. Apoptosis and cell cycle arrest in the S phase of MCF-7 cells induced by 1a rather than 3a. Cells were seeded into six well plates and treated with drug vehicle (0.1% DMSO) alone or the indicated concentrations of 1a (A) and 3a (B) for 12 h. Cells were harvested and assayed by flow cytometry. At least 10000 events were counted for each sample. Bars are means of duplicates from three independent experiments. Significance is indicated by asterisks: **, P < 0.01, OriginPro 7.5.

changes in chromatin structure activate ATM in the absence of DNA strand breaks, since ATM substrates such as γ-H2AX that would typically be phosphorylated at the site of DNA breaks fail to become phosphorylated. Together, these results indicated that 1a and 3a can induce DNA conformational changes in vivo. Furthermore, 1a stimulated the phosphorylation of ATM more efficiently than 3a. 1 a-Induced p53-Dependent Apoptosis. In our previous report, we illustrated that 1a, but not 3a, could lead to apoptosis. This indicates that there may be different underlying mechanisms that determine the fate of cells when exposed to either 1 a or 3a. Here, the activation of ATM as a consequence of 1a and 3a treatment demonstrated that the DNA damage signaling pathway may respond to DNA conformational alterations. In particular, the different degrees to which ATM was activated as a function of the varying extents of DNA double helix unwinding presented the possibility that these drugs may be exploited to probe the various signal factors involved in the pathways triggered in response to these drugs. The results of flow cytometry analysis indicated that 1ainduced apoptosis and S phase arrest occurred in a dosedependent manner (Figure 4). However, consistent with our previous studies, no apoptotic induction was observed for 3a. To exploit the mediator of apoptosis induction by 1a, we measured the effects on p53 since p53 plays a pivotal role in the response to a wide spectrum of DNA damage and is known to induce the apoptotic process. After MCF-7 cells were cultured with increasing concentrations of 1a and 3a (2, 4, 6, and 10 µM) for 12 h, we subsequently counted 10000 cells from each sample

using a FACScan flow cytometer and analyzed these data using the CellQuest software. As shown in Figure 5A, we determined that 1a was an efficient trigger for p53 accumulation, whereas 3 a could not trigger the accumulation of p53. Interestingly, p53 started to be upregulated at the same concentration that apoptosis occurred (2 µM), and p53 levels increased 3.99%, which was significantly higher than that observed in the control (P < 0.01). A dose-dependent trend was observed for analogue 1a, where the stimulation of p53 reached 15.24% at a concentration of 6 µM. Both the time points and the concentration frame are consistent with the occurrence of apoptosis. In the case of the 3 a treatment, no p53 activation was observed even at a concentration of 10 µM after 12 h of exposure. Collectively, these data indicate that only analogue 1a induced p53-dependent apoptosis. We also measured the phosphorylation of p53 under our experimental conditions. We performed flow cytometry analysis and immunofluorescence staining to assess the phosphorylation of p53 at amino acid residues Ser15 and Ser20 in chemically treated MCF-7 cells. As shown in Figure 5B,C, 1a-induced phosphorylation of p53 residues Ser15 and Ser20 was observed at a concentration of 2 µM, and Ser15 and Ser20 phosphorylation increased to 12.99 and 10.51% at 6 µM, respectively. No obvious phosphorylation was observed upon exposure to 3 a. The immunofluorescence staining of p53, which allowed us to determine the phosphorylation state following treatment with 1 a and 3a, is shown in Figure 5D. These results further support the conclusion that p53 is a mediator of 1a-induced apoptosis.



Figure 5. Activation and phosphorylation of p53 following 1a treatment. MCF-7 cells were grown on six well plates. At 40% confluence, cells were treated for 12 h with 1a and 3a, respectively. (A) Activation of p53. Cells were fixed and stained with an anti-p53 antibody (DO-1) for flow cytometry analysis of mean cellular fluorescence relative to an untreated control sample. The graphic represents a quantification of cells activated. The means and standard errors for three independent experiments are shown. Significance is indicated by asterisks: **, P < 0.01, OriginPro 7.5. (B and C) Phosphorylation of p53. Cells were fixed, permeabilized, and labeled with an anti-p-p53 antibody. The expression of p-p53 (Ser15) and p-p53 (Ser20) was assayed by flow cytometry. (D) Immunofluorescence staining of p-p53. MCF-7 cells were treated for 24 h with 4 µM 1a and 10 µM 3a, respectively. Cells were fixed, permeabilized, and labeled with an anti-p-p53 antibody. Images were taken by an inverted microscope. A 20× object lens was used. (E) Western blot detection of activation and phosphorylation of p53 following 1a and 3a treatment. Total protein extracts from cells previously treated with compounds at indicated doses for 12 h. 1a but not 3a caused concentration dependence of the p53 response.

488


Figure 6. Western blot detection of 1a-induced phosphorylation and degradation of the level of RNAPII. IIo and IIa are the phosphorylaeted and unphosphorylated forms of the large subunit of RNAPII, respectively. (A) 1a but not 3a caused concentration dependence of RNAPII phosphorylation and degradation. Total protein extracts from cells previously treated with 2, 4, and 6 µM 1a and 4, 6, and 10 µM 3a, respectively, for 12 h. (B) Compound 1a but not 3a caused timedependent phosphorylation and degradation of the level of RNAPII. Total protein extracts from cells previously treated with 2 µM 1a and 10 µM 3a for the indicated time courses.

Phosphorylation and Degradation of RNAPII in a-Treated Cells. The pathway triggered by 1a appears to be a very interesting pathway. Our results indicate that 2 µM is the lowest concentration of 1a that can induce the p53-dependent apoptotic pathway. This is interesting since this concentration is lower than the concentration that is required to activate ATM, which we found to be 4 µM (Figure 2). Although the p53 protein is only one of the substrates that can be activated by ATM, it is not likely that ATM activation could account for the accumulation of p53 and phosphorylation at a 1a analogue concentration of 2 µM. Thus, we sought to identify other factors or machinery that may be involved in the 1a-induced apoptotic pathway. We hypothesized that RNAPII may be phosphorylated in the initial steps of DNA damage and may thus account for the aforementioned disparity. We were therefore interested in determining whether or not 1a and 3a have different abilities in phosphorylating and degrading RNAPII. We subsequently subjected MCF-7 cells to media containing various concentrations of 1a and 3a. The concentration ranges were chosen according to the results from the p53 simulation experiments. Twelve hours after the addition of the specific chemical, the cells were harvested. As shown in Figure 6A, after a 12 h incubation period with 1a, the RNAPII phosphorylation state was monitored by observing a shift from the unphosphorylated IIa form of the polymerase (pol IIa) to the phosphorylated IIo form (pol IIo). This shift occurred at 2 µM 1 a, which is the lowest concentration at which we observed a reduction in the relative amount of pol IIa. Interestingly, with an increase in the concentration of 1a, we observed that the amount of pol IIa decreased until the concentration of 1a reached 6 µM at which the pol IIa form was completely undetectable. No significant variation of either the pol IIa or the pol IIo levels was observed when the concentration of 3a was varied incrementally from 4 to 10 µM (Figure 6A). Subsequently, we performed a series of experiments to explore the effect of the exposure time on the RNAPII phosphorylation state. We subjected MCF-7 cells to media containing 2 µM 1a or 10 µM 3a. The cells were harvested at different time points ranging from 0 to 36 h of exposure. As 1

Zhang et al.

indicated in Figure 6B, we observed that 1a treatment induced significant phosphorylation of RNAPII that was followed by degradation of RNAPII. However, treatment with 3a did not cause any notable changes in the phosphorylation state of RNAPII (Figure 6B). Thus, 1a triggers both the phosphorylation and the reduction of the level of RNAPII, while 3a causes no visible change in either the phosphorylation state or the protein levels of the enzyme. Because the ubiquitylation and subsequent degradation of RNAPII are a general response to stalled or arrested elongation complexes (25), we were now curious to know if transcriptional blockage is involved in the signaling pathway triggered by a change in DNA conformation. Inhibition of RNAPII Translocation along the c-Myc Gene Following Exposure to 1a. We expanded the analyses to probe the dynamics of RNAPII distribution along a transcribed gene following treatment with each of the two analogues. The chromatin binding site of RNAPII was mapped using ChIP, and the DNA recovery levels were determined using real-time PCR. We chose the proto-oncogene c-Myc as a target since premature termination and pausing of RNAPII have been observed with this gene in vivo (8). We examined the RNAPII levels bound to the P2 promotor fragment (A), exon fragment (B), and intron gene fragment (C). Because the phosphorylation and degradation of RNAPII were observed after 12 h of exposure to 2 µM 1a, we therefore used the same concentration (2 µM) and shorter time periods, ranging from 8 to 12 h of exposure to 1a, to prevent significant protein loss due to degradation. First, the amount of RNAPII observed at the P2 promoter was significantly larger than that found at other regions along the transcribed sequence (Figure 7). This is in agreement with the presence of a pause site at the P2 promoter. Following a plateau at 8-10 h of 1a treatment, RNAPII levels on all of the fragments probed significantly increased at 12 h of treatment. This time point is precisely when our previous experiments indicated that the phosphorylation of RNAPII occurred. The tendency of RNAPII to accumulate along the transcribed DNA, including accumulation at the promoter, the exon, and the intron fragments, was observed. It has been well-established that a transcriptional stress response that is followed by a change in the phosphorylation status of RNAPII can occur at the early steps of the response to DNA damage. Camptothecin, for example, has been shown to induce a change in the dynamics of the RNAPII distribution within a few minutes. Under our conditions, we did not observe the phosphorylation and transcriptional blockage of RNAPII until 12 h post-treatment with 1a. A possible explanation for this may be that the DNA damage was so insignificant that only double helix unwinding occurred without the generation of any DNA strand breaks. While these changes were statistically significant, the relatively small increase in RNAPII levels on the P2 promotor may also due to a low level of DNA damage. In stark contrast to the 1a analogue, no significant change in the RNAPII levels was observed at any of the target fragments of the c-Myc gene, even at 3a concentrations up to 10 µM and after 12 h of treatment. Thus, we have observed 1a-induced transcriptional blockage. To verify this phenomenon further, we performed measurements of RNA synthesis via incorporation of radiolabeled uridine. As shown in Table 1, RNA synthesis in tumor cells was significantly inhibited by 1a at 2 µM, which is consistent with the results of both Western blot analyses and CHIP. These results convincingly demonstrated that transcriptional blockage occurred after exposure to 1a.



Figure 7. Dynamics of RNAPII levels along the c-Myc gene during 2 µM 1a (A) and 10 µM 3a (B) treatments, respectively. RNAPII bound to fragments a, b, c, and control (white bar) following chemicals treated for indicated time periods are shown. DNA recovery is normalized relative to the fragment recovery of untreated cells. The means and standard errors for three independent experiments are shown. Significance is indicated by asterisks: **, P < 0.01, OriginPro 7.5.

Table 1. Inhibitory Effect of 1a on the Uptake of [5-3H]Uridine (*P < 0.05 vs Control Group) 1

a (µM)

[5-3H] uridine incorporation (cpm)

control

1

2

4

6

2897 ( 223

2789 ( 234

1679 ( 120*

1520 ( 267*

1300 ( 340*

Altogether, our findings have indicated that the molecular responses to 1a and 3a were characterized by significantly different effects on the level of RNAPII localized to the transcribed gene, and the 1a analogue specifically caused transcriptional blockage. This molecule not only prohibited the transition of RNAPII from the initiation form to the elongation form but also inhibited the translocation of RNAPII along the entire length of the gene. Thus, it is the transcriptional blockage that allows for the accumulation and phosphorylation of p53, which in turn induces apoptosis in the case of 1a treatment. Transcriptional Blockage Independently Triggered p53 Accumulation. We have also concluded that the transcriptional blockage triggered p53-mediated apoptosis at 2 µM 1a exposure, a concentration at which ATM was not activated. To estimate the contribution of ATM activation to apoptosis, we designed experiments over a range of concentrations (4 µM and upward) and performed further studies by using the ATM inhibitor KU55933 (2-morpholin-4-yl-6-thianthren-1-ylpyran-4-one). We exposed MCF-7 cells to either 1a alone or to 1a in the presence of KU-55933. Subsequently, p53 positive cells were analyzed via flow cytometry. Interestingly, the presence of KU-55933 did not have any effect on the accumulation of p53. After treatment with 1a (4 µM), 9.1% of cells were p53 positive, while 9.86% of cells were positive in the presence of KU-55933 (Figure 8). At concentrations of 6 µM 1a, the accumulation of p53 in the presence or absence of KU-55933 was unchanged overall, and 15.24 and 15.05% of cells were positive, respectively. Thus, we can conclude that ATM did not participate in 1 a-induced p53 accumulation even when ATM is known to have been activated.

Herein, we have found that the unwinding of the DNA double helix caused by intercalating agents may function as the signal to ATM in living cells. Besides ionizing radiation and ultraviolet light damage, topological damage that results in various conformational perturbations in DNA is a new type of DNA damage that can be detected by ATM. Moreover, we have demonstrated that the amount of ATM protein that becomes phosphorylated depends on the degree to which the DNA helix is unwound. This is illustrated by our finding that the 1a analogue, which unwinds the double helix more significantly than 3a, activates ATM at a relatively lower concentration than 3 a, and the number of ATM foci induced by 1a is significantly greater than that induced by 3a. In our previous report, we demonstrated that 1a, but not 3a, can induce apoptosis. This property contributes to the higher toxicity of 1a and the more favorable potency as an antitumor agent. Because both of these analogues can activate ATM, a question then arises regarding the different mechanisms medi-

Discussion In this study, two analogues that are able to unwind the DNA helix to different extents were exploited to probe how ATM, p53, and RNAPII function when they encounter topological damage in the DNA double helix. ATM initiates the signaling pathways in mammalian cells following exposure to agents that introduce DNA damage. We have studied ATM as a first step in elucidating the cellular mechanisms that deal with topological damage to the DNA double helix.

Figure 8. Compound 1a-induced activation of p53 in the absence and presence of ATM-specific inhibitor KU-55933. MCF-7 cells were grown on six well plates. At 40% confluence, cells were treated for 12 h with the indicated concentrations of 1a alone or indicated concentrations of 1 a and a certain concentration of KU-55933 (10 µM). Cells were then fixed and stained with an anti-p53 antibody (DO-1) for flow cytometry analysis of mean cellular fluorescence relative to an untreated control sample. The graphic represents a quantification of cells activated.

490


ated by the two chemicals during the induction of apoptosis. Therefore, we investigated the possible mechanisms involved in the induction of apoptosis by these two chemicals. As one of the downstream substrates of ATM, p53 is expected to be phosphorylated following the activation of ATM. While it is not completely clear why the activated ATM does not result in the downstream phosphorylation of p53 in the case of 3a treatment, it is conceivable that the relatively lower amount of activated ATM is not sufficient to have an effect on p53. Bakkenist and co-workers proposed that the amount of ATM phosphorylation may reach a maximum at seemingly lower doses of IR where all of the ATM in cells has been activated (4). Our study raised the possibility that there may be a minimal amount of ATM activation that is required to trigger the downstream effectors. Surprisingly, 1a triggered the p53-dependent apoptosis pathway at lower concentrations than it activated ATM. We speculate that transcriptional blockage may contribute to 1ainduced apoptosis. On the one hand, previous research has demonstrated that transcriptional blockage may be involved in p53-dependent apoptosis that is related to S phase arrest (26). It has also been well-established that blockage of transcription may phosphorylate p53 at Ser15 (27), as we observed in this study. On the other hand, we wanted to determine whether or not RNAPII could recognize DNA conformational damage as well as DNA strand breaks (6). Thus, the time- and dose-dependent phosphorylation of RNAPII and its subsequent degradation after 1a treatment were assayed. Interestingly, a shift of the pol IIa form to the IIo form was detected at a concentration as low as 2 µM. Intriguingly, this was the concentration that could induce the p53 response, including both p53 accumulation and phosphorylation. Simultaneously, RNAPII accumulated along the c-Myc gene. These results strongly suggested that the transcriptional blockage indeed occurred as a consequence of 1a treatment (8, 28). However, until now, the initial molecular event has been proposed to be transcriptional blockage due to 1a exposure via conformational damage to chromatin at a concentration of 2 µM. Subsequently, the blockage triggered the phosphorylation and accumulation of p53, which then led to apoptosis. We attempted to estimate the contribution of ATM activation to p53-dependent apoptosis at concentrations of 4 and 6 µM 1a where ATM was activated. However, these results indicated that ATM did not participate in the 1a-induced accumulation of p53. On the one hand, these data validated our hypothesis that there is a threshold of ATM phosphorylation that is required to trigger p53, such that if the activation does not exceed the threshold, the downstream effectors will not be signaled. However, on the other hand, we have concluded that transcriptional blockage is sufficient to determine whether or not cells undergo apoptosis. Interestingly the results obtained from experiments involving 3 a also support this conclusion. In the absence of transcription blockage, ATM activation itself is unable to trigger apoptotic events. The activation of ATM is required for most, if not all, of the relevant cellular responses to irradiation; thus, the investigation of the responses to DNA damage should not focus simply on events occurring at the DNA break but rather should include a consideration of more complex nuclear events (4). Herein, we have provided evidence not only for the idea that subtle structural elements can influence ATM activation but also that there is a possible threshold for ATM to signal its effectors. These results may provide new insight into ATM activation.

Zhang et al.

The results that we present here suggested an interesting aspect of the DNA damage signaling pathway during the unwinding of the DNA double helix in vivo. Blattner and coworkers proposed that following IR exposure and the subsequent formation of DNA strand breaks, ATM plays a key role in triggering the p53 response (29). Following UV light irradiation, however, p53 has been shown to be triggered by blockage of transcription (30). Our current study provided a mechanism for the damage pathway triggered by two DNA intercalators. In addition, both ATM activation and transcriptional blockage can be detected under these conditions. However, transcriptional blockage is the primary factor that influences this event. The damage by itself is not sufficient to induce apoptosis until it is able to induce transcriptional blockage. Thus, these studies support the concept of transcriptional blockage serving as a therapeutic target for anticancer therapy. The development of antitumor agents that act by DNA intercalation should target transcriptional blockage. Acknowledgment. This work was supported by the National Natural Science Foundation of China (30772622).

References (1) Ljungman, M. (2005) Activation of DNA damage signaling. Mutat. Res. 577, 203–216. (2) Meek, D. W. (2004) The p53 response to DNA damage. DNA Repair 3, 1049–1056. (3) Stiff, T., Walker, S. A., Cerosaletti, K., Goodarzi, A. A., Petermann, E., Concannon, P., O′Driscoll, M., and Jeggo, P. A. (2006) ATRdependent phosphorylation and activation of ATM in response to UV treatment or replication fork stalling. EMBO J. 25, 5775–5782. (4) Bakkenist, C. J., and Kastan, M. B. (2003) DNA damage activates ATM through intermolecular autophosphorylation and dimer dissociation. Nature 421, 499–506. (5) Ljungman, M., and Lane, D. P. (2004) Transcription- guarding the genome by sensing DNA damage. Nature 4, 727–737. (6) Brueckner, F., and Cramer, P. (2007) DNA photodamage recognition by RNA polymerase II. FEBS Lett. 581, 2757–2760. (7) Scicchitano, D. A., Olesnicky, E. C., and Dimitri, A. (2004) Transcription and DNA adducts: What happens when the message gets cut off? DNA Repair 3, 1537–1548. (8) Khobta, A., Ferri, F., Lotito, L., Montecucco, A., Rossi, R., and Capranico, G. (2006) Early effects of topoisomerase I inhibition on RNA polymerase II along transcribed genes in human cells. J. Mol. Biol. 357, 127–138. (9) Giaever, G. N., and Wang, J. C. (1988) Supercoiling of intracellular DNA can occur in eukaryotic cells. Cell 55, 849–856. (10) Lavelle, C. (2007) Transcription elongation through a chromatin template. Biochimie 89, 516–527. (11) Straney, D. C., and Crothers, D. M. (1987) Effect of drug-DNA interactions upon transcription initiation at the Lac promoter. Biochemistry 26, 1987–1995. (12) Lee, J. H., and Paull, T. T. (2004) Direct activation of the ATM protein kinase by the Mre11/Rad50/Nbs1 complex. Science 304, 93–96. (13) Mochan, T. A., Venere, M., Jr., and Halazonetis, T. D. (2003) 53BP1 and NFBD1/MDC1-Nbs1 function in parallel interacting pathways activating ataxia-telangiectasia mutated (ATM) in response to DNA damage. Cancer Res. 63, 8586–8591. (14) Goodarzi, A. A., Jonnalagadda, J. C., Douglas, P., Young, D., Ye, R., Moorhead, G. B., Lees-Miller, S. P., and Khanna, K. K. (2004) Autophosphorylation of ataxia-telangiectasia mutated is regulated by protein phosphatase 2A. EMBO J. 23, 4451–4461. (15) Svejstrup, J. Q. (2007) Contending with transcriptional arrest during RNAPII transcript elongation. Trends Biochem. SCI 32, 165–171. (16) van Gijssel, H. E., Mullenders, L. H., van Oosterwijk, M. F., and Meerman, J. H. (2003) Blockage of transcription as a trigger for p53 accumulation by 2-acetylaminofluorene DNA adducts. Life Sci. 73, 1759–1771. (17) Brana, M. F., and Ramos, A. (2001) Naphthalimides as anticancer agents: Synthesis and biological activity. Curr. Med. Chem.-AntiCancer Agents 1, 237–255. (18) Li, H. H., Aubrecht, J., and Fornace, A. J., Jr. (2007) Toxicogenomics: Overview and potential applications for the study of non-covalent DNA interacting chemicals. Mutat. Res. 623, 98–108.

DNA Damage Signaling Pathway Triggered by a DNA Intercalator (19) Zhang, Z., Yang, Y., Zhang, D., Wang, Y., Qian, X., and Liu, F. (2006) Acenaphtho[1,2-b]pyrrole derivatives as new family of intercalators: Various DNA binding geometry and interesting antitumor capacity. Bioorg. Med. Chem. 14, 6962–6970. (20) Gaiddon, C., Jeannequin, P., Bischoff, P., Pfeffer, M., Sirlin, C., and Loeffler, J. P. (2005) Ruthenium (II)-derived organometallic compounds induce cytostatic and cytotoxic effects on mammalian cancer cell lines through p53-dependent and p53-independent mechanisms. J. Pharmacol. Exp. Ther. 315, 1403–1411. (21) Reitsema, T., Klokov, D., Bana´th, J. P., and Olive, P. L. (2005) DNAPK is responsible for enhanced phosphorylation of histone H2AX under hypertonic conditions. DNA Repair 4, 1172–1181. (22) Wang, C. G., and Chen, J. D. (2003) Phosphorylation and hsp90 binding mediate heat shock stabilization of p53. J. Biol. Chem. 278, 2066–2071. (23) Rogakou, E. P., Pilch, D. R., Orr, A. H., Ivanova, V. S., and Bonner, W. M. (1998) DNA double-stranded breaks induce histone H2AX phosphorylation on Serine 139. J. Biol. Chem. 273, 5858–5868. (24) Rogakou, E. P., Boon, C., Redon, C., and Bonner, W. M. (1999) Megabase chromatin domains involved in DNA double-strand breaks in ViVo. J. Cell. Biol. 146, 905–916.

Chem. Res. Toxicol., Vol. 22, No. 3, 2009 491 (25) Luo, Z., Zheng, J., Lua, Y., and Bregman, D. B. (2001) Ultraviolet radiation alters the phosphorylation of RNA polymerase II large subunit and accelerates its proteasome-dependent degradation. Mutat. Res. 486, 259–274. (26) Garssen, J., van Steeg, H., de Gruijl, F., de Boer, J., van der Horst, G. T. J., van Kranen, H., van Loveren, H., van Dijk, M., Fluitman, A., Weeda, G., and Hoeijmakers, J. H. J. (2000) Transcription-coupled and global genome repair differentially influence UV-B-induced acute skin effects and systemic immunosuppression. J. Immunol. 164, 6199–6205. (27) Ljungman, M., O’Hagan, H. M., and Paulsen, M. T. (2001) Induction of ser15 and lys382 modifications of p53 by blockage of transcription elongation. Oncogene 20, 5964–5971. (28) Weber, A., Liu, J., Collins, I., and Levens, D. (2005) TFIIH operates through an expanded proximal promoter to fine-tune c-myc expression. Mol. Cell. Biol. 25, 147–161. (29) Blattner, C., Sparks, A., and Lane, D. (1999) Transcription factor E2F-1 is upregulated in response to DNA damage in a manner analogous to that of p53. Mol. Cell. Biol. 19, 3704–3713. (30) Ljungman, M. (2000) Dial 9-1-1 for p53: Mechanisms of p53 activation by cellular stress. Neoplasia 2, 208–225.

TX800288V

DNA Double Helix Unwinding Triggers Transcription Block-Dependent

Recommend Documents