|Home | About | Journals | Submit | Contact Us | Français|
Acrolein (Acr) is a potent cytotoxic and DNA damaging agent which is ubiquitous in the environment and abundant in tobacco smoke. Acr is also an active cytotoxic metabolite of the anti-cancer drugs cyclophosphamide and ifosfamide. The mechanisms via which Acr exerts its anti-cancer activity and cytotoxicity are not clear. In this study, we found that Acr induces cytotoxicity and cell death in human cancer cells with different activities of p53. Acr preferentially binds nucleolar ribosomal DNA (rDNA) to form Acr-deoxyguanosine adducts, and induces oxidative damage to both rDNA and ribosomal RNA (rRNA). Acr triggers ribosomal stress responses, inhibits rRNA synthesis, reduces RNA polymerase I binding to the promoter of rRNA gene, disrupts nucleolar integrity, and impairs ribosome biogenesis and polysome formation. Acr causes an increase in MDM2 levels and phosphorylation of MDM2 in A549 and HeLa cells which are p53 active and p53 inactive, respectively. It enhances the binding of ribosomal protein RPL11 to MDM2 and reduces the binding of p53 and E2F-1 to MDM2 resulting in stabilization/activation of p53 in A549 cells and degradation of E2F-1 in A549 and HeLa cells. We propose that Acr induces ribosomal stress which leads to activation of MDM2 and RPL11-MDM2 binding, consequently, activates p53 and enhances E2F-1 degradation, and that taken together these two processes induce apoptosis and cell death.
Acrolein (Acr) is a ubiquitous environmental contaminant that predominantly arises from incomplete combustion such as cooking and tobacco smoking . Acr contains a carbonyl group and an α,β-unsaturated double bond which owing to their reactivity with different cellular components such as nucleic acids and proteins can induce mutagenic DNA adducts and induce protein dysfunction . Acr has been proposed to be carcinogenic via DNA adduct induction and impairment of DNA repair function [2–5]. In addition, Acr also has a potent cytotoxic effect; it can induce cell death via both apoptosis and necrosis pathways [1, 6, 7]. Acr has been shown to be a major cause of tobacco smoke related chronic obstructive pulmonary diseases (COPD) and asthma ; it has been proposed that the apoptotic and necrotic effects of Acr elicits these diseases [6, 7].
Acr is a major metabolite of the antitumor drugs cyclophosphamide and ifosfamide. Acr cytotoxicity is believed to be the major antitumor activity of these drugs [9, 10]. Hence, understanding the Acr-induced effects - DNA adduct formation, protein modifications, and cell death - may enhance not only our understanding of how Acr induces different diseases but also help to elucidate the anti-tumor activity of these drugs. While it is well understood of how Acr adducts DNA and proteins, the cellular processes by which Acr elicits cell death are not clear.
Mapping Acr-induced DNA adduct formation at the DNA sequence level we have found that Acr-DNA adducts are preferentially formed at GC rich sequences [2–5]. This finding raises the possibility that the nucleolus is also a preferential target of Acr since ribosomal DNA (rDNA) in nucleolus is GC rich . If this is the case, then it is possible that Acr-rDNA binding elicit cell death signals since it is well established that rDNA damage is the major cellular stress response hub [12, 13]. In this study we tested this possibility and delineate the Acr-induced stress pathway. Using an immunofluorescent staining method, we found that Acr-DNA adducts are indeed preferentially formed in the nucleolus. Acr induces oxidative damage in both rDNA and rRNA. Acr interrupts rRNA transcription and processing, as well as polysome formation and global protein translation.
It is well understood that the nucleolus is the site of ribosome biogenesis which is an essential and energy consuming cellular process [12, 13], and that impairment of ribosome biogenesis causes ribosomal stress (also known as nucleolar stress) [12, 14, 15]. The correlation of the DNA damage response with the nucleolus has shown that the nucleolus acts as a sensor for cellular stress signals through stabilization of p53 by ribosomal protein (RP)–MDM2/HDM2 interactions, which induces cell cycle arrest or apoptosis [14, 16–20]. Intriguingly, we found that Acr induces the same extent of apoptosis and cell death in human lung adenocarcinoma A549 cells and human cervical cancer HeLa cells with active p53 and inactive p53, respectively. These results raise the question of “what are the apoptosis signals induced by Acr in these cells?”.
We found that Acr induces ribosomal stress resulting in disintegration of ribosome, and enhancing RP11-MDM2 interactions. Consequently, Acr reduces binding of activated p53 proteins in A549 cells, and reduces binding of E2F-1 with MDM2 causing E2F-1 degradation in A549 and HeLa cells. We propose that Acr cytotoxicity occurs via ribosomal stress which activates p53 and enhances E2F-1 degradation, both of which can cause cell death.
Although Acr induces cell death via both apoptosis and necrosis process, the signals that induce these processes are not well understood. It has been established that Acr mediates antitumor activity of cyclophosphamide and ifosfamide through its cytotoxicity [9, 10]. Since p53 plays a center role in apoptosis and 50% of human cancer cells either have inactive p53 function or carry mutant p53, it is important to determine the toxicity of Acr in cells with different activities of p53 and the cell death mechanisms. We chose to determine the Acr induced cytotoxicity and cell death processes in human lung adenocarcinoma cells A549 which carry wild type p53 gene and human cervical cancer HeLa cells in which p53 is inactive [21, 22]. The result in Figure Figure1A1A shows that both cells have the same sensitivity toward Acr-induced cytotoxicity. We also found that Acr induces mainly apoptosis in A549 and HeLa cells (Figure (Figure1B1B and and1C).1C). These results indicate that Acr induces both p53 dependent and independent apoptosis.
Previously, we found that Acr can damage genomic DNA to induce mutagenic Acr-dG adducts in different human cells, and that Acr-dG adducts preferentially occurred in run's of G sequences [4, 5, 23]. Since rDNA is rich in GC content it is possible that rDNA in the nucleolus is a preferential target for Acr [15, 24–26]. Using an immunofluorescence assay with a specific anti-Acr-dG antibody, we found that Acr-dG adducts were preferentially observed in the nucleoli in both A549 and HeLa cells (Figure (Figure2A2A and Figure S1A). It has been found in cell culture that, Acr can trigger lipid peroxidation and production of intracellular reactive oxygen species, as shown in Figure S2, which can induce oxidative DNA damage . Results in Figure Figure2B2B show that Acr indeed induces DNA damage that was recognized by anti-8-oxo-dG antibodies (Figure S1B). The formation of Acr-dG and 8-oxo-dG in nucleoli was further confirmed by the results in Figure Figure2A2A and and2B2B and Figure S1A and S1B showing that pre-incubating anti-Acr-dG or anti-8-oxo-dG antibody with a 15 to 20-fold excess of soluble Acr-dG and 8-oxo-dG abolished their ability to detect the Acr-dG or 8-oxo-dG formation in the nucleoli.
Since the nucleolus is a large aggregate consisting of rDNA, precursor and mature rRNAs, it is possible that in addition to rDNA, the rRNA is also a target of Acr . To test this possibility, using DNase and RNase treatment followed by the immunofluorescence assay with anti-Acr-dG or anti-8-oxo-dG, we found that the 8-oxo-dG antibody recognizing oxidative damage was sensitive to both DNAse and RNAse digestion, the same as damage induced by H2O2. These results indicate that Acr-induced oxidative damage can occur in both rDNA and rRNA (Figure (Figure2C2C and Figure S1C).
We also determined the effect of Acr treatment on nucleolus morphology and molecular re-arrangements and compared these changes to those induced by actinomycin D (Act D), a well-characterized nucleolar disruptor . Results in Figure Figure33 show that Act D treatment (20 ng/ml, 3 h) induced RNA polymerase I (Pol I) or upstream binding factor (UBF) segregated into caps around the DAPI-sparse nucleoli such that nucleophosmin (B23) and nucleolin (NCL) were mislocalized over the nucleoplasm (Figure (Figure3A3A and and3B3B and Figure S3A and S3B). Acr (75 μM 3 h) induced movement of RNA Pol I or UBF, but not onto the nucleolar cap (Figure (Figure3D3D and Figure S3D). Rather, Acr induced translocation of B23 into the nucleoplasm was only observed in cells treated with Acr for a short period (3 h) (Figure (Figure3C3C and Figure S3C). These results raise the possibility that Acr may have effects on different levels of rRNA synthesis, including inhibition of rRNA processing.
Previous studies have shown that inhibition of rRNA synthesis is related to disintegration of nucleolar structures . Using immunofluorescence detection we found that Acr treatment diminished nucleolar 5-fluorouridine incorporation into nascent rRNA indicating that Acr treatment inhibits rRNA synthesis (Figure (Figure4A4A and and4B4B and Figure S4A). Chromatin immunoprecipitation assay results in Figure Figure4C4C show that Acr treatment decreased binding of RNA Pol I and UBF on the promoter region of rDNA in cells treated with Acr. These results are consistent with the decreased co-localization of RNA Pol I and UBF showing in Figure Figure3D3D and Figure S3D. The quantitative real-time RT-PCR analysis results show that Acr decreased the expression of 45S pre-rRNA, but had no effect on 18S rRNA expression (Figure (Figure4D4D and Figure S4C). This result indicated that while Acr inhibited the synthesis of rRNA, it impaired the processing of rRNA only modestly.
In order to further understand whether Acr influenced ribosome biogenesis, we used the ribosomal profiling assay to assess whether the large and small ribosomal subunits could be assembled normally to perform their respective translation functions. These results in Figure Figure5A5A and and5B5B and Figure S4B show that indeed Acr dramatically diminished the formation of polysomes (Figure (Figure5A5A and and5B5B and Figure S4B). Consistent with this reduction of polysome formation, Acr treatment also caused a dose-dependent decrease of global protein synthesis in both A549 and HeLa cells. (Figure (Figure5C5C and Figure S4D). However, no difference in the amount of 28S and 18S rRNAs was found among cells treated with or without Acr (Figure (Figure5D).5D). Together, these results suggest that Acr inhibited global mRNA translation by interfering with rRNA synthesis (reduction of Pol I and UBF loading at 45S promoters) and consequentially polysome assembly in both A549 and HeLa cells.
It is well established that nucleolar transcription is inhibited under DNA damage induced stress [12, 14, 15], during which several proteins regulate rRNA transcription or processing. The nucleolus acts as a sensor for cellular stress signals through stabilization of p53 by RP–Mdm2/HDM2 and ARF–Mdm2/HDM2 interactions, which induce cell cycle arrest or apoptosis [16–19]. We found that Acr treatment caused an increase of both phosphorylated and total p53 protein levels in p53-active A549 cell in dose and time-dependent manner (Figure (Figure6).6). The total protein and the phosphorylated MDM2 levels were also increased up to 8 h incubation. After 24 h incubation the levels of MDM2 and phosphorylated MDM2 decreased while the levels of p53 and phosphorylated p53 continuously increased. These results indicate that the decrease of MDM2 is due to a p53-MDM2 feedback loop and that p53 is a sensor for Acr-induced ribosomal stress via MDM2 activation.
Since Acr also induces ribosomal stress in HeLa cells which have p53 nullified by viral E6 [21, 22], we then determined the signal pathway of this ribosomal stress in p53-inactive cells. Results in Figure Figure66 show that while Acr treatment modestly increase total p53 levels after no phosphorylated p53 was detected. Acr also failed to stimulate the expression of its downstream targets, p21 in these p53 inactive cells (data not shown). These results indicate that Acr-induced ribosomal stress does not activate p53. However, Acr treatment enhances both total and phosphorylated MDM2 indicating that MDM2 play a p53 independent role in Acr-induced ribosomal stress response.
Since E2F-1 is also known to be involved in regulating rRNA transcription and coordinating DNA damage and nucleolar stress , we next measured the expression levels of E2F-1 in Acr-treated A549 and HeLa cells. As can be seen (Figure (Figure7A7A and and7B),7B), E2F-1 was reduced in a time-dependent fashion in A549 and HeLa cells. However, no change in the mRNA levels of E2F-1 occurred at that time (Bar graphs of Figure Figure7A7A and and7B),7B), indicating a post-transcriptional mechanism for the downregulation of the E2F-1 protein. The reduction of E2F-1 protein in Acr-treated cells was partially restored by MG-132, a well-known proteasome inhibitor (Figure (Figure7C7C and and7D).7D). These results suggest that Acr directly and/or indirectly via ribosomal stress response caused E2F-1 protein degradation via a proteasome pathway.
The mechanism of p53 stabilization (Figure (Figure6)6) or E2F-1 degradation (Figure (Figure7)7) after perturbation of ribosome biogenesis is possibly the consequence of changes in functional and physical interactions of these proteins with MDM2. Previous studies have shown that MDM2 negatively controls p53 activity in two ways: by binding to the protein and interfering with its transactivation activity, and by facilitating p53 proteasomal degradation thereby acting as an E3 ubiquitin ligase . On the other hand, it has been shown that MDM2 binds to the E2F-1 protein and protects it from proteasome-mediated degradation . As a consequence of reduced ribosome biogenesis, several RPs such us L5, L11, L23 and S7 are no longer used for ribosome generation but instead binding to MDM2 to relieve its inhibition on p53, as well as its protection toward E2F-1 [32–36]. Interestingly, our results showed that in Acr-treated A549 cells, the amounts of p53 and E2F-1 complexed with MDM2 were markedly reduced in comparison with controls cells (Figure (Figure8A).8A). By contrast, the amount of RPL11 associated with MDM2 was significantly increased in Acr-treated cells (Figure (Figure8A).8A). We interpret these results as indicating that Acr-induced rRNA synthesis inhibition can cause a disintegration of the polysome and the integrity of ribosomal structure. Consequently, ribosomal proteins including RPL11 will be released from the ribosomal structure. RPL11 is able to bind to MDM2. The RPL11 bound MDM2 loses its function in mediating p53 degradation and protecting E2F-1 against proteasomal degradation.
Considering that both p53 and E2F-1 are crucial regulators of cell apoptosis, we next examined the subsequent signaling responsible for Acr-induced apoptosis in both A549 and HeLa cells observed in Figure Figure1B1B and and1C.1C. As can be seen, a time and dose-dependent cleavage of caspase 3, caspase 9, and PARP were observed in these cells (Figure (Figure8B8B and and8C),8C), suggesting apoptosis induced by Acr. Interestingly, PUMA, a pro-apoptotic gene induced by p53  was increased in dose and time-dependent manner in p53-active A549 cells, but not in p53-inactive HeLa cells. However, Bcl-2, an anti-apoptotic gene regulated by E2F1  was decreased in dose and time-dependent manner in both cells. This indicates that Bcl-2 plays a major role in Acr-induced apoptosis in p53-inactive HeLa cells.
Since A549 and HeLa cells are derived from lung and cervical cancer, respectively, it is possible that apoptosis pathway induced by Acr is due to cell differences, not p53 activity. In order to further confirm the role of p53 in Acr-induced apoptosis pathways, we used siRNA to knockdown p53 in A549 cells. Results in Figure Figure9A9A and and9B9B show that Acr induced phosphorylation of MDM2 and reduction of E2F1 in p53-knockdown cells, which is similar to HeLa cells (Figure (Figure6).6). Results in Figure Figure9C9C and and9D9D show that Acr can also induce apoptosis pathway in p53 knockdown cells, but the extent of apoptosis is much lower than in A549 shown in Figure Figure9E9E and and9F.9F. Furthermore, Acr causes a decrease of Bcl-2 in these cells (Figure (Figure9C9C and and9D).9D). Taken together, these results indicate that via Bcl-2 regulation p53 plays a crucial role in Acr-induced apoptosis in p53-active cells.
Although Acr is a ubiquitous environmental contaminant, it also is an active cytotoxic metabolite of the anti-cancer drugs cyclophosphamide and ifosfamide. It is generally accepted that the antitumor activity of these drugs resides in the cytotoxicity of their major metabolite Acr [9, 10]. However, the mechanisms via which Acr exerts its anti-cancer activity and cytotoxicity are not clear. Our previous studies have shown that Acr-dG adducts were preferentially formed in DNA runs of G sequences [4, 5, 23]. The rDNA is organized in the form of tandem repeats with high GC content in the nucleolus and we demonstrate here that indeed the nucleolus is a preferential target of Acr. It is well established that the nucleolus is the major hub for sensing DNA damage-induced stress [12, 13]. In this study we found that Acr-induced ribosomal stress cascades via disintegration of the polysome and ribosomal structure, freeing the ribosomal proteins (Figures (Figures33 and and8).8). Consequently, the freed ribosomal protein RPL11 binds to MDM2 to cause MDM2 dysfunction. In p53-active A549 cells, the activated p53 proteins are stabilized which cause E2F-1 degradation. In p53-inactive HeLa cells, MDM2 proteins are unable to prevent E2F-1 from proteosomal degradation due to RPL11 binding. Reduction of E2F-1 prevents activation of cell cycling and sensitizes cells to DNA damage-induced cell death (Figure (Figure7).7). We believe these two processes are the major causes of Acr cytotoxicity in human cancer cells with different activities of p53 (Figure (Figure99).
We found that Acr induces not only Acr-dG adducts but also 8-oxo-dG adducts in rDNA. These results are consistent with previous studies showing that Acr can trigger lipid peroxidation to induce reactive oxygen species which can cause additional oxidative DNA damage . This conclusion was further supported by the result that DNase treatment eliminates all sites recognized by the 8-oxo-dG antibody in the nucleolus. It has been found that oxidative damage can also occur in RNA [39–41]. Results from RNase treatment in Acr-treated cells show that while RNase treatment does not affect the integrity of nuclear DNA it eliminates all 8-oxo-dG antibody recognition sites in the nucleolus. We believe that this is due to the conformation of 8-oxo-dG containing DNA in the “A” form similar to RNA conformation , therefore being sensitive to RNase digestion. Our results are consistent with previous findings that RNA is vulnerable to oxidative damage [39–41]. Oxidative modification of RNA results in disturbance of the translational process and impairment of protein synthesis, which can cause cell deterioration or even cell death [43, 44]. Maintaining the integrity of rDNA is crucial for rRNA synthesis which is the initial step of ribosome biogenesis.
Previous studies have shown that inhibition of rRNA transcription and early rRNA processing steps, but not of late rRNA processing steps, coincides with the loss of nucleolar integrity . Translocation of B23 was only observed at the short exposure time (3 h) for Acr treatment (Figure (Figure3C3C and Figure S3C), but not for longer time treatments with Acr (3–24 h), indicating that Acr may affect different levels of rRNA generation. This is consistent with our observations that rDNA/rRNA damages induced by Acr lead to the abolition of rRNA transcription, ribosomal assembly and eventually global translation (Figures 4–5 and Figure S4). However, localization of B23 in the Act D-exposed nucleoli was diminished and nucleoli were further increased in compactness and segregation to form nucleolar caps due to inhibition of rRNA synthesis (Figure (Figure3A3A and and3B3B).
A widely accepted mechanism of p53 checkpoint activation after alteration of ribosome biogenesis is that it causes leakage of ribosomal proteins including L5, L11, L23 and S7, and that these RPs might bind to MDM2 causing MDM2 dysfunction in mediating p53 degradation [32–36]. In this study, we found that in p53-active A549 cells, inhibition of rRNA transcription by Acr induced p53 stabilization, increased binding of RPL11 with MDM2 and induced E2F-1 degradation (Figure (Figure8A);8A); these results are consistent with the current understanding that in p53 proficient cells altered ribosome biogenesis inhibits cell proliferation through the activation of the RP-MDM2-p53-E2F-1 pathway. We found Acr treatment causes reduction of E2F1 in p53-active A549 cells and p53-inactive HeLa cells. In p53-inactive cells, MDM2 does not bind with E2F-1 but binds with RP11 after Acr treatment. The degradation of the E2F-1 protein has previously been shown to be hindered by its interaction with MDM2, which acts by inhibiting E2F-1 ubiquitylation and subsequent proteasomal degradation . Our results show that the downregulation of E2F-1 protein in Acr-treated cells could be prevented by blocking 26S proteasome, indicating that MDM2 binding protects E2F-1 from degradation (Figure (Figure7).7). We found that Acr treatment does not affect mRNA levels of E2F-1 supporting the notion that Acr-induced reduction of E2F-1 levels is mediated by translational or a posttranslational control (Figure (Figure77).
The inhibition of rRNA synthesis is not always associated with downregulation of E2F-1. For example, Act D, cisplatin and etoposide – all drugs that inhibit rRNA transcription – cause accumulation of E2F-1 protein by phosphorylation in response to DNA damaging agents [45, 46]. It is worth noting that these drugs are DNA-damaging agents which could activate the checkpoint kinase 2 to phosphorylate E2F-1, resulting in the increases of both its half-life and transcriptional activity . In contrast, it has been demonstrated that silencing TIF1A without using DNA damaging agents, not only inhibits rRNA transcription, but also causes a downregulation of E2F-1 protein . We show that Acr-induced ribosomal stress causes a reduction of E2F-1 and apoptosis in cells with active or inactive p53 (Figure (Figure11 and Figure Figure7).7). While it is well established that reduction of E2F-1 can prevent cell proliferation , these results support the finding that Acr-induced reduction of E2F-1 inhibits Bcl-2 expression resulting in apoptosis in p53-inactive HeLa cells or in p53-knockdown A549 cells (Figure (Figure8B8B and and8C8C and Figure Figure9C9C and and9D).9D). However, the extent of apoptosis is much lower in p53-knockdown cells than in control cells. This result indicates p53 plays a crucial role in Acr-induced apoptosis in p53-active cells.
Acr has been shown to trigger either apoptotic or necrotic pathways depending on the cell types, culture conditions, or even medium composition used . For example, Acr has been demonstrated to induce apoptosis in Chinese hamster ovary (CHO) cells, either through the intrinsic pathway which involves cytochrome C release  or through the extrinsic pathway by activating death receptors . On the other hand, Acr has also been reported to cause necrotic death in murine FL5.12 proB lymphocytes cells cultured in serum depleted medium .
In summary, we demonstrate here that Acr-induced rDNA/rRNA damages hindered rRNA transcription and processing, which results in MDM2-RPL11 binding, E2F-1 degradation and cellular apoptosis in p53-active A549 cells and p53-inactive HeLa cells. We propose that in p53-active cells activated p53 triggers apoptosis while in p53-inactive cells E2F1 degradation results in suppression of Bcl-2 consequently enhancing Acr-DNA damage induced apoptosis (Figure (Figure10).10). Our results not only enhance our understanding of the molecular mechanisms of Acr mediated antitumor activity but also may enable the development of better therapeutic strategies for killing cancer cells regardless of their p53 status.
Human cervical carcinoma cells (HeLa) and human lung adenocarcinoma cells (A549) (American Type Culture Collection, Manassas, VA) were grown in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% FBS and in RPMI 1640 Medium supplemented with 10% FBS, respectively. Acrolein (Acr) stock solution (Sigma-Aldrich) was prepared freshly before use. Cells at 70% confluency were washed with PBS buffer (137 mM NaCl, 2.7 mM KCl, 8 mM Na2HPO4, 1.46 mM KH2PO4, pH 7.0) and treated with different concentrations of Acr (0–100 μM) in complete culture medium for different times as indicated at 37°C in the dark.
RNA interference in A549 was carried out according to the manufacturer's protocol using GenMute siRNA Transfection Reagent (SignaGen Laboratories). The sequence of the p53 siRNA targets p53 mRNA (NM_000546.5) is GACUCCAGUGGUAAUCUAC and siRNA was synthesized by Sigma (St. Louis, MO, USA). The p53 siRNA and control siRNA were transfected at a final concentration of 30 nM for 24 h following by Acr treatment as described above.
The Acr cytotoxicity was determined using a modified 3-(4,5- dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium (MTT; Sigma, St. Louis, MO) assay  as described previously. Briefly, for MTT assay, cells (5 × 103/ well) were seeded in 96-well plates overnight, and then treated with Acr (0–250 μM, 24 h). The resulting formazan dissolved with DMSO was measured at 570 nm and results were presented as the percentage of the control values. For Lactate dehydrogenase (LDH) leakage assay, cells (2.5 × 104/well) were seeded in 24-well plates overnight and then treated with Acr (0–250 μM, 24 h). LDH activities were measured using the LDH detection kit (Sigma) as described in the manufacturer's protocol. These experiments were performed in triplicates and were repeated at least three times. The extent of cellular damage was calculated based on the percent LDH activity in the supernatants relative to that in the cell lysates.
The generation of pro-oxidants was measured as described previously , with modifications. Briefly, A549 or HeLa cells were treated with Acr (0–250 μM) or hydrogen peroxide (H2O2, 1 mM) as a positive control for one hour. Then, 2′7′- dihydrodichlorofluorescein diacetate (H2DCFDA, 10 μM, Sigma) was added and incubated for 30 min. Levels of pro-oxidants in 10,000 cells were determined by flow cytometry using the FL-1 detector as described previously .
For detection of Acr-dG or 8-oxo-dG adducts, immunofluorescent staining were performed as described previously . The following antibodies were used at the noted dilutions: Acr-dG (1:100), 8-oxo-dG (1:100, Abcam. Ab62623), nucleolin (1:500, Cell signaling, #14574), B23 (1:250, Zymed), UBF (1:100, Santa Cruz, sc-9131) or RNA Pol. I (RPA194, 1:100, Santa Cruz, sc-28714). The appropriate fluorophore-conjugated secondary antibodies (1:200, FITC or Rhodamine; Molecular Probes) were used and immunofluorescent images of the fixed cultures were viewed with a fluorescence laser-scanning confocal microscope (Olympus FV10i, Center Valley, PA). The immunofluorescent image acquisition times for the DAPI, Rhodamine and FITC channels, respectively, were kept constant over all samples. Staining specificity was determined by pre-incubating anti-Acr-dG or anti-8-oxo-dG antibody with a 15 to 20-fold excess of soluble Acr-dG and 8-oxodeoxyguanosine (Berry & Associates, Inc., city and state). To investigate if Acr-dG or 8-oxo-dG adducts were preferentially formed in DNA or RNA, Acr-treated cells on coverslips were pretreated with DNase I (0.1 mg/ ml in PBS for 10 min at RT; Sigma) or RNase (1 mg/ ml in PBS for 10 min at 37°C; Invitrogen) before fixation. After enzyme treatment, the remaining DNA or RNA was evaluated by immunofluorescence assay for Acr-dG and 8-oxo-dG adducts as described above.
The polysome assay was carried out using cell lysate from HeLa or A549 cells (1 × 108) treated with Acr (0–100 μM, 3 h) as described previously . The lysate was prepared in buffer A containing 25 mM HEPES (pH 7.5), 400 mM KOAc, 5 mM Mg (OAc)2, 2% TritonX-100, 0.2 mM cycloheximide, and 40 U/ml RNasin (Invitrogen). A 12-ml 10–50% sucrose density gradient containing buffer (20 mM Tris–HCl, pH7.5; 50 mM KCl; 3 mM MgCl2) was used for the analysis. The polysome profile was monitored by absorbance at 254 nm using an ISCO auto fractionation instrument and starting with the free material followed by 40S ribosomal subunit detection and continuing through to polyribosome complexes. Aliquots were taken from each fraction and subjected to the phenol–chloroform extraction to allow rRNA analysis.
HeLa or A549 cells were treated with Acr (0–100 μM, 3 h), cell lysates were prepared and analyzed as described previously . Briefly, blots were probed with a monoclonal antibody against MDM2 (1:1000, Abcam, ab178938), p-MDM2 (Ser166, 1:1000, Cell signaling, #3521), E2F-1 (1:1000, Cell signaling, #3742), p-p53 (Ser15, 1:1000, Cell signaling, #9284), p53 (1:1000, Calbiochem) and RPL11 (3A4A7, Thermo Scientific) at 4°C for overnight following by horseradish peroxidase-conjugated secondary IgG (1:3,000; Millipore) for 1 h at room temperature. The immunoreaction was visualized using Enhanced Chemiluminescence (ECL) (Millipore Corporation, Billerica, MA). The bound primary and secondary antibodies were stripped by incubating the membrane in stripping buffer (100 mM 2-mercaptoethanol, 2% SDS) for 30 min at room temperature. The membrane was then re-probed with β-actin antibody (1:5,000; Millipore [clone C4]).
Immunodetection of nascent rRNA was performed by incorporation of 5-fluorouridine (5-FU), according to the method described . Briefly, cells growing on coverslips were incubated with 2 mM 5-FU (Sigma) for 15 min, then washed with cold PBS and fixed in 4% paraformaldehyde and 1% Triton X-100 in PBS for 10 min. Subsequently, the cells were immunofluorescently stained with a specific monoclonal antibody for halogenated uridine (1:400, Sigma [BU-33]). Mounting and nuclei counterstaining and immunofluorescent image were performed as describe as above. Quantification of incorporation of 5-FU into rRNA was using Olympus cellSens™ Dimension software (Olympus Life Science).
For ChIP assays, cells (5 × 106) were grown in 15-cm dishes overnight. After treatment of Acr, DNA was cross-linked with 1% formaldehyde at room temperature for 10 min and assays performed using 106 cells per immunoprecipitation as described in Chromatin Immunoprecipitation (ChIP) Assay Kit (Millipore). Cells were lysed in the presence of protease inhibitors and chromatin was fragmented to 200–1000 bp by sonication (high power, 20 cycles of 30 seconds with 30 seconds between pulses). Immunoprecipitations were performed with 2 mg of UBF (1:100, Santa Cruz, sc-9131) or RNA Pol I (RPA194, 1:100, Santa Cruz, sc-28714), or 2 mg of normal rabbit IgG (Santa Cruz, sc-2027). Complexes were collected with protein G agarose (GE, 17-0618-01). De-crosslinked DNA was purified and eluted in 50 ml of elution buffer of which 2 ml was used for PCR. Primers (5′-3′) were CGATGGTGGCGTTTTTGG and CCGACTCGGAGCGAAAGATA for the rRNA promoter region; and CGACGACCCATTCGAACGTCT and CTCTCCGGAATCGAACCCTGA for the rRNA transcribed region. Samples were analyzed in triplicate using the SYBR green dye on the StepOnePlus™ Real-Time PCR System (Applied Biosystems). To calculate the percentage of total DNA bound, unprecipitated input samples from each condition were used as reference for all qPCR reactions.
For IP studies, cells were washed and scraped in PBS, then suspended in IP lysis buffer (20 mM Tris-HCl (pH7.4), 170 mM NaCl, 13 mM MgCl2, 0.5% NP40) and dounced 30 times in a Dounce homogenizer. After centrifugation at 16,100 × g, 4°C for 15 min, the protein concentrations in the supernatant were measured using BCA protein assay kit (Pierce, Rockford, IL). IP procedure was followed according to manufacturer's instructions (Dynabeads® protein G, Invitrogen). Briefly, 2 μg of MDM2 (ab16895, Abcam) or IgG (Santa Cruz Biotechnology Inc) antibody in 300 ml of IP lysis buffer was incubated with 100 μl of dynabeads protein G for 2 h on rotating platform at 4°C. After removing antibody solution using the Dynal magnet system, 1 mg of protein samples in 300 ml of IP lysis buffer were added and incubated overnight on rotating platform at 4°C. Beads were then washed (using the Dynal magnet system) three times with 0.5 ml of ice cold IP buffer. After the last wash, beads were centrifuged and last traces of buffer were removed using a micropipette. Antibodies/protein complexes were eluted with 1× SDS sample buffer followed by western blotting.
Total RNA of A549 or HeLa cells treated with Acr (0–100 μM, 3 h) was extracted using Trizol (Invitrogen) according to manufacturers' instructions. The 28S and 18S RNA subunits were visualized by loading in a 1% agarose gel stained with ethidium bromide an equal fraction (10%) of the total quantity of the obtained RNA. The intensity of the bands was evaluated with the densitometric software UVP™ Doc-It™ LS Image Analysis Software.
Total RNA was isolated from harvested cells using TRIzol® Reagent (Thermo Fisher Scientific). Reverse transcription was RevertAid Reverse Transcriptase (Thermo Fisher Scientific) according to manufacturers' instructions. Subsequent real-time RT-PCR analysis of cDNA were performed in triplicate using the SYBR green dye on the StepOnePlus™ Real-Time PCR System (Applied Biosystems). The primers (5′-3′) were CTCCGTTATGGTAGCGCTGC and GCGGAACC CTCGCTTCTC for 45S; CGACGACCCATTCGAAC GTCT and CTCTCCGGAATCGAACCCTGA for 18S; GCCACTGACTCTGCCACCATAG and CTGCCCATC CGGGACAAC for E2F1; CCGTCTAGAAAAACCTGCC and GCCAAATTCGTTGTCATACC for GAPDH. To calculate the relative RNA expression, GAPDH was used as an internal control for all qRT-PCR reactions and compared with control groups.
Cells (5 × 105/ 6-well plate) treated with Acr (75 μM, 0–24 h) as previously described . After harvested cells were washed twice in ice-cold PBS and fixed in ice-cold 70% ethanol for 30 min or overnight at 4°C. Cells were washed in PBS and digested with DNase-free RNase A (50 U/ ml) at 37°C for 30 min. Before flow cytometry analysis, cells were resuspended in 500 μl propidium iodide (PI, 10 μg/ ml; Sigma) for DNA staining. PI staining was used to measure for cell cycle status using a Becton-Dickinson FACScan instrument and Cell Quest software.
Cells (1 × 106/ 10-cm dish) treated with Acr (0–100 μM, 24 h) were analyzed by FITC Annexin V Apoptosis Detection Kit I (BD Biosciences Canada, Mississauga, ON, Canada) according to manufacturers' instructions. Briefly, Acr-treated cells were washed with PBS and resuspended in 1 ml of binding buffer (10 mM HEPES/ NaOH, pH 7.5, 140 mM NaCl, and 2.5 mM CaCl2). A volume of 500 μl of cell suspension was incubated with 5 μl of Annexin V-FITC and 10 μl of propidium iodide (PI) for 10 min at room temperature in the dark. Cells (10,000) were then analyzed by flow cytometry. Annexin V-FITC fluorescence was detected on the FL-1 detector and PI fluorescence on the FL-2 detector. Four populations of cells were analyzed: live control cells (Annexin V−/PI−); early stage apoptotic cells (Annexin V+/PI−); late stage apoptotic cells (Annexin V+/PI+); necrotic cells (Annexin V−/PI+). The results are reported as the fraction of total apoptotic cells (early and late stage apoptosis) and necrotic cells.
The global protein synthesis was analyzed using Cayman's protein synthesis assay Kit (Cayman chemical, # 601100), following the manufacturer's protocol. Briefly, the protein synthesis rate was evaluated by incorporation of puromycin analog o-propargyl-puromycin (OPP) added to the Acr-treated HeLa or A549 cells resuspended in complete medium. Upon application to cells, the OPP probe incorporates into the C-terminus of translating polypeptide chains, thereby stopping translation. These truncated C-terminal alkyne-labeled proteins are then subsequently detected via copper-catalyzed click chemistry using 5 FAM-Azide following by flow cytometry. Histograms show the values (mean ± s.d.) of three independent experiments.
Student's t-tests were used to determine statistical significance, and two-tailed P-values are shown. A minimum of three independent replicate experiments was performed to justify the use of statistical tests.
We thank Dr. Anya Maan-Yuh Lin, National Yang-Ming University, and Dr. Zee-Fen Chang, National Taiwan University for their material support and helpful advice. We appreciate Dr. Shieh, Sheau-Yann, Academia Sinica for her kind gift of p53 siRNA. We also thank Dr. Catherine Klein, New York University Langone School of Medicine, and Dr. Yeu Su, National Yang-Ming University, for critically reviewing this manuscript.
CONFLICTS OF INTEREST
No potential conflicts of interest were disclosed by the authors.
This work was supported by Ministry of Science and Technology, Taiwan. [Grant# 103-2320-B-010-042, Grant # 104-2320-B-010-040-MY3 (H-t Wang) and NIH CA165980 and CA190678 (M-s Tang).