| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 38, 39686-39696, September 17, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
**
From the
Laboratory of DNA Repair, Health and Environment Unit, and ¶Glycobiology Laboratory, Research Center for Infectious Diseases, Laval University Medical Center, CHUQ, Faculty of Medicine, Laval University, Ste-Foy, Quebec G1V 4G2, Canada
Received for publication, March 10, 2004 , and in revised form, June 18, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Within the automodification domain, there are 15 glutamic acid residues that are potential ADP-ribose polymer attachment sites (11). The length of ADP-ribose polymers, which are highly negatively charged homopolymers, can reach over 200 ADP-ribose residues (12). Thus, when PARP-1 is automodified, the affinity of the zinc fingers of PARP-1 for DNA breaks is reduced by the negative charges of the polymers attached to the automodification domain, leading to the dissociation of PARP-1 from DNA breaks (13). During the 1980s, PARP-1 was considered a DNA repair enzyme (2). However, it is unlikely that PARP-1 itself has enzymatic activity to repair DNA breaks, as breaks can be efficiently repaired in the absence of PARP-1 (3, 14–17). Alternatively, recent models suggest that the ADP-ribose polymers have a function in recruiting DNA repair enzymes to DNA damages (18, 19). We have also recently proposed an alternative model in which PARP-1 automodification plays a role in transcription regulation at the level of elongation, because the zinc finger motifs of PARP-1 can bind to RNA with high affinity, leading to negative regulation of RNA synthesis by RNA polymerase II, and, following PARP-1 automodification, the binding affinity of the zinc fingers for RNA is reduced, resulting in up-regulation of transcription (20).
In addition to these studies, which focused on poly(ADP-ribosyl)ation, accumulated evidence in recent years suggest that PARP-1 has roles in these various fundamental cellular processes through interaction with other factors and enzymes. For example, PARP-1 has been reported to form complexes with transcription initiation factors, including Oct-1 (21), TEF-1 (22), B-MYB (23), YY1 (24), NF-
B (25, 26), Tax (27), retinoid X receptor (28), p53 (29, 30), and AP-2 (31, 32). In addition, co-expression of PARP-1 with B-MYB or NF-B leads to increased transcription levels of specific genes (23, 25). Thus, PARP-1 has been proposed to be a factor involved in regulating transcription initiation (33). The role of PARP-1 in DNA replication has also been proposed mainly based on the observation that PARP-1 can be co-purified with DNA polymerase 
(34, 35). Other lines of evidence suggest that PARP-1 plays a role in DNA repair through interaction with XRCC-1 (36, 37), DNA ligase III (19), and Werner syndrome protein (38). Finally, PARP-1 has been reported to bind to topoisomerase I (Topo I) (39, 40), which is an enzyme required for DNA replication and transcription (41, 42). Taken together, these observations suggest that PARP-1 is an enzyme playing roles in transcription initiation, DNA replication, and DNA repair through interaction with several different factors and enzymes. Interestingly, in some cases, the interacting domain of PARP-1 with these factors and enzymes has been identified; the interactions between PARP-1 and XRCC-1 (37), NF-B (25), YY1 (24), and Topo I (40) occur through the automodification domain of PARP-1. Thus, ADP-ribose polymer formation on the automodification domain is likely to have an effect on the interactions between PARP-1 and these factors and enzymes. In fact, Chang and Alvarez-Gonzalez (26) suggested, using a gel retardation assay, that the interaction between NF-B and PARP-1 can be disrupted by automodification of PARP-1 in response to DNA damage. Furthermore, von Kobbe et al. (38) recently reported that automodified PARP-1 cannot bind to the Werner syndrome protein in vitro. As disruption of the interaction between PARP-1 and these various factors and enzymes likely has an effect on transcription, DNA replication, and DNA repair, automodification may play critical roles in the regulation of these processes in response to DNA damage. However, little is known about the effect of PARP-1 automodification on the interactions between PARP-1 and these various factors and enzymes.
To obtain a better understanding of the functional roles of the interactions between PARP-1 and these other factors and enzymes, we have created fluorescent protein-tagged PARP-1 to study PARP-1 in live cells. As a factor that interacts with PARP-1, we focused on Topo I, because: 1) Topo I is an abundant nuclear enzyme that is present at similar amounts as PARP-1 in vivo (43, 44); 2) Topo I has been reported to interact with the automodification domain of PARP-1 in vitro (40); 3) Topo I is involved in both DNA replication and transcription, both at the level of initiation and elongation (42, 45); and 4) Topo I activity is known to be promoted by PARP-1 in vitro (46), which allows us to investigate the effect of disrupting the PARP-1-Topo I interaction on the activity of Topo I. Thus, we have also constructed fluorescent protein-tagged Topo I and carried out co-expression experiments using live cells. Here we demonstrate that PARP-1 co-localizes with Topo I in the nucleus and that this co-localization is maintained throughout the cell cycle. To confirm that the co-localization is, in fact, because of direct protein-protein interactions, the bioluminescence resonance energy transfer (BRET) assay was employed, and results suggest that PARP-1 directly interacts with Topo I in vivo. These results demonstrate that Topo I is one of the enzymes that indeed interacts with PARP-1 in vivo. Interestingly, in G1 phase cells, PARP-1 and Topo I were localized in the nucleoplasm and nucleoli, whereas both enzymes were concentrated into S-phase foci during the S phase. As Topo I is an enzyme required for both transcription and DNA replication, results from these co-localization experiments suggest that PARP-1 indeed has functions in both transcription and DNA replication. In response to DNA damage, PARP-1 de-localization because of PARP-1 automodification was observed. BRET signal was also reduced by the automodification, suggesting that PARP-1 automodification results in the disruption of the interaction between PARP-1 and Topo I. Because Topo I activity is reported to be promoted by addition of PARP-1 into DNA supercoil relaxation assays (46), we then investigated the effect of disrupting the PARP-1-Topo I interaction on Topo I activity, and we found that the disruption results in a reduction in Topo I activity as compared with Topo I activity measured under conditions allowing the PARP-1-Topo I interaction. Taken together, these results suggest that PARP-1 automodification has a function to regulate the interaction between PARP-1 and Topo I and, consequently, the Topo I activity, in response to DNA damage.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Antibodies, Enzymes, Cells, and Recombinant Proteins—Anti-PARP-1 antibodies, C-II-10 (automodification domain) and F1–23 (second zinc finger), anti-ADP-ribose polymer antibody, LP-96-10, and poly- (ADP-ribose) glycohydrolase were obtained from G. G. Poirier. Anti-Topo I antibody (Sarcoma sera) and purified human Topo I were purchased from TopoGen. Recombinant Topo I was obtained from Sigma. GMO6315A cells were purchased from the NIGMS Human Mutant Cell Repository (Camden, NJ).
Analysis of PARP-1 and Topo I by Western Blotting—COS-7 cells in 10-cm dishes were transfected with 1.6 µg of pPARP-1-EYFP, pPARP-1-DsRed, pTopo I-EYFP, or pTopo I-RLuc, each mixed with 2.4 µg of pMACS KK.II (a mouse H-2KK expression construct) using Effectene (Qiagen). After 48 h, transfected cells co-expressing mouse H-2KK were purified by colloidal superparamagnetic microbeads (MACSelect KK MicroBeads (Miltenyi Biotec)) (47, 48). These purified cells were used for Western blotting with C-II-10, anti-Topo I, or anti-enhanced green fluorescent protein (EGFP)/EYFP (Clontech) antibodies. Alternatively, extracts were prepared from these cells and used for poly(ADP-ribosyl)ation assays.
Cell Extract Preparation—Cells were suspended in 10 mM Tris-HCl, pH 7.8, 0.5 M NaCl, 1 mM dithiothreitol, and an anti-protease mixture. In certain samples, 10 mM EDTA and 50 µM ADP-dihydroxypyrollidine (Calbiochem) were included. After freeze-thawing, cellular debris was spun down and the supernatant was used for poly(ADP-ribosyl)ation, Western blotting, immunoprecipitation, and DNA supercoil relaxation assays.
Poly(ADP-ribosyl)ation Assay—Cell extracts (2 µg/ml) were incubated in 10 mM Tris-HCl, pH 7.5, 5 mM MgCl2, 50 mM NaCl (EX Buffer), containing 1 µg/ml 
X174 DNA-HaeIII digest, 32 nM NAD+, and 10 µCi/ml [32P]NAD+, in the presence or absence of 250 µM NAD+, for 15 min at 37 °C. The reaction mixtures were fractionated on an SDS-7.5% polyacrylamide gel. 32P activity was visualized by autoradiography.
Confocal Microscopy—Cells were seeded in Delta T4 Culture Dishes (Bioptechs) or on 40-mm round coverslips (Bioptechs). pPARP-1-DsRed (1.5 µg/ml) and pTopo I-EYFP (0.5 µg/ml) were transfected by the standard calcium phosphate method. In some cases, pPARP-1-DsRed or pPARP-1-EYFP (2 µg/ml) alone, pPARP-1-DsRed (1.9 µg/ml) and pPCNA-EYFP (0.1 µg/ml), p89 kDa-EYFP (1.0 µg/ml), and p24 kDa-DsRed (1.0 µg/ml), or pPARP-1-EYFP (0.5 µg/ml) and pTopo I-DsRed (1.5 µg/ml) were transfected instead. The cells were cultured for another 48 h. For live cell analyses of less than 2 h, the Delta T4 Culture Dishes were mounted onto the stage of an Olympus IX70 inverted microscope equipped with a Delta T4 Culture Dish Controller (Bioptechs) and kept at 37 °C in HEPES-buffered minimal essential medium, in the presence or absence of 500 µM 1,5-dihydroxyisoquinilone (DHQ). Then, 100 µM N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) was added to the dish. Images acquired through a x60 1.4 NA objective were captured by the FLUOVIEW FV300 confocal scanning unit (Olympus). For multiple-channel imaging, fluorescence from each channel was imaged sequentially to eliminate cross-talk between channels. EYFP fluorescent protein was exited with a 488-nm argon-ion laser line and the fluorescence was imaged using a 570-nm beamsplitter and the combination of a 510–530-nm band pass emission filter and 510-nm long pass emission filter. DsRed fluorescent protein was exited with a 543-nm helium-neon laser line and the fluorescence was imaged using a 570-nm beamsplitter and a 575–630-nm band pass emission filter. In this configuration, cross-talk between EYFP and DsRed was below detectable limits (see Fig. 2F, co-expression experiments with 24- and 89-kDa PARP-1 fragments tagged with DsRed (24 kDa-DsRed) and EYFP (89 kDa-EYFP), respectively). To minimize toxicity potentially introduced by laser lines, the output from argon and helium-neon lasers was attenuated to 10 and 40%, respectively. Excitation of fluorophores by laser could produce reactive oxygen species, which could induce DNA damage. However, the localization of both PARP-1 and Topo I was not affected by laser scanning itself, under similar experimental conditions and scanning frequencies that we used (data not shown). Green, red, and merged color images were created by FLUOVIEW 300 version 3.3 software (Olympus). Then, color contrasts were adjusted using the Photoshop version 6.0 software. For live-cell analyses of up to 24 h and for cell cycle phase determination, the 40-mm coverslips were assembled into an FCS2 live cell chamber (Bioptechs), in which complete Dulbecco's modified Eagle's medium was constantly replenished at a flow rate of 1 ml/h and kept at 37 °C by an FCS2 controller (Bioptechs). Mitotic or G1 cells were selected and images were captured at various time points and processed as described above. To reduce phototoxicity, minimal laser intensities, image capture times, and image resolution were used for this analysis.
|
-D-thiogalactoside. After sonicating the E. coli, the proteins were purified by Ni-NTA-agarose chromatography (Qiagen) (47, 49, 50). Preparation of Automodified PARP-1—Automodified PARP-1-His was prepared following a method described previously (20). Briefly, PARP-1-His was applied to double-strand DNA cellulose resin. PARP-1-His that was retained on the resin was incubated with EX Buffer containing 250 µM NAD+, respectively. Automodified PARP-1-His, released from the resin, was used for assays. To prepare 32P-labeled PARP-1-His and 32P-labeled automodified PARP-1-His, PARP-1-His that was retained on the resin was incubated with EX Buffer containing 32 nM NAD+ and 250 µCi/ml [32P]NAD+. 32P-Labeled PARP-1-His was eluted with EX Buffer containing 1 M NaCl. Alternatively, the resin was incubated with EX Buffer containing 250 µM NAD+ to prepare 32P-labeled automodified PARP-1-His.
Immunoprecipitation—HeLa S3 cell extracts (30 µg of protein) containing 10 mM EDTA and 50 µM ADP-dihydroxypyrollidine were mixed with either anti-Topo I antibody (100-fold dilution) or non-immune sera for 1 h on ice in 50 µl of 10 mM Tris-HCl, pH 7.8, 50 mM NaCl, 1 mM dithiothreitol, 10 mM EDTA, 50 µM ADP-dihydroxypyrollidine, and an anti-protease mixture (IP buffer). Alternatively, purified 89-kDa fragment (300 ng) was mixed with various amounts of Topo I in 50 µl of 10 mM Tris-HCl, pH 7.8, 50 mM NaCl, and 1 mM dithiothreitol at 4 °C for 30 min, and the mixture was incubated with anti-Topo I antibody (100-fold dilution) for 1 h on ice. Protein-G-Sepharose (5 µl) was added to the mixtures containing the cell extracts or purified proteins and incubated for 1 h at 4 °C. After washing the protein G-Sepharose with 100 µl of IP buffer, protein retained on the resin was analyzed by Western blotting with C-II-10, F1–23, or anti-Topo I antibodies.
PARP-1-Topo I Interaction Assay—Topo I (10 units) was mixed with 2 µg of 32P-labeled PARP-1-His or 32P-labeled automodified PARP-1-His in a 100-µl mixture containing 5 µl of Ni-NTA-agarose resin in EX Buffer and incubated for 1 h at 4 °C. The resin was spun down, washed with EX Buffer (10 volumes), and resuspended in 25 µl of EX Buffer containing 250 mM imidazole. The eluates were loaded onto SDS-7.5% polyacrylamide gels. 32P-Labeled PARP-1 and 32P-labeled automodified PARP-1 were visualized by autoradiography. Topo I was analyzed by Western blotting with anti-Topo I antibody.
BRET Assays—The donor construct (pRLuc-N1 or pTopo I-RLuc, 0.5 µg/ml) and acceptor construct (pPARP-1-EYFP (1.5 µg/ml)) were transfected into COS-7 cells by the standard calcium phosphate method. After 48 h, cells were trypsinized, washed, and incubated in HEPES-buffered minimal essential medium, in the presence or absence of 500 µM DHQ, for 20 min at 37 °C. Then, cells were incubated in the presence or absence of 100 µM MNNG for 30 min at 37 °C. The RLuc substrate, coelenterazine h (Molecular Probes), was added at 1 µg/ml prior to sequential emission detection in the 440 to 500 nm (emission 1) and the 510 to 590 nm (emission 2) windows, using the Fusion Universal Microplate Reader (Packard Biosciences). The BRET ratio was defined as [(emission 2) – (emission 1) x C]/(emission 1), where C corresponds to (emission 2)/(emission 1) when the donor construct is transfected alone with pBluescript k/s+ (51).
Cell-free DNA Repair Assay—Plasmid DNA containing 
-ray-induced single-strand DNA breaks was prepared and repair reactions were performed with 300 ng of substrate DNA, 50 µg of cell-free extract prepared from GMO6315A lymphoblastoid cells, and 2 mM NAD+ as described (15). Reactions were terminated by 50 µM ADP-dihydroxypyrollidine and 1 mM EDTA. Then, the mixture was used for DNA supercoil relaxation assays or for Western blotting with C-II-10 or anti-Topo I antibodies.
DNA Relaxation Assay and Analysis of Topo I-DNA Complex by Nicking Assay—Relaxation assays were carried out for 20 min at 37 °C with purified enzymes, or with cell extracts in a 20-µl reaction mixture containing 10 mM Tris-HCl, pH 7.5, 100 mM NaCl, 5 mM EDTA, 50 µg/ml tRNA, and 12.5 µg/ml pBluescript k/s+. The reactions were terminated by addition of 0.5% SDS, 5 mM EDTA, and 250 µg/ml proteinase K. After a 30-min incubation at 37 °C, plasmid DNA was precipitated by ethanol, dissolved in 10 mM Tris-HCl, pH 8.0, and 1 mM EDTA, and applied to a 1% agarose gel for analysis. Plasmid DNA was then stained with 0.5 µg/ml ethidium bromide and visualized by UV.
For the nicking assay, the reaction was carried out in the same reaction conditions as a relaxation assay, in the presence of 250 µM camptothecin, and the reaction was terminated by addition of 0.5% SDS, 5 mM EDTA, and 250 µg/ml proteinase K. After precipitation of DNA, nicked circular plasmid DNA, which was formed by the digestion of Topo I that was covalently complexed with DNA, was separated from closed circular DNA by ethidium bromide, 1% agarose gel electrophoresis. DNA was visualized by UV.
Gel Filtration—Superose 12 HR 10/30 gel filtration column chromatography was carried out using fast protein liquid chromatography. Proteins were eluted with 50 mM Tris-HCl, pH 7.5, 2 mM MgCl2, and either 50 or 100 mM NaCl at a flow rate of 0.4 ml/min. Blue dextran 2000, ferritin (440 kDa), catalase (250 kDa), E. coli topoisomerase I (100 kDa) (50), albumin (65.4 kDa), and ovalbumin (48.9 kDa) were used as molecular mass markers. Fractions were collected at 1-min intervals.
SYPRO Ruby Protein Staining—Protein staining with SYPRO Ruby (Bio-Rad) was carried out following the method of Malone et al. (52). Bovine serum albumin was used as a standard.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
1.5 times higher than that of full-length endogenous Topo I. Localization of PARP-1 and Topo I in Living Cells—To study the localization of PARP-1 and Topo I in living cells, PARP-1-DsRed and Topo I-EYFP were transiently co-expressed in either COS-7 or HeLa S3 cells. On a microscope stage, these cells were maintained in HEPES-buffered minimal essential medium at 37 °C, and fluorescence emissions from PARP-1-DsRed and Topo I-EYFP were sequentially captured. As shown in Fig. 2, the majority of PARP-1-DsRed and Topo I-EYFP were concentrated into nucleoli in COS-7 cells (Fig. 2, A and B) and HeLa S3 cells (Fig. 2E), consistent with previous observations obtained by immunostaining experiments (data not shown and Refs. 54 and 56). As DsRed has a tendency to form aggregates and has a longer maturation time than EYFP, we tested whether the fusion of PARP-1 with DsRed had an effect on the localization of PARP-1-DsRed. PARP-1-EYFP and Topo I-DsRed were co-expressed in HeLa S3 cells instead of PARP-1-DsRed and Topo I-EYFP. Similar to PARP-1-DsRed and Topo I-EYFP, we found both PARP-1-EYFP and Topo I-DsRed to be concentrated into nucleoli (data not shown). Thus, the type of fluorescent proteins fused to PARP-1 or Topo I is unlikely to have any major effect on the localization of these enzymes in living cells.
In the nucleoplasm of interphase cells, PARP-1-DsRed was distributed in a manner similar to Topo I-EYFP (Fig. 2B). To further investigate the localization of PARP-1 and Topo I in living cells, we analyzed COS-7 cells at different phases of the cell cycle. We found that the fluorescence emission from PARP-1-DsRed found in the nucleoli and nucleoplasm of cells in interphase (Fig. 2B) was gradually concentrated into over 100 granules as the cell cycle progressed (Fig. 2C). Within 6 to 10 h, the cells containing these granules entered mitotic division. It has been shown that enzymes required for DNA replication are concentrated into similar granules during S-phase (57). One of these enzyme is PCNA, and both immunostained PCNA and EGFP-tagged PCNA have been shown to be concentrated into replication foci (58, 59). As shown in Fig. 2G, PARP-1-DsRed co-localized with PCNA-EYFP in the granules, suggesting that those granules formed because of the concentration of PARP-1 into replication sites in S-phase cells. During mitotic division, PARP-1-DsRed was found on condensed chromatin (Fig. 2D), consistent with previous observations made from immunostaining experiments (60). Topo I-EYFP was also concentrated into over 100 granules during S-phase (Fig. 2C), and associated with condensed chromatin during mitotic division (Fig. 2D). The localization pattern of Topo I-EYFP (Fig. 2, B to D (Overlay)) was almost identical to that of PARP-1-DsRed, indicating that PARP-1 and Topo I are co-localized throughout the cell cycle.
Localization of PARP-1 in Cells Exposed to DNA Damaging Agents—When cells are exposed to a DNA damaging agent, PARP-1 is automodified. To investigate whether automodification has any effect on the localization of PARP-1 and Topo I in damaged cells, we employed time-lapsed analysis of HeLa S3 cells that co-expressed PARP-1-DsRed and Topo I-EYFP (Fig. 3A). Poly(ADP-ribosyl)ation was induced by treating cells with an alkylating agent, MNNG, which has been shown to induce this modification in vivo (10). Before treatment, both PARP-1-DsRed and Topo I-EYFP were concentrated in nucleoli (Fig. 3A, 0 min). As early as 5 min post-exposure, automodification of PARP-1, showing a reduced mobility on an SDS-polyacrylamide gel, was observed (Fig. 3B), and the fluorescence emission pattern of PARP-1-DsRed had changed (Fig. 3A, 5 min). After 17 min, the PARP-1-DsRed that had been concentrated in the nucleoli at time 0 min diffused away into the nucleoplasm. In contrast, Topo I-EYFP remained concentrated in nucleoli even after 30 min of exposure to MNNG (Fig. 3A). Similar results were also obtained when the experiments were carried out using PARP-1-EYFP and Topo I-DsRed (data not shown). As with HeLa S3 cells, automodification of PARP-1 in COS-7 cells was observed at 5 min post-exposure to MNNG (Fig. 3E). However, the turnover rate of ADP-ribose polymers in COS-7 cells seemed to be slower than in HeLa S3 cells, as the initial automodification rate in COS-7 cells appeared to be less extensive and significantly more automodified PARP-1 was still found after 30 min (Fig. 3E). Possibly reflecting the slower turnover rate of ADP-ribose polymers, de-localization of PARP-1-DsRed from nucleoli to the nucleoplasm in COS-7 cells required longer incubation times compared with HeLa S3 cells (Fig. 3D). When automodification of PARP-1 in HeLa S3 or COS-7 cells was inhibited by DHQ, PARP-1-DsRed remained concentrated in the nucleoli in the presence of MNNG (Fig. 3, C and F). Together, these results suggest that PARP-1 is delocalized as a result of its automodification, which leads to the separation of PARP-1 from Topo I in damaged cells.
|
2-adrenergic receptors (dimers) (51, 61), our results suggest that PARP-1 directly interacts with Topo I in vivo.
|
Interaction of PARP-1 with Topo-1 in Vivo and in Vitro—To confirm results from our time-lapsed studies (Fig. 3) and BRET assays (Fig. 4A), we then carried out immunoprecipitation experiments. Extracts were prepared from control HeLa S3 cells or cells exposed to 100 µM MNNG for 5 min to allow poly(ADP-ribosyl)ation to take place (Fig. 3), and then the extracts were incubated with an anti-Topo I antibody. As shown in Fig. 4B, lanes 2 and 4, while Topo I was immunoprecipitated with protein G-Sepharose from both control and MNNG-treated cells, PARP-1 was only co-precipitated from control cell extracts. To further investigate the effect of poly(ADP-ribosyl)ation, we carried out an analysis with purified histidine-tagged PARP-1 (PARP-1-His) and Topo I. It has been reported that purified PARP-1 can interact with Topo I in vitro (46). Consistent with these observations, we found that purified Topo I was retained on a Ni-NTA-agarose resin when mixed with PARP-1-His (Fig. 4C, lane 2), because of the interaction of PARP-1 with Topo I. We then studied the effect of automodification of PARP-1 on its interaction with Topo I. Instead of unmodified PARP-1-His, automodified PARP-1-His was mixed with Topo I and applied onto a Ni-NTA-agarose resin. As shown in Fig. 4C, lane 3, the automodified PARP-1-His was retained on the resin, whereas Topo I was no longer found in the retained fraction, indicating that Topo I cannot bind to automodified PARP-1. Together, these results suggest that automodification of PARP-1 results in the disruption of PARP-1-Topo I interaction.
Effect of PARP-1 Automodification on Topo I Activity—Our results thus far have indicated that PARP-1 interacts with Topo I in undamaged cells, and that, in response to DNA damage, PARP-1-Topo I interaction is disrupted. We then investigated the effect of this interaction on Topo I activity. Several lines of in vitro evidence have suggested that the Topo I activity to relax DNA supercoils is promoted by addition of purified PARP-1 into a supercoiled DNA relaxation assay (46). Thus, we first have tested the effect of PARP-1 addition to DNA supercoil relaxation assay with Topo I. As we have recently reported the presence of Mg2+-dependent Topo I-like activity in PARP-1 (50), reactions were carried out in the presence of EDTA to suppress the activity and to observe the effect of PARP-1 on Topo I activity. Consistent with reports demonstrating the promotion of Topo I activity by PARP-1, the addition of unmodified PARP-1 increased the relaxation activity of Topo I of supercoiled DNA (Fig. 5A, lanes 5–9). This effect was, however, not observed by the addition of the 24-kDa fragment of PARP-1 (Fig. 5A, lanes 17-21), which contains the DNA binding domain, whereas the 89-kDa fragment, containing the automodification domain, promoted the relaxation activity of Topo I of supercoiled DNA (Fig. 5A, lanes 12–16). Furthermore, it has been reported that a 36-kDa proteolytic fragment of PARP-1 containing part of the automodification domain promotes the activity of Topo I (40). Thus, the interaction between the automodification domain of PARP-1 and Topo I likely promotes the relaxation activity of Topo I of the supercoiled DNA.
|
Topo I and the 89-kDa fragment were then used in immunoprecipitation experiments. However, first, various amounts of Topo I and 89-kDa fragments were quantified by SYPRO Ruby protein staining (Fig. 5B, SYPRO) and Western blotting with C-II-10 on the same amounts of the 89 kDa was also carried out (Fig. 5B, Western). Using these data, we established the relationship between the amount of the 89-kDa fragment and Western blotting signal intensity. Then the 89-kDa fragment (200 ng) was incubated with various amounts (up to 25 ng) of Topo I and precipitated with anti-Topo I antibody (100-fold dilution, which could precipitate over 95% of the Topo I in the mixtures). The precipitates were used for Western blotting with C-II-10 to visualize the 89-kDa fragment that precipitated with Topo I (IP-Western). Then, the precipitated amount of the 89-kDa fragment (Fig. 5B, IP-Western) was determined from the above relationship and the average molar ratio of the 89-kDa fragment to Topo I was calculated (using 70 kDa as molecular mass of Topo I (TopoGen)) to be 1.3 (Fig. 5B). These results suggest that Topo I interacts with PARP-1 at an 1 to 1 ratio in vitro. As both Topo I and PARP-1 interact also with other factors, an equimolar amount of PARP-1 and Topo I may be present as components of a high molecular weight protein complex in vivo, e.g. DNA replication complex (62).
Camptothecin, a Topo I inhibitor, stabilizes Topo I-DNA covalent complexes formed after incision of DNA by Topo I (63). Proteolytic digestion of Topo I covalently linked to plasmid DNA results in the production of nicked circular DNA (open circular form) (63). As shown in Fig. 5C, lane 4, 8 units of Topo I produced nicked circular DNA in the presence of camptothecin, whereas only a negligible amount of nicked circular DNA was produced by 2 units of Topo I (lane 6). On the other hand, addition of PARP-1 to the reaction mixtures with 2 units of Topo I in the presence of camptothecin increased the formation of nicked circular DNA (Fig. 5C, lane 8), suggesting that PARP-1 promoted the formation of Topo I-DNA covalent complexes. Increased binding of Topo I to DNA in the presence of camptothecin has also been reported using Topo I-DNA covalent binding assays (46). Thus taken together, promotion of Topo I activity by PARP-1 occurs by increasing the binding efficiency of Topo I to its substrate DNA, possibly because of structural alterations of Topo I.
Because our data suggest a disruption of PARP-1-Topo I interaction after automodification of PARP-1 (Fig. 4), we then carried out DNA supercoil relaxation assays with automodified PARP-1 and Topo I. As shown in Fig. 5A, we found that automodified PARP-1 did not promote the activity of Topo I to relax supercoiled DNA (lane 10), whereas automodified PARP-1 treated with poly(ADP-ribose) glycohydrolase (lane 11), which degrades the polymers on automodified PARP-1 (64), promoted the relaxation of supercoiled DNA. Thus, these data suggest that the effect of PARP-1 to promote the activity of Topo-1 to relax supercoiled DNA is disrupted by automodification of PARP-1.
To obtain more conclusive evidence for the roles of PARP-1-Topo I interaction and the disruption of this interaction in the relaxation of supercoiled DNA, we used an in vitro DNA repair assay with cell-free extracts, in which the extent of poly(ADP-ribosyl)ation can be controlled. Previously, we demonstrated that unmodified PARP-1 is converted into the automodified form in response to DNA break formation (16). In this assay system, the polymer content is transiently increased and reaches maximal levels during the first 15 min of incubation (16). Thus, a DNA repair reaction mixture (15) that was incubated for 15 min was used for the relaxation assay of supercoiled DNA. When PARP-1 was not automodified, we found that the DNA repair reaction mixture had the expected activity to relax supercoiled DNA (Fig. 5D, lanes 2, 4, and 5). In the presence of NAD+, PARP-1 was converted into the automodified form (less PARP-1 was recognized by anti-PARP-1 antibody because of a reduced mobility of automodified PARP-1 on an SDS-polyacrylamide gel and a reduced affinity of anti-PARP-1 antibody for automodified PARP-1 (53)) (Fig. 5D, PARP-1, lane 3). In this DNA repair assay mixture, the activity to relax supercoiled DNA was significantly reduced (Fig. 5D, Relaxation, lane 3). However, neither the level of Topo I detected by anti-Topo I antibody nor any obvious molecular weight shift of Topo I was found in this mixture (Fig. 5D, Topo I, lane 3). In support for this observation, an anti-ADP-ribose polymer antibody did not recognize any proteins having a similar mobility to Topo I on an SDS-polyacrylamide gel (data not shown). Therefore, based on these results, we conclude that PARP-1-Topo I interaction has a promoting effect on the activity of Topo I to relax DNA supercoils, and that the disruption of this interaction by automodification of PARP-1 leads to the reduction of the activity.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
PARP-1, Topo I, and Other Factors Reported to Interact with PARP-1—Topo I has been purified to near homogeneity and PARP-1 has never been reported to be found in the purified fractions. PARP-1 has also been purified to over 99% homogeneity, suggesting that PARP-1 can be separated from other proteins and enzymes, although some Topo I activity is still found in highly purified fractions (39). Thus, the PARP-1-Topo I interaction that we reported here is likely not stable under conventional column chromatography conditions, possibly because of high salt concentrations, and could only be observed under physiological conditions. Previously, using live cells under physiological conditions, Christensen et al. (54) suggested that Topo I interacts with nuclear components, as the diffusion rate of Topo I in the nucleoplasm is slower than the rate of free diffusion. One such component could be PARP-1, which is in agreement with the co-localization of PARP-1 and Topo I in live cells (Fig. 2). However, the interaction between PARP-1 and Topo I likely does not lead to a stable complex formation, as Topo I is still mobile in live cell nuclei (54). Therefore, at the molecular level, PARP-1 and Topo I possibly only form complexes with a short half-life through the automodification domain of PARP-1.
PARP-1 may also form short half-life complexes with other factors and enzymes, as PARP-1 can be separated from these factors under conventional column chromatography conditions using high salt concentrations. In the reports demonstrating co-precipitation of PARP-1 with other factors and enzymes, in fact, 50–100 mM salt concentrations were used at 0–4 °C, which is a mild condition that likely does not disrupt relatively weak protein-protein interactions. However, if PARP-1 only forms complexes with a short half-life with various factors and enzymes, including highly abundant enzymes such as Topo I, how PARP-1 is capable of complexing with over dozens of these factors and enzymes in vivo could be explained; i.e. the domain of PARP-1 that binds to other factors and enzymes is not permanently occupied by certain factors and enzymes, and thus PARP-1 can interact with all of them.
As shown in Fig. 2F, the domain of PARP-1 that contains the zinc fingers (24-kDa fragment) is sufficient for the subnuclear localization of PARP-1, because the zinc fingers have a high affinity for DNA loops and RNA (20, 65), whereas the automodification domain, which is located next to the zinc fingers, is responsible for the interactions between PARP-1 and certain factors and enzymes. Thus, perhaps PARP-1 acts as a "universal molecular adapter" that concentrates itself into actively transcribed or replicating DNA regions through binding of the zinc finger motifs to DNA loops or nascent RNA, and recruits various factors and enzymes to these regions by forming complexes of short half-life with these factors and enzymes.
Transcription and DNA Replication—Multiple factors and enzymes interact with PARP-1, including transcription initiation factors (6, 33) and DNA polymerase (34, 35). Based on the interaction between PARP-1 and these factors and enzymes, it has been proposed that PARP-1 has a function in transcription initiation and DNA replication, although the function of PARP-1 in these distinct cellular processes is not yet clear. As shown in Fig. 2B, PARP-1 showed distinctive localization patterns between the G1 phase and S phase cells. In G1 phase cells, PARP-1 co-localized with Topo I, and about 80% of these enzymes were concentrated into nucleoli, whereas the remaining 20% were in the nucleoplasm (Fig. 2B). As Topo I is an enzyme required for transcription (42, 45), co-localization and interaction between PARP-1 and Topo I suggest that PARP-1 also has a role in transcription by RNA polymerase I in nucleoli and by RNA polymerase II in the nucleoplasm. In relation to transcription, we previously proposed that PARP-1 plays a role in transcription regulation by binding to nascent RNA (20, 49). Thus, the interaction between PARP-1 and Topo I, which is an enzyme required to remove DNA superhelical tension produced during the progression of transcribing RNA polymerases (42, 45), suggests that PARP-1 also plays a role in the regulation of RNA synthesis at the level of elongation, in addition to transcription initiation as proposed by others. In S phase cells, a role for PARP-1 in DNA replication is supported by the observation that PARP-1 and Topo I are concentrated into S phase foci (Fig. 2C), which are the cellular centers for DNA replication (57). In conclusion, the localization of PARP-1 appears to be cell cycle dependent, and PARP-1 may have a role in transcription during G1 phase, and in DNA replication during S phase.
Another group of factors and enzymes that interact with PARP-1 is DNA repair enzymes, including XRCC-1, DNA ligase III, and Werner syndrome protein (19, 36–38). Although these DNA repair factors appear to have functions in DNA repair or DNA damage responses, they are also present in non-damaged cells. Thus, if these factors and enzymes form complexes with PARP-1 in vivo, they may also be concentrated into actively transcribed and replicating DNA regions by PARP-1. As DNA damage formation on actively transcribed or replicating regions would have significant effects on cell viability, one possible functional relevance of the interactions between PARP-1 and DNA repair factors could be to pre-concentrate these factors and enzymes to actively transcribed and replicating regions to efficiently repair DNA damages produced in these regions.
PARP-1 and Topo I Activity—Previously, Bauer et al. (50) reported that addition of PARP-1 purified from calf thymus into supercoiled DNA relaxation assays with Topo I results in the promotion of Topo I activity (46). As we recently reported the association of a Mg2+-dependent Topo I-like activity in PARP-1 (50), we initially considered that the promotion reported by Bauer et al. (50) was because of the additive effect of Topo I activity and Topo I-like activity of PARP-1 (see "Discussion" in Ref. 50 for the relationship between Topo I-like activity of PARP-1 and Topo I). However, we also found that PARP-1 promoted Topo I activity (Fig. 5A) even in the presence of EDTA, which inhibits the Topo I-like activity of PARP-1 (50). Thus, this promotion is not only because of the simple additive effect of Topo I-like activity and Topo I activity. Furthermore, in the presence of camptothecin, PARP-1 promoted the formation of Topo I-DNA covalent complexes (Fig. 5C) despite the fact that Topo I-like activity does not result in PARP-1-DNA covalent complex formation (50). Thus, taken together, we concluded that PARP-1 can indeed promote Topo I activity.
Several previous reports have already suggested a link between PARP-1 and Topo I. Jongstra-Bilen et al. (39) in fact, had reported the presence of Topo I activity in highly purified PARP-1 fractions, and activation of PARP-1 resulted in reduction of supercoiled DNA relaxation activity present in these purified PARP-1 fractions. Boothman et al. (66) later reported a reduction in DNA supercoil relaxation activity in extracts prepared from cells exposed to DNA damaging agents. Whereas Jongstra-Bilen et al. (39) and Boothman et al. (66) concluded that the reduction was because of the poly(ADP-ribosyl)ation of Topo I, no direct evidence demonstrating poly(ADP-ribosyl)ation of Topo I was presented (39, 66). We were also unable to obtain any direct evidence for the modification of Topo I in extracts used in our experiments (Fig. 5D). In vitro, Topo I can be poly(ADP-ribosyl)ated (40, 67). Thus, while our results do not exclude the possibility of Topo I modification by ADP-ribose polymers, the reduction in Topo I activity can occur due to automodification of PARP-1 without poly(ADP-ribosyl)ation of Topo I, because we observed the reduction in Topo I activity by incubating Topo I with automodified PARP-1, and the promoting effect was restored by addition of poly(ADP-ribose) glycohydrolase (Fig. 5A, lanes 10 and 11). In summary, one of the functional roles of the interaction between PARP-1 and Topo I could be to control Topo I activity in response to DNA damage.
Protein-Protein Interaction and Poly(ADP-ribosyl)ation—In cells treated with MNNG, which activates PARP-1 in vivo (Fig. 3), PARP-1 no longer co-localized with Topo I. In addition, the BRET signal for the PARP-1-Topo I interaction was reduced in cells treated with MNNG (Fig. 4A), suggesting that the intermolecular distance between PARP-1 and Topo I is increased following PARP-1 automodification. Furthermore, Topo I was not co-immunoprecipitated from extracts prepared from cells exposed to MNNG (Fig. 4B), and also automodified PARP-1 did not interact with purified Topo I (Fig. 4C). Thus, these results suggest that PARP-1 automodification results in the disruption of the interaction between PARP-1 and Topo I both in vivo and in vitro. Previously, Chang and Alvarez-Gonzalez (26) also suggested that automodification of PARP-1 affects its interaction with NF-B, based on observations obtained mainly from gel retardation assays. They showed that NF-B is unable to bind to its target sequence in the presence of PARP-1 because of PARP-1-NF-B interaction, while, when PARP-1 is automodified, NF-B is released from PARP-1 and is allowed to bind to its target sequence (26). Thus, they proposed that PARP-1 has a negative effect on NF-B-dependent gene expression. On the other hand, Hassa et al. (25) demonstrated that the interaction between PARP-1 and NF-B is required for NF-B-dependent gene expression. The reason for this discrepancy is not yet clear. Another example demonstrating the effect of automodification of PARP-1 on its interactions with other factors and enzymes has recently been reported by von Kobbe et al. (38). They showed that Werner syndrome protein interacts with PARP-1, whereas automodified PARP-1 does not bind to purified Werner syndrome protein in in vitro binding assays (38). Interestingly, overexpression of a Werner syndrome protein fragment that is involved in PARP-1 binding inhibits ADP-ribose polymer formation (38). Thus, von Kobbe et al. (38) proposed that PARP-1 automodification is required for the regulation of the DNA damage response mediated by Werner syndrome protein. Similar to Werner syndrome protein, Masson et al. (37) reported that the binding of XRCC-1 to PARP-1 results in inhibition of PARP-1 automodification, whereas the effect of the automodification on the interaction between PARP-1 and XRCC-1 is not known. Taken together, these observations and our results obtained using live cell analysis suggest that one functional relevance of PARP-1 automodification is to regulate the interactions between PARP-1 and various factors and enzymes in response to DNA damage.
Base Excision Repair and PARP-1—Previously, we demonstrated that DNA breaks that can trigger poly(ADP-ribosyl)ation are produced by enzymes of the base excision repair pathway (16), which repairs modified DNA bases induced by reactive oxygen species or alkylating agents (68). Therefore, poly(ADP-ribosyl)ation is a post-translational protein modification that likely occurs as part of the base excision repair pathway. As factors and enzymes that interact with PARP-1, including Topo I, play roles in transcription and DNA replication, one of the functions of poly(ADP-ribosyl)ation is, perhaps, to link base excision repair to transcription and DNA replication to transduce a DNA damage signal from the base excision repair pathway to these key cellular processes, leading to efficient cellular responses to DNA damage.
|
FOOTNOTES |
|---|
Present address: Graduate School of Bioscience and Biotechnology, Tokyo Institute of Technology, 4259 Nagatsuta-cho, Midori-ku, Yokohama 226-8503, Japan.
|| Supported by the Fonds de la Recherche en Santé du Québec.
** Supported by the National Cancer Institute of Canada and the Canadian Institutes of Health Research. To whom correspondence should be addressed. Tel.: 418-656-4141 (ext. 47340); Fax: 418-654-2159; E-mail: Masahiko.sato{at}crchul.ulaval.ca.
1 The abbreviations used are: PARP-1, poly(ADP-ribose) polymerase-1; Topo I, topoisomerase I; BRET, bioluminescence resonance energy transfer; PCNA, proliferating cell nuclear antigen; EYFP, enhanced yellow fluorescent protein; RLuc, Renilla luciferase; DsRed, red fluorescent protein; DHQ, 1,5-dihydroxyisoquinilone; MNNG, N-methyl-N'-nitro-N-nitrosoguanidine; EGFP, enhanced green fluorescent protein; Ni-NTA, nickel-nitrilotriacetic acid.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
