A Topoisomerase IIß-Mediated dsDNA Break Required for Regulated Transcription

See allHide authors and affiliations

Science  23 Jun 2006:
Vol. 312, Issue 5781, pp. 1798-1802
DOI: 10.1126/science.1127196


Multiple enzymatic activities are required for transcriptional initiation. The enzyme DNA topoisomerase II associates with gene promoter regions and can generate breaks in double-stranded DNA (dsDNA). Therefore, it is of interest to know whether this enzyme is critical for regulated gene activation. We report that the signal-dependent activation of gene transcription by nuclear receptors and other classes of DNA binding transcription factors, including activating protein 1, requires DNA topoisomerase IIβ-dependent, transient, site-specific dsDNA break formation. Subsequent to the break, poly(adenosine diphosphate–ribose) polymerase–1 enzymatic activity is induced, which is required for a nucleosome-specific histone H1–high-mobility group B exchange event and for local changes of chromatin architecture. Our data mechanistically link DNA topoisomerase IIβ–dependent dsDNA breaks and the components of the DNA damage and repair machinery in regulated gene transcription.

Regulated activation of gene transcription requires a network of sequentially exchanged cofactor complexes (13). Recently, a signal-inducible exchange of coregulator complex and poly[adenosine diphosphate (ADP)–ribose] polymerase–1 (PARP-1), a nicotinamide adenine dinucleotide (NAD+)–dependent enzyme that detects and repairs damage to DNA, has been linked to these multistep events for regulated activation of gene transcription (47). Therefore, we explored whether the components of the DNA damage and repair apparatus and the enzymatic activity of DNA topoisomerase IIβ (TopoIIβ) might be critical to the multistep events required for regulated gene expression, in response to ligand- or signal-dependent stimuli.

Using chromatin immunoprecipitation (ChIP) analysis, we examined the time course of cofactor exchange on the pS2 promoter in 17β-estradiol (E2)–treated Michigan Cancer Foundation (MCF)–7 cells. We observed basal levels of PARP-1 and TopoIIβ along with nuclear receptor corepressor (N-CoR) and histone deacetylase 3 (HDAC3) components of the corepressor complex but no detectable changes in that of the adenosine 3′,5′-monophosphate response element–binding protein (CBP) coactivator and RNA polymerase II (Pol II) in the absence of the ligand (Fig. 1A and fig. S1A). Previously identified components of the PARP-1 corepressor complex, including nucleolin, nucleophosmin, and human heat shock protein 70 (HSP70) (7), were present on the pS2 promoter in unstimulated cells but were rapidly depleted in E2-treated cells (Fig. 1A). There was a rapid increase of TopoIIβ and PARP-1 recruitment together with the CBP coactivator and subsequent recruitment of Pol II and the elimination of N-CoR and HDAC3 corepressors in response to E2 (Fig. 1A and fig. S1A). PARP-1 was detected with the components of DNA damage and repair machinery (8, 9), and Western blot analyses revealed that TopoIIβ, DNA-dependent protein kinase (DNA-PK), and Ku86 and Ku70 were copurified with PARP-1 (fig. S1B). DNA-PK is a nuclear serine-threonine protein kinase that forms a complex with the regulatory DNA binding subunits Ku86 and Ku70 in DNA damage and repair (9). After E2 treatment, all components of the TopoIIβ/DNA-PK/Ku86/Ku70 complex were stably recruited to the pS2 promoter but not to the coding region of the pS2 gene (10) (Fig. 1B and fig. S1C). No increased recruitment of TopoIIα was detected on promoter region (Fig. 1B). After an initial ChIP with an antibody to TopoIIβ, we performed a second step of ChIP analysis to confirm the mutual presence of all of the components of this TopoIIβ/PARP-1 complex on the pS2 promoter in E2-treated MCF-7 cells (Fig. 1C).

Fig. 1.

TopoIIβ/PARP-1 complex–mediated transient DNA break formation is required for E2-stimulated gene activation. (A) E2-dependent dynamic occupancy of TopoIIβ, PARP-1, and the coregulators on the pS2 promoter. Immunoglobulin G (IgG) was used for negative control. Input, 5% total. (B) E2-dependent occupancy of TopoIIβ, DNA-PK, and Ku86 and Ku70 on the promoter of the pS2 gene. The antibody to x-ray repair cross-complementing group (XRCC-1) was used for negative control. (C) Corecruitment of PARP-1, DNA-PK, and Ku86 and Ku70 in E2-treated MCF-7 cells. (D) Detection of biotin incorporation (DNA break) in the promoter or coding region of the pS2 gene in E2-treated MCF-7 cells. ORF, open reading frame. (E) Time course DNA break–labeling ChIP analysis.

We next tested the hypothesis that the enzymatic activity of TopoIIβ that alters the topology of the DNA (11) might be a required component for signal-dependent activation of gene transcription. We developed a protocol that detects DNA break formation in the promoter region using a combination of biotin-11–deoxyuridine triphosphate (dUTP) labeling by terminal deoxynucleotide transferase (TdT) and subsequent ChIP analysis. In MCF-7 cells stimulated with E2 for 10 min, we detected putative DNA breaks in the pS2 promoter but no signal in the coding region of the pS2 gene (Fig. 1D). In contrast to the strong signal in the pS2 promoter, no increased enrichment of biotin incorporation was observed either at the promoter region of a gene encoding the ribosomal protein L13a, which is expressed in MCF-7 cells in an E2-independent fashion, or at the promoter of the neuronal-specific sodium channel II (NaChII), which is not transcribed in MCF-7 cells (fig. S1D). We did not detect breaks in the DNA on the pS2 promoter in the E2 antagonist, 4-hydroxytamoxifen (4-OHT)–treated MCF-7 cells (fig. S1E).

Evidence that the regulated DNA breaks occur only in a transient fashion in the pS2 promoter region was obtained by performing a time course of DNA break–labeling ChIP assay in E2-stimulated MCF-7 cells, after the release from α-amanitin treatment to achieve transcriptional synchronization of cells (Fig. 1E). These results suggest that transcriptional activation can be triggered, at least in part, by the same molecular machinery responsible for the relief of torsional tension on the DNA by creating transient double-stranded breaks in response to signals for transcriptional gene activation.

Conversely, a two-step ChIP assay using an antibody against biotin as a mark of DNA breaks revealed the presence of the TopoIIβ/PARP-1 complex in the vicinity of the DNA break formation within the pS2 promoter region (Fig. 2A). These data suggest that specific components of the DNA repair machinery may also participate in nuclear receptor–dependent transcriptional activation in the vicinity of DNA break sites upon ligand treatment. This finding is consistent with several reports of the presence of Ku86 and Ku70 subunits and DNA-PK on the Pol II transcription units (1215).

Fig. 2.

Functional roles of enzymatic activity of TopoIIβ on E2-stimulated gene activation. (A) Corecruitment of the TopoIIβ/PARP-1 complex at DNA break sites on the pS2 gene promoter. (B) Inhibition of TopoII activity by 100 μM Mer blocked DNA break formation (top section) and blocked E2-stimulated pS2 transcription (bottom section). GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HPRT1, hypoxanthine-guanine phosphoribosyltransferase 1. (C) Mer inhibits E2-dependent recruitments of CBP coactivator and Pol II on the pS2 gene promoter. (D) Requirement of enzymatic activity of TopoIIβ for ERα reporter activation. β-Gal, β-galactosidase; wt, wild type; mut, mutant. Error bars indicate SEM.

To test whether TopoII is mediating the formation of DNA breaks in the pS2 promoter region, we treated cells with the TopoII inhibitor merbarone (Mer) (5-N-phenylcarboxamido-2-thiobarbituric acid), which inhibits the cleavage activity of TopoII without damaging the DNA or the stabilizing DNA-TopoII cleavable complexes. Treatment with Mer prevented the DNA breaks on the stimulated pS2 promoter and blocked pS2 transcriptional activation (Fig. 2B). These results are consistent with the observation that the inhibitor of TopoII ultimately blocked recruitment of coactivators such as CBP and Pol II in E2-treated MCF-7 cells (Fig. 2C). To further clarify the requirement of enzyme activity of TopoIIβ, we performed transient transfection assays using enzymatically inactive mutant mouse TopoIIβ (Pro723 → Leu723) (16) with human-specific TopoIIβ small interfering RNA (siRNA) (Fig. 2D and fig. S2A). Consistent with the results of the specific inhibitor, enzymatic activity of TopoIIβ was required for transcriptional activation in response to ligand treatment.

To further examine the mechanisms of TopoIIβ/PARP-1–dependent pS2 gene activation events, we refined our analysis to the level of single nucleosomes by examining the area of the surrounding estrogen response element (ERE) present in the pS2 promoter. Chromatin was digested to mononucleosomes and examined by ChIP analysis (nucleosome-ChIP assay). Occupancy on individual nucleosomes was assessed by means of nucleosome-specific polymerase chain reaction (PCR) primers (17). After treatment with E2, PARP-1 localized to the nucleosome containing the cognate estrogen receptor binding site (NucE) of the pS2 promoter, whereas the adjacent nucleosomes (NucU and NucT) exhibited a decrease of PARP-1 occupancy (Fig. 3A). In contrast, TATA-binding protein (TBP) was associated with the TATA box–containing NucT (Fig. 3A). Therefore, PARP-1 binds weakly to all three nucleosomes as a corepressor complex in the unstimulated pS2 promoter region but, after treatment with E2, becomes localized to NucE alone. Similar results were found for TopoIIβ localization (fig. S3A).

Fig. 3.

Nucleosome-specific recruitment of the TopoIIβ/PARP-1 complex to the pS2 promoter. (A) Decreased recruitment in PARP-1 on NucU and NucT but increased recruitment on NucE in response to E2. (B) Antibody to ERα immunoprecipitated material from E2-treated MCF-7 cells as assessed by Western blot analysis. (C) NucE-selective recruitment of ASC2 coactivator in E2-treated MCF-7 cells. (D) Identification of TopoII-mediated DNA break site (arrows) upon E2 stimulation in MCF-7 cells. Dideoxy-terminated sequence ladders were generated with the same primers. (E) Proportion of the pS2 promoter that contains dsDNA breaks in E2-treated MCF-7 cells. qPCR, quantitative real-time PCR. Error bars indicate SEM.

These data suggest that recruitment of the TopoIIβ/PARP-1 complex in the NucE might reflect the actions of a specific coactivator complex recruited to the liganded receptor. Indeed, PARP-1, DNA-PK, and Ku86 and Ku70 have been noted to be able to bind to the C terminus of an Leu-X-X-Leu-Leu–containing thyroid hormone receptor-binding coactivator, alternatively referred to as protein-activating signal coregulator 2 (ASC2) (18). ASC2 is also capable of interacting with many liganded nuclear receptors and other classes of transcription factors (18, 19). ASC2, TopoIIβ, PARP-1, DNA-PK, and Ku86 and Ku70 coimmunoprecipitate from the total extract of E2-treated MCF-7 cells with an antibody to estrogen receptor α (ERα) (Fig. 3B). Consistent with this observation, the nucleosome-ChIP assay revealed selective recruitment of ASC2 to NucE (Fig. 3C). siRNA-mediated reduction of ASC2 abrogated the E2-dependent activation of ERα reporter activity (fig. S3B). Overexpression of an activation function 2 helix-deleted ERα mutant that cannot bind Leu-X-X-Leu-Leu motif–containing coactivators (20) similarly blocked E2-dependent DNA break formation of the pS2 promoter in HeLa cells (fig. S3C). Taken together, these data indicate a nucleosome-specific recruitment of a TopoIIβ/PARP-1 complex to the pS2 promoter by liganded ERα and its associated coactivators. In addition, we observed reduced recruitment of TopoIIβ/ASC2 in 4-OHT–treated MCF-7 cells (fig. S3D), consistent with a failure of TopoII-mediated DNA break formation in response to 4-OHT (fig. S1E).

To identify endogenous TopoII-mediated DNA cleavage sites on the pS2 promoter region upon E2 treatment, we stabilized the transient covalent TopoII-DNA intermediate (a cleavage complex) by using etoposide (VP-16), which is capable of blocking TopoII ligase activity, and sequenced the cleavage sites by primer extension of genomic DNA (21). This strategy revealed that the position of the DNA break within pS2 promoter was centered at an adenine-thymine–rich linker region between NucU and NucE; no extension products were detected in unstimulated MCF-7 cells (Fig. 3D). These data suggest that TopoII is required for transient, site-specific dsDNA break formation during nuclear receptor-mediated transcriptional activation. To investigate the proportion of pS2 promoters in E2-stimulated MCF-7 cells that harbored the specific dsDNA break, we stabilized the transient covalent TopoII-DNA intermediate by using VP-16 and performed PCR using specific primer pairs around the cleavage site. The quantitative PCR data showed that >50% of the pS2 promoters exhibited the specific dsDNA break in response to E2 (Fig. 3E).

Despite the several lines of evidence indicating that PARP-1 may exert its function in transcriptional regulation, the molecular determinants for PARP-1 recruitment and function in the absence of genotoxic stress have remained obscure (47, 2226). To further define the potential molecular mechanisms by which PARP-1 may function as a coregulator for nuclear receptors, we measured the pS2 transcript. The presence of PARP inhibitors (3-AB or PJ-34) blocked ERα-dependent gene activation (Fig. 4A and fig. S4, A and B). Single-cell microinjection of either an antibody to PARP-1 or specific PARP-1 siRNA (fig. S2B) similarly blocked ERα-dependent reporter gene activation (fig. S4C). Inhibition of ERα-dependent gene expression by human PARP-1 siRNA could be efficiently restored by expression of wild-type murine PARP-1 (fig. S4D). However, a catalytic mutant (Glu988 → Ala988) of PARP-1 did not rescue the activation of ERα-dependent reporter (fig. S4D). These data confirm that the catalytic activity of PARP-1 is required for transcriptional activation by the ER.

Fig. 4.

TopoIIβ/PARP-1–dependent exchange of histone H1 for HMGB 1 on NucE in E2-treated MCF-7 cells. (A) PARP inhibitor (3-AB or PJ-34) blocked activation of pS2 transcription. (B) Loss of histone H1 in E2-treated MCF-7 cells on NucE. (C) E2-dependent recruitment of HMGB1 on NucE. (D) Inhibition of E2-dependent exchange of histone H1 for HMGB1 on NucE by 3-AB– or Mer-treated MCF-7 cells.

Because PARP-1 can use histone H1 as a substrate for poly(ADP-ribosyl)ation (6, 27) and because histone H1 is widely viewed as a repressor of transcription (28), we investigated whether the poly(ADP-ribosyl)ation activity of PARP-1 regulates histone H1 modification in a nucleosome-specific fashion during ligand-dependent transcriptional activation. A nucleosome-ChIP analysis indicated that histone H1 dismissal may occur locally on the single nucleosome containing ERE upon E2 treatment, whereas the two adjacent nucleosomes did not exhibit a decrease of histone H1 occupancy (Fig. 4B). In contrast to dismissal of histone H1, ChIP analysis revealed that high mobility group B 1/2 (HMGB1/2) (formerly known as HMG1/2) was recruited on the NucE after E2 treatment, suggesting that histone H1 is replaced by HMGB1/2 upon E2 stimulation (Fig. 4C and fig. S4E). These results suggest that the replacement of histone H1 by HMGB1/2 on the single ERE-containing nucleosome may be part of the ERα-mediated transcriptional activation of the pS2 gene, because HMGB1/2 has been viewed as an activator (29), and may even stabilize the binding of ERα to cognate DNA sites (30, 31). Inhibitors of either PARP-1 (3-AB) or TopoII (Mer) enzymatic activity blocked exchange of histone H1 for HMGB1/2 on NucE, which is consistent with suppression of the pS2 gene expression (Fig. 4D and fig. S4F).

Events analogous to those observed for ERα-dependent promoters were also observed for androgen receptor (AR), retinoic acid receptor (RAR), thyroid receptor (T3R), and activating protein 1 (AP-1)–dependent transcriptional activation, all of which recruited the TopoIIβ/PARP-1 complex that caused the DNA break–dependent activation of PARP-1 and histone H1-HMGB exchange (fig. S5). Consistent with suggestions that the TopoIIβ-dependent DNA break represented a widespread strategy for regulated gene transcription, DNA break–labeling ChIP assays in AR-, RAR-, T3R-, and AP-1–dependent transcription events revealed increased incorporation of biotin in their promoters in response to the signal and ligand (Fig. 5A).

Fig. 5.

The general roles of the TopoIIβ/PARP-1 complex in nuclear receptor-mediated gene regulation. (A) TopoIIβ-mediated transient dsDNA breaks occurred on endogenous promoters of AR [prostate-specific antigen (PSA)], RAR (RARβ), T3R (Dio1), and AP-1 [matrix metalloprotease 12 (MMP12)]. (B) TPA-dependent activation of AP-1-mediated pS2 gene expression in MCF-7 cells (left). TPA-induced DNA break formation was observed on pS2 promoter (middle) but not on the silent NaChII promoter (right). (C) Molecular mechanism of TopoIIβ/PARP-1–dependent regulated gene transcription. DHT, dihydrotestosterone.

To address whether identical molecular events operated in response to different signaling pathways targeting the same nucleosome, we investigated the actions of AP-1, which binds to a site that is also located in NucE (32, 33). In addition to E2, AP-1 can mediate pS2 gene activation in 12-O-tetradecanoylphorbol 13-acetate (TPA)–stimulated MCF-7 cells (Fig. 5B). Analogous to the effects imposed by estrogen treatment, TPA-treated MCF-7 cells also exhibited a transient DNA break in the pS2 promoter region (Fig. 5B).

Taken together, these data suggest that the transient TopoIIβ-mediated dsDNA break formation creates the signal that, either directly or indirectly, results in the activation of a PARP-1 enzymatic function, underlying a nucleosome-specific histone H1-HMGB exchange and is likely to serve as a general mechanism for regulated initiation of gene transcription upon ligand- or signal-dependent stimulation (Fig. 5C).

The TopoIIβ-containing complex functions as an additional component of the cascade of coactivator complexes required for regulated gene transcription. Thus, a component of the genome-wide DNA-damage surveillance machinery also serves in the multistep process of regulated, gene-specific transcription, functioning to facilitate dynamic changes in chromatin organization in a localized fashion with the precision of a single nucleosome. Our data are consistent with the report that TopoII inhibitor treatment caused notable changes at or near promoters, whereas TopoI inhibitors caused transcription complexes to stall in the midst of transcription units (34).

Collectively, our data reveal that a transient dsDNA break occurs at multiple regulated transcription units. This raises questions regarding the interplay between molecular machineries that are involved in the repair of dsDNA breaks and the activation of the gene transcription.

Supporting Online Material

Materials and Methods

Figs. S1 to S5


References and Notes

View Abstract

Navigate This Article